play the french 3rd edition
by John Watson EVERYMAN CHESS Gloucester Publishers pic www.everymanchess.com
First published in 2003 by Gloucester Publishers pIc (formerly Everyman Pub lishers pIc), Gloucester Mansions, 1 40A Shaftesbury Avenue, London WC2H 8HD Copyright © 2003 John Watson The right of John Watson to be identified as the author of this work has been asserted in accordance with the Copyrights, Designs and Patents Act 1 988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, elec trostatic, magnetic tape, photocopying, recording or otherwise, without prior permission of the publisher. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library.
ISBN 1 85744 337 3 Distributed in North America by The Globe Pequot Press, P.O Box 480, 246 Goose Lane, Guilford, CT 06437-0480. All other sales enquiries should be directed to Everyman Chess, Gloucester Publishers pIc, Gloucester Mansions, 1 40A Shaftesbury Avenue, London WC2H 8HD (tel: 020 7539 7600 fax: 020 7379 4060) email:
[email protected]; website: www. everymanchess.com Everyman is the registered trade mark of Random House Inc . and is used in this work under license from Random House Inc .
EVERYMAN CHESS SERIES (formerly Cadogan Chess)
Chief advisor: Garry Kasparov Commissioning editor: Byron Jacobs Typeset and edited by First Rank Publishing, Brighton. Cover design by Horatio Monteverde. Production by Navigator Guides. Printed i n the United States by Versa Press.
Contents
Bibliography
4
Introduction
5
1
Advance Variation: Introduction
7
2
Advance Variation: 5 . . . .i.d7
19
3
Advance Variation: 5 . . . Wb6
36
4
King's Indian Attack
55
5
Exchange Variation
70
6
Tarrasch Variation: Introduction and 3 . . . c5
81
7
Tarrasch Variation: 3 . . . .i.e7
1 01
8
Winawer Variation: Fourth Move Alternatives
121
9
Winawer Variation: Fifth Move Alternatives
1 54
10
Winawer Variation: Main Line with 7 'iVg4
175
11
Winawer Variation: Positional Lines
1 92
12
Winawer Variation: Black Plays 6 . . . 1Wc7
212
13
Classical Variation: 4 e5
227
14
Classical Variation: 4 .tg5 ( 4 . . . dxe4 5 tiJxe4 J.. e 7)
246
15
Odds and Ends
256
Index of Variations
264
Bibliography
Books Advance French: Alternatives to 4 c3, Jim Bickford (Syzygy 2002) Chess Informant 1-86, Chess Informant C1 8- 1 9 French Defence, Viktor Korchnoi (Chess Informant 1993) Die Franzoesische Verteidigung 3.Sc3, Hagen Tiemann (Reinhold Dreier 1 998) Four Gambits To Beat the French, Tim Harding (Chess Digest 1998) Franzoesisch Winawer: 7 Qg4 0-0, Stefan Kindermann & Ulrich DiIT (Chess gate 2001) French Classical, Byron Jacobs (Everyman Chess 2001) French Defence 3 Nd2, Lev Psakhis (Batsford 2003) French Defence Main Line Winawer, John Moles (B T Batsford 1975) French Defence, Alexander Kalinen (Russian Chess House 2002) French Defence, Neil McDonald, (ChessPublishing.com) French Defense 2, Nikolai Minev (Thinker's Press 1998) French Winawer, Neil McDonald (Everyman Chess 2000) New In Chess Magazine and Yearbook through 2003, New In Chess The French Tarrasch, John Emms (B T Batsford 1 998) Encyclopaedia of Chess Openings C, 4th Edition, Chess Informant 2000 The Main Line French: 3 Nc3, Steffan Pedersen (Gambit 2001) The Complete French, Lev Psakhis (Batsford 1 992) Databases and software MegaBase2003 TWIC 1 -457 (This Week in Chess, games) Correspondence Database 2002; ChessBase, 2002 MegaCorr2 ; Tim Harding, 2001 ChessBase Magazine (CBM) through #89; CD-ROM; ChessBase
Introduction
It may surprise the reader that in my largest database, mostly consisting of games from the last 10 years, the French Defence has been played in about 6. 75% of all chess games (the French Defence figure excludes King's Indian Attacks that can also be reached by a Sicilian Defence order) . This is less than 1 e4 e5 and less than the always-popular Sicilian Defence, but more than the Najdorf, Dragon, Rauzer, Accelerated Fianchetto and Taimanov Si cilians combined! In correspondence chess, the percentage rises to 8.2%, quite a chunk of all games played. The popularity of the French derives from its flexible and double-edged character. One can play it safely or with the most aggressive intentions . Black's share of the centre is greater than in most other defences to 1 e4, arguably only less than that after 1 . . . e5 (which tends to cede space early on anyway) . Thus the risk of being overrun is small and at the same time White is hard-pressed to make his first-move advan tage count at all. Finally, Black can achieve dynamic counterchances in every main line. I associate today's modern French Defence with specialists and innovators such as Bareev, Morozevich, Radj abov, Mikhail Gurevich, Korchnoi, Lputj an, Psakhis, Khalifman, Dreev, Nikolic, Uhlmann, Chernin, Kiriakov, Djurhuus, Yusupov, Sergey Ivanov, Ulybin, Gulko and Kindermann. Other leading players like Shirov, Ivanchuk, and Anand also use the French from time to time. Only a very few openings are played with such regularity at the grandmaster level. As with the first two editions of Play the French, this is a repertoire book that offers the reader two or more distinct variations to play against all im portant white systems . The idea is to provide maximum flexibility should some line of play prove unsatisfactory or in need of repair. As always, 8 years of practice since the last edition have drastically altered the theory on practi cally every variation. I have therefore thoroughly revised most older lines, but also presented new solutions. For example, I have added the Classical System (1 e4 e6 2 d4 d5 3 liJc3 liJf6), the Tarrasch Variation with 3 . . . i.. e 7 (af5
Play the French ter 1 e4 e6 2 d4 d5 3 t2Jd2) , a set of new answers to the King's Indian Attack (1 e4 e6 2 d3) , a new Main-Line Winawer Variation (3 t2Jc3 �b4 leading to 7 'ili'g4 0-0), and so forth. This is not to say that any major systems in the last edition are bad; on the contrary, none of them has been discredited. The in clusion of these new lines, however, provides some fresh air and illustrates the breadth of playable variations in the French Defence. Of special note: Chapter 1 2 expands upon the lines of the 6 .. :VJlic7 Winawer that were pre sented in the second edition. Full credit for this chapter goes to Norwegian FM Hans Olav Lahlum, who updated all of the variations therein. He used sources that I wouldn't have had access to and the insights of many strong players to put together a lucid exposition of that system. See also that chap ter's introduction. An opening book that tries to cover every useful byway will necessarily be dense and in some spots difficult to read. Thus one may want to skim over chapters before settling in to a closer examination. But I would urge the reader to pay particular attention to logical alternatives to what are (tempo rarily) the 'main lines' . This not only prepares one to meet moves that are likely to be played, but also informs one of key ideas that may not be explic itly re-described as the main line unfurls. There are too many people to thank individually; I am in debt to the many, many players who sent me material and questions via email. I received by far the most help from the above- mentioned Hans Olav Lahlum, who apart from Chapter 12 sent me many games and ideas in numerous other variations. Hans Olav wants to thank his friends who helped with the chapter on 6 . . . 'iIi'c7; they are cited in his exposition. Two players that were particularly helpful were my friend Joachim Wintzer, who made pertinent suggestions, and the insightful Antti Koponen who provided me with some excellent new ideas . French expert Stefan Kindermann nudged me in the right direction at the right time. Finally, my editor Byron Jacobs deserves credit for shaping a rough manuscript into a coherent and hopefully readable book. John Watson, September 2003
6
Chapter One Advance Variation : Introd u ction
1 e4 e6 2 d4 d5 3 e5 Historically, the Advance Variation and the Exchange Variation were the fIrst popular responses to the French Defence. The theory of 3 e5 was stimulated by the provocative ideas of Nimzowitsch but no leading advocate took his place and the Advance Variation was in remission for most of the modern era. In the last two decades, however, 3 e5 has taken a quantum leap in popularity. With the increasing number of top-level play ers using the French and with White's inability to find promising positions in many traditional lines with 3 lDc3 and 3 lDd2, leading grand masters have employed 3 e5 to get a different type of central advantage. They have devised new strategies to support that centre, notably by flank advances on both wings. These innovations have enjoyed con siderable success at the top levels; at the moment, the Advance Variation is so popular that it barely lags be hind 3 lDd2 in frequency (with 3 lDc3 still very much the leading choice, of course). 3 ...c5
The principled continuation, at tacking the centre at its base (d4) . Now White has four main choices. 4 c3 is the main line and subject of Chapters 2 and 3. Here we look at: 1 . 1 4 dxc5 1.2 4 'iVg4 1.3 4 lDf3
Others are either unchallenging or weak: (a) 4 b4?! is a sort of Wing Gambit that is popular at lower levels (com pare Chapter 1 5) ; Black continues 4 . . . cxd4 5 a3 and now: (al) 5 . . . lDc6 6 f4 (6 lDf3 iIIc 7 7 .i.f4 7
Play the French ttJge7 intending . . . ttJg6) 6 . . . ttJh6 (6 . . . a5 7 b5 ttJce7 8 ttJf3 ttJf5 9 J.. d 3 ttJgh6=i= with the ideas . . . i..c 5 and . . . ttJe3 Frenkel-Rao, Boston 1988) 7 ttJf3 (7 b5 ttJe7 8 ttJf3 ttJef5+) 7 . . . a5 8 b5 ttJe7 9 ttJxd4 ttJef5+ with multiple ideas including .. :iVb6 and . . . i.c5; (a2) 5 ... a5!? is another good solu tion, e.g., 6 b5 ttJd7 7 ttJf3 (or 7 f4 ttJh6 8 ttJf3 ttJf5 9 J.. d 3 ttJe3 and . . . J..c 5) 7 . . . Wic7 8 J.f4 ttJe7 9 i.e2 ttJg6 10 J.. g 3 J..c 5+ Cherkasov-Nikolenko, Czesto chowa 1992; (b) 4 J.. e 3 'iVb6 5 b3!? (5 dxc5 J.. xc5 6 J..xc5 'ilVxc5 is comfortable, e.g., 7 'ilVg4?! 'ii'xc2! 8 "iYxg7 'i'c 1 + 9 �e2 'ii'xb2+ 10 ttJd2 "iYxa1 11 'i'xh8 �f8) 5 . . . ttJc6 6 ttJf3 cxd4 7 i.. xd4 ttJxd4 8 'ilVxd4 i..c 5 9 'i'd2 ttJeTi= Renaud-Hook, Dubai 1986; (c) 4 f4 is probably the most inter esting of these deviations, but it takes time and opens White's centre to at tack after 4 . . . ttJc6 (or 4 . . .cxd4 5 ttJf3 ttJc6 and 6 ttJxd4 'i'b6 or 6 ttJbd2 ttJh6 7 i.. d 3 ttJg4 8 ttJb3 i..b 4+ 9 �e2 f6 10 h 3 ttJ e 3 1 1 i.. x e3 dxe3 12 c3 .i.e7=i=) 5 c3 Wb6 6 ttJf3 ttJh6 7 b3 (two lines that leave Black with a pair of bish ops and White with a miserable one are 7 'i'b3 cxd4 8 'ii'xb6 axb6 9 cxd4 ttJf5 10 i..b 5 i.. d 7 1 1 i.. xc6 bxc6 Pul nikov-Iljin, Togliatty 200 1 ; and 7 i.. d 3 .i.d7 8 'iVb3 'ii'xb3 9 axb3 cxd4 10 ttJxd4 ttJxd4 Ozkan-Ozturk, Ankara 2002, when simplest was 1 1 cxd4 ttJf5 12 i.xf5 exf5=i=) 7 . . . cxd4 8 cxd4 .i.b4+ 9
(d) 4 ttJd2!? is the equivalent of the Tarrasch variation 3 ttJd2 c5 with the strange-looking 4 e5!? Black can play 4 . . . cxd4 (or 4 . . . ttJc6, e.g., 5 ttJgf3 'i'b6 6 dxc5 J.. xc5 7 'i¥e2 Balogh-Rittner, corr 1958, and now the easiest move is 7 .. .f6 8 ttJb3 ! fxe5 9 ttJxc5 1i'xc5 1 0 ttJxe5 ttJ f6 with development and ac tivity; Black even has the idea of . . . ttJxe5 and then . . . 'i'xf2+) 5 ttJgf3 ttJc6 6 ttJb3 'i'c7 7 i.b5 (7 i.f4 i.b4+) 7 . . . J.. d 7 8 i.xc6 (8 'i'e2 J..b 4+ 9 J.. d2 J..xd2+ 10 ttJbxd2 ttJge7=i=) 8 . . . bxc6 9 'i!Vxd4 c5 10 'i'f4 ttJe7 with the two bishops and intending 1 1 ttJg5 f6! 1 2 exf6 'ilVxf4 1 3 i.. xf4 gxf6 with a mas sive centre; (e) 4 i.. f4 cxd4 (4 . . . 'i'b6 5 b3 ttJc6=i=) 5 ttJd2 'i'b6 6 ttJb3 i.b4+ (Harding). Then 7 i.. d2 ttJc6 8 ttJf3 f6 is good for Black; (f) 4 i..b 5+ will trade White's 'good' bishop for Black's 'bad' one, but the real problem is that it helps Black to develop : 4 . . . i.. d 7 (or 4 . . . ttJc6, since 5 ttJf3 'iVb6 gains a tempo unless White plays the depressing 6 i.. xc6+ bxc6) 5 i..x d7+ ttJxd7 6 ttJf3 (6 c3 'iVb6 fol lowed by . . . ttJe7-c6 and/or .. .f6 depending upon White's plan) 6 . . . cxd4 7 'ilVxd4 i..c 5 8 'i'd3 f6! 9 i.. f4 'i'b6 1 0 0 - 0 ttJe7!? (1O . . . g5 ! , e. g. , 1 1 i.. g 3 g4 12 ttJfd2 fxe5) 1 1 exf6 gxf6 12 'i'b3!? 'i'c6 13 .i.e3 e5 and Black has his typical central advantage, T.Barnes-Sporn, email 1998. 1.1 4 dxc5 One of White's most popular op tions to 4 c3. Rather than defend d4, White shifts his attention to bolster ing e5 and bringing his pieces to more active squares than in other lines. The drawbacks to 4 dxc5 are that White gives up the pillar of his centre and Black develops rapidly. I find
Advance Variation: Introduction this variation truly instructive in that one will see the combination of dxc5 and e5 throughout this book in re markably many variations exhibiting similar strategies. 4 ...liJc6 Avoiding the awkward 4 . . . i.xc5 5 'ii' g 4. 5 liJ f3! White should avoid 5 i.f4?! i.xc5 6 i.d3?, which practically loses after 6 . . . 1li'b6! 7 liJc3 (7 liJf3 i.xf2+ 8 'it>e2 'ii'xb2+) 7 . . . i.xf2+! 8 <;£?fl i.d4 and Black stands to win. Also not impres sive are 6 c3 liJge7 7 liJf3 liJg6 8 i.g3 f6! 9 b4 i.b6 10 exf6 'ii'xf6 11 i.d3 0-0=1= and 6 'iVg4 liJge7! with the idea 7 'ii'xg7 (7 i.d3 1IVb6) 7 . . . .l:!.g8 8 'ii'xh7 i.xf2+! 9 'iitxf2 1IVb6+ 10 'it>e2 'ii'xb2 . 5 ... i. xc5
6 i. d3 (a) 6 a3?! is too slow after 6 . . .f6: 7 b4 i.b6 8 b5!? (8 exf6 liJxf6 9 i.e2 0-0 10 0-0 e5! 11 b5 liJd4) 8 . . . liJxe5 9 liJxe5 fxe5 10 'iVh5+ 'iitf8 1 1 'ii'x e5 i.c7 and Black wins the centre again, e.g., 12 '1We3 'fIf6 13 l:.a2 liJe7 14 i.b2 e5; (b) 6 i.f4?! allows 6 . . . 1li'b6 (6 . . . liJge7 and 6 .. .f6 are also fine) 7 i.g3 'fIxb2 8 liJbd2 liJge7 with White having insuf ficient compensation for a pawn. 6 ... £6 The clearest course, since it liqui-
dates the centre. For those wishing for a game of a more closed nature, 6 . . . liJge7 is a good alternative:
(a) 7 0-0 liJg6 8 .l:!.el (a good plan for Black is illustrated by 8 'i'e2 0-0 [or 8 . . . i.b6 9 c3 i.c7 10 i.g5 'ii' d 7 1 1 i.xg6 fxg6] 9 c 3 i..b 6 10 liJa3 i.c7 Avirovic-Subaric, Novi Sad 1945; now White should try 1 1 i.. xg6 fxg6, but Black has at least equal play, espe cially since in many lines . . . l:i.xf3 can give him a pawn, mobile centre, and attack for the exchange) 8 . . J!Vc7!? (8 ... 0-0 may be the easiest course, intending . . . i.b6, . . . i.d7, . . . i.c7 and responding to i.. xg6 with . . . fxg6) 9 'ii' e 2 liJd4 (Czerniak suggests 9 . . . i.. d 7 followed by . . . 0-0-0) 10 liJxd4 i.xd4 1 1 i.xg6 fxg6!? (the most interesting move) 12 c3 �b6 (12 . . . �c5! is more accurate; Black follows up as in the game, or by . . . 0-0, . . . �d7 and dou bling on the f-file) 13 liJd2 0-0 14 liJf3 i.c5 (a tempo down, but still satisfac tory) 1 5 i.e3 b6 16 �d4 a5 1 7 'ii'e 3 i.a6 18 .l:!.adl h6 with interesting play, Westermeier-Reefschlaeger, Menden 1 974. Black has the idea of . . . g5 and again doubling rooks; (b) 7 liJc3 a6!? (to secure a7 for the bishop; 7 . . . liJg6 is equal) 8 0-0 liJg6 9 �xg6!? (9 .l:!.el 0-0 and White lacks for a plan) 9 . . . hxg6 10 liJe2 'iVc7 1 1 i..f4 9
Play the French Biti-Thi Thanh Houng, Moscow 1994, when 1 1 . . .i.. d 7 12 'Wd2 0-0-0 was natural and good; (c) 7 .i.f4 �6 (the most ambitious; 7 . . . tDg6 is also logical) 8 0-0 tDg6 9 i.. g3 (9 'Wc l tDxf4 10 "it'xf4 - Keres, but Black can grab a pawn and re duce White's attacking possibilities by 10 . . :Yixb2 1 1 tDbd2 f5 ! ; in P. Short Heidenfeld, Castleconnell 2000, White tried the gambit 9 iLxg6 fxg6 10 tDc3, but Black can simply grab a pawn by 10 . . . 'Wxb2, since 1 1 tDa4 'itb4 12 tDxc5 'Wxf4 gives White very little) 9 . . :�xb2 10 tDbd2 tDgxe5! (10 . . :�a3 1 1 h4 'iVa4!? with the idea 12 h5 tDf4 is also interesting) 1 1 tDxe5 tDxe5 12 l:i.b l 'iVc3 13 l:i.b3 'i'd4 14 Ji.xe5 (14 i..b 5+ tDd7 1 5 .txd7+ i.. x d7 1 6 ':'xb7 Pachman, when 16 . . . i.. c6 1 7 lIc7 "YWa4 consolidates the pawn) 14 . . . 'i¥xe5 1 5 i..b 5+. This keeps the king in the cen tre but Black has the bishop pair and mobile central pawns: 15 . . . 'it>e7 1 6 l:i. e l 'iVd4 1 7 l:tee3 a 6 1 8 Ji.f1 .l:i.d8 1 9 l:tbd3 'iVh4 2 0 l:i.g3 g 6 2 1 c4 � f8 and White had no compensation, T.Johan sson-E.Berg, Bergen 200 1 .
7 'We2 7 iLb5? 'iYb6 intending 8 'iVe2 .txf2+ is weak, and the others are fun for Black: (a) 7 exf6?! tDxf6 8 0-0 0-0 9 c4 (9
10
lIel e5!; 9 'i'e2 e5! 1 0 tDxe5 tDxe5 1 1 'Wxe5 tDg4 1 2 'i!Vh5 g6+) 9 . . .'iVd6 (9 . . . e5!? 10 cxd5 e4 1 1 dxc6 'Wxd3 1 2 'iVxd3 exd3:j: Van Scheltinga-Van der Tol, Netherlands 1 946) 10 tDc3 a6 1 1 i.. g5?! tDg4!+ intending . . .l:txf3 Camil leri-Debarnot, Malta 1980; (b) 7 i.. f4? practically loses by force: 7 . . . fxe5 (or 7 . . .'iVb6) 8 .txe5 (8 tDxe5? 'Wf6+ , e.g., 9 tDxc6 'i'xf4 10 'i¥h5+ g6 1 1 .txg6+ �f8 12 tDe5 tDf6 13 'We2 'Wc l + 14 'i'dl 'i¥xdl + 15 �xdl i.d4- + Rohel-Keitlinghaus, Badenweiler 1990) 8 . . . tDf6 (or 8 . . . tDxe5! 9 tDxe5 'i'g5) 9 0-0 0-0 (threatening . . . ltJg4) 10 h3 ltJe4! 1 1 "YWe2 l:txf3! 12 gxf3 "YWg5+ 13 �hl ltJxf2+, etc. 7 fxe5 Here Black can play ultra-dynam ically by 7 .. :W1c7!? 8 i..f4 g5! 9 .tg3 g4. ...
This wins a key central pawn at the cost of exposing Black's king. Thus far this move has done reasona bly well and I see nothing wrong with it. Here are two examples: 10 tDfd2 (10 ltJh4 ltJxe5 1 1 0-0 Ji.d6 12 ltJc3 a6 was solid enough in Van der Nat-Van Oirschot, corr 1 976) 1O . . . ltJxe5 1 1 .i.b5+ �f8!? (1 1 . . .i.d7 1 2 Ji.xd7+ �xd7 is also unclear. White wants to at tack, Black to consolidate his extra pawn by, e.g., . . . l:te8 and . . . �c8) 12 ltJc3 Ji.d4 13 0-0 a6 14 i.. d 3 h5 with
Advance Variation: Introduction the idea . . . h4, Thomson-Caamano, corr 1 988. Black's intended . . . h4 will interfere with White's only dangerous plan of �hl and f3. The game contin ued 15 l:t.fe l h4 16 .i.xe5 �xe5 1 7 '1i'xg4 ..txh2+ 18 �hl ..td6 19 �f3 .i.d7 and White didn't have enough for the pawn. S ttJxe5
It's remarkable that this position might even turn out to be advanta geous for Black. His central majority is very important and White cannot maintain his outpost on e5. S ttJxe5 This leads to a degree of simplifica tion, but there remains an imbalance in the position. Instead 8 . . . ttJf6 can be played in order to equalise or even win if Black is willing to take some risks. He threatens . . . ttJxe5 and . . .�xf2+, and play might go 9 .i.f4 (9 O-O?! ttJxe5 10 "iYxe5 0-0 1 1 c4 'tWb6 again threatens . . . i..xf2+: 1 2 "iYe2 .i.d7 1 3 ttJd2 l:rac8:j: Makropoulos-Hug, Nice 1974; 9 ttJxc6 bxc6 10 ttJd2 0-0 1 1 ttJf3 �c7! planning . . . .i.d6 and in some cases . . . ttJg4) and now: (a) 9 . . . 0-0 heads for equality: 10 0-0 ttJe4 11 ttJxc6 bxc6 12 .i.e3 �xe3 !? 13 '1i'xe3 ttJf6 14 ttJd2 Becker-Maroczy, Karlsbad 1929; now Becker gives 14 . . . 'tWb6! when a plausible line is 1 5 ...
'iVxb6 (15 'iVe5 �8!? 16 ttJf3 ttJg4 1 7 'iVh5 ttJf6 1 8 �e5=) 1 5 . . . axb6 16 ttJf3 c5 17 ttJe5 (17 c4 :a4!? 18 b3 .l:ta3) 17 . . . c4 18 i.. e 2 ttJd7 19 ttJxd7 i.. x d7 20 f4 �f7= intending . . . 'it>e7-d6. (b) 9 . . . �6!? is more ambitious and practically forces White to gambit by, for example, 10 0-0 (or 10 ttJxc6 'iVxb2 1 1 'iVe5 'iVb6) 1O . . . �xb2 1 1 ttJxc6 bxc6 12 ttJd2 'iVa3, when White has some thing for the pawn but I'm not sure that it's enough. 9 "iYxe5 'iVf6 This is Black's best chance for ad vantage: he tries to get to an ending that will be favourable to him by vir tue of his central majority. Again 9 . . . ttJf6 suffices for easy equality and perhaps a tiny bit more. White must take care, e.g., 10 0-0 (10 �f4? .i.xf2+; 10 .i.b5+?! 'it>f7 when . . .'iVb6 and . . . .l:tf8 with . . . 'it>g8 will follow: 1 1 0-0 �6 12 .i.e2 - versus ... .i.xf2+ with a ... ttJg4+ fork in mind - 12 . . .11f8 13 ttJd2 'it>g8 14 'iVg3 e5!) 10 . . . 0-0 1 1 c4?! .i.d6 12 'iVe2 �d7 13 ttJc3 "fIc7 14 h3 .l:tac8!? 1 5 cxd5 exd5 16 i.. g5 llce8 1 7 'iYd2 Gallego-Mawed, Bled 2002, and one course was 17 . . . i.. c 6 18 �xf6 .l:txf6 1 9 .l:tael lief8 with a small edge due to the bishops.
10 'iYe2 White needs take more care than
11
Play the French Black due to the latter's central majority and in fact he has few desirable moves: (a) 10 'iVc7?! is speculative after 10 . . . 'i!Vxf2+ 1 1 �dl .i.d7! 12 lifl lic8 13 'i!Vxb7 'iVxg2+ ; (b ) 10 i.. f4 i.. d 4!? 1 1 i..b 5+ 'iitfl 12 'i!Vc7+ 'tJe7 1 3 c3 i..b 6;; (c) 10 i..b 5+ i.. d 7!? (or 1O . .. �fl and 1 1 'iVg3 'tJe7 or 1 1 1i'c7+ 'ile7 12 'ilg3?! 'tJf6 1 3 0-0 e5 or 1 3 . . . i.. d 6, but White can doubtless play better) 1 1 i.. x d7+
1.2 4 'iVg4
Nimzowitsch's radical idea. White abandons d4 and intends to overpro tect e5 by means of 'tJf3, 'iVg3, and later .tf4 and/or liel . 4 'iVg4 is in some ways less of a challenge for Black than 4 dxc5. Bringing the queen out before the other pieces is not necessarily bad; but if she loses more time and fails to create weak nesses in the enemy camp such a foray can hardly be justified. In the previous edition I gave 4 . . . 'tJc6 5 'tJf3 cxd4 6 i.. d3 'i!Vc7 (see the next note). Another popular and successful line is 4 . . . cxd4 (or 4 . . .f5 5 'iVg3 cxd4 6 'tJf3 'tJc6, often transposing) 5 'tJf3 f5 6 'i!Vg3 and 6 . . . 'tJc6 or 6 . . . 'tJe7 intending . . .'tJd7-c5. Here as my main line I will show a less tested and generally sim pIer response that solves Black's problems and creates ones for White. It was played by Botvinnik as far back as 1937 and has been revived to good effect in recent times: 4 ...'iVa5+ Meeting a queen move with a queen move! Black's idea is to be sure that White's queen will lose time to . . . 'tJh6 by diverting the cl bishop from a potential capture on h6. At the right moment . . .f5 can also be useful. The line 4 . . . 'tJc6 5 'tJf3 cxd4 (note that
Advance Variation: Introduction here 5 . . :ii'a5+ 6 i.. d 2 'ib6 attacks b2, and otherwise 7 c3 cxd4 transposes to our 4 .. :ilVa5+ main line) 6 .i.d3 filc7 was analysed in detail in the second edition.
Here let me present a brief update for Black covering the main continua tions: (a) 7 .tf4 liJge7 is fine for Black, but 7 . . . liJb4 has scored 80% in my database! Of course things aren't really so bad for White, but Black is certainly content after 8 0-0 (8 liJa3 liJxd3+ 9 cxd3 'ib6 10 ..tc1 liJh6! 1 1 Wixd4? .tc5 is extremely strong, e.g., 12 Wia4+ .td7 1 3 Wic2 .tb4+!? 14 .i.d2 l:i.c8 1 5 1i'd1 ..tc5 1 6 .i.e3? i..x e3 0 - 1 , in view of 17 fxe3 Wixb2, Cornelison Dirr, email 1 994; a well-known trick is 8 liJxd4?! liJxd3+ 9 cxd3 1lVb6 10 liJb3 'ib4+ 1 1 liJ 1 d2?? g5 winning a piece) 8 . . . liJxd3 9 cxd3 Wic2 10 liJxd4 Wixd3 !? (10 . . . Wixb2 is more likely to yield an advantage, when 1 1 liJb3 h5! chases the queen to a square from which Black's development is not so hindered, for example, 12 Wif3 .i.d7 1 3 liJ1d2 .i.a3!) l l l:i.d1 ! (if l l liJb3 h5! and the queen has no good square) 1 1 . . :�'g6 (1 1 . ..Wia6!?) 12 'i!Ve2 M.Pruess-Kumme row, Dortmund 1991; and now Black's best course was 12 . . . i.. d 7!= with the idea 13 ttJb5 l:i.c8;
(h) 7 Wig3 f6 with: (h I) 8 .i.xh7?! liJxe5 9 i.. g6+ <;t>d8 (9 . . . 'itt d 7! is even better) 10 liJxd4 (10 liJxe5 Wixe5+ 11 'i!Vxe5 fxe5+ with an extra pawn and huge centre) 1O . . . liJe7 11 .i.d3 liJxd3+ 12 Wixd3 J:[xh2 1 3 0-0 l:i.h8+ Toth-Maillard, corr 1990; (h2) 8 exf6 'ilVxg3 9 f7+ 'ittxf7 10 hxg3 liJf6 1 1 liJg5+ 'it>g8 12 f4 liJb4 (or 12 . . . h6 13 ttJf3 ttJe4) 1 3 liJd2 liJxd3+ 14 cxd3 Linskens-Munoz Izcua, Mon tevideo 1954; and best was 14 . . . ttJg4! 1 5 liJb3 e5!+; (c) 7 0-0 f6!? (7 ... i.. d 7 8 l:.e1 0-0-0 has also been successful) and: (c 1) 8 .i.xh7 liJxe5 9 liJxe5 fxe5 10 .i.g6+ (10 l:i.e1 liJf6 1 1 i.. g6+ 'itt d 7!) 10 . . . 'itt d 8 11 h3 (1 1 'ilVg3 ..td6) 1 1 . . .liJf6 12 'i!Vd1 Smyslov-Lisitsin, Moscow 1942; 12 . . . i.. c 5!� and Black will soon mobilise with . . . e4; (c2) 8 SLf4 fxe5 9 liJxe5?! (9 �5+ Wif7 10 Wixf7+ <;t>xf7 1 1 liJxe5+ liJxe5 12 .i.xe5 liJe7!� intending . . . liJc6 and . . . .i.c5) 9 . . . liJf6! 10 .tg6+ (10 liJxc6 liJxg4 1 1 .i.xc7 bxc6; 10 filg3 ..td6) 10 . . . hxg6 (or 10 . . . 'itt d 8 1 1 filg3 hxg6 12 liJxg6 e5 etc.) 1 1 Wixg6+ 'it>e7+ .
5 c3 The most popular reply. 4 . . :�Va5+ is a pragmatic move that counts upon the drawbacks of each response. Much of what follows is analysis
13
Play the French without examples, but the main ideas should be clear: (a) 5 ttJd2 ttJc6 (5 . . . ttJh6 is also fme) 6 ttJgf3 (6 dxc5 ttJh6 7 'iVf4 i.. e 7! 8 c3 [8 h4 f6! 9 exf6 i..xf6 and . . . 0-0] 8 . . . g5! 9 'iWe3 ttJf5 10 'iVe2 'iVxc5 with supe rior development) 6 . . . ttJh6
Here's the drawback to 5 ttJd2. Black exploits the fact that the c 1 bishop n o longer covers h 6 : 7 'iWf4 ttJb4 (forcing the king to move) 8 <;t>d1 (8 ttJb3? 'ilVa4) 8 . . . c4 9 a3 (after 9 ttJb 1?! ttJxa2, there is no way to ex ploit the pin and Black threatens to play . . . ttJc3+; 9 c3 'iWa4+ 10 b3 cxb3 1 1 ttJxb3 ttJxa2 1 2 'iii>c 2 i.. d 7 1 3 ttJfd2 .l:tc8 14 i..b 2 �b4 15 'iii>b 1 i..xc3 16 l:i.xa2 i..xd2 0 - 1 Nei-Gleizerov, Osterskars 1995; Black ends up with a couple of extra pawns, although White could have played on) 9 . . . �d7 10 b3 l:tc8! 1 1 �b2? (yet 1 1 bxc4 dxc4 12 ttJxc4 'iWa4 is also good for Black: 13 'iWd2 ttJg4! 14 ttJe1 ttJd5 1 5 ttJb2 - forced thus far - 1 5 . . . ttJxf2+! 16 'i¥xf2 ttJc3+ 1 7 'it>d2 ttJe4+ 18 'it>e3 ttJxf2 19 ttJxa4 ttJxh1 20 ttJb2 g6-+) 11 ... ttJxc2 12 'iii>xc2 c3+ 13 'iii>b 1? cxb2 0- 1 Pusch-Kern, Reckling hausen 2000; (b) 5 i.. d2 ttJh6! (5 . . . 'iVb6 is less clear but still promising after 6 �c3 ttJh6 7 'iWf4 ttJf5 8 ttJf3 ttJc6 or 6 ttJc3 ttJh6, e.g., 7 �b5+ 'iVxb5 8 ttJxb5 ttJxg4 14
9 ttJc7+ 'iii> d 7 10 ttJxa8 cxd4 1 1 ttJf3 ttJc6 and . . . b6; or, finally, 6 b3 cxd4 7 ttJf3 ttJc6 8 i.. d 3 f5 9 exf6? ttJxf6 10 'iWg3 �d7 1 1 a4 ttJe4+ Kaarne-Sisatto, Finland 1998) 6 'iWg3 (6 �xa5 ttJxg4 presents White the double problem of . . . ttJc6 and defending his centre, so 7 h3 ttJh6 8 ttJf3 ttJc6 9 �c3 ttJf5:j: could follow) 6 . . . 'iVb6 (hitting b2 and d4) 7 'ilVb3 'i¥xb3 8 axb3 ttJc6 9 ttJf3 (9 �xh6?! gxh6 10 dxc5 i.. xc5!? 1 1 ttJf3 ttJd4!; here 10 . . . ttJxe5 or a move ear lier 9 . . . ttJxd4! was also strong) 9 . . . ttJf5 10 dxc5 �xc5 with the freer play. Black is somewhat better. 5 ... cxd4 6 ttJf3!? It seems odd, but trying to gambit a pawn may be best. The alternative 6 'iVxd4 ttJc6 is easy for Black to play, as in the stem game for 4 . . . 'iWa5+, Rabinovich-Botvinnik, Moscow 1937: 7 'iVf4 ttJge7 8 .i.d3 (8 ttJf3?! ttJg6 9 'iVg3 'iVc7 10 .i.f4 f6 1 1 �b5 fxe5 12 �xe5 ttJgxe5 1 3 'iWxe5 �d6+) 8 ... ttJg6!? (here 8 ... d4 9 ttJf3 ttJd5 is also inter esting) 9 �xg6 hxg6 10 ttJf3 'iVa6 1 1 ttJg5 ttJd8 12 ttJd2 'i¥d3 !? 13 ttJb3 b 6 14 �d2 �d7 1 5 0-0-0 :c8:j:.
6 ... ttJ c6 Black could accept the gambit by 6 . . . dxc3, but that isn't necessary, es pecially since he also has the option of 6 .. .f5, for example, 7 'iWg3 ttJc6 8 b4
Advance Variation: Introduction Buchnicek-P.Kucera, Mlada Boleslav 1 993; and here 8 . . . 'iWa4! was strong in view of 9 b5 "iVc2 ! 10 'iVf4 'iVe4+ 1 1 'iVxe4 fxe4 (or 1 1 . . . dxe4 1 2 bxc6 exf3 13 cxd4 tLle7!:j:) 12 bxc6 exf3 . Then White has serious difficulties, e.g., 13 cxb7 (13 .ltb5 c;i;>f7 14 cxb 7 fxg2 1 5 l:!. g l .ltxb7 16 cxd4 tLle7 1 7 lhg2 tLlf5+) 13 . . . .ltxb 7 14 cxd4 fxg2 1 5 .ltxg2 tLl e 7 16 .lth3 (versus . . . tLlf5) 16 . . . tLlc6 1 7 i.b2 (17 i.e3 tLlxe5!) 1 7 ... tLlh4! 18 'iit> d 2 l:.c8 19 tLla3 'iit> d 7 20 l:.hc 1 i.e7 threatening . . . i.g5+ as well as . . . l:.hf8. 7 tLlxd4 7 .ltd3 would be a poorly-timed sac rifice due to 7 . . . dxc3 8 tLlxc3 (8 bxc3 d4 9 0-0 dxc3 10 tLla3 and now 10 . . . h5 11 'iWg3 h4 12 'iWg4 h3! or 10 . . . tLlh6!? 11 .ltxh6 gxh6 12 tLlc4 "iVc5 13 l:!.ab 1 tLlb4! 14 l:i.fd1 i.. d 7+ , Jimenez Villena Shabalov, Linares 2000) 8 . . . d4!? (or 8 . . :�b4, since White cannot avoid the exchange of queens due to . . . d4) 9 tLlxd4 tLlxe5 10 'iVg3 tLlxd3+ 1 1 'iWxd3 .ltd 7 with a solid extra centre pawn.
7 ... tLlxe5!? This wins a key central pawn at the risk of some attack by White. Black can also play effectively by 7 . . . Wic7 8 i..f4? (8 tLlxc6 is best, al though 8 . . . bxc6 strengthens Black's centre, e.g., 9 'iWg3 tLle7 10 .ltd3 tLlg6
1 1 f4 WVb6! preparing . . . iLe7 and/or . . . i.a6) 8 . . . tLlxd4 9 cxd4 'iWb6!, hitting b2 and d4. If 10 i.c 1 , 10 . . . tLlh6 1 1 'iWf4 tLlf5 wins the d-pawn. S 'iWg 3 8 'iWh5?! tLld7! 9 iLd3 (not 9 tLlxe6? tLlgf6 winning a piece) 9 . . . tLlgf6 and White has nothing for the pawn. S . . tLlg6 There's nothing wrong with 8 . . . tLlc4, and 8 . . .f6 is promising. In the latter case Black intends to give the pawn back for development and dark squares after 9 i..f4 (9 i.e2 c;i;>f7 un tangles and prepares . . . iLd6) 9 . . . �6 10 b3 (10 b4 'fic7 11 iLb5+
f7! , preparing . . . i.. d 6 and giving back the pawn for develop ment and dark squares. .
9 h4 White lacks compensation after 9 tLlb5 e5 10 i.d3 f5 ! or 9 i.. d 3 i.. d 7 10 0-0 tLlf6. 9...tLlf6 10 h5 tLle4 1 1 Wie 3 tLle7 12 tLld2 12 i.d3 e5 1 3 tLlf3 tLlc6! 14 i.. xe4 dxe4 15 'iYxe4 i.e6 16 0-0 0-0-0 with space, two bishops and a clear advan tage. 1 2 ... tLlxd2 Or 12 . . . tLld6 with . . .tLlef5 in mind.
15
Play the French Then 13 �d3 'iVb6 14 liJ2f3 f6! ; pre pares . . . e5. 1 3 ii.xd2 ..td7 14 ..td3 liJc6 1 5 0-0 liJxd4 16 'i¥xd4 'iVc5 17 'iWe5 f6 1 8 'i'g3 'iWd6 and Black has consolidated the ex tra pawn with insufficient counter play for White. Notice the number of good options the second player had along the way. The whole 4 . . . 'i'a5+ variation is only lightly tested, but appears to offer Black more than adequate chances. 1.3 4 liJf3
Of White's alternatives to 4 c3, this is the most common. 4 liJc6 5 �d3 5 dxc5 is 1 . 1 and 5 c3 is the subject of the next two chapters. The move 5 �b5?! has been played in over 200 database games, but loses time after either 5 . . . �d7 (threatening . . . liJxe5) , which has scored 73 % , or 5 . . . 'iVb6!? which has scored a cool 80% for Black, e.g., 6 liJc3! (6 i.xc6+ bxc6 7 0-0 - else . . . �a6 - 7 . . . cxd4;; 6 'iWe2 cxd4 7 0-0 �d7 - intending . . . liJxe5 8 ..td3 f6l:j:) 6 . . . �d7 7 .i.xc6 bxc6 8 dxc5 .i.xc5 9 0-0 'iWc7 10 liJa4 �b6 1 1 b 4 liJe7 12 c 3 0-0 1 3 c4?! c5! 1 4 bxc5 ..txa4 15 'iWxa4 (15 cxb6 axb6 16 'iWe2 dxc4+) 15 . . . �xc5 16 cxd5 liJxd5 1 7
�d2 ':'fc8 18 Wh1 l:tab8+ Brans M.Nikolic, Dresden 2000. 5 ...cxd4 6 0-0 f6 An interesting juncture. 6 . . . .i.c5 was analysed in the last edition and is in good theoretical shape. 6 . . . f6 immediately challenges e5. An ob scure but promising alternative is 6 . . . g6!?, as suggested in the last edi tion. There are few examples, but play might proceed 7 l:te1 (7 a3?! ii.g7 8 �f4 of Kobelev-Lobach, USSR 1 988 invites 8 . . .f6! 9 .l:r.e1 fxe5 10 liJxe5 liJge7 with an extra pawn and supe rior activity; 7 ..tf4 i.g7 8 liJbd2 f6 probably transposes but here 7 . . . 'i'b6 is an option) 7 . . . ii.g7 8 �f4 (8 liJbd2 f6 9 exf6 liJxf6 10 liJb3 0-0 with the idea 11 liJbxd4 liJxd4 12 liJxd4 e5! 13 l:txe5 liJg4) 8 . . .f6 9 liJbd2 (9 ..tb5 liJge7 10 liJxd4 0-0; 9 exf6 liJxf6 10 .i.b5? 0-0 1 1 ..txc6 bxc6 12 liJxd4 'i'b6 or here 12 . . . liJe4) 9 . . . liJge7! (or 9 . . .fxe5 10 liJxe5 liJf6=) 10 exf6 (10 'iWe2 fxe5 1 1 �xe5 liJxe5 1 2 liJxe5 0-0 1 3 liJdf3 liJc6;) 10 . . . i.xf6 1 1 liJe5?! ( 1 1 ..th6 'iYd6 and . . . e5; 11 liJb3 0-0 12 'iWd2;) 1 1 . . .liJxe5 12 �xe5 �xe5 13 l:txe5 0-0 14 liJf3 'i'd6 and White has nothing for the pawn.
.•.
16
7 'i¥e2 (a) 7 :e1? loses the centre to 7 .. .fxe5 8 liJxe5 liJxe5 9 l:!.xe5 liJf6 10
Advance Variation: Introduction i.b5+ (10 i.g5 i.d6 1 1 i.b5+ <Jilfl 12 �el h6 1 3 .i.h4 g5! 1 4 i.g3 i.. xg3 15 hxg3 'iWb6+) 1 0 .. .'iit fl 1 1 'iVxd4 i.d6 12 .l:le3 (the best try) Kogan-Perez Gar cia, Dos Hermanas 2000; now 12 . . . 'iVc7! 13 'iWh4 g5! 14 'iVh6 'iVxc2 and White is in bad shape; (b) 7 1.f4? g5! (7 . . .fxe5 8 lDxe5 lDf6 9 'i!t'e2 transposes to 7 'iVe2) 8 i.g3 g4 9 lDh4 (9 lDxd4!? lDxd4 10 YWxg4 lDc6 1 1 'iWh5+ \it1d7 12 lDc3 'iVe8 and White has some attack, but not enough for a piece) 9 . . .fxe5 10 'ilfxg4 lDf6+ Black has a pawn and huge centre, and in tends moves like . . . i.d6, .. :ii'b 6, and . . . e4; (c) 7 i..b 5 i.d7 8 i.xc6 bxc6 9 'it'xd4 c5 (or 9 . . . f5 first, e.g., 10 'it'f4 lDh6 1 1 lDg5 .i.e7 12 'ii' g 3 lDfl! 1 3 lDxfl �xfl 14 c4 d4 1 5 lDd2 c5 16 lDf3 h6 with two bishops and moves like . . . l:[b8, . . . ..tc6, and . . . g5 in store, Gamback Kiriakov, Skelleftea 2001) 10 'iVf4 f5 1 1 c4 (otherwise Black's two bishops and central majority give him the advantage) 1 1 . . .d4 12 b4 (12 lDbd2 a5! 1 3 b3 i.e7 14 lDel Arnason-J.Watson, Gausdal 1978; 14 . . . lDh6! :j: and . . . lDfl) 12 . . . cxb4 13 lDxd4 (13 'i!t'xd4 lDe7 14 i.f4 lDc6 1 5 'i!t'e3 1Vb6 with the idea . . . i.c5) 13 . . . lDe7 14 'iVe3 lDg6!? (or 14 . . . lDc6 1 5 i.b2 lDxd4 16 i.xd4 "fic7) 15 f4 (15 lDd2 'iVb8 16 lD2f3 ..tc5 1 7 .i.b2 0-0) 1 5 . . . i.c5 1 6 lDd2 a 5 ( 1 6 . . . i.a4!?) 1 7 lD2b3 i. a 7 1 8 a3 a 4 19 axb4 axb3 20 c5 0-0 with the idea . . . lDe7-d5, among others, Bickford Pals, corr 1997/8. 7 ...ilc7 I use this as the main line because it is probably the best way to play for a win and also because there is less theory than after 7 . . . fxe5 8 lDxe5, when both 8 . . . lDf6 and 8 . . .lDxe5 were analysed in the second edition of this book. The former keeps more pieces
on the board; its main line goes 8 . . . lDf6 9 i.f4 (9 i.g5 lDxe5 10 'it'xe5 i.d6 1 1 i.b5+ 'iit fl 1 2 'iWxd4 'iWc7:j: ; 9 f4 ..td6, e.g., 10 lDd2 0-0 1 1 lDdf3 'i!fb6) 9 . . . ..td6 10 lDd2 0-0 1 1 lDdf3 'ilfc7 (1 1 . . .'ii'e 8!? 12 .i.g3 is complex after Bickford's 12 . . . lDh5= or 12 . . . i.d7 1 3 lDxc6 ..txg3 14 lDcxd4 ..td6 1 5 l:tae l 'it'h5 Hodgson-Speelman, Brighton 1 984) 12 1.g3 (a cute miniature fol lowed 12 l1ael lDe4 13 lDxc6 i.xf4 14 lDcxd4 e5 1 5 lDb3 i.g4 16 \it1hl?? lDg5 0 - 1 Aagaard-McDonald, Budapest 1 996) . At this point I suggested 12 . . . g6 unclear, with the idea . . . lDh5; in practice, 12 . . . i.d7 has also held the balance.
8 i.f4!? 8 1.b5 is safer, but Black has an easy time of it after 8 . . . fxe5 9 lDxe5 lDf6 (9 . . . .i.d6!? 10 'i!Vh5+ g6 1 1 lDxg6 "fifl=) 10 lDd2 (10 i.g5 .i.d6 1 1 ..txf6 O-O! is a nice trick; if White tries to win material by 12 i.xg7 "fixg7 1 3 lDxc6 bxc6 14 i.xc6, then the bishops and huge centre dominate following 14 . . . l:tb8, e.g., 1 5 i.b5 l:.b6 16 lDd2 e5) 10 . . . ..td6 11 lDdf3 (1 1 f4 0-0 12 lDdf3 lDe4) 1 1 . . . 0-0 with a pawn and active play; 12 llel falls short after 12 . . . lDe4 1 3 lDxc6 bxc6 14 ..td3 lDc5 ! . 8 g5! 9 i.g3 9 1.xg5!? fxg5 10 lDxg5 is an enter...
17
Play the French prising piece sacrifice that falls short after 10 . . . Vi'xe5 (or even 10 .. :i¥g7 1 1 "ikh5+ �e7 1 2 f4 h6 1 3 liJf3
18
10 ... fxe5 ! This takes over the centre (four pawns to none!). A safe alternative is 10 . . .f5!? 1 1 liJd2 (1 1 f3 liJh6) 1 1 . . . liJge7!? (1 1 . . .liJh6) 12 f4 liJg6 13 liJb3 ..It e7 14 liJxg6 hxg6 and a well timed . . . g5 (probably after . . . 0-0-0) will keep the advantage. 1 1 'ii'xg4 1 1 �b5 Ji.g7 12 'i'xg4 liJf6 13 Vi'xd4 O-O! 14 i..xc6 liJe4! threatens the bishop on c6 and . . . liJxg3. 1l ... liJf6 12 Vi'e2 ..Itd6 13 ..Itb5 0-0 14 liJd2 and 14 . . . e4 or 14 . . . liJa5!?, with the advantage in both instances. As in so many lines of the Advance Variation, Black's central pawn mass is the most important factor on the board.
Chapter Two Advance Variation : 5 i.d 7 . . .
1 e4 e6 2 d4 d5 3 e5 c5 4 c3 White maintains the pawn chain. This is played in over 85% of games after 3 e5. 4 .. .ttJc6 5 liJf3
This is the main line of the Ad vance Variation. White supports d4 before undertaking operations. As we shall see, Black can attack either the 'base' of the pawn chain at d4 or its 'head' at e5, or both. I will recom mend two systems, beginning with: 5 . . . Sl.d7 (this chapter) and 5 . . :iib6 (Chapter 3) . White has some little-known devia tions at this point, of which only one is moderately popular.
(a) 5 'iii g4? cxd4 6 cxd4 1lfb6 7 liJf3 liJh6! 8 'iiif4 (8 �xh6? 'iii xb2) 8 . . . liJf5+ Weiss-Halberditz, corr 1933; (b) 5 f4 'iiib 6 6 liJf3 liJh6 transposes to 4 f4; (c) 5 a3!? could transpose after 5 . . :iib6 6 liJf3 to 3 . 1 6 a3. Black also has 5 . . . a5!? (or 5 . . .f6!? 6 liJf3 "ikc7, e.g., 7 .tf4 cxd4 8 cxd4 g5 9 .tg3 g4 with the idea 10 exf6 'iii a 5+ 1 1 liJfd2 liJxf6=) 6 liJf3 cxd4 7 cxd4 liJge7 8 liJc3 (8 Sl.d3 liJf5 9 .txf5 exf5 10 liJc3 �e6 is a standard kind of position that tends to be okay for Black) 8 . . . liJf5 9 ..tb5 (9 g4 liJh4) 9 . . . ..td7 10 Sl.xc6 ..txc6 1 1 0-0 ..te7= ; (d) 5 liJe2 'iiib 6 (5 . . .f6 6 f4 fxe5 7 dxe5 liJh6 8 liJg3 .td7 9 .te2 'iiib 6 10 0-0 0-0-0 was comfortable for Black in Romero Holmes-Korchnoi, Pamplona 1990) 6 g3 (consistent, now that the knight on e2 blocks the bishop) 6 .. .f6 7 exf6 (7 liJf4?! g6 8 exf6 cxd4) 7 . . .liJxf6 8 .tg2 cxd4 9 cxd4 .td6 10 0-0 0-0 11 liJbc3 Sl.d7= Neretljak Brynell, Rodeby 1998; (e) 5 �e3!? is Kupreichik's move, to reinforce d4. It is the most popular of these options to 5 liJf3 and leads to original play: 19
Play the French gives 9 tt'lxe5 tt'lf6 10 0-0 i.d6 1 1 f4 0-0=) 8 tt'lxe5 tt'lf6 (8 . . . tt'lxe5 9 dxe5 tt'le7!?) 9 �d3 �d6 10 dxc5 .ltxc5 1 1 i.xc5 'iYxc5 1 2 tt'lxc6 bxc6 13 0-0 0-0= . 5 i.d7 Sometimes called 'Euwe's move', 5 . . . i.d7 was played by Greco in 1620(!), revived by Korchnoi in the 1970s, and is now a main line. It is a very flexible move that waits for White to commit to a plan. Black rea sons that . . . i.d7 will be part of almost any set-up. There are also several concrete move order advantages. If White plays 6 i.d3, Black can play 6 . . . cxd4 7 cxd4 Wb6 with the Milner Barry Gambit (a line that strong masters don't believe in). In the case of the move 6 �e2, there are many variations in which Black can hit the front of the pawn chain by . . .f6 (gen erally that is a less effective move with a bishop on d3) . The manoeuvre . . . tt'le7-f5 also has more effect after �e2 because White doesn't have the option of an immediate .ltxf5. We will also see the idea . . . tt'lge7-g6 become effective. Finally, Black can reply to 6 a3 with a more flexible strategy, e.g. , the queen might go to c 7 instead of b6. Playing .. .f6 is a theme in every case; this contrasts with the old strategy of attacking the base at d4, and has become a normal way of pro ceeding. I should mention that 5 . . . tt'lge7, with similarly flexible ideas, has also been used with considerable success in the last decade. Mter 6 i.d3, Black can play 6 . . . tt'lg6, but a great amount of theory is now at tached to 6 . . . cxd4 7 cxd4 tt'lf5. An other unique approach is 5 . . . tt'lh6, which has done rather well in the past two years. I think that it's fair to say that all of these ideas without the traditional 5 . . :iVb6 have taken some •••
(e l) 5 .. .lbh6 hasn't been played much, but I think that it is logical since White's bishop would have to move twice to capture on h6: 6 tt'lf3 (6 �d3 Wb6 7 1Wd2 tt'lg4 8 tt'lf3 tt'lxe3 9 fxe3 �e7 10 0-0, and now 10 . . . 0-0 1 1 b3!? f6 1 2 exf6 �xf6= or 1 0. . . c4 1 1 �e2 0-0 1 2 �hl f6 1 3 exf6 �xf6= Shaked-J.Watson, Los Angeles 1 995) 6 . . . tt'lf5 (6 .. :iVb6 7 'iYd2 �d7 8 �e2 tt'lf5 9 0-0 tt'lxe3 10 fxe3 �e7 unclear, Sengupta-Barua, Calcutta 2002) 7 g4?! tt'lxe3 8 fxe3 f6 9 exf6 gxf6 10 l:tgl 'iYb6 11 b3 cxd4 12 exd4 �d7 1 3 g5 �e7 14 �d3 O-O-o:i: B. Orton-J.Wat son, Las Vegas 1998. (e2) The normal move is 5 .. .'ii'b 6 6 1\Vd2 f6 (two good options are 6 . . . cxd4 7 cxd4 i.d7 8 tt'lc3 l:tc8 9 tt'lf3 tt'la5 10 licl �b4 11 �d3 tt'lc4 12 �xc4 l:txc4= Starostits-Glek, Baden-Baden 200 1 , and 6 . . .� d 7 7 tt'l f3 f6 8 �d3 fxe5 9 tt'lxe5 tt'lf6 10 0-0 �d6 1 1 f4 0-0= Ulibin and Lysenko) 7 tt'lf3 (an in structive example was 7 f4 tt'lh6!? 8 exf6 gxf6 9 Jt.e2 tt'lf5 ! 10 i.h5+ �d8 1 1 �f2 cxd4 1 2 cxd4 tt'lfxd4 1 3 tt'lc3 �c5 1 4 0-0-0 .ltd7 15 �b l tt'lb3?! 1 6 axb3 �xf2= Filimonov-Shorokhov, Smo lensk 2001) 7 . . . fxe5 (or 7 . . . i.d7 8 �d3 fxe5 9 dxe5 tt'lh6 10 0-0 tt'lf7 11 i.f4 �e7 12 l:.el 0-0-0=1= Westerinen Ulibin, Benidorm 1993; here Ulibin 20
Advance Variation: 5 . i.d 7 ..
of the sting out of the Advance Varia tion, although both sides willingly enter this line for the unbalanced play that it produces.
From the diagram we consider: 2.1 6 dxc5 2.2 6 a3 2.3 6 iLe2
Of course White can play almost anything, but the following are also logical: (a) As mentioned, 6 �d3 cxd4 7 cxd4 'i!Vb6 will generally lead to the Milner-Barry Gambit, discussed in Chapter 3 via 5 . . . 'i!Vb6 6 �d3 cxd4 7 cxd4 i.d7. Another response is 6 .. :�c7 with the idea 7 a3 c4 8 i.c2 0-0-0, having in mind either . . . f6 or . . .f5/ . . . h6/ . . . g5. Also played is 6 . . . .l:r.c8, e.g., 7 a3 (better is 7 dxc5, although it hasn't achieved much after 7 . . .iLxc5 8 0-0 liJge7 9 �g5 h6 10 �h4 g5 1 1 iLg3 liJf5= with the idea 12 �xf5?! exf5 1 3 'i¥xd5 liJ e 7 and . . . f4; o n 7 0-0, 7 . . . cxd4 8 cxd4 liJb4= wins the two bishops) 7 . . . cxd4 8 cxd4 'i!Vb6 9 �c2 (9 0-0 liJxd4 10 liJxd4 'ii'x d4 11 liJc3 of Klausner-Rivas, Marbella 1 982 com pares very poorly for White with a main-line Milner-Barry gambit after 1 1 . . . 'ii'x e5 12 l:tel 'i!Vb8) 9 . . . liJxd4! 1 0
liJxd4 iLc5 1 1 liJe2 iLxf2+ 12 'i£tfl f6 Bjornsson-M.Hansen, Gausdal 1 994, with two pawns and a powerful at tack for the piece; (b) 6 liJa3 f6 (or 6 . . . cxd4 7 cxd4 iLxa3 8 bxa3 unclear, e.g., 8 . . :iVa5+ 9 'iVd2 l:tc8) 7 liJc2 fxe5 (7 . . :iib6 with the idea . . . 0-0-0) 8 dxe5 'ii'c 7 9 i.f4 0-0-0 (or 9 . . . liJge7 10 h4 h6 1 1 'ii'e 2 0-0-0 unclear, Balinas-Espig, Odessa 1976) 10 �e2 liJh6 1 1 'ili'd2 liJf7 12 'ii'e 3 �e7 planning . . . g5, Dominguez Topalov, Zaragosa 1992; (c) 6 iLe3 is logical and related to the Kupreichik idea 5 iLe3 above. There are several reasonable an swers, including: (c l) 6 . . . cxd4 7 cxd4 liJge7 (7 . . .f6!?) 8 liJc3 liJf5 9 iLd3 liJxe3 10 fxe3 'i¥b6 (10 . . . g6!? 1 1 0-0 iLh6 12 'ii'e 2 0-0 in tending . . .f6) 1 1 'ili'e2 iLe7 12 0-0 0-0= Roeder-Weih, Bundesliga 1986; (c2) A simple but previously ne glected idea was 6 . . . liJge7 7 dxc5 liJf5! 8 �d4 liJfxd4 9 cxd4 b6! 10 cxb6!? (10 �b5 bxc5 11 'iVa4 'i!Vb6;) 1O ... 'iIi'xb6 1 1 'ii'd 2 Iib8 1 2 liJc3 (12 b 3 iLb4 1 3 liJc3 'ii'a 5) 12 .. :iWxb2 13 .ub I 'ili'xd2+ 14 <;itxd2 i.b4 15 i.d3
21
Play the French An important idea from Svesh nikov. 6 dxc5 used to be a main line after 5 . . . �d7 and helped to revive the Advance Variation, but it is doing rather poorly these days. Black de velops quickly and White's positional idea of securing the centre (normally by b4-b5) is countered by queenside play and .. .£6 ideas. This is crucial for understanding the apparently slow 5 . . . i.d7, so I'll devote some attention to it. 6 ... i.xc5 7 b4 This thrust was the main reason for the popularity of 6 dxc5 in modern times. Since 7 i.e2 instead concedes the centre after 7 .. .£6, the alternative is 7 i.. d 3, which is less weakening and probably sounder:
14 0-0 e5= Sveshnikov-Savon, Lvov 1978) 9 b5 ltJxe5 10 ltJxe5 fxe5 1 1 "iVh5+ Wf8 1 2 i.a3+ ltJe7 13 'ii'x e5 Wg8 14 ltJd2 ltJg6 15 'i'g3 "iVf61=i= with ideas like . . . e5, .. :ikxc3, and . . .ltJf4, Ciem niak-Schmidt, Bundesliga 1993; (b) 7 ... ltJge7 8 0-0 (8 b4 i.b6 9 b5 ltJa5 transposes to 7 b4; 8 i.f4 ltJg6 9 i.g3 f51? 10 h4 'i'c7 1 1 b4 i.b6 12 "iVe2 h51 13 0-0 ltJce7= - Knaak) 8 . . . ltJg6 9 "iVe2 (9 .l:tel 'JlIc7 10 i.. xg6?1 fxg61 1 1 i.. f4 0-0 1 2 i.. g 3 ltJe7 1 3 ltJbd2 ltJf5 14 ltJb3 i.. a 4=i= Sveshnikov-Balashov, USSR Ch 1976; a good example of the . . . fxg6 idea in the Advance Variation, very often effective if White's bishop ends up stranded on g3) 9 . . . "iVc7 10 l:1.e l 0-0 (10 ... f61? 11 exf6 gxf6 12 c4 d4 13 a3 Pachman-Voicelescu, Bucha rest 1953, and 13 . . . a5 is unclear) 1 1 i.g5 (1 1 h 4 f61; 1 1 b 4 i.b6 1 2 a4 a 5 1 3 b 5 ltJce7 14 �a3 �c5 with play against the White queenside) 1 1 . . .£6 12 exf6 gxf6 1 3 �h6 ltJf4 14 i.xf4 �xf4= Paredes-Goldschmidt, 1992. 7 .. i..b 6 8 b5 8 i.d3 is effectively answered by 8 ... ltJge7, but also good is the imme diate liquidation of White's centre by 8 . . .f6 9 b5 (9 i.f4 fxe5 10 ltJxe5 "iVf6 1 1 ltJxd7 iLxf2+ 1 2 c;;t>xf2 liVxf4++ G.Lee Cooper, Ayr 1978) 9 . . . ltJxe5 (9 . . . ltJa5 has also been played) 10 ltJxe5 fxe5 11 'iVh5+ c;;t>f8 12 i.. a 3+ (12 'i'xe5 ltJf6 1 3 0-0 l:1.c8 14 i.. g 5 'itf7= G.Wall A. Summerscale, Torquay 2002) 12 . . . ltJe7 1 3 "iVxe5 c;;t> g8 (or 1 3 . . . Wf7 14 'iVh5+ ltJg6 and Black with the centre and c-file stands well) 14 ltJd2 ltJg6 1 5 "iVg3 "iVf6 16 0 - 0 "iVxc3 1 7 ltJb3 'JlIe5=i= Bastian-Luther, Binz 1995. 8 . ltJa5 9 iLd3 .
(a) 7 .. .£6 8 b4 (8 �f4 g51? intends . . . g4 and provokes 9 ltJxg51? fxg5 10 'iVh5+ Wf8 11 i.. xg5 i.. e 71 12 i.h6+?1 ltJxh6 13 'iVxh6+ Wg8+ ; 8 "iVe2 fxe5 9 ltJxe5 ltJxe5 10 'iVxe5 ltJf6 1 1 0-0 0-0 and White has his usual problems with both Black's development and centre pawns: 12 i.. e 3 i.b6 13 i.d4 i.. c 7 14 'i'g5 h6 15 'i'h4 e5 16 i.c5 e4 17 i.e2 lU7 with a clear advantage, Campora-Speelman, Menor 1 994) 8 . . . i.b6 (8 . . . i.e7 is also fully playable after 9 b5 ltJxe5 10 ltJxe5 fxe5 1 1 'iVh5+ Wf8 1 2 "iVxe5 i.f6 1 3 "iVd6+ ltJe7 22
..
(see following diagram) 9 ...ltJ e7 The main move, but not necessarily
Advance Variation: 5. . . i.d 7 best, since White has nothing special versus Black's alternatives, both em phasising the open c-file:
(a) 9 . . . lLlc4!? 10 a4 "i8c7 1 1 "i8e2 a6 12 bxa6 �xa6 13 0-0 lLle7= Kharlov Dreev, Moscow 199 1 ; (h) 9 . . . iVc7! is a multi-purpose move, taking over the c-file, hitting e5, and preparing to castle queenside: (hI) 10 iVe2 lLle7 1 1 h4!? was Khar lov-Kramnik, USSR 1 99 1 ; but it's not clear what White's idea was after 1 1 . . .lLlc4, e.g., 12 h5 !? (12 a4 a6! 1 3 bxa6 �xa6) 1 2 . . . i.xb5 1 3 h 6 gxh6! 1 4 i.xh6 0-0-0 etc.; (h2) 10 O-O!? lLle7 (again, 10 . . . lLlc4! hits b5 and e5; since 11 i.xc4 iVxc4 seems good for Black and 1 1 a4 a6! takes the initiative, play might pro ceed 1 1 �e l!? i.xb5 12 lLld4 i.d7, when 13 iVh5 is unclear but 13 "i8g4? ! lLlxe5! 14 iVxg7 lLlxd3 15 iVxh8 0-0-0 1 6 �dl lLlxc l 1 7 �xc l iVf4 is too strong, e.g., 18 lLla3 lLlf6 19 iVg7 lLlg4) 1 1 a4 lLlg6 12 �el i.c5!? (12 . . . lLlc4= ; 12 . . . f6!?) 1 3 �a2 0-0-0 14 i.e3 i.xe3 15 �xe3 f6 16 �ae2 'ii?b 8 unclear, Sveshnikov-Popovic, Palma de Mal lorca 1989. 10 0-0 �c8 Crazy play followed 1 0 ... lLlg6 1 1 �el (1 1 i.a3) 1 1 . . . 0-0 1 2 i.g5 iVe8 1 3 a4!? �c8 14 h4 h6= , when White got
too ambitious with 15 h5? hxg5 1 6 hxg6 fxg6 1 7 �a2 iVfl 1 8 iVc2 iVf4 1 9 i.xg6 g4 2 0 lLlh4 g3 ! 2 1 i.h7+ 'ii?fl 22 iVg6+ rJJ e 7 23 iVxg7+ 'iit> d 8-+ Davies Benson, Port Erin 1 998.
1 1 a4 Perhaps the most logical move is 1 1 iVe2, to reinforce e5 while covering c4. But Black has no particular diffi culties after 1 1 . . .h6 (or 1 1 . . .0-0, since 12 i.xh7+?! rJJ xh7 13 lLlg5+ 'ii? g 6 14 iVd3+ lLlf5 15 g4 �c4! 16 gxf5+ exf5 is fme, as given by Psakhis) 12 g3!? iVc7 13 rJJ g2?! lLlc4 14 a4 lLlg6! 15 �el (psakhis gives 15 i.xg6 fxg6 16 iVc2 O-O! 1 7 i.xh6!? �xf3! 18 'iit> xf3 lLlxe5+ 19 'iit> g2 gxh6 20 f4 lLlfl 21 iVxg6+ 'iit>f8�) 15 . . . 0-0= Giaccio-ARodriguez, Buenos Aires 2000. 11 ...lLlg6 12 �e1 12 i.a3 i.c5= was the game Svesh nikov-Balashov, Lvov 1978. Mter 12 �e l, White intends �a2-e2 to rein force e5. 12 ... lLlc4 1 3 �a2 iVc7 14 �ae2 Now 14 . . . a6 15 bxa6 bxa6 has been tried several times, but it's also safe enough to play 14 . . . 0-0. Then 15 h4!? can be met by 15 . . .f6! 16 i.xg6 hxg6 17 exf6 (17 �f4 a6!? 18 bxa6 bxa6 19 g3 iVc5=) 1 7 ... �xf6 and Black has very active pieces which make up for his weakened pawn structure.
23
Play the French 2.2 6 a3
A very popular move. I think that it is less dangerous for Black than the analogous 5 . . . VWb6 6 a3. For example, there are more immediate threats against d4 after 5 . . . 'iWb6 6 a3, but Black's queen sometimes gets in the way and has to retreat to c7 or even d8. Also, 5 . . . �d7 is more flexible in lines like the one below where Black attacks e5 by . . .f6. Then he might want to play . . . Wic7 and perhaps . . . 0-0-0. Finally, after the continua tion 5 . . . �d7 6 a3, the b6-square can be useful for a knight (i.e . . . . liJge7-c8b6) . 6 ... f6!? A direct approach suggested in early editions. By attacking the cen tre, 6 .. .f6 tries to prove that 6 a3 is too slow. It has been scoring well in master practice and offers double edged play. Black has several other valid alternatives which have also gotten good results, including 6 . . . c4 and 6 . . . a5. Because it is instructive, I'll use 6 . . . liJge7 as a second recom mendation. This flexible move (reserving the choice between . . . liJf5 and . . . liJg6) can resemble a number of other Advance Variation lines. An abbreviated over view of the key lines: 24
(a1) 7 �d3 cxd4 8 cxd4 liJf5 (8 . . . 'iWb6 9 �c2 liJf5 10 i.xf5 exf5 1 1 liJc3 i.e6 has also proven playable) 9 i.xf5 exf5 10 liJc3 i.e6 1 1 b4 i.e7 12 h4 h6 1 3 �f1 :c8 14 'i'd3 'i'd7 1 5 i.f4 a6! 1 6 liJd2 liJa7 1 7 ':h3 'i'c6 18 :'c 1 liJb5 19 liJdb 1 i.d7 Y2-Y2 Bruno Minzer, Buenos Aires 1993; (a2) 7 b4 is the consistent and most important try. Then 7 . . . cxd4 8 cxd4 liJf5 is a type of position that players on both sides of the Advance Varia tion should study. It is very similar to that with 5 . . . 'iYb6, but the queen might want to stay on d8 to save a tempo and support moves like . . . liJh4 or even . . . i.e7 and . . . g5. Here are some examples that illustrate the main themes in this often-arising structure. They show why Black has a relatively easy time with it: (a2 1) 9 i.b2 'iWb6 transposes to 5 . . . 'iWb6 6 a3 lines, which is not bad but also unnecessary. The more logi cal 9 . . . i.e7 and 9 . . . .l:.c8 are considered equalisers, and the ambitious thrust 9 . . . b5!?, intending . . . a5, has done sur prisingly well. On 10 �xb5, 1O . . . liJxe5 1 1 dxe5 i.xb5 12 liJc3 i.c4 is an effec tive course, for example, 1 3 g4!? liJe7 1 4 h4 h5!? (14 . . . liJg6) 15 liJg5 �d7 16 gxh5 liJf5+ M.Petrov-Degraeve, Metz 2002;
Advance Variation: 5. . . .i.d 7 (a22) Similarly, the line 9 �e3 b5!? (9 ... ttJxe3 10 fxe3 g6!? might be con sidered, intending . . . .i.h6, . . . 0-0, and . . .f6) 10 ttJc3 a5! 1 1 ttJxb5 axb4 12 a4 JJ.. e 7 13 .J.d3 0-0 14 0-0 ttJa5 gave Black good play in Van der Hoeven Michiels, Amsterdam 200 1 ; (a23) 9 g4!? i s sometimes too loos ening. Here 9 . . . ttJh4 (9 . . . ttJh6 is also fine, since 10 JJ.. xh6!? gxh6 leaves Black with two bishops and open lines; then . . . .l:tc8 and in some cases .. Jlg8 and . . . h5 will follow) 10 ttJbd2 ttJxf3+ (10 . . . Irc8!) 1 1 ttJxf3 'iVb6 12 �e2 �e7 1 3 .tb2 lIc8 was equal in Shmyrina-Paulet, Halkidiki 200 1 ; (a24) 9 ttJc3 .l:tc8 (Black can also launch a standard kingside attack after, e.g., 9 . . . �e7 10 .i.b2 .l:'tc8 1 1 iLe2 h5 12 0-0 g5!, Yilmaz-Dreev, Berlin 1991) 10 �b2 ttJh4! 1 1 ttJxh4 (1 1 ttJa4?! ttJxf3+ 12 'iWxf3 ttJxe5) 1 1 . . ."iVxh4 12 ttJa2 (12 ttJe2? ttJxb4-+) 12 . . .'iWe4+!? (or simply 1 2 ... JJ.. e 7) 13 'iWe2 'iWg6!? lliescas Cordoba-Speel man, Linares 1 992. This is probably about equal, although White is rather disorganised.
7 JJ.. d 3 The most challenging move, trying to exploit the kingside looseness cre ated by .. .f6. Nevertheless there are a few other reasonable options:
(a) 7 b4!? follows up on White's 6th. Because a3 and b4 take time, how ever, Black is able to break up White's centre and to develop quickly: 7 . . . fxe5 8 b5 (8 ttJxe5 ttJxe5 9 dxe5 'iWc7 and . . . 0-0-0; 8 bxc5?! e4 9 ttJg5 ttJf6:j:; 8 dxc5?! e4 9 ttJd4 ttJf6:j: Pop chev-Dolmatov, Polanica Zdroj 1987; fmally, after 8 dxe5 "iVc7 9 1i.f4 ttJge7 10 .td3, 10 . . . g6!? is particularly at tractive since Black's pieces come out smoothly, e.g., 1 1 'i'd2!? �g7 12 0-0 0-0 1 3 �g3?! cxb4 14 cxb4, and now 14 . . J%xf3 ! is obviously sound but Black took another effective route via 14 . . . ttJf5 1 5 .l:te1 ttJxg3 16 hxg3 .l:txf3! 1 7 gxf3 �xe5 1 8 .l:'ta2 ttJd4 19 'iWe3 lIfB 0-1 in Castaneda-Morales, Guay mallen 200 1) 8 . . . ttJxd4!
This sacrifice for three pawns has been played for over a decade and looks sound. Black gets a mobile cen tral pawn mass after 9 cxd4 (9 ttJxe5 ttJf5 10 'i!Vh5+ g6 1 1 ttJxg6 ttJf6 12 'i!Vh3 Irg8 13 ttJxf8 'it?xf8 14 .td3 ttJe4!? [bet ter is simply 14 ... 'i'e8] 15 ttJd2? ttJg5 16 'i!Vh5 'iWf6 17 .tb2 c4 18 .tc2 'i'e5+ 19 'i'e2 'i'xe2+ 20 'it?xe2 .txb5 21 a4 .J.c6 22 h4 ttJe4 23 ttJxe4 dxe4 24 g3 :d8 Y2 - Y2 Movsesian-Luther, Cappelle la Grande 1998 - Black was a little better throughout) 9 . . . exd4 with dy namic play, e.g., 10 �f4 ttJf6 1 1 ttJbd2 25
Play the French 'iib 6 (1 1 . . :�a5 12 �e2 liJe4 13 0-0 liJc3 14 'iVe l i.e7=) 12 liJe5 �a5 1 3 g4 (13 liJxd7 d 8 strongly favours Black) 9 . . . liJf7 10 l:tel Movsesian-Luther, Groningen 1 998. Now I like simply 1O . . . fxe5 1 1 liJxe5 (1 1 dxe5 g5 intends both . . .iLg7 and . . . g4, e.g., 12 b4 i.g7 13 �b2 g4 14 b5 ! gxf3 15 bxc6 �xc6 16 'iVxf3 �6 and . . . 0-0 or . . . 0-0-0) 1 1 . . .liJcxe5 ...
26
12 dxe5 i.c5 13 i..f4 1lib6! 14 l:1.e2 i.b5 and Black has an easy game; (h) The most popular option is 7 .. .'fic7: 8 0-0 (8 exf6?! liJxf6 9 0-0 c4 10 iLc2 �d6 1 1 J:l.el 0-0-0 12 �e2 l:I.dfB 13 liJe5 i.e8 14 f4 g5 ! 15 fxg5 i.xe5 16 dxe5 i.h5+ Branford Martin, Westergate 1992; 8 i..f4 c4 9 i.c2 0-0-0 10 liJbd2 liJh6 1 1 i.g3 f5 12 i.f4 liJf7 1 3 h4 i.e7 14 'iNe2 g6 1 5 g3 h6, preparing ... g5, Kodric-Soln, Bled 2001) 8 . . . 0-0-0 (or 8 .. .fxe5!? 9 liJxe5 liJxe5, when 10 i.f4?! 'iib 6 1 1 i.xe5 'iNxb2 12 'iNh5+ 'iit d 8 doesn't give White compensation) .
At this point White has three in structive continuations: (h I) 9 'uel c4 (9 . . . h6!?) 10 .i.f1 !? liJge7 1 1 liJbd2 fxe5 (1 1 . . .liJg6 12 exf6 gxf6 1 3 b3 cxb3 14 'iVxb3 was unclear in Mek-Schaefer, Budapest 1993) 12 liJg5 (12 liJxe5 liJxe5 1 3 dxe5 liJg6 14 liJf3 i.c5:j:) 12 ... exd4! 1 3 liJf7 dxc3 14 bxc3 liJg6 1 5 liJxh8 liJxh8:j: Genin Epishin, Leningrad 1 980; (b2) 9 �f4 c4 10 i.c2 h6! 11 h4 (1 1 .tg3 f5 !:j:, since 12 .txf5? exf5 1 3 e6 f4 1 4 exd7+ .uxd7 1 5 �h4 g5 is decisive) 1 1 . . . iLe8! (1 1 . . .g5!? 12 1Lh2 f5 13 hxg5 i.e8 with compensation, Buchal Fischer, Germany 1997) 12 b3?! cxb3 13 i.xb3 i.h5 14 liJbd2 fxe5 1 5 dxe5 i.c5+ Adams-Epishin, Ter Ape1 1 992;
Advance Variation: 5. . . i.d 7 (b3) 9 'iWe2 h6 10 b4 c4 1 1 .i.c2 f5 (intending . . . g5) 12 tiJh4 .i.e8 13 f4 .i.e7 14 tiJxf5?! (overambitious, but 1 4 tiJf3 allows 14 . . . �h5 when Black will soon attack by . . . g5, at least equalis ing) 14 . . . exf5 15 .i.xf5+ �b8 16 'iWg4 g5! 17 fxg5 hxg5 18 .i.xg5 ..th5 19 1li'g3 ..txg5 20 'ilVxg5 tiJge7 (White has three pawns for the piece but Black has two open fIles aimed against the king) 21 tiJd2 .l:tdg8 22 'ilVe3 tiJxf5 23 l:1.xf5 'iVh7 24 :f6 .i.e2 ! 25 h3 .i.d3 26 'it>h2 tiJe7 27 tiJf3 tiJf5 28 'iWf4 'iit a 8 29 l:1.g1?! 'iWh5! 30 e6 .i.e4 3 1 l:1.f1 l:1.g3 32 l:1.xf5? 'iWxh3+! 0-1 Grischuk-Short, Reyjavik 2000.
8 dxe5 8 tiJxe5 tiJxe5 (8 . . . tiJf6!? 9 .i.f4 'ilib6 10 tiJxc6 bxc6=) 9 dxe5 and now: (a) 9 . . . g6 10 c4 (10 0-0 'ilVc7 1 1 f4 c4 12 .i.c2 i.c5+ 13 �h1 tiJe7 1 4 tiJd2 0-0-0 15 b4 cxb3 16 tiJxb3 ..tb5= Afek Hebert, Paris 1995) 1 O . . :iic 7 1 1 0-0 i.g7 12 cxd5 i.xe5 1 3 dxe6 i.xe6 1 4 �a4+ � d 7 1 5 �b5 0 - 0 - 0 16 tiJ c 3 a6 1 7 i.xd7+ l:1.xd7 Y2 - Y2 Charbonneau Barsov, Montreal 2002; (b) I suggested 9 .. :1Wh4!? in the previous edition but it remains un tried. The queen move prevents 'iVh5+ and prepares . . . tiJh6 and . . . .i.c6, or . . . c4 with . . . .i.c5. For example, 10 tiJd2 (10 g3 "iWh3 1 1 .i.e2 h5) 10 . . . tiJe7 1 1
tiJf3 'i!Vh5 (hardly clear, but Black looks well off) 12 .i.e2!? 'iff7! 13 tiJg5 �g8 planning . . . h6, . . . 0-0-0, and . . . i.e8-g6. 8 'iWc7 ...
This position has been extremely kind to Black. Also satisfactory are 8 . . . c4 9 .i.c2 .i.c5 and 8 . . . tiJh6 9 0-0 tiJf7, when the strength of a knight on f7 shows itself, e.g., 10 I!e1 g5 1 1 h3 .i.g7 12 tiJbd2?! (12 c4 tries to break up Black's centre, but would cost a critical tempo after 12 . . . dxc4 1 3 i.xc4 h5!) 12 .. :iVe7 13 c4 0-0-0 14 l:1.b 1 h5 and Black's pressure on e5 and the kingside proved decisive in Druon Boissel, corr 1993. 9 'iWe2 The idea of . . . tiJh6-f7 is a major theme since the king can escape via . . . 0-0-0, whereas .i.xh6 is awfully risky due to Black's attack down the g-fIle and pressure on e5: (a) 9 0-0 0-0-0 (9 ... g6 is an increas ingly popular idea in such positions, e.g. , 10 :tel .i.g7 11 1li'e2 tiJge7 12 b4 c4 1 3 .i.c2 a6!? 14 a4 0-0 1 5 tiJa3 tiJf5 16 .i.g5 :f7:� with doubling on the f fIle, Ocampo-Munoz, Guiines 1998) 10 l:te1 tiJh6 1 1 i.xh6 gxh6 12 c4 d4 13 tiJbd2 Bosboom-Langeweg, Eind hoven 1988, and here 13 . . . .i.g7 14 'iVe2 i. e8 intending . . . i.h5 o r . . . �b8 27
Play the French and . . . i.g6 was fully satisfactory; (b) 9 i.. f4 0-0-0 10 ltJbd2 ltJh6 1 1 'iVe2 ltJrl 1 2 O-O-O?! (12 0-0 i.e7 1 3 h4 c4 14 i.c2 l:.df8 and Black is at least equal) 12 . . . c4 13 i.. c 2 b5!? (13 . . . i.c5 ! looks stronger) 1 4 .:the 1 a5 1 5 ltJg5 ltJxg5 16 i.xg5 :e8 intending . . . b4 with a meaningful attack, S.Lalic King, London 1 989. 9 ...ltJ h6
has a drawback) 16 g3 (White has similar problems after 16 h5 g5 and 16 i.g3 g5) 16 . . . g5 17 hxg5 ltJxg5=i=. 2.3 6 i.. e 2
The main line. White develops quickly, planning 0-0 and l:te l . I will propose two solutions:
10 0-0 (a) 10 i.f4 ltJrl 1 1 h4 (clamping down on . . . g5) 1 1 . . .0-0-0 12 ltJbd2 i.e7 (12 ... c4 1 3 i.. c 2 i.c5 develops more smoothly) 13 b4 c4 14 i.c2 .:tdf8 1 5 g3 g6 16 0-0 .:thg8! (the idea is . . . h6 and . . . g5) 17 b5 ltJa5 18 a4 Panagopoulos Sigalas, Halkidiki 2002; now 1 8 . . . h6! would have led to a clear advantage; (b) 10 c4 is at best harmless, e.g., 1 0 ... dxc4 (or 10 ... ltJrl 11 cxd5 ltJcxe5) 1 1 i.xc4 ltJf5 12 i.f4 (12 0-0 ltJcd4 1 3 ltJxd4 ltJxd4) 12 . . . ltJcd4 1 3 ltJxd4 ltJxd4 14 'i¥d1 g6!? 15 0-0 �g7 16 .l:te1 0-0 1 7 i.g3 i.. c 6 with wonderful pieces and . . . ,Uad8 on the way. After 10 0-0, Mek-Haak, Vlissin gen 2002 continued 10 . . . ltJrl 1 1 ]::t e l c4 12 i.. c 2 i.. e 7 1 3 h 4 0-0-0 (or simply 1 3 . . . 0-0, since the moves . . . i.c5 and .. :iVb6 will tie down White's pieces down) 14 i.f4 .l:r.dg8!? 15 ltJbd2 'iiV d 8 (Now h4 is a target and every defence
28
2.31 6 ...ltJge7 2.32 6 ... f6 2.31 6 ...ltJge7
My favourite. This knight is flexi bly placed to assist in various attacks on White's centre. We look at: 2.3 1 1 7 0-0 2.312 7 ltJa3
Mter 7 dxc5 ltJg6, Black will win back the pawn on e5 with a nice cen tre, e.g., 8 i.e3 ltJcxe5 9 ltJxe5 ltJxe5 10 0-0 .i.e7 (10 . . . .tc6!? with the idea 1 1 f4?! ltJc4!) 1 1 .i.d4 (1 1 f4 ltJg6!? 12 f5 exf5 13 'ilixd5 f4! 14 i.xf4 ltJxf4 1 5 .l:txf4 0-0 with great compensation) l 1 . . .'ilic7 12 f4 ltJg6 (or 12 . . . ltJc6) 1 3 f5 ltJf4!=i= . 2.3 1 1 7 0-0 ltJg6 A move considered anti-positional
Advance Variation: 5. . . iJ.. d 7 for a century or so, simply because it attacks the 'wrong' part of the chain, and because the knight is subject to harassment. It is now a favourite of M.Gurevich and P.Nikolic and has been very successful in recent years. The idea is simple: . . . f6 and the de struction of White's pawn chain. 7 . . . lLlf5 is also played and worth considering if you don't like 7 . . . lLlg6. It intends . . .cxd4 with a standard at tack on the d4 square. White can try to exploit this move order by S �d3!?, e.g., S ... cxd4 (S ... lLlh4!? looks quite reasonable) 9 �xf5 exf5 10 lLlxd4 (10 cxd4 �e7 11 lLlc3 �e6 is a standard position that I consider equal) 10 . . . �e7 1 1 .l:lel (1 1 'iVb3 'iic 7 12 'iix d5 �xe5 13 'i!Vxe5 lLlxe5 with the bishop pair, Hendriks-M.Gurevich, Amsterdam 200 1) 1 1 . . .0-0 with the idea 12 'i!Vb3 �a5; Black has satisfac tory play. 8 g3
A slightly unconventional move in tending h2-h4; this already acknowl edges that simple development won't prevent Black's plan from being real ised, for example: (a) S ii.. e 3 iJ.. e 7 (Black has also scored very well after S . . . cxd4 9 cxd4 f6) 9 lLla3 (9 lLlbd2 is more common, when after 9 . . 0-0 10 lLlb3 c4!? 1 1 .
lLlbd2, Black has the normal 1 1 . . .f6 or 1 1 . . .b5) 9 . . . 0-0 10 lLlc2 cxd4 1 1 cxd4 f6 12 exf6 ..ixf6 with plenty of activity; (b) S ..id3 ..ie7 9 ..ie3 'iVb6 10 'ii d2 c4 1 1 ..ie2 0-0 12 g3 (12 'iic 2 'iic 7 1 3 lLlbd2 f6) 12 . . . 'iid S!? 13 lLl e l f5 1 4 f4 (14 exf6 �xf6+) 14 . . . b5 1 5 lLlg2 .l:lbS (or 15 . . . b4) 16 b4 a5 17 a3 .l:laS+ Jonkman-Hertneck, Saint Vincent 2000. 8 f6 A fully playable option is S . . . �e7 9 h4 (the point of S g3, but . . . ) 9 . . . 0-0! 10 h5 lLlhS ...
This offside knight (having no moves!) is now actually well placed in the corner, because it supports the attack against the head of the pawn chain by .. .f6 and . . . lLlf7. Korchnoi first played a similar idea some years back via 9 . . cxd4 10 cxd4 0-0 1 1 h5 lLlhS 12 h6 g6 13 lLlbd2 f6. The inter polation of 9 . . . cxd4 is perhaps inaccu rate here because White gains the extra option of lLlc3, but that's not clear. From the diagram, play contin ues 1 1 h6 g6 (actually, 1 1 . . .g5!? seems to work out well, with ideas like . . . f6 and . . . lLlg6 or . . . lLlf7, e.g., 12 �e3 cxd4 13 cxd4 f6 14 exf6 ..txf6 1 5 lLlc3 lLlf7) 12 dxc5 f6 (12 . . . ..txc5 13 c4 f6 is an other idea that looks fine, e.g., 14 cxd5 lLlxe5 1 5 lLlxe5 fxe5 16 lLlc3 'iVb6 29
Play the French 17 ttJe4 �d4=i=) 13 exf6 (trying to win a pawn; 1 3 c4 will probably transpose into the last note) 13 . . . ..txc5 14 ..tg5 ttJf7 15 .th4 Wic7. White is a pawn up, but Black is active with a potentially mobile centre: 16 'iVd2 (16 ttJbd2? g5! intending 17 ttJxg5 ttJxg5 1 8 i.xg5? 'iUxg3+) 16 . . . ttJce5 17 ttJxe5 'iUxe5 1 8 'itJh2 Timman-P .Nikolic, Amsterdam 1999, and now 18 . . . g5 ! would have been good for Black due to 19 f4 'iUxf6 (or 19 . . . gxf4 20 gxf4 Wif5) 20 fxg5 (20 i.xg5? ttJxg5 2 1 fxg5 'ilVxfl ! 22 i.xfl lIxfl wins) 20 . . .'iVe5 2 1 i.d3 'iUe3 ! 22 g6 ttJe5! 23 gxh7+ �h8 24 Wixe3 i.xe3 and Black will recover both pawns with advantage.
9 i. d3 Black is able to break up the centre after 9 h4 fxe5 10 h5 e4! 1 1 ttJh2 (1 1 hxg6 exf3 and 12 ..td3 'ief6 1 3 gxh7 c4 14 i.c2 0-0-0 1 5 ttJd2 g6 16 ttJxf3 lIxh7 with good play, or 12 i.xf3 hxg6 13 lIe1 'iVf6 14 dxc5 ..txc5 1 5 'iVxd5 0-0-0 16 'iVe4 e5 1 7 ..te3 i.f5 18 'iUa4 i.xe3 19 .l:txe3 e4 20 i.g2 'iUg5+) 11 . . . ttJge7 (threatening . . . cxd4 and . . . ttJf5) , e.g., 12 dxc5 ttJf5 ! (or 12 . . . e5 with the idea 13 ttJg4 ..te6 14 b4 d4!=i=) 13 g4 (13 b4 ..te7 14 ttJa3 0-0 15 i.f4 a5 16 b5 ttJa7 1 7 b6 ..txc5!+) 1 3 . . . ttJh4 14 i.e3 (14 b4 ..te7 15 h6 g6=i= with ideas like . . . 0-0 and . . . ttJe5) 14 . . :�c7
30
15 ttJa3 a6 16 c4 d4 (16 .. .'i¥e5+) 1 7 i.xd4 ttJxd4 18 'ilVxd4 ..txc5 19 'iUxe4 0-0-0 and Black is overwhelming White with the threats of . . . Wig3+ and . . . .tc6. 9 cxd4!? Good, but 9 . . .fxe5! looks very strong and is likely the best move: ...
(a) 10 ttJxe5 ttJgxe5 1 1 dxe5 ttJxe5 (or 1 1 . . .g6) 12 Wih5+ ttJf7 13 lIe1 (13 ..txh7 Wif6) 1 3 ... Wif6 with the idea . . . 0-0-0. (b) 1 0 ..txg6+ hxg6 11 dxe5 (1 1 ttJxe5 ttJxe5 12 dxe5 ..tb5 13 .l:.e1 'i'd7 14 'i'c2 'iUf7; 1 1 'iVc2? e4) 1 1 . . .i.e7!? (contemplating ... g5; the alternative is 1 1 . . :i¥c7 12 ..tf4 0-0-0 with the idea 13 ttJg5 i.e7! 14 ttJf7 g5 15 ttJxd8 .txd8 16 i.e3 ttJxe5 17 ttJa3 'iVb6!? 1 8 b4 g4!=i=) 12 h 4 ( 1 2 c 4 dxc4 1 3 ttJbd2 ttJb4 14 ttJxc4 0-0; 12 'iVc2?! 0-0 1 3 'itJg2 g 5 14 h 3 i.e8 1 5 ttJa3 d4; 12 ttJa3 g5 13 g4 0-0 14 'itJg2 ..te8 15 Wid3 d4=i=) 12 . . . 0-0!? 13 ..tf4 (13 ttJbd2? lIf5 14 Wie2 Wic7 15 .:tel .l:.af8+) 13 .. J:rxf4!? (13 ... 'iVb6 14 'ilVc2 .:tf5 1 5 ttJbd2 lIaf8=i= is easier, but less fun) 14 gxf4 'iUf8, e.g., 15 Wid2 Wif5 16 �h2 Wih5 ! 17 c;t;g3 .td8! with the idea . . . ttJe7-f5. 10 cxd4 Or 10 exf6 Wixf6 1 1 cxd4 �d6 with good activity. 10 ttJb4 ...
Advance Variation: 5. . . i.. d 7 pawn chain. Black often plays the reasonable alternative 7 . . . cxd4 8 cxd4 lDf5 9 lDc2
1 1 i.. e 2!? White has two moves that are probably better: (a) 1 1 i..xg6+ hxg6 12 a3 lDc6 1 3 lDh4!? (13 �d3 lDe7!? 14 lDc3 ! , when 14 . . . g5 with the idea .. JWb6 is un clear) 13 . . .';;?;? f7 (13 . . . g5!?) 14 �d3 g5!? This is double edged, one idea being 15 �g6+ g2 lDxc l won the game. Instead, the line 1 7 i.. e 2 lDxe5 (17 . . . 0-0-0!? 1 8 'il¥c l !) 18 lDc3 would have tried to keep Black's king in the centre, although I'd still prefer to be the second player after the move 1 8 . . . �f6. .
.
•••
2.312 7 lDa3 lDg6 Again aiming at the head of the
Note 5 . . . cxd4 6 cxd4 lDge7 7 lDa3 lDf5 8 lDc2 i.d7 transposes if White plays 9 i.e2, but with that order White also has 9 i.. d 3, thinking about a capture on f5. Then Movsesian Lyrberg, Neum 2000 showed one line: 9 . . . i.e7 10 h4 'iYb6 11 'itfl !? a5! 12 �xf5 exf5 13 h5 f4! 14 i.. xf4, and in ChessPublishing both 14 . . . i.. g4 and 14 . . . i.f5 ! (intending 15 lDe3 i.. e4) are recommended, with compensation. From the diagram, there is a great deal of theory, but in general one should be wary of early simplification that favours White's centre. A brief overview of other ideas: (a) 9 . . . �6 and: (al) 10 0-0 a5 1 1
Play the French (a3) 10 h4 f6! 1 1 g4 lZ'lfxd4! (1 1 . . .lZ'lfe7 is unclear) 12 lZ'lcxd4 (12 lZ'lfxd4 lZ'lxe5 or 12 . . .fxe5 1 3 lZ'lxc6 �xc6) 12 . . . lZ'lxe5 13 g5 .tc5 14 0-0 lZ'lxf3+ 1 5 lZ'lxf3 'i!i'b4! 1 6 lZ'le 1 ! ? (16 �e3 'ilg4+ 1 7 'it>h1 'iWh3+ is a draw) 16 . . . 'iWxh4 1 7 lZ'lg2 'i!i'b4 18 �d3 0-0-0 and with 3 pawns for the piece and White's weakened kingside, Black can be quite happy, Movsesian-M. Gu revich, Bosna 2000. (b) 9 . . . �e7 10 0-0 g5 ! (Black can play on the flank as well) 1 1 g4 lZ'lg7!?
A fianchettoed knight, and without a square to move to! But it can be effective in supporting . . . h5 and/or .. .f5: 12 b4 a6 13 l:tb 1 f5 !? (this open more lines with unclear effects; 1 3 . . . h5!?) 14 exf6 .txf6 1 5 a4 h6 1 6 lZ'lce 1 !? 0 - 0 1 7 ..i e3 ..ie 8 1 8 lZ'ld3 �g6= Dvoirys-Zakharevich, Elista 200 1 . 8 h4 Given time, Black will attack the centre with . . . f6, for example, 8 lZ'lc2 .i.e7 9 g3 0-0 10 h4 cxd4 1 1 cxd4 f6! 12 h5 lZ'lh8 (now the idea is . . .lZ'lf7 to force the exchange of the e5 pawn) 1 3 h 6 g 6 14 exf6 �xf6 1 5 0 - 0 lZ'lf7 and Black has nice pressure on the centre with ideas of . . . lZ'ld6-f5/e4. White tried to free himself by 16 lZ'le3?! lZ'lxh6 1 7 lZ'lxd5 exd5 18 ..ixh6 l:te8 1 9 ne 1 , but 19 . . . �g4!+ threatened .. Jixe2 in 32
Jonkman-I.Botvinnik, Tel Aviv 2000. 8 ...cxd4 9 cxd4 i..b 4+
9 . . . i.. xa3!? 10 bxa3 weakens the white pawns but cedes the dark squares. Nevertheless, after 10 . . . h6! 1 1 h5 lZ'lge7 12 0-0 of Grischuk-Graf, Bled 2002, Black could have tried 12 . . . 0-0!? intending moves like . . . lZ'lf5, ... f6 and eventual play on the queen side. This is only for the brave. 10 Wn The knight on g6 comes in handy in another way following 10 ..id2 'i'b6 1 1 �xb4 'iYxb4+ 12 'i'd2 0-0 1 3 h5 lZ'lf4! 1 4 .i.f1 g6 15 hxg6 fxg6= . 1 0 ... h61 Not only stopping h6, but prepar ing a reorganisation of the knight to f8 and then to h7! l l lZ'lc2 1 1 h5 lZ'lf8 resembles the game. 1 l ... .te7 1 2 h5 (a) McDonald suggests 1 2 g3, al though 12 . . . 0-0 13 Wg2 lZ'lh8! and . . . f6 is an effective idea seen in the note to White's 8th move; (b) 12 ..id3 has been played to dis courage . . f6. Then Black can go in for 12 . . . 0-0 1 3 ..ie3 f5 !? 14 exf6 nxf6. 12 ...lZ'lf8 13 l:i.h31 lZ'lh71 The point. Now Black has the idea . . . lZ'lg5 in addition to the usual .. .f6. This all looks odd, but White hasn't .
Advance Variation: 5 i.d 7 . . .
done much either, and in all these positions the pawn on h5 can eventu ally become a target.
14 .i.d3 1 4 J:i.g3 might be met by 1 4 . . .'�Jg5 15 ttJxg5 hxg5 16 f4 "iWc7! with the idea 1 7 fxg5 f6! and otherwise . . . 0-0-0 with attack. 14 ... 0-0 15 l:tg3 �h8 16 .i.f4 f51 1 7 exf6 Instead, McDonald gives the line 1 7 "iWd2 �e8 18 l:th3 ttJg5 with black advantage (the h-pawn has become a problem) . Mer 1 7 exf6, Movsesian-P.Nikolic, Istanbul 2000 continued 1 7 . . .ttJxf6 1 8 ttJe5 ttJxe5 19 i.xe5 i.e8 2 0 l:th3 i.. d 6!? 21 ttJel ! ? (2 1 �e2 looks logical but allows 2 1 . . .i.xe5 22 dxe5 ttJe4! 23 i.. xe4 dxe4 intending 24 �xe4 'iVd2) 2 1 . . .i.xe5 22 dxe5 ttJe4! 23 lDf3 l:tf5 ! 24 ii'e2 (24 g4? protects the h-pawn but loses all the key squares after 24 . . JH4) 24 . . .l:txh5 25 .l::i.xh5 .i.xh5 26 i.xe4 dxe4 27 �xe4 m6+ . Black's king is safer, his pieces are better, and . . ..l:If8 is coming.
ommended in earlier editions. It con tains many ideas which haven't yet been looked at by theory. 7 0-0 The only serious move. 7 exf6 ttJxf6 with the idea . . . ii.d6, . . . 0-0 is a stan dard idea but even better for Black with the bishop on e2. And 7 .i.f4? m6! intending . . . ii'xb2 or . . . g5-g4 allows of no good answer. Finally, 7 c4?! cxd4 8 cxd5 exd5 9 exf6 ttJxf6 10 ttJxd4 .i.c5+ was Kupreichik-Dolma tov, USSR Ch 1 980.
7 :VIic7 6 . . . "iWc7 7 0-0 f6 reaches the same position. 7 . . . �c7 tries to exploit the fact that Black's queen is not yet committed to b6 and so can hit e5; this may also lure White's pieces to the wrong squares. ••
2.32 6 ... f6 Some readers will not be comfort able with the . . . lDge7-g6 ideas shown above. As an alternative I present a repertoire with 6 . . .f6, which was rec-
33
Play the French 8 � f4 The most successful move thus far. 8 l:!.e1 tries to defend e5 tactically, when Black has two effective re sponses: (a) 8 . . . fxe5 9 ctJxe5 ctJxe5 10 dxe5 0-0-0. This is a hard position to assess which McDonald thinks is 'dynami cally equal': 1 1 .i.g4 ctJh6!? (1 1 . . .g6!? is a logical idea, intending 12 .tf4 h5 13 .th3 .i.h6=) 12 ..txh6 (12 ..tg5 l:!.e8 13 .th5 g6 14 .tf6 :g8 15 ..tf3 maybe be slightly in White's favour) 12 . . . gxh6 1 3 ctJd2 b8 1 4 'ii'e2 l:!.g8 Strijbos unclear play, with Hochstrauer, Rotterdam 1 998; (b) 8 . . . 0-0-0 is a promising alterna tive, e.g., 9 �f4 (9 .tf1 fxe5 10 ctJxe5 ctJxe5 1 1 dxe5 g6!?; 9 �e3 fxe5 1 0 dxc5 ctJh6!?) 9 . . . g 5 (9 . . . h6!?) 1 0 �e3?! (10 �g3 'iVb6! intends . . . g4; There could follow 1 1 b4 cxb4 12 exf6 ctJxf6 1 3 ctJxg5 l:!.g8 14 ctJf'7 l:!.e8 and . . . l:!.g7 with a very slight white advantage) 10 . . . g4 1 1 ctJfd2 cxd4 12 cxd4 Broek man-Kleinhenz, Triesen 2000; and the easiest course was 12 . . . fxe5 1 3 dxe5 'iVxe5 14 .txg4 .td6 with a n ob vious advantage. 8 ...ctJ ge 7
9 .t g3 Other moves don't look dangerous: (a) 9 ..td3 f5 ! 10 :e1 c4 (10 . . . 'iVb6!?)
34
1 1 �c2 ctJg6 12 ..tg3 (12 .i.e3 .i.e7) 12 . . . �e7 1 3 h4 0-0= Galdunts Komarov, USSR 199 1 ; (b) 9 ctJa3 a6!? ( 9. . . ctJg6! 10 �g3 fxe5 1 1 .i.d3 cxd4 12 cxd4 'iVb6 with plenty of activity) 10 .tg3!? (10 .td3) Bastian-Lobron, Bundesliga 1 990; now 1 0 . . . 'iVb6 looks best, e.g., 11 'iVd2 cxd4 12 cxd4 ctJf5=. 9 .. :ti'b6! Even though this whole system is relatively unexplored, it's surprising that such a simple more has only been tried a couple of times. Black tries to exploit the fact that the g3 bishop isn't able to defend the centre or queenside.
10 'ilVd2! (a) 10 exf6 gxf6 11 'iVb3 'iVxb3 12 axb3 cxd4 1 3 ctJxd4 ctJxd4 14 cxd4 ctJf5 1 5 l:!.d1 ctJxg3 16 hxg3 .i.b4:j: with two bishops and the better pawn struc ture; (b) 10 ctJa3 ctJf5! (or 10 . . .cxd4 1 1 exf6 ctJf5!) 1 1 exf6 gxf6:j: . 10 ... cxd4!? 10 . . . ctJf5! 1 1 exf6 gxf6 looks good. 11 exf6 11 cxd4? ctJf5 12 :d1 ctJxg3 13 hxg3 fxe5 1 4 ctJxe5 (14 dxe5 .tc5 1 5 :£1 0-0 1 6 ctJc3 ctJd4 17 :ac1 :f5!+) 14 . . . ctJxe5 15 dxe5 .tc5 16 .th5+ g6 17 .tf3 0-0 1 8 ctJc3 ,U,f5 1 9 l:!.e1 l:!.af8+ Montoro-De
Advance Variation: 5. . . iLd 7 Dovitiis, Buenos Aires 1 998. 1 1...gxf6 12 tZlxd4 e5
13 tZlb5 1 3 tZlxc6 iLxc6 is at least equal for Black because of his centre. 13 ...tZla5! 14 tZl la3! Otherwise Black consolidates after 14 tZl5a3 iLe6. 14 ... a6 1 5 'iVe3! d4 16 cxd4 tZlf5! 16 ... axb5? 1 7 dxe5 'iVxe3 18 fxe3± . 17 'iVe4 tZlxg3 18 hxg3 axb5 19 dxe5 0-0-0 with unclear complications. White might be a little short of com pensation because Black's pieces are active. A lot of the above is pure analysis, but at this point 6 . . .£6 and 7 . . . 'iVc7 seems a good practical line with no major drawbacks.
35
Chapter Three Adva nce Variation :
S :iVb6 .
.
1 e4 e6 2 d4 d5 3 e5 c5 4 c3 liJc6 5 liJf3 'i!fb6 This introduces our alternate sys tem against the Advance Variation. Black immediately attacks the centre and will tend to play more direct moves than with 5 . . . �d7. The sys tems have some overlap, since the line 5 . . . �d7 6 �d3 cxd4 7 cxd4 'fIb6 transposes to 5 . . Jib6 6 .i.d3 cxd4 7 cxd4 .i.d7 below. Material divides into:
move has been the hottest line in the Advance Variation and used heavily at the top levels of international play. In fact, my most comprehensive da tabase has over 7000 games with 6 a3, considerably more than the tradi tional 6 iLe2. Since most French players still use 5 . . . iVb6, one could say that 6 a3 is the most critical line for modern Advance Variation theory.
3 . 1 6 a3 3.2 6 �e2 3.3 6 �d3
Others: (a) 6 dxc5?! .i.xc5 7 'iVc2 f6 destroys White's centre, e.g. , 8 b4 (8 exf6 liJxf6 9 iLf4 0-0 10 �d3 e5!) 8 . . . .te7 9 iLd3? fxe5 1 0 i.xh7 e4 1 1 �g6+ �f8 12 liJg5 i.xg5 13 iLxg5 liJe5+ ; (b) 6 g3?! f6! 7 �g2 (7 liJh4 g6!) 7 . . . cxd4 8 cxd4 iLb4+ 9 �f1 (9 liJc3 fxe5 intending . . . d4) 9 .. .fxe5 10 dxe5 liJge7 1 1 'iVe2 liJf5+. 3 . 1 6 a3 Over the past 1 5 years or so, this
36
6 ...liJh6!? There is an undeniable logic to de veloping rapidly versus the slow 6 a3. Black intends . . . cxd4 and . . . liJf5, win ning a pawn. The main idea is to fol low up with straightforward devel opment ( . . . .i.d7, . . . .i.e7, . . . :c8, and an
Advance Variation: 5. . :fib6 eventual . . .f6). In the second edition, I called the main move 6 . . . ltJh6 below a 'slightly irregular answer which I believe has been underestimated'. Now there are some 1250 games in my database, and practically every top French player has used it. Black has of course a wide variety of alter natives. 6 . . . c4 is the oldest move and is holding up well, but it is also less flexible. And I have the impression that the enormously popular line with 6 . . . i.d7 7 b4 cxd4 8 cxd4 .l::i.c 8 is falling a little short against some pre cise move orders by White. Of course that could change, but it seems to me that 6 . . . ltJh6 isn't as dependent upon special tactics as 6 . . . .i.d7 is. However, I do think that Black has an excellent weapon that has only been lightly explored, i.e. 6 . . .f6!? This, a sugges tion from the fIrst edition, still strikes me as sound and worthwhile.
dxe5!? lbh6! 9 .i.xh6 gxh6 10 i.d3 i.g7 1 1 0-0 0-0+ Guigonis-Renaudin, Paris 2000) 7 . . . c4 (also noteworthy are 7 . . . g6!? 8 b4 [8 exf6 lbxf6 9 "fic2 "fic7] 8 . . . cxd4 9 cxd4 i.g7; and 7 . . .fxe5, e.g., 8 dxe5 lbh6!? 9 0-0 c4 10 i.c2 lbf7, Mek-Murey, St Quentin 1 999) 8 .i.c2 g6!? (or 8 . . . .i.d7 9 0-0 0-0-0) 9 0-0 fxe5 (9 . . . i.g7 is also rea sonable) 10 lbxe5 (10 dxe5?! lbh6 1 1 "fie2 ltJf7+) 1O . . . lbxe5 1 1 dxe5 lbh6 12 "fie2 i.g7 1 3 i.e3 "fic7 14 f4 0-0 and the play is balanced. Perhaps the 6 . . .f6 variation should be seen more often. 7 b4 (a) 7 i.xh6? "fixb2 8 .i.cl "fixal 9 "fic2 is unsound: 9 . . . c4! intending . . . lba5, . . . i.d7, and . . . lbb3; (b) 7 i.d3 doesn't go well with a3: 7 ... cxd4 8 cxd4 i.d7 9 a-a?! (a strange form of Milner-Barry Attack, trading a3 for . . . lbh6; this should favour Black) 9 . . . lbxd4 10 lbxd4 "fixd4 1 1 lbc3 a6 1 2 "fie2 "fig4!? 1 3 f3? i.c5+ 14 r;;t>h l "fih4+ 1 5 g3 iVh3 16 "fiel i.c6 17 i.f4 0-0 with an extra pawn and ac tive pieces, Dvoirys-Lputian, Istanbul 2003. 7 ... cxd4 Here two moves have been tested for White: 3 . 1 1 8 i.xh6 3.12 8 cxd4
Black can (like his opponent) play on both wings here: on the queenside with . . . 0-0-0, . . . r;;t>b 8, . . . ltJa5 and . . . .l::i. c 8 or on the kingside with ideas like . . . fxe5, . . . ltJh6-f7, . . . g6 and . . . i.g7. Let's look at the briefest of overviews: 7 i.d3 (the most common reply; 7 b4 has been answered by 7 . . . c4 and 7 . . . cxb4 8 cxb4 fxe5, but the most in teresting alternative was 7 . . . fxe5!? 8
3 . 1 1 8 i.xh6 gxh6 9 cxd4 This is a fascinating line with which White trades off his important cl bishop to damage Black's pawn structure and secure his d4 pawn. Black gets two bishops and has good development in return for the broken pawn structure on his kingside. The plan is to play the moves . . . .i.g7, . . . 0-0, and .. .f6.
37
Play the French f6 14 exf6 'iVxf6 15 l:Ib 1 .l:tb8! 16 'ii' d 3 l:.g4 17 l:.d1 b6! with mutual chances) 12 . . . l:.g4? (12 . . .f6; 12 . . .f5!?) 13 �e2 (or 13 h3! l:.e4+ 14 i.. e 2±) 13 . . . a5 14 b5 llJa7 1 5 'i!Vb3 l:.e4 1 6 llJc3 a4 1 7 'i!fb2 'iWa5 18 O-O± D.Olafsson-A.Karlsson, Reykjavik 2000.
9 �d7 I like the fact that this move pre vents 10 �d3 (due to 10 . . .llJxd4) and it probably needs to be played any way. 9 . . . �d7 also clears the c-file, so that a move like 10 llJc3 can be met by 10 . . . llJxb4 1 1 axb4 �xb4 (see be low, and note that the bishop hasn't gone to g7 in this case) . I recommended 9 . . J:tg8!? in the last edition; but now we know that it gives White the advantage for posi tional reasons, which for complete ness I should specify: 10 llJc3! (10 b 5 llJ a 5 1 1 llJ c3 i s not very convincing in view of White's queenside weak nesses; for example, 1 1 . . .i.. d 7 12 llJa4 'iVc7! is strong) 10 . . . i.. d 7 (my main move in the second edition 1O . . . l:.g4? looks bad for several reasons, one of them being 1 1 h3 .l:.f4 12 i.b5 a5 1 3 O - O ! axb4 14 axb4 .l:.xa 1 1 5 'iVxa1 lixf3 16 gxf3 Wi'xd4 17 'iVa8± 'iit d 7? 18 llJa4 �xb4 19 ':'c 1 cJi;c7 20 i.. xc6 1-0 Budapest Keitlinghaus-Blauert, 1 998) 1 1 llJa4! (1 1 l:.c 1?! llJxb4 12 axb4 �xb4 gave plenty of compensa tion in Kretschmer-Wrba, corr 1993; 11 g3? .l:tg4! 12 i.. h 3 l:.xd4 1 3 llJxd4 'ii'xd4 14 'iixd4 llJxd4 15 0-0-0 llJf3 1 6 i.. g2 ':' c8 1 7 cJrb2 llJxe5 1 8 �xd5 llJg4!:j: Hakuc-Djabri, email 1 999) 1 1 . . .'ilid8 12 g3! (12 llJc5 �xc5 13 bxc5 ...
38
10 �e2 This solid move is best. 10 llJc3 al lows 10 . . . llJxb4! 1 1 axb4 .txb4
when Black's attack on c3 has no adequate answer, showing the value of delaying the f8 bishop's develop ment. One line goes 12 'iVb3 I1c8 1 3 :tc 1 0 - 0 (with the idea o f doubling on the c-file, but also of . . .f6) 14 cJre2 'iVa6+ 15 cJrd2 (15 cJre3? 'iVa3) 15 . . . 'iWa5 16 .td3 l:tc7 (or 16 . . . .ta4 17 'iVb2 ':'c6) 1 7 llJg1 �a4 18 'iWb2 f6! 19 exf6 l:txf6 20 f3 e5! 2 1 llJge2 .l:.fc6- + .
Advance Variation: 5. . . 'iVb6 10 ... .:.c8 Black agam retains the . . . lL'lxb4 idea while making a useful move. But 1O . . . i.g7 is also fine, since 1 1 lL'lc3?! (1 1 0-0 ':'c8 transposes) l 1 . . .f6 12 lL'la4 'iVd8 13 0-0 of Pisa-De la Villa Garcia, Spanish Ch 1993 should have been met by the simple capture 13 . . . fxe5 ! 1 4 lL'lxe5 lL'lxe5 15 dxe5 �xe5 1 6 �h5+
12 'iVd2!? Surprisingly, this runs into trou ble. Since moves by the b l knight lose the d-pawn, here are the alternatives: (a) 12 'iVd3 isn't really satisfactory after 12 . . .f6! 13 b5? (13 exf6 �xf6 1 4 ':'dl 0-0=1=) Degraeve-C.Bauer, Mont pellier 1993, and now 13 . . . lL'lxe5! 14 dxe5 fxe5+, since the advance ... e4 is next; (b) 12 b5 is probably best, although Black stood well after 12 . . . lL'le7 13 a4 f6 14 'iVd2!? fxe5 15 lL'lxe5 lL'lf5 1 6 �h5+
objectively no better, but doesn't give White any respite.
1 3 lL'lxd4! Better than 13 'iVxd4 'iVxd4 14 lL'lxd4 i.xe5 15 lL'lb3 i.xal 16 lL'lxal �a4 17 lL'ld2
39
Play the French
9 f61 Right now this older line looks promising to me. It turns out that Black doesn't have to accede to a likely drawing line but can try for more, whereas White (in order to avoid that line) , risks being left at a disadvantage. (a) To enter into a closed manoeu vering game, Black may want to try to block the kingside. One way would begin with 9 . . . ..te7: ...
Play here has gone 1 0 �d3 (10 �e2 0-0 11 lLlc3 f6 12 lLla4 'i!fd8 1 3 exf6 i.xf6 1 4 0-0 e5= Prang-Boehle, 1993; 10 lLlbd2 lLlxe3 11 fxe3 f5 12 exf6 ..txf6 1 3 ..td3 e5!=i=; 10 g4 lLlxe3 1 1 fxe3 f6! =i= ; 10 b5 lLla5 intending a quick . . . .td7 and . . . :c8) 10 . . . lLlxe3 1 1 fxe3, and now l 1 . . .f5 ! tries to set up a com40
plex manoeuvring struggle after 12 0-0 0-0 (threatening .. .f4) 13 lLlc3 'i!fd8! 14 lLle2 ..td7 15 lLlf4 a5! 16 b5 lLla7 17 a4 lLlc8 intending . . . lLlb6, . . . :c8, . . . lLlc4. The two bishops are temporarily passive but where is White's play coming from? Alternatively White could keep the position open: 12 exf6 ..txf6 1 3 lLlc3 (13 lLlg5 ..txd4! , e.g., 14 exd4 'i!fxd4 1 5 "i8h5+ �e7! 16 'i1kf7+ �d6-+ ; 1 3 0-0 0-0 1 4 lLlc3 .td7) 1 3 ... 0-0 14 0-0 (14 'i1kc2 g6 with the idea 1 5 �xg6!? hxg6 16 'i!fxg6+ iLg7 1 7 lLlg5 .l:!.f6 18 'ifu7+ 'iit>f8 19 "i8h5 'iit> e 7 20 'iVh7 'it>f8=) 14 . . . g6 15 'i!fc2 'it>g7?! (15 . . . lLle7 fol lowed by . . . .td7 and . . . Il.ac8 100ks fine) 16 lLla4 'i!fd8 1 7 lLlc5;!; Carlsen Lahlum, Gausdal 200 l . (b) Another blocking strategy is seen in 9 . . . iLd7 10 lLlbd2 (10 iLd3 could be answered by 10 . . . lLlxe3 1 1 fxe3 f5) 1O . . . lLlxe3 1 1 fxe3 f5 1 2 ..td3 (12 exf6 gxf6 13 .td3 e5!? 14 b5 lLla5!) H.Akopian-J.Watson, Los Angeles 1 998; and Black's simplest course was something like 12 . . . ..te7 13 0-0 0-0 1 4 lLlb3 'i1kd8 with the idea 1 5 lLlc5 b6 1 6 lLlxd7 'i!fxd7 17 l:tcl (17 iLb5 a6) 1 7 . . . a6 18 b 5 lLlb8 19 bxa6 lLlxa6= ; (c) 9 . . . lLlxe3 10 fxe3 g6 1 1 iLd3 .th6 12 'i1ke2 0-0 has been played a fair number of times; one idea for White might be 1 3 g4!? (13 0-0 f6 or 13 . . . ..td7) 1 3 . . . ..td7 14 h4 with ideas like 14 . . . J:tac8 15 lLlbd2 lLle7 (15 . . . ..tg7 16 h5 lLle7 17 'i!fh2) 16 h5 g5 17 lLlb3 etc. 10 .td3 Interesting and tactical. White may have difficulties in what used to be a drawing line: (a) 10 b5 lLlxe5! 1 1 dxe5 lLlxe3 12 fxe3 'i1kxe3+ 1 3 'fWe2 (13 ..te2? fxe5 is bad, leaving Black with two bishops and the centre) 13 . . . 'i1kc 1+!? 1;2 - 1;2 Ro-
Advance Variation: 5. . . �b6 manishin-Lputian, Yerevan 1 988, in view of 1 4 �d1 �e3+ 1 5 �e2 �c1+ etc. This verdict has been unchal lenged but there are other possibili ties. First, 14 . . . �2!? 15 tLlbd2 fxe5 1 6 �b 1 �xa3 1 7 tLlxe5 �e3+ 1 8 �e2 is analysis by Paul Cumbers, when Neil McDonald suggests 18 . . . �c5, 'though I still prefer White'. This should be looked at; perhaps White's grip on e5 slightly outweighs the two bishops. It seems to me that Black has a better winning try than Cumbers' by playing 1 3 . . . �xe2+ 14 �xe2 fxe5 1 5 tLlxe5 �d6, since then h e gets the centre but prevents the bind on e5. Something like 1 6 tLlf3 0-0 1 7 0-0 i.d7 1 8 tLlbd2 .u.ac8 might follow, intend ing . . . e5. Black only has two pawns for the piece but they are mobile passed central pawns and his bishop pair is powerful. There's also the im mediate problem that White can't yet challenge the c-file. I think that this is worth playing and that therefore 10 b5 doesn't even necessarily secure a draw; (b) 10 exf6 keeps the pieces on, al though it has scored poorly after 10 . . . gxf6 1 1 tLlc3 (1 1 �d3 tLlxe3 12 fxe3 e5 [or 12 . . . �d7] 1 3 tLlc3 �e6 14 �3 0-0-0 1 5 tLla4 �c7 1 6 0-0 cJ:;>b8 17 �b5 e4 18 tLld2 i.d6:j: G.West-Barlow, CAPA email 1997) 1 1 . . .tLlxe3 12 fxe3 �h6 and: (hI) 13 �d2 0-0 14 �d3 cJ:;>h8 1 5 � d 1 (15 0-0 tLlxd4!) 1 5 . . . �d7 16 0-0 �e8 17 cJ:;>h1 tLle7; (b2) 1 3 cJ:;>f2 0-0 14 �d3?! tLle5 1 5 h3 .td7 1 6 .u.e1 .u.ac8 1 7 .u.c1 tLlxd3+ 18 �xd3 a6 19 cJ:;>gl .u.c7+ Nilsson-Fer nandez Romero, Dos Hermanas 2002; (b3) 13 �d3 0-0 1 4 �e2 �d7 1 5 0-0 (15 tLla4 tLle5! 16 tLlxb6 tLlxd3+ 1 7 �xd3 axb6 18 cJ:;>e2 Gruic-Arkhipov, Belgrade 1 990, and simplest looks
1 8 . . . e5:j:) 1 5 . . . .te8!? (15 . . . tLle7!) 16 cJ:;>h1 (16 tLla4 �d8 17 tLlc5 �e7=) 16 . . . �g6 1 7 �d2 tLle7 18 tLlh4 �g5 1 9 tLlxg6 hxg6= Petrov-Fernandez Romero, Linares 2002. 10 ...tLlxe3 11 fxe3
1 1 ...fxe5 Or 1 1 . . . .td7!?, e.g., 12 exf6 gxf6 13 tLlc3 �h6 14 �e2 tLle7 (simpler seems 14 . . . 0-0 15 0-0 �f7 and . . . cJ:;>h8 next) 1 5 0 - 0 (15 tLle5 fxe5 1 6 �h5+ cJ:;> d 8 1 7 �xh6 .u.c8 was unclear in Jonkman San Segundo Carrillo, Mondariz 2000) 15 . . . 0-0 16 cJ:;>h1 .u.ac8 17 tLld1?! �g7!? 18 tLlf2 e5 19 dxe5 fxe5 20 e4 d4:j: Morozevich-Bareev, Monaco 2002. 12 b5 tLlxd4! Not 12 . . . exd4!? 13 exd4! and White stands much better, nor 12 . . . e4?! 13 bxc6 exd3? 14 tLle5± Prie-De la Villa Garcia, Leon 1 99 1 . 1 3 exd4 13 tLlxe5? tLlf5 hits e3 and favours Black. 1 3 ... e4 This has become a well-known po sition: I think that Black is distinctly better. 14 �xe4! Weak is 14 tLle5? �xd4, and 14 O-O? exf3 leads to 1 5 �xf3 �xd4+ or 15 .u.f2 �d6 with a winning position.
41
Play the French 14 ... dxe4 15 tZ'le5
15 ... g6 This is good, but apparently sim pler is the recent 15 . . . �d7! of Potkin Filippov, Togliatti 2003, which in tends 16 'iVh5+ g6 17 tZ'lxg6 hxg6 1 8 �xh8 0-0-0 with excellent play. S o the game went 16 0- 0 0-0-0 with Black better. This needs analysis and tests. 1 6 0-0 16 tZ'lc3 iLg7 17 tZ'lxe4 0-0 and White can't castle, whereas 18 tZ'lf3 �xb5 is also bad. 16 ... iLg7 17 hl iLxe5 18 dxe5 'iVxb5 19 tZ'lc3 'iVxe5 20 'iVb3 b6 Or perhaps 20 . . . e3!? After 20 . . . b6, Hurley-T. Clarke, Ireland 1 996 went 2 1 .l::f. a dl , and now Cumbers and McDonald have combined to produce the sample line 2 1 . . . .i.b7 22 tZ'lb5 (22 �a4+ b5 23 tZ'lxb5 iLc6) 22 . . . .:.d8 23 .l::f.xd8+ xd8 24 tZ'lxa7 'iVc5 25 �xe6 c7 26 �f6 .l::f. e 8 27 �f4+ �e5 28 �c 1 + 'iit>b 8 29 tZ'lc6+ iLxc6 30 �xc6 �e6, a pawn up. This is all up for grabs, but it's hard to believe that Black isn't in general able to use his extra pawns to consolidate. And White has to contend with the newer 1 5 . . . .i.d7 as well. 3.122 9 iLb2 iLd7
42
There are several move order is sues here. The text prevents 10 .i.d3 in view of the hanging d-pawn. Al though 9 . . . iLe7 is played consistently as well (e.g., by Lputian), it doesn't prevent that move, and I also prefer to have the e7 square to retreat the knight to after White's g4. Still, 9 . . . iLe7 has it's own advantages and my choice here is as much a matter of taste as an objective assessment. 10 g4 Whether or not this is the best move, it is the most critical, and the most frequent except for 10 iLe2. White's space-gaining ideas like h4 and g4 are the most dangerous ones, and several of the alternatives show that White cannot develop slowly without affording his opponent good chances: (a) 10 b5?! tZ'la5 gives up the c4 square too easily: 11 tZ'lc3 (1 1 a4 a6 12 tZ'lc3 .i.b4 13 bxa6 bxa6 14 iLe2 .l::f. c 8+ Hitzker-Kaland, Pinneberg 1 992) 1 1 . . . .l::f. c 8 12 tZ'la4 �d8 13 l:r.cl l:r.xc l 14 �xcl a6 (14 ... tZ'lc4! 1 5 iLxc4 �a5H) 15 tZ'lc5? (15 tZ'lc3:j:) 15 . . . iLxb5 16 .i.xb5+ axb5 17 0-0 tZ'lc4+ Marszalek Beyen, Leningrad 1960; (b) 10 'iVd3 nc8! threatens . . . ..txb4, so a sample line might go 11 �c3 a5!? (best is probably 1 1 . . .�e7 12 .i.e2 0-0
Advance Variation: 5 . :fib6 .
1 3 0-0 f6) 12 b5 0,a7 13 a4 .i.b4 with mutual chances; (c) 10 h4 is a modern treatment, the idea being to play h5 and g4 to establish a large space advantage on the kingside. In the meantime, the Black knight is denied access to g6 after a retreat to e7. Nevertheless, Black has well-placed pieces and two options:
(c 1) 10 . . .h5 1 1 g3 ..ie7!? (1 1 . . . .l:!.c8 intending . . .0,a5 may be even better, e.g., 12 'iVd2 j.,e7 13 .i.h3 'fVb5 !? 14 0,c3 'iVc4 with good play) 12 .i.h3 l:tc8 (12 . . . a5 !?) 13 'iVd2 (13 0-0 0,a5) 13 . . . a5 1 4 j.,xf5 exf5 15 0,c3 0,a7! 1 6 0-0 (16 0,xd5? �c6) 16 ... .i.e6 with approxi mate equality. Notice that in these lines all 7 White pawns are on the colour of his 'bad' bishop, but that this is more of an advantage than a problem for him! (c2) 1O .. JIc8 (quick development is the cure for Black's problems here) 1 1 g4 and: (c2 1) 1 1 . . .0,h6 12 J:!.gl i.e7 (12 . . . 0,a5!? 13 0,fd2?! j.,e7 14 i.c3 0,c4 15 0,xc4 dxc4 16 0,d2 O-O! un clear, Kozak-Antonio, Las Vegas 1 999, due to 17 0,xc4? l:!.xc4! 18 .i.xc4 'iVc7+) 13 0,c3 0,a5 14 0,a4 'iVc6 15 0,c5 0,c4 16 .i.c 1 0,g8! 17 .i.d3 �xc5 1 8 dxc5 b6 19 .i.xc4 dxc4 20 .i.e3
Sveshnikov-Dizdar, Dubai 2002, and 20 . . . 0,e7! 21 Iic1 b5 was good, intend ing moves like . . . 0,d5 and . . :iWa6 with . . . .i.c6. (c22) 1 1 . . .0,fe7!? 12 0,c3 0,a5! 1 3 0,a4 'iWc6 14 0,c5 0,c4 1 5 .i.c1 h5! 16 gxh5 0,f5 1 7 .i.h3 was Movsesian Haimovich, Panormo 200 1 . Here Mov sesian suggests 17 . . . b6 'and Black shouldn't be worried about the future at all' . (d) 10 .i.e2
(d1) 10 . . . �e7 11 0-0 0-0 12 g4! (12 'iWd2?! f6 1 3 exf6 .i.xf6 14 0,e5 i.. e 8!+ S.Lalic-N.Berry, Birmingham 2001) 12 ... 0,h4 (or 12 ... 0,h6 1 3 h3 f6) 13 0,xh4 i.xh4 14 f4 f6!? unclear; (d2) 10 . . . l::t c 8!? (with the idea . . . 0,a5) 1 1 0-0 i.. e 7 12 'iVd2 h5 13 l:td1!? g5! 14 0,c3! (14 0,xg5 0,cxd4 15 .txd4 0,xd4 16 'iVxd4 'iVxd4 17 lIxd4 .txg5 18 f4 i.d8! 'threatening .i.b6 when Black has the two bishops and control of the c-f'lle' - McDonald) 14 . . . g4 15 0,e 1 Grosar-Tukmakov, Bled 200 1 ; and now simply 15 . . . 0,cxd4! (15 .. .'iWd8!? 16 0,c2 Ji.g5 17 'iVe 1 0,ce7 was played) 16 0,xd5 0,xe2+ 1 7 �xe2 exd5 was very strong, since White nearly has to enter the variation 18 e6 'ii'x e6 19 'i'xe6 j.,xe6 20 i.. xh8 f6 21 J:i.ac 1 1Ib8! followed by . . .
Play the French 10 ... lLlh6 Also playable is 1O . . . lLlfe7
1 1 lLlc3! ( 1 1 .i.d3 h5 12 g5 g6 1 3 h4 lLlf5 14 .i.xf5 gxf5 1 5 lLlc3 lLle7! 16 0-0 lLlg6; Ancheyta-Nogueiras, Santa Clara 2001 ; 1 1 h3 lLla5 [or 1 1 . . .h5] 12 lLlbd2 lLlg6 1 3 .i.cl lLlc4! Novikova Matlakov, St Petersburg 2002, in tending 14 lLlxc4 dxc4 1 5 .txc4 .l:i.c8 1 6 .i. d 3 i.c6) 1 1 . . .lLla5 ! 12 lLl d2 (12 lLla4 "iVc6 13 b5 "iVc7 14 lLld2 [14 .l:i.cl lLlc4 15 'iVb3 'iVa5++] 14 . . . lLlc4 15 lLlxc4 dxc4 16 lLlc3 lLld5; Lindfelt-P.H.Niel sen, Nyborg 2001) 12 .. .lic8 13 licl lLlg6 1 4 h4 .i.e7 15 g5 (15 h5 lLlf4 1 6 "iVf3 i.g5= Grischuk-Kruppa, Elista 2000) 15 . . . h6 (or simply 15 . . . 0-0! intending 16 h5 lLlf4 or even 16 . . . "iVxd4! 17 hxg6 fxg6, so Potkin Hug, Istanbul 2003 continued 1 6 ligl !? lLlxh4 1 7 i.. d 3, when 1 7 . . . lLlg6 18 'iVg4 lic7! 1 9 lihl lifc8 20 'iVh5 �f8 2 1 �xh7 intending 2 1 . . . .i.xg5 would have been strong) 16 gxh6 Grischuk Radjabov, Wijk aan Zee 2003, and here McDonald (and later Grischuk!) mentions 16 . . . gxh6!? 17 h5 lLlh4 1 8 'ii'g4 lLlf5, which looks okay. 1 1 h3 White loses a pawn after 11 .i.h3? a5 ! ; and 1 1 .l:i.gl allows 1 1 . . .f6 (or 1 1 . . . i.. e 7 and . . . .ll c 8) 12 exf6 gxf6 1 3 lLlc3 lLlf7 ( 1 3 . . . i.. d6) 14 lLla4 ( 1 4 licl 44
i.h6! 1 5 lic2 lLle7; Lautier-Bauer, Val d'Isere 2002) 14 . . . "iVc7 1 5 licl 'iWf4! 16 lLlc5 i.xc5 1 7 dxc5 lLlce5 1 8 lLlxe5 Short-Lputian, Batumi 1999; and McDonald notes that 18 . . . fxe5! is quite strong. 1 1 ... f6
Shirov-Sadvakasov, Astana 200 1 continued 1 1 . . J:tc8 1 2 lLlc3 lLla5 1 3 lLla4 'iVc6 14 11cl lLlc4= . 1 2 lLlc3 12 i.. d 3 can as usual be met by 12 . . . fxe5 13 lLlxe5 lLlxe5 14 dxe5 lLlf7: 15 0-0 g5!? (or 15 . . . a5 16 b 5 .tc5) 1 6 lLlc3 i.g7 1 7 : e l 0 - 0 1 8 'ii'd2 lLlh8?! (18 . . . liac8;) 19 lLldl l:tf3 20 i.. d 4 'iVd8;t 12 - 12 Hollrigl-Van der Weide, Liech tenstein 1999. 12 ... fxe5 1 3 dxe5 13 lLla4 is counterproductive after 1 3 . . . 'iVd8 14 lLlxe5 lLlxe5 1 5 dxe5 .i.e7 with dark square control. 1 3 ... i. e7 Or 1 3 ... lLlf7. After 1 3 ... i.. e 7, Svesh nikov-Lputian, Sochi 1 993 continued 14 lLla4 'ii'd 8 15 licl 0-0 16 lLlc5 i.xc5 17 l:lxc5. Apart from 17 . . . lLlf7, as played, Black had 17 . . . 'i!Ve7! intending . . . a5, when the second player has a slight edge, a risky line going 1 8 litcl?! (better i s 1 8 ':'c3 lLlf7 1 9 11e3 lLlg5;) 18 . . . lLlxb4! 19 g5! (19 axb4 'iVxb4+ 20 'iVd2 'iVe4+) 19 . . . lLlf5 20
Advance Variation: 5. . . �b6 axb4 ttJh4! 2 1 ttJxh4 �xb4+ 22 �d2 �e4+ 23 �e3 �xh 1 +. This has been lot of detail on one variation, but it has been the most important line of the Advance Varia tion for many years now. 3.2 6 .lie2
The traditional main line, some what superceded by 6 a3 but still very important and extremely com plicated. 6,..ttJh6 After this flexible move, White must attend to the threat of . . . cxd4 and . . . ttJf5. He can do so by: 3.21 7 b3 3.32 7 .lixh6
Others: (a) 7 dxc5? i.xc5 8 0-0 ttJg4+ ; (b) 7 ttJa3? cxd4 8 i.xh6 gxh6 9 cxd4 �xb2 10 ttJb5 'iWb4+ 1 1
16 lIc1 ttJxd4 17 �xd4 i.g7 and Black is better - Boey; (e) 7 i.d3 tries to prove that in a Milner-Barry sort of position, the ex tra move . . . ttJh6 actually hurts Black: 7 . . . cxd4 8 cxd4 .lid7 (or 8 . . .ttJf5 9 i.xf5 exf5 10 ttJc3 i.e6 1 1 ttJe2 h6 12 h4 0-0-0= Benj amin-Shaked, Kona 1 998) 9 i.c2 (in view of the threat . . . ttJxd4) and Black can play simply 9 . . . ttJb4= , or more interestingly: (e 1) 9 . . .f6 exploits Black's consider able lead in development to break down White's centre: 10 ttJc3 (10 i.xh6? gxh6 11 b3 i.b4+ 12
45
Play the French point performance rating advantage for Black in 55 games! (Of course, that is by no means proof of objective worth, and the sample sizes are re duced by unrated games, but still . . . ) A typically entertaining line is 1 1 g3 0-0 12 'it>g2 f6 13 a3 i.e7 and the white king still isn't safe: 14 b4 fxe5 15 dxe5 i.c5! 16 g4 lLlfe7 1 7 lLlc3 i.xf2! 18 lLla4 'iVe3 19 .l:tfl lLlg6 20 .l:txf2 lLlf4+ 21 'i.t>fl lLlh3! 22 'iVe l lLlxe5 23 i.xe5 i.xa4 24 i.d4 'iVe4 25 'iVd2 'iVg6? (25 . . . lLlxf2 ! 26 'i.t>xf2 .l:txf3+ 27 i.xf3 .l:f8 with the idea 28 'iVe3 llxf3+ 29 'iVxf3 'iVxd4+) 26 'iVe3? 0- 1 Marti novic-Pavlovic, Igalo 1994. White is lost after 26 . . . lLlxf2.
l l lLlc3 This developing move avoids weak nesses and is best. Alternatives: (a) 1 1 i.d3 f6! 12 i.xf5 (12 a3 i.e7 13 i.xf5 exf5+ leaves White with a stranded king and weaknesses) 12 . . . fxe5!? ('! !' - Tsouros, but obj ec tively better is 12 . . . exf5!+) 13 i.c2? (13 i.h3! e4 14 lLlel e5 15 i.xc8 .l:taxc8 - Tsouros; but here 14 lLlg5! im proves) 13 . . . e4 14 lLlel e5 15 a3 i.d6 16 b4 exd4 17 i.b3 i.e6 18 f3 exf3 1 9 lLlxf3 .l:txf3+!-+ was Nicoleris-Tsouros, Greece 1973; (b) 11 g4 is the old and most dan gerous line, which deserves some 46
study:
1 1 . . .lLlh6 and White has: (b l) 12 h3?! f6 (or 12 . . . i.d7 13 a3 i.e7 14 b4 f6! 1 5 exf6 .l:txf6 1 6 g5? .l:txf3) 13 exf6 .l:txf6 14 a3 (for 14 g5 .l:txf3, compare 12 ttgl) 14 . . . i.d6 15 g5 .l:txf3 16 i.xf3 lLlf5 17 lLld2 i.d7! 18 i.g4 lLlcxd4 1 9 b4 .l:tfB 20 i.xd4 lLlxd4 2 1 lLlf3 i.b5+ 22 �el l:If4 23 lLld2 lLlf5 24 f3 'iVe3+ 0- 1 Gutierrez Castillo Matamoros Franco, Malaga 200 l . (b2) 12 l:!.gl f6 1 3 exf6 (a beautiful game followed 1 3 a3 i.e7 14 exf6 .l:txf6 1 5 g5 .l:txf3 16 gxh6 .l:tf7 1 7 i.d3 g5! ! 18 'iVh5 i.d7 19 lLld2 .l:taf8 20 f3 'iVc7 2 1 .l:te l 'it>h8 22 .l:te2 i.f6 23 .l:teg2 'iVf4 24 .l:te2 i.xd4- + Kiselev-Bilokha, Kiev 1 999) 13 . . . .l:!.xf6 14 g5 (the move 14 lLlc3 can be met conservatively by 14 . . . i.xc3 15 i.xc3 lLlf7, or by 14 . . . lLlf7 15 g5 .l:tf4 16 lLla4 'iVc7 17 i.cl e5! 18 i.xf4 i.h3+ 1 9 .l:tg2 exf4+) 14 ... .l:txf3 and: (b21) 15 gxh6?! . .l:tf7 16 i.d3 e5 (or 16 . . . g6! - Botterill) 1 7 �5 e4 1 8 i.e2 Botterill-Ligterink, London 1 978, and here Botterill gives 18 . . . g6!+ which in fact looks winning. ; (b22) 1 5 i.xf3 lLlf5 16 .l:tg4 i.d7 1 7 lLlc3 .l:tfB, e.g., 18 lLla4 'iVc7!? 1 9 'i.t>gl i.e8 20 lLlc5! (20 'iVe2 i.h5 ! 21 'iVxe6+ 'i.t>h8 22 'iVxd5 i.xg4 23 i.xg4 'iVf4 24 'iVg2 lLlcxd4 25 .l:tdl lLlh4+) T.Reich-
Advance Variation: 5" .1Wb6 P.Meister, Bayern 1988, and best was 20".tLlfxd4+ ; (c) 1 1 a3 iLe7 12 b4 f6 13 g3 fxe5 14 dxe5 a5! 1 5 b5 iLc5 1 6 g4 tLlfd4- + Campora Perez-Matamoros Franco, Seville 200 1 ; Returning to 1 1 tLlc3: 11 ... f6 1 2 tLla4 12 g4 tLlh6 (or just 12. "tLlfe7 aiming at f4) 13 exf6 tLlxg4!? 14 fxg7 Mf7 with complex play, Gudbrandsen-Lahlum, Oslo 2000. Mter 12 tLla4, the main line goes 12" .1Wd8 13 h4 (13 a3 iLe7 14 b4 fxe5 1 5 dxe5 iLd7+ Sorokin-Sokolsky, USSR 1951) 1 3 " .fxe5 14 dxe5 iLe7!? (or 14" .iLd7 1 5 iLd3 iLe7! 16 h5 Mc8) 15 iLd3 iLd7! with the idea " .iLe8-h5. This is advantageous, since 16 g4? tLlxh4 only exposes White, and if 1 6 Mel, 16" .iLe8 1 7 h5 iLg5 activates Black's pieces. 3.22 7 iLxh6
This is the most critical line after 6 iLe2 tLlh6. White gives up the bishop pair in order to inflict damage upon Black's pawns. 7 ... gxh6 Commentators have given a wide variety of conflicting analysis about 7" .1Wxb2? ! ; ultimately it seems infe rior after 8 iLe3! 1Wxa1 9 1Wc2 cxd4 10
tLlxd4! iLd7! 1 1 O-O! (1 1 tLlb3 tLlb4) 1 1 . . .tLlxe5 12 tLld2 1Wxf1+ 13 iLxfl± with the idea 1Wb3 and tLlb5. 8 1Wd2 This is better than 8 b3? cxd4 9 cxd4 f6 ! 10 exf6 iLb4+ 1 1 <J;;>f1 0-0 with attack down the f-file, Karlsson Helmers, Reykjavik 1 982 : 12 a3 iLd6 13 tLlc3 Mxf6 14 tLla4 1Wc7 15 b4 1Wf7!? (15" .1Wg7!+) 16 Mc1 e5!? (16" .iLf4) 17 Mxc6! (17 dxe5 tLlxe5 18 tLlc3 iLe6+) 17" .bxc6 18 dxe5 Mxf3 19 iLxf3 iLxe5+. 8 ... iLg7 9 0-0 0-0 10 tLl a3 White has no other good way to de velop his pieces. 10 ... cxd4 10" .iLd7 1 1 tLlc2 Mac8 12 dxc5!? (12 b4! is the main line) 12" .1Wc7!?= . 1 1 cxd4 iLd7 Noteworthy is the unique order 1 1 . . .f6 12 exf6 Mxf6 13 tLlc2 (13 Mad1 iLd7 14 tLlc2 iLe8 15 b4 a6 16 a4 iLg6= Patience-Pye, Eastleigh 2001) 1 3 " .a5 (to prevent b4; compare 1 1 . . .iLd7 12 tLlc2 a5) . Here typical play might go 14 tLle3 ! (14 Mab 1 iLd7 15 b4 axb4 16 tLlxb4 1Wa5=) 14" .h5! (my idea in several lines, preventing tLlg4) 1 5 tLlc2 iLd7 16 Mab 1 iLe8= and " .iLg6. 12 tLlc2
The basic position. White wants to 47
Play the French expand on the queenside by b4 and a4, and/or attack Black's kingside (e.g. by 'Lle3-g4) . Black will play .. .f6, after which he hopes that the two bishops, f-fIle, and/or pressure on d4 will yield good play. 12 f6 The normal move. But a very im portant alternative is 12 . . . a5!?, by which Black temporarily prevents b4: ...
1 998, and now the easiest course was 16 . . . i.. x e5 17 dxe5 'Llxe5 18 iLxh5 ..tb5 19 i.. e 2 �xe2 20 'ilixe2 J:iae8= . 1 3 exf6 13 iLd3 fxe5! 14 'Llxe5 (14 dxe5 .l:!.xf3 15 gxf3 'Llxe5+) 14 . . . 'Llxe5 15 dxe5 l:tf7+ (or 1 5 . . .'i'xb2) . 13 l1xf6 14 b4 White shouldn't allow Black to de velop uninterrupted by . . JbfB and . . . ..te8-g6. The alternative is 14 'Lle3 J:tafB 15 l1adl l1f4 16 'Llc2 i.. e 8!i 011Rohde, New York 1 994. 14 11af8 The standard move, but it leads to positions that are difficult to play. So close attention should be paid to 14 . . . i.. e 8!?, a promising move that represents a new plan in the position: 15 b5 (15 a4 i.. g6 16 a5 "fIic7 17 l:tfcl a6=) 1 5 . . .'Lld8! ('!' for originality: in stead of heading for g6 or f5, Black envisions a knight on f7 that covers e5 and can look for greener pastures via d6) . ...
...
(a) 13 a4 f6 (13 . . . 'Llb4 1 4 'Lle3 h5 1 5 .J:!.fdl f6 with good counterplay) 1 4 exf6 l:1.xf6 1 5 lIfb l �e8! 1 6 b4?! iLg6 17 bxa5 'iYxa5 18 'iVdl 1:.f7 (18 .. :iVa7!i) 1 9 l1b5?! Trygstad Brynell, Gausdal 200 1 ; and now 19 .. :ii'a7! leaves White no good de fence against . . . ii.xc2 and/or . . . l:.afB. (b) 13 l1ab l f6 14 exf6 l1xf6 15 b4 axb4 1 6 'Llxb4 'iYa5=, as above; (c) 1 3 'Lle3 h5! and: (c l) 14 g4!? f6! 15 exf6 l1xf6 1 6 gxh5 (16 g5? lH4 1 7 'Llg2 .l:!.f7i) 1 6 . . . .J:!.af8 (16 . . . ..te8=) 17 'Lle5 iLe8 1 8 'Llxc6 ( 1 8 .J:!.adl l:tf4!+) 1 8 . . . bxc6 19 'Llc2 e5 20 dxe5 11f4!i; (c2) 14 a4 f6 1 5 i..b 5 fxe5 16 iLxc6 bxc6 17 'Llxe5 .i.e8 18 f4 l:tb8 19 l:tf2 c5i D . Fernando-Brynell, Bled 2002; (c3) 14 l1adl f6 15 exf6 ..txf6 (15 . . . l:txf6 16 'Lle5 ..te8 1 7 'Llxc6 bxc6=) 16 'Lle5 (16 'Llc2 ..te8 and . . . i.. g6) Rajlich-J.Watson, Las Vegas 48
Now White has: (a) 16 'Lle5 'Llf7 with these plausi ble continuations: (a l) 17 f4? 'Llxe5 18 fxe5 11xf1 + 19 1:.xfl i.. x e5 with the idea 20 'iVxh6? �xd4+ 2 1 'Llxd4 'iYxd4+ 22 'i£?hl i.. g 6+ ; (a2) 1 7 'Llxf7 l:txf7! 1 8 a4 I1c8 19 a5
Advance Variation: 5. . . 'iib6 'ii d6 20 g3 (20 Itfcl Itfcn) 20 . . . e5= ; (a3) 1 7 lbg4 J:.g6 1 8 a4 Itc8! 19 Itfc l lbd6! 20 lbe5 lbe4! (20 . . . i.xe5 2 1 dxe5 lbe4 22 'iie 3!?) 2 1 'iie l (2 1 'iie 3? Itc3) 2 1 . . .:f6! 22 a5 'iid6 23 lbf3 lbg5=; (b) 16 a4 lbf7 1 7 lbe3 h5 1 8 .l:!.ac l (18 i.d3?! .l:!.xf3 ! 19 gxf3 'i!i'xd4) 18 . . . lbd6! 19 'ilYb4 (or 19 lbe5 lbe4 20 'ii d 3 .l:!.f4!=) 19 . . . lbe4 20 a5 'i!i'd8!? (20 . . . 'i!i'd6!=) 21 b6 i.f8 22 'i!i'b2 i.d6 23 lbe5 axb6 24 'iixb6? (better is 24 f3 lbg5 25 'i!i'xb6 'i!i'xb6 26 axb6 .l:!.f4=) 24 . . . 'i!i'xb6 25 axb6 lbd2 26 .l:!.al .l:!.b8 27 .l:!.fdl lbb3 28 Ita7 lbxd4 29 Itxd4 J.xe5+ Fernando-Gdanski, Cappelle la Grande 2002. The line with 14 . . . i.e8 and the idea . . . lbd8-f7 IS a fun and little investigated way to create counter chances and avoid the theoretical 14 . . . Itaf8. 15 b5 lbe7 This has been played in many games. Probably White is slightly better in a theoretical sense, but so far Black has more than held his own in terms of results. Another sound option is 15 . . . lba5!?, my suggestion from the second edition whose main drawback is that it can easily lead to a draw. It has now been tested: 1 6 lb e 5 ( 1 6 Itab l Itf4! keeps White tied down, and 16 'i!i'e3 Itf4 [16 . . . Itc8!?] intending 1 7 lbe5 i.xb 5 1 8 Itfb l? Ite4 winning for Black) 16 . . . i.xb5! 1 7 Itab l (17 lbg4 Itf4 18 Itab l lbc4+ ; 1 7 Itfb l?? Itxf2; 1 7 'i!i'e3 Itf4!) 1 7 . . . i.xe2 18 Itxb6 i.xf1 19 Itb l Itxf2 20 'i!i'xf2 .l:!.xf2 21 'iit>xf2 J.a6 (thus far my analysis of 14 . . . lba5; Black has two bishops and two pawns to compen sate for the exchange and his weak rook pawns) 22 .l:!.el (22 lbb4 J.xe5 23 dxe5 J.c4, about equal and leading correctly to a draw, Jonkman Tiggelman, Vlissingen 1999) 22 . . . i.b5
23 Ite3 a6 24 a3 i.a4 25 lbel Rajlich Hummel, North Bay 1 998; and al though 25 . . . lbb3 was still well inside the drawing range, simpler was 25 . . . lbc4 26 lbxc4 dxc4 27 lbf3 'iit>f7 28 lbe5+ 'iit> e 7= . 16 lbe5 J.e8
1 7 g3 (a) 17 J.d3 lbg6 18 i.xg6 hxg6 19 a4 Thipsay-Konguvel, London 200 1 ; and Black should try 1 9 . . . g5! (versus f4, and clearing the way for a future . . . i.g6) 20 Itae l a6 2 1 bxa6 'iixa6 with excellent play; (b) 17 a4 (a popular move with the idea Ita3-g3, but Black can challenge the centre first) 1 7 . . . lbg6 1 8 lbg4 It6f7 19 liJxh6+?! (19 g3 h5 20 liJh6+ J.xh6 21 'i!i'xh6 liJf4! 22 i.xh5? [22 'iig 5+ Itg7 23 'i!i'e5 i.g6=] 22 . . . Itf6 23 'i!i'g5+ Itg6! 24 'i!i'e5 liJxh5 0- 1 Kun-Szuk, Budapest 2000) 19 . . . i.xh6 20 'i!i'xh6 lbf4 2 1 i.f3 Itg7 22 g3 'i!i'd8! (threat ening . . . Itg6) 23 'iit>h l .l:!.g6 24 iVh4 'i!i'xh4 25 gxh4 liJd3+ Adams-Lobron, Amsterdam 1 994. 1 7 ... h5 1 8 a4 Or 18 liJe3, e.g., 18 . . . liJf5!? 19 liJxf5 Itxf5 20 f4 J.xb5!? 2 1 Itfb l Itxe5! 22 fxe5 i.xe5 23 'iit> g2 'iix d4 24 'iix d4 i.xd4 25 i.xb5 Itf2+ 26 'iit>h 3 J.xal 27 Itxal 'iit> g 7, perhaps with ... h4 next. Black's two passed central pawns
49
Play the French look good enough to hold, but this demonstrates the marginal nature of Black playing this line past his 1 4th and 1 5th move options. IB ... h4
This position was drawn in Magem Badals-Vallejo Pons, Mondariz 2000, but Topalov-Bareev, Novgorod 1997 showed that White was somewhat better: 19 i.. d 3 hxg3 20 hxg3 h5? (20 .. .ltJg6! with only a small advan tage for White and play for both sides) 2 1 ctJe3 i..h 6?! (2 1 . . :i'd8 22 l:[acl ctJg6;!;) 22 a5 'ifd8 23 f4± . Over all, 15 . . . ctJe7 is playable, but 15 . . . ctJa5 should completely equalise. Looking back over this variation for better winning chances, Black should strongly consider the promis ing alternatives 12 . . . a5 and 15 . . . i.. e 8. I think that he can be happy with both of these more ambitious moves. Also interesting is the lesser-known 1 1 . . .f6. 3.3 6 i.. d 3 This normally leads into the Mil ner-Barry Gambit, in which White gives up a pawn for rapid develop ment. This gambit has now been relegated to a rare sideline, but we must also look at new ideas, e.g., on White's 8th and 9th moves. 50
6 ... cxd4 7 cxd4 7 0-0 has been answered by 7 . . . i.. d 7 in several games, transposing after 8 cxd4 ctJxd4. Also possible is 7 . . . dxc3 8 ctJxc3 i.. d 7, 8 . . . ctJge7, 8 . . .f6!? 7 ...i.. d 7
Not 7 . . . ctJxd4?? 8 ctJxd4 'iWxd4?? 9 i..b 5+. B 0-0 Gambiting the d-pawn is normal because the options are so dispiriting: (a) 8 i.. e 2?! is a full tempo down on older 6 i.. e 2 lines after 8 . . . ctJh6! 9 b3 ctJf5 10 i..b 2 i..b 4+:j: 11 'it>f1 0-0 12 ctJc3 f6 and Black will wreak havoc down the f-file; (b) 8 i.. c 2 !? is slow. After 8 . . . ctJb4, White has lost any chance for advan tage (he has scored 22% in my data base, although only weaker players are involved) : 9 ctJc3 (9 i.. a 4 'iWa6!:j: Schwarz, e.g., 10 i.. x d7+ 'it>xd7 and . . . ctJd3+ follows; 9 0-0 ctJxc2 10 'i'xc2 .l:i.c8 1 1 ctJc3 transposes; 9 i..b 3? i..b 5) 9 . . . ctJxc2+ 10 'i'xc2 l':tc8 (or 1O . . . ctJe7 1 1 i.. e 3 ctJc6 12 0-0 .i.e7 intending . . . 0-0 and . . . f6) 1 1 0-0 ctJe7 12 i.. e 3 (12 'i'd3 ctJc6 1 3 a3 ctJa5 14 b4 ctJc4:j: Jona than-Soln, Szeged 1994) 12 . . . ctJc6 (or 12 . . . ctJf5!) 13 'i'd2 (13 a3 ctJa5) 1 3 . . . i..b 4! 14 a3 Stadler-Ilchmann, corr 1958-9; 14 . . . i.. xc3! 15 bxc3 ctJa5+ - Heemsoth;
Advance Variation: 5. wt'b6 . .
(c) 8 ttJc3 ttJxd4. Now 9 ttJxd4 'iVxd4 is the main line gambit. White also has: (c1) 9 ttJg5 !? resembles the 9 0-0 ttJxd4 10 ttJg5 gambit line below. But this order may be flawed due to 9 . . . ttJc6!, since .:te l is not available, e.g. 10 0-0 (10 'i!Ve2 loses a pawn after 10 . . . d4! 1 1 ttJb5 [1 1 ttJd1 '1i'a5+ and . . . wt'xe5] 1 1 . . . a6 12 ttJd6+ .i.xd6 1 3 exd6 wt'b4+ 14 .i.. d2 llYxd6+) 10 . . . ttJxe5 1 1 .l:r.e1 (1 1 ttJxd5 llYd6!) 1 1 . . . ttJxd3 12 wt'xd3 ttJf6+ with the idea 1 3 ttJxd5 ttJxd5 14 wt'xd5 ..tb4 1 5 11e2 ..tb5; (c2) 9 0-0 ttJxf3+ (9 ... ttJc6:j:) 1 0 �xf3 ttJe7 1 1 'iWg3 �c6! 12 a4 ttJf5 1 3 ..txf5 exf5 14 'iWg5 d4+ Sobolevsky- Enders, Muelhausen 2000. 8 ... ttJxd4 Now: 3.31 9 ttJg5 3.32 9 ttJxd4
Other ways to gambit the pawn are tricky but insufficient: (a) 9 ttJbd2 ttJc6 (9 . . . ttJxf3+ is not bad either) 10 ttJb3 ttJge7 (or 10 .. .f6!? with the idea 11 .i.. f4 fxe5 12 ttJxe5 ttJf6) 1 1 1l.e3 (1 1 ..tf4 ttJg6 12 ..tg3 a5 1 3 a4 f5 14 exf6 gxf6 1 5 l1e1 'iiff7) 1 1 . . :�c7 12 .l:r.c 1 ttJg6 13 ttJc5!? (13 lIe1 ttJgxe5!) 1 3 ... ..txc5 14 ..txc5 ttJgxe5 1 5 ':' e 1 ':'c8! 16 .i.. a 3 ttJxd3 1 7 wt'xd3 ttJe5 !? 18 lIxc7 (18 '1i'e3!? ttJxf3+ 1 9 gxf3 �d8 20 l1xc8 .i..xc8 2 1 'i¥c5 f6) 1 8 . . . ttJxd3 1 9 llxc8+ .i.xc8 20 lld1 ttJf4 2 1 g3 ttJg6 22 11c 1 �d7 23 ttJg5 f6 and Black was still two pawns up in Nun Schmittdiel, Prague 1 990; (b) 9 ttJc3!? is mentioned by Bickford as a transpositional device, e.g. 9 . . . a6 10 ttJxd4 'i¥xd4. The alter natives are 9 . . . ttJc6 and 9 . . . ttJxf3+! 1 0 �xf3 ttJe7, e.g., 1 1 '1i'g3 (1 1 � h1 ttJg6 12 llYe2 .i.. c 5+) 1 1 . . . .i.c6 12 a3 (12 a4
ttJf5 13 ..txf5 exf5 is Sobolevsky Enders above; 12 lle1?! g6 13 .i.. g 5 �g7) 12 . . . a6 13 b4 ttJg6+ Maderna Naj dorf, Mar del Plata 1942. 3.31 9 ttJg5
This is known as Sorensen's Gam bit. It has undergone a spurt of popu1arity and White has enjoyed a num ber of nice attacking victories, but it is dubious at best if the opponent is prepared. 9 ..ttJc6! The simplest course, fighting for the initiative by attacking e5. 10 lIe 1 This is thought to be best and has been played almost exclusively. White's other plausible moves are: (a) 10 llYe2 ttJb4 (K.Krantz's solu tion) 1 1 .i.xh7 ..tb5 12 wt'f3 ttJh6 1 3 l:te1 �e7 14 a 3 ttJc6 1 5 ..tc2 ttJd4 16 wt'd1 ttJg4! and Black is winning. Black has also gotten a clear advan tage after 10 . . . g6 1 1 ttJc3 .i.. g 7, as in Wuttke-Lieb, Berlin 1 998; (b) 10 ..tf4? wt'xb2 (or 1O ... ..te7 1 1 wt'h5 g6 1 2 wt'e2 ttJb4!) 1 1 ttJd2 ttJxe5 12 l::tb 1 wt'd4 1 3 ..txe5 'i!Vxe5 1 4 ttJdf3 wt'c7 1 5 wt'e2 ..td6- + , e.g. 16 wt'b2 (16 l:tfc 1 ..tc6 1 7 ttJd4 ttJe7-+ Storey A.Whiteley, Newcastle 1995) 16 . . . ttJf6 17 wt'xb7 wt'xb7 18 l:txb7 h6 1 9 ttJh3 .
51
Play the French g5!- + , winning a piece! (c) The sacrificial try 10 ttJc3?! ttJxe5 11 ttJxd5 is met by l 1 . . .'iIt'd6! (as given Krantz) with Black a clear pawn up after 12 ttJf4 ttJf6. 10 .. i.c5! This is not the only solution, but it seems to virtually refute White's gambit. In my database, 10 . . . i.c5 has scored 7 1 % with a 3071 to 2 106 Per formance Rating advantage! Of course that's silly, but White really doesn't seem to have a thing for his pawn here. 1 1 'ilt'f3 .
':xe5 i.xe5 16 i.xd8 lIxd8-+ Cal zolari-Poupar, Clichy 1 999) 13 . . . ttJe7!? 14 ttJe2 ttJg6 15 i.xg6 fxg6!? 16 ttJxd4 'iYxd4 17 lIad1 'fixb2! (17 . . . 'iVb6 1 8 'iYa3 ! i s given b y Sorenson expert Lennart Hansson) 18 lib 1 'fixa2 1 9 lIxb7 ttJf5+. 1 3 'ikf4 1 3 'ilt'd 1 ttJhf5 14 'iVh5 g6 1 5 'tWd1 ':'c8+ Bailen Canales-Barr, corr 1998, or 1 5 . . . 0-0 ! . 1 3 ... ttJdf5 Or 13 . . . ttJhf5 14 g4 h6+ . After 1 3 . . . ttJdf5, Schmedders-Wilde, Sueder 1996 went 14 i.d2 0-0 15 b4! i.e7! 16 ttJxh7!? 'it>xh7 1 7 g4 'fixb4 18 ttJe4 'fia3-+ . 3.32 9 ttJxd4 'fixd4
1 1 . ttJh6 The most common, but 1 1 . . .0-0-0! looks even stronger: 12 ttJc3 (12 ttJxf7?? .:tf8-+) 12 . . .f6! (12 ... ttJh6 isn't bad either, with 1 3 i.f4 ttJd4 or 1 3 h3 i.d4) 13 exf6 (White blundered in one game by 13 ttJa4?? 'iYa5; also bad is 1 3 ttJ rl ttJxe5! 14 ':xe5 fxe5 1 5 ttJxh8 ttJf6 1 6 ttJrl e4! 1 7 i.xe4 lIf8!+) 1 3 . . . ttJxf6 1 4 ttJrl ttJg4! (or 14 . . . ttJb4) 1 5 'ilt'xg4 (15 ttJdl ':hf8+) 1 5 . . . .i.xf2+ 16 'it>f1 i.xe l 1 7 'it>xe1 'ilt'gl + 18 i.f1 ttJd4- + . 12 ttJc3 12 h3? i.d4! wins. 12 .. ttJd4 Again 12 . . . .i.d4 is good: 13 .i.f4 (13 ttJxd5!? 'iYa5! 14 ttJc3 ttJxe5 1 5 1We4 'fic5+; 13 ttJxh7 0-0-0 14 .i.g5 ttJxe5 1 5 ..
.
52
The Milner-Barry proper. 10 ttJc3 (a) Like most lines in this section, 10 'tWe2 has scored overwhelmingly for Black, e.g., 1O . . . ttJe7 (10 . . . f6 is my suggestion; one line is 1 1 exf6 ttJxf6 12 1Id1 'fih4!+, and not much better is 1 1 'fih5+ 'it>d8 12 'firl ttJe7 1 3 exf6 'fixf6 14 'fih5 g6 1 5 'fie2 ttJf5, when two central pawns and development outweigh king position) 1 1 �h1 (1 1 ttJc3 ttJc6 is normal, intending 12 ttJb5 'ikxe5!; unnecessary but fun was 1 1 . . .ttJg6 12 i.. e 3 'iYxe5 13 f4 'iWb8 14
Advance Variation: 5. . . �b6 f5 .ltd6! 1 5 ii.b5 .i.c6 16 fxg6 i.xh2+ 17 ..t>h1 hxg6-+ Lemmers-Pliester, Amsterdam 1993) 1 1 . . .'Llc6 (1 1 . . .�4 is also good) 12 f4 .i.c5 !? 13 a3 a6 1 4 .:ta2 0 - 0 1 5 b4 i.. a 7 16 'Ll d2 'ii'e 3+ J.Brown-Hummel, Las Vegas 1 997; (b) 10 .:te l?! 'Lle7 1 1 'Llc3 a6 12 iLe3 (12 'iWf3 'Llc6 1 3 'iWg3 0-0-0 1 4 .ltfl .i.e7!+ Lundquist-Dijkstra, corr 1962) 12 . . . 'iWxe5 13 'iWf3 'Llg6 (13 . . . d4!?+ ; 13 . . . �c6+) 14 .ltxg6 hxg6 1 5 .ltf4 �5+ Parkyani-Anka, Budapest 1 990. 10 a6 The main line: it rules out 'Llb5 and clears the a7 square for the dark squared bishop. Given the priority of other subjects I'm going to give only one solution here. I should mention that 10 . . . 'iWxe5 is doing brilliantly theoretically but it is more difficult in practice than 10 . . . a6. ...
The traditional move although in fact 1 1 . . .�4!? has scored 70% for Black in limited outings, e.g. , 12 f4 i.c5+ 1 3 ..t>h1 h5!? 14 ii.e3 ii.xe3 1 5 'ii'x e3 'Llh6 1 6 %:tf3 'Llg4 1 7 �gl l:i.c8+ Duggan-Gunter, St Heliers 1997. 12 'it>h1 The older 12 .:td1 isn't seen much. Black continues 12 . . . 'Llc6 (12 . . . �4!?) and: (a) 13 i..x a6 �xe5 14 �xe5!? (14 i.xb7 �xe2 1 5 'Llxe2 l:r.a7 1 6 i.xc6 .ltxc6 1 7 'Lld4 i.d7 18 iLe3 .:tb7 19 b3 .ltd6; McCann-Jestadt, California 1993) 14 . . . 'Llxe5 15 .i.xb 7 l:ta7 16 .i.xd5 exd5 1 7 l:1.e1 f6 18 f4 i.. c 5+ 19 �h1 d4+ Mnatsakanian-Monin, USSR 1979 (and several later games) . (b) 1 3 .i.b5 �6!? (or 1 3 . . . 'iVh4 14 i-xc6 i-xc6 1 5 i-e3 �c4!) 14 .lta4 ii.c5 ! + . 12 'Ll c 6 13 f4 Unsound looks 1 3 .lte3 �xe5 (13 . . . �4 is good, threatening . . . d4, e.g., 14 g3 �d8 15 �f4 i.. e 7+) 14 f4 �d6 1 5 f5 (15 %:tad1 .lte7 16 f5 �e5) 15 . . . 'i'e5! 16 %:tae 1 iLd6+ . 13 ii.c5 ...
•..
11 'iVe2 11 l:i.e 1 'Lle7 transposes to 1 0 l:te1 above. Others: (a) 1 1 ..t>h1?! 'iVxe5 12 f4 (12 ::te l 'i'd6 13 �f3 ii.c6+) 12 . . . 'i'd6 1 3 �e3 Ji.. e 7 14 �f3 ii.c6 15 i.. d4 'Llf6 1 6 lIae1 g6!+ Eley- Whiteley, Great Britain 1978; (b) 11 �f3 �xe5 1 2 .ltf4 �f6 13 �g3 'Lle7 and 14 iLg5 'Llf5 or 14 .:tfe1 i.c6. 1 1 'Lle7 ...
Yet another line with an over whelming score (+350 performance rating points for Black) . The reason ing behind . . . .ltc5 is to develop, pre vent iLe3, and allow . . .�f2 after 53
Play the French White's l1d1 . 14 iLd2 This both discourages . . . �4 and prepares 15 1:1.f3 with the idea .te3 (14 lIf3?? allows mate on gl). It's slow, but everything seems bad here: (a) 1 4 1:1.dl isn't played much. After 14 . . . lif2 1 5 'iWg4 (15 'iWh5 liJb4!) , how ever, no one has tried my 15 . . . 0-0-0! with the idea ... h5, e.g., 1 6 .td2 h5 17 W6h3 g5! . Instead, 1 5 ... g6 has led to even results in spite of some Black advantage; (b) 14 a3 liJa5! (not the only move; 14 . . . .ta7 1 5 .td2 g6 16 1:1.f3 �6 is quite good) 1 5 .td2 (15 b4 'iWxc3! 1 6 .td2 'iVd4 1 7 bxc5 liJb3 1 8 .te 3 'iWc3 19 1:1.ab l d4!-+ L.Kristensen-Jorgensen, Esbjerg 1996) 15 . . . liJb3 16 1:1.ae l .ta7 1 7 1:1.f3 liJxd2 1 8 'iWxd2 g6 1 9 liJe2 �6 20 licl 1:1.c8 21 1:1.xc8+ .txc8=t Niccoli Naumkin, Bergamo 2002. 14 ... .ta7 1 4 . . . liJb4!? 15 .tb l .ta7 1 6 a3 liJc6 17 .td3 g6 18 1:1.acl (18 1:1.f3 �6!; 1 8 b4 �6 19 .l::t ab l liJ d4 20 �dl 1:1.c8=t Soylu-Lputian, Manila 1 992) 18 . . . �6 19 b4 liJd4 20 'iWdl liJf5!? 2 1 liJa4 'iWd8
54
22 liJc5 iLc6! 23 liJxa6 liJe3 24 .ixe3 .ixe3=t with the bishop pair and queenside pressure, Golod -Barsov, Dieren 1998.
1 5 1:1.ae l 1 5 .l::i.f3 liJb4 16 .l::t afl (16 .tb l?? �gl mate; 16 a3 liJxd3 17 .l::tx d3 'iUf2) 16 . . . liJxd3 1 7 1:1.xd3 �c4+ , e.g., 18 f5 exf5 19 b3 �g4! . 1 5 . . .g 6 1 6 1:1.£3 'i'b6 1 7 ..te3 'iUa5 18 .txa7 liJxa7 19 'i'f2 liJc6 20 'iVh4 lid8 21 �h6 �e7=t . This is from the game L.Hansen-Jorgensen, Holbaek 200 1 . White has little if anything for the pawn.
Chapter Fou r King ' s Ind ian Attack
1 e 4 e6 This chapter will deal with the classical King's Indian Attack and its most popular offshoot. Hence: 4.1 2 d3 4.2 2 "ife2 4.1 2 d3 This introduces the King's Indian Attack proper, so called because White sets up the same basic position as Black often does in the King's In dian Defence to 1 d4. The sequence of White moves tends to be e4/d3/lLld2/lLlgf3/g3/�g2/0-0, with lIel and/or c3 to follow. Over the years, the 2 d3 system has been used by some leading players, notably Fischer (although he didn't stick with it long, preferring the more dynamic 2 d4 and 3 lLlc3). Pal Benko's creative practice of the opening should be noted; it was his games that inspired Fischer to take up the KIA. And although few leading players today employ it con sistently, Lev Psakhis remains one of the KIA's high-level modern adher ents, whereas Morozevich plays both it and the variant 2 "ife2 . And of
course numerous strong players will occasionally dabble in 2 d3 when they wish to avoid a game with a lot of critical theory, as Short, Bareev, and Anand have done. In spite of its pas sive reputation, French players will be defending against the KIA a great deal of the time because it appeals to so many average players, not least because they can play their fIrst 7 or 8 moves without thought. I proposed the same two systems against The King's Indian Attack in the fIrst two editions. Those are still completely valid, but for fun I have added a few new set-ups and eliminated one of the old ones. Most black defences equal ise against the KIA (accounting for its lack of popularity at the very top) , but a reason for my new choices is to give the French player some ideas that are relatively unexplored and there fore difficult for your opponent to re spond to automatically. 2 ... d5 Very often the line presented in 4.12 is entered by playing 2 . . . c5, which is a personal favourite given in the previous editions, but some French players may not wish to play 55
Play the French the Closed Sicilian positions arising from 3 �f3 �c6 4 g3 g6 5 �g2 i.. g 7 6 0-0 �ge7 7 �c3, or those from 3 g3 �c6 4 .tg2 g6 5 �c3 .tg7 6 f4. Need less to say, those relatively harmless systems need not deter one from play ing such a sound and unbalanced sys tem as Black. 3 �d2 I'll discuss move order issues both here and after Black's 3rd moves, for example: (a) 3 g3?!, trying to save the move �d2 and keep open 'iWg4, has a couple of drawbacks, but the main one is 3 . . . dxe4! (3 . . . .td6!? 4 'iWg4 g6 5 "iVe2 �c6= is not as strong but it is in structive, e.g., 6 �f3 dxe4 7 dxe4 e5 8 .tg2 �f6 9 0-0 Ji.g4 10 c3 h6 1 1 h3 .i.e6 and Black is at least equal) 4 dxe4 'iWxdl + 5 'iitx dl �f6! (5 . . . iLc5 also leaves White short of equality, e.g., 6 f3 �c6 7 �d2 e5;) 6 f3 (6 �c3 .tb4!? 7 �b5 .ta5 8 .i.d2 Ji.b6 9 f3 0-0+) 6 . . . �c6 (6 . . . .tc5;) 7 Ji.e3 (7 iLb5 i.. d 7 8 i.. e 3 0-0-0 9 �d2 a6 10 Ji.e2 .te7 11 c3 e5 12 'iit c 2 .te6;) 7 . . . b6!? 8 �d2 Ji.b7 9 .tb5 0-0-0;; (b) 3 'iWe2 will sometimes transpose into 2 'iWe2, but White must also deal with lines like 3 . . . i.. c 5!? 4 exd5 (4 �f3 �e7 5 d4 .tb6 would be experimental, e.g., 6 e5 0-0 7 g3 c5 8 dxc5 .i.xc5 9 .tg2 b6! intending . . . i.. a6) 4 . . . �f6 5 dxe6 i.. x e6 (5 . . . 0-0!? 6 .te3!?) 6 Ji.e3 i.. xe3 7 fxe3 �c6 8 �f3 'ile7 with the idea . . . 0-0-0 and . . . l:.he8, soon forcing central concessions. Mter 3 �d2, the play splits: 4.11 3 ... �f6 4.12 3 .. c5 .
4.11 3 ...�f6 4 �gf3 4 g3 again intends to leave White's options open, for example, 4 . . . c5 5
56
�gf3 (or 5 i.g2 and 6 f4) takes Black out of the system he wants. Black could try 4 . . . dxe4 5 dxe4 e5 here, but then the fl bishop can still develop along the fl -a6 diagonal in conjunc tion with attacks of the e-pawn. That's a decent sequence, but Black has two more interesting ideas: (a) 4 . . . �c6 5 i.g2 (5 f4?! dxe4 6 dxe4 e5 and White has major weak nesses; 5 �gf3 dxe4 6 dxe4 i.. c 5! ; 5 c3 and Black has 5 . . . e5= or 5 . . . a5) 5 . . . dxe4 6 dxe4 e5
(al) 7 �e2?! is rather passive be cause it takes the pressure off e5. Black can play directly by 7 . . . .tc5 or try to force weaknesses by 7 ... .tg4 8 0-0 (8 f3 i.e6 9 0-0 i.. c 5+ 10 �hl 'fie7 1 1 �b3 l:.d8;) 8 . . . �d4 9 f3 �c5 10 'iith l i.e6 11 c3 �c6 12 �b3 Ji.b6; ; (a2) 7 �gf3 is the position that Black has been looking for: White's bishop is passive on g2 while both of Black's bishops develop freely, and in the absence of i..b 5 his e-pawn is simple to defend. See the instructive example of this position below via the order 4 �gf3 �c6 5 g3 dxe4 6 dxe4 e5 7 i.. g2. (b) 4 . . . b6 intends to occupy the a6fl diagonal once White plays .i.g2, or to settle for . . . Ji.b7 if White decides to play passively and waste the g3
King's Indian Attack move. One sequence would be 5 1l.g2 (5 12Jgf3 dxe4 6 dxe4 1l.c5) 5 . . . l2Jc6 6 12Jgf3 dxe4 7 dxe4 �a6! 8 c3 (8 e5 12Jd7; 8 c4 �c5) 8 . . . 'il¥d7 9 'il¥a4 �d3 ! :j:. After 4 12Jgf3 I will give n o less than 3 systems: 4. 1 1 1 4 ...l2Jc6 4.112 4 ... �c5 4. 1 1 3 4 ... b6 4. 1 1 1 4 .. l2Jc6 .
fence; it is an example of how having a tempo less can mean getting more information about what the opponent is doing, in this case g3, and then putting it to use) 7 .i.g2 (this invites Black to try to occupy the a6-fl di agonal by . . . b6 and . . . .i.a6; on the other hand, putting the fl bishop elsewhere would leave White's king side light squares undefended, e.g., 7 �b5 .i.c5 and 8 0-0 i.g4 or 8 .i.xc6+ bxc6 9 'il¥e2 'fWe7 10 12Jc4 .i.a6 1 1 12Jfxe5 0-0 intending . . ..i.d4, or 1 2 c3 'ife6 1 3 b3 .i.xc4 1 4 12Jxc4 12Jxe4 1 5 0-0 12Jxf2 1 6 'iix e6 12Je4+ 17 'It>hl fxe6:j:) 7 . . .tc5 8 0-0 0-0 .
5 c31 Called 'best' by nearly everyone over the last 5 years, 5 c3 is the only reason that 4 . . . l2Jc6 has been some what under a cloud. But I think that Black has some original ideas in these lines. Against moves other than 5 c3, Black will usually play for . . . dxe4 and . . . e5: (a) 5 .i.e2 .i.d6 (maybe an original idea? Black covers e5 and waits on . . . e5; possible too is 5 . . . 1l.c5!? 6 0-0 e5!? 7 exd5 �xd5=) 6 0-0 0-0 7 c3 a5 8 'fWc2 e5 9 exd5 12Jxd5 10 12Jc4 a4!? or 10 . . . .i.e6 1 1 l2Jg5 .i.f5 = ; (h) 5 g 3 dxe4 6 dxe4 e5 (as men tioned above, the idea is that g3 weakens the light squares and loses a tempo if White plays .i.c4 or .i.b5. Compare this with the Philidor De-
(h I) 9 c3 a5 10 �c2 b6 11 12Jc4 .i.a6:j: 12 I:.dl 'il¥e7 13 12Je3 l:!.ad8! 14 12Jd5 12Jxd5 15 exd5 e4! 16 'uel? (16 12Je l? 12Je5! 1 7 i.xe4 12Jg4! - Fappas; 16 dxc6 is best) 16 . . . i.d3 17 'fWb3 "iVf6 18 dxc6 (18 .i.g5 'ilVf5 19 .i.xd8 exf3+) 18 . . . exf3 19 .i.hl .i.e2-+ Fargags Fappas, Germany 1979; (h2) 9 'iie 2 a5 10 12Jc4 'iie 7 (10 ... .i.g4 11 c3 'ilVc8!?) 11 12Je3 (1 1 .i.g5 .i.g4 12 c3 h6 1 3 .i.xf6 'iixf6 14 h3 .i.e6; 1 1 c3 b6 and . . . i.a6, when the queen is not happy on e2) 1 1 . . .l2Jd4! (1 1 . . .'ud8) 12 12Jxd4 exd4 1 3 12Jc4 1l.g4 1 4 �d3 12Jd7 ( 1 4 . . . l:!.fe8 is mentioned by Wintzer; then 15 e5 12Jd7 16 .i.xb7 lIab8 17 .i.c6 'ike6 1 8 57
Play the French .i.xd7 1!i'xd7 gives Black more than enough for a pawn) 1 5 a4 f6 16 i.d2 b6 1 7 �hl l:tad8 1 8 .i.f4? (18 f4 .l:tfe8:j:) 18 . . . i.b4! 19 f3 (19 1!i'xd4?? t'De5) 19 . . . t'Dc5 20 1!i'e2 .te6 21 .te l (2 1 .l:tfdl d3 22 cxd3 t'Dxd3 !+) 2 1 . . .d3 22 cxd3 l1xd3-+ 23 t'Da3 .l:tfd8 24 .tf4 .i.b3 25 t'Db5 t'De6 26 .te3 i.. c 5! 27 i.. xc5 t'Dxc5-+ Matthaei-Wintzer, Bundes liga West 1997; (c) 5 e5 t'Dd7 6 d4 f6 and because of the wasted d3-d4 Black is a whole tempo up on the Guimard Variation (2 d4 d5 3 t'Dd2 t'Dc6 4 t'Dgf3 t'Df6 5 e5 t'Dd7) . Thus: (cl) 7 exf6 'ilVxf6 8 t'Db3 i.. d 6 (or 8 . . . e5 9 dxe5 t'Ddxe5 10 t'Dxe5 1!i'xe5+ 11 .i.e2 i.. d 6:j:) 9 i..b 5 e5! 10 dxe5 t'Ddxe5 1 1 'ilVxd5 .te6 12 1!i'e4 t'Dxf3+ 13 'ilVxf3 'ilVxf3 14 gxf3 .i.d5 1 5 t'Dd4 0-0 1 6 i.. xc6 bxc6 17 i.. e 3 c5- + . (c2) 7 i..b 5 fxe5!? 8 dxe5 i. e 7 9 i.xc6 bxc6 10 t'Dd4 (10 0-0 O-O:j: with the bishop pair and . . . c5 to come) 10 . . . t'Dxe5 1 1 'iVh5+ t'Df7 12 t'Dxc6 'ilVd6 13 t'Dxe7 "ikxe7 14 0-0 0-0 15 t'Db3 e5!:j:.
5 i.d6!? This is one of those ideas that I thought I'd come up with myself but when I sat down to write this chapter I found 13 master games with it! It tries to address the dual issues of White's moves e5 and b4, and I have .••
58
supplied some analysis to indicate that it succeeds in doing so. Normally Black plays 5 . . . a5 to stop b4, which is supposed to favour White but that's not clear either and deserves a close look: 6 e5! (in my opinion the only chance for an advantage; others in clude 6 'ilVa4 .te7 7 e5 t'Dd7 8 d4 0-0 9 i..b 5!? t'Da7 10 i.. d 3 c5 1 1 h4!? f5 12 t'Dfl cxd4 1 3 cxd4 t'Db6 14 'ilVdl .i.d7:j: Friedman-Shulman, Connecticut 2003; and 6 i.e2 i.. e 7 7 0-0 0-0 8 'ilVc2 e5 9 b3 b6 10 i..b 2 .i.a6 1 1 .l:tfe l 'ilVd7 12 I!.adl .l:!.ad8= was a typical re versed Philidor's Defence, Sergeev Lipka, Czech Rep 2002; the typical Philidorian 6 'iVc2 e5 7 i.e2 .i.e7 is equal, and has been played in many games; finally 7 . . . g6!? is an interest ing option) 6 . . . t'Dd7 7 d4 f6 and: (a) 8 i.b5 is given '!' by McDonald and has done well, but may not do much: 8 . . . fxe5 9 dxe5 (9 t'Dxe5? t'Dcxe5 10 dxe5 c6 1 1 i.d3 t'Dxe5:j:) 9 . . . i.e7 10 0-0 0-0 1 1 1!i'e2 t'Dc5?! (1 1 . . . 1!i'e8!? in tending . . . 1!Vh5 - Kindermann; a good idea that should be tried!) 12 t'Db3 t'Dxb3 1 3 axb3 .i.d7 14 i.. d 3 with a small but definite advantage, Kin dermann-Farago, Budapest 1 988; (b) 8 exf6 'ilVxf6 9 .i.b5 i.. d 6 10 t'Dfl 0-0 1 1 t'De3. This position has arisen many times in practical play. Black might try 1 1 . . .t'Db6!?, to guard d5 and prepare . . . e5, e.g., 12 0-0 (12 .i.d3? e5) 12 . . . i.d7 (12 . . . e5?! 1 3 t'Dxd5 ! t'Dxd5 14 i.c4 i.e6 1 5 dxe5 t'Dxe5 1 6 .i.xd5 t'Dxf3+ 17 i..xf3;!;) 13 i.d3 (13 i.xc6 .i.xc6 14 t'Dg4?! 1!i'f5 15 t'Dge5 i.. x e5 16 t'Dxe5 .i.b5+) 13 ... e5 14 t'Dxd5 t'Dxd5 1 5 i.. c 4 exd4 16 i.. x d5+ '>t>h8 1 7 cxd4 i.. g4= . 6 'iVe2 Threatening e5. Black seems to consolidate his position in every case: (a) 6 g3 0-0 7 .i.g2 e5 (7 . . . dxe4 8
King's Indian Attack dxe4 e5 unclear, e.g., 9 'ii' e 2 i.. g4 1 0 ttJc4 a5 1 1 h3 i.. e 6 12 ttJxd6 'iVxd6 1 3 0 - 0 .l:i.ad8) 8 0 - 0 i.. e 6= with the idea 9 ttJg5 i.. g4 10 m3 �e2; (b) 6 b4 e5 (or 6 . . . a6 7 �b2 e5, or 6 . . . a5!? 7 b5 dxe4 8 dxe4 ttJe5 9 ttJxe5 i..x e5 10 "iVc2 Portisch-Ljubojevic, Montreal 1979, and 1O . . . i.. d 6 1 1 ttJc4 �c5;!; was best) 7 exd5 (7 b5 ttJe7 8 �b2 0-0=) 7 . . . ttJxd5 8 b5 (8 i..b 2 0-0 9 ttJc4 �g4=) 8 . . . ttJce7 and Black has no problems; (c) 6 d4 dxe4 7 ttJg5 e3!? 8 fxe3 0-0 9 i.. d 3 e5= ; (d) 6 .te2 0-0 7 0-0 a5 8 l:tel l:te8 (8 . . . e5!?) 9 h3 h6 10 b3 b6 1 1 .tfl e5 1 2 a3 d4= Baklan-Gdanski, Magde burg 2000. 6 e5
4.112 4 ...� c5 I have to admit that not so long ago when I first noticed a game with this move I thought it a very irregular one. But my database has over 1 1 0 games with 4 . . . .tc5 and they don't show me any convincing reason not to recommend it. The idea is a provoca tive one: to lure White's central pawns forward in order to undermine them. Barring that, the bishop will be an aggressively placed piece on c5.
•••
7 exd5?! White should prefer either 7 b4 a6 (or 7 . . . 0-0 8 b5 ttJe7) 8 �b2 0-0= or 7 g3 0-0 8 .tg2 .te6= . 7 . ttJxd5 8 ttJc4 0-0 9 ttJxd6?! Best is 9 i.. d2 .tg4 10 h3 �h5 1 1 0-0-0 'iVf6 1 2 g4 .tg6=i= . 9 . .1lVxd6 10 i.. d 2 i.. f5 1 1 0-0-0 .l:i.ad8+ with excellent play against the backward d-pawn. It seems to me that Black has sufficient play after 4 . . . ttJc6, and in particular both 5 . . . a5 and 5 . . . i.. d 6 should hold their own against 5 c3. ..
.
5 e5 By far the most important move, as seen from: (a) 5 g3? dxe4 6 dxe4 (6 ttJxe4 ttJxe4 7 dxe4 �xf2+ 8 'iitxf2 'ii'x dl 9 i..b 5+ 'ii' d 7+) 6 ... ttJg4, and White (who has fallen for this trap more times than one could possibly imagine) is objec tively lost; (b) 5 �e2 dxe4 (5 . . . ttJc6!? 6 0-0 e5) 6 ttJxe4 (6 dxe4!? e5, or Black can try the funny Philidor-like 6 . . . ttJg4 7 0-0 �xf2+ 8 lIxf2 ttJe3 9 'ii'e l ttJxc2 10 'ii' d l ttJxal 11 b4! with very unclear play) 6 . . . ttJxe4 7 dxe4 'ii'x dl + 8 i.. x dl ttJc6 9 .tf4 .tb6 (9 .. .f6!? 10 i.. xc7?! e5) 10 0-0 f6 l 1 l:r.el?! (1 1 e5=) 1 1 . . .e5 12 i.. e 3 .te6=i= Gushpit-A.Ivanov, Bar linek 2002; (c) 5 d4 is almost never played and may very well transpose to Chapter 59
Play the French 7: 5 . . . dxe4!? is an independent re sponse (5 . . . i.b6 6 i.. d 3 c5 7 dxc5 i.. xc5 transposes to the 3 . . . i.. e 7 Tarrasch chapter; 5 . . . i.e7 is a direct transposi tion to that same chapter with an extra move on both sides, i.e. 2 d4 d5 3 ltJd2 i.. e 7 4 ltJgf3 ltJf6) 6 dxc5 exf3 7 'iWxf3 ltJc6 and Black's free play makes this appear equal, e.g., after S ltJe4!? (S i.d3 ltJe5; S c3 0-0 9 i.. e 2 e5) S . . . ltJxe4 9 'iWxe4 0-0 10 i.. e 2 (10 i.. f4?! e5 11 i.. d 3 f5 12 'iWc4+ 'iith S; 10 i.. d 3 f5 1 1 �e3 e5) 10 . . . e5 1 1 0-0 i.. e 6= .
5 . . ltJfd7 6 d4 (a) 6 g3!? isn't played, but demands some analysis: 6 . . . ltJc6 (maybe 6 . . . i..b 6 7 i.. g2 c5 is a good sequence, e.g. , S 0-0 ltJc6 9 l1e1 i.. c 7 1 0 ltJb3 ltJcxe5 1 1 ltJxe5 ltJxe5 12 ltJxc5 0-0=) 7 ltJb3 i.. e 7 (a fascinating line would be 7 . . . i.b6!? S d4 f6!? 9 c4! dxc4 1 0 i.. xc4 fxe5 ! 1 1 i.. g5 ltJf6 12 ltJxe5 ltJxd4 un clear) S d4 f6!? (S . . . O-O) 9 exf6 (9 i.h3 fxe5 ! 10 i.xe6 e4:j:) 9 . . . ltJxf6 10 i.b5 i.d7 1 1 �e2 (1 1 0-0 0-0 12 i.xc6 i.xc6 13 ltJe5 'ii'e S 14 i.f4 i.b5 15 .:tel i.d6=) 1 1 . . .0-0! 12 i.xc6 i.xc6 13 'iWxe6+ 'iith S with at least sufficient compensation; (h) 6 c3 is subtle: where does Black wish to put his pieces? (h I) 6 . . . 0-0?! 7 d4 i.. e 7 (7 . . . i.b6 S i.d3 c5 9 ltJf1 !? f5 10 ltJg3 ltJc6 1 1 ltJe2 .
60
cxd4 12 cxd4;;1; Kasparov-Rendle, London 1 995) S i.. d 3 c5 9 0-0;;1;. Black has castled too early in this variation that can also arise in Chapter 7; (b2) 6 ... i.b6! 7 g3?! (7 b4!? f6 S d4 fxe5 9 ltJxe5 ltJxe5 10 dxe5 'i!Vh4!? 1 1 g3 �e7 12 i.. d 3 0-0= Bruno-Playa, Buenos Aires 1 994; for 7 d4 c5 see the main line with 6 d4 i..b 6) 7 .. .f6 S i.h3?! (S exf6 'iYxf6) S . . . fxe5 9 i..xe6 ltJc5! 10 i.. x cS ltJxd3+ 1 1 'it>e2 ltJxc 1+ 1 2 l:.xc 1 'ii'xcS+ .
6 ...i.b6!? That Black's system is fundamen tally sound is shown by the fact that 6 . . . i.. e 7 7 c3 (or 7 i.. d 3 c5 S c3) 7 . . . c5 transposes to a Chapter 7 main line in which Black has equal chances! But 6 . . . i.b6 is also full of interest. 7 ltJb3! A logical move played in order to neutralise . . . c5 and to get the knight out of the way of its own pieces. 7 i.. d 3 c5 presents no problems for Black after S c3 (S c4?! ltJc6 9 0-0 cxd4 10 cxd5 exd5 1 1 ltJb3 ltJfS?! [1 1 . . .0-0 12 .l:te1 .:teS 13 i.f4 ltJfS!;] 1 2 i.g5 Wfic7 13 a4!? ltJe6 14 .:tel i.d7!?= Ree U.Geller, Netanya 1965) S . . . ltJc6 9 dxc5 i..x c5 (or 9 . . . ltJxc5 10 i.. c 2?! ltJd7 11 'ilVe2 "ikc7 12 i.a4 0-0 13 i.xc6 bxc6;) 10 ltJb3 i.b6 1 1 'iVe2 'iYc7 12 i.f4 (12 i.b5 0-0 1 3 i.xc6 bxc6 14 0-0
King's Indian Attack a5! 1 5 l:!.e1 �a6=1= Orsagova-Kiss, Ri mavska Sobota 1996) 12 . . . f6! 13 O-O!? (13 exf6?! 'iYxf4 14 fxg7 .l:lg8 1 5 �xe6+ ttJe7; 13 ttJbd4! ttJdxe5 14 ttJxe5 fxe5 15 ttJxc6 exf4 16 ttJe5 0-0=1=) 1 3 . . . ttJdxe5=1= Dulik-Wrba, email 1997. 7 ... c5 8 dxc5 8 c3 ttJc6 9 i.. e 3 cxd4 (9 . . . c4!? 10 ttJbd2 f6=) 10 cxd4 i.. a 5+? (10 .. .£6!) 1 1 ttJxa5 1!t'xa5+ 1 2 i.. d2 1!Vb6 1 3 i.. c 3 0-0 1 4 .td3± Khalifeh-Kristinsdottir, Is tanbul 2000. 8 ...ttJxc5 9 i.. e 3 9 i.. g5 1!t'c7! - Kindermann. 9 ...1!t'c7 10 ttJxc5 .txc5 1 1 �xc5 'iWxc5 12 c3 i.. d 7 13 ft3 To prevent . . . .tb5. Kindermann analyses 13 �d3 .tb5 14 0-0 �xd3 1 5 'iYxd3 ttJc6= . 1 3 ... ttJc6 14 .td3 f6!? During the game Kindermann ex pected 14 . . . d4! , liquidating the centre; after 1 5 .te4 dxc3 1 6 1!t'xc3 1!t'xc3+ 1 7 bxc3 ttJa5 (17 . . . l:tc8 may b e still bet ter) 18 .l:lb 1 �b8 19 ttJd4 b6, Black has at least equalised. 1 5 exf6 gxf6 16 'iYxb7 16 1!t'c2 0-0-0 1 7 0-0-0 e5 18 .tf5 ttJe7 with the advantage - Kinder mann. 16 ....l:lb8 17 'iYc7 lhb2 Black could apparently have drawn by force following the move 1 7 . . . e5!?, which cuts off the queen from the kingside after 18 0-0 (18 0-0-0 ':'c8 1 9 1!Vb7 .l:lb8 20 1!t'a6 e 4 2 1 lthe 1 i.. c 8 2 2 1!t'a4 0-0) 1 8 . . . .l:i.c8 (18 . . . e4 1 9 .l:i.ael) 19 'iWb7 ltb8 20 1!t'a6? (20 �c7 .l:i.c8=) 20 . . . e4+ . 18 0-0 'ii'x c3 19 .l:ladl and White had significant but un clear compensation in Kindermann Sarana-Hungeling, Bad Wiessee 2002. A cute idea is 19 . . .ltxa2 20 ttJd4! �a5! 21 ttJxe6!. 4. 1 1 3 4 ...b6
This is now a well-established solu tion to the KIA. Black's c8 bishop can go to b7 or a6 depending upon what White does, and in particular the de sired g3/i.. g2 set-up will encourage an early . . . i.. a 6. 5 c3 (a) 5 g3 dxe4 6 dxe4 i.. c 5! (Black has scored 58% with this move)
7 .tb5+ (7 �g2 .i.a6 8 c4 1!t'd3 9 'iVb3 'ifxb3 10 axb3 i..b 7 1 1 ttJe5 ttJc6! 12 ttJd3 �e7 1 3 e5 ttJd7 1 4 f4 a5 1 5 0-0 0-0-0 and Black had a solid grip on the position, Heinz-Germer, Will ingen 200 1 ; 7 i.. d 3 e5 8 0-0 i..h 3 9 Mel ttJg4 10 .l:le2 1!t'f6 1 1 c3 g5! 12 'iVel a5 1 3 b3 h5 1 4 a4 h4+ Rolletschek Kindermann, Graz 2001) 7 . . . i.. d 7 8 i.. d 3 ttJc6 9 c3 (9 ttJb3 i.. d 6 10 c3 0-0 11 h3 e5=1= Bhat-G.Rey, San Francisco 61
Play the French 2000) 9 . . . e5 10 b4 �d6 1 1 0-0 0-0 12 �c2 iLg4 1 3 iLe2 a5 1 4 b 5 liJb8!? 1 5 liJc4 liJbd7 16 l:t d 1 "ile7 1 7 liJh4 �e6 1 8 liJf5 iLxf5 19 exf5 h6 20 liJxd6 cxd6 2 1 c4 liJc5 Morozevich-Dreev, Yalta 1995. In this battle of elite grandmas ters, Black has achieved a powerful knight outpost on the open c-file to match up against the bishops. He went on to win, but this position is hard to assess; (b) 5 e5 liJfd7 6 g3 c5. This is a standard . . . b6 position that produces interesting play based about the at tack on e5. Black has done very well here, mainly because the d2 knight has no good squares to go to in order to prepare �f4: 7 �e2 'iic 7 (7 . . . liJc6 8 liJb3 Wlic7 9 �f4 c4!? [9 . . . a5 was a good choice] 10 dxc4 �a6 1 1 liJbd4?! �xc4 12 'i¥e3 �xf1 13 �xf1 liJxd4 14 liJxd4 ..tc5 15 c3 0-0=1= R.Montgomery A. Shaw, corr 1 983) 8 c4 �b7 9 ..tg2 dxc4!? 10 liJxc4 b5 1 1 liJa3 a6 12 0-0 liJc6 1 3 .l:Ie1 �e7 14 liJc2 0-0 1 5 b3 liJb6=1= Lushnikov-Grishanovich, St Petersburg 1999. 5 ... c5
6 e5 (a) 6 'i¥a4+ i.d7 (6 ... �d7 7 �c2 'iVc7! 8 ..te2 liJc6 9 0-0 �d6 10 lIe1 0-0=1= Bauer-Bareev, Cannes 2001) 7 "iVc2 "iVc7 8 ..te2 �e7 9 0-0 0-0 10 lIe1
62
liJc6 1 1 ..tfl .l:Iac8= Van Delft-Hert neck, Tegernsee 2003; (b) 6 g3 is cooperative, because White cannot abandon d3: 6 . . . iLa6!? 7 exd5 (7 c4!?) 7 . . . �xd5 (7 . . . exd5 8 d4 ..txf1 9 liJxf1 ..te7=) 8 liJc4 .txc4!? 9 dxc4 'iYxd1+ 10 �xd1 liJc6 1 1 ..tg2 ':c8= or maybe a touch better for White's bishop pair, Short-Bareev, Sarajevo 1 999. 6 ... liJfd7 A position contested by some top players. 7 d4
7 ... �e7 Black has also played: (a) 7 . . . ..ta6!? 8 ..txa6 liJxa6 9 0-0 �e7 1 0 liJe1 b5 (a battle of wing at tacks begins) 1 1 �g4 g6 (1 1 . . . 0-0 12 liJdf3 �h8=) 12 liJdf3 h6 13 h4!? �6 1 4 �e3 c4= Van Weersel-De Jong, Wijk aan Zee 2002; (b) 7 . . . liJc6 8 �b5 �b7 9 0-0 a6 10 iLa4 g5!? Chudnovsky-Bhat, Catons ville 2000. There should follow 1 1 liJb 1 ! �e7 1 2 �e3;!; . Objectively better was 10 . . . lIc8 intending . . . b5, . . . cxd4, . . . liJb4, and . . . "ilc7. 8 iLb5!? ..ta6!? 8 . . . 0-0 9 0-0 i.. a 6 10 a4 ..txb5 1 1 axb5 a5! = . 9 a4! i.b7! 9 . . . cxd4 10 cxd4 0-0 11 0-0 "iVc8 12
King's Indian Attack :tel liJc6 1 3 .l:i.e3;!; Adams-Bareev, Frankfurt 2000. Mter 9 . . . �b7 !, Anand-Dreev, Lon don [rapid] 1995 continued 10 0-0 liJc6 11 :tel .l:.c8 (1 1 . . .a6!?) 12 liJf1 c4 1 3 liJg3 h5! 14 b4! cxb3 1 5 'ili"xb3 liJa5 1 6 'iVc2 liJc4!?= . 4.12 3 ... c5 This move introduces what is ar guably the most solid counter to the King's Indian Attack that still retains great flexibility and ambition. Black wants to play . . . liJc6/. . .�d6/ . . . liJge7 and create a thoroughgoing imbal ance in the position. He may play in the centre for . . . e5 and/or . . . d4 or on the queenside with . . . b5. I have rec ommended this system for two edi tions and since it is a second (or really, fourth) system in this book I will present the basics without much detail, since the ideas are similar in most subsystems. 4 liJgf3 4 g3 has independent significance:
This has become a popular move order to prevent the . . . liJc6, . . . �d6 move order by hitting d5 before . . . liJge7 can be played. It can lead to: (a) 4 . . . ..ll d6 (this is necessary if Black wants to play the . . . ..ll d 6, . . . liJc6, . . . liJge7 set-up) and White has:
(a1) 5 �g2 liJe7 6 f4 (6 'ili"g4 0-0) 6 . . . 0-0 7 liJgf3 liJbc6 8 0-0 f6 9 'iVe2 J..c 7 10 liJb3 b6 1 1 liJh4 a5! (1 1 . . .e5!?) 12 c3 a4 13 liJd2 b5 (13 . . . :ta7=) 14 f5 b4! 1 5 fxe6 J..xe6 16 exd5 ..llx d5= was the game Sheremetieva -Fomina, De brecen 1 992. (a2) 5 f4 liJc6 6 liJgf3 is similar, al though 6 . . . liJh6!? threatens . . . liJg4 in some lines (after 7 .i.g2?! , for exam ple) and supports . . . e5, e.g., 7 h3 0-0 8 ..ll g2 f5!? 9 e5 .i.c7 10 0-0 liJf7 with a solid position and ideas of . . . g5 and . . . b5; (a3) 5 'iVg4 is the consistent move, playing to weaken Black's kingside. Now there is another split: (a3 1) 5 . . . liJf6!? 6 'iYxg7 l:tg8 7 'ifu6 liJc6 with a large lead in development and unclear compensation, e.g., 8 c3 �d7 9 liJgf3 'ili"e7 and . . . 0-0-0. The only practical example that I could find went 10 'ii' e 3 �c7 1 1 Jtg2 0-0-0 12 'iVe2 h5!? (White was doubtless somewhat better about here, but both sides' moves can be criticised) 13 b3 dxe4 1 4 dxe4 h4!? 1 5 liJxh4 liJg4 16 �b2 liJce5 1 7 c4?! Jtc6 1 8 Jtxe5 liJxe5 19 liJhf3? liJd3+ 20 'it'f1 'iYf6 21 Ir.d1 l:i.d7 22 h4 l:i.gd8 23 l:i.b 1 Jta5 when the extra pawns mean little or noth ing, Vlassov-Kobylkin, Yalta 1995; (a32) 5 ... g6! looks weakening, but it can actually be useful in some posi tions and White's queen loses time because of the threat . . . e5, e.g., 6 'i'e2 liJc6 7 liJgf3 liJge7 8 �g2 (8 h4!? could be met by 8 .. .f6!? with the idea 9 h5?! g5; but also fine is the modest 8 .. :ilc7 9 liJb3 e5 10 h5?! ..ll g 4! 1 1 hxg6 fxg6 12 exd5 liJxd5 13 .lih3 .lixh3 14 l:txh3 'iVd7 15 .l:i.h4 0-0-0 16 .i.h6 liJd4 with a solid edge, Vorobiov-S.Ivanov, St Pe tersburg 2003) 8 .. .f6 (or 8 . . . b6 9 exd5 exd5; or 8 . . . 'iYc7 9 0-0 .i.d7 10 exd5 exd5 1 1 c4 0-0 12 cxd5 liJxd5 13 liJc4 63
Play the French %:tfe8=i=) 9 exd5 (9 0-0 0-0 10 exd5 exd5 1 1 c4) 9 . . . exd5 10 0-0 0-0 and Black's development is freer than White's. (b) 4 . . . ltJc6 5 JLg2 dxe4! (White's point is 5 . . . JLd6? 6 exd5 exd5 7 JLxd5 ltJge7 8 JLg2±) 6 dxe4 b6 and Black's bishop will be unopposed on the a6-f1 diagonal:
games with it in my database. 7 0-0 Black can try to benefit from doing without this move: (a) 7 . . . �c7 8 ltJh4 (8 %:tel .lid7 9 �e2 f6 - compare 7 . . . 0-0 8 %:tel �c7; 8 ltJe1 a-a!? 9 f4 f6 10 ltJef3 �d7 1 1 %:tel %:tac8 12 ltJf1 b5= Rolletschek-Linn, Dearborn 1991) 8 . . . g5!? (or 8 . . . iLe5 9 f4 �f6 10 ltJhf3 g6) 9 ltJhf3 h6 intend ing . . . JLd7, . . . 0-0-0. 9 . . .f6!? is also in teresting; (b) 7 . . . b6 8 ltJh4 (8 %:tel �c7 9 ltJf1 iLb7 10 h4 h6 1 1 c3 dxe4 12 dxe4 �xd1 13 %:txd1 %:td8= Lederer Ma.Tseitlin, Beersheva 1990) 8 . . . �e5!? 9 f4 JLf6 10 ltJhf3 g6 11 e5 �g7 seems more effective than the 7 . . . 0-0 8 ltJh4 iLe5 line below, because now 12 c3 iLa6! prevents d4. •••
7 ltJe2 �a6 8 c3 (8 0-0 �d7 9 c4 %:td8 10 �a4 ltJb4 1 1 'iVb3 ltJe7 12 a3 ltJd3 13 %:td1 ltJc6 with an edge) 8 . . . �d7 9 0-0 %:td8 10 f4 ltJge7 1 1 �e1 e5!? 12 �f2 exf4 13 gxf4 ltJg6! 14 ltJf3 iLe7 1 5 .lie3 0-0 16 %:tae1 �d3 1 7 e 5 f6+ Dutschak-Fischer, Bundesliga 1 998. 4 ...ltJc6 5 g3 �d6 6 iL g2 ltJ ge7 7 0-0
Mter 7 . . . 0-0, White usually plays: 4.121 S %:te l 4.122 S ltJh4
S �e2 and 8 c3 often transpose, but 8 �e2 e5!? 9 exd5 ltJxd5 is an option (10 ltJc4 JLg4 1 1 c3 f6), and 8 c3 b6 9 %:tel allows 9 . . . dxe4 10 dxe4 .lia6= . The basic position for this system, which is solid but gives plenty of winning chances. These appealing aspects account for the roughly 1 500 64
4.121 S %:te l This i s far and away the most popular move. White eyes the possi-
King's Indian Attack bilities of c3 and e5. 8 ... 'i!i' c 7 Here 8 . . . b6 and 8 . . . f6!? (rare but in teresting because it's so flexible) are also played. But the main alternative is 8 . . . .i.c7 9 c3 (9 e5? ttJg6 10 ttJb3 b6 11 d4 cxd4+ ; 9 exd5 ttJxd5 1 0 ttJe4 b6= or here 9 ... exd5; 9 ttJh4 e5!=; finally, 9 'iVe2 can be answered by 9 . . . b6 intending the plan . . . a5 and . . . .i.a6, or 9 . . . e5!?)
A popular position. Black has used some 7 different moves here, all of them playable! These are 9 . . . d4, 9 . . .f6, 9 . . . dxe4 10 dxe4 b6, 9 . . . ttJg6, and: (a) 9 . . . b6 (easily the most fre quently-played move) 10 e5 (10 exd5 ttJxd5 1 1 ttJc4 �b7= , or again 1 0 . . . exd5=) 1O . . . a5 (or 10 . . . lt>h8! with the idea . . .f6) 1 1 ttJf1 (1 1 d4 .i.a6 [or 1 1 . . .ttJf5 12 ttJf1 f6 1 3 exf6 'ifixf6=] 12 ttJf1 cxd4 1 3 cxd4 'iWd7 1 4 h4 ttJb4 15 l:r.e3 l:r.fc8 16 a3 ttJbc6 1 7 b3 ttJf5; Dzhumaev-Saltaev, Abu Dhabi 2000) 1 1 . . .i.a6 1 2 h4 d4 1 3 c4 'i!i'd7 14 ttJlh2 f5?! (14 .. .f6!) 15 exf6? (15 a4;!;) 15 . . . gxf6 16 ttJg4 (16 �h3 ttJf5 17 ttJg4 It>h8 - Karpov) 16 . . . e5 17 �h3 'i!i'e8 1 8 .i.h6 l:i.f7 1 9 i. d 2 It>h8; Sznapik Karpov, Skopje 1972. (b) 9 ... l:r.b8 10 a3 (10 e5 b6 11 d4 ttJf5 12 ttJf1 f6 1 3 exf6 'iVxf6= Ostojic Pytel, Kikinda 1976; 10 'i!i'c2 b5 1 1
ttJb3 c4! 12 dxc4 bxc4 13 ttJc5 l:i.b5 14 �e3 iLb6 1 5 b4 cxb3 16 ttJxb3 �xe3 1 7 ':'xe3 'iYc7=) 1O . . . a5 1 1 a4 b6 12 'iVc2 i.. a 6 1 3 .i.fl 'ifid7 14 ttJb l ttJg6; intending .. .f5, Makropoulos-Grivas, Athens 1989; (c) 9 . . . a5 10 a4 b6 11 exd5 exd5 12 ttJb l (12 d4 .i.g4= Medina-Portisch, Palma de Mallorca 1967) 12 . . . �g4 13 ttJa3 d4= Vaganian-Sokolov, Minsk 1986. 9 c3 (a) 9 'iVe2 f6 10 c3 might transpose after 1O . . . �d7, but Black also has 10 . . . d4 1 1 ttJc4 e5 12 .i.d2 (12 a4 �e6) 12 . . . b5! 13 ttJxd6 \liVxd6; RosenbergBukhman, USSR 1 974; (b) 9 b3 �d7 10 i.b2 d4! 11 ttJc4 (1 1 c3 dxc3!? 12 i.. xc3 e5 1 3 ttJc4 f6, at least = , Pacis-Galego, Manila 1 992) 1 1 . . .e5 12 a4 b6 13 'iVd2 f6; Calvo Karpov, Madrid 1973; (c) 9 ttJh4 g5!? (9 ... �d7 10 f4 f6=) 10 ttJhf3 f6 11 d4 ttJxd4 12 ttJxd4 cxd4 1 3 exd5 e5 Foygel-Serper, Seattle 2003; I prefer Black here.
9 .. i.d7 Or 9 . . . b5 10 'ifie2 f6, e.g., 11 a3?! l:b8 1 2 d4 cxd4 1 3 cxd4 'ifib6 14 ttJb3 b4 1 5 e5? fxe5 16 dxe5 ttJxe5! 1 7 ttJxe5 bxa3+ was V. Georgiev-Peev, Tsarevo 200 1 . 10 'iVe2 f6 ! .
65
Play the French dxc4 bxc4 17 exd5 exd5 18 .te3 �f5; Mooses-Noroozi, Tehran 200 l . 1 0 ... i.d7 Not the only move. Also effective was 10 . . . d4 1 1 a4 i.. d 7 12 liJc4 e5 1 3 �d2 dxc3! 14 bxc3 .te6 1 5 l:t f2 l:i.ad8 16 �f1 'ilVd7 17 i.e3 b6 18 f5 i.. f7; in Nicholson-Fernandez Romero, Seville 2003.
Now White can't play e5, whereas Black has several plans including . . . b5 and . . . d4 with . . . e5. This position has done well for Black, for example, 1 1 liJf1 (1 1 a3 can be met by 1 1 . . .a5 12 a4 �e8!? intending . . . i.f7 or 1 1 . . .d4!? 12 cxd4 cxd4 1 3 b4 e5) 11 . . . d4 (1 1 . . .i.e8 12 liJh4 .tf7=) 12 .i.d2 (12 a4 e5 13 liJ3d2 a6 14 liJc4 b5 15 liJxd6 'ilVxd6 16 axb5 axb5 17 .l:lxa8 l:txa8+ Savic-Kalevic, Belgrade 2001) 12 ... e5 1 3 .l:i.ec l l:tfc8!? 1 4 h3?! .te6 15 liJ lh2? (but these positions are noto riously hard to play) 1 5 . . . \\!Vd7 16 h4 b5 1 7 cxd4 cxd4 18 liJel a5 19 f4 .tb4+ A.Zaitsev-Gufeld, USSR Ch 1969. 4.122 8 liJh4 'ilVc7 One option is 8 .. .f5!? 9 f4 b5 1 0 c3 l:tb8 1 1 exf5 exf5 12 liJdf3 b4 13 c4 d4 14 l:tel h6 15 i.d2 \\!Vc7= Meier Boensch, Berlin 1 992. 9 f4 f6 This position has discouraged play ers from 8 liJh4. Now neither e5 nor f5 achieve anything, so the whole plan looks artificial. 10 c3 Black's manages to get both queen side expansion and central play after 10 liJhf3 b5 1 1 c3 .td7 12 'iVe2 .l:i.ae8!? 1 3 �el liJg6 14 liJf1 a5 15 h4 c4 16
66
l l liJb3!? 11 \\!Vh5!? .te8 12 \\!Ve2 .i.f7 1 3 liJdf3 .l:i.ae8!? (13 . . . dxe4 14 dxe4 e5) 14 e5? (14 .i.e3 was about equal) 14 .. .fxe5 1 5 fxe5 liJxe5 16 liJg5 liJ7c6 1 7 .i.h3 h6! 1 8 �xf7 liJxf7 1 9 liJxe6 liJcd8!-+ Rigo Sax, Magyarorszag 1976. 1 1 ...d4! 12 c4 a6 1 3 .i.d2 :ab8 14 l:tcl i.. e 8 A typical manoeuvre for Black, re routing the bishop to eye the queen side. 1 5 'iVe2 Or 15 g4 .tf7. After 15 'ii'e 2, Nevednichy-Horvath, Odorheiu Secu iesc 1993 continued 15 . . .�f7 16 liJf3?! (16 'it>hl e5!) 16 . . . e5! 17 f5 b5; with the idea 1 8 cxb5 l:txb5. 4.2 2 \\!Ve2 An old move, tried by Chigorin and revived in recent years by Morozevich and others. I don't really think that objectively 2 'ilVe2 poses many prob-
King's Indian Attack lems for Black but it has scored bril liantly versus the unaware and you'll probably see it some time since there are over 5000 games with it in my database! The main point of 2 'iVe2 is to meet 2 . . . d5?! with 3 exd5 'iVxd5 4 tLlc3. Otherwise White usually enters a King's Indian Attack set-up where the queen is oddly placed except for early e5 thrusts. One problem for White is that Black can choose sys tems in which e5 is either impossible or ineffective. Two of these are:
4.2 1 2 ... tLlf6 4.22 2 ... c5 4.21 2 ...tLlf6
I think that this has been underes timated. Black prepares ... d5 and de-
velops, so to make sense of 2 'ife2 White has to commit his centre, ar guably too early. 3 e5 Natural but not very challenging. (a) Mter 3 d4, Black can of course play 3 . . . d5, but I like 3 . . . c5!
and: (al) 4 d5 d6 5 c4 exd5 6 cxd5 iJ.. e 7 (6 . . . g6 aims for a Benoni system with the queen misplaced on e2, so logical is 7 tLlc3 'ife7 8 tLlf3 i.. g 7 9 i.g5 h6 10 .th4 0-0 11 0-0-0 g5 [or 11 . . J:te8!] 12 iJ.. g 3 tLlh5 with standard counterplay) 7 tLlc3 0-0 8 tLlf3 tLla6!? (8 . . . .l:te8; 8 . . . i.g4) 9 'ifdl tLlc7 10 a4 ,Ub8 11 i.d3 a6= . (a2) 4 e5 tLld5 5 c4 (5 tLlf3 cxd4 6 tLlxd4 d6 7 tLlf3 tLlc6:j:) 5 . . . tLlb4!? (or 5 . . . tLlc7) 6 dxc5 (6 tLlf3 cxd4 7 tLlxd4 tLl8c6:j:) 6 . . . .txc5 7 tLlf3 tLl8c6!? 8 tLlc3 (8 a3?! 'ii' a 5 9 .td2? tLlc2+ 10 Wdl "iVa4 1 1 tLlc3 'i¥b3-+) 8 . . . 'ifa5! 9 a3 tLld4 10 tLlxd4 i.xd4 11 .tf4 tLlc6 and e5 falls; (b) 3 tLlf3 is arguably the main line. Rustemov's treatment was instruc tive in the following game: 3 . . . d5 4 e5 (4 d3 is one line White was presuma bly not very interested in playing when he chose 2 'ife2. Black's easiest solution is to develop by 4 . . . .te7 5 g3 0-0, but 5 . . . tLlc6 is also interesting, to meet 6 i.g2 by 6 . . . dxe4 7 dxe4 e5, and 67
Play the French 6 e5 ttJd7 7 �g2 0-0 8 0-0 by 8 .. .f6) 4 . . . ttJfd7 5 d3 (5 d4?! c5 6 c3 b6 Rustemov; 5 g3 b6 6 �g2 iLc5 !? 7 d3 ttJc6 8 c3 a5 9 0-0 �a6 10 :el �e7 1 1 'i!Vdl ttJc5 1 2 �f1 d4! 1 3 c4 �b7:j: Troianescu-Cappello, Bari 1 9 71) 5 . . . c5 6 g3 ttJc6 7 �g2 �e7 (7 . . . b5!?)
'iWc2!? looks safe enough) 14 l:r.b l a5 15 h5 �d7 16 0-0 0-0 1 7 b3 (here Rustemov offers 1 7 a3?! a4+ and 1 7 l:r.fdl l:Ifd8:j:) 1 7 . . .l::t fd8 18 a 3 �c6 1 9 b4 �xg2 20 'if;>xg2 ttJa4 (20 . . . axb4!? 2 1 axb4 ttJa4 2 2 c5 ttJc3 23 'fIf3 'ii'b 5+ or 24 .l:.b3 .l:.a4! - Rustemov) 21 c5 Fe dorov-Rustemov, Vilnius 1997. Black was better after 2 1 . . :iVc6+, but Rustemov finds the stronger 2 1 . . .ttJc3! 22 'fIf3 (22 cxb6 ttJxe2) 22 . . . 'ii'b 5+; (c) 3 f4 doesn't make much sense. One game went 3 . . . d5 4 e5 ttJg8!? (4 ... ttJfd7 5 ttJf3 c5) 5 ttJf3 ttJh6!? (or 5 . . . c5) 6 ttJc3 c5 7 ttJdl ttJc6 8 g3 ttJf5 9 c3 �e7 10 �h3 :b8 1 1 d3 b5 with a queenside attack underway, Khou seinov-Villamayor, Bled 2002. 3 .ttJd5 4 ttJf3 (a) 4 d4 c5 transposes to 3 ttJf3 c5 4 e5 ttJd5; or there's 4 . . . d6 'unclear' Rustemov; (b) 4 c4 ttJf4 (or 4 . . . ttJb4 5 d4 c5= in tending 6 a3 'iVa5!) 5 'i'e4 ttJg6 6 ttJf3 d6 (or 6 . . . c5) 7 exd6 �xd6 8 d4 c5 9 ttJc3 cxd4 10 ttJxd4 ttJd7!? 1 1 �e3 ttJf6= Motwani -Nogueiras, Istanbul 2000; (c) 4 g3 d6 5 d4 c5 6 c3?! cxd4 7 cxd4 b6! (7 . . . ttJc6 8 ttJf3 dxe5 9 dxe5 ii.c5:j:) 8 �g2 ii.a6+ Arques Lopez Barria, Alicante 2000. 4 ... c5! 5 c4 Instead, 5 ttJc3 ttJc6! 6 ttJxd5 exd5 is a very nice version of a Nimzowitsch Sicilian Defence - 5 g3 ttJc6 6 �g2 d6 equalises on the spot. 5 ...ttJb4! Disrupting White's development. The alternative was 5 . . . ttJc7 6 d4 cxd4 7 ttJxd4 d6!? 6 d3 d5 7 'fIdl ! Best. 7 a3? 'i¥a5! 8 ttJc3 d 4 9 llb l dxc3 10 axb4 'iVa2 is awful. 7 ... dxc4 Black is slightly better after __
8 h4 (8 0-0 g5! - Rustemov; then . . . g4 is threatened to win the e-pawn, and �f4 is prevented, so 9 h3 h5 1 0 g 4 hxg4 11 hxg4 'fIc7 12 .l:.e l f6! 1 3 exf6 ttJxf6+ could follow, when Black intends . . . e5 and can answer 14 �xg5 by 14 . . . ttJxg4 1 5 iLxe7 ttJxe7 [or 1 5 . . . ttJd4!? 16 'ii' d l 'if;>xe7+] 16 ttJe5 'iWxe5 17 'fIxe5 ttJxe5 18 .l:.xe5 ttJg6 19 l:r.el ttJf4+ with . . . iLd7 and . . . 0-0-0) 8 . . . h6 (or 8 . . . b5) 9 c4 (Rustemov as sesses 9 ii.f4!? b5 10 c4 bxc4 11 dxc4 iLa6 12 ttJbd2 as unclear, although the aggressive 12 . . . ttJb4! with the idea 13 0-0 (what else?) 13 . . . g5! 14 iLe3 d4 looks very strong) 9 . . . dxc4 (9 . . . 0-0 is less committal and probably equal) 10 dxc4 ttJd4 1 1 ttJxd4 (Rustemov mentions 1 1 'iWe4, when 1 1 . . .0-0! 12 ttJxd4 cxd4 1 3 'lWxd4 is met by 13 . . . 'iVa5+) 1 1 . . .cxd4 12 �f4 (12 0-0 'iWc7 1 3 l::[ e l g5!?) 12 . . :iYb6 1 3 ttJd2 (13 0-0 g5 14 hxg5 hxg5 15 i.cl 'iVc5:j: ; 13 b3 'fIa5+!? [or 1 3 ... ttJc5] 1 4 ttJd2 iLb4 15 0-0 iLc3) 13 . . . ttJc5 (a sensible con tinuation, although 13 . . .'iVxb2 14 .l:.b l 68
King's Indian Attack 7 . . . Jl.. e 7. S dxc4 "iVxdl+ 9 �xdl b6 10 a3 liJ4a6 1 1 liJc3 ii.b7 1 2 ii.e2 liJc6 and Black, intending . . . 0-0-0 and . . . liJd4, has a nice advantage. The 2 . . . liJf6 system guarantees an imbal ance and looks like fun.
8 . . . d6 and, for example, 9 liJbd2 h6 10 a4 e5!, transposing to 7 ... e5. 8 lDbd2 0-0 9 liJc4 d6 10 a4 h6 11 c3 .i.e6
4.22 2 ... c5 A good reply: as I asked in PTF2: Who wants the queen on e2 in a Sicil ian? I will look at just a small num ber of examples.
3 liJf3 3 f4 lDc6 4 lDf3 1J.. e 7 5 liJc3 d5 6 d3 (6 e5 liJh6 intending to continue . . . liJf5) 6 . . . liJf6 7 e5 liJd7 8 g3 b 5 ! 9 1J.. g2 'iVb6 10 0-0 b4 1 1 lDdl c4+! 12 ii.e3 (12 d4 c3! with the idea ... 1J.. a6) 12 . . . 1J.. c 5 13 cJilhl Jl.. a 6:j: Lendauf Uhlmann, Graz 1 99 1 . 3. . .lDc6 4 g 3 g 6 5 .i.g2 1J.. g 7 6 0-0 liJge7 7 d3 e5 7 ... 0-0 is a normal Sicilian: If White plays 8 liJc3, Black has either 8 . . . d6, with a standard position except for the strange and committal "iVe2; and if he plays 8 c3 Black can again play
Black is well centralised and White has no breaks, whereas the prospects for . . . d5 and/or kingside attack are good. 12 l:tdl b6?! Weakening. Instead, 12 . . .f5! threatens . . .f4 and 1 3 exf5 lDxf5 clamps down on d4 while preparing the attack. 13 .i.d2 'i¥c7!? Better is 1 3 ... g5 ! . 14 b4! 1 4 .l:!.ab l?! d5! . 14. . .cxb4 1 5 cxb4 d5 16 exd5 liJxd5 1 7 l:r.ac1 liJd4! I S 'iWe l? 1 8 liJxd4 was better when 1 8 ... exd4 is equal. 1 8 ... liJxf3+!? Correct was 1 8 . . . liJb3! 1 9 .:r.b l liJxd2 20 liJcxd2 .l::i.fc8 with a solid advan tage. 19 .i.xf3 .l:!.acS 20 b5 fts 21 'iVe2 l:r.fdS with double-edged play, Galego Baches Garcia, Havana 2003.
69
Chapter Five Exchange Variation
1 e4 e6 2 d4 d5 3 exd5 The Exchange Variation was probably the most popular variation of the French Defence in the late 1 9th century. It is not a particularly imaginative line. 3 ... exd5 The exchange of pawns has actu ally deterred some players from using the French Defence. With only one open file, they reason, both sides will mindlessly exchange their heavy pieces and a drawish position will ensue. But there is a huge hole in this reasoning: in most cases, neither side's rooks belong on the e-file. Why? Because there will be no points of penetration available for them along the file, e.g., squares like e2, e3 and e4 for White and e7, e6, and e5 for Black will customarily be covered 2 or 3 times by that side's pawns, knights, and bishops . Moreover, only one such defender is really necessary. So if, for example, White doubles or triples along the e-file while Black is blithely mounting a kingside pawn storm (backed by his rooks and queen) , the second player has every chance of winning. Furthermore, Black can al-
70
ways make the struggle imbalanced should he so choose. This has been pointed out and demonstrated by great players for at least 70 years. Thus my comments from the first two editions and they still hold true. Although the Exchange Variation appeals to players who are trying to draw against stronger players, allow ing equality on the third move as White may not be the way to go about that. Be aware that it's a strategy that has failed miserably throughout the years. White's main moves are: 5 . 1 4 c4 5.2 4 lDf3 5.3 4 .td3
He has other choices, but Black can always achieve an imbalance if he wants to: (a) 4 lDc3 .tb4 is dealt with in Chapter 8, under the order 3 lDc3 i.b4 4 exd5; (b) 4 c3 is slow: 4 . . . i.. d6 (or 4 . . . lDc6; then 5 .tb5 might look appealing, but c3 doesn't go well with this move, i.e. lDc3 is preferable; compare the re-
Exchange Variation versed positions in Chapter 8). Now if 5 i.. d 3, 5 . . . ltJc6 transposes to 4 .id3 ltJc6 below. On 5 ltJf3, 5 . . . �g4 will often tranpose to 4 ltJf3, and 5 . . . ltJc6 6 ltJf3 .tg4 is also possible; (c) 4 .te3 is passive. Then 4 . . . ltJe7 intending . . .ltJf5 and 4 . . . �d6 5 ltJc3 ltJe7 are dynamic continuations; in practice, 4 . . . ltJf6 has also scored well; (d) 4 'iVf3 has been played a bit more recently, but nearly every game develops independently and Black seems to get equality with 4 . . . ltJf6, 4 . . . i.. d 6, 4 . . . c6, or even 4 . . . ltJc6. A nice example of the latter was 4 . . . ltJc6 5 .ib5 ltJf6 6 ltJe2 .ig4 7 'iVc3 .id6! 8 .ixc6+ bxc6 9 �xc6+ .id7, and Black has more than enough compensation: 10 Wa6 ltJg4!? (10 . . . 0-Q:i:) 1 1 h3? 'iVf6! 12 ltJbc3 (to stop . . . i.. b 4+) 12 . . :i!Vxf2+ 13 'it>dl Wxg2- + Bykhovsky-Ulibin, Aaland-Stockholm 1997. I like the straightforward 4 ... i.. d6 5 c3 (5 �d3 ltJc6) 5 . . . ltJe7. A sequence which has come up more than once follows 4 . . .ltJf6 5 h3 (versus . . . .ig4, which is often followed by . . . .th5-g6) . Black has won several games from this position, for example, 5 . . . c5 (or 5 . . . .id6 6 i.. d 3 0-0 7 ltJe2 c5 8 dxc5 i.. xc5 9 0-0 ltJc6 10 .tf4 .l:le8 1 1 ltJbc3 ltJe5 12 .ixe5 1:txe5 13 l:[adl .td7 1 4 ltJg3 �6 =i= Kos-Mohr, Ptuj 2000) 6 �b5+!? ltJc6 7 ltJe2 �6! 8 'i'd3 !? c4 9 'i'e3+ .ie6 10 �a4 .id6+ with a big lead in development, Hermann Vaganian, Bundesliga 1992; (e) After 4 �f4, Black has equalised with 4 . . . .if5, 4 . . . c5, and: (e l) 4 . . . ltJf6 5 .id3 c5 6 dxc5 (6 'iVe2+ .te7! 7 dxc5 0-0) 6 . . . i.. xc5 7 ltJf3 0-0 8 0-0 ltJc6 9 c3 .l:le8 10 ltJbd2 ltJe4= Meyers-Gleserov, Tzrkva 1 990; or 10 . . . i.. g4! ; (e2) 4 . . . i.. d6 led t o a nice example in Prie-Vaisman, Nimes 1990: 5 �xd6
Wxd6 6 ltJc3 (6 c3 ltJf6 7 �d3 0-0 8 ltJe2 b6 9 0-0 c5 10 ltJd2 �g4!? 1 1 f3 �d7 12 'i'el l:te8 13 �f2 ltJc6=i= Ste fanova-Kindermann, Vienna 1996) 6 . . . i.. f5 7 .id3 .txd3 8 'i'xd3 ltJd7!? (8 ... ltJc6=) 9 ltJf3 0-0-0 10 0-0-0 ltJgf6 1 1 h3 (White would lose material af ter 1 1 .l:lhe l l:the8 12 'it>b l ltJe4! 1 3 ltJxe4?! dxe4 14 1:txe4? ltJc5! 1 5 dxc5 'iVxd3 16 cxd3 l:txe4) I 1 . . J::th e8 12 l:[de l ltJe4! 1 3 ltJxe4? dxe4 14 .l:lxe4 ltJc5 15 dxc5 n6+ 16 b l (16 'iVe3 l:[xe4) 16 . . . l:r.xd3 17 l:txe8+ .l:ld8 18 .l:lxd8+ �xd8 1 9 ltJe5 'i'f4-+ Piroth Apicella, Sautron 2003 . 5 . 1 4 c4
This move has become one of White's main options. It can trans pose to 4 ltJf3 and 5 c4, but also has independent significance. I will give two solutions. 4 ... �b4+ A move this book helped to bring to attention. I think that it casts into doubt the value of 4 c4 as a practical weapon; in fact, Black has scored 53% with a clearly superior performance rating after many games. Also good is 4 . . . ltJf6, which is played about 10 times as often! Then White has: (a) 5 ltJf3 .tb4+ (or 5 . . ..ig4) 6 ltJc3 (6 �d2 .txd2+ 7 ltJbxd2 tends to be 71
Play the French bad in such positions because in the isolated pawn position following . . . .i.xd2 and . . . dxc4, the exchange of bishops favours Black, e.g., 7 . . . 0-0 8 .i.e2 dxc4 9 liJxc4 liJbd7 10 0-0 liJb6 1 1 liJe3 .i.e6 1 2 liJe5 'iVd6'+ Ellenbroek Stellwagen, Wijk aan Zee 2001) 6 . . . 0-0 7 �e2 dxc4 8 0-0 (8 �xc4 .i.g4) 8 . . . �xc3!? 9 bxc3 .i.e6 10 �b 1 lLlbd7! 11 .i.g5 liJb6'+ Herrera- Campora, Ar gentina 1987; (b) 5 liJc3 �b4 6 cxd5 (6 a3 ..i.xc3+ 7 bxc3 0-0 8 .td3 lLlc6 9 liJe2 dxc4 1 0 �xc4 liJa5=) 6 . . . liJxd5 7 .td2 0- 0 8 �d3 c5! 9 dxc5 liJxc3 10 ..i.xc3 ..i.xc3+ 11 bxc3 liJd7 12 c6 liJc5 ! 1 3 ..i.c2 lIe8+ 1 4 liJe2 "iie 7+ Kuijf-Korchnoi, Tilburg 1992. 5 liJc3 liJe7 5 ... liJf6 has been played often with success and an appealing idea is 5 . . :iVe7+!? 6 .i.e3 (6 'iVe2? ! liJc6 7 liJf3 �g4) 6 . . . lLlf6, e.g., 7 ..i.d3 .i.e6 8 c5 0-0 9 h3 b6,+ Perrin-Knott, USA 1 857(!).
6 liJ f3 6 a3 .i.xc3+ 7 bxc3 0-0 8 liJf3 lLlbc6= intending as always . . . liJa5 and/or . . . ..i.g4. Then 9 ..i.d3!? dxc4 10 .i.xc4 lLld5 1 1 .i.xd5 l:!.e8+ 12 ..i.e3 'iVxd5 1 3 h 3 (13 0 - 0 .i.g4) 1 3 . . . 'iVc4 1 4 'iVc2 b6+ was the game Speck-Luther, Liech tenstein 1993. 6 0-0 ...
72
6 . . . ..i.g4! will usually transpose and pretty much amounts to the same thing. But I actually prefer this less common order because it puts imme diate pressure on the centre. These useful early pins are characteristic of the Exchange French.
7 a3 (7 'ii' a4+ liJbc6 8 lLle5 �e6= ; 7 �e2 lLlbc6 S a3 ..i.xc3+ 9 bxc3 would transpose to this note, and here 7 . . . dxc4 S .i.xc4 0-0 9 0-0 liJbc6 is the main line) 7 . . . ..i.xc3+ S bxc3 liJbc6 (S . . . O-O is again the main line) 9 ..i.g5 (9 a4 liJa5'+; 9 cxd5 liJxd5 10 ..i.d2 "iie 7+ 1 1 'iVe2 �xf3 12 'ii'x e7+ liJcxe7 1 3 gxf3 lLlg6'+) 9 . . . 0-0! 10 .i.e2 (10 cxd5 "iixd5 threatens capture on f3, hitting the g5 bishop) 10 . . . f6 (10 . . . 'iWd7 may be more accurate: 1 1 0-0 dxc4 12 .i.xc4 lLld5 with the idea ...lLla5, ... b 5 etc.) 1 1 il.. d2 lLla5= . Black's control of the light squares compensates for the bishop pair: if 12 c5, 12 . . . b6!; and 12 cxd5 'iWxd5 13 nb 1 (13 0-0 liJc4 14 .l:.e1 nfeS 1 5 l:t.b 1 b 5 16 a4 a6) 13 ... lLlc4! plans . . . b5. 7 a3?! 7 �e2 dxc4 8 .i.xc4 .i.g4 9 0-0 liJbc6 is a typical position in which Black has pressure on the d-pawn and light squares. Several games have gone 10 .t e 3 lLlf5!? (10 . . JlbS!? 1 1 a3 ..i.xc3 12 bxc3 lLld5 1 3 �d2 lLla5 14 .ta2 b5 and
Exchange Variation the light squares balance White's bishops) 1 1 Wd3 tDd6: 12 �d5 (12 ..tg5!? 'ii'c 8 13 ..td5 ..tf5!? 14 'ii'd 1 l:le8=) 12 . . . ..tf5 13 'iid 1 (13 'iVe2 tDa5 1 4 i.. g 5 'iic 8 15 l:lfc1 l:.e8 1 6 'i¥f1 c6 1 7 a 3 i.xc3 1 8 l:txc3 i.. e 6 19 ..txe6 'iVxe6 20 b4 tDac4:j: Ashley-Remlinger, New York 1994) 13 . . . tDe7 14 ..tb3 c6 1 5 tDe5 'it>h8 16 ]:tc1 f6 1 7 tDd3 ..txc3 1 8 bxc3 tD d 5 1 9 tDf4 tDxe3 2 0 fxe3 'iVe8:j: Waitzkin-Shaked, Mermaid Beach Club 1997. 7 ... i.xc3+ 8 bxc3 tDbc6 5.21 4 ... ..tg4 5.22 4 ... ..td6
Of course, 4 . . . tDf6 is fully playable (and a current favourite among grandmasters). I should point out that 5 c4 (a key white plan) then al lows 5 . . . ..tb4+ 6 tDc3 tDc6, which is of particular interest because the light square theme after . . . ..tg4 can be ef fective. 9 �e2 9 c5 b6!:j:, since 10 cxb6 axb6 gives Black the a-file and all the light squares, perhaps even with . . . ..ta6 next. 9 ... dxc4 10 ..txc4 tDd5 1 1 i.. d 2 z:te8+ 1 2 .1i.e2 'fWe7+ 13 c4 tDb6 Or 13 . . . ..tf5!? with the idea 14 cxd5? i.c2. 14 i.e3 i.g4 1 5 0-0 I:ad8 16 d5 ..txf3 1 7 gxf3 tDe5 1 8 ft3 tDbxc4! 19 ..txc4 tDxf3+ 20 'it>g2 We4! with an overwhelming attack, Santo Ro man-Shaked, Cannes 1997. 5.2 4 tDf3 White's most common move; it is very natural but allows double-edged play because of the possibility of a . . . ..tg4 pin at some point. Black has:
5.21 4 ... ..tg4
5 h3 (a) 5 ..tf4 ..td6 6 i.xd6?! 'i'xd6 7 c3 We7+ 8 'it>d2 tDf6+ Karolyi-Vaisser, Clichy 1991; (b) 5 ..te2 ..td6 (5 ... tDc6 6 i.f4 i.d6 7 'iid2 tDf6 8 'iie 3+?! tDe4! 9 tDbd2 0-0
73
Play the French 10 ttJxe4 dxe4 1 1 ttJe5 ttJb4 12 �d2 i.xe2 13 'ilVxe2 f6 14 'iVc4+
A position that has been played of ten and is crucial for assessing 4 . . . i.. g 4. The lines are very concrete. 9 g4 White does well to chase this bishop immediately and eliminate
74
some options for Black after 9 0-0-0 such as 9 .. .f6, 9 . . . h6, 9 .. .f5, and 9 . . .'ilVb4!?, e.g., 10 g4 i.g6 1 1 tL'lh4 (1 1 ttJe5?! ttJxe5 12 dxe5 d4 13 'iVd2 .te7!+) 1 1 . . .'ilVa5 12 ttJxg6 hxg6 1 3 'iVb5! .tb4 ( 1 3 . . . ttJf6) 14 'ilVxa5 ttJxa5 15 ttJb l ttJf6 16 c3 .td6= Rutter S.Williams, London 1993. 9 ... i.. g6 10 0-0-0 ttJb4!? A direct tactical solution that has barely been tried. It may be no better than the options, but leads to wonder ful complications. Black has also played 10 . . . 'i¥e8 here, but I prefer 10 .. .f6, covering e5 and allowing . . . 1.f7 after ttJh4. Black's main idea is . . . tL'la5: 1 1 .tg2!? (1 1 ttJh4 .tf7 12 a3 Vi'd7 !? clears the bishop on f8: 13 i.g2 ttJa5!? 14 f4 ttJc4 1 5 .l:1.he l ttJh6= , e.g., 1 6 'iVd3 .td6! 1 7 ttJxd5? Wib5 1 8 b3 .txa3+ 1 9
1 1 a3 This is supposed to be the refuta tion.
Exchange Variation (a) The natural 1 1 l:i.d2 allows 1 1 . . .f6!? (to cover e5; also, . . . i.f7 in some lines strengthens c4 while pro tecting d5) 12 �g2 ! �f7 (another idea would be 12 . . . �d7 13 a3 ttJc6 14 ttJh4 i.. f7 , e.g., 15 .l:.e1 ttJa5 16 'iVf3 ttJc4 1 7 .l:.de2 ttJe7!? 18 g 5 ttJc6 1 9 gxf6 gxf6 20 'iYxf6 .l:.g8 with dynamic compensa tion) 13 .l:.e1 c6!? (13 . . . 'ilVd7=) 14 'iVd1 'ilVc7 15 a3 ttJa6 16 l:.de2 .td6 17 'iVd2 ttJe7= (aiming for f4) 18 I1d1 ttJg6 1 9 .l:.ee 1 ttJb8!? ( 1 9 . . . ttJf4=i=) 2 0 ..tf1 ttJd7 21 ttJa4 l:1.de8 Y2 - Y2 Mimon-Sanchez Jimenez, Spain 1993. Black still stands a little better; (b) 1 1 ttJe1 ttJf6 12 a3 ttJc6 13 �g2 (13 ttJf3 ttJe4) 13 . . . ttJe4 14 ttJxe4 .txe4 1 5 f3 i.. g6 16 f4 �e4= , or here 16 . . . .l:.e8 17 f5 'ilVxe3+ 18 '1Wxe3 �xe3 19 fxg6 hxg6 20 �xd5 ttJd8= . 1 l ... ttJxc2 12 �f4 This is Ulibin-Vilela, Santa Clara 199 1 . Instead of the tempting 12 . . . ttJa1?! 13 b4!, Black might have tried: 12 ...ttJxa3! which results in fascinating com plications.
13 bxa3 13 �xe7 i.. x e7 14 bxa3 i.xa3+ 1 5 'it>d2 ttJf6 transposes. 1 3 .. :iVxa3+ 14 'i¥b2 .tb4 1 5 'ilVxa3 Too slow is the alternative 1 5
i.. d2?! h5! (15 . . . 'ilVa5!?) 16 i.g2 ttJf6 1 7 g 5 'ilVxb2+ 18 cJilxb2 i.xc3+ 19 i.xc3 ttJe4 20 i.e 1 l:.he8=i= . 1 5 ... .ixa3+ 1 6 'it>d2 ttJf6 Black has three pawns and enough play for the piece, for example: 17 ttJe5! ..tb4! 18 f3 c5 19 ttJxg6 hxg6 20 'it>c2 c4 2 1 .tg2 b6 Planning . . . 'it>b7-c6 followed by ad vancing the 3 connected passed pawns . 22 :b l ! 2 2 iL e 5 cJilb7 23 f4 iLxc3 24 cJilxc3 ttJe8! 25 f5 gxf5 26 gxf5 'it>c6=i= . 22 ... i.xc3 23 'it>xc3 'it>b7 24 l1he l 'it>c6 Riskier is 24 . . . l:i.de8!? 25 i.. e 5! cJilc6 26 f4 a5 27 g5 ttJd7 28 i.xg7 l:i.xe1 29 .l:.xe 1 .l:.g8 30 i.. e 5 b5, although the pawns are dangerous enough to hold the balance. 25 11e2 Now 25 iLe5 a5 26 g5 ttJe8 achieves nothing. 25 ...l:i.de8 26 l:ieb2 lie7= There follows . . JIhe8, and the idea of . . . ttJh7, .. .f6, . . . g5 can be surpris ingly effective. This is of course just sample analysis, but it includes sev eral exact moves for White that let him escape from being overrun by the passed pawns. I think that the ver dict is an objective one. 5.22 4 ...i.d6 5 c4 (a) 5 ttJc3 is rare because 5 . . . c6 re stricts the White knight, and 5 . . . ttJe7 is also fine; (b) 5 .td3 ttJe7 (5 . . . ttJc6 might transpose to 4 .id3 ttJc6) 6 0-0 .tg4 (or, again, 6 . . . ttJbc6) 7 �g5 (7 h3 �h5 8 c4 dxc4 9 .txc4 0-0 10 g4 i.,g6 1 1 ttJe5 c5!? 1 2 ttJxg6 ttJxg6 1 3 dxc5 .txc5 14 'i8xd8 l:i.xd8 15 �g5 .l:.d7= Thesing Knaak, Bundesliga 1992; and here 1 1 . . .ttJbc6 would have been more con-
75
Play the French frontational) 7 . . . ltJbc6 8 c3 'iWd7 9 ltJbd2 f6 10 .th4 h5 1 1 b4 ltJd8 (or 1 1 . . .0-0-0) Mokcsay-L.Portisch, Zala karos 199 1 ; Black's coming kingside pawn storm in this type of position will be seen again under 4 iLd3. 5 .. c6 This has the advantage of reserv ing a later . . . ltJe7, when Black's knight will be better placed than White's in the event of cxd5. 5 . . .ltJf6 6 c5 i.e7 is also played. 6 ltJc3 6 cxd5 cxd5 7 i.d3 (7 ltJc3 ltJe7 transposes, and here 7 i.g5 f6! only helps Black) 7 . . . ltJe7 8 i.g5 h6 9 ..th4 'i!lb6!= Holubcik-Vavrak, Tatranske Zruby 2000. 6 ltJe 7 .
•••
man) 13 . . . l:.fe8 14 ..tb2 ltJg6! 15 g4 (15 g3 'iVe6 ! ; 1 5 .tc1 .l:.xe2!) 1 5 ... ltJf4 16 ltJ e 5 i.g6 1 7 i.. f3 ltJxh3+ 18 �h1 ..te4- + Johansson-Renman, Sweden 1 982; (c) 7 c5 tends to be harmless in these positions because it releases the pressure on the centre, e.g., 7 . . . i.. c 7 8 .td3 (8 i.. g5 f6 9 .te3 0-0 10 i.. d 3 i.f5=) 8 . . . 0-0 (8 . . . i.f5 is the more common move order, to force the kind of position that arises next) 9 0-0 h6 10 h3 (10 'iWc2 ..tg4! 1 1 ltJe5 .txe5 12 dxe5 ltJd7:j:) 10 ... i.f5 (10 ... .te6 would be more enterprising) 1 1 lie1 i.. x d3 12 'iWxd3 ltJd7 13 ltJh4!? (versus . . . ltJg6) 13 . . . .l:.e8 14 iLd2 ltJf8 1 5 b4 'iYd7 1 6 a4 ltJfg6 1 7 ltJxg6 ltJxg6= George-Hutchings, England 1986. 7 dxc4 8 iLxc4 0-0 9 0-0 9 i.e3?! ltJd7 10 'iVd2 ltJb6 1 1 iLb3 ltJbd5 12 0-0-0 a5! 13 ltJxd5 ltJxd5:j: with the idea . . . a4-a3, Jellison J.Watson, Denver 1 974. 9 ltJd7 10 .l:1.el 10 i.. g 5 ltJb6 11 i..b 3 ( 1 1 i.. d 3 f6m l 1 . . .�h8 (l 1 . . .iLg4!? has a similar idea, e.g., 12 h3 .th5 13 g4 i.g6 14 ltJe5 'it>h8! intending ... f6) 12 'i¥d2 (12 i.. xe7 'iYxe7 1 3 .l:1.e1 'iYc7 14 d5 i.. g4=) 12 ... f6 (this plan is worth noting, since it has not been played in several recent games with this variation) 1 3 i.f4 i.b4!? ( 1 3 . . . i.xf4:j:) 14 a 3 i.. a 5 1 5 h 3 (15 i.. a 2 ltJbd5:j:) 1 5 . . . ltJed5 16 i.. g 3 i.e6 17 i.a2 ltJa4 18 'iWe2 i.xc3 1 9 'iYxe6 :e8 2 0 'iYf5 ..txb2-+ Frederick J.Watson, New York 1981. 10...ltJb6 1 1 i.b3 ltJbd5 11 . . . ltJed5 12 ..tg5 f6 13 i.h4 .l:1.e8 14 i.. g3 .txg3 15 hxg3 .te6 16 'iVd2 � - � was Neubauer-Borges Mateos, Santa Clara 2002 . 12 h3 .tf5 12 . . . YWb6!? 13 i.c2 f6 keeps things complicated. ••.
...
7 iLd3 (a) 7 iLe2 might transpose after 7 . . . dxc4, but Black might also simply develop by . . . 0-0 and . . . i.f5, in view of the passivity of White's bishop; (b) 7 cxd5 cxd5 yields a static cen tral pawn structure where knights are at least the equal of bishops. Thus White is ill-motivated to waste two tempi to win the d6 bishop by 8 ltJb5?! 0-0 (8 . . . ltJbc6) 9 ltJxd6 'iWxd6 10 .te2 .tg4! 11 0-0 ltJbc6 (with the idea . . .ltJf5) 12 h3 .th5 13 b3? (but 13 i.e3 f5! or 13 g4 ..tg6 14 ltJh4 f5 ! - Ren-
76
Exchange Variation 1 3 lDe4 .i.xe4 14 .l:i.xe4 Y2 - Y2 Leko Varga, Budapest 1993. I prefer Black here after 1 4 ... 'ii'c 7 and . . ..l:i.ad8, but it's a matter of taste. 5.3 4 .i.d3
Formerly the main move and cer tainly an important one. Here an old semi-arranged draw used to continue 4 . . . i.. d 6 5 lDe2 lDe7 6 0-0 0-0 7 .i.f4 i..f5 or something similar. I recom mend lines which immediately break the symmetry. This section is mostly unchanged from earlier editions, as I see no major changes in the theory. 4 ...lDc6 One of the main reasons that 4 lDf3 has been replacing 4 .i.d3 is the move 4 . . . c5, against which White hasn't succeeded in finding any edge.
White's problem is that the bishop on d3 is not placed effectively for play against an isolated pawn. I will treat these lines briefly, since nothing much has occurred since the last edi tion: (a) 5 'ii'e 2+ i.. e 7 6 dxc5 lDf6 7 lDf3 0-0 8 0-0 i..x c5 9 i.. g 5 lDc6 10 lDbd2 h6 1 1 i.. h 4 i.. g4 12 J:tfel g5! 13 i.. g 3 lDh5:j: Mantovani-King, Lugano 1 989; (b) 5 c3 lDc6 6 lDf3 (6 lDe2 c4!? 7 .i.c2 .i.d6= with the idea . . . lDge7) 6 . . . .i.d6 7 .i.e3!? c4 and 8 .i.e2 i..f5= or 8 i.. c 2 .tg4= ; (c) 5 lDf3 lDc6 6 1IVe2+ .te7 7 dxc5 lDf6 8 h3?! 0-0 9 0-0 .txc5 10 c3 l:t.e8 1 1 'iVc2 'iVd6! 12 lDbd2? 'i'g3! 13 .i.f5? (13 lDg5!:j:) 13 . . . .l:i.e2-+ Tatai-Korch noi, Beersheva 1978; (d) 5 dxc5 (the main move) 5 ... ..txc5 6 lDf3 lDf6 (or 6 . . . lDe7) 7 0-0 0-0 8 lDbd2 (8 .i.g5 lDc6 9 lDc3 .i.e6=) 8 . . . lDc6 9 lDb3 .tb6 10 c3 .tg4 11 i.. e 2 1IVd6= Moldobayev-Orlov, Belgorod 1 989.
5 c3 5 lDf3 .i.g4 6 c3 i.. d 6 transposes. 5 lDe2 is natural and can't be bad, but Black can play aggressively by 5 . . . i.. g4 (or 5 . . . .td6) and: (a) 6 c3 i.. d 6!? 7 1IVb3 lDge7 8 lDa3!? (8 'iWxb7 l:tb8 9 'ii' a 6 0-0 10 0-0 J:tb6 1 1 'iWa4 lDg6 followed by . . .'i'h4) 8. . . a6 9
77
Play the French f3 i.f5 ! 10 i.xf5 liJxf5 1 1 0-0 0-0 12 'iYc2 'iYf6+ Lau-J.Watson, Philadel phia 1978; (b) 6 f3 i.h5 7 c3 (7 liJf4 i.g6; 7 liJg3 'iYh4 8 0-0 -td6 9 'iYe l + �d7! 1 0 f4 .l:1e8 planning . . . liJf6-g4) 7 . . . i.d6 8 .if4 (8 0-0 'iYf6!? or 8 . . . i.g6 with the idea 9 f4 f5 ! and . . . liJf6-e4) 8 . . . liJge7 9 0-0 i.g6 10 i.xd6 'iYxd6 1 1 f4?! (1 1 liJa3 0-0-0 12 'iYd2 h5!? 1 3 liJf4 i.f5 1 4 i.xf5+ liJxf5 1 5 liJd 3 a6 1 6 liJc2 h4= Lenic-Karer, Kranj 2001) 1 1 . . .'iYe6! 12 .l:1f3 liJf5 (12 ... i.h5 !) 1 3 �f2 liJd6 1 4 .l:1 e 3 i.e4 1 5 liJg3 f5=i= Sollid-J.Watson, Gausdal 1981. A particularly mischievous move is 5 i.b5, which transposes into a Winawer with colours reversed (White has wasted a move) ! As we show in the line 3 liJc3 i.b4 4 exd5 exd5, this is objectively equal, with chances for both sides (Black must be careful not to overextend) . 5 i.d6 6 liJf3 (a) 6 liJe2 'iYh4! (6 . . . i.g4 is often played; and Alekhine suggested 6 .. :tWf6!?, for example, 7 0-0 liJge7 8 liJd2 i.f5 9 liJf3 i.xd3 10 'iYxd3 h6 planning . . . 0-0-0 and a kingside at tack) ...
Here there are some fascinating ideas: (a l) 7 liJg3 liJh6!? (7 . . . i.g4!) 8 i.xh6 78
"iVxh6 9 0-0 liJen Westerinen-Kava lek, Solingen 1986; (a2) 7 g3 with: (a2 1) 7 . . . 'iVh3 8 liJf4 i.xf4 9 i.xf4 'iYg2?! (9 . . . liJf6 10 "iVe2+ �d8 unclear) 10 'iYe2+! .ie6 (1O . . . �d8 1 1 'iYfl) 1 1 'iYfl 'iYxf1 + 1 2 .l:1xfl 0-0-0 1 3 liJa3! i.h3 14 .l:1g1 .l:1e8+ 1 5 �d2;!;; (a22) 7 . . . 'iVh5 8 i.f4 (8 liJf4 'iYxdl + 9 �xdl liJf6 10 .l:1el + liJen Apsenieks Alekhine, Buenos Aires 1939, or 9 . . . i.g4+=) 8 . . . -tg4 (8 . . . i.xf4 9 liJxf4 'iYxdl + 10 �xdl i.g4+= - McDonald) . I called 8 . . . i.g4 clearly better for Black, but Justin Horton points out the line 9 i.xd6 cxd6 10 iVb3 ! . Then 10 . . . 0-0-0 1 1 liJf4 .l:1e8+ 12 �d2 "iVh6! (with the idea . . . liJf6-h5) is interesting but hardly a clear advantage for Black: 1 3 �c2 ! (13 'iYxd5? liJf6 14 iVb5 a6 1 5 'iYc4 i.f3+ ; 13 liJa3? liJf6! and 14 h4? liJh5 o r 14 �c2 -tf3 1 5 .l:1hfl g5 16 liJb5 gxf4 1 7 liJxd6+ �c7 18 liJxe8+ .l:1xe8- +) 13 . . .liJf6 14 liJd2 �b8 15 h3 .l:1c8! 1 6 'iYa4 i.d7 unclear; (a3) 7 liJd2 i.g4 (several readers and players have discussed this line, and some of the material is based upon that discussion; thanks to all) 8 'iYc2 (8 iVb3 0-0-0 9 'iYxd5 liJf6+ Alekhine) 8 . . . liJf6 (8 . . . 0-0-0 9 liJfl , and Black is better after either 9 . . . liJf6 or 9 . . . g6 10 i.e3 liJge7 1 1 0-0-0 i.f5 12 liJeg3 i.xd3 1 3 .l:1xd3 h6=i= Alekhine Winter, Nottingham 1 936) 9 liJfl 0-0 10 i.e3 i.h5 l l liJeg3?! i.xg3 12 liJxg3 i.g6 13 h3 .l:1ae8 14 liJfl i.xd3 1 5 'iYxd3 liJe4+ Ledwon-Bednarski, Po land 1975; (b) 6 'iYf3 liJf6!? (6 ... i.e6 is a solid alternative, or 6 . . . liJce7 7 i.f4 liJf6 8 liJe2 -tg4 9 'iVg3 i.xe2 10 i.xe2 i.xf4 1 1 'iYxf4 0-0 12 i.d3 liJg6 1 3 i.xg6 .l:1e8+ 14 �dl hxg6=i= Lazic-Kosten, Varallo 1991) 7 i.g5 (7 h3 0-0 8 liJe2 .l:1e8 9 i.g5 i.e7 10 i.e3 liJe4! 1 1 i.xe4
Exchange Variation dxe4 12 'iWxe4 tLlb4!+ Malanyuk-Psak his, Moscow 1983) 7 . . . �g4 8 'iVxg4! (J.Horton's suggestion; 8 �xf6? .txf3 9 .txd8 i.. xg2-+; 8 'ii' e 3+?! 'iit d 7! 9 tLle2 %:te8=i=) 8 . . . tLlxg4 9 i.. x d8 �xd8 (9 . . . %:txd8 10 tLlf3 0-0 1 1 0-0=) 10 tLlf3 %:te8+ 1 1 �fl tLlf6 (1 1 . . .g6 12 g3 �d7 13 tLlbd2 f5 14 �g2=) 12 g3 tLle4 1 3 tLlbd2 f5 14 �g2 �d 7 1 5 Itae 1 = . 6 ... i.. g4 6 . . . tLlge7 usually transposes. 7 0-0 tLlge 7
14 'iVxd3 'ii'f5=). What follows comes from a news group discussion: 1 1 . .. f6 12 tLlg3 i.. g6 13 'iVc2 (13 .tfl h5) 13 . . . �xd3 14 'ii'x d3 h5!? (14 . . . i.. xg3! 1 5 fxg3 tLlf5 16 g4 tLld6, analysis by C. Schulien) 1 5 tLlh4 g5! (15 . . . �xg3 16 fxg3 g5 1 7 tLlg6 Hagerty-Jeffrey, Par sippany 1996) 16 tLlhf5 (16 tLlg6 .l:.h6 17 tLlxe7+ tLlxe7 18 h4 .l:i.g8 with good play) 1 6 . . . �xg3 17 tLlxe7+ tLlxe7 18 fxg3 g4 unclear - Mannien. 8 'iVd7 9 tLlbd2
8 Itel (a) 8 i.. g 5 'iWd7 9 tLlbd2 f6 10 .te3 (10 .th4?! O-O-O=i= intending ... tLlf5 and, e.g., . . . g5, . . . h5-h4 is a position where Black's attack plays itself) 10 . . . 0-0-0 11 c4 (1 1 'ii' a 4 �b8 12 b4? �xf3 13 tLlxf3 tLlxd4+ Trzaska-Koepf, Bundes liga 1988; 1 1 b4 J:1dg8 12 tLlb3 tLld8! with the idea 13 tLlc5 'iVe8=i=, when Black covers a4 and threatens the move . . . 'ii'h 5) l 1 . . .tLlb4!? (1 1 . . . dxc4! 12 tLlxc4 'iitb 8 13 tLlxd6 'iVxd6=i=) 12 i.. e 2 dxc4 13 tLlxc4 tLlbd5 14 �3 i.. e 6!? 1 5 Itfc1 �b8 16 tLlfd2 tLlf4 1 7 � f3 .td5 1 8 tLl e 4 tLle2+!=i= Marcal-J.Watson, St Paul 1 982; (b) 8 h3!? ('inaccurate' - Psakhis) 8 . . . �h5 9 %:tel 'iVd7 10 tLlbd2 0-0-0 (or 10 . . . 0-0!?, since 1 1 i.. xh7+? �xh7 12 tLlg5+ �g6 doesn't work) 11 tLlfl (1 1 'iWc2!? Itde8 12 b4 .tg6 13 tLlb 3 i.. xd3
This one of the recurring positions of the French Exchange Variation. 9 ... 0-0-0 An all-out move. Also perfectly good is 9 . . . 0-0 10 h3 .tf5 1 1 tLlf1 %:tae8 as in Tal-Korchnoi, USSR 1955, when White tried 12 �xf5!? (12 �g5= ECO; but I think Black is slightly better here also, for example, 12 . . . h6 13 i.. h 4 tLlg6 14 i.. g3 tLlf4) 12 . . . 'iVxf5 (12 . . . tLlxf5!?) 1 3 tLle3 'ili'd7 14 i.. d2 tLld8 15 'iVc2 c6=i= . 10 b4 tLlg6 10 . . . Itde8 1 1 b5 tLld8 12 'iVa4 �b8 13 i.. a 3 f6 also worked well for Black in Kholmov-Psakhis, Kiev 1 984. 11 tLl b3 11 b5!? tLlce7 and Black is well off; a typical line was 12 'ili'a4 �b8 1 3 �a3 ( 1 3 tLle5 i.. x e5 14 dxe5 f6!) 1 3 . . . tLlf4 (13 . . ..txa3 14 'iYxa3 f6, e.g.,
...
79
Play the French 15 l:!.ab l?! .l:.he8 16 .l:.b4 tLlf4 17 lla4 tLlc8+) 14 �f1 f6! 15 �c5 tLlc8=i= Z.Nik olic-Barlov, Yugoslavia 1 986. I 1 ..Jlde8
A position from Spielmann-Mar oczy, Sliac 1932. It is easy for White to go wrong here, although equality
80
seems the fair result. The obvious 12 i.. d2? fails to 12 ... tLlh4 1 3 i.. e 2 �xe2! 14 'iixe2 �xf3 1 5 gxf3 �3. A better line is 12 �e3! tLlh4!? 13 �e2 tLlf5 14 �d2 .l:.xe2!? 1 5 'iixe2 tLlh4 Havski Vladimirov, USSR 1956. This is given as unclear in ECO. The only defense to . . . .i.xf3 and . . . 'iVh3 is 16 b5!, when 16 . . . tLlb4!? (16 . . . i.. xf3? 17 bxc6 holds due to the counterattack) 17 tLle5! (17 cxb4 �xf3 18 gxf3 'iih 3-+) 1 7 ... �xe2 18 tLlxd7 tLlc2 19 l:txe2 tLlxal 20 tLldc5 tLlxb3 21 axb3 tLlf5 leaves no advan tage for either side. Objectively, the Exchange Varia tion is of course equal. But we have seen that there is ample leeway for the stronger player to outthink his opponent and win. This is all we can ask for from any opening.
Chapter Six Tarrasch Variation : I ntroduction a n d 3 . . . c 5
1 e 4 e 6 2 d4 d 5 3 lbd2 With apologies for the digression, here are some general remarks that could be of particular importance to the readers of previous editions. After many years of modern practice with the Tarrasch Variation, I feel that theory has coalesced enough to come to some general conclusions. First, it will come as no surprise that 3 lbd2 does not threaten the survival of the French Defence. The knight on d2 exerts no pressure on d5 and to some extent gets in the way of the devel opment of White's other pieces. Lead ing players are using the Tarrasch Variation much less than 3 lbc3 and only somewhat more often than 3 e5. On the other hand, it is equally clear that 3 lbd2 is a solid continuation which, if handled well, does not ex pose White to great risk and gives some prospects for advantage. How does theory stand? It now seems to me that all of Black's most popular responses either equalise or stay within the bounds of a normal small edge for White; none, however, is clearly preferable. Since choosing a system (or two) against the Tarrasch
is among the most important deci sions that a French player is faced with, let me briefly try to characterize the main options: (a) 3 . . . c5 4 exd5 exd5 has a long history of being safe, and it hovers near theoretical equality. Players like Bareev, Short, Vaganian and Dreev have been holding their own and reaching that equality in the line 5 �b5+ �d7 (or 5 lbgf3 lbf6 6 i.b5+ �d7) , but it offers few winning chances between opponents of equal strength and is hardly attractive to the average player; (b) 3 . . . dxe4 (also played via 3 lbc3 dxe4) is popular among the world's leading players at the moment (com pare the related lines in Chapter 1 4) . This i s also in the 'safe but dull' cate gory; (c) the variation from the previous two books, 3 . . . lbf6 4 e5 lbfd7, is still going strong, but I feel that the the ory now runs too deeply to present in a repertoire. Not only are the lines forcing, but most readers will prefer something requiring less memorisa tion. I'm also not excited about re peating most of the same material 81
Play the French from the previous edition, but those who wish to use 3 . . .ltJf6 can use it as a base from which to work. One can also choose 3 . . . ltJc6 and 3 . . . a6 to meet the Tarrasch, but they are marginal moves that have never become completely respectable. For this book, I have chosen two other systems. This chapter examines an ambitious variation from the previous editions, 3 . . . c5 4 exd5 'i!Vxd5. There is still much to be discovered here. In the next chapter, a repertoire is of fered with 3 . .. .i.e7, a fascinating move that is far from being understood, much less worked out. It is a leading defence at present and has thus far proven impervious to White's as saults. 1 e4 e6 2 d4 d5 3 ltJd2 c5
Black's philosophy here is to break up the centre immediately even at the cost of some time. This frustrates White's attempt to cramp Black (by e5), and in our main line leads to a Sicilian-like position in which Black hopes that his central majority will counteract White's his superior de velopment. Those are essentially po sitional considerations but in practice the play can be rather sharp . White has two main replies: 82
6.1 4 ltJgf3 6.2 4 exd5
Others are relatively rare: (a) 4 J..b 5+ ltJc6 5 exd5?! (5 ltJgf3 cxd4 6 ltJxd4 J.. d 7 transposes to 4 ltJgf3) 5 . . . 'i!Vxd5 6 ltJgf3 cxd4 7 .i.c4 Valenti-Huss, Lugano 1 980. White is a full tempo down (i.b5-c4) on the main line. Most ambitious is 7 . . . 'iWc5!? (7 . . . 'iWd6) 8 'iWe2 e5 9 ltJg5 ltJh6 10 ltJge4 'iWe7; (b) 4 dxc5 iLxc5 (more active than 4 . . . ltJf6 5 exd5 'iWxd5= Bronstein Khasin, Kislovodsk 1 968, with the idea 6 ltJb3 'iWxdl + 7 Wxdl ltJa6!=) 5 iLd3 (5 ltJb3 iLb6 6 .i.b5+ ltJc6 7 exd5 exd5 with an ideally active isolated IQP position, since Black's bishop came to c5 in one move: 8 a4 ltJge7 9 ltJe2 0-0 10 0-0 .i.g4; Bronstein Makarichev, Reykjavik Open 1990) 5 . . .ltJf6 (or 5 . . . ltJc6) 6 'iVe2 (6 e5 ltJfd7 7 'iWg4 and 7 . . . 0-0 8 ltJgf3 'ilc7 or 7 . . . ltJxe5!? 8 'iVxg7 ltJxd3+ 9 cxd3 :f8 10 ltJgf3 ltJd7 1 1 iVxh7 'ilf6 12 'ilh4 'iWg6 with compensation, Feldman Botvinnik, Leningrad 1931) 6 . . . ltJc6 7 ltJgf3 0-0 and we have arrived at a position with Black a full tempo ahead of the corresponding line with 3 . . . ..te7 4 .i.d3 c5 5 dxc5 etc. as de scribed in the next chapter. This makes life easy: (h I) 8 e5?! 8 . . . ltJg4 9 0-0 f6 10 ltJb3? (but 10 exf6 ltJxf6 is comfortable and at least equal; also possible is 10 . . . 'iWxf6 1 1 ltJb3 ..tb6) 10 . . . ..txf2+ 1 1 ':'xf2 ltJxf2 1 2 'i!Vxf2 fxe5-+ Landen bergue-Huss, Silvaplana 1993; (h2) 8 0-0 'iic 7 (or 8 ... .i.b6!) 9 e5 (9 c3 .i.b6) 9 ... ltJg4 10 .i.xh7+ (10 ltJb3? .i.b6 1 1 ..txh7+ Wxh7 12 ltJg5+ Wg8 1 3 'i!Vxg4 ltJxe5 14 'ilh5 'iWxc2) 10 . . . Wxh7 1 1 ltJg5+ Wg8 12 'ilVxg4 'iVxe5 1 3 'ilh5 'tWf5;;
Tarrasch Variation: Introduction and 3. . . c5 (c) 4 c3 is played more often than it should be. Among several answers Black has 4 . . . cxd4 (or 4 . . . ctJc6 5 ctJgf3 cxd4) 5 cxd4 ctJc6 6 ctJgf3 (6 exd5? 1!i'xd5 wins a pawn) 6 . . . 1i'b6! (more pointed than 6 . . . dxe4 7 ctJxe4 �e7 8 Ji.d3 ctJf6=) 7 exd5 (7 1!i'a4?! �d7 threatens . . . ctJb4 and . . . :c8: 8 �b5 dxe4 9 ctJxe4 a6+ Suechting-Alekhine, Carlsbad 1 9 1 1) 7 . . . exd5 8 ctJb3 (8 1!Vb3 1!i'xb3 9 ctJxb3 ctJf6 10 a3 .i.d6 1 1 �d3 0-0 12 0-0 i.. g4:j: Hedenstroem Nygren, Skelleftea 1 999) 8 . . . i.. g4!:j: (only 8 ... ctJf6 has been seen here) 9 Ji.e2 ctJf6 10 0-0 i.. d6 1 1 h3 .i.xf3 12 Ji.xf3 0-0 and as so often in these op posing d-pawn positions, the knights are better than White's bishops. 6.1 4 ctJgf3
This is a common choice. White wishes to sidestep our main line, since now 4 . . . ctJc6 5 exd5 'iix d5 6 i.. c 4 1!i'd6? loses its point after 7 ctJe4. But Black has two good answers: 4 ... cxd4 4 . . . ctJf6 is also fine: a) 5 i..b 5+ i.. d 7 6 �xd7+ ctJbxd7 7 e5 ctJe4 8 ctJxe4 dxe4 9 ctJg5 cxd4 10 'li'xd4 1!i'a5+ 11 i.. d2 'iix e5= Van der Wiel-P.Nikolic, Lucerne 1 989; b) 5 e5 ctJfd7 6 c3 ctJc6 7 Ji.d3 i.. e 7 transposes to 3 ctJd2 �e7 4 ctJgf3 ctJf6
5 e5 ctJfd7 (5 . . . ctJe4!?) 6 Ji.d3 c5 7 c3 ctJc6, and here Black also has the popular options 7 . . . g6 and 7 . . . 1!Vb6 8 0-0 g6; c) 5 exd5 ctJxd5 (5 . . . exd5 6 �b5+ Ji.d7 7 i.. x d7+ ctJbxd7 8 0-0 �e7 9 dxc5 ctJxc5 10 ctJb3 ctJce4 has held up well recently: 1 1 ctJbd4 'iid 7 and 1 1 ctJfd4 1!i'd7 1 2 f3 ctJd6 1 3 ctJc5 'ilc7 14 ctJd3 0-0 1 5 �hl .l:tfe8 M.Pavlovic Bukal, Saint Vincent 2002) 6 ctJb3 (6 c4 ctJf6 7 ctJb3, and best seems 7 . . . ctJa6!? - Psakhis; 6 ctJe4 can be met by 6 . . . cxd4 7 ctJxd4 and 7 . . . i.e7 or 7 . . . a6 intending . . . e5; Black also has 6 . . . ctJd7; finally, Black can answer 6 dxc5 by 6 . . . �xc5 7 ctJe4 i.e7 8 c4 ctJb4! = or by 6 . . . ctJd7 7 c4 ctJ5f6 8 ctJb3 1!i'c7 9 Ji.e3 ctJxc5=) 6 . . . ctJd7 (or 6 . . . cxd4 7 ctJbxd4 Ji.e7, e.g., 8 g3 0-0 9 Ji.g2 ctJc6 10 ctJxc6 bxc6 1 1 0-0 �a6 12 .l:i.el i.f6= Djuric-Uusi, Tallin 1981) 7 c4 (7 Ji.g5 Ji.e7 8 �xe7 1!i'xe7 9 Ji.b5 cxd4 10 1!i'xd4 0-0 11 0-0-0 a6= Pavlovic Short, Catalan Bay 2003) 7 . . . ctJ5f6 8 dxc5 (8 i.e2 �e7 9 0-0 0-0 10 i.f4 b6 11 ctJe5 i.b 7 12 �f3 1!i'c8= Arnason Gulko, Groningen 1 990) 8 . . . 1!i'c7 9 g3 ctJxc5 10 �g2 ctJxb3 1 1 1!i'xb3 �d6= Belyavsky-P.Nikolic, Barcelona 1989. 5 ctJxd4 Here 5 exd5 \1!Vxd5 would be the main line of the chapter. 5 .. . ctJf6 5 . . . ctJc6 is often played as well. The most critical line is 6 �b5 (6 ctJxc6 bxc6 is solid; after 7 �d3 Black has several moves but I like 7 . . . ctJf6 8 1!i'e2 [8 e5 ctJd7 transposes to the main line note to White's 8th] 8 . . . i.. e 7 9 0-0 0-0 1 0 e5 ctJd7 11 ctJf3 i.b7= with the idea . . . ctJc5 or . . . c5) 6 . . . i.d7 7 ctJxc6 bxc6 (or 7 . . . i.xc6 8 i.. xc6+ bxc6 9 c4 Ji.c5 !? 10 1!i'a4 ctJe7 11 exd5 exd5 12 ctJb3 i.b6 13 0-0 0-0= Pixton-Gulko, Seattle 2003) 8 �d3 .lid6 (no one has tried 83
Play the French my suggestion 8 . . . m8!? with the idea 9 'iWe2 i.d6 10 tllf3 tll e 7 1 1 e5 J.. c 7) 9 'iVe2 (but the similar 9 0-0 'iVb8! 10 h3 tll e 7 was tried in Pert-Levitt, Telford 2003; Black has no problems, espe cially since 1 1 tll f3 can be answered by l 1 . . .e5!) 9 . . . tll e 7 10 e5 (10 tll f3 tll g 6!=) 1O . . . i.c7! 1 1 0-0 tll g6 1 2 tll f3 f6 1 3 .i.xg6+ (13 exf6 'iVxf6 1 4 tll g5 e5) 13 . . . hxg6 14 i.f4?! g5 15 i.g3 f5 ! :;: . Let's return t o the position after 5 . . . tllf6:
�4+) 8 . . . .i.xe6 9 fxe5 tll c 6 10 tll f3 i.c5+ . 7 ..tll c 6 8 i.b5!? 8 tllxc6 bxc6 9 i.d3 J.. a 6! 10 0-0 J..x d3 11 "iWxd3 J.. e 7 12 c4 0-0= Svidler-Dreev, Rostov 1993; and 8 i.f4 'iVb6! is at least equal. .
8 .'iVb6! This is more straightforward than 8 . . . 'iVc7!? 9 0-0 a6 (9 . . . i.e7!?) 10 tll xc6 bxc6 1 1 .i.a4 which has arisen in sev eral games and seems very slightly in White's favour. 9 c3 (a) 9 .i.e3 tllx d4 10 tllx d4 i.c5 !? (or 10 . . . a6 1 1 .txd7+ JLxdn) 1 1 "iWg4 .i.xd4 12 .txd7+ i.xd7 13 J..x d4 'iVb4+!:;:; (b) 9 0-0 tll x d4! (9 ... i.c5 has equal ised in several games) 10 tllxd4 .tc5 11 J.. e 3 transposes to note 'a'. 9 ... J.. c 5 9 . . . tllxd4!? is still reasonable in view of 10 tllx d4 (10 i.xd7+ i.xd7 1 1 tllxd4 .tc5 12 "iWg4 O-O-O!) 10 . . . i.c5 1 1 i.a4 (1 1 'iVa4 a 6 1 2 i.d3 0-0) 1 1 . . .0-0 12 f4 f6! 13 exf6 tllxf6 with a healthy counterattack. Mter 9 . . . i.c5, a possible line is 10 'iYa4 O-O! 11 JLxc6 bxc6 12 "iWxc6 J.. x d4 13 tllx d4 tllx e5 ! :;: , because 14 'iVxa8?? i.a6 15 "iWxf8+ Wxf8 is hopeless for White. ..
6 e5 The only dangerous move. (a) 6 .i.b5+ J.. d 7 7 i.xd7+ tllb xd7 8 exd5 tllx d5 9 0-0 J.. e 7 10 c4 tll 5f6 1 1 b 3 tll c 5= Chiburdanidze-Luther, Graz 1991; (b) 6 i.d3 i.c5 7 tll 4b3 dxe4 (7 . . . J..b 6!?) 8 tll xc5 exd3 9 tllx d3 0-0 1 0 0-0 tll c 6= Mi.Tseitlin-Teske, Polanica Zdroj 1988; (c) 6 exd5 tllx d5 (6 . . . "iWxd5!? 7 tllb 5 tll a 6 8 tll c 4 "iWxd1 + 9 Wxd1 i.c5 - Ad ams; this looks nice for Black, e.g., 10 tll c d6+ We7 11 tllxc8+ llaxc8 with the ideas . . . .txf2, . . . tll g4 and . . . tllb 4) 7 tll 2 f3 .i.e7 8 JLd3 (8 .i.e2 a6! = plan ning . . :ii1c7 Eingorn-Dreev, Lucerne 1993) 8 . . . tll d 7 9 c4 tll 5 f6= Scherzer Remlinger, New York 199 1 . 6 ...tll fd 7 7 tll 2 f3 Not 7 f4? tllxe5 ! 8 tll xe6?! (8 fxe5
84
Tarrasch Variation: Introduction and 3 . c5 . .
6.2 4 exd5 'iVxd5
This line was obscure when I chose it for the first edition. Now 4 . . .'iVxd5 may well be the most popular line of the Tarrasch! Thus one can no longer expect any surprise value from the queen recapture; but on the bright side 4 . . . 'iVxd5 has survived all chal lenges and the fact that so many of the world's top players use this sys tem is an indication of its essential soundness. Black intends, by trading his c-pawn for White's d-pawn, to end up with an extra centre pawn. This is as if, in an Open Sicilian Defense, he had played . . . d5 and replied to exd5 with a piece capture. As in the Sicil ian, violent White attacks are possi ble, based on superior development, but Black's structural advantage can give him a long-term edge. 5 � gf3 The only serious alternative is 5 dxc5 .1i.xc5 6 �gf3 �f6 and: (a) 7 .1i.d3 �c6 (7 . . . 0-0 8 'iVe2 �bd7 9 �e4 b6 10 �xc5 'iVxc5 1 1 .1i.e3 "Wic7 12 Ji.d4 Ji.b7 13 0-0-0 �c5! = Kas parov-Anand, Reggio Emilia 1 992) 8 'iVe2 �b4! 9 Ji.b5+ .ltd7 1 0 .1i.xd7+ �xd7 1 1 �e4 1Ic8 12 0-0 Ji.e7= Chow Remlinger, Chicago 1992; (b) 7 .i.c4 "Wic6! 8 0-0 0-0 9 'iVe2 a6 10 �e5 'iVc7 11 �df3 b 5 12 .1i.d3 .1i.b7=
Kudryashov-Petrosian, Moscow 1967. 5 ... cxd4 6 .i.c4 'iVd6
The trademark move of Black's system. The queen covers e5, pre vents .1i.f4, and keeps the back rank clear for rapid development; one might note that the older move 6 . . . 'iVd8 did none of these things, al though it sometimes transposed if . . . 'iVc7 followed. 7 0-0 A common oversight is 7 �b3? 'iVb4+ (8 �bd2 �c6 and White has lost time) . 7 'iVe2!? has taken a leap popular ity recently, especially since the main lines below have proven satisfactory for Black. White's main idea is to get a rook to dl quickly in order to cap ture on d4 with more effect. The move �e4 can be useful, and the bishop on c4 won't necessarily lose a tempo if and when Black plays the move . . . 'iVc7: (a) 7 . . . �f6 8 �b3 �c6 9 .1i.g5 (see following diagram) This is an important position these days. Black's solutions are relatively unexplored; I particularly like the alternative idea in the recent game in 'a2': 85
Play the French
(al) 9 . . . a6!? 10 0-0-0 b5 1 1 i.d3 iLe7 is rather loose, but if Black gets . . . i.b7 in he will be ideally active: (a l l) 12 lit>b l?! i.b7 1 3 liJfxd4 liJxd4 1 4 liJxd4 i.d5! 1 5 liJf3 0-0 16 h4 b4 1 7 liJe5= Illescas Cordoba-Rustemov, Mondariz 2002; (aI2) 12 liJbxd4 may be best, e.g., 12 ... liJxd4!? 13 liJxd4 �d5 1 4 iLxf6 iLxf6 1 5 c;;t>b l and 1 5 . . . .i.xb7 1 6 liJxb5 or 1 5 . . . "iVc5 1 6 i.e4 11a7 1 7 iLc6+ with the idea 1 7 . . . i.d7 18 liJxe6! . Black should try 12 . . . i.b7; (aI3) 12 .:thel i.b7 (12 .. .'ili'c7 13 liJbxd4 liJxd4 14 liJxd4 iLb7=) 13 liJbxd4 liJxd4 14 liJxd4 iLd5= ; (a2) 9 . . . 'ilVb4+! is a logical and ap parently effective new move from Ad ams-Akopian, Enghien les Bains 2003: 10 iLd2 (10 liJfd2 iLe7 1 1 0-0 0-0 12 a3 'i!Vb6) 10 . . . 'ilVb6 (now the queen won't be exposed down the d fIle after 0-0-0) 1 1 0-0-0 iLd7 12 iLg5 iLc5 !? 13 �b l (13 iLxf6 gxf6 1 4 "iWe4 f5 15 "iWh4 iLe7; , e.g. , 16 'iYh6 0-0-0 1 7 liJbxd4 liJxd4 1 8 liJxd4 l:.hg8!? 19 'fixh7 i.f6 !; 13 liJe5 liJxe5 1 4 �xe5 iLe7=) 13 . . . 0-0-0?! (13 . . . h6 14 iLf4 o-m:) 14 liJe5! liJxe5 1 5 it'xe5 iLd6 1 6 'iVe2 (16 'fixd4 "iWxd4 1 7 liJxd4 h6 1 8 iLxf6 gxf6=) 16 . . . h6= ; (a3) 9 . . . i.e7!? (unpinning the e pawn in the case of . . . e5; Black's life 86
seems slightly difficult after this popular move) 10 0-0-0 e5 1 1 i.b5! i.g4 1 2 'iHxe5 �xe5 1 3 liJxe5 iLxdl 1 4 liJxc6 i.e2! 1 5 i.. xe2 bxc6 16 liJxd4 unclear (Firman); White is an ex change down for a pawn but he has two bishops and will likely win an other pawn, so I prefer his chances. (b) For some reason, few players have paid much attention to the unique order 7 . . . liJc6 8 liJb3 (8 liJe4 'iHc7 9 0-0 liJf6 has proven solid in many games, e.g. , 10 i.. g 5 liJxe4 1 1 'iYxe4 iLd6 1 2 lUe l 0-0 13 liJxd4!? i.xh2+ 14 lit>hl liJxd4 1 5 �xd4 i.d6 16 .l:tadl i.c5 17 "iWh4 iLxf2!? 18 'fixf2 �xc4 19 i.f6! 'iHg4 20 .:td3 S.Pedersen-Ostenstad, Aars 1 999, and Black should play 20 . . gxf6 21 J:tg3 �g6 with . . . e5 next and a small material advantage), and now 8 . . . a6! . In spite of Black's slow development I see nothing wrong with his position. White will have to waste time to re gain the pawn and won't always suc ceed, e.g. , 9 0-0 (9 i.d2?! b5 10 .lid3 liJf6 1 1 0-0 .i.b7; Einvik-Lahlum, Gausdal 1994) 9 . . . liJge7!? (9 . . .liJf6 10 JLg5 �c7 11 :ladl b5 12 iLd3 Char bonneau-Sarkar, internet game 200 1 ; and 12 . . . i.e7 13 liJbxd4 liJxd4 1 4 liJxd4 i.b 7 looks about equal, but there are many options here) 10 .l:.dl b5 11 .lid3 e5 12 c3 f5 1 3 iLc2 iLe6 14 cxd4 e4 15 liJg5 iLc4 with ideas like . . . liJb4-d3, . . . liJd5, and . . . h6/ . . . g5 with . . . iLg7. Given some recent White suc cesses with 7 "iWe2, this is an idea worth noting. 7 liJf6 7 . . . liJc6 8 liJe4!? (8 liJb3 usually transposes) 8 . . :iVd8 is risky, but per haps satisfactory. White can vary with 9 l:.el or even 9 c3!? dxc3 10 liJxc3 with compensation. 8 liJb3 .
...
Tarrasch Variation: Introduction and 3 . . c5 .
Recently, Kasparov revived the old move 8 lIe l , but it hasn't really caught on and I'm not sure why. One possible (but untested) solution might be 8 .. .'�Jc6 (8 . . . i.e7 9 ttJe4 ttJxe4!? 10 lIxe4 ttJc6 11 ttJxd4 0-0 12 i.f4 with advantage was Kasparov-Gelfand) 9 ttJe4 ttJxe4 (Rustemov recently tried 9 . . . 'ilVd8!? and achieved a draw) 10 lIxe4 i.d7 (to get ... 0-0-0 in as fast as possible in some lines) 1 1 i.f4 'ilVc5 (1 1 . . . 'ilVb4 12 i.b3 'ilVb6 13 ttJxd4 0-0-0 14 i.g5! lIe8 1 5 i.e3 i.c5;!;) 12 i.b3 i.e7 13 ttJxd4 i.f6= . 8 ...ttJc6
9 ttJbxd4 (a) 9 �e2 !? risks losing a pawn: (al) 9 . . . i.e7 10 lIdl e5 is solid: 1 1 i.b5 i.g4 12 ttJbd2 ttJd7 1 3 h 3 i.xf3 14 ttJxf3 0-0 1 5 c3 i.f6 16 i.e3 Lo bron-Henley, Indonesia 1983; and now 16 . . . lIfe8!?=i=; (a2) 9 ... a6 10 a4?! (10 lIdl! b 5 1 1 i.d3 and 1 1 . . .i.d7 1 2 c 3 'ilVc7= or 1 1 . . .'ilVc7 12 a4 b4 1 3 i.c4 i.b 7 14 ttJbxd4 ttJxd4 1 5 lIxd4 i.c5= Geller Dolmatov, Moscow 1992) 10 . . . i.e7 1 1 g3 (1 1 i.g5?! e 5 1 2 lIfe l i.g4 1 3 h3 i.xf3 14 'ilVxf3 0-0+ Siklosi-Brinck Claussen, Copenhagen 1988) 1 1 . . .e5 12 ttJg5 0-0 13 f4 Di Lao-Terenzi, corr 1991, and now 13 . . . i.g4! 14 ttJxf7 (14 'ilVf2 e4!) 14 ... i.xe2 1 5 ttJxd6+ i.xc4 16
ttJxc4 e4+ - analysis in CCYB#7; (b) 9 lIel i.d7 (9 . . . a6 10 a4 'ilVc7 1 1 ttJbxd4 i.d7 1 2 'ilVe2 ttJxd4 1 3 ttJxd4 i.c5 14 ttJb3 i.d6= Lane-Levitt, Brit ish Ch 1987) 10 g3 i.e7 1 1 ..if4 �4 12 'ilVd3 lIc8 (12 . . . 0-0 13 i.c7 i.d8! 14 a3 'ilVe7= Winsnes-Lein, Gausdal 1990) 13 a4 0-0 14 a5 ..id8 15 iLd2 'ilVd6 16 ttJbxd4 ttJxd4 1 7 ttJxd4 e5 18 ttJb5 'ilVc5= Ljubojevic-P.Nikolic, Til burg 1 987. 9 ...ttJxd4 10 ttJxd4 At this point, White can choose to go into an endgame with 10 'ilVxd4
This option is obviously important but has grown less popular by the year. Essentially, White hopes that his slightly freer development (Black's c8 bishop is still restricted) will give him an edge, whereas Black counts upon his 4-3 kingside majority in conjunction with the c-file to equal ise and perhaps lead to a long-term Sicilian-like advantage. Indeed, when 4 . . . 'ilVxd5 fIrst hit the scene, Black managed to win several of these end ings; later, players of White got their revenge by pressuring the queenside (e.g., by i.e2-f3, i.f4, and a4-a5 or ttJb3-a5) . Now we know that the end ing should be drawn, but also that either player can try for more if his opponent slips up, with White need87
Play the French ing to be more careful. Little has es sentially changed since the second edition, and Black has three reason able moves: lO . . . a6, lO . . . 'i'xd4, and 10 . . . .i.d7. I believe that all three are playable; but to keep things clear I will limit myself to the latter, which develops quickly: 10 . . . .i.d7
In this position, for what it's worth, Black has a higher performance rat ing than White! I will try to show cor rect play: (a) 1 1 .i.e3 'i'xd4 (or 1 1 . . J::t c8) 12 J.xd4 l:!.c8 1 3 ltJe5! �c5 Gust for fun, Black can achieve equality by means of the bizarre line 13 . . . ltJg4!? 1 4 ltJxd7 I1xc4 15 :ad1 �d6 16 iLxg7 J.xh2+ 17 �h1 :g8 18 ltJf6+ ltJxf6 19 J.xf6 J.c7 20 :d4! .l:r.g4!) 14 l:!.ad1 c;2:;e7 (14 . . . �xd4 1 5 .l:r.xd4 c;2:;e7 1 6 l:!.fd1 I1hd8=) 1 5 ltJxd7 �xd4 16 l:!.xd4 l:!.hd8!? (16 . . . ltJxd7 17 I1fd1 ltJe5 1 8 �b3 :hd8=) 1 7 l:!.fd1 (17 ltJb6 ,Uc6!) 17 .. J:txd7 18 l:!.xd7+ ltJxd7 1 9 ..tb5 ltJe5 20 c3 a6 21 i.e2 ltJc4= Brodsky; (b) 1 1 .te2 'ikxd4 12 ltJxd4 llc8 (12 . . . �c5=) 13 �f3 .tc5 1 4 ltJb3 .i.b6 15 c3 (15 �xb7 l:!.xc2) 15 . . . i.c6= ; (c) 1 1 .i.f4 'ilVxd4 12 ltJxd4 .l:r.c8 has been well tested: (c 1) 13 �b3 iLc5 (White's bishop bites on granite; how awkward this 88
position has proven for him is indi cated by the 200-point performance rating advantage that Black has here after more than 100 games! 13 . . . ltJe4!? was a creative solution in Eismont-Glek, Katowice Open 1993: 14 :fdl ltJc5 1 5 l:td2 �e7 1 6 c3 f6!? 17 .te3 e5 18 ltJe2 ltJxb3 1 9 axb3 a6 20 f4 J.e6:j:) 14 lIad1 0-0 1 5 .l:!.fel (15 h3 l:tfd8= 16 c3 �f8 17 .i.e3 h6 18 ltJc2?! ltJe4 19 �d4 �xd4 20 I1xd4 ltJc5:j: Lindenann-Luther, Boblinger 2000) 15 . . JIfd8 16 c3 iLe8 1 7 ltJc2 a5 1 8 a3 �c6= Ermenkov-Eingorn, St John 1988; (c2) 13 .i.d3 ltJd5!? (13 . . . iLc5 14 ltJb3 �e7 1 5 'ufe1? ltJd5 16 �d2 ltJb4:j: Tagnon-Lane, France 1989) 14 �g3 i.c5 1 5 ltJb3 .i.e7 16 c3 0-0= Wikman Prystenski, corr 1990; (c3) 1 3 .i.e2
and both Black moves are instruc tive: (c3 1) 13 . . . .i.c5 14 ltJb3 i.b6 15 c4 (15 c3 0-0 16 .tf3 .i.c6 17 .i.xc6 l:!.xc6 18 I1adl ltJd5 19 iLg3 :fc8 Y2-Y2 was the game Sedina-Naumkin, Tunis 2000) 1 5 . . . ltJe4!? 16 .:tac1 c;2:;e7 1 7 J.f3 �c6 18 �xe4 �xe4 19 :fe 1 �c6 20 c5 i.c7 21 .txc7 l:txc7 Y2 - Y2 Vogt-Hug, Zurich 1999; (c32) 1 3 . . . ltJd5 14 J.g3 h5!? (14 . . . �c5 is common and equal, but
Tarrasch Variation: Introduction and 3. . . c5 this is more exciting; also of note is 14 . . . g5!?, by which Black prepares a possible . . . lLlf4, as well as . . . i.g7, e.g., 1 5 c3 lLlf4 16 .:radl!? lLlxe2+ 1 7 lLlxe2 f6 18 f4! .i.c5+ 19 iLf2 gxf4= Jacoby Weidemann, Bundesliga 1986) 1 5 c4 (15 h4 iLc5 16 lLlb3 ..t.b6 17 c3 lLle7! 18 iLd3 f6! 19 a4 a6 20 i.d6 e5 21 a5 iLa7 22 .i.e4 ..t.e6=i= Acs-Shaked, Bu dapest 1997) 15 . . . h4!? ( 1 5 . . . lLlb6! 16 b3 h4 17 .i.f4 f6) 16 cxd5 hxg3 1 7 hxg3 e5 18 lLl f3 iLd6 with two bishops and activity for a pawn, Womacka Glek, Cattolica 1993. 10 lLlxd4 a6
The main position of the 4 .. :iWxd5 variation. Black prevents any tricks on b5 and prepares . . . 'fiIc7 and . . . ..t.d6, to catch up on development. He also has the idea of developing the queen's bishop aggressively by . . . b5 and . . . iLb7. The analogies with the Sicil ian continue, as . . . a6 is part of the Najdorf, Scheveningen, Taimanov, and other Sicilian variations. From the diagram, White has tried a vari ety of plans: 6.2 1 1 1 i.b3 6.22 1 1 c3 6.23 1 1 b3 6.24 1 1 :e l Others are less critical:
(a) 1 1 iLe3 "fic7 12 ..t.b3 (12 "fie2 .i.d6 13 h3 b5 14 iLb3 0-0= was Ma tanovic-Eliskases, Stockholm 1952) 12 . . . i.d6 13 h3 (13 lLlf3? lLlg4=i= ; 13 g3 h5!? or 13 . . . b6, which is often better than . . . b5 in these positions because the b-pawn isn't subject to attack by a4) 13 . . . 0-0 14 c3 (14 "fif3?? e5! 1 5 lLlf5 [ 1 5 lLle2 e4] 1 5 . . . e4 16 lLlh6+ �h8-+) 1 4 . . . e5 (this tends to be safe when lLlc2-e3 is not handy) 15 lLlf3 e4 (15 . . . h6!? intending . . . i.f5) 16 lLld4 'WIe7 17 lLle2 l:td8 18 'iVc2 i.f5= Al masi-Luther, Kecskemet 1993; (b) 11 a4 'iVc7 (1 1 . . .i.d7 has also had excellent top-level results) and: (b l) 12 i.b3 i.d6 13 lLlf3 b6 14 lIe 1 0-0 (14 . . . i.b7 is probably more accu rate, with a typical line being 15 h3 l:!d8 16 'fiIe2 h6 17 c3 0-0) 15 c3 i.b7 16 h3 l1ad8= LDahl-Lahlum, Asker 1 994. (b2) 12 'iVe2 ..t.d6 13 h3 (13 lLlf5 .i.xh2+ 14 hl f8! 1 5 lLlg3 h5=i= Van der Wiel- Glek, Tilburg 1 994) 13 . . . 0-0 1 4 i.g5 (14 lIdl b6 1 5 i.g5 .i.h2+ 16 �hl i.e5 - now that f4 is not possible - 17 l:1.a3 i.b7 18 l:1.e3 i.f4= Ljubo jevic-Hiibner, Wijk aan Zee 1 988) 14 . . . lLle4! (14 . . . b6 is also fine, since 1 5 iLxf6 gxf6 gives Black two bishops and the open g-fIle, a recurring theme) 1 5 iLe3 ( 1 5 'fiIxe4 'iYxc4=i= with two bishops and . . . e5 in store) 15 . . . b6 1 6 c3 iLbn Pleyer-Keitlinghaus, Bad W6rishofen 1997; (c) 11 iLd3 iLd7 12 lLlf3 "fic7 13 "fie2 .i.d6 14 .l:.el lLld5 15 a3 (15 c4 lLlf4 16 i.xf4 i.xf4 1 7 g3 iLh6 1 8 lLle5 g6= with . . . 0-0 and . . . ..t.g7 to come) 15 . . .lLlf4 16 i.xf4 .i.xf4 17 g3 .i.h6!? 1 8 Iiadl l:1.d8 1 9 lLle5 g 6 20 lLlxd7 was agreed drawn in Berelovich-Boro vikov, Ukrainian Ch 200 1 in view of the opposite-coloured bishops . 6.2 1 1 1 i.b3 89
Play the French 13 . . . i.e5 14 c3 .td7 1 5 .tg5 !? (15 ii.e3 0-0=) 15 . . . i.xd4! 16 cxd4 i.c6 1 7 �e3 lZJd5= or slightly better, Ernst-Orn stein, Swedish Ch 1980. 14 i.g5
This takes the bishop off the ex posed c4 square and waits one move before committing the white pieces to their posts. 1 l .. :i'c7 12 'i'f3 1 2 ':e1 and 12 c3 transpose to 1 1 :e1 and 1 1 c 3 below, respectively, whereas 12 a4 ii.d6 transposes to the note on 1 1 a4 above. 12 ii.g5 ii.d6 (or 12 . . .'�Je4! 13 .Jle3 lZJc5= , getting rid of the b3 bishop) 13 .Jlxf6!? gxf6 14 'ilfu5 (14 J:te 1? i.xh2+ 1 5 'it>h1 h5!+ Jansa Hubner, Bundesliga 1989) 14 . . . 'iVc5 ! (This idea pops up throughout the 10 . . . a6 variation. Black will be better in an ending due to his two bishops unless White can use his develop mental edge) 15 'i¥h4! �g5!? 16 �e4 �f4 1 7 "iWxf4 .Jlxf4 18 llad1 ii.d7 19 g3 i.c7= , but with slightly better practi cal chances for Black. 12 ... i.d6 13 h3 13 �h1 0-0 14 i.g5 lZJd7!? (14 . . . i.e5 15 l:.ad1 .td7= looks easier) 15 c3 lZJe5 1 6 'ilfu5 lZJg6 17 .ic2 h6!? 1 8 lZJf3 (18 .te3 lZJf4 19 'iVf3 Potkin-Rodriguez Guerrero, Linares 2002; and perhaps 19 . . . .td7 was best, to get developed) 18 . . . b5 19 l:.ad1 i.f4 20 i.xf4 lZJxf4 2 1 "iWe5 "iWxe5 2 2 lZJxe5 i.b7 23 f3 IUd8= Ivanchuk-Anand, Reggio Emilia 1 992. 1 3 ... 0-0 An important theme appears after 90
14 b5!? Ornstein's bold gambit. Black can also equalise by 14 . . . lZJd7 15 c3 b5! 16 I1ad1 (16 'iYxa8? ii.b7 1 7 'iVxf8+ lZJxf8 1 8 .Jld1 lZJg6+ with the idea . . . h6, . . . lZJh4 Ernst-Wiedenkeller, Swedish Ch 1 989; 16 .Jlc2 i..b 7 17 'ilfu5 g6 1 8 'i¥h4 i.h2+ 1 9 �h1 .te5= Djurhuus Lahlum, Bergen 2000; 16 l:.fe 1 .ib7 1 7 'iVh5 lZJc5 1 8 i.c2 g6 1 9 'iVh4 l:.fe8 20 I1e3 .Jld5= Asrian-Savchenko, Ohrid 2001) 16 . . . lZJc5 (16 . . . ii.b7 1 7 'iVg4 lZJc5 1 8 i.f6 g 6 19 l:r.fe1 lZJxb3 2 0 axb3 1Ife8 2 1 I1d3 i.h2+ 22 'it>h1 i.f4= M.Adams-Levitt, London 1 989) 1 7 .tc2 .Jlb7 1 8 'i¥h5 .Jle4 1 9 i.xe4 lZJxe4= Hellers-Hubner, Wijk aan Zee 1 986. 1 5 i.xf6 Not 15 'iYxa8? i.b7 16 "iWxf8+ 'it>xf8 because Black's attack continues. 15 . gxf6 1 6 "ii'xf6 .Jle5 1 7 ii'h4 .Jlb7 18 c3 'it>h8 1 9 f4 Heretofore considered a kind of refutation, but Black has been given a new life recently. Instead, 1 9 .ic2?! f5 20 .l:lae1 was Nolan-Spiegel, corr 1 986; and Black should have played ..•
..
Tarrasch Variation: Introduction and 3. . . c5 20 . . . i.h2+! 2 1 �hl 'iWg7 22 f3 i.g3:j: . 19 ... .l:Ig8 20 Itf2 .td6! A creative idea of the strong corre spondence player John Knudsen. 20 . . . i.xd4?! 21 cxd4 l:tg6 22 f5 ! exf5 23 d5 was better for White in Yudasin Ornstein, Trnava 1983.
Here's some analysis, with themes that apply to the whole system: (a) 2 1 a4 lig7 22 axb5? liag8 23 'i'f6 i.e7 24 'iWe5 'iWxe5 25 fxe5 i.xg2 and there are no good defences against . . . i.f3+ and ideas like . . . i.g5e3; (b) 21 "iVf6+ .l:i.g7 22 l:!.el (22 i.xe6 .ten 23 "iVh6?? l:[g6; 22 g4 i.e7 23 'fie5 'i¥xe5 24 fxe5 i.h4 25 l:[f£1 h5 26 lLlf3 i.g3! 27 g5 J:.ag8:j:) 22 .. .lhg8 23 l:!.ee2 (23 g4 i.e7 24 "iVe5 i.d6 25 'i!Ve3!? i.e7! 26 l:!.dl? i.h4 27 lie2 h5+) 23 . . . i.e7 24 "iVh6 b4! and White is very tied up, e.g., 25 ':'c2? 'i'c5!-+ threatening . . . l:!.g6 and answering 26 f5 with 26 . . . .tg5; (c) 21 i.c2 lig7 22 liel l:tag8 23 i.e4 .i.c8! (preparing .. .f5 and . . . i.b7; 23 . . . .txe4 24 lixe4 b4!?= contemplates . . . "iVa5 or . . . 'iWc4) 24 l:tee2!? (24 'i!Vf6! J.e7 25 'it'h6 f5 26 i.c2 l:tg6 27 'it'h5 i.c5=) 24 . . . f5 25 i.c2 'iWf7! (intending . . . i.e7) 26 11£1 (26 i.b3?? .te7 27 'iYh6 lig6) 26 . . . i.b7 27 lLlf3 'iWc7! 28 i.b3 J.e4! and f4 is hard to defend, e.g., 29
lLlel (29 lLlg5 i.d3!) 29 . . . .te7 30 'i!Vh5 'iWc6. 6.22 1 1 c3
By protecting the knight on d4, White frees his queen to go to f3 or e2, and he awaits developments be fore committing his other pieces. 1 1 ... "iVc7 Because 11 c3 is rather slow, 1 1 . . .i.d7 comes under consideration, e.g., 12 'fif3 "iVc7 13 .tb3 .td6 14 h3 i.h2+ (in the same vein, Remlinger suggests 14 . . . 0-0 1 5 i.g5 i.h2+ 16 'ifo>hl i. e5! 1 7 liadl h 6 1 8 i.h4 l:!.ae8 intending . . . i.c8, . . . b5, and . . . i.b7) 1 5 'ifo>hl i.e5 16 i.e3 0 - 0 1 7 a 4 Larsen Seirawan, Mar del Plata 1982, and here I proposed 17 . . . b6! 18 i:ladl l:tad8= to be followed by . . . .tc8-b7.
12 'i!Ve2 (a) 12 i.d3 i.d6 13 h3 b6!? (a less aggressive method is 13 . . . i.d7 14 'fif3 0-0 15 i.g5 .i.h2+ 16 'ifo>hl i.e5= Tivi akov-Kramnik, USSR 1991) 14 'i¥e2 i.b7 15 a4!? 0-0 16 .i.g5 lLld5 17 I!fe l (17 'it'h5 g6 18 'it'h6 .i.h2+ 1 9 Whl i.f4!=) 17 . . . .tf4 (or 17 . . . lLlf4!?) 18 'iVg4 .txg5 19 'fixg5 h6 20 'it'h4 lLlf6 2 1 l:te3 lifd8! 22 l:tae l (22 l:!.g3 Wf8=) 22 . . . lid5! (threatening . . . l:th5) and Black stood well in Maahs-Lahlum,
91
Play the French Hamburg 2002; (h) 12 i.b3 i.d6 13 h3 0-0 (or 1 3 . . . b6) and: (h I) A standard tradeoff occurred after 14 i.g5 .if4! 15 i.xf6 gxf6 in Nunn-Speelman, London 1984, with Black's weaknesses compensated by his bishops and g-flle: 16 �g4+ �h8 17 �4 i.g5 18 �5 :g8= 19 liJf3?! i.f4 20 ii.c2 lig7 21 :tadl?! ii.d7=i= in tending . . . i.c6; (h2) 14 .l:.el b5 15 i.g5 i.b7 1 6 ii.c2 (16 i.xf6 gxf6 17 'fig4+ �h8 1 8 'fih4 i.h2+ 19 �hl 'fif4!=) 16 . . . liJd5 1 7 'iVh5 g 6 18 � 4 i.h2+ 1 9 �hl .if4= M.Adams-Gulko, Groningen 1993. 12 ... i.d6 1 3 h3 The alternatives are instructive: (a) 13 �hl 0-0 14 .ig5 allows the trick 1 4 . . . liJe4!= ; (b) 1 3 liJf3 (watching e5 and often planning to go there) is often White's best idea in these lines. Here are two responses: (b l) My second edition suggestion was 1 3 . . . .id7!? 14 .id3 (14 I1el i.c6 15 liJe5 .ie4!=i=; 14 i.g5? i.c6 1 5 i.xf6 gxf6 threatens . . . ii.xf3, which is hard to meet) 14 . . . .ic6 15 h3 0-0-0 or here 15 . . . 0-0 16 i.g5 liJd5= ; (b2) 1 3 . . . 0-0 14 ii.g5 liJd5 (on 14 . . . b 5 1 5 ii. d 3 liJ d 5 , 16 a4;!; has t o be dealt with) 1 5 liadl (probably more accu rate would be 1 5 Itfdl, leaving the idea of a4 open, or 1 5 i.d3 with the idea 'fie4) 15 . . . b5 ! 16 i.d3 h6 17 i.cl Koehn-Lahlum, Goteborg 2000 (by transposition), when Black can equal ise by simply 17 . . . i.b7. 13 ... 0-0 14 i.g5 The most natural move, Others don't do much, e.g., (a) 14 ii.d3 ii.h2+ (14 ... b6!? trans poses to the note on 12 ii.d3 above) 1 5 �hl ii.f4 16 'i¥f3 ii.xcl 1 7 liaxcl l:r.e8 1 8 life l 'iVa5 19 a3 e5= Lansin92
Gurjatinski, USSR 1 990; (h) 1 4 lidl e5! 15 liJf3 (15 liJc2 !? b5 16 i.b3 i.b7 1 7 liJe3) 1 5 . . . b5 16 .ib3 i.b7 17 i.g5 liJe4 18 i.h4 �ae8= Jo vovic-Bohak, corr 1 990. 14 ... liJe4 1 5 i.h4!? 15 �xe4 �xc4 yields nothing after 16 ii.f4 i.xf4 1 7 'iVxf4 i.d7= Palac Keitlinghaus, Prague 1 990. 15 ... liJd2!? 1 6 'i¥xd2 �xc4= Ko sashvili-Holzke, Bie1 1 989. In general, 11 c3 is a solid move that doesn't do more than equalise. 6.23 1 1 b3
This is one of White's better plans: he exerts pressure on the kingside via the long diagonal, protects his bishop on c4, and might in some cases ex pand on the queenside by .td3 and c4. 11 ...�c7 12 i.b2 Sometimes White prefers to delay the fianchetto by 12 'iVe2!? This is a multipurpose move, keeping a queen out of f4 and contemplating attack by liJf5 or c4, but keeping the idea liJf3e5 in reserve:
(a) 12 . . . i.c5 13 .i.b2 0-0 14 liJf3! (14 lladl b 5 1 5 .id3 .i.b7 1 6 liJf3 'i¥f4! 1 7 liJ e 5 'i¥g5=i= Kotronias-Kindermann, Debrecen 1989) 14 . . . b5 15 i.d3 i.b7
Tarrasch Variation: Introduction and 3. . . c5 (15 .. :iVf4!? 16 .i.e5 'iVg4 1 7 a4!) 16 Cbe5 (16 Cbg5?! 'iVc6!) 1 6 . . JIad8 1 7 a4 (17 Mad1 .i.a8! intends . . . 'lib7) 17 . . . .i.d4 18 .i.xd4 Mxd4 19 axb5 axb5 20 Ma7 'lib8 2 1 Mfa 1 .i.e4 22 Cbf3 .i.xd3 YZ - YZ Ochsner-Brinck Claussen, Denmark 2000; (b) 12 . . . .i.d6 1 3 Cbf5!? (13 .i.b2! .i.xh2+ 14 'iith 1 0-0 1 5 g3!? e5! 1 6 'iit xh2 exd4 1 7 .i.xd4=) 1 3 . . . .i.xh2+ 14 'iith 1 0-0 1 5 Cbxg7! (15 .i.b2? exf5 16 .i.xf6 gxf6 1 7 g3 Tavcar-Bohak, Bled 200 1 ; and 17 . . . .i.xg3 looks good) 1 5 . . . .i.e5! (15 . . . 'iVe5?! Rubin-Glek, corr 1 989; 16 .i.h6! 'iVxe2 17 .i.xe2 .i.e5= Glek, but 18 Madl±) 16 .i.h6 .i.xa1 1 7 Mxa1 'iVc5 (17 . . . Md8 1 8 Cbh5 Cbxh5 19 'iVxh5 f6 20 'iVh4 'iVf7 2 1 .i.g5! Mf8 22 .i.h6 Md8 23 .i.g5 Mf8 YZ - YZ G.Perez Pietra, email 2000) 18 Md1 b5 19 .i.e3 'iVe5 20 f4 Tsheshkovsky-Stirenkov, Belorechensk 1 989; and instead of 20 . . . Cbe4 2 1 fxe5 Cbg3+ 22 'iith 2 !;!; , Black might risk 20 . . . 'iVc7. 12 ... .i. d6 12 . . . b5 13 .i.d3 .i.d6 transposes.
1 3 Cbf3 (a) 1 3 h3 b5 14 .i.d3 0-0 1 5 Cbf3?! .i.b7 16 'iVe2? Tal-Hubner, Brussels 1987; 16 . . . Cbh5! 17 Mfe 1 Cbf4 18 'iVe3 Mac8+ with the idea 19 Mac1? .i.c5; (b) 1 3 Mel 0-0 14 Cbf3 b5 1 5 .i.d3 .i.b7 1 6 a4 Mfd8 1 7 'iVe2 .i.b4! 1 8 M£1
bxa4 1 9 Mxa4 a5= Psakhis-Chernin, USSR Ch 1 987. 13 ...b5 14 .i.d3 0-0 Or 14 . . . .i.b7 15 Mel 0-0 16 Cbe5 Mad8 17 'iVe2 Cbd5! 18 'iVg4 f5 19 'iVh4 Cbb4:j: Tiviakov-Psakhis, Rostov 1993. 1 5 Me l Skrobek -Ornstein, Pamporovo 1981. This improves upon 1 5 .i.xf6?! gxf6 16 'iVe2 f5 - Ornstein, as well as 1 5 'iVe2 Cbd5! :j: Kaiszauri-Ornstein, Eksjo 1980, intending 16 .i.xh7+? 'iitxh7 17 Cbg5+ 'iit g8 18 'iVe4 (18 'iVh5 'iVxc2) 18 . . .f5 19 'iVh4 Cbf6-+ . After 1 5 Mel , Ornstein suggests the ingenious manoeuvre 15 . . . .i.b7 16 .i.xf6?! (16 a4) 1 6 ... gxf6 1 7 .i.e4 Mad8 18 'iVe2 .i.c8!= The point is that White cannot safely prevent . . .f5 and then . . . .i.b7, e.g. 1 9 h3 (19 c4 may be best) 19 . . .f5 20 .i.d3 .i.b7 21 a4 (2 1 Cbe5 .i.b4!) 2 1 . . . .i.b4 22 Med1 .i.c3:j: . 6.24 1 1 Mel
This direct move is by far the most popular among top players. 1 1 Mel is the most active move and introduces ideas such as Cbf5 or sacrifices on e6 . Theory has expanded dramatically since the second edition and this is a line that one should know intimately to get the most out of this 4 . . . 'iVxd5 repertoire with 10 . . . a6. I will there93
Play the French fore devote a great deal of analysis to it. 1 l :iVc7
12 .td6 ...
..
12 i..b 3 (a) 12 i.. d 3?! used to be critical but is now considered solved: 12 . . . i.. d6 1 3 lLlf5 (the only ambitious try) 1 3 . . . i.. xh2+ 14 'it>hl 'iitf8 (14 . . . h5? 1 5 g3± Shamkovich-Seirawan, U S Ch 1980) 15 g3 (15 lLlxg7? h5!) 15 . . . exf5 1 6 'it>xh2 (16 i..f4 �c6+ 1 7 'iitxh2 i.. e 6 18 c3 h5� Brindza-Bohak, COIT 1989) 16 . . . h5 17 i.. f4 'iVb6 18 'iit g2 i.. e 6 19 f3 (19 c4 h4! 20 f3?? hxg3 2 1 i.. xg3 f4-+) 1 9 ... h4 20 llhl lLld5 2 1 iVd2 h3++ Geller-Thesing, Dortmund 1 992; (h) 12 'i¥e2 has two adequate re plies: (h I) 12 . . . i.. c 5 1 3 c3, and although 1 3 . . . 0-0 isn't bad, the direct answer is 1 3 . . . b5! 14 i..b 3 0-0 1 5 i.. g 5 i.. b 7!= Speelman; (b2) 12 . . . i.. d 6 13 i.. g 5!? (13 lLlf3 b5 1 4 i.. d 3 i.b7 1 5 lLle5 0-0 1 6 i.. f4 lLld5 1 7 i.. g3 lLlb4 18 liadl :ad8� Peters Lakdawala, Los Angeles 200 1) 13 . . . 0-0 14 g3 (14 i.. xf6 gxf6 15 i.. d 3 i.. xh2+ 16 �hl i.. f4= Emms-Knott, Millfield 2000) 14 . . . lLle4! 15 'i'xe4 ii'xc4 16 i.. f4 i..xf4 17 gxf4! (17 'ili'xf4 f6� Van der Wiel-Nogueiras, Rotter dam 1 989) 17 . . . i.. d 7= Tal-Korchnoi, Brussels 1 988.
94
The ambitious move; everything else is slightly worse for Black. 13 lLlf5 This is the critical and most popu1ar move, initiating amazing tactics. Others tend to lose the initiative: (a) 13 h3!? 0-0 (13 . . . i.. f4 is also equal) 14 i.. g5 b5 15 i.. xf6 (15 c3 i..b 7 16 i.. c 2 lLld5!? 1 7 'i¥g4 ];tfe8 18 'tWh4 g6 19 i.. e 4 f5=) 15 . . . gxf6 16 'iWh5 i..b 7, and 1 7 'iWh6 i.. e 5! or 1 7 c3 'iith 8 18 'tWh6 11g8 19 iVxf6+ 11g7 20 f3 11ag8!= (intending . . . e5 or ... i.. e 7 with attack), V.Akopian-Levitt, Groningen 1 990; (h) 13 g3 !? O-O!? (or 13 . . . e5!, e.g., 14 i.. g 5 i.. g4 15 f3 i.. h 3) 14 i.. g 5 b5!? (14 ... e5 is also thematic; then Fur hoff-Backe, Sweden 2001 was drawn after 1 5 i..xf6 gxf6 16 'iWf3! exd4 1 7 'ili'xf6. White will get a perpetual check, which Black can force by 17 .. :ii'c6) 15 i.. xf6 gxf6 with astonish ingly effective counterplay: 16 'i'd2 !? (two other ideas are 16 'i'g4+ �h8 17 'i\Vf3 f5 1 8 'i¥xa8 i.b7 1 9 'i'xf8+ .txf8 20 lladl i.. a 8! 2 1 f3 !?; and 16 'i'f3 i.. e 5! 1 7 'i¥xa8 i..b 7 1 8 'i'xf8+ �xf8 1 9 c 3 , when 1 9 . . . a5 2 0 11acl 'i'b6 looks fully satisfactory for Black) 16 . . . ..te5 1 7 c3 A.Ivanov-Brunner, Gausdal 199 1 ; here 1 7 . . .f5!? 18 'iWh6 i.. g 7 19 'iWg5 �h8 is a reasonable continua-
Tarrasch Variation: Introduction and 3 . c5 ..
tion (hut not 1 9 . . . .tb7? 20 lIxe6!). 13 ... �xh2+ 14 �hl 0-0 14 . . . �f8?! was debated for several years until it was found that 1 5 'i'd4!? was strong, e . g . , 1 5 . . . exf5 ! 16 'i'xf6! h6! 1 7 'ii'd4 .td6 1 8 .Jid2! threatening .tc3 and .tb4. 1 5 tiJxg7 I:i.d8! move. Black cannot Gulko's attack counter after White's 1 5 . . .'�xg7? 16 "iVd4! (and even 1 6 'i'd2 is difficult) .
16 'i'f3 (a) Se.lvanov mentions 1 6 tiJh5 l:.xdl 17 tiJxf6+ �h8 18 lIxdl �e5 19 .tg5 'i'e7, but this must favour Black; (h) 16 'i'e2?! (avoiding . . . .Jib7 with tempo) hasn't been decisively refuted (although see 'b2'), but isn't worth playing; 16 . . . �xg7 1 7 g3 i.xg3 and: (h I) 18 lIgl? (after this White seems too exposed) 18 . . . b5 1 9 lixg3+ �h8 20 .Jig5 .tb7+ 21 �h2? (2 1 �gl tiJe4! 22 .txd8 Ihd8+) 2 1 . . .tiJe4 22 'ilke3 l:.g8 and Black was winning in Kobalija-Kasimdzhanov, Wijk aan Zee 1 998; (b2) 1 8 fxg3 allows 3 good but only one possibly decisive move. The try for a kill is 18 . . . .td7! (18 . . . 'iYc6+ gives some advantage, e.g., 1 9 �h2 h5!? 20 'iVe5 'iWb5 ! 2 1 .th6+ �g6:j: Mannion Bryson, Scottish Ch 1993; 18 . . . 'ilkc5 19
.tf4 'i'h5+!? 20 'iWxh5 tiJxh5 2 1 .tc7 l:!.d 7 22 .te5+ f6 23 g4 fxe5 24 gxh5 Efler-Grulich, corr 1999, and 24 . . . .l:i.d2:j:) 19 .Jif4 (19 'iWe3 �c6+ 20 �gl tiJe4! 2 1 'i'h6+ �g8 22 .Jif4 e5 or 22 .. :�Ve7) 1 9 . . . 'i'a5! (I had HiArcs' help here; 19 . . . 'iYc5 20 �e5 �c6+ 2 1 'it>h2 'it>g6 2 2 .Jixf6 has been analysed to a draw) 20 .Jie5 .i.c6+ 21 'it>gl lId2! 22 'iWe3! (22 'i'g4+?? �f8 23 �xf6 "iVc5+; 22 .i.xf6+ �xf6 23 :f1 + �g7 24 "iVg4+ �h8 25 1If2 l:!.g8-+) 22 . . . �g6 23 l1e2 1Iad8!-+ . 1 6 ... �xg7
17 �h6+ 17 g3? is complex but is now estab lished as extremely good for Black. A few lines to illustrate this: 1 7 . . . b5! 18 �xh2 (18 .tf4? .Jib 7 19 �xc7 .txf3+ 20 �xh2 tiJg4+ 21 'it>gl l:!.d2 22 l:!.fl ':c8 23 .ta5? .l:tc5 0 - 1 Ye Jiangchuan Dolmatov, Moscow 1 992) 18 . . . .tb7 1 9 'iVe2 (19 .th6+ 'it>g6 and now a trick is 20 'iVf4 'i'c5! 21 'u'e5? 'i'xe5!, and 20 .tf4 'i'c8! 21 'i'e2 e5!+ as given by Lahlum isn't much better; 19 'i'f4 'i'c6! 20 lIgl lIdl ! 2 1 �e3 l:!.xal 22 'i'g5+ 'it>f8 23 'iWc5+ �e8 24 'i'xc6+ �xc6 25 .uxal tiJg4+ 26 'it>h3 tiJxe3 27 fxe3 l:!.d8:j: Kotronias-Chernin, Mu nich 1 993) 19 . . . lld4! 20 f4 (20 .tf4? lIxf4; 20 f3?! lIh4+ 2 1 �g2 tiJg4!; 20 .tg5 tiJg4+ 21 'it>h3 'i'c6+) 20 . . . l:.ad8+ ,
95
Play the French e.g., 21 Wh3 l:te4 22 .te3 e5!. 1 7 'iitt g 6 1 8 c31 (a) I S :ad1? .l::1.x d1 1 9 .li:!.xd1 e5! wins for Black, e.g., 20 Wxh2 (20 g3 liJg4 21 i.c1 i.f5 22 c3 e4-+ Rado vanovic-Vakhidov, Chania 2000) 20 . . . liJg4+ 2 1 Wg1 'it>xh6 0- 1 Zaw Win Lay-Khalifman, Bali 2000, in view of 22 lLxf7 (22 'ilVg3 i.f5 23 .l:.d4 'iitt g5 !-+) 22 ... e4 23 'i!i'h3+ 'it>g7-+ ; (b) I S c 4 tries t o prevent . . . liJd5, but then l S ... e5! is similar to the main text and even better in varia tions where White would like to play ii.xe6 or where Black uses the d4 square. Black can also play lS . . . liJh5, when 1 9 lLe3? f5 20 g4 (20 g3 b6 21 Wxh2 i.b7+) 20 ... liJf6 2 1 gxf5+ exf5 22 'ilVg2+ liJg4 23 f3 i.d7-+ was Gufeld-Ravi, Calcutta 1994. 1 8 e51? ...
...
The most direct move and most promising, although lS . . . liJh5 may be okay as well. The game Adams-Dreev below (after l S . . . liJd5 1 9 liad1) was given enormous publicity after it ap peared in 1 996, and has been given as the main line of this variation for many years (implying that the line was bad for Black) . I found lS . . . e5 and decided that it was okay just a few weeks after becoming aware of Adams-Dreev, and yet it only recently 96
received serious attention. This illus trates the way that a single move in one top-level game (a move that would attract only mild notice if played by a lower grandmaster) can discourage even strong players from discovering obvious improvements. There are alternatives to lS . . . e5: (a) l S ... liJd5?! is considered bad due to the aforementioned Adams-Dreev, Wijk aan Zee 1 996: 19 .l:.ad 1 ! f5 20 .tc1 i.d6 2 1 i.xd5 exd5 22 11xd5 i.d7 23 'tin3 i.f8 24 .l::1. e 3 Wg7 25 l:I.g3+ 'iitth S 26 'iWh4 i.e6 27 i.f4 ii.e7 2S ii.xc7 1-0. Improvements for Black were long sought and not found, though they may be available; (h) lS . . . liJh5 is a valid alternative and has now been played in at least 60 games. Since Black seems to end up with equality after either lS . . . e5 or lS . . . liJh5, this is a reasonable choice: 19 lLc l (19 lLe3 f5 ! has won several games for Black, but the clever 19 l:1.e4! seems to draw: 1 9 . . . 'iittxh6 20 .l:.h4 'i'e5! 2 1 l:1.xh2 i.d7 22 'i'xf7 i.c6!? [22 . . . .teS 23 'ilVf8+ 'iWg7 24 'i'f4+ 'ii'g 5 draws] 23 f4 'i'f5 24 'iVxf5 exf5 25 .tf7 l:1.d2 26 lLxh5 Lu ther- Schlecht, Boblinger 2000, and now 26 . . . .l:.adS was fully equal) and: (h I) 19 .. .f5 20 g4! (this probably doesn't yield much, but it is more in teresting than 20 i.xe6 liJg7 21 lLxcs .l::1. axcS 22 g3 ..txg3 23 fxg3 'i'c6= Womacka -Harikrishna, Pardubice 2002) 20 . . .liJf4 (20 . . .liJf6? 2 1 lLxe6 <j;g7 22 'iVxf5 with attack; but 20 . . . i.d7!? 2 1 gxh5+ cJi;g7 22 'ii' g2+ 'it>hS 23 i.g5 i.e5 24 i.xdS 'iYxdS was dynamically equai in the game Blehm-S.Ivanov, Krynica 1 997, one line being 25 iLxe6!? 'i!Vh4+ 26 'iitt g 1 lLxe6 27 .l:.xe5 .i.d7! with complica tions) 21 gxf5+ exf5 22 .l::1. g 1+! 'iitt f6 23 lLxf4 'iWxf4!? 24 'i!Vh5 iLe6 25 .txe6
Tarrasch Variation: Introduction and 3. . . c5 c;t>xe6 26 l:.tae1 + g1 I!d2 ! 25 ':'c2 l:tad8 26 l:!.xd2 ':xd2 27 'i¥xd2 'iVxg3+ 28 \t>f1 'iVh3+ with repetition. 19 i.c2+! 19 h1 (20 f1 .lif5) 20 . . . c;t>xh6 21 l:!.e4 'iVc6- + and White has no attack: 22 l:!.ae1 YWg6 23 i.xf7 Forberg-Lahlum, Nor way 2002; and the cutest win was 23 . . :i!i'xe4! 24 l:!.xe4 l:!.d1 + 25 'ii'xd1 lLlxf2+ etc. 1 9 ... e4
I have resisted the temptation to split up this 11 l:le1 material at sev eral points, but now it seems neces sary: 6.241 20 i.xe4+ 6.242 20 g3 6.243 20 l:!.xe4 6.241 20 i.xe4 lLlxe4 2 1 l:!.xe4 'ii'c 6!
Threatening . . . \t>xh6 and . . . ..ltf5. Black is playing for the win. 2 1 . . . ..ltd6 seems to lead to a pawn down but drawn opposite-coloured bishops end ing after 22 llh4 'iVc5 23 l:ld1! .lie6 24 i.e3 'iVf5 25 l:!.h6+ <J;; g 7 26 'i!Vxf5 i.xf5 2 7 l:!.hxd6 l:!.xd6 28 l:lxd6 i.e6 29 a4 .l:.e8. Amazingly, after I had done all of this analysis, a game A. Olsson Leer Salvesen, Stockholm 2002 was sent to me with exactly the same moves, indeed leading to a draw!
22 �e3!? This may barely suffice, but White has a more promising choice: (a) 22 xh6 23 "iixf7 'iVxe4 24 'iVf6+ 'it>h5 25 'iVxd8 i.h3 ! 26 'iVa5+ b5 27 f3 "iVe7! + ; (b) 22 l:!.h4? 'iVxf3 23 gxf3 .lic7!? 2 4 ,U g 1 + \t>f5 25 �h5+ d7+ with the idea 27 i.g5 Ite8 28 l:!.d1+ c;t>c6; (c) 22 'iVe3! leads to an interesting attack for a piece: (c 1) 22 .. :fib5 23 lIh4 l:!.e8 24 'iVd2 (24 a4!?, and 24 .. :iVe5?! 25 'iVf3 ! 'iVf5 26 <J;;xh3 'i!Vxf3 27 gxf3 is difficult for Black, even if drawable, so much bet ter is 24 . . . l:!.xe3! 25 axb5 lie2 26 f6 28 .lig7+
97
Play the French 29 i.xe5 �xd3 30 i.xh2 l:[xf2 3 1 l:1.el + �f6 which is well within drawing bounds) 26 . . . �g3 ! 27 fxg3 'i'xg3 and Black has a very large advantage af ter 28 l:1.h2 �g4 29 'i'gl l1d8! 30 l:1.f1 l:1.xb2; (c2) 22 .. :iVd5 looks like a forced draw (although you never know) : 23 l:1.h4 .id6 24 f4! (24 .l:1.el f6) 24 . . . .i.f5 25 'i'g3+ �f6 26 .l:tel (26 c4 'ii' d 3; 26 i.g5+ �e6 and not 27 l:1.e1+? Wd7, but 27 i.xd8! l:1.xd8 28 'i'g5, when Black's bishops may still give him a slight advantage after 28 . . .f6 29 l:th6 l:i.f8) 26 . . . l:tg8 27 .1i.g7+! (27 �g5+ Wg7 28 �e7+ .1i.g6) 27 . . . l:1.xg7 28 lth6+ l:tg6 29 'Ml4+ Wg7 30 l:1.xh7+ Wg8 3 1 l:1.h8+ �g7 32 .l:1.h7+ with a draw by perpet ual check. 22 . . f5 Not 22 . . . .ic7?? 23 l:1.g4+! �xg4 24 'i'xg4+ �f6 25 i.g5+± . 23 l:.h4 'i'xf3 24 gxf3 �e5 Black has to be careful here, even being in possession of an extra piece: 24 . . .i.d6? 25 l1g1+ Wf7 26 l:1.xh7+ �e6 27 l:i.e l ± . Mter 24 . . . i.e5, White can stop Black from consolidating his material by 25 l:[gl +! (25 l:1.el? h5! 26 .l:1.g1 + r3;f7 27 .l:1.xh5 l:1.h8) 25 . . . r3;f7 26 .l:1.xh7+ r3;e6 27 :th6+! (27 l:1.el l:1.d7 28 l:1.h5 .i.f6! and while things are still not clear, Black will apparently be able to give up the f-pawn and retain his advan tage) 27 . . . �d7 28 .l::!. d l + r3;e8 29 l:1.el �f7 30 .l:1.h7+ Wg8 (30 ... Wf6 3 1 l:1.h6+ achieves nothing) 31 �e7! .1i.f6 32 l:1.g1 + Wh8 33 l:i.f7 �e5 34 l:1.e7 �f6 drawn. .
6.242 20 g3!? This was played recently and is ex tremely complicated. According to my analysis, it leads to a probable draw, but Black has some chances:
98
20 . .l:1.e8 This was suggested to me by Lahlum. It is good, but perhaps not the best way to get winning chances. Alternatives: (a) 20 . . . lll g4?? 21 l:1.xe4! wins, al though in Battaglini-Lahlum, Kec skemet 2003, Black survived after 2 1 . . .Wh5 when White missed the in stantly winning 22 'ilkf6!; (b) 20 ... �xg3 is the critical move: 2 1 fxg3 ! (2 1 l:.gl lll g 4! with 22 ltxg3 Wxh6, 22 'ii'x e4+ Wh5 ! ! , or 22 �e3 f5 23 'ii'xg3 'iixg3 24 l1xg3 h5 and Black consolidates his pawn with only slight compensation for White) and here one more split is necessary: (h I) 2 1 . . :iYd7? 22 g4! (22 .ie3 'iYg4+) 22 . . . lllxg4 (22 . . . 'i!t'xg4 23 i.xe4+ lllx e4 24 l:1.g1 and regardless of the material Black's king is too exposed) 23 'iif4! lllf2 + 24 �g2! + - (analysis by Lahlum); (b2) 2 1 . . .l:1.e8! is the way to play for a win, but is difficult to analyse. One line would be 22 i.f4 (22 'i'e3 'i'e5; 22 �e3 .i.g4 23 'ilVf2 'i'e5) 22 . . . 'i'e7 23 'iie 3 (23 �gl .if5 24 'iVhl lll g 4!) 23 . . ..i.f5 24 .l:.adl (24 i.g5? lll g4) 24 . . . l:1.ad8, and Black has the advantage, but it isn't a pawn's worth be cause the bishops count for some thing; ..
Tarrasch Variation: Introduction and 3. . . c5 (b3) 2 1 . . :iie5 is the safe backup : 22 i.f4 'iif5 (or 22 . . . 'iie 6 23 i.. xe4+ liJxe4 24 l:!.xe4 'iVg4) 23 i..xe4 liJxe4 24 'iixe4 'iYxe4+ 25 ':'xe4 i.f5 and the opposite coloured bishops practically ensure a draw. 21 i.f4 1lt'c5 ! 2 1 . . .1lt'b6!? 22 i.e3! avoids various drawn bishops-of-opposite-colours endings, although it's hardly clear. 22 i.e3!? 22 'it>xh2 i.g4= is easy for Black, e.g., 23 1lt'g2 'it>g7 24 'it>gl !? i.f3 25 1lt'h3 liJg4! hitting f2 with ideas like . . . l:!.ad8-d5-h5; then 26 Ji.e3 1lt'f5 27 i.d4+ f6= is best play, when White has nothing special to do and . . . .l:.ad8 follows. 22 .. :�d5! 22 . . :iVb5 is playable but less accu rate.
23 i.d4!? White should avoid 23 'i'f4?? i¥h5 as well as 23 .l:.ad1?! 'iVe5 ! intending 24 i.. d 4? i.. g4 or 24 'iii'xh2? i.. g4 25 'iWh1 1lt'h5+ 26 �gl 1lt'xh1 + 27 �xh1 i..f3+ 28 �gl .l:.e5. Better is 23 1lt'g2! 1lt'h5 24 'iVxh2 i.. h 3!= (with the idea . . . liJg4) 25 i.. d 1 ! liJg4 26 'it>gl liJxh2 27 i.xh5+ 'it>xh5 28 �h2 i.. g4 and only Black has chances, although this is surely drawn. Mter 23 i.. d 4, a logical continua-
tion would be 23 . . . i.. f5 24 'iif4 (24 1lt'g2 liJg4) 24 . . . liJh5! 25 'iie 3 (25 'i!fu4?? e3+) 25 . . . i.. xg3! 26 fxg3 i.. g4 27 �h2 f5 and Black looks solid enough; the bishops have their influence but don't fully compensate for the extra pawn. 6.243 20 .l:.xe4 Probably the safest line. 20 ... Ji. f5 Black's only way to play for compli cations, although White can appar ently bail out. Also leading to 'only' a draw for Black is 20 . . . liJxe4 2 1 �xe4+ �xh6 22 1lt'xh7+ 'iii' g5 with a perpet ual check, in view of 23 g3? Ji.xg3 24 'iYg7+ �h5 25 'it>g2 i.. h 3+! 26 'it>xh3 f5! 27 'iYxg3 1lt'xg3+ 28 'it>xg3 .l:.d2+.
2 1 .l:.g4+! Heading for a perpetual. Other moves seem to favour Black: (a) 2 1 'iie 3 liJxe4 22 i..xe4 f6 23 i.xf5+ 'it>xf5 24 g3 i.xg3 25 fxg3 1lt'e5, and Black is simply an exchange up; (b) 2 1 J:tc4!? 'iYe5!? (2 1 . . .Ji.xc2 ! 22 .l:.xc7 Ji.xc7 would be a winning try) 22 Ji.xf5+ 1lt'xf5 23 'iVxf5+ 'it>xf5, and not 24 �xh2?? liJg4+ but 24 g3 ! .l:.d3!=, a sample line going 25 i.. f4 liJg4 26 .l:.£1 .l:.e8 27 .l:.c5+ 'it>f6 28 l:!.g5 h5! 29 .l:.xh5 l::tf3 30 .l:.h4! �f5 31 I:th5+ �f6= ;
99
Play the French (c) 21 :ae1 is very messy, but I think that Black has good play, e.g., 2 1 . . Jld5!? 22 ..tf4 (22 l:1.g4+? liJxg4 23 .i.xf5+ Wxh6 24 'iVh3+ �g7 25 'iVxg4+ �h8 26 'iVh3 .llxf5 27 'iWxf5 'iVf4- + with both . . . 'iWh6 and . . . 'iVxf2 in mind) 22 . . . ..txf4 23 :xf4 'iVxf4! 24 'iVxf4 i.xc2 25 l:ie3 (25 Ile7 l:iad8-+) 25 ... h5! + ; if 26 .l:!.g3+?, 26 ... liJg4 wins. 2 1 . . liJxg4 22 'iWxf5+ If instead 22 'iixg4+? �xh6 23 'iVxf5 'iie 5! 24 'iVxh7+ Wg5 and the attack is over. 22 ... �xh6 23 'iVxh7+ 'i£?g5 24 'iWf5+ �h6= .
1 00
This is all extremely long analysis and can doubtless be improved upon. Nevertheless, the fate of the whole 4 . . . 'iVxd5 line with 10 . . a6 may be linked to the assessment of Black's 18th moves, so I think that this de tailed look is justified and practically obligatory. My impression is that both 18 . . . e5 and 18 . . . liJh5 equalise, and in the former case White had bet ter know some specifics or he can eas ily come out worse. In conclusion, Black's game is holding up against 1 1 .l:!.e 1 , which is the only serious try at refuting 10 . . . a6 and thus 4 . . . 'iWxd5.
Cha pter Seven Tarrasch Variation : 3 i.. e 7 . . .
1 e 4 e 6 2 d4 d 5 3 lLld2 .i.e7 Only a few years ago this lame looking move was an obscure sideline. Stretching far into the past, Black attacked the centre by 3 . . .lLlf6 or 3 . . . c5 followed by active development. Even 3 . . . lLlc6 and 3 . . . dxe4 seem to have more to do with the position than 3 . . . ..te7. It even looks like a be ginner's idea, developing the bishop to a passive square before it knows where it wants to go and making no attempt to free Black's game or re solve the centre. But of late (mainly in the last five years), 3 . . . i.. e 7 has become a hot main line. In fact, the move in many ways reflects the spirit of the times. It can be viewed as a sort of useful waiting move, and to some extent a prophylactic one. As we shall see, every logical move that White can now make commits him to a position in which Black can play a set-up that he might not have been able to reach via another third move. This is in line with the information theoretic idea that one learns more about the opponent's intentions by doing less! Of course, any tempo used by Black will give White new oppor-
tunities as well. This kind of interplay can only be demonstrated in practice. The most important moves are: 7.1 7.2 7.3 7.4
4 c3 4 e5 4 lLlgf3 4 i.. d 3
White can also play 4 g3, a slow move that allows several answers, e.g., 4 . . . lLlf6 (the rarely-played 4 . . . c5!? is likely to produce dynamic play be ginning with 5 dxc5 lLlf6 6 e5 lLlfd 7 7 lLlgf3 lLlc6 8 .i.b5 lLlxc5=) and: (a) 5 i.. g2 0-0 6 lLle2 c5 7 dxc5! (7 exd5?! exd5 8 c3 lLlc6 9 0-0 Baran Bobrowska, Suwalki 1999, and one way to advantage is 9 . . . i.. g4! 10 f3 i.. f5=i=) 7 . . . dxe4 8 b4 (8 0-0 'iVd5 !? 9 lLlc3 'iVxc5 10 lLldxe4 lLlxe4 1 1 lLlxe4 'ikc7=) 8 . . . a5 9 c3 axb4 10 cxb4 lLlc6 1 1 'i!Vb3 lLld5 (or 1 1 . . :iVd3 1 2 a3 e5=i=) 12 i.. a 3?! (12 0-0 f5) 12 ... .:.xa3! 13 'ii'x a3 lLldxb4 14 i..x e4 f5 !; Oumnov Stimpel, Bundesliga 1 999, since 1 5 i.. xc6 runs into 1 5 . . . lLld3+; (b) 5 e5 lLlfd7 6 �g2 (6 f4 c5 7 c3 lLlc6 7 lLldf3 would be a standard 3 1 01
Play the French lDd2 lDf6 variation which is fine for the second player. Black can also play 7 . . . b6 8 lDdf3 cxd4 9 cxd4 a5 10 �d3 i.. a 6= Rapp-Lenz, Hauenstein 1991) 6 . . . c5 7 c3 lDc6 S lDe2 'Wb6 (8 . . . 0-0 9 0-0 cxd4 10 cxd4 f5!? 1 1 f4!? lDb6 12 lDf3 a5 intending ... a4; this is a stan dard set-up in the 3 lDd2 lDf6 4 e5 lDfd7 5 f4 Tarrasch) 9 lDf3 cxd4 1 0 cxd4 f6 1 1 exf6 lDxf6 12 0 - 0 0-0= in tending . . . lDe4.
version of a 3 lDd2 c5 Tarrasch line) 6 . . . lDc6 (or 6 . . . 1l.d7 7 1l.xd7+ lDxd7 S dxc5 lDxc5=) 7 lDgf3 'Wb6! and Black has already equalised, e.g., S 'iVe2 c4! 9 �a4 lDf6 10 0-0 �e6 11 lDg5!? i.. f5! 12 l:.el (12 'ti'e5 ..id3 1 3 l:.e1 O-O=i=) 12 . . . 0-0 1 3 lDdf3 ..id6=i= ; (b) 5 lDgf3 lDf6 (5 . . . cxd4 6 lDxd4 dxe4 7 lDxe4 lDf6) 6 exd5 exd5!? 7 dxc5 1l.xc5 S lDb3 �b6 9 i..b 5+ lDc6 10 0-0 0-0 1 1 .tg5 i.. g4= . 5 ... 1l.xc5
7.1 4 c3
A fairly harmless move after which Black should be able to equalise eas ily, but he can also enter into compli cations should he choose to do so. 4 ... c5 This is a good move if Black wants positive chances. 4 . . . dxe4 5 lDxe4 is a kind of Rubinstein Variation that can't be too bad for Black, since the inclusion of c3 is not always appro priate; nevertheless the text is more exciting.
5 dxc5 (a) 5 exd5 exd5 (the easiest course, but 5 ... 'iWxd5 has also equalised after 6 dxc5 i.. xc5 7 lDgf3 lDf6; a simple answer to 6 lDgf3 is 6 . . . cxd4 7 1l.c4 'Y!Vh5 8 cxd4 lDf6=) 6 .tb5+ (6 lDgf3 lDc6 7 dxc5 i.. xc5 S lDb3 �b6 is a nice
1 02
6 lDb3 (a) 6 lDgf3 lDf6!? (6 ... lDc6 7 lDb3 .tb6 S exd5 exd5= gives Black active play in an isolated queen's pawn posi tion) 7 e5 (7 exd5 exd5 8 lDb3 �b6) 7 . . .lDfd7 (a safe move; 7 . . . 'iVb6!? 8 lDd4 lDfd7 9 'Y!Vg4!? was unclear in Onis chuk-Morozevich, Bundesliga 1999; the best idea was probably just 9 ... g6, but the game got wild following 9 . . . 0-0!? 10 lD2f3 lDc6?! 1 1 .i.h6 g6 12 0-0-0 lDdxe5 1 3 'iVf4! f6!? 14 �xfS?! i..xf8 15 Wxf6 lDg4 16 'iVh4 e5 and Black had more than enough com pensation) 8 lDb3 ..ib6 9 lDbd4 lDc6 10 i..b 5 'Wic7= Vajda-Mkrtchian, Bled 2002, a critical line going 1 1 �f4 1l.xd4 1 2 lDxd4 lDdxe5 13 0-0 0-0 1 4 We2 f6 1 5 l:tae l 'Wb6= . But 6 . . . lDc6 is probably easiest; (b) 6 ..id3 lDf6 7 We2 resembles one
Tarrasch Variation: 3 ... �e 7 of the main 3 liJd2 .i.e7 4 ii.d3 c5 5 dxc5 lines, but White has committed to c3 rather early on. One idea for Black is 7 . . . liJc6 8 liJgf3 O-O!? 9 0-0 (9 e5 liJg4! 10 0-0 f6=) 9 . . . ii.b6!? 1 0 e5 (10 b3 e5!) 10 . . . liJd7 1 1 liJb3 £6 12 'ikc2 (12 exf6 liJxf6 13 i.g5 'ikd6! with the idea 14 'ikc2? e5! 1 5 �xf6 'ikxf6 16 .i.xh7+ �h8+) 12 ... liJdxe5 1 3 liJxe5 fxe5 14 .i.xh7+ �h8 15 .i.g6 (15 'ikg6?? 'iVh4) 1 5 ... e4 with at least equality. 6 ... i.b6 7 exd5 7 e5?! liJc6 8 liJf3 liJge7 9 .i.d3 liJg6 10 'ike2 'fic7 1 1 .i.xg6 fxg6+. 7 ...exd5 8 liJf3 8 'ike2+ liJe7 9 .i.e3 i.xe3 10 'ikxe3 0-0 1 1 i.e2 liJf5 12 'iVd2 .l:!.e8 13 liJf3 'ike7 1 4 liJfd4 liJxd4 1 5 liJxd4 liJc6 16 liJxc6 bxc6 1 7 0-0 , V h Zatonskih Shulman, Lindsborg 2002. 8 ... liJf6 9 i.b5+ liJc6 10 0-0 0-0 1 1 i.g5 a6 Black can also play 1 1 . . .i.g4, e.g., 12 .i.xf6 (12 'ikd3 'ikd6=) 12 ... 'ikxf6 13 'ikxd5 .i.xf3 14 'ikxf3 'ikxf3 1 5 gxf3 liJe5 1 6 �g2 liJg6= . 12 i.e2 .l:.e8= Langer-Atalik, Las Vegas 200 1 . Black can unpin by . . . 'ikd6 or in some cases even . . . h6 and . . . g5. 7.2 4 e5
A fascinating and critical variation that prevents Black's knight from developing naturally to f6 (and e7 is already occupied) . White's Advance Variation centre is subject to a seri ous attack (his knight on d2 is not ideally placed) , but the idea of 'ikg4 is difficult to meet and gives the strug gle an added dimension. 4 ... c5 This is almost alwllYs played, but I think that the move 4 . . . liJh6!? de serves serious consideration.
It brings out a piece, prevents 'ikg4, prepares castling, and plans to de velop the knight to f5 (or in some cases to f7 after the move .. .£6). White will of course try to exploit the knight on the rim and consolidate his pawn wedge in the centre: (a) 5 i.d3 c5 (another plan is 5 . . . b6 6 liJdf3 liJf5 7 liJe2 .i.a6) gives Black time for standard counterplay against d4, e.g., 6 c3 (6 dxc5 liJd7 7 liJb3 liJxe5) 6 . . . cxd4 (6 . . . liJc6 7 liJe2 cxd4 8 cxd4 liJf5 9 liJf3 0-0 10 0-0 £6) 7 cxd4 liJc6 8 liJgf3 (8 liJdf3 'i'a5+!? 9 .i.d2 �6 or simply 8 . . . liJf5) 8 . . . 'i'b6 9 liJb3 a5 1 0 a4 (10 i.. xh6 gxh6 1 1 a4 �4+! 12 �e2 £6+ opening lines for attack) 10 . . . .i.b4+ 1 1 �e2 (1 1 i.d2 .i.xd2+ 12 liJbxd2 liJxd4) 1 1 . . .liJf5 12 g4 liJfe7 1 3 .i.e3 .i.d 7 intending . . . l:.c8 and either 1 03
Play the French . . . h5 or . . . 0-0 and .. .f6; (b) 5 �gf3 O-O!? (a little committal; safer would be 5 . . . �f5, and very in teresting is 5 . . . �d7 6 �b3 [6 .td3 c5=] 6 . . . �f5 7 .i.d3 a5!?; then we could see 8 a4 h5 intending . . . b6, or some thing wild like 8 g4!? �h4 9 �xh4 .txh4 10 g5 ..txg5 1 1 'iWg4 J.xc l 12 'iWxg7 ..txb2 1 3 1iVxh8+ �f8 1 4 l1b l J.c3+ 1 5 'it>e2 �4 with a pawn and plenty of play for the exchange!) 6 ..td3 f6 7 �b3 �f7 8 exf6 .i.xf6 9 0-0 �c6 and . . . e5 next, except after 10 .tf4 g5!? 11 ..tg3 g4 12 �e5 J.xe5 13 dxe5 �cxe5 14 "iVe2 Vg5 when White definitely has compensation but this is probably only sufficient and not more; (c) 5 �df3 (the most obvious move and the one played in most contests with 4 . . . �h6 thus far - everything above is analysis that I did a couple of years ago) 5 . . . c5!? (ignoring the 'threat' of ..txh6; or 5 . . . �f5 6 .i.d3 �c6!?, e.g., 7 c3 f6 8 exf6 .i.xf6 9 �e2 0-0 10 O-O?! e5! 11 dxe5 �xe5 12 �xe5 ..txe5 13 .tf4 J.xf4 14 �xf4 c6 15 c4?! 12 - 12 Gajewski-Bartel, Laczna 2002; Black keeps the advantage with either 1 5 . . . �d6 or 1 5 . . . �h4! 1 6 g 3 "iVg5) 6 ..td3 ( 6 c 3 �c6 7 dxc5 ..txc5 8 ..txh6 gxh6= ; 6 J.xh6!? gxh6 7 'iWd2 �c6 8 'ilVxh6 'iVa5+!? 9 'iWd2 'iWxd2+ 1 0 �xd2 �xd4 11 �xd4 cxd4 12 � f3 O-O! 1 3 �xd4 f6 with compensation) 6 . . . �c6 7 .txh6 gxh6 8 'iVd2 (this looks more critical than 8 dxc5 J.xc5 9 �h3 'iWc7 10 0-0 ..td7!? l 1 l:te l 0-0-0 12 c3 f6! 13 'ilVe2 �xe5 1 4 �xe5 fxe5 1 5 'ii'x e5 ..td6; Czarnota-Olszewski, Laczna 2002) 8 . . . 'i!Vb6 9 0-0-0 .i.d7 with the idea 10 Vxh6? c4! 11 'i¥g7 (1 1 .i.xh7 �b4! with the idea of meet ing 12
5 'iWg4
There are two other important moves here: (a) 5 c3 is rather slow but Black still needs to play actively: 5 . . . cxd4 (5 . . . �c6 6 �gf3?! cxd4 7 cxd4 'iVb6! is a favourable version of certain Ad vance Variation lines: 8 �b3 a5 9 a4 .i.b4+ 10 ..td2 ..txd2+ 1 1
and now:
Tarrasch Variation: 3. . . .i.e 7 (a1) 7 . . :iVb6! 8 .i.d3 (8 liJe2 liJh6!? 9 .i.xh6 gxh6 10 'i'd2 f6! 1 1 exf6 .i.xf6 12 0-0-0 0-0 with the idea 13 'i¥xh6 e5 unclear; here 8 . . . f6 might be a good option) 8 . . . .i.b4+! looks fully equal, with the idea 9 �d2 liJxd4 10 liJxd4 'i'xd4 1 1 'i'a4+ .i.d7 12 'iYxb4 'i'xd3 13 'i'xb7 l:[c8:j: ; (a2) 7 . . . .i.b4+ 8 .i.d2 'i'a5 (or 8 . . . 'i'b6, which transposes to 7 . . . 'i¥b6 above after 9 .i.d3?; best is 9 �xb4 'i'xb4+ 10 'iVd2 .i.d7! and not instead 1 1 'i'xb4?! liJxb4 12 'ot>d2 lIc8, but 1 1 1:tc 1 ! liJh6!? 1 2 'i¥xb4 liJxb4 1 3 a 3 liJc6 14 .i.d3 f6, which is approximately equal)
1 7 c;t;?d3 dxe4+ 1 8 f8) 6 . . . Jixc5 7 liJb3 �b6 8 �d3 f6! 9 'i'e2 (9 .i.b5 .i.d7 10 i.xc6 bxc6=) 9 . . . fxe5 10 liJxe5 liJf6 1 1 0-0 liJxe5 12 'i'xe5 0-0 1 3 .i.g5 (Ftacnik analyses 1 3 �g5 e5! 14 'i'xe5 liJg4 1 5 � 5 g 6 16 �xg6 hxg6 1 7 'i'xg6+
and now: (a2 1) 9 �d3 f6!? 10 exf6 liJxf6 1 1 a3 �xd2+ 12 Vi'xd2 'i¥xd2+ 13 f8 15 'i!Va3+ liJe7 16 liJf3 'i'b6 17 �d3 �d7 1 8 0-0 f6 unclear; (a23) 9 liJe2 .i.xd2+ 10 'i'xd2 'i¥xd2+ 1 1 �xd2 f6 12 exf6 liJxf6!? 1 3 liJ c 3 0 -0 (or 1 3 . . . liJg4 1 4 .i.b5 ! ? 0- 0 1 5 l:!.afl .i. d 7 16 h 3 liJh6=) 1 4 .i.b5 liJe4+ 15 'ot>e3 liJe7!? 16 .i.d3 (16 liJxe4? liJf5+
This hasn't been popular but strongly deserves consideration. It weakens the kingside somwhat but avoids moving the king. Whether the weaknesses can be exploited is an other issue; I think that the benefits outweigh the problems. 6 dxc5 (a) 6 liJdf3 liJc6 7 dxc5 .i.xc5! 8 �d3 f6! 1 05
Play the French
This position is highly relevant to the whole system, as it transposes to the line with 6 dxc5 ltJc6 7 ltJdf3 .i.xc5: (al) 9 'ilVh3 fxe5 ! (9 ... 'it'f7 may also suffice, e.g., 10 'i!Vg3 'i!Vc7 1 1 �f4 'i!Vb6) 10 �xg6+ 'it'd 7 1 1 .i.g5 (1 1 .i.f7 'i'f6!) 1 1 . . .ltJf6 1 2 .i.f7 iLxf2+! 1 3 �e2! �d6! 1 4 1:.f1 1:.f8:j: ; (a2) 9 'ilVh4!? fxe5 ! (9 . . . ltJxe5 1 0 ltJxe5 fxe5 11 i.. xg6+ 'it> d 7 12 .i.g5 'i!Vb6 13 0-0-0 ..td4! 14 c3 .i.xf2 with great complications) 10 .i.xg6+ 'it>d7 11 'i!Vxd8+?! ( 1 1 iLg5 ! �6 12 0-0-0 i..xf2 1 3 'iVh5 ltJge7=) 1 1 . . .'it>xd8 12 ltJg5 (12 .i.h5 ltJf6 1 3 .i.g5 11fS 1 4 ltJe2 e4+) 12 . . . hxg6 13 ltJf7+ 'it>e7 1 4 ltJxh8 'it>f6 and the knight does not escape, so Black stands better; (b) 6 c3 ltJc6 7 ltJgf3?! , although played several times without penalty, is actually a mistake after 7 . . . cxd4 (also good is 7 . . . h5 8 'iVg3 h4 9 'iVh3 cxd4 10 cxd4 'i'b6) 8 cxd4 ltJh6! 9 'i!Vh3 ltJf5 10 .i.b5 'i!Vb6 1 1 i.. xc6+ bxc6 12 ltJb3 �a6+ . 6 ...ltJd7 To get a knight to c5. 6 . . . ltJc6 is playable although White is a touch better: 7 ltJgf3 i.. xc5 8 ltJb3 iLb6 9 .i.d3 W1c7 10 i.. f4 ltJb4! 1 1 ltJfd4! ltJxd3+ 12 cxd3 i.. xd4 13 ltJxd4 �6 1 4 ltJb3 iVb4+ 15 'it>e2�. 1 06
7 ltJgf3 7 J.b5 �xc5 and there's a crazy line following the natural 8 ltJgf3: 8 . . . .i.xf2+! 9 'it'xf2 'i!Vb6+ 10 ltJd4 ltJh6 11 'i!Vf4 g5! and Black stands well. 7 ... ltJxc5 8 .i.e2 8 h4 h5!? 9 'i'f4 ltJh6 was unclear in Nykopp-Hoi, Reykjavik 1984. 8 ... ltJh6 9 'i'h3 ltJf5 10 g4 ltJg7! These knight fianchettoes are in creasingly common in the French! As usual, Black has ideas of . . . h5 and .. .f5 or . . . f6. Now Jaracz-Lputian, Is tanbul 2003 continued 1 1 'ilVh6 i..fS 12 h4 .i.d7 13 llh3?! 'i'c7 14 'i'f4 f5! 1 5 ltJd4? fxg4 16 l::[ c 3 ltJh5+. 7.22 5 ...'it>f8
The most popular move, by which Black avoids weakening his position
Tarrasch Variation: 3. . . Ji.e 7 and bets that White's precarious cen tre and the awkward position of the queen will compensate him. 6 dxc5 (a) 6 ttJdf3 ttJc6 7 dxc5 Ji.xc5 8 Ji.d3 f6 transposes to the 6 dxc5 ttJc6 7 ttJdf3 line below, but Black also has 7 . . .£6 8 Ji.f4!? Ji.xc5 9 Ji.d3 �6! 10 ttJh3! �4+ 11 c3 �xb2 12 0-0 �xc3 1 3 Ji.b5 ! fxe5 14 l:i.fc l �4 1 5 l:i.ab l ttJf6!?, which led to an unclear, titanic struggle of material versus initiative in Pokorna-Matveeva, Bled 2002. Here 1 5 . . . i.xf2+ 16 'it>xf2 �e4 17 Ji.xc6 ttJf6 is a good option given by Shipov; (b) 6 ttJgf3 h5!? (6 . . . ttJc6 7 dxc5 Ji.xc5 again transposes to 6 dxc5) 7 �g3 h4 8 'iWf4 g5 9 �g4 ttJc6 10 dxc5 (10 ttJb3!? is worth thinking about. Then 10 . . . c4 1 1 ttJc5 b6 12 ttJa4 ttJb4 13 'it>dl f5! 14 exf6 i.xf6 gives a lot of play, e.g., 15 i.xg5 e5! 16 Ji.xf6 �xf6 1 7 dxe5 .i.xg4 18 exf6 ttJxf6 1 9 a3! ttJc6 20 �e2 l:te8 with complications) 1O . . . ttJh6 1 1 n5 f5 12 exf6 Ji.xf6 1 3 ttJb3 ttJf5 14 �g4 e5 1 5 �a4 e 4 1 6 ttJ g l Ji. d 7 1 7 .i.b5!? (17 c3 had previ ously been played) 1 7 . . . l:i.h7! 18 ttJh3 ttJcd4 19 �xd7 'ii'x d7= Schaar Andeer, email 2000; (c) 6 c3 ttJc6 7 ttJb3?! (7 dxc5) 7 . . . cxd4 8 cxd4 a5 (8 . . . ttJb4!?) 9 ttJf3? (9 Ji.d2) 9 . . . ttJb4! 10 �dl Rogovoi Kashtanov, St Petersburg 1 999; and simply 1O . . . �c7 1 1 ttJel Ji.d7 intend ing . . . l:i.c8 looks very strong; (d) 6 ttJb3 is a straightforward and perhaps underrated move: 6 . . . ttJc6!? (6 . . . cxd4 7 ttJf3 ttJc6 8 ttJbxd4;l; ; 6 . . . c4 7 ttJd2 f6 8 ttJgf3 ttJc6 may be best) 7 ttJxc5 ttJxe5 8 dxe5 i.xc5 9 Ji.d3 Ji.d7 Wells-Pert, Southend 2002; and the natural 10 ttJf3 should yield a nice advantage if only because Black has little to show for his king position.
6 ...ttJc6
6 . . . h5!? has been played in many games and is still considered unclear, but I'm not convinced that Black gains by committing himself. 7 ttJgf3 Natural, but this is a juncture at which White has improvements: (a) 7 �g3 is a good option: 7 . . :�i'c7!? 8 ttJgf3 (8 ttJdf3 Ji.xc5 9 Ji.d3 ttJge7=) 8 . . . Ji.xc5 9 ttJb3!? (9 �d3! ttJb4 10 O-O;l; is a problem) 9 . . . �b6 10 Ji.d2?! f6 1 1 Ji.c3 fxe5 1 2 0-0-0 ttJf6! 1 3 ttJxe5 ttJe4 1 4 �f3+ 'it>g8 15 ttJxc6 bxc6=i= ; (b) 7 �e2 !? also has good points: 7 . . .i.xc5 8 ttJb3 i.b6 9 ttJf3 f6 10 �e3! fxe5 11 Ji.xb6 (better is 11 ttJxe5! ttJxe5 12 Ji.xb6 axb6 13 �xe5 �f6 14 �d6+ 'it>f7 1 5 c3;l;) l L .axb6 12 ttJxe5 Y2 - Y2 Andreev-Itkis, Alushta 2000; 12 .. :�f6 would follow with an equal game, since Black's centre becomes mobile; (c) 7 ttJdf3 f6 8 �g3 (8 �f4 i.xc5 9 Ji.d3 �6!? was seen under the move order 6 ttJdf3 ttJc6 7 dxc5) 8 . . . Ji.xc5 9 Ji.d3 fxe5 10 ttJxe5 ttJxe5 1 1 �xe5 ttJf6 12 ttJf3!? i.xf2+ 1 3 'it>dl (13 'it>e2 !?) 1 3 . . . i.c5 14 l:i.f1 i.d6 1 5 �e2 �c7 and White seemed short of compensation in the game V.Akopian-Pelletier, Au bervilliers 2002, although Black still had to resolve his king position. 107
Play the French 7 h5! ...
This move has been quite success ful for Black and makes one wonder whether White does well to continu ally enter into this line instead of choosing a deviation earlier. His problem is that there is no safe place from which the queen cannot be at tacked without giving up the centre. Black continually presses for the ini tiative and has thus far succeeded. 8 'i!ff4 (a) 8 'i!fa4!? 'fic7 9 'i'f4 (paradoxical, but now White has discouraged . . . g5 by diverting Black's queen to c7) 9 . . . f6 1 0 ltJh4 (10 ltJb3 ltJxe5 1 1 �e3 ltJh6 looks about equal) 10 . . . 'i'xe5+ 1 1 'itxe5 ltJxe5 1 2 f4 g5 1 3 fxe5 gxh4 1 4 ltJ f3 .i.xc5 1 5 ltJxh4?! (The play has been forced to this point. Perhaps White could reinforce e5 with 15 .tf4, but his advantage is small in any case) 1 5 . . . �g7 16 .i.f4? (16 ltJf3) 16 . . . .i.d4! 1 7 ltJf3 .ixb2 1 8 11b l .i.c3++ Liss-I.Botvinnik, Ramat Aviv 2000; (h) 8 'itg3 h4 9 iVf4 g5! 10 'iVe3!? (The alternative is 10 'iWa4, when one of the stem games for this variation went 1O . . . ..td7 1 1 .i.b5 a6 12 .i.xc6 .txc6 13 'iWd4 ltJh6 14 ltJb3 ltJf5 1 5 'iWd3 d4! 16 .l:tgl 'fic7 1 7 ltJfxd4 'i'xe5+ 1 8 ..te3 'fixh2 19 0-0-0 with dynamic equality, Adams-Morozevich, Sara1 08
jevo 1999. But there are many op tions along the way and White may want to return to this line) 1O . . . ltJh6 (threatening . . . ltJf5 with tempo) 1 1 .td3 ltJg4 1 2 'iVe2 'i¥c7!? (12 . . . .txc5 1 3 ltJb3!? ..txf2+ 1 4 �f1 ..tb6 1 5 .txg5 'iWc7 is fine for Black) and: (hI) 13 ltJb3 lIg8! 14 0-0 (14 .i.d2 ltJgxe5 15 0-0-0 ltJxd3+ 16 'ili'xd3 b6! 1 7 cxb6 axb6 gave Black more than enough compensation with his two bishops and open lines on the queen side in P .Popovic-Del Rio Angelis, Halkidiki 2002) 14 . . . ltJgxe5 15 .td2 ltJxd3 16 'ili'xd3 llg6 17 l:tfel g4 1 8 ltJfd4 ltJxd4 1 9 ltJxd4 'itxc5 20 ..tf4 ..td7+ P.Popovic-Kosic, Yugoslavia 2002; (b2) 13 ltJxg5! may the best try. It tries to keep Black's king exposed. Then 13 . . . 'i'xe5?! is mentioned by McDonald along with the strong reply 14 ltJb3 ! , but I think that 13 . . . ltJgxe5! makes a better impression. Black can get his pawns moving after 14 ltJb3 ltJxd3+ 15 cxd3 a5! intending 16 a4 b6! 17 .te3 bxc5 18 licl f6 19 ltJf3 d4 20 .td2 'i!Vb6! etc. 8 g5! 9 'iVe3 ltJ h 6 So we have the same position as we did after 8 'ili'g3 except that Black's pawn isn't on h4 (White went to f4 in one move) . This makes the move 9 . . . d4!? more plausible: 10 'fie2 (10 'ili'e4 g4! 11 ltJgl f5 ! 12 exf6 ltJxf6 13 'ili'd3 ..txc5;) 10 ... g4 11 ltJgl ..txc5 and Black has a space advantage. His problem is as usual the king. Play might go 12 h3! (12 ltJb3 'i'd5! 1 3 ltJxc5 'i!fxc5 14 h3 ltJxe5 1 5 hxg4 ltJxg4 and Black is temporarily a pawn ahead) 12 . . . g3 !? (to keep lines closed) 13 f4 ltJb4 14 ltJe4 .i.e7 15 a3 'fia5 16 c3! ;t . 10 .i.d3 1 0 ltJb3 ltJf5 11 'i!fd2 g4 12 ltJfd4 ...
Tarrasch Variation: 3 . . i.. e 7 .
ttJfxd4 1 3 ttJxd4 ttJxe5 wins a key cen tral pawn and should give adequate play.
1 0 ... ttJg4 Black has also tried 10 . . . 'i!Vc7 on several occasions: 1 1 ttJb3 (Black's idea is to trade a flank for centre pawn after 1 1 ttJxg5?! ttJg4 12 1IVf4 'ii'x e5+ 1 3 'ili'xe5 ttJgxe5 1 4 ttJb3 ttJxd3+ 15 cxd3 11g8 and Black wins his pawn back) 1 1 . . .g4?! (1 1 . . .ttJg4 looks best) 12 ttJfd4 'iYxe5 1 3 ttJxc6 'lWxe3+ and here Todorovic-Antic, Vrsac 2001 saw 14 fxe3! (14 J.xe3 bxc6 1 5 ttJa5 i.d7 16 0-0 e5 1 7 a3 f5 ! was more than satisfactory in Wang Yu-Kamble, Mumbai 2003) 1 4 . . . bxc6 15 e4 (or 1 5 O-O!) 15 . . .f6 (15 . . . .i.h4+!? 1 6 g3 .i.f6 - Antic) 1 6 i.e3 with an extra pawn, although Black has al most enough central mobility to com pensate. 1 1 'i!Ve 2 ..txc5 12 ttJb31 12 0-0 ttJxf2 ! 1 3 l:txf2 g4 14 ttJel YWb6 1 5 ttJb3 i.xf2+ 1 6 'iixf2 'iixf2+ 17 'iitxf2 ttJxe5 and Black's extra central pawns give him equal chances. 12 ...i.xf2+ 13 'it'd1 13 �f1 Jl.b6 14 ..txg5 'i!Vc7 15 h3 ttJgxe5 16 ttJxe5 ttJxe5 17 i.f6 ttJxd3 18 i.. xh8 ttJf4 19 Jl.e5 ttJxe2 20 ..txc7 i.xc7 21 �xe2 is assessed as slightly better for White by Antic, but this is
hard to believe because Black has the bishop pair, a pawn and a mobile cen tre in return for the exchange. We are following Blehm-Hausner, Ostrava 1999. At this point 13 . . . 'iib 6 was promising, the likely course of play being 14 i.. xg5 J.e3! 15 J.xe3 'lWxe3, when Black actually wins a pawn, although 16 'ilxe3! ttJxe3+ 17 'it'd2 ttJxg2 18 l:.hgl gives White some compensation, e.g., 18 . . .ttJf4 19 :afl ! ttJxd3 20 cxd3 .i.d7 2 1 ttJc5 'it'e7=i= . To play 5 . . . 'it'f8 one has to be will ing to sacrifice a pawn for long-term compensation. There is generally a reward for doing so, but not always. In addition White seems to achieve small edges in some of the early side lines. None of this is too worrisome but it makes me inclined to lean to ward 5 . . . g6 (I don't see a problem with it yet), and truly experimental types could look into 4 . . . ttJh6. 7.3 4 ttJgf3
White develops naturally and leaves open the question of what to do with his bishop. In return, Black uses the commitment of White's knight as an inducement to switch strategies. 4 ...ttJf6 Provoking the closing of the centre and a pawn chain struggle. 1 09
Play the French 5 e5 The only try for advantage. (a) 5 .i.d3 c5 is a subtle move order. Now 6 e5 lLlfd7 transposes to the main line, and White has nothing else particularly dangerous:
la Paz-Arribas, Santa Clara 2001) 9 0-0 0-0 10 iLe3 (10 �c2 lLlb4 I l lLlxf6+ gxf6! 12 'i'dl lLlxd3 1 3 'iYxd3 'ilthS=) 1O . . . lLlb4 (or 1O . . . lLld5=) 1 1 .i.b l b6 12 lLlxf6+ (12 lLlg3 ..tb7 13 a3 lLlbd5 14 'i1id3 'i1id6; Kosintseva-Matveeva, Elista 2002) 12 . . . ..txf6 13 �e4 lLld5= ; (b) 5 exd5 exd5 is a comfortable Exchange Variation for Black, since White's knight gets in the way on d2: 6 .i.d3 0-0 7 0-0 ..tg4 S c3 c5!? 9 dxc5 ..txc5 10 lLlb3 .i.b6 1 1 .l:!.el (1 1 ..tg5 lLlc6 12 h3 .i.h5 13 l1el 'ifd6=) 1 1 . . .lLlc6 12 ..te3?! .l:teS 13 ..txb6 'i'xb6 14 .te2 a5! 1 5 J:.b l ..tf5 16 �d3 l1xe l + 1 7 lLlxe 1 lLle4; was the game Akhmadeev-Rustemov, Moscow 1 999; Black intends . . . lLle5 and . . .lieS. 5 lLlfd7 ...
(al) 6 dxc5 lLlc6 or 6 . . . 0-0 would transpose to 4 iLd3 c5 5 dxc5 etc.; but Black has a nice solution in 6 . . . dxe4 7 lLlxe4 lLlxe4 S ..txe4 'iYxdl + 9 �xdl ..txc5=, a position dealt with in the next section via 4 .i.d3 c5 5 dxc5 lLlf6 6 lLlgf3 dxe4 etc.; (a2) 6 exd5 'i!Vxd5 7 dxc5 transposes to the order 4 �d3 c5 5 dxc5 lLlf6 6 exd5 'i'xd5 7 lLlgf3 in the next section, and 6 . . . exd5 is also fine: 7 dxc5 0-0 S 0-0 (S b4 a5 9 c3 axb4 10 cxb4 lLlc6 1 1 l:tb l �xa2=; S lLlb3 lLlbd7 [S . . .a 5 is more interesting, but this will do] 9 ..te3 'iic 7 10 0-0 lLlxc5=) S . . . ..txc5. An isolated que en's pawn position arises in which Black has taken two tempi to capture on c5, but White's bishop is not ideally placed on d3. In Chow J.Watson, Chicago 2003, play contin ued 9 lLlb3 .i.b6 10 h3 lLlc6 1 1 c3 .l::i. e S= ; (a3) 6 c3 cxd4 (6 . . . lLlc6 transposes to the main line after 7 e5 lLld 7) 7 cxd4 dxe4 S lLlxe4 with an easy-to play isolani position: S . . . lLlc6 (or S . . . lLlxe4!? 9 ..txe4 lLld7 10 0-0 0-0= De 110
6 .i.d3 (a) 6 c4 has been played a bit but is easily met: (al) 6 . . . c5 breaks up the centre, e.g., 7 cxd5 exd5 S 'iYb3 cxd4 9 "iVxd5 lLlc6 10 iLb5!? lLlb4 1 1 "iVe4 a6 12 ..txd7+ 'ifxd7 1 3 a3 lLlc6 14 lLlb3 d3 1 5 0 - 0 0 - 0 and . . . 1IdS; (a2) 6 . . . 0-0 asks White what he's up to: 7 cxd5 (7 .i.d3!? c5) 7 . . . exd5 S .i.d3 c5 9 'i'c2 h6 10 dxc5 (10 0-0 lLlc6;) 10 . . .lLlxc5 11 0-0 lLlc6 and White has no good place for his important light squared bishop .
Tarrasch Variation: 3. . . �e 7 (b) 6 c3 c5 7 a3!? is an interesting attempt that strongly resembles the a3-b4 ideas in the Advance Variation: 7 . . . a5 (7 . . .'�Jc6 8 b4 cxd4 9 cxd4 seems to fall in with White's plans, but Black has a clever change of plans in mind: 9 . . . lLlb6! 10 �d3 �d7 with queenside light square pressure, L. Cooper-Cobb, Newport 200 1 . One idea is . . . a6, . . . lLla7, and . . . �b5) 8 i.b5?! (this doesn't work out well, but 8 �d3 b6 and . . . �a6 is one easy solu tion, and 8 . . . a4!? is another way to fight for the light squares) 8 . . :�Yb6 9 'ilfa4 0-0 10 0-0 lLlc6 1 1 l:tel (White has little to do) 1 1 . . .f6 12 exf6 lLlxf6 1 3 i.d3 i.d7 14 �c2 cxd4 1 5 c4 h6 16 b3 .l:tac8 1 7 �b2 ii.d6+ Rozentalis S.Ivanov, Halkidiki 2002. 6 c5 7 c3 lLlc6 B 0-0 Other moves aren't of much inter est here, and most of them can be met by a plan with . . .f6, e.g., 8 b3?! f6 9 0-0 fxe5 (9 . . . cxd4 10 cxd4 fxe5 1 1 lLlxe5 lLldxe5 1 2 dxe5 lLlxe5 1 3 'iVh5+ lLlf7:j:) 10 lLlxe5 lLldxe5 1 1 dxe5 lLlxe5 12 'i!Vh5+ lLlf7 13 �xh7 Sutor Vyoshin, Krakow 1999, and easiest was to get to the queenside by 13 . . .�f6 14 �b2 'iVa5 15 lLlf3 �d7 16 :fe l 0-0-0 1 7 'ilfxf7 lhhn.
c3 lLlc6 7 �d3 �e7!? 8 0-0. Normally White would rather have his queen's knight go to f3 and his king's knight to e2, but it does take him one less tempo to arrange his knights this way and there are various attacking ideas that can stem from the lLlgf3 move. Furthermore, Black has com mitted his king's bishop on e7 whereas in the corresponding posi tion he had 7 . . . 'iVb6 8 0-0 g6 or 7 . . . g6 Thus both sides have made small concessions from what they might ideally like and this tradeoff leads to unique play in fairly unexplored ter ritory. B a5 ...
...
This position arose in the old days via 3 lLld2 lLlf6 4 e5 lLlfd 7 5 lLlgf3 c5 6
A subtle and multipurpose move that has become one of the main lines here, but no one knows what's best. Perhaps it shouldn't come as a sur prise that the two moves receiving the most attention are 8 . . . a5 and 8 . . . g5, both flank moves that don't develop a piece or even prepare to develop one. These days centres are both protected and attacked from the wings. Black has two other exciting ap proaches: (a) It is worth mentioning the old line 8 . . . �6!? 9 .l:tel (9 dxc5! is the most critical move) and now an at111
Play the French tractive mixing of systems is 9 . . . g5! 1 0 liJf1 (10 c4 g4 11 cxd5 exd5! + , e.g., 12 e6 fxe6 1 3 liJe5 liJdxe5 1 4 dxe5 c4 1 5 i.. c 2 ii.c5-+) 10 ... g4 11 dxc5 �xc5 12 liJd4 liJdxe5 13 i.. f4 liJxd3 14 'ifxd3 liJxd4 15 cxd4 �xd4 16 �e3 �xe3 1 7 liJxe3 0-0 1 8 liJxg4 f6 and Black has stood well in a couple of games (giv ing back the pawn via . . . e5 is a theme); (h) An older, likely better, and still hotly contested line is 8 . . . g5!?, to chase the f3 knight away from de fence of the centre and/or to grab space.
I won't attempt to cover all the dense theory of this variation, but here are a few lines that should indi cate roughly where things stand: (h I) 9 b4?! cxd4 10 cxd4 liJxb4 1 1 �b 1 b 6 1 2 a3 liJc6 1 3 liJb3 .ta6 1 4 .:tel �c4:j: Conquest-A. Summerscale, England 200 1 ; (h2) 9 h 3 i s one o f the oldest moves, but Black seems fine after 9 . . . h5 10 g4 hxg4 1 1 hxg4 'iib 6 12 'iia 4 'iic 7 (or 12 . . . cxd4 13 cxd4 f6) 13 !:tel liJb6 1 4 'iV d 1 cxd4 1 5 cxd4 �d7 1 6 liJ b 3 O-O-O! 17 c;i(g2 f6 with good kingside pros pects, Saltaev-Poldauf, Plovdiv 1988; (b3) 9 c;i(h1 !? is the latest try: 9 . . . h5 (9 . . . '1i'b6!?) 10 dxc5 i.. xc5 (or 10 . . . g4! 1 1 liJd4 liJdxe5!? 12 i..b 5 (Dembo) ; 1 12
then 12 . . . .i.xc5! 1 3 f4 i.. x d4 14 cxd4 liJd7:j:) 1 1 i..b 5 g4 12 .txc6 bxc6 1 3 liJd4 Dembo-Goczo, Magyarorszag 2002. Dembo analyses 13 . . . liJxe5 14 f4 .i.xd4!? 1 5 'iia 4 .i.b6 16 fxe5 'with at tack on the king' but this doesn't look at all convincing; (b4) 9 a3!? (another recent move) 9 . . . cxd4 10 cxd4 'iib 6 (McDonald sug gests 10 . . . g4 1 1 liJe1 f5!? 12 exf6 liJxf6 1 3 liJc2 i.. d 7, which can fairly be de scribed as 'unclear' after 14 f3 'iVb6) 1 1 liJb3 g4 12 liJe1?! (12 liJg5! liJf8!? threatens ... h6 and pretty much commits White to the unclear piece sacrifice 13 .i.c2! h6 14 lLlxf7 rJ;;xf7 1 5 �xg4 with considerable compensa tion and something like dynamic equality after 1 5 . . . h5. Both this posi tion and earlier alternatives need to be looked at) 12 . . . liJxd4 13 liJxd4 'iYxd4 14 liJc2 �xe5 15 'iixg4 h5 16 'iYh3 lLlc5 1 7 .i.b5+ .i.d 7 and White lacked compensation for the pawn and centre in Hossain-Barsov, Dhaka 2003; (b5) 9 dxc5 and now: (b5 1) 9 . . . liJdxe5 10 i..b 5 (10 liJxe5 lLlxe5 1 1 .i.b5+ i.. d 7 12 'iVe2 i.xb5 1 3 'i'xb5+ 'ifd7 14 'ife2 liJg6=) 10 . . . i.. d 7 11 'iie 2 "ikc7 1 2 lie1 liJg6 13 liJb3 g4 14 liJfd4 e5 with at least equality, Rublevsky-Volkov, Ohrid 200 1 ; (b52) 9 . . . g 4 10 liJd4 liJdxe5 1 1 liJ2b3 (1 1 .i.b5 ! - Dembo; but then 1 1 . . .Jt.xc5 looks equal after, e.g., 12 .l:.e1 .lid7 1 3 .i.xc6 liJxc6 14 �xg4 �f6 1 5 lLl2b3! .i.b6 1 6 .tf4 0-0-0) 1 1 . . .liJxd3 12 �xd3 e5 13 liJxc6 bxc6 14 .i.h6 (14 f4 gxf3 15 'iixf3 f6 16 �5+ 'i2i>d7 - Izoria) 14 . . f6 1 5 f4 gxf3 ( 15 . . . e4!?) 16 'iixf3 .:tg8 1 7 'iWh5+ !:tg6 18 i.e3 �d7! 1 9 h3 (19 'iVxh7 'iig 4 20 g3 .tf5= - Izoria) 19 . . . �d8= Gasanov-Izoria, Baku 2002. 9 lie 1 Other moves are no better: .
Tarrasch Variation: 3. . . �e 7 (a) 9 b3 a4! 10 bxa4 c4 (or 10 . . . 'iVa5 !) 1 1 �c2 'iVa5 12 'Db l (12 �b2 'Db6) 12 . . . h6! 1 3 �a3 'Db6 14 h4?! (14 �xe7 'Dxe7 15 'Da3=) 14 . . . �dn Sulskis Lputian, Las Vegas 200 1 ; (b ) 9 'iVe2 cxd4 10 cxd4 'Db6!? (1O ... g5! ; 10 ... iVb6) 11 .l:i.dl a4 1 2 'Dfl 'Db4 1 3 �b l 'Dc4 14 'De3 b5= Van der Wiel-Korchnoi, Antwerp 1993. 9 ... cxd4 An experiment was 9 . . . a4!? 10 dxc5 'Dxc5 1 1 �c2 f6 12 exf6 �xf6 13 'Dfl 0-0 14 'De3 a3! unclear, Ulibin Shulman, Calcutta 1999. 10 cxd4 'iVb6 10 . . . g5!? is a recent treatment: 1 1 g4?! (White stops . . . g4, but as in sev eral such cases, his kingside becomes vulnerable; better is 1 1 h3! h5 12 'Dfl ! g4 13 hxg4 hxg4 1 4 'D3h2, when 14 . . . 'Dxd4 1 5 'iVxg4 and 14 . . . iVb6 1 5 �e3 are both unclear) 1 1 . . . h 5 12 h3 hxg4 13 hxg4 iVb6 14 'iVa4!? Gor mally-McDonald, London 200 1 ; and now Gormally suggests 14 . . . 'Df8! (14 . . . f6!? McDonald) 1 5 'Dfl �d7; then 16 �e3 (in view of 1 6 �b5? 'Dxe 5! ; and 16 'iVdl 'Dxd4) 16 . . . 'iVxb2 17 .l:i.ab l 'Db4! gives Black an extra pawn and a seeming advantage.
1 1 'iVa4?! This leads to a cute miniature. More critical moves are:
(a) 1 1 a3 g5 ! 12 h3 h5 1 3 'Dfl (13 g4? hxg4 14 hxg4 'Dxd4) 1 3 . . . g4 14 hxg4 hxg4 1 5 'D3h2 'iVxd4 1 6 'Dxg4 'Dc5 (16 . . . �c5!?) 17 �c2 'iVxdl 1 8 .l:i.xdl b6 19 �e3 �b7 (19 . . . �a6 2 0 'Df6+ �d8!? 2 1 .l:i.abl? �xfl ! 2 2 �xfl 'Dxe5+ Nisipeanu-Lputian, Batumi 1999) 20 .l:i.d2 0-0-0 2 1 .l:i.adl .l:i.dg8 22 'Dfh2 d4+ Nouro-Gleizerov, Stock holm 1 998; (b) 11 'iVe2 g5! 12 h3 h5 13 'Dfl 'Dxd4!? (13 . . . g4! 14 hxg4 hxg4 1 5 'D3h2 'iVxd4) 14 'Dxd4 'iVxd4 1 5 'Dh2 g4! 16 hxg4 hxg4 17 'Dxg4 .l:i.g8 18 'Dh6 .l:i.g7= Rublevsky-Lputian, Mon tecatini Terme 2000; (c) 11 'Db l ! (a nice attacking gam bit) 1 1 . . . 'Dxd4 12 'Dxd4 'iVxd4 13 'Dc3 �c5! (an aggressive replacement for the usual 13 . . . iVb6) 14 'Db5 (14 �e3 'iVh4 15 g3 'iVe7 16 'Db5 0-0 17 'iVc2 b6!? 18 �xh7+ �h8 19 i.xc5 'Dxc5 20 �d3 'Dxd3 21 'iVxd3 i.a6, at least equal) 14 . . .'iVxf2+ 15 �hl 0-0 16 �g5 �b4!? 17 .l:i.e3 g6 (or 17 .. .£5) 18 .l:i.f3 'iVxb2 19 �f4 'Dc5 20 �c2 �d7 2 1 'Dd6 �a4 2 2 �xa4 'Dxa4 23 .l:i.b3 'iVf2 24 .l:i.f3 iVb2 YZ - YZ B.Vuckovic-Solak, Belgrade 2000. 1 1...g5 12 'Db3? Remarkably, this practically loses by force. But McDonald's excellent suggestion after 12 h3 h5 13 g4 is 13 . . . 'Df8! 'planning . . . i.d7, . . . 'Dg6, etc.' 12 ...g4 1 3 'Dfd2 'iVc7! 14 'Dfl? But there's really no salvation. McDonald analyses 14 �b l? b5! 1 5 'iVxb5 �a6 16 'iVa4 'Db6; also bad are 14 �fl 'Db6 15 iVb5 a4 16 'Dc5 'Dd7! 1 7 'Dxd7 �xd7- + ; and 14 i.e2 'Db6 1 5 iVb5 �d7!- + . 14 ... 'Db6 1 5 'iVb5 a4 0-1 This was the game Kwiatkowski Rendle, Hastings Challengers 2000. White resigned in view of 16 'Dc5 .l:i.a5. 113
Play the French 7.4 4 i.d3 c5 5 dxc5
This is currently the main line of the 3 . . . .te7 variation at the top levels. White hopes the e7 bishop will move twice while he develops smoothly via \!ie2, ttJgf3, 0-0 and very often e5. Black does has the advantage of two centre pawns versus one, and it is often true that the bishop on d3 can be subject to tempo-winning attacks which allow Black to catch up in de velopment. The main lines see a lot of dynamic pawn sacrifices for Black, but he also has calmer solutions. 5 ... ttJf6 5 . . . ttJd7 is increasingly played and would be an ideal answer to this sys tem, but I don't believe in the theo retical solutions to the rare sequence 6 exd5 (or 6 b4) 6 . . . exd5 7 b4. One highly tactical line is 7 . . . a5 8 a3! axb4 9 .i.b2 ttJxc5 10 .txg7 i.. f6 1 1 iLb5+! ..td7 12 i.. x d7+ ttJxd7 13 iLxh8! (que ried by commentators) 13 . . . ..txh8 1 4 l:ta2 l::tx a3 and now I think that White can control the a-pawn after 1 5 .l:!.xa3! bxa3 16 ttJe2 and 16 ... ttJc5 1 7 0-0 a2 1 8 c3 or 16 ... a2 1 7 c3 'iVa5 1 8 0-0. If the reader can answer the b4 idea, then I recommend 5 . . . ttJd7 as a good alternative. 6 'iVe2 This is definitely the most danger114
ous move for Black. The alternative tries seem rather timid in compari son: (a) 6 exd5 'i!Vxd5 (6 . . . exd5 is also equal, and is treated under the order 4 ttJgf3 ttJf6 5 .id3 c5 6 exd5 exd5) 7 ttJgf3 ttJbd7! (to put the knight on c5 rather the bishop or queen) 8 ttJb3 (8 0-0 ttJxc5 9 .ic4 'i!Vd6 10 'i!Ve2 0-0 1 1 ttJe5 b 6 1 2 ttJdf3 ttJd5 1 3 l:tdl .ib7= Starostits-Drasko, Cutro 2002; 8 b4 a5 9 i.. c4 \!ih5 10 c3 ttJd5 11 iLxd5 'ilfxd5 12 'iVb3 i.f6 13 .l::i.b l ttJe5 14 0-0 ttJd3=l= Kholmov-Morozevich, Perm 1998) 8 . . . ttJxc5 9 ttJxc5 'iVxc5 10 0-0 0-0 11 ..te3 'ilc7 12 i.. d 4 b6 1 3 "ii'e2 iLb7 Bellia-Drasko, Italy 1999. This kind of position with 4 kingside pawns to 3 and a queenside minority tends to be favourable for Black unless White has a nice lead in development; (b) 6 ttJgf3 allows Black to simplify by 6 . . . dxe4! 7 ttJxe4 ttJxe4 8 i.. x e4 'iYxdl + 9 'it>xdl i..xc5 10 'iit e 2 ttJd7= ; again Black has the 4-3, 2-3 advan tage, although White doesn't have any problems after the reply 1 1 i.. e 3! (alternatively 11 .l:!.dl ttJf6 12 .ig5 i.. d 7! 13 i.. xf6 gxf6) 1 1 . . .i.. x e3 12 �xe3 ttJc5 13 ttJd2 (13 l:tadl?! ttJxe4 14 �xe4 b6 15 'iit e 3 ..tb7 16 c4 'iit e 7 1 7 l:td4 l:tad8 1 8 :hdl l::txd4 1 9 l:txd4 ':'c8=l= Furhoff-Djurhuus, Sweden 2002) , for example, 13 . . . ltJxe4 14 ttJxe4 �e7 (14 ... ..td7!?) 1 5 :hdl b6 16 l::t d 2 .tb7 1 7 l1adl :hd8 1 8 ':xd8 .l:!.xd8 19 .l:!.xd8 'iitxd8 20 ltJd6 i..xg2 2 1 ltJxf7+ �e7 22 ltJe5; (c) 6 b4?! doesn't work this time due to 6 . . . a5 7 c3 dxe4 8 ttJxe4 ttJxe4 9 .txe4 "ii'x dl + 10 'it>xdl axb4 1 1 cxb4 1 1 . . .f5 !? (or 1 1 . . . .tf6 12 .l:.b l litxa2) 12 i.. c 2 (12 i.. f3 .tf6 13 llb l l::t xa2 14 i.. d2 0-0 with an early . . . e5) 12 . . . ttJc6 1 3 i.. d 2 i.. f6 1 4 .l:!.b l ltJd4=l= .
Tarrasch Variation: 3 i.e 7 . . .
6 'iVe2 initiates one of the most hotly debated lines in the French De fence. I will examine one relatively calm solution and one that is well known but on the edge: 7.41 6 ... 0-0 7.42 6 liJc6
A position that is surprisingly easy for Black: (h I) 9 c3 a4 10 liJd4 liJxc5 1 1 Si.c2 f5 12 exf6 i.xf6 1 3 f4 (13 liJgf3 e5!) 1 3 ... b6!? 14 liJgf3 (14 'iVe3!? e5! 15 fxe5 i.xe5 16 liJgf3 i.f4:j:) 14 ... i.a6 15 'iVe3 'iVd7:j: , e.g., 16 i.d2 liJc6 1 7 liJxc6 "xc6 1 8 liJd4 iLxd4 19 'iVxd4 i.c4!+; (h2) 9 a4 liJa6!? 10 iLxa6 lha6 1 1 liJf3 liJxc5 1 2 liJxc5 iLxc5 1 3 0-0 �a8 1 4 c4 b6:j:; (h3) 9 �5?! g6 10 \\!Ve2 liJc6 and White is in trouble due to problems with the e-pawn; (h31) 1 1 i.h6 l:1e8 12 liJf3 a4 1 3 liJbd4?! ( 1 3 liJbd2 a 3 14 b3 liJdxe5!) 1 3 ... liJdxe5! 14 liJxe5 liJxd4 etc.; (h32) 11 liJf3 a4 1 2 liJbd2 (12 liJbd4 liJdxe5) 12 . . . 'iVc7!:j: or 12 . . . a3!?:j: . 7 ... a5!?
•••
7.4 1 6 ... 0-0 I recommend the following lesser known sequence as a first system and a sound way to avoid having to keep up with the latest nuances of theory. 7 liJgf3 Others are easier to meet: (a) 7 b4?! a5 8 c3 axb4 9 cxb4 liJc6 10 .l:!.b l Itxa2+ (or 1O . . . b6! 1 1 cxb6 liJxb4+); (h) 7 e5 liJfd7 8 liJb3 a5
I think that this is the most effec tive of the move orders by which Black advances the a-pawn and pre vents b4 in order to capture on c5 with a knight. . .. a5 also anticipates White's move liJb3 (to which . . . a4 will be the reply) and opens up the two possibilities of . . . liJa6 (hitting c5) or . . . b6 and . . . i.a6. The negative side of 7 . . . a5 is of course that it doesn't de velop a piece, but these flank pawn moves are increasingly acceptable in modern chess. In a similar vein, there 115
Play the French are two little-known plans that de serve attention: (a) 7 . . . tDc6 8 0-0 (8 exd5!? exd5 9 tDb3 a5! 10 a4 l:Ie8 1 1 .i.e3 d4 12 0-0-0 dxe3 1 3 J..xh7+ �xh7 1 4 lhd8 i.xd8 15 fxe3 �g8! unclear) 8 . . . tDd7!? (the point; compare our main line) 9 exd5 (9 tDb3 tDxc5 10 tDxc5 J..xc5) 9 . . . exd5 10 tDb3 tDxc5 (10 . . . a5!?) 1 1 tDxc5 J..xc5; (b) 7 . . . tDfd7 8 tDb3 (8 b4 a5 9 c3 axb4 10 cxb4 tDc6; 8 exd5 tDxc5!? 9 dxe6 tDxd3+ 10 'iVxd3 J..x e6 1 1 'ili'xd8 llxd8 12 0-0 tDc6 with compensation in the form of activity and the two bishops) 8 . . . a5 (or 8 . . . tDxc5 9 tDxc5 �xc5 1 0 0-0 tDc6) 9 exd5 exd5 1 0 .i.e3 a4 11 tDbd4 tDxc5 12 0-0 a3 13 b3 tDxd3 14 'i'xd3 tDc6 with a reasonable position. S 0-0 The most natural move but as usual there are serious alternatives here: (a) 8 a4 tDa6 9 e5 tDd7 10 tDb3 (10 0-0 tDdxc5 11 J..b 5 tDc7= ; 1 0 i.xa6 :'xa6 gives up the powerful d3 bishop and should not be dangerous: 1 1 tDb3 tDxc5 12 tDxc5 �xc5 13 �e3 b6=) 1 O . . . tDaxc5 11 tDxc5 tDxc5 12 0-0 tDxd3 (or 1 2 ... i.d7 1 3 J..b 5 J..xb5 14 axb5 'i'c7 15 J..f4 tDe4) 1 3 'i'xd3 b6 14 l:!.e1 �a6 15 'i'd1 'uc8; P. Carlsson Kruppa, Halkidiki 2002; (b) 8 c3 tDfd7 9 exd5 tDxc5! 10 J..b 5!? exd5 11 0-0 tDc6 (or 1 1 . . .tDba6 12 :tel J..f6 13 tDd4 tDc7 14 tD2f3 .i.g4!? 15 h3 J..h 5 16 J.. e 3 .i.xd4!? 1 7 �xd4 tD5e6= - Radjabov) 12 tDd4 Almasi-Radjabov, Pamplona 2001 ; and although 1 2 . . . i.d7 equalised with time, Radjabov suggests 12 . . . tDxd4! 13 cxd4 tDe6 14 tDf3 tDc7 (I like 14 . . . i.f6! 1 5 i.e3 tDc7 1 6 i.d3 i.g4=) 15 .td3 i.g4 16 'i'e3! .td6 17 tDe5 J.. e 6 unclear; 116
(c) 8 e 5 tDfd7 9 tDb3 a4 1 0 tDbd4 tDxc5= . After the bishop on d3 falls, Black has no problems. S tDa6 This seems to equalise rather eas ily. Black wants to play . . .tDxc5 and capture the bishop on d3, but he leaves d 7 free for the other knight in case of e5. In fact, it can go there im mediately: 8 . . . tDfd7 has been played several times with respectable re sults, e.g., 9 exd5 exd5 10 :'e 1 (10 tDb3 a4 11 tDbd4 tDxc5 12 i.f5 tDc6= Grabarska-Mkrtchian, Warsaw 2001) 10 . . . tDc6 11 i.b5 (1 1 tDb3 a4 12 tDbd4 tDxd4 13 tDxd4 .i.xc5 14 J.. e 3 l::1. e 8 1 5 c3 tD e 5 16 i. b 5 i. d 7 1 7 :ad1 tDg4=) 1 1 . . .i.f6 12 c3 tDxc5 13 tDb3 tDe6!? (13 . . . tDxb3 14 axb3 d4) 14 i.e3 a4 15 tDbd4 Lakos-H.Richards, Bled 2002; and 15 . . . tDexd4 16 tDxd4 tDxd4 17 .txd4 i.xd4 18 cxd4 'i'b6 was equal. ...
g exd5 9 .i.xa6?! :'xa6 doesn't help White (see the note to White's 10th) ; so the main alternative is 9 e5 tDd7 (Black gets the move he's been wanting: . . . tDxc5) 10 c3 (10 tDd4 tDdxc5 1 1 f4 'ili'b6 12 tD2f3 i.d7; Kotronias-Barsov, Montreal 2002) 10 . . . tDaxc5 1 1 .tc2 b6 12 .l:.e1 Jens-Papa, Deizisau 2003; and a change of scene by 12 .. .f6! seems indicated: (12 ... .ta6 1 3 'i'e3
Tarrasch Variation: 3. . ..te 7 .
f6!?) 1 3 exf6 .i.xf6=, due to Black's mobile centre and f-file. 9 ... exd5
10 ...l::t e 8! 1 1 �b3 �xc5 A good alternative is 1 1 . . .a4! 12 �bd4 �xc5 13 ..tb5 ..td7 14 ..txd7 (or 1 4 ..te3=) 14 . . :ii'x d7 1 5 �e5?! 'fic8 and Black threatens . . . .i.d6 with good play. After 1 1 . . .�c5, play might go 12 ..tg5 ..td7 13 �xc5 .i.xc5 14 'ii'd 2 'fib6= (another typical isolated pawn position in which Black has nice ac tivity), when 15 ..txf6 'iVxf6 16 c3 ..tg4= is a natural continuation. 7.42 6 ... �c6 7 �gf3 �b4!?
This is sound and well tested. 9 .. .'iWxd5 has been played a lot, but I'm not happy with the response 10 �c4! .i.xc5 1 1 c3 of P.Carlsson Kjartansson, Hallsberg 2003. 10 l:t.e1! This untried move looks best, as it tries to either stop Black from captur ing with the knight on c5 or to lure a rook to e8 where the bishop can at tack it. (a) 10 �b3 a4 (1O . . . �xc5 1 1 �xc5 ..txc5=) 1 1 �bd4 �xc5 12 .i.b5 �fe4 (or 12 . . . ..td7) 1 3 a3 .i.g4 14 h3 ..th5 (14 . . . ..txf3 15 'fixf3 'fib6!) 15 �f5 ..tf6 16 I:tdl l:ta5 and Black was doing quite well in Womacka-Djurhuus, Gausdal 2002; (b) 10 ..tb5 �xc5 l 1 l:tdl 'fic7= ; (c) 10 ..txa6 llxa6 leaves Black the two bishops and develops his rook actively, all for the sake of preventing a knight capture on c5. Black is probably already better, as in 1 1 �d4 l::t e 8 12 'fib5 'fic7 13 �2f3 .i.d7!? (or 1 3 . . . ..txc5=i=; Black is active and better developed) 14 c6! ..txc6 15 �xc6 Lenic- Erdos, Pula 2003; and even now 15 . . :i!Vxc6! 16 'fixc6 bxc6 is better for Black.
For a second system I offer this of ten wild and tactical line championed by Morozevich. It has survived scru tiny for the last decade, so the likeli hood of a refutation is remote. How ever, I do feel that a couple of lines are slightly but stably better for White. In them, White is a pawn up but Black has the bishop pair versus either the knight pair or a knight and bishop . One can argue that the bishop pair offers full compensation, and strong players including Moro zevich are still playing 7 . . . �b4 with out fear; nevertheless I don't feel that this system offers quite the positive chances that 6 . . . 0-0 and . . . a5 does, so I will try to give a helpful repertoire without providing total detail. 117
Play the French Interestingly, 7 . . . a5!? is also play able here and unrefuted. Briefly, the main line runs 8 0-0 (8 e5 ttJd7 9 0-0 ttJxc5 when the knight has gotten to c5 with no resistance) 8 . . . 0-0 9 c4 a4! 10 IIdl (10 cxd5 exd5 1 1 exd5 'i!Vxd5 frees Black's c8 bishop) 10 . . . �xc5 1 1 exd5 exd5 1 2 cxd5 ttJb4! 1 3 ttJe4 ttJxd3 14 ttJxf6+ (14 l:.xd3 ttJxe4 15 'i!Vxe4 .l:i.e8 16 'ilVf4 'ilVb6 when White had to give back the pawn by 1 7 �e3 with an unclear position which looked at least equal for Black after 1 7 . . :iVxb2 1 8 l:.el 'iVb4 19 'i!Vxb4 i.xb4 20 l:.cl iLf5 Marinkovic-Drasko, Vrnjacka Banja 1999) 14 . . . 'ilVxf6 15 'i!Vxd3 iLf5 and the bishops offer nice compensa tion: 16 'ilVb5 b6 17 iLg5 (17 iLe3 l1a5 18 'ilVe2 iLxe3 19 fxe3 a3 with near equality and a quick draw, Rublev 2003) sky-Pelletier, Poikovsky 1 7 . . :iVg6 1 8 i.e3 i.xe3 1 9 ttJh4 'ilVf6 20 ttJxf5 .i.xf2+ 21 �xf2 "iYxf5+ with ap proximate equality (soon drawn), As rian-Lputian, Yerevan 2000. 8 ttJb3 Not necessarily best. Here are some options: (a) 8 0-0 has become very popular and often yields a small positional edge.
As usual, Black must play sharply: (al) 8 . . . ttJxd3 9 cxd3 iLxc5 10 ttJb3 118
i.e7 1 1 i.g5 (1 1 e 5 ttJd7 1 2 ttJfd4 0-0 1 3 i.f4 ttJc5 14 "iYg4 �h8 1 5 ttJxc5 iLxc5 16 ttJf3 i.e7 1 7 l:.fel i.d7 18 'iVh5 f6!= Svidler-Radjabov, Moscow 2002) 1 1 . . . h6 12 i.h4 i.d7 13 e5 (13 .i.xf6 .i.xf6 14 exd5 i.b5! :j: ; 1 3 .l:i.acl dxe4! 1 4 dxe4 "iVb6 1 5 lifdl i.a4 16 'i!Vc4 i.xb3 1 7 axb3 0-0= Ponomariov Morozevich, Istanbul 2000) 13 . . . ttJg8 14 �xe7 (14 i.g3! h5 15 .tf4 ttJh6 16 ttJfd4!) 14 ... ttJxe7 1 5 ttJc5 .i.c6 1 6 ttJd4 'i!Yb6 17 .l::tfcl 0-0 18 a3 a5 19 l:.c3 .l:.fc8 20 .l:i.ac l .te8 2 1 'ilVd2 l:.c7 22 .l:i.3c2 a4 was only very mildly better for White in Luther-Duppel, Boblinger 2000; (a2) 8 ... 0-0 9 e5 (9 ttJb3 ! a5 10 a4 b6 11 e5 ttJd7 12 c6 ttJxc6;t Sadykov Bhat, Oropesa del Mar 2001) 9 . . . ttJd7 10 ttJb3 a5 1 1 a4 ttJxd3 12 cxd3 ttJxc5 13 ttJxc5 i.xc5 14 d4 (14 i.e3 d4 1 5 �f4 iVd5:j: Mrva-Kostenko, Istanbul 2000) 14 . . . i.e7 15 i.d2 b6= Heim L.Johannessen, Bergen 200 1 ; (b) 8 e5 ttJd7 9 ttJb3 ttJxd3+ 10 iVxd3 !? (10 cxd3 ttJxc5 11 ttJxc5 i.xc5 12 d4 i.e7=) 1O . . . a5! 1 1 �e3 (1 1 a4 b6! 12 cxb6 i.a6 13 'ilVe3 [13 'ii'd l 'i!Vxb6] 1 3 . . . 0-0 14 ttJxa5 .i.c5 McDonald; White is probably some what better after 15 ttJd4 'ilVxb6 16 ttJab3 i.e7, although it's hard t o fmd a positive plan) 1 1 . . .a4 12 ttJbd2 0-0 1 3 0-0 ttJxc5 14 'i!Vd4 .i.d7 1 5 "iVg4 'it>h8 16 .l::t a dl .l::t c 8 17 c4 f5 ! Zagrebelny Morozevich, Moscow 2003; Black is very active and stands better. Note that Morozevich is still willing to de fend his system as this book is being written; (c) 8 exd5 'ii'x d5 9 ttJb3 i.d7 10 0-0 (10 .tc4 'i!Ve4= ; 10 ..te3! ttJxd3+ 1 1 'i!Vxd3 'ilVxd3 1 2 cxd3 iLb5 - Golosh chapov; a touch better for White, but not much) 10 . . . ttJxd3 1 1 'ilVxd3 iVxd3 12 cxd3 �b5 13 .l::t d l .i.a4! 14 d4 ttJd5 !? with a blockade, Ganguly-Gol-
Tarrasch Variation: 3. . i.e 7 .
oshchapov, Sangli 2000; (d) 8 i.b5+?! i.d7 9 O-O!? tiJxc2 10 i.xd7+ tiJxd7! (McDonald) gives Black a substantial advantage. 8 tiJxd3+ A number of well-known world class struggles stemmed from this move. Also seen is 8 . . . a5!?, but 9 �b5+! (9 i.g5 a4 10 tiJbd2 a3!? 1 1 b3 0-0 12 0-0 h6, when again 1 3 i.e3;!; may be best) 9 . . . i.d7 10 i.xd7+ tiJxd7 Kanovsky-Jurek, Roznov 2002; and I think that 1 1 c3!? with the idea 1 1 . . .tiJa6 12 exd5 exd5 13 c6! bxc6 1 4 lllb d4! and tiJf5 i s very promising. 9 cxd3 a5 ...
10 �g5 Objectively 10 a4 seems to yield some advantage but practice is a dif ferent matter: 1O . . . b6!? 1 1 e5 tiJd7 12 c6 tiJc5 1 3 tiJbd4 i.. a 6 1 4 tiJb5 (14 I1a3 tiJe4!=) 14 . . . tiJb3 (14 . . . d4 15 tiJfxd4! .i.xb5 16 tiJxb5 tiJxd3+ 17 'it>fl fa voured White in Barua-Ravi, New Delhi 200 1) 1 5 l:tb l .i.b4+ 1 6 'it>f1 0-0 17 d4 f6! 18 'ilVe3?! (18 'ilVd3! seems to improve, e.g., 1 8 . . . i..xb5 19 axb5 a4 20 g3, and if 20 . . . tiJxcl?! 2 1 l:i.xcl 'ilVe8 22 exf6 gxf6 23 �g2±) 18 . . . �xb5+! 19 axb5 a4 20 g3 �e8! 2 1 'iVd3 (a crucial tempo down on the note to White's 18th) 2 1 . . .tiJxc l 22 ':'xc l a3 23 b3 (23 bxa3 I1xa3 24 'ilVe2 \lVh5) 23 . . . a2 24
�g2 �g6! 25 �xg6 hxg6 and Black had the better chances (again White lacks a plan) in Adams-Morozevich, Wijk aan Zee 2000. 1 0 a4!? 10 ... 0-0 is recommended by McDon ald: 1 1 0-0 (1 1 e5?! tiJd7 12 i.. x e7 'ilVxe7 13 d4 b6! 14 c6 .i.a6 15 �d2 4:Jb8 16 l:.cl Itc8+) 1 1 . . . b6! 12 4:Je5 (12 e5) 12 ... dxe4 'with messy play, but Black looks fine', according to McDonald. Still, 1 3 tiJc6 exd3 14 �e3 'ilVc7 15 tiJxe7+ �xe7 16 cxb6 probably favours White somewhat. The b-pawn is dangerous and d3 is vulnerable. 1 1 tiJbd2 h6 1 1 . . . dxe4 12 dxe4 tiJd7 13 i.xe7 'ilVxe7 14 J:r.c l 0-0 1 5 'ilVe3! I1a5 16 0-0 .l:!.xc5 1 7 'ilVa3 b5 1 8 tiJe5 ! .l::!. e 8 19 .l:!.xc5 �xc5 20 'ilVxc5 tiJxc5 21 .l:.c1 with an irritating pull, Almasi-Timman, Pamplona 1 999. ...
12 i.xf6!? 12 i.e3! 'ilVa5 1 3 0-0 i.xc5 14 i..x c5 'ilVxc5 15 I1fcl 'iVb6 16 e5 tiJg8 17 'iVe3 'ilVxe3 18 fxe3 appeared to favour White in Navara-Duppel, Pardubice 2000; an interesting sequence. 12 i..xf6 13 e5 Black also gets counterplay after 1 3 exd5 a3! (13 . . . 0-0 14 dxe6 �xe6 1 5 tiJe4 a 3 1 6 tiJxf6+ 'ilVxf6 1 7 tiJe5;!;) 1 4 d 4 ( 1 4 d 6 axb2 1 5 l:i.b l b 6 ! 16 d 4 bxc5 ...
119
Play the French 17 dxc5 ':xa2 - Psakhis) 14 . . . axb2 1 5 Vib5+ 'ii'd 7 16 'iYxb2 'ilixd5 1 7 'i'b3 'with the superior chances' - Psakhis; but 17 . . . 'iYxb3 (or 17 . . . 0-0 18 0-0 .td7 1 9 'iYxd5 exd5 20 .l:.fb l .l:.a7) 1 8 liJxb3 i.d7 (or 18 ... .l:.a4) 1 9 0-0 0-0 20 liJe5 i.b5 2 1 �fb l .l:.a4 and the two bishops fully compensate for the sacrificed pawn. 13 ... ..te7 14 'ile3 'ila5 15 ':cl bS!? 1 5 . . . ..td7 16 0-0 0-0 1 7 .l:1.c2 l::tfc8 18 .l:.fcl l:tc7 1 9 d4 b6!? Pogonina Matveeva, Elista 2002; yet another of those funny pawn-down two bishops lines if White plays 20 cxb6 llxc2 2 1 l:1xc2 'ilxb6; maybe it's enough, but
1 20
I'm not sure. IS cxbS?! 1 6 c6 must be better; then 16 . . . 'i'b5 1 7 ':c2 i.a6 18 'iii' e 2 has been sug gested. IS ... .td8! 17 'ilic5 'i!fxc5 18 ':xc5 i.xbS 1 9 l:ic2 0-0 20 liJf1 ..taS with strong bishops and plenty of pressure, Godena-Morozevich, Istan bul 2000. The 7 . . . liJb4 systems are still scoring well at the top levels, which says a lot; you should defi nitely consider them if you're a fan of the bishop pair. I prefer the 6 . . . 0-017 . . . a5 plan and I like the fact that it's still relatively unexplored.
Chapter Eight Winawer Variation : Fourth M ove Alternatives
1 e4 e6 2 d4 d5 3 ttJc3 iLb4 3 . . . .tb4 is the Winawer Variation, one of the most fruitful and exciting openings in chess. Black puts pres sure on e4 and prepares to double White's c-pawns with consequent po sitional pressure. It is one of the premier chess variations because it embraces so many types of closed manoeuvring positions and wide-open tactical ones, often both within the same game. For this edition, I have added an extra repertoire for the move 3 . . . ttJf6, hopefully an interesting one. But it's hard to match 3 . . . ..tb4 for complexity and excitement. After 3 . . . �b4, White's main move is 4 e5. Then White has to suffer through a counterattack on his centre and in most cases doubled pawns on c2 and c3. His rewards for this sacrifice are great in terms of both attacking chances and opportunities to domi nate his opponent with his bishops, so 4 e5 continues to be easily the most popular choice. But some players pre fer to avoid this kind of commitment and choose one of the alternatives in this chapter instead. Theory on and practice with these variations have
exploded and the well-prepared French player must know something about each:
8.1 8.2 8.3 8.4 8.5 8.6 8.7
4 a3 4 'iVg4 4 i.. d 2 4 ttJe2 4 ..td3 4 exd5 4 'ii'd 3
8. 1 4 a3 �xc3+ 4 . . . i.. a 5!? probably isn't objectively the best way to deviate, but might appeal to those who want original play and/or don't like 5 . . . ttJe7 of the next note. Here's just a sample of
121
Play the French what might happen: 5 b4 i.b6 6 e5 tDe7 (6 . . . a5?! 7 b5 c5? 8 tDa4! cxd4 9 �g4! �f8 10 tDxb6 �xb6 1 1 a4! ± Mortensen-Fant, Copenhagen 2001) 7 tDa4 tDf5 (7 . . . 0-0 is more flexible, e.g., ... tDd7 and . . .f6 or ... f5 might follow) 8 tDf3 0-0 9 J.g5 f6 10 exf6 gxf6 1 1 tDxb6 axb6 1 2 i.. f4 tDc6 1 3 b 5 (13 i.. d 3 e5!? 1 4 dxe5 fxe5 15 tDxe5 tDh4!? 16 �h5 .l:i.xf4 17 �xh7+ �f8 18 'iWh6+ �e8 19 tDxc6 tDxg2+ 20 cJifl �4 with a real mess) 13 . . . tDce7 intending . . . tDg6 and perhaps . . . e5. Probably White gets an edge somewhere in this line with perfect play (the two bish ops must be worth something), but that advantage will be small and doesn't negate the value of this idea for achieving a playable imbalance. 5 bxc3
Used by Alekhine, Smyslov, and Fischer, 4 a3 intends to prove that White's doubled pawns are not as important as the dark square vulner ability of Black's kingside. It has never become very popular in top level play, however (even the above mentioned players mostly employed other anti-Winawer systems), be cause White loses a tempo and Black develops freely. Nevertheless, it is extremely difficult to play against if one is unprepared, and to play the 122
main lines well requires as much study as any variation in this chap ter. 5 . .,dxe4 If one wants to avoid the prepara tion of a specialist in 4 a3, 5 . . . tDe7 is a good option, especially if you're will ing to transpose:
Then 6 e5 c5 is the main line of the Winawer, and 6 'ilVg4 0-0 (with the idea . . . e5) 7 e5 can be answered by 7 . . . c5, which is another main line (7 'i'g4 0-0 of Chapter 10); but Black also has some original alternatives, e.g. 7 . . . tDd7 with the idea .. .f5, secur ing the kingside before attacking via . . . c5. Here are some other tries for White: (a) 6 .i.d3 c5 7 exd5 (7 dxc5 tDd7= , or 7 . . . dxe4 and . . . 'i'xdl +; 7 'i'g4 0-0 8 e5 is a main line, and instead, both 7 . . . c4 and 7 . . . tDbc6!? are interesting; finally, after 7 tDf3 c4 8 .i.e2, Black has 8 . . . 'ilVa5!?= or 8 . . . dxe4! 9 tDg5 tDd5 with the idea 10 i.xc4 'i'c7 or 10 . . . e3, or here 10 tDxe4 f5 1 1 i.g5 'ilVc7 12 i.h5+ g6 13 tDf6+ tDxf6 14 i.xf6 0-0 15 i.e5 'ilVa5 with good play Wintzer) 7 . . . exd5 8 dxc5 'iVa5 (or 8 . . . tDd7 9 i.e3 'ilVc7 intending 10 'ilVg4 tDxc5 1 1 'i'xg7 tDxd3+ 12 cxd3 :g8=) 9 tDe2 tDd7 (a sequence that has oc curred in several games) 10 ':'b l
Winawer Variation: Fourth Move Alternatives ttJxc5!? (10 . . . a6! 1 1 0-0 ttJxc5 12 ttJd4 0-0) 1 1 i.b5+ i.d7 12 i..x d7+ ttJxd7 1 3 l:txb7 ttJc5 14 l:tb4 0 - 0 (White's extra pawn is balanced by his static weak nesses) 15 0-0 l:tac8!? (15 . . . l:tfe8) 1 6 l:th4 (16 c4 l:tfd8) 16 . . . ttJe4= Sariego Arencibia, Bayamo 1990; (b) 6 exd5 exd5 7 ttJf3 (7 i.. d 3 0-0 8 ttJe2 i.f5=) 7 . . . i.g4 8 i.. e 2 0-0 9 0-0 ttJg6 1 0 a4 ttJc6 11 h3 i.f5 12 i.a3 l:te8 13 �d2 ttJa5 and White struggled with light square weaknesses, O.Pedersen-Tonning, Gausdal 1997; (c) 6 ttJf3 dxe4 7 ttJg5 c5!? (7 ... 0-0 8 ttJxe4 e5! 9 i.. g 5?! exd4 10 cxd4 �d5 !) 8 ttJxe4 cxd4 9 cxd4 ttJbc6 10 c3 e5 1 1 i.. c4 i..f5!? (1 1 . . .0-0 1 2 i.g5?! exd4 1 3 cxd4 b5! 1 4 i.. a 2 i.. f5) 1 2 ttJg3 (12 ttJg5 0-0=1=) 12 . . . exd4 13 0-0 0-0 14 i.. g 5 Ba nas-Herzog, Keszthely 1981; when the light squares would again give Black the advantage after the con tinuation 14 . . . i.e6! 15 i.xe6 fxe6 1 6 cxd4 �d5. Mter 5 . . . dxe4, we will look at the main traditional move and a popular alternative: 8. 1 1 6 f3 8. 12 6 �g4 8.1 1 6 f3 The Winckelmann-Riemer Gambit, which has grown considerably in popularity since the second edition. The idea is to get something resem bling the Blackmar-Diemer Gambit (1 d4 d5 2 e4 dxe4 3 ttJc3 ttJf6 4 f3) with Black's dark-squared bishop no longer available for defence. In the last edition I recommended 6 . . . b6, which is reasonable but not very inci sive. Since then I have successfully tested one lesser-known idea, and one solution that has been generally rec ommended also looks good:
8. 1 1 1 6 .. ttJd7 8. 1 1 2 6 .. e5 .
.
8. 1 1 1 6 ...ttJd7 This is a very logical move because it gets out a piece, prepares central breaks ( . . . c5 and . . . e5), and doesn't commit to a fIxed scheme of develop ment. There is little theory about 6 . . . ttJd7, and I like it very much.
7 ttJh3 I think that this is best because White has few other choices that get a piece out effectively. Nevertheless, a knight on h3 is not ideally placed for a Blackmar-Diemer sort of posi tion (it belongs on f3 watching over e5). (a) 7 fxe4? �4+ forces White's king to move; (b) 7 i.e3 ttJgf6 (or 7 . . . exf3, since the bishop shouldn't be on e3 in an attacking system) 8 fxe4 ttJxe4 9 �g4 ttJdf6 10 �xg7 l:tg8 1 1 �6 l:tg6 12 �f4 ttJxc3+; (c) 7 a4 ttJgf6 8 i.a3 c5!? (8 ... e3 !?=I=) 9 dxc5 ttJd5! 10 �d4 �4+ 1 1 g3 �f6 12 fxe4 ttJxc3 and Black threatens . . . ttJxe4 while White's pawns are a mess. 7 .. ttJgf6 Also good looks 7 . . . exf3 8 �xf3 �h4+!? 9 g3 �f6, preventing �g3 and .
1 23
Play the French asking the queen where it's going, e.g., 10 'i'e2 tDe7 1 1 ..tg5 YWf5 12 ..tg2 tDf6 1 3 l:tf1 'i'a5;. 8 fxe4 tDxe4 9 'iWg4 tDdf6 1 0 'ilkh4 Black's developmental advantage gets even bigger after 10 "fixg7 .J:lg8 11 'i'Vh6 tDxc3 (or simply 1 1 . . .c5 with the idea . . . 'iVa5 : 12 .i.b5+ �e7!? 1 3 0-0 'iWa5 1 4 ..td3 YWxc3 1 5 .i.xe4 tDxe4 16 'iVxh7 'iVxd4+ 1 7 �h1 tDf6) 12 .i.g5 lig6 1 3 'i'Vh4 tDce4 14 .txf6 tDxf6+ . 1 0 ...tDxc3 Slower but still effective enough would be 10 . . . c6!? 1 1 .i.d3 'i'a5 1 2 0-0 "fixc3 1 3 1Ib l (13 ..te3 c5!) 13 ... tDd2 14 ..txd2 "fixd2 1 5 'iVg3 'i¥h6+ . 1 1 ..td3 tDfe4 12 "fif4 f5 1 3 ..te3 0-0 14 0-0 "fid6 15 'iVh4 ..td7 and Black is two pawns up, Tahiri-J.Watson, Berlin 1997. 8. 1 12 6 ... e5 This is the solution recommended by just about everyone. It frees the c8 bishop and therefore compels White to try to gain compensation quickly before Black develops.
(a) 7 ..tb5+ c6 8 .i.c4 'iWa5+ (peder sen), since 9 .i.d2 e3! wins; (b) 7 a4 exd4 (7 . . . ..te6 has been played several times and must be good as well) 8 cxd4 c5 puts the ques tion to White's centre: 9 dxc5 (9 d5 tDf6 1 0 ..tc4 0-0) 9 . . . 'iYxd1 + 10 <3o>xd1 .tf5 (or 10 . . . tDf6 11 ..tb2 tDbd7 12 ..tb5 0-0 13 c6 bxc6 14 ..txc6 lib8;) 11 g4 .i.e6 12 .i.b2 f6 ! + ; White's pawns are extremely weak. 7 tDc6 Or 7 . . . .tf5; or 7 . . . exd4 8 cxd4?! (8 'iYxd4! 'iVxd4!? 9 .i.xd4 f6 10 fxe4 tDc6;) 8 . . . tDh6! 9 fxe4 (9 .i.xh6 'ith4+ 10 g3 'iVxh6+) 9 . . . 'iYh4+ 10 �d2 'i'xe4 1 1 tDf3 tDf5 12 ..tb5+ tDc6 13 lIe1 0-0 with a decisive advantage, Gra barczyk-Gdanski, Lubniewice 1993. 8 ..tb5 Or 8 d5 tDb8 9 .i.c4 tDf6. After 8 ..tb5, Oparaugo-Cech, Passau 1997 went 8 ... .td7 9 d5 tDce7 10 .i.xd7+ "fixd7+ (White hasn't developed and his pawn structure is damaged) 1 1 fxe4 tDf6 1 2 'ii'd 3 'iYg4! 1 3 'iYb5+ tDd7! 1 4 �f2 YWxe4 15 c4 'i'f5+ 16 <3o>g3 'iWxc2 1 7 ':f1 tDf5+ 18 l:ixf5 'i'xf5 19 'iWxb7 0-0 20 tDf3 l:r.ab8 2 1 'i'xa7 l:ib2 22 'iVxc7 tDf6 0- 1 . •.•
8 . 1 2 6 "fig4 tDf6 7 "fIxg7 :g8 8 'iVh6 tDbd7
7 ..te3 Black has a 200-point performance rating advantage after 6 . . . e5, mainly because White has little to do. Again 7 fxe4? 'i¥h4+ is undesirable, and these alternatives don't help much: 124
Winawer Variation: Fourth Move Alternatives I think that this solution is the most fun. It takes some risks, espe cially in the main line below, so I will present some options along the way. 8 . . . c5 is an established alternative that I gave in the previous edition, but I'll forego that and discuss similar lines with . . . c5 below. After 8 . . . liJbd7, White has: 8. 1 2 1 9 fa 8.122 9 liJh3 8.123 9 liJe2 The latter move is the current fa vourite. The alternatives are less challenging, but they help to illus trate the basic themes: (a) 9 .ib2 is rarely seen anymore because of 9 . . . liJb6! (or 9 . . . c5 10 0-0-0 cxd4 11 cxd4 and one idea is 1 1 . . .b5! 12 .i.xb5 :b8 1 3 c4 a6 1 4 �xd7+ i.xd7 and White's king is exposed) 10 c4 (10 a4 .td7 11 a5 liJa4 1 2 'iVcl c5 13 liJe2 'iWxa5 14 dxc5 'iWxc5 1 5 .ia3 �e5 1 6 g3 Iic8!+ Quinteros-J.Watson Mexico 1 976) 10 . . . liJa4 1 1 0-0-0 �d7 12 f3 'fie7 13 fxe4 (13 'fie3 Ji.c6 14 fxe4 .txe4!+) 1 3 . . . liJxe4 1 4 'iVe3 (14 'fixh7? 'iWg5+) 14 ... f5 1 5 liJf3 c5! 16 .:tgl 0-0-0 1 7 'iVb3 liJxb2 1 8 'iWxb2 cxd4 19 11xd4 .ic6 20 .l:.xd8+ ':xd8 2 1 .ie2? �c5 0- 1 Foguelman-R.Byrne, Buenos Aires 1 964; (b) 9 h3 b6 (a legitimate but under investigated move is 9 . . . c5, e.g., 10 g4 'iWa5 1 1 .td2 'fia4 12 g5 �c6! 13 .tg2 cxd4 14 cxd4 11g6 1 5 �h4 h6 16 gxf6 ':'xg2:j:) 10 g4 (10 g3 'iWe7 1 1 i.g2 J:r.g6 12 �4 .ia6!? is an idea of Amador Rodriguez; then 13 .ixe4 loses the light squares, e.g., 13 . . . liJxe4 14 'i!Vxe4 0-0-0 1 5 'iWa8+ liJb8 16 'iWxa7 e5+) 1O . . . iLb7 11 .ig2 (1 1 g5 .l:.g6 12 �4 and 12 . . . liJg8!? 1 3 liJe2 h6! 14 .l:.gl ! c5 15 liJf4 l:tg7 16 liJh5 l:tg6= or
12 . . . liJd5!:j:) 1 1 . . .'iVe7!? (a possible im provement is 1 1 . . .c5 12 liJe2 'iWc8!? thinking about both . . . cxd4 and . . . e3; again the light squares are weak af ter 1 3 .ie3 .ia6!) 1 2 g5 'iVf8! 1 3 'iWxf8+ (13 'iWh4 h6!) 13 . . . liJxf8 14 h4 h6 1 5 f4 (15 liJe2 liJ8h7+) 1 5 . . . 0-0-0 1 6 liJe2 liJh5 ! 17 .ie3 f5:j: intending . . . liJg6 and . . . .ia6, Vorotnikov-Uhlmann, Lenin grad 1 984. The sidelines look safer, however; (c) 9 iLg5 �e7!? (9 . . . .lIg6 10 'i!Vh4 c5 1 1 liJe2 'ii'a 5 12 .td2 'i!Va4 gave Black excellent play in Trnovec-Gyarmati, Balatonbereny 1997) 10 f3 liJg4 1 1 .ixe7 liJxh6 1 2 �h4 liJf5 1 3 �f2 Re issmann-Jolles, Dieren 1988, and easiest was 13 . . . e3! 14 iLg3 c5:j: . 8.121 9 fa
This is tricky move that forces Black to react in the centre: 9 ... c5 9 . . . l:tg6 was analysed briefly in the second edition but only played once that I know of. It still looks fine after 1 0 'iVh4 (10 'iWe3 b6! 1 1 fxe4 l:tg4 12 .id3 i.b7= , e.g., 1 3 'iVe2 c5! threaten ing . . . c4) 1O . . . c5 1 1 fxe4 (else 1 1 . . .�a5) 1 1 . . . .:tg4 12 �6 liJxe4 (12 . . . l:txe4+!?:j:) 13 liJe2 liJdf6:j: . 10 i..b 2 (a) 1 0 'iVe3 and 1O . . . liJd5 11 'i!Vxe4 1 25
Play the French lLlxc3 12 �xh7 lLlf6 with advantage or 10 . . . 'iVc7 1 1 fxe4 cxd4 12 cxd4 lLlxe4 (or just 12 . . . \\Wxc2) with advantage; (b) 10 fxe4 cxd4 1 1 cxd4 lLlxe4 with the idea of . . . �a5+, and if 12 'iVxh7?? lLldf6 13 �h4 �xd4. 1 0 ...'iVa5!? A cleaner and better solution ap pears to be 10 . . . cxd4! 1 1 cxd4 'iVa5+ and 12 c3 b6 13 fxe4 �b7! or 12 'iVd2 \\Wxd2+ 13 'it>xd2 b6. 11 0-0-0 b6 12 g3 Better is 12 fxe4 cxd4 1 3 cxd4 lLlxe4! 14 �xh7!? (14 d5 .i::[ g6 15 �e3 exd5�) 14 . . . 'i¥g5+ 15 'it>b l lLldf6 1 6 'iVh3 lLld2+ (16 . . . lLld5 !?) 1 7 � al lLlxfl 1 8 .i::[xfl ii.b7 19 lLlf3 \\Wxg2= . After 12 g3, Temmink- Stul1, corr 1983 went 12 . . . ii.b7 13 ii.g2 cxd4 1 4 cxd4 (14 .i::[xd4 l:1.c8) 14 . . . .i::[ c 8 1 5 'iVe3 'iVa4 with the initiative. 8.122 9 lLlh3 This move is logical and very im portant, but has declined in popular ity. White prepares i.g5 or lLlg5 while leaving the bishop on f1 free to move. 9 ... b6 A good alternative is here is the continuation 9 . . . c5 10 �e2 �a5 1 1 ii.d2
1 1 . . .l:1.g6 (this line got a bad reputa tion from Ljubojevic-Korchnoi, Til1 26
burg 1986 after 1 1 . . . .i::[xg2?! 12 lLlg5! cxd4 13 'iVg7 ! ; but 1 1 . . .cxd4 12 cxd4 'iVf5 intending . . . b6, . . . ..Iib 7, . . . 0-0-0 is unclear - Ftacnik) 12 �4 .i::[xg2 1 3 lLlg5 cxd4 14 cxd4 �d5 !? 1 5 i.e3 'iVf5 1 6 �f1 .i::[ g 4! launches a pretty attack: 17 ii.xg4 lLlxg4 18 'iVg3 h6 19 lLlh3 b6!? 20 lLlf4 .ta6+ 21 'iif;? g2 0-0-0 22 .i::[h dl .i::[ g8 23 'iif;? h l lLlgf6 24 �4 i.e2 25 .i::[ g l ii.f3+ 26 lLlg2 .i::[ g4 27 'iVxh6 lLlg8 28 �8 �g6 29 'ikh3 lLle7 30 'iVh8+ 'it>c7 3 1 d5 J:[xg2 0 - 1 Nieder maier-T.Martin, Bayern 1987.
10 lLlg5!? (a) 1 0 ..Iig5 ii.b7 11 ..Iib5 .i::[ g6 12 iVh4 h6! 1 3 ..Iixd7+ 'it>xd7 14 i.xf6 'iVxf6 1 5 'iVxf6= - Byr, although Black is quite active; (b) 10 i.e2 i.b7 1 1 0-0 "WIe7. Now Lane- Tisdall, Gausdal 1987 saw Black develop a textbook attack: 12 a4 0-0-0 1 3 i.a3 c5 14 'iVh4 �d6!? 1 5 .i::[fdl ! ? (15 a 5 e3 16 f3) 1 5 . . . e3! 16 f3 .i::[xg2+! 1 7 �xg2 .i::[ g8+ 1 8 lLlg5 h6 19 .i::[ g l? (19 dxc5! hxg5 20 cxd6 gxh4+ 2 1 �f1 lLle5 unclear) 19 . . . hxg5 20 �g3 'iVd5 21 �f1 lLlh5 22 'iVg2 lLlf4 23 �g3 lLlxe2 24 �xe2 'iVc4+ O- I . 1 0 . . ..i::[ g6 1 1 "iYh 4 i.. b 7! 12 lLlxh7 lLlxh7 1 3 �xh7 \\Wf6 This position has undergone quite a few tests; it seems to lead to a dy namic balance.
Winawer Variation: Fourth Move Alternatives Black seems to have plenty for a pawn: 19 dxe5 ctJxe5 20 �5 a6 2 1 'iVa4 ctJg4 22 'iVd4 'u'gg8 23 � 4 'ii'h.4 ! + 24 .i.d4 e3 25 .i.xe3 l:te8 with a deci sive attack, Leeuw-Mathews, IEee email 1999.
14 .i.e3 (a) 1 4 h4?! O-O-O! 1 5 �g5 l:th8 16 �xf6 l:txh7 1 7 i.e5 ctJxe5 1 8 dxe5 ,U,g5; Degraeve-Djurhuus, Arnhem 1988; (b) 14 'i'h3 0-0-0 1 5 'ii' e 3 'Ylig7 16 �b2 ctJf6 1 7 0-0-0 ctJg4 1 8 'ii'e 1 'i!fu6+ 19 �b1 'iVf4; Yeo-Knott, Birmingham 1999. 14 ... 0-0-0 15 'ii'h 5 e5 1 6 :dl Novotny-Kopecky, COIT 1989 went 16 0-0-0, and now Black had several interesting ideas, e.g., 16 . . . 'ii'c 6 and 16 . . . :h8 17 'iVe2 'iVe7! 18 �b2 c5 19 d5 with the idea 1 9 ... b 5 ! or 1 9 dxe5 ctJxe5 intending . . .b5, in both cases with attacking chances and a complex game ahead. 16 ... :h8 17 'ii'e 2 'i!Ve7 18 'i!Vc4 f5
White's king is in the centre and
8.123 9 ctJe2 The old main line which is again receiving considerable attention of late. I suspect that it is the most ef fective move because it challenges Black to speculate. See his option in the next note, however.
9 ... b61? A very interesting juncture. 9 ... b6 leads to fundamental and entertain ing play, but if Black wants to avoid the somewhat speculative play that can follow from this move, he should seriously consider 9 . . . c5. This is an attractive transpositional trick. Nor mally this position would be arrived at via 8 . . . c5 9 ctJe2 ctJbd7, and in that case 9 ctJe2 is not the move that most tests Black. 9 . . . c5 has in fact scored very well for the second player over the years. Black intends 10 ctJg3 'iVa5= or 1O . . . 'iVc7= , and the try 10 a4 can be met simply by 10 . . . b6 1 1 a5 J.b7= . Lastly, 10 g3 b6 11 .tg2 .i.b7 12 0-0 'iVe7 1 3 a4 ctJg4! was strong in Kir.Georgiev-Psakhis, Sarajevo 1986 127
Play the French due to 14 'it'xh7 0-0-0 15 'ifu5 1 0 liJg3!? The currently popular White protects g2, attacks prepares to complete his ment. The main option is 1I¥e7!?
f5 ! . choice: e4, and develop 10 ..tg5
16 i..h 4! 118g6 1 7 1i¥h8+ 'ug8 18 'iWh6= , but not here 16 i.xf6 liJxf6 1 7 i.e2 I14g6 18 'it'e3 liJd5 19 'it'd2 f5!) 1 5 ... liJg8!? (or 1 5 ... c5; both sides have a lot of options hereabouts) 1 6 'ilVe3 f5 1 7 liJh5!? e5 18 dxe5 liJxe5 19 O-O-O?? (but the suggested 19 liJf4 allows 19 . . . .l:txd2 ! and 20 'i'xd2 e3! 2 1 'i!Vxe3 11xf4! or 20 'iitxd2 'iVd6+ 2 1 'it>e2!? liJxc4 22 'it'c3 i.a6 23 'iit e 1 'it'xf4 24 i.xc4 i.xc4 25 'i!Vxc4 lhg2 with an ongoing but unclear attack) 19 . . . liJd3+ ! . Black is winning. Romero Holmes-Matamoros Franco, Elgoibar 1997 continued 20 cxd3 'iVxa3+ 2 1 'iit c 2 exd3+ 2 2 i.xd3 'it'a2+ 23 'iit c 3 l:txd3+! 24 'ilVxd3 'iVa3+ 25
11 1i¥h4 (1 1 liJg3?? liJg4 0- 1 was I.Andersen-Matamoros Franco, Co penhagen 2002; 1 1 a4 ..tb7 12 liJg3 of Zagorovsky-Kahn, corr 1 980 seems well met by 12 . . . .l:.g6 13 'i!Vh4 e5!, e.g., 1 4 liJf5 'iVe6) 1 1 . . . ..tb7 1 2 liJg3 (12 liJf4 0-0-0 1 3 i.e2 h6! , and 14 .i.xf6 'i'xf6 at least equalises, whereas 14 Ji.xh6 11h8 is an extremely awkward pin for White) 12 . . . h6!? (Black follows the the play in the famous game Fischer Kovacevic, Rovinj -Zagreb 1 970) 1 3 i.. d2 (13 'it'xh6? liJg4! 1 4 i.. x e7 liJxh6 15 ..th4 l:.g4- + ; 13 i.xh6!? 0-0-0 un clear) 13 . . . .l:!.g4!? (Kovacevic played 13 . . . 0-0-0 14 i.e2 liJf8 and won, but the game is so overanalysed that 13 . . . .l:.g4 is an appealing practical try. After Kovacevic's 14 . . . liJfS, McDonald does much analytical work with 1 5 liJh5 Itxg2! to support my contention that Black is fine, but there have been whole articles about alterna tives. My guess is that White has a small theoretical edge) 14 'ilVxh6 0-0-0 15 c4!? (Finkel suggests 1 5 i.g5! .l:.g8 1 28
This long-known position has been the starting point for a number of relevant correspondence and email contests. 1 1 i.e2 (a) 11 i..b 2 'i!Ve7 12 ..te2 0-0-0 13 0-0-0 liJg4! 14 i.xg4 l:txg4 1 5 'iVxh7 'iVg5+ 16 'iitb 1 .l:.h4 1 7 "ilxf7 11fS 18 1I¥xe6 i.d5=F Lindinger-Sprotte, Ger many 1999; (b) 11 i.g5 is probably equal, but has done horribly here. 1 1 . . .'iVe7 threatens . . . liJg4, and 12 'ifu4 h6 1 3 i.d2 0-0-0 has proven strong, e.g., 14
Winawer Variation: Fourth Move Alternatives .i.e2 e3 1 5 .i.xe3 i.xg2 1 6 l:tgl .tb7 1 7 a4?! e5 1 8 a 5 exd4 1 9 cxd4 'i!Vb4+ 20 i.d2 'i!Vb2 21 l::t c l liJe4 22 .te3 'ifc3+ 23 'iitf1 liJd2++ Benatar-Sukhov, IECG email 2000. 11 ...'iYe7 12 a4 (a) 12 0-0 0-0-0 13 f3 I1g6 14 'ml4! exf3 (14 . . . .:te8 15 fxe4 liJxe4=) 15 .i.xf3 l:tdg8? (15 . . . ii.xf3! 16 l:!.xf3 l:1dg8, per haps still slightly better for White) 16 ii.xb7+? (16 .th5 ! wins the exchange) 16 . . .';Pxb7 1 7 a4 'ii' d6! ; (b) 12 ii.g5 :g6 1 3 'iWh4 �d6! 1 4 i.f4 e5 1 5 i. e3 'i'c6; Zedekar-Spear, corr 1 988. 12 ... 0-0-0 Probably best. 12 . . . l:tg6!? 13 'ii'h 4 c5!? (13 . . . 0-0-0 is the old move) 14 0-0 e5!? 1 5 i.b2 'iVe6 1 6 a5! bxa5 with complications, although White should have some advantage, De Waard Lahlum, corr 2003. 13 0-0 liJg4!? 14 .i.xg4 Or 14 'iVh3 f5 15 i.f4 h5! (intending to advance further with . . . h4) 16 a5 (16 liJxh5? I1h8) 16 . . . h4 1 7 liJhl e5 1 8 dxe5 liJdxe5 with promising attacking chances. After 14 .i.xg4, Marez-D.Hardy, IECG email 2000 continued 14 . . . I1xg4 15 i.a3 �e8 (15 . . . c5 is also playable, and if 16 1\Vxh7?, 16 . . . �f6! 1 7 'ml5 1\Vg7! 18 h3 liJf6 19 'i'e5 I1d5-+) 16 a5 f5 (a key move in these lines) 17 axb6 axb6 18 i.el 'i'g6 19 h3 1\Vxh6 20 .txh6 Itg6 Y:! - K I think that the 8 . . . liJbd7 variation is playable and exciting. However, there is massive theory attached to it, and the main lines with 9 liJe2 b6 may not appeal to those who are nervous about sacrificing pawns. To them I would recommend looking into 9 . . . c5 (less ambitious but solid) or even earlier deviations such as 5 . . . liJe7.
8.2 4 'iWg4 The 'Blitz Variation'. White tries to save a tempo on 4 a3 by bringing his queen out immediately. The problem is that he hasn't bolstered his centre (as bxc3 did above) , and Black can therefore counterattack more quickly in that sector. Although debate still rages about 4 'ifg4, I think that accu rate play by Black makes it difficult for White to even equalise. 4 .. .eiJf6 5 'iYxg7 .:tg8 6 'iWh6 c5! It turns out that this move is far better (not just a more accurate move order, as has been stated) than the immediate 6 . . . I1g6 In the latter case, a critical line goes 7 'ii'e 3 c5 8 .i.d2 ! and 8 . . . liJc6 9 liJge2 (see the notes below about how 6 . . . c5 avoids this) or 8 . . . liJg4!? 9 'ii' d 3 liJc6 10 liJge2 (10 a3? c4 1 1 'ii'h 3 liJxd4+) 10 . . . cxd4 1 1 liJxd4 unclear. Current theory is fairly nice to White here. I have suggested 1 1 . . .'i!Vb6, but I don't think that it's been tested. 7 liJge2 This is supposed to be the best move and it is in fact the main line, because after 7 . . . lig6 8 'i'e3 we have transposed to a position that is also reached via 6 . . JIg6 7 'ii'e 3 c5 8 liJge2. If then 8 . . . liJc6 9 .td2, we reach the traditional main line of 4 'iWg4. But we shall see that things aren't so easy. Here are White's alternatives: (a) 7 i.g5? I1g6 8 'i!Vh4 cxd4 9 e5 l:txg5! 10 exf6 'iWxf6 11 liJge2 liJc6 12 a3 .i.. a 5 13 b4 liJxb4! 14 0-0-0 liJxc2! 1 5 liJb l .i.. d 7! 16 I1d3 liJel 0- 1 Crowl Harris, corr 1 938; (b) 7 e5!? has been played a lot: 7 . . . cxd4 8 a3 .i.f8 (also promising looks 8 . . . 'iWa5 !? 9 exf6 dxc3 10 l:tb l , and McDonald's idea 10 ... liJd7!?, since 1 1 axb4 'iWa2 is good, or 10 . . . cxb2+ 1 1 axb4 bxcl'iW+ 1 2 'iVxc l �c7 with a 1 29
Play the French positional advantage) 9 'i!Vxf6 (9 �4 dxc3 10 exf6 lDd7 1 1 �b5 cxb2 12 i.xb2 'i!lia5+) 9 . . . 'i!lixf6 10 exf6 dxc3�
11 lDe2 (1 1 bxc3 lDd 7 12 �b2 lDxf6 1 3 c4 i.. g n Y.Hernandez-S.Bjerke, Oropesa del Mar 1 999) 1 1 . . .lDd7 12 lDxc3 �d6 1 3 g3 lDxf6 1 4 JLg2 (14 lDb5 i..b 8 1 5 c4 a6 16 lDc3 dxc4 17 i.xc4 b5 1 8 �e2 �b7 1 9 0-0 h5 McDonald) 14 . . . �d7 1 5 0-0 0-0-0 1 6 � e 3
1 30
At this point, three bad mistakes that have happened over the board are: 9 e5? lDg4 10 Wd3 cxd4, 9 lDf3? dxe4 10 lDe5 cxd4 1 1 �b5+ 'iit e 7!, and 9 exd5? lDxd5 10 'i!lid3 cxd4 1 1 'i!lixd4 e5! 12 'i!lid3 l:.d6 ! . So White has: (c 1) 9 lDh3 lDxe4 (9 . . . e5!?; 9 . . . lDc6 10 �b5 dxe4+) 10 b4 lDxc3 1 1 'ii'xc3 cxd4 1 2 'iWxd4 e5!+ Plasman-Steen bekkers, Dieren 1 998; (c2) 9 lDge2 lDc6 (or 9 . . . cxd4 10 lDxd4 lDxe4) 10 i.. d 2 cxd4 (10 . . . �b6!? 11 lDxd5!? lDg4�) 11 lDxd4 lDg4 transposes to 9 i.d2; (c3) 9 i.d2
9 . . . lDc6 (9 . . .cxd4 10 'ii'xd4 lDc6 1 1 'i!lid3 lDxe4 1 2 lDxe4 dxe4 1 3 'ii'xd8+ i.. x d8 is 'only' equal) 10 lDf3 ! (10 lDge2?! cxd4 11 lDxd4 transposes, as (probably) does 10 . . . lDg4; but 1O . . . i.b6! is strong; 10 i.b5!? cxd4 1 1 'iVxd4 i..b 6 1 2 Wd3 lDxe4 1 3 lDxe4 dxe4 14 'iVxd8+ xf2 i.b6 14 �e3! ( 1 4 exd5 i.. xd4+ 1 5 i.. e 3 'i!lif6+ 16 f2 lDxd4+ 19 el lDxd4�) 15 . . . lDe5! 16 'ii'b 5+! �d7 17 i.. xb6
Winawer Variation: Fourth Move Alternatives .i.xb5 18 i.xb5+ f8 19 i.c5+ Wg8 intending the moves . . . ':'xg2+ or . . . dxe4 and . . . tDg4+. In spite of the approximate material equality, Black stands better; (d) 7 �d2? cedes the advantage immediately to 7 . . . cxd4! (7 . . J:tg6?! 8 'ife3 is theory's main line) 8 e5 (8 tDb5 tDg4! 9 'ilixh7 �xd2+ 1 0 xd2 tDf6!) 8 . . . dxc3 9 bxc3 'ilic7!�.
From the diagram, the normal move here has been 7 . . . l:rg6, when 8 'i!Vh4!? is a try to avoid the usual move 8 'iVe3. The latter can be met by 8 . . . cxd4! (8 . . . tDc6 9 i.d2 is the main line of the 4 \\!kg4 Winawer, consid ered dynamically equal) 9 tDxd4 tDxe4 and I don't see anything wrong with Black's position. This could be the first benefit of playing 6 . . . c5 instead of 6 . . .l::t g6. 7 ... cxd4 Here's another benefit! Pederson says that 'it is obviously not favour able for [Black] to exchange so early on d4'. While that is a logical asser tion because 7 . . . cxd4 frees White's pieces, it happens that there are con crete tactical compensations. Inter estingly, 7 . . . tDc6!? also hasn't been sufficiently tested, the main line go ing 8 exd5 tDxd4 9 tDxd4 cxd4 10 a3 .i.f8 1 1 �b5+ and if nothing else,
1 1 . . .We7!? 12 d6+ (what else?) 12 . . . 'iVxd6 13 'ifxf6+ 'ifi>xf6 14 tDe4+ �g7 15 tDxd6 �xd6 16 0-0 e5 is equal, whereas 1 1 . . .i.d7 tries for more: after 12 i.xd7+ of Polland Rubinsky, corr 1983, 12 . . . Wxd7!? 13 dxe6+ rtJe7 is an awfully speculative piece sacrifice for White. So 7 . . . tDc6 may be much better than its reputa tion. 8 tDxd4 e5!
This looks very promising, e.g., 9 i.b5+ i.d7 10 i.xd7+ tDbxd7 1 1 tDf5 d4 12 tDg7+ (12 0-0 i.f8-+) 12 . . . :xg7 13 \\!kxg7 dxc3 14 0-0 cxb2 (or 14 . . . i.f8+) 15 i.xb2 'iVe7!? (15 .. :iYc7+) 16 f4 tDxe4 17 h1 0-0-0 18 fxe5 i.c5 1 9 h3 i.f2 20 l:rfd1 i.h4 21 l:1.d4 tDg3+ 22 'ifi>h2 tDf5 23 l:rc4+ b8 24 'i!Vxh7 tDxe5 25 .l:.e4 tDf3+ 26 gxf3 l:rd2+ 27 'ifi>gl 'ilig5+ 0 - 1 Bartel-Szelag, Brzeg Dolny 200 1 . For all the lengthy ar guments about the 4 'iVg4 main lines, it appears that 6 . . . c5! avoids them completely and for several separate reasons resolves the variation fa vourably for Black. 8.3 4 i.d2 Here again White's idea is 'i!Vg4, and play will resemble 8. 1 and 8.2. The main difference is that White develops more rapidly, but he forfeits 1 31
Play the French any pretensions to holding his centre and in fact gambits a pawn. His re ward is rapid development. 4 .i.d2 has some fascinating byways, and is quite as interesting as more popular options; I will treat it in some detaiL It is notoriously difficult to avoid forc ing drawing lines versus 4 i.d2, so I'll give three unbalanced solutions with two based upon the same sequence.
4 ... dxe4 4 . . . liJc6!? is also possible for origi nal play: 5 �g4 (another idea was 5 a3 i..x c3 6 1i.xc3 liJf6 7 e5 liJe4 B i.. d 3 liJxc3 9 bxc3 "fie7 followed by . . . i.. d 7, and . . . 0-0-0, Ljubojevic-U.Andersson, Palma de Mallorca 1971) 5 . . . liJf6 6 Wixg7 l:!.gB 7 'iVh6 l:!.g6! B "fih4?! (B 'iVe3 dxe4 again transposes to the main line, and here B . . . liJg4 is a good alternative) B . . . l1g4 9 Wih6 liJxe4!? (9 . . . liJxd4�) 10 'iVxh7? (10 liJf3 liJxd2 1 1 'iYxd2 e5!) 10 . . . liJxd2 1 1 'it>xd2 'ikg5+ 12 'it>el l1e4+! 13 liJge2 liJxd4+ Petzold-Fuchs, Berlin 1965. 5 'iYg4 5 liJxe4? "fixd4+ , e.g., 6 .i.d3 .i.xd2+ 7 'iYxd2 "fixb2 B l:tdl liJd7. 5 . . liJf6 6 'iYxg7 .i:i.g8 7 "fih6 And now: .
8.31 7 ..."fixd4 8.32 7 ... l:!.g6 1 32
7 . . . liJc6 has independent value if one is aiming for a draw. This can arise after B 0-0-0 (B liJge2!?) B .. JIg6 9 'iWh4, transposing to 7 . . . :g6 B 'i¥h4 liJc6 9 0-0-0 in B.32. 8.31 7,..'iYxd4! Fearlessly grabbing a pawn. The ory on this move has expanded con siderably and I think that it offers ' Black excellent chances for the ad vantage, thus calling 4 i.. d2 into some doubt. 8 0-0-0 White plays B liJge2!? fairly often, but Black has plenty of play after B . . :ife5 9 0-0-0 (or 9 .i.f4 "fif5� Keres; then 10 liJg3 "fig6 1 1 "fixg6 i..xc3+! 12 bxc3 .l:txg6 1 3 i.xc7 liJc6 offers Black some advantage, and Halldorsson- Schmitz, Reykjavik 2000 went 10 i.xc7 "fixf2+! 1 1 'It>xf2 liJg4+ 12 'it>e l liJxh6 13 a3 i.xc3+ 14 liJxc3 f5 1 5 liJb5 liJa6� intending . . . liJf7; with out central breaks the bishop pair doesn't fully compensate for his extra pawn)
Here Black can play 9 . . . i.. fB, which I won't deal with here, or choose be tween: (a) 9 . . . J::t g 6 10 'i¥h4 .l:i.g4!? 1 1 'iYh3 liJc6 12 f3 exf3 13 'iYxf3 a6 (13 . . . i.. d 7 is natural and good) 14 .i.f4 'iYa5 1 5
Winawer Variation: Fourth Move Alternatives i.xc7!? "fixc7 16 'iYxf6 iLf8! 1 7 g3 .i.g7 18 'iVf2 i.. d 7 19 i..h 3 .l:!.c4 and Black exerted queenside pressure m Mineur-Abenius, corr 1 988; (b) McDonald thinks that Black is doing well after 9 . . . liJbd7! and that seems right, e.g., 10 liJg3 (10 i.f4 'iVf5 or 10 . . :i'a5 are both good according to McDonald) 10 . . . liJg4!? (or 1 O . . . l:1g6 intending 1 1 'iVe3 .i.c5 or 1 1 'iWf4? i.xc3 12 bxc3 'iVa5 1 3 'iitb 2 e5 with a winning attack, Brouwers-Van Gee men, corr 1 985) 1 1 'iWxh7 liJdf6 12 'iWh4 .i.e7 13 l:.e l?! liJxf2 14 i.f4 "fia5 15 l:rgl iLd7 16 .i.e2? 0-0-0 and White's queen was embarrassed m Kasten-Kirwald, corr 1986. Mer 8 0-0-0, Black has: 8.3 1 1 8 ..liJg4 8.3 1 2 8 ... i.f8 .
8.3 1 1 8 ... liJg4!?
An obscure move that I've looked into and believe works. There have been almost no games with 8 . . . liJg4 (two ended in mate 4 moves later!), but it seems to give full-fledged chances with oodles of non-forced play. At least it avoids drawish lines. 9 'iWh4! Others are either disastrous or un promising:
(a) 9 'iWf4? e5 10 "fig3 'iVxf2 1 1 'iYh3 liJc6 12 'iWxh7 liJf6 13 'iYh6 i.f8!? (also 1 3 . . . .i.g4- +) 14 'iWe3 liJg4 (14 . . . i.c5 !-+) 1 5 'iWxe4 i.f5 16 'i!Va4 0-0-0 and Black's attack was too strong in Cross-Playa, Roque Saenz Pena 1997; (b) 9 'iYxh7?? l:th8 10 �xe4 (or 10 liJge2 'iVxd2+ 11 .l:!.xd2 .l:!.xh7 also win ning) 1 0 . . . .i.xc3-+ ; (c) 9 'iVh5 liJxf2! (9 . . . liJd7!? also looks good in view of 10 .i.e3 'iWe5 or 10 liJb5 .i.xd2+ 1 1 .l:!.xd2 'itb6 12 'iWxh7 liJdf6 1 3 'i'h4 i.. d 7+) and: (c l) 10 liJge2? 'i'f6 1 1 liJb5 (1 1 'iYxh7 .l:!.g4! threatens . . . .l:!.h4 and guards the e-pawn: 12 'iWh5 .i.xc3 13 liJxc3 "fig6-+) 1 1 . . .iLxd2+ 1 2 ':'xd2 llg5 13 liJxc7+ �f8 14 'i'xh7 liJxhl 1 5 liJxa8 l:tf5-+); (c2) 1 0 liJb5! .i.xd2+ 11 ':xd2 'iWe3 12 liJxc7+ �f8 13 liJxa8 liJxhl is messy, but I think that Black is al ways slightly better, e.g., 14 'i'xh7 (14 liJh3 liJc6) 14 ... liJf2 1 5 liJe2 liJc6 16 'iWh4 liJe5! (16 . . . �g5!? 1 7 'iYxg5 .l:!.xg5:j:) 17 liJc7 i.d7!+ (17 . . . liJc4 18 'iYd8+ r:3i;g7 19 'iWd4+ 'iVxd4 20 .l:!.xd4 liJe3 21 liJg3 f5:j:).
9 . .liJd7!? Not necessarily best, this leads to dynamically balanced play. There are several alternatives, one of which is not 9 . . . 'iVxf2?? 10 'iVd8+! �xd8 1 1 .
1 33
Play the French �g5+ �e8 12 l:td8 mate (as occurred in at least two games!). Instead there are: (a) 9 . . . 'iVf6 10 �g3
(b) 10 �g5 'i'xf2! l 1 lbxe4 (1 1 �xg4 i.xc3 12 bxc3 �c5 13 lbh3 h6- +) 11 ... �xh4 12 ..txh4 i.. e 7 (12 . . . f5! 13 lbg5 l:txg5 14 Sl.xg5 lbf2=F is also good) 13 �xe7 Wxe7=F . 10 .. JWe5 1 1 �b ll White correctly avoids 1 1 'iWxh7?! J::th 8 12 'iWxe4 �xc3 13 �xg4 �xb2+ 1 4 �b 1 i.a3 15 c3 lbc5 ! =i= . 1 1 ... �e7 1 2 �xh7 12 'iWg3 �xg3 1 3 hxg3 allows 1 3 . . . e3! 14 fxe3 i.c5. 12 lbdf6 1 3 'iWh4 e31 14 fxe3 lbd5 15 'iWell Not 1 5 'i'h7? l:th8; or 1 5 'i'g3? 'iWxg3 16 hxg3 lbxc3+ 17 i.xc3 lbxe3=F . 15 ... lbdxe3 1 6 i.xe3 lbxe3 The play has been pretty forcing thus far and this looks equal. It might continue 1 7 lid3 (17 :d2 .i.b4 1 8 l:te2? i.xc3 1 9 lIxe3 i.xe1 20 l:txe5 �b4 leaves Black with two strong bishops) 1 7 . . . lbc4 18 'ifxe5 lbxe5 1 9 lbb5!? i.. d 8! = . I believe that the ob scure move 8 . . . lbg4 is fully playable, although the precise details have to be worked out. •••
and: (a1) 1O . . . i.. xc3 1 1 i.. xc3 e5! 12 lbh3 (versus .. :iixf2) 12 . . . lbc6 13 .i.b5 �d7 14 l:the 1 0-0-0 15 l:txe4 'iWe6! 16 �c4 �6+ (or 16 .. :iVe7!? 17 l:tee 1 ! lbf6 1 8 � f3 .i.g4 1 9 l:!.xd8+ l:txd8 20 'iVg3 �e6=) 17 i.. d2 �5 with at least equality; Black is ready for . . . f5; (a2) 10 . . . e3!? 1 1 �xe3 (1 1 fxe3 .i.d6=) 1 1 . . .�xc3 12 bxc3 'i'e7! is complex and hard to assess, but I think that White's strong bishop pair are more or less balanced out by his weaknesses. (b) 9 . . .lbxf2 is extremely compli cated but looser than 9 . . . 'i'f6 or 9 . . . lbd7. A key line is 10 lbb5 (10 lbf3 .i.e7! 1 1 Wrfxe7+ �xe7 12 lbxd4 lbxhU:; probably 10 lbxe4! i.e7 1 1 �xf2 Wrfxe4 1 2 lbf3 lbc6 1 3 �d3 is best) 10 . . . i.. e 7! l 1 lbxc7+ �d7 12 Wrff4! lbxd1!? (12 . . . e5!? 13 'iWxf7 lbxd1 1 4 lbxa8 :f8 1 5 'i!fb3 e3!?) 1 3 lbxa8 e5 1 4 �xf7 :f8 seems to work out well enough, e.g., 15 i.b5+ �d8 16 m3! �e6 ! = . 1 0 lbh31 (a) 10 lbb5 i.xd2+ 1 1 l:txd2 l\Yb6 in tending . . . e3 looks fine; 1 34
8.3 1 2 S ... i.f8 9 �4 Worse is 9 Wrff4?! i.d6 10 lbge2 �xf4 l 1 lbxd4 i.. e 5=i=. 9 ...l:tg4 10 'iWh3 'iWxf'2
1 1 �e2
Winawer Variation: Fourth Move Alternatives Hitting the rook and threatening :f1 . Others are worse: (a) 11 lLlb5? lLla6 12 'ii?b l Ji.d7 13 Ji.e3 'fif5 14 lLld4 'iig 6+ Boleslavsky Bronstein, Moscow 1950; (b) 11 Ji.b5+? c6 12 .l:[fl 'iixg2 13 .l:.xf6 cxb5+ Mossung-Vrbata, Prague 1989; (c) 11 Ji.e3?! 'fih4! 12 'iixh4 (12 g3 'iixh3 1 3 lLlxh3 lIg6 14 lLlf2 �d7 1 5 �g2 �c6 16 lLlcxe4 lLlxe4 1 7 �xe4 Jansen-Alger, email 1997; and best looks 17 . . . f5 18 .txc6+ lLlxc6+) 12 .. .l:1xh4 13 g3 (13 .tg5? .th6) and here Black has the simple escape 13 . . . I1h5 ! , when he'll remain a pawn up after the e-pawn falls (compare the main line). 1 l ...l1h4 This was once considered a blun der, but is now the main line. I'm not sure if anyone has ever repeated 1 1 . . .Wxg2!? of Redolfi-ldigoras, Mar del Plata 1 956, but it looks promis ing: 12 �xg4 (12 lLlb5 lLla6) 12 . . .'i¥xg4 1 3 1lVxg4 lLlxg4 14 lLlxe4
14 . . . .td7! (or 14 . . . lLld7!?; Black has two bishops, material equality and the two unopposed central pawns that he gets in the main line, al though his pieces aren't quite as well placed after 15 lLlf3 b6!? 16 .l:thgl h5 17 h3 .tb7 18 hxg4 Ji.xe4 19 lLlg5
.tg6=) 15 h3 �c6 16 hxg4 �xe4 1 7 .l:.h3 ( 1 7 lIh4 lLlc6 18 g 5 �g6:j:) 17 .. .ltJd7:j: (or 17 . . . lLlc6:j:). I don't know why 1 1 . . :fixg2 hasn't been touched in so many years, especially considering its similarity to lines that follow. 12 'i'xh4! 1lVxh4 13 g3 The queen is trapped, but as com pensation for losing the exchange Black will have the bishop pair and strong central pawns.
l3 ... e3!? This underrated pawn advance (threatening . . . 'i!Vb4) has been around for a long time but has been over shadowed by the other trade-off of White's dark-squared bishop via 13 . . . 1lVh6 14 �xh6 �xh6+ 15 'ii?b 1 . Nevertheless, the latter position is justifiably popular for Black. Apart from the usual 1 5 . . . e5 (about equal) and 15 . . . �d7 (unfathomable, but still doing well), a new move for Black is 15 . . . a6!? I like this modest-looking continuation, which plans to develop quickly by . . . b5 and . . . �b7 while toss ing . . . b4 into the mix of ideas. Two sample games: (a) 1 6 h4 b5 17 lLlh3 (17 g4 b4 1 8 lLla4 � e 3 19 g 5 � d 7 2 0 gxf6 Ji.xa4 2 1 lLlh3 lLld7:j:) 1 7 . . .b 4 18 l:t df1 (18 lLla4 �e3! 19 .l:.hf1 <:Jite7) 18 . . . bxc3 19 .l:txf6 �g7 20 l:i.ff1 f5 2 1 .th5+ 'ii? e 7 22 g4?! 1 35
Play the French (but 22 b3 liJd7+) 22 . . . cxb2 23 gxf5 exf5 and the central pair has sur vived with a three-pawn advantage, so White was effectively lost in the game Korepanov-Skomorokhin, Po dolsk 1 993; (b) 16 liJh3 is natural: 1 6 ... b5 17 l1hf1 liJbd7 18 g4 b4 19 g5 (19 liJa4 liJd5 20 g5 .i.f8 21 �h5 liJe5 22 l1de 1 i.. d 7 23 l1xe4 liJg6 24 c4 - forced, be fore or after .i.xg6 - 24 . . .bxc3 25 liJxc3 liJxc3+ 26 bxc3 .tc6 27 �g4 l1b8+ 28 �c2 (28 �a1 �d6!:j:) 28 . . . �a3 29 .l:i.b 1 J::txb 1 30 �xb 1 .i.d6!:j:) 1 9 . . . �xg5 20 liJxg5 bxc3 21 l1f4 cxb2 22 .tc4 (22 liJxe4 liJxe4 23 l1xe4 'ub8) 22 . . . i.b7+ Rantanen-Raisa, Tampere 1 989, due to 23 liJxf7? i.. d 5!, winning material. 14 gxh4 exd2+ 15 �hl Said to be best by nearly everyone. The obvious 15 �xd2 exposes the king to tempo-gaining attacks after 1 5 . . . .i.d7 (or 15 . . . i.h6+ 1 6 �e 1, and now 16 . . . �e7 17 liJf3 .i.f4! with un clear prospects or 16 . . . .i.d7 17 liJh3 .tc6 18 l1f1 liJbd7 Hort-Nogueiras, Reggio Emilia 1 986) 16 �c1 (16 liJf3 .td6 1 7 �c1 �f4+ 18 �b 1 liJc6 1 9 11hf1 l1d8 20 liJg5 � e 5 2 1 liJce4 liJxe4 22 liJxe4 �e7 and Black's central pawns are still dangerous) 16 . . . .i.c6!? (16 ... liJc6 1 7 liJf3 0-0-0) 1 7 .tb5 (17 .i.f3 .th6+ 18 �b 1 liJg4!? 19 liJe4 liJd7 20 .i.xg4 .txe4 21 �f3 liJf6 22 !:tel 0-0-0=) 17 . . .i.h6+ 1 8 �b 1 �e7 19 i.. xc6 liJxc6 20 liJh3 .l:i.g8 21 .l:thg1 l1xg1 22 1:!.xg1 .te3 23 J::t g 3 .td4 24 liJg5? ! (24 liJe2 i.b6 25 �c1) 24 . . . h6 25 liJge4 liJxe4 26 liJxe4 f5 (simplifica tion tends to help Black in this line, because it frees his pawns to ad vance) 27 liJd2 'it>f7!? 28 h5 .i.e5 29 J::t g2 ..tf6 30 c3 e5 31 'iit c 2 e4 and Black's two passed pawns are begin ning to dominate, D.Larsson-T.Karl sson, Vaxjo 1992. 1 36
15 ... .th4 I suggested this years back; Black is trying to isolate every White pawn! Whether that occurs or not, the es sence of this position is Black's bishop pair and two passed central pawns versus White's temporary develop mental edge. A similar and potentially impor tant idea would begin with 15 . . . i.d7!, again using the ... ..tb4 motif and counting upon the passed central pawns.
Some sample analysis: (a) 16 l1xd2?! .i.b4 1 7 l1d3 (17 .tf3 .i.xc3 18 bxc3 i.c6) 17 . . . �xc3 18 bxc3 (18 l1xc3 liJe4 wins material) 18 . . . liJe4 19 liJh3 .i.b5 etc.; (b) 16 a3 .i.c6 1 7 liJf3 i.h6! (17 . . . liJbd7 1 8 l1xd2 i.h6 19 l:tdd1 �e7
Winawer Variation: Fourth Move Alternatives is also fine) 18 .i.b5 cJ;; e 7 19 i.. xc6 ttJxc6 20 ttJxd2 liJg4! 2 1 liJde4 liJe3 22 lId2 liJf5 ! and White's rooks lack squares; best is 23 11f2 .i.e3 24 J:tffl l:tg8 25 ttJdl �b6, when one prefers Black; (c) 16 ttJf3 �b4! 17 ttJxd2 .i.xc3 1 8 bxc3 ttJc6 1 9 .l:I.hgl � e 7 20 .l:t g7 l1g8! 21 l:tg5 h6 22 l1xg8 ttJxg8 and again Black's pawns are the long-term fac tor, so White can hope for equality at best; (d) 16 ttJb5 �xb5!? (or 16 . . . liJa6 fol lowed by . . . liJe4) 1 7 �xb5+ c6 1 8 �d3 (18 .i.e2 ttJe4) 18 . . . ttJbd7 (18 . . . i.. h 6 19 ttJf3 liJg4! 20 lIhgl liJf2:j:) 19 l1xd2 ttJe5 20 ttJe2 l:td8 2 1 11f1 .i.h6 22 J:tddl cJ;; e 7 leaves White looking for a plan.
16 .i.f3 (a) 1 6 11xd2?? �xc3 17 bxc3 ttJe4-+ ; (b) 16 ttJb5 �a5 ( 1 6 . . . ttJa6 1 7 c3 i..c 5 18 ttJf3 ttJe4) 17 ttJf3 ttJc6 intend ing . . . e5, and answering 18 liJxd2 by 18 . . . a6 19 liJc3 (19 liJa3? �xd2 20 l1xd2 liJe4) 19 . . . Axc3 20 bxc3 e5 2 1 .i. f3 �e6! 2 2 �xc6+ bxc6 23 lIde l liJd7. 16 ....i.xc3 1 7 bxc3 liJbd7 1 8 l:!.xd2 liJe5 19 Ag2 ttJc4 20 l1e2 e5!? Also interesting are 20 ... �d7 and even 20 . . . c6!?, e.g. , 21 ttJh3 ttJd5 22 .i.xd5 cxd5 23 11g1 �f8 24 :f2 ttJd6. After 20 .. e5, play might continue
2 1 cJ;; c l i.. e 6!? (or 2 1 . . .c6 22 liJf3 i.. g4) 22 .i.xb7 :b8 23 .i.c6+ cJ;; e 7 24 ttJf3 Wd6 25 J.e4 (25 i.. a 4 ttJd5 26 i.. b 3 ttJxc3:j:) 25 . . . i.. g4 with various threats and at least equal play. 8.32 7 ... l1g6 This is a solid move that forces White to commit at an early stage. It could serve as a backup if the 7 .. :VWxd4 lines look shaky or intimi dating. 8 'YWe3 The alternative is 8 'YWh4 when Black can apparently force a draw, but also has options: (a) 8 . . . ttJc6 9 0-0-0 .l:tg4 (9 . . . i.. xc3 10 �xc3 'iVd5 !? lost a famous game be tween Keres and Botvinnik; 10 .. :iVd6 is more interesting) 10 'YWh6 (10 'iVh3 l1g6 intends . . . e5; then 1 1 'YWe3?! �xd4 12 'i'xd4 ttJxd4 is considered good for Black, but so is 1 2 liJb5 'iVxe3 1 3 i..x e3 �a5:j:) 1O . . . l:tg6 1 1 'YWh4 l:rg4 with a draw; (b) 8 . . . 'YWxd4 as a winning try has the problem 9 ttJf3 'i'c5 10 liJxe4= ; (c) 8 . . . ttJbd7 is obscure but appeal ing, covering everything and aiming to develop rapidly and prepare . . . 0-0-0: 9 0-0-0 (9 ttJxe4 l:!.g4 10 ttJxf6+ ttJxf6 1 1 'iVh6 is messy after 1 1 . . . .i.f8 12 'Wh3 'iVxd4 1 3 O-O-O! 'YWxf2 14 ttJf3 �d6 and Black is temporarily a pawn up with . . . �d7 and . . . 0-0-0 planned) 9 . . . �xc3 10 �xc3 b6 1 1 i..b 5 i..b 7 12 ttJe2 (12 'YWh3 a6) 12 ... liJd5 !? (simply 12 . . . c6 13 �a4 b5 14 .i.b3 a5 looks effective; in several lines . . . l:txg2 will follow) 13 'iVxh7 'YWf6 14 'iVh3 (14 i.. x d7+?? Wxd7 and White's queen is trapped after 1 5 'iVh5 l:rg5 16 'iVh4 l1h8) 14 . . . 0-0-0 1 5 �el Graf-Uhl mann, Wuerzburg 1992, and here 15 . . . c6 16 �c4 :h8 gives compensa tion, e.g., 17 'iVa3 lhg2 18 'iVxa7 137
Play the French l:lhxh2 19 l:lxh2 l:txh2 and White has nothing on the queenside so Black is at least equaL 8 ... ltJc6 9 ltJge2 Minev gives the line 9 ltJxe4 ltJxe4 (9 . . . i..xd2+ 10 ltJxd2 �xd4=) 10 'ii'xe4 (10 i..xb4? ltJxb4 1 1 'i¥xe4 'ii'xd4!) 10 . . . i..xd2+ 11 �xd2 ltJxd4 (1 1 . . .'flixd4+ 12 'i¥xd4 ltJxd4 followed by . . . e5 is equal) 12 l:tdl ! . Then 12 . . . c5 13 �cl �a5 14 lIVe5?! (14 'it>b l i.. d 7=) 14 . . . 'i¥xa2! 15 i.b5+ i.. d 7 16 �S+ (16 i..xd7+ 'iitx d7�) 16 ... 'iite 7 1 7 'iVxaS �al + IS 'iit d2 �a5+ 19 'it>cl 'iVxb5 gives more than enough for the ex change; Black is threatening . . . �c4. 9 ... i.. x c3 A safe line which gives a reason able game, as opposed to the theoreti cal 9 . . . e5?! 10 dxe5 ltJg4 1 1 �f4 ltJcxe5?, when there's a huge hole in the traditional line 12 h3 i.. d6 1 3 �xe4 f5 14 'flia4+! i.. d 7, namely 1 5 ltJb5! and White i s a pawn ahead with the better position.
10 i..x c3 (a) 10 �xc3 is also playable: 10 . . . ltJd5 1 1 �a3 (1 1 �c5!? �d7! 12 0-0-0 b6 1 3 �a3 i..b 7=) 1 1 . . .�f6! (ty ing White to the f-pawn; this is given as equal by the books, but it's not clear; here's some analysis:) 12 c4! ltJb6 13 �c3 i.d7 1 4 .te3 0-0-0 1 5 1 38
0-0-0 ltJe7!? ( 15 . . . 'it>bS 16 ltJf4 l:tggS 1 7 i.e2 ltJe7) 16 h4! i.. c 6! 1 7 i.. g 5 :xg5 IS hxg5 'flixg5+ 19 'flid2 (19 'it>b l lIVf6 20 �4 ltJf5) 1 9 . . . e3! 20 'iVxe3 'iVxe3+ 21 fxe3 ltJxc4 22 l:lxh7 (22 l:td3 ltJd5 23 l:lh3 ltJb4!) 22 . . .ltJxe3 23 IIel ltJc2! 24 l:.dl ltJe3= ; (b) 10 bxc3 e5!? 1 1 h3 (1 1 dxe5 ltJxe5 12 ltJf4 l:.gS 13 f3 �e7! 14 fxe4 ltJfg4 15 'iVe2 .te6=) 1 1 . . . .tf5!= intend ing 12 g4 ltJd5. 10 ...�d6!? Another move is 10 . . . ltJe7, e.g., 1 1 ltJg3!? (but 1 1 ltJf4! looks better: 1 1 . . . .u.gS 12 0-0-0 'i!Vd6;!;) 1 1 . . . ltJed5 12 'i¥e2?! (12 'flid2 h5! 1 3 h4? e3 14 'i!Vd3 'i¥d6+) 12 . . . ltJxc3 13 bxc3 'iWd5 14 'i¥b5+ c6 (or 14 . . . i.d7) 15 �4 h5 16 i..c 4 �dS 17 l:tgl h4 I S ltJf1 b5 19 i.e2 ltJd5 20 'i¥c5 1!Vb6 with a nice advan tage, Brendel-Poldauf, Tege1 200 1 .
l l ltJg3 At this point I'll give some sample moves. Instinctively White looks very slightly better, but that has to be proven. One old line is 1 1 0-0-0 ltJb4, but instead 1 1 . . .ltJg4! wins material after 12 'iYf4 'i¥xf4+ 13 ltJxf4 ltJxf2 14 ltJxg6 ltJxd1 1 5 Wxdl hxg6. 1l ... ltJb4 12 'flie2 12 i..b 5+!? c6 13 i.a4. 12 ... i.. d 7!? 13 ltJxe4 ltJxe4 14 'iVxe4 i.. c 6 15 'flie2 ltJd5 16 i.. d 2 ltJf4 1 7
Winawer Variation: Fourth Move Alternatives .i.xf4 ii'xf4 and Black regains his pawn. Unlike 7 . . . 'i!Vxd4, 7 . . . .:.g6 plays only for equality; it is nevertheless a prac tical line that avoids much theory and yields a playable game. 8.4 4 ttJge2
with advantage. Rustemov analyses 14 f5 i.. a 6! 15 ttJd6 i.xd6 16 exd6 'i!Vb6 1 7 ttJa4 'iVa5 18 ttJc5 (18 ttJc3 ':'xb2! 1 9 �xb2 l:.b8+) 1 8 . . . c3. This should fa vour Black, e.g., 1 9 bxc3 'i!Vxc3 20 ttJxa6 ,Ub2 2 1 l:td2 'iYxa3 and 22 ':'f2 lia2+ 23 �d2 lial-+ or 22 I:!.d3 'i!Val + 23 �d2 'i!Vxa6+ . 8.41 4 ...ttJc6 This was suggested in the fIrst edi tion and has grown increasingly popular since. One theme of 4 . . . ttJc6 is that the b4 bishop can now react to a3 by . . . i.a5-b6, putting pressure on d4. Another is to attack the centre by . . .f6 after e5 is played.
Still a popular alternative to 4 e5, mainly because White feels that this is a safe line. The avoidance of dou bled c-pawns has its price, however, in that White confmes his own devel opment and allows his centre to be challenged quickly. Here I suggest the following: 8.41 4 ...ttJc6 8.42 4 ... dxe4 Yet another solution is the recently revived 4 . . . ttJf6 5 e5 (5 i.g5 dxe4 6 a3 i.e7 7 ttJg3 0-0 with the idea 8 ttJcxe4 ttJxe4 9 i.xe7 ii'xe7 10 ttJxe4 l:.d8) 5 . . . ttJfd7, e.g., 6 'ifd3 0-0 7 a3 i.e7 8 ii'g3 �h8 9 i.e3 c5 10 0-0-0 ttJc6 1 1 f4 b5! 12 ttJxb5 (12 dxc5 b4 1 3 axb4 ttJxb4 intending . . . 'i!Va5, . . . :b8 - Rus temov) 12 . . . .:.b8 13 ttJec3 (13 ttJbc3 c4!) 13 . . . c4! , and here R.Perez-Ruste mov, Villa de Albox 2002 went 1 4 .i.e2?! f5 1 5 'i¥f2 'i¥a5 1 6 ttJd6 i.a6 17 'i!Ve l i.xd6 1 8 exd6 ttJf6 19 i.f3 'i!Vb6
We look at: 8.4 1 1 5 e5 8.412 5 a3 (a) 5 exd5 exd5 just frees Black's bishop, for example, 6 g3 i.g4! 7 i.g2 ttJge7 8 0-0 'i!Vd7 9 a3 i.xc3 10 bxc3 ttJa5 1 1 l:.el 0-0 12 f3 .tf5; B.Jacobs Remlinger, Philadelphia 1995; (b) 5 'iVd3 dxe4 6 'i!Vxe4 ttJf6 7 'i¥h4 ttJe7 (or 7 . . . 0-0, e.g., 8 i.g5 i.xc3+ 9 bxc3 e5 10 i.xf6 'ii'xf6 1 1 'i!Vxf6 gxf6 12 0-0-0 lid8;) 8 i.g5 ttJed5 9 a3 .i.e7 is equal, Simon-Hofmair, Austrian Teams 1997. 1 39
Play the French 8.4 1 1 5 e5 This has proven popular. Previ ously I argued that it is less flexible than 5 a3 �a5 6 e5 of 8.412, but then again White may not care for the bishop going to a5 and b6. 5 ... f6 Now 6 a3 �a5 transposes to 5 a3. But White has other approaches: 6 exf6 (a) 6 liJf4 is similar to 5 a3 .lia5 6 e5 f6 7 liJf4, but here the bishop can come back to help the king: 6 . . .fxe5 (6 . . . g6!? 7 �b5 fxe5 8 liJd3! iLxc3+ 9 bxc3 e4 10 .i.xc6+ bxc6 1 1 liJe5 .i.a6! is unclear but probably better for White, who has good attacking pros pects) 7 iVh5+ with: (a1) 7 . . . 'it>f8 8 dxe5 (8 liJg6+? hxg6 9 'iVxh8 exd4 10 a3 .i.a5 1 1 b4 liJxb4 12 axb4 .i.xb4 13 iLa3 1iVe7- +) 8 ... d4 9 1iVf3 !? liJf6! (or 9 . . . 'it>e8 10 iVh5+ 'it>f8=) 10 exf6 'ii'xf6 1 1 .i.d2 dxc3 12 bxc3 .lid6 13 iLe2 e5 14 liJh5 'ii'xf3 15 �xf3 i.f5=; (a2) 7 ... g6 8 liJxg6 liJf6
Black has just enough initiative to counteract his weaknesses: (a2 1) 9 Wi'h4!? exd4 10 liJxh8 dxc3 1 1 b3 'it>e7 12 �b5 .td6 (12 . . . l:!.b8! 1 3 .txc6 bxc6 14 'ii'xf6+ 'litxf6 1 5 �g5+ 'it>xg5 16 liJf7+ 'it>f6 17 liJxd8 .td7) 1 3 .txc6 bxc6 14 'iWxf6+!? 'it>xf6 1 5 �g5+ 1 40
'it>xg5 1 6 liJf7+ 'it>f6 1 7 liJxd8 c5:j: 18 liJc6 .i. d7 1 9 liJa5 l:.b8 20 a3 (20 a4 c4) 20 . . Jlb5 21 b4 cxb4 22 axb4 .l:1.xb4 23 'it>e2 i.b5+ 24 'litd1 0 - 1 Barczay-Ditt mar, Oberwart 1995; (a22) 9 'ii'h 6 liJxd4 10 .lid3! l:!.g8! 1 1 liJxe5 �d6= ; (a23) 9 'iVh3!? liJxd4 10 �d3 e4! 1 1 liJxh8 e 5 1 2 iVh4 exd3 1 3 �g5 .txc3+ 14 bxc3 liJxc2+ and 15 'litfl liJxa1 16 .txf6 'iVd7 1 7 1iVh5+ will draw by repetition. White can try for more by 15 �d1 i.g4+ 16 f3 but he runs into 16 . . . d4! ! intending 17 .i.xf6? liJe3+, so a wild line goes 17 fxg4 dxc3 18 'iWh6 lLlxa1 19 .i.xf6 'iWd4 20 iVh5+ 'ii? d 7 2 1 'ii'f5+ 'ii? c 6 22 .txe5 'iWa4+ 23 'lit e 1 d2+ 24 'ii? e 2 'i!Vb5+ 25 'litd1 c2+ 26 �xd2 l:!.d8+ -+; (b) 6 f4?! liJh6 7 1iVd3 0-0 8 �e3 �d7 (or 8 . . . liJf5) 9 a3 (9 g3 .te8 has the idea . . . .tg6) 9 . . . .ta5 10 liJc1? (10 0-0-0 .i.e8!) 10 ... fxe5 11 fxe5 liJf5 12 liJb3 �b6 13 0-0-0 liJxe5 14 "iVd2 liJxe3 0-1 Brock-Hummel, Las Vegas 1999. 6 ...'iWxf6!? Quite as good seems 6 ... liJxf6, for example, 7 a3 �a5 (or 7 . . . �d6) 8 b4 �b6 9 liJa4 e5! 10 b5 .ta5+ 11 .td2 .i.xd2+ 12 'iWxd2 liJe4+ Ivanec-Kovace vic, Ljubljana 200 1 .
7 �f4 (a) 7 'iWd3 liJge7 8 �d2 0-0 9 f4 liJf5
Winawer Variation: Fourth Move Alternatives 10 a3 Ji.a5 1 1 Ji.e3 tL'lxe3 12 'i'xe3 Ji.d7 1 3 g3 (13 0-0-0 tL'le7) 1 3 . . . tL'le7! 14 Ji.g2 tL'lf5 1 5 'i'f2 (15 'i'd3 .i.e8 16 0-0-0 �g6) 1 5 ... Ji.b6! 1 6 O-O-O!? (16 ltdl .l:i.ac8 1 7 b4, when the moves 17 . . . a5 and 17 . . . tL'ld6!? both favour Black) 1 6 . . . .l::!. a c8! 1 7 g4 tL'ld6 1 8 "ifg3 tL'lb5 ! ? (18 . . . c5 may be better) 19 g5 tL'lxc3 20 'iWxc3 'i¥e7!? 2 1 .l::!.he l c5=1= Bhat-Hummel, Irvine 1997; (b) 7 .i.e3 tL'lge7 8 a3 �a5 9 b4 .tb6 10 tL'la4 0-0 1 1 tL'lxb6 axb6=1= Rodina Hacat, Ontario 2002; in most lines . . . e5 follows. 7 ...tL'lge7! The game Guliev-Lieb, Berlin 1998 saw 7 . . . �a5 8 'i'd2 tL'lge7 9 0-0-0 0-0= . After 7 . . . tL'lge7! , Herbrechtmeier Lieb, Bundesliga 1988 continued 8 i.xc7 0-0 9 f4 tL'lf5 10 'i'd3 'i'e7?! (10 . . . tL'lfxd4!=I=) 1 1 i.e5 tL'lxe5 12 fxe5 'i'g5 with good play for the pawn.
6 "if d3 is far less challenging after 6 . . . dxe4 7 'i'xe4 tL'lf6 (or 7 . . . tL'lge7 8 �g5 'i¥d5 !? 9 'iWxd5 exd5=) 8 'i'h4 O-O!? (8 . . . tL'le7) 9 �g5 �xc3+ 10 bxc3 h6 (10 . . . e5! 1 1 Itdl exd4 12 cxd4 .l::!. e 8 13 .txf6 "ifxf6 14 'i¥xf6 gxf6=) 1 1 i.xf6 'iWxf6= Barczay-McCambridge, Dort mund 1 982. 8.4121 6 e5 f6!? Black plays the most challenging move. But it's not the only satisfac tory one, e.g., 6 . . . �d7!? is an interest ing semi-waiting move and 6 . . . tL'lge7 intends . . . tL'lf5: 7 tL'lg3 (7 b4 i.b6 8 tL'la4 tL'lf5 9 tL'lxb6 axb6 10 c3 0-0 and . . .f6) 7 . . . 0-0 8 i.e2 i.b6! 9 �e3 tL'lf5?! (9 . . . tL'lxd4! 10 i.xd4 c5 11 i.e3 d4 12 i.d2 dxc3 1 3 i.xc3 tL'ld5=) 10 tL'lxf5 exf5 1 1 tL'la4 f6 12 tL'lxb6 axb6 1 3 f4 fxe5 1 4 dxe5 i.e6 1 5 c3 tL'la5 16 0-0;1;; Schramm-Kindermann, Dresden 1999.
8.412 5 a3 Ji.a5
While we've now had a good deal of experience with 4 . . . tL'lc6, the theory of this position is wide open. The play diverges here (with apologies for the mess) : 8.4 1 2 1 6 e5 8.4122 6 b4 8.4123 6 �e3
7 tL'lf4 A direct attacking move. White has these options: (a) 7 exf6?! 'i'xf6 (or 7 . . . tL'lxf6 and . . . 0-0 with some advantage) 8 i.e3 tL'lge7 9 'i'd2 0-0 10 f4 tL'lf5 11 i.f2 tL'ld6! 12 g3 Ji.d7!? 13 i.g2 i.e8 14 0-0 Ji.h5=1= Bolwerk-Pliester, Netherlands 200 1 ; (b) 7 f4 fxe5 (or 7 . . . tL'lh6!) 8 fxe5 1 41
Play the French liJh6 9 .i.xh6 'i!Vh4+ 10 g3 'li'xh6 1 1 'li'd2 'ii'x d2+ 1 2 �xd2 O-O=i= Medunova Szymanski, Olomouc 200 1 ; (c) 7 b4 SLb6 8 b5 liJa5 9 liJf4 fxe5 ! 10 'i!i'h5+ (10 dxe5 'iVg5) 1 O . . . g6 1 1 1!i'xe5 ( 1 1 liJxg6?! liJf6 1 2 'iWh3 l:.g8 1 3 dxe5 l:.xg6 14 exf6 'ii'xf6=i=) 1 1 . . :iVf6 12 1!i'xf6 liJxf6 1 3 .te3 c6 14 liJa4 i.c7 1 5 liJc5 e5=i= Heyken-Kindermann, Dort mund 1993. 7 ... g6!? A very interesting juncture; this position needs to be investigated more. Black's other options are 7 . . :fie7, 7 . . . fxe5, and 7 . . .1!i'd7, after which I gave 8 i..b 5 fxe5 9 'ii'h 5+ 'ii'f7 10 'ii'x e5 ttJge7 in the first edition, but here 10 i.xc6+ is better. 8 i.b5 8 exf6 'li'xf6 9 i..b 5 ttJge7 1 0 0-0 0-0 with a balanced struggle in prospect. 8 ... fxe5 9 1.xc6+ 9 dxe5 ttJge7 leads to a complex but balanced game, e.g., 10 0-0 0-0 1 1 b4 i.b6 12 ttJd3 liJd4 13 1.a4 SLd7 1 4 SL g 5 "iWe8= . After 9 1.xc6+, Olsson-Hector, Lin koping 200 1 continued 9 . . . bxc6 10 dxe5 c5 11 0-0 ttJe7 12 1!i'g4 ttJf5 1 3 ttJ d 3 i..b 6 14 1.g5 'iVd7 1 5 1.f6 0-0= .
As usual, White has several op tions here: (a) 7 e5 f6 8 b5 (8 liJa4 fxe5 9 liJxb6? axb6 10 dxe5 ttJxb4 Hoiberg Antonsen, Aarhus 1 997; 8 exf6 'iVxf6 9 1.e3 ttJge7=i=) 8 . . . liJa5 9 ttJa4 fxe5 10 ttJxb6 cxb6! 11 dxe5 1!i'c7=i= Podesta Giaccio, Buenos Aires 1995. Black has strong play along the c-file; (b) 7 i.e3 ttJf6 (7 . . . dxe4 8 ttJxe4 e5!? 9 d5 i.xe3) 8 e5 (8 f3?! dxe4 9 fxe4 e5!=i=) 8 . . . ttJg4 with a complex position in which Black looks fine, e.g., 9 ttJf4 (9 i.f4 f6=i=) 9 . . . ttJxe3 10 fxe3 0-0 1 1 1.d3 ( 1 1 1.b5 f6 1 2 i.xc6 bxc6 1 3 exf6 lIxf6 1 4 0-0 e5) 1 1 . . :ili'g5 12 'i!i'h5 (12 O-O? ttJxe5) 12 ... 1!i'xh5 1 3 ttJxh5 f6 14 exf6 g6! , playing for . . . e5; (c) 7 ttJa4 dxe4!? Now White must take time to recover his pawn, and by angling for . . . c6, Black will force White to exchange on b6 and then use his restraint of the centre to ad vantage.
8.4122 6 b4 1.b6
7 1.b2 1 42
White has two important replies: (c 1) 8 i.e3 ttJce7! 9 c4 c6 10 ttJxb6 axb6 1 1 ttJc3 (1 1 ttJg3 ttJf6 12 SLe2 liJf5) 1 1 . . .f5!? (1 1 . . .ttJf6=) 12 i.e2 ttJf6 13 0-0 0-0 14 f3 b5!? (14 . . . exf3 1 5 .txf3 ttJg6 16 'iVb3 'iWc7!=) 1 5 fxe4! ( 1 5 cxb5 ttJed5) 1 5 . . . bxc4 16 i.xc4 fxe4 (16 . . . liJxe4 17 ttJxe4 fxe4 18 l:.xf8+ 'ii'xf8=) 17 1.g5 �h8! 18 i.xf6?! (18
Winawer Variation: Fourth Move Alternatives iLb3 lDf5 1 9 lDxe4 lDxd4=) 18 . . . gxf6 19 lDxe4 lDf5, winning the d-pawn; (c2) 8 iLb2 lDf6! (8 . . . lDce7 9 c4 a5 !?) 9 c4 a5 10 b5 (10 lDxb6 cxb6=) 1O . . . lDe7 1 1 c5 iLa7 and Black's con trol of d5 ensures a decent game, be cause he can always play for . . . e5 to free his queen's bishop: 12 �d2 (12 �b3 only temporarily stops ... e3: 12 ... lDed5 1 3 lDg3 0-0 14 iLc4 c6 1 5 b6 iLb8 16 0-0 e3!+; 12 lDg3 lDed5 13 iLc4 e3!+; 12 iLc3 lDed5 1 3 iLxa5 iLxc5+) 12 . . . 0-0 13 lDg3 lDed5 (or 13 . . . iLd7 14 iLc4 c6) 14 iLe2 c6! 1 5 b6 iLb8 1 6 0-0 iLf4+ and . . . e3. 7 ... lDge7 7 . . . lDf6 has also been played, but 7 . . . lDh6 is the most interesting option: (a) 8 lDa4 �h4! 9 lDxb6 axb6!? 10 e5 (10 lDg3 dxe4 1 1 d5 exd5 12 iLxg7 Mg8 13 iLxh6 �xh6 14 �xd5 Mxa3!? 1 5 �xe4+ iLe6+) 10 .. .f6 11 exf6 0-0 12 lDg3 (12 g3 �e4 1 3 fxg7 Mxf2!) 12 . . . �xf6 13 �d2 lDe7= with the ideas . . . iLd7, . . . b5, . . . lDhf5-d6; (b) 8 �d3 0-0 9 lDa4 �h4! 10 exd5 (10 lDxb6 axb6 1 1 e5 f6 12 exf6 �xf6=) 10 . . . exd5 1 1 �g3 �xg3!? (1 1 . . .lDf5 12 �xh4 lDxh4=) 12 hxg3 Me8= . 8 �d3 (a) 8 g3 e5! 9 exd5 lDxd4 10 iLg2 iLg4+ Volokitin-Stellwagen, Gronin gen 1 999; (b) 8 lDa4 0-0 9 lDg3 (9 lDxb6 cxb6 10 lDg3 f5!? 1 1 e5 lDg6 ! , hitting f4 and thinking about . . . f4-f3) 9 . . . dxe4!? 10 lDxb6 axb6 1 1 lDxe4 lDf5 12 c3 e5!? 13 d5 (13 b5 lDa5 14 dxe5 �e7=) 13 . . . lDce7 14 c4 lDd6 15 �e2!? f5 16 lDxd6 cxd6 with mutual chances. 8 ... 0-0 9 0-0-0 Not 9 g3? e5!. 9 ...a5! 10 b5 lDa7 and Black will play . . . c6 to open up the c-file.
8.4123 6 iLe3 lDf6!? It's is not at all clear what is best here. This is the most popular move and has enjoyed success, but it is also more difficult to handle than the al ternative of putting the knight on e7 instead. (a) Here a neglected idea is 6 . . . dxe4 7 b4 iLb6 8 lDxe4 lDf6 9 lDxf6+ �xf6 10 c4 a6 1 1 c5 iLa7. This is quite in teresting: White's d5 is weak and he should do something before Black occupies it and unravels his pieces, say, 12 lDc3 0-0 13 �d2 (13 iLd3!? Md8 14 lDe4 �e7 1 5 lDg5 h6 16 iLh7+ 'it'h8 17 iLe4 lDxd4! unclear) 13 . . . e5!? 14 lDd5 �d8 1 5 dxe5 lDxe5 with coun terplay, e.g., 16 0-0-0 c6 17 �c3 iLg4! 18 f3 cxd5 19 �xe5 Me8 20 �d4 iLd7= ; (b) The safest move is 6 . . . lDge7: 7 e5 (7 g3 e5! 8 iLg2 iLg4 9 f3 iLh5 10 dxe5 dxe4 11 f4 lDf5 [or 1 1 . . .lDd5+] 12 iLc5 iLf3 1 3 0-0 Y2-Y2 Miladinovic Atalik, Elista 1 998; Black is slightly better after 1 3 . . . iLxc3 14 �xd8+ Mxd8 1 5 lDxc3 iLxg2 16 'it'xg2 e3)
7 . . . 0-0!? (Short's choice and not bad, but I prefer that Black take some ini tiative with 7 . . . lDf5, e.g. , 8 lDf4 0-0 9 iLb5 f6 10 exf6 �xf6! intending . . . e5, and anticipating 11 iLxc6 bxc6 12 b4? e5!; there's a lot to be analysed here) 8 1 43
Play the French lDg3 f6 9 f4? (9 exf6 lIxf6 10 �d3 h6 aiming for . . . e5 or . . . ..ixc3 and . . . lDa5 etc.) 9 . . . fxe5 (McDonald mentions the standard idea 9 . . . .td7 10 �d3 �e8 unclear) 10 fxe5 i.d7 11 �g4!? (White has to get castled. Maybe he should try 1 1 �d2, when HiArcs anticipates 0-0-0 by 1 1 . . .a6!? 12 0-0-0 lDa7!, in tending . . . lDb5 or . . . c5, e.g. , 13 �d3 c5 14 dxc5 lDec6 15 b4 �c7 16 �g5 �e8:j:) 1 1 . . .lDf5!? (1 1 . . .�b6! would prepare this) 12 lDxf5 exf5 13 �f3 (13 �f4 lDe7 is unclear) 13 . . . .lte6 14 .ltb5 hc3+ 15 bxc3 f4 16 .tf2 �d7! (threatening . . . l2Jxe5!) 17 O-O? (17 ..ie2 ! ; then maybe 17 .. .I1:Je7 18 0-0 b6, playing for . . . lDg6 and . . . c5) 17 . . . lDxd4 18 cxd4 �xb5+ E.Berg-Short, Malmo 2002; Black is a pawn up and his bishop has better prospects than White's. 7 e5 lDg4
Black has enjoyed much success here, but White may gain some edge by S lDf4! The alternatives are also signifi cant: (a) 8 i.d2 f6 (8 . . . 0-0 9 f3 lDh6 10 i.xh6 'i!Vh4+ 1 1 g3 'iYxh6= Van Mil B.Martin, Oakham 1 994) 9 h3 lDxf2 ! 10 'it>xf2 fxe5 with a strong attack for Black; (h) 8 i.f4 f6 9 lDg3 (9 h3? fxe5) 1 44
and now there are two very tactical ideas: (h 1) 9 . . . lDxf2!? 10 'iitxf2 ..ib6 11 'iit e 1 i.xd4 12 i.b5?! (12 'i¥h5+ g6 1 3 1iVh6 fxe5 14 i.d2 'iY'e7=) 12 . . . 0-0 13 exf6! (13 i.xc6 fxe5 is good for Black) 13 . . .'iY'xf6 14 :f1 i.xc3+!? 15 bxc3 �xc3+ 16 ..id2 �e5+ 17 i.e2!? �d7 18 :b 1 b6 19 lIb3, when McDonald likes White, but Black plays for the central advance . . . e5 by 19 . . . 'iY'd6 (or instead 19 . . . lIxf1 + and . . . �d6) ; (b2) 9 . . . fxe5!? seems like a good al ternative, since after 10 �xg4 exf4 White's best is probably to head for simplification by 1 1 iYxg7 (1 1 lDh5? .ltxc3+ 12 bxc3 g6; 11 �xf4 e5! 12 dxe5 ':f8) 1 1 . . .1::[f8 12 lDh5 (12 i.e2 �e7 1 3 �h5+ 'iit d 8 14 �xe7+ 'iitx e7 15 lDge2 �xc3+ 16 bxc3 'iit d 6:j: and . . . e5) 12 . . . �en or 12 . . . .:.f7!? 1 3 'iVg8+ r:J;; e 7 14 �g5+ 'iit d 7:j: ; (c) 8 b4 i.b6 9 lDa4 lDxe3 (9 . . :i!i'h4 has also been successful) 10 fxe3 f6 1 1 lDf4 0-0 12 lDxb6 axb6 13 exf6 �xf6 with a strong initiative, McDonald-B.Martin, London 1 994; (d) 8 .tc 1 ! ? 0-0 9 h3 lDh6 10 .i.xh6 gxh6 1 1 b4 i.b6 12 lDa4 f6 with good activity for Black, McDonald B.Martin, London 1994. S lDxe3 9 fxe3 O-O!? This move is certainly natural but ...
Winawer Variation: Fourth Move Alternatives as far as I know untried. Practice has revealed some serious but not neces sarily insoluble problems with the alternative 9 . . :iVh4+ 10 g3 'iVh6 1 1 'iVd3!? (1 1 b 4 i.b6 1 2 ttJa4! also keeps some advantage) 1 1 . . .i.d7 12 i.g2 i.b6 1 3 O-O-O;!; Illescas-Taddei, France 2000.
tion on either wing. Given the nature of the play, however, I think that af ter 6 i.e3, 6 . . . ttJge7 promises more positive play than 6 . . . ttJf6. This has been a long look at 4 ttJge2 ttJc6. It's holding up very well in theoretical terms, and as an extra benefit it offers room for experimen tation and development. 8.42 4 ... dxe4 5 a3 i.e7 6 ttJxe4 ttJf6
10 i.d3! Now 10 'iVf3? is pointless after 10 . . .£6, and 10 g3 f6 1 1 exf6 'iVxf6 12 i.g2 allows the interesting sequence 12 . . . i.xc3+ 13 bxc3 g5!? (13 . . . i.d7 14 0-0 'iVg5 1 5 'iVd3;!;) 1 4 ttJh3 i.d7 15 'iVh5 h6 and I see nothing wrong with Black's position. 10 . . f5 With this move Black accedes to letting White's e-pawn survive, but he stops White's attack and retains play on the queenside. 1 1 0-0 1 1 b4 i.b6 12 ttJa4 (12 O-O? ttJxd4) 12 . . . ttJxe5!? 13 dxe5 i.xe3 14 'iVf3 i.d4 wins a third pawn and leaves the situation unclear. 1 1 ... i.d7 White should have a small advan tage here; play might continue 12 �hl ttJe7 1 3 ttJce2 (13 g4 c5!) 1 3 . . . c6 14 b4 i.c7 1 5 c4!? dxc4 16 i.xc4 ttJd5 1 7 'iVd2 'iVe7;l; , but Black stands sol idly enough and can contemplate ac.
Some will like this clear solution. 6 . . .ttJf6 resembles the . . . dxe4 lines of the Classical variation in Chapter 14, and those who are familiar with the order 3 ttJc3 dxe4 4 ttJxe4 ttJd7 5 ttJf3 ttJgf6 followed by . . . i.e7 will realise that White has lost more than he has gained by comparison, since putting a knight on g3 or c3 in two moves can hardly be as good as having the knight on f3 in one. 7 ttJ2g3 This has done the best recently. I don't think that any move is particu larly dangerous versus natural play: (a) 7 ttJxf6+ i.xf6 8 i.e3 (the slow 8 c3 can be met in several ways including the immediate 8 . . e5= . Also possible is 8 . . . ttJd7 with . . . c5 or in some cases . . . e5 next) 8 . . 0-0 (8 . . . c5!? 9 dxc5 'iVxdl + 10 z:txdl i.xb2= was Ambroz Uhlmann, Trencianske Teplice 1979) .
.
1 45
Play the French 9 'it'd2 e5!? (9 . . . a6!? 10 0-0-0 b5) 10 0-0-0 l2Jc6= was Kurajica-Petrosian, Zagreb 1 970, when instead of the strange 1 1 f3?, White could have played 1 1 d5 l2Je7 12 l2Jc3 l2Jf5 1 3 'i£?b l = ; (b) 7 l2Jg5?! c5 ( 7 . . . e5=) 8 l2J f3 l2Jc6 9 g3 cxd4?! (9 . . . e5!+) 10 l2Jexd4 l2Jxd4 1 1 'it'xd4 'i!Vxd4 1 2 l2Jxd4 Moreda Lageyre, Clermont Ferrand 200 1 ; and Black would still have some advan tage after 12 . . . �d7 13 Sl,g2 e5 1 4 l2Je2 0-0-0; (c) 7 l2J2c3 l2Jc6!? (7 ... 0-0 8 �c4 l2Jc6= is easier) 8 �e3 (8 �b5 l2Jxe4!? 9 �xc6+ bxc6 10 l2Jxe4 �a6 is un clear) 8 . . . 0-0 (or 8 . . . l2Jxe4 9 l2Jxe4 e5= - Euwe; then possible is 10 d5 l2Jd4!? 11 �c4! l2Jf5 12 �d2 l2Jd6 1 3 l2Jxd6+ cxd6 14 0-0 0-0 with .. .f5 to come) 9 'it'd2 (9 �e2 l2Jxe4 10 l2Jxe4 was Lasker-Capablanca, Moscow 1935, and Black missed the chance for 10 . . .f5! 1 1 l2Jc3 f4, picking up the d pawn at no cost) 9 . . . a6!? 10 �e2 ! b 5 1 1 �f3 �b7 12 g4?! l2Jxe4 1 3 l2Jxe4 Auwerswald-Fritz, corr 1989, and best was 13 . . . e5 14 d5 l2Jd4 ! ; (d) 7 'iWd 3 i s a popular move i n this type of position, preparing quick cas tling on the queenside. Then 7 . . . a6!? and ... b5 is of interest, while 7 . . . l2Jbd7 is a flexible move that has enjoyed some success, e.g., 8 �f4 l2Jxe4 9 'i!Vxe4 l2Jf6 10 'i!Vf3!? (10 'i!Vd3 0-0 1 1 0-0-0 �d7 with the idea . . . �c6, but also . . . b5/ . . . a4/ . . . b4) 10 . . . 0-0 1 1 g3?! �d7! 12 .l:!.gl?! (but 12 'i!Vxb7 .l:!.b8 1 3 1i'xa7 .l:!.xb2 is depressing) 12 . . . �c6 1 3 'ifb3 'iVd7 14 i.. g2 �xg2 1 5 .l:!.xg2 c5 1 6 f3!? 'il'c6 1 7 c4 .l:!.ac8+ McDonald-Pert, Telford 2003. 7 0-0 Similar to the main line is 7 . . . l2Jbd7 8 i.. d3 c5 9 dxc5 l2Jxc5 1 0 l2Jxc5 .i.xc5 1 1 0-0 'i!Vc7!? 12 'iVe2 �d7 intending ••.
1 46
. . . iLc6 and perhaps . . . 0-0-0, Llaneza Vega-Foisor, Istanbul 2003. 8 c3 l2Jbd7 This is a rare but solid move, con templating forthcoming breaks in the centre. 9 iLd3
9 c5! The break that corresponds to the Classical Variation. 10 dxc5? Better is 10 l2Jxf6+ l2Jxf6 11 dxc5 .i.xc5= - Rogers. Then Black has a 43 central pawn majority vs 3-2 on the queenside, which is always an inter esting imbalance. 10 ...l2Jxe4 11 l2Jxe4 f5! 1 2 l2Jg3 l2Jxc5 13 i.. c 4 1i'xd1+ 14 'it>xd1 iLd7+ Black intends . . .iLa4+ and . . . l:tad8, while . . . b5 and . . . l2Jd3/b3 is another plan, e.g., 15 i.. f4 b5 16 i.. a2 l2Jd3 1 7 �e3 f4 18 'i£?c2 fxe3 1 9 'it>xd3 .l:!.xf2 and Black is winning. ...
8.5 4 �d3 This lacks punch and is seldom seen at the top levels any more. I ha ven't revised much here, but there are people who play this regularly and you still need to know some lines. 4 ... dxe4 4 . . . c5 is also equal: 5 exd5 (5 a3 �xc3+ 6 bxc3 c4 7 i.. e 2 l2Jf6 8 e5 l2Je4
Winawer Variation: Fourth Move Alternatives 9 �b2 t'bc6=i= Akhmedov-Fatullaev, Baku 2000) 5 . . . exd5 6 dxc5 (6 t'bf3 c4! 7 �e2 t'be7=i= ; 6 i..b 5+ t'bc6 7 t'bge2 can be answered by 7 . . . c4 or 7 . . .t'bge7 8 a3 i.xc3+ 9 t'bxc3 0-0= with the idea 10 dxc5?? d4) 6 . . . t'bc6 7 t'bf3 �xc5 8 0-0 t'bge7 9 h3 0-0 10 a3 f6!? (a standard French idea, to support a knight on e5; but Black could also play simply by 10 . . . a6 or 10 . . . h6 1 1 �f4 �e6) 1 1 lle l t'be5!? 12 t'bxe5 fxe5 1 3 lIxe5 'ilVd6!? 14 J:!.e2 (14 'i!Vh5 g6 15 Wlg5 i.xf2+ 16 'it>hl i.h4 ! ; 14 �e3!?) 14 . . . �xh3 1 5 t'be4! 'ilVg6! 16 t'bg5 �f5=i= Sariego-Paneque, Pinar del Rio 1990. 5 i.. x e4
9 . . . 0-0 10 0-0 cxd4! 1 1 cxd4 e5) 10 i.g5 (10 'ifd3 cxd4 11 cxd4 exd4! =i= , since White will have t o play 1 2 i.xc6+ and lose the light squares)
1O . . . exd4 (or 10 . . . cxd4 1 1 O-O!? dxc3 12 Wixd8+ t'bxd8 13 t'bxc3 ..te6=) 1 1 cxd4 (1 1 i.xc6+ bxc6 1 2 cxd4 cxd4 1 3 Wixd4 Wia5+ 14 ..td2 'iVd5= Hort Pietzsch, Kecskemet 1964) 1 1 . . .h6 12 i.xf6 Wixf6 1 3 c3?! (13 dxc5 0-0 14 0-0 lId8=) 1 3 ... 0-0 14 0-0 l1d8 1 5 'ilVa4 cxd4! (this looks better than 1 5 . . . i.d7!?=i= Ortega-Uhlmann, Polani ca Zdroj 1967) 16 i.. xc6 bxc6 17 cxd4 (17 t'bxd4 c5!) 17 . . . c5! 18 dxc5 i.a6. 5 t'bf6 A good alternative is the immedi ate 5 . . . c5 (compare the main game), e.g., 6 a3 i..x c3+ 7 bxc3 t'bf6 8 i.f3 (8 i.d3 'ilVc7! 9 dxc5 t'bbd7 10 t'be2 t'bxc5 1 1 f4 ..td7 12 0-0 t'ba4!=i= intending . . . 'ilVc5+ and . . . 0-0 Bungo-E.Pedersen, corr 1 986; or here 8 . . . t'bbd7 9 t'bf3 'ilVc7! 1 0 0-0 c4 1 1 ..te2 t'bd5+ as in the game Barlov-Sahovic, Vrnjacka Ban ja 1 984) 8 . . . t'bc6 (8 . . . 'ilVc7 is a useful way to avoid simplification, for exam ple, 9 dxc5?! 'ilVxc5 10 'ilVd4 Petkovic Gavric, Yugoslavia 199 1 , and instead of 10 . . . 'ilVxd4 1 1 cxd4 t'bc6= , 1O . . . t'bbd7 1 1 i.e3 Wic7 would have given better chances for advantage) 9 t'be2 e5 (or ...
6 �f3 (a) 6 �d3 c5 7 t'bf3 is Djuric's spe cialty, although it seems harmless (others: 7 a3 Jixc3+ 8 bxc3 Wic7= with the idea . . . c4; 7 dxc5 t'bbd7 8 i.d2 147
Play the French lZ'lxc5 9 ..tb5+ .i.d7 is perfectly good and even gives Black a lead in devel opment) 7 . . . 0-0 (or 7 . . . cxd4! 8 lZ'lxd4 e5 9 lZ'lde2 0-0 10 0-0 ..tg4 11 f3 Djuric Drasko, Niksic 1 996; 1 1 . . . ..te6 12 a3 ..te7 'looks very solid for Black' Pedersen; I agree, and the kingside 43 majority may come in handy later) 8 0-0 (8 a3 ..txc3+ 9 bxc3 Djuric Prata, Lisbon 1 999, and McDonald recommends 9 . . . lZ'lbd7 10 0-0 'iVc7) 8 . . . lZ'lc6 9 dxc5 �xc3 10 bxc3 e5 (10 ... 'tlfa5) 1 1 lZ'lg5 !? (1 1 i.g5) 1 1 . . .'tlfa5 12 �e3 'ikxc3 with complex and bal anced play, Roitburd-S.Bjerke, Oro pes a del Mar 2000; (b) 6 .i.g5 is an innocuous line from the MacCutcheon French. Black can play, for example: 6 . . . c5! (6 . . . lZ'lbd7 is also good) 7 lZ'lf3 (7 dxc5 'iVxdl + 8 ':'xdl lZ'lbd7 9 �f3 �xc3+ 10 bxc3 lZ'lxc5= intending 1 1 .i.e3 lZ'lfd7) 7 . . . cxd4 8 lZ'lxd4 (8 ..txf6 'iVxf6 9 'i'xd4 'iVxd4 10 lZ'lxd4 ..txc3+ 1 1 bxc3 lZ'ld7=l=) 8 . . .i.xc3+ 9 bxc3 'ii'a5 10 .i.xf6 'iWxc3+ 1 1 'i'd2 'i'xd2+! 12 'ittxd2 gxf6:j: Honfi Lengyel, Budapest 1957. 6 c5 7 lZ'lge2 7 a3 ..txc3+ 8 bxc3 is 5 . . . c5 above, and 7 .i.e3 is easily met by 7 . . . cxd4 8 'iWxd4 'i'xd4 9 i.xd4 lZ'lc6= , or by 7 . . . lZ'ld5 8 .i.d2 .i.xc3 9 bxc3 lZ'lc6. 7 lZ'lc6 It seems that liquidating the centre is almost always reasonable in this line, e.g., 7 . . . cxd4 8 'iVxd4 (8 lZ'lxd4?! e5 9 lZ'ldb5 a6 10 'i'xd8+ 'ittxd8) 8 . . . 'i'xd4 9 lZ'lxd4 a6 10 0-0 lZ'lbd7 1 1 lZ'la4!? 0-0 12 a 3 i.d6 1 3 c 4 i.c7 1 4 ..tg5 lZ'le5!? 1 5 i.e2!? (15 i.xf6 gxf6 1 6 i.e2 b6=) 1 5 . . . lZ'le4 16 i.e3 i.d7 1 7 f3 ..txa4 1 8 fxe4 lHd8 with a positional advantage (the e5 square), Gra barczyk-Gdanski, Polish Ch 199 1 . 8 � e 3 cxd4 9 lZ'lxd4 Best may be 9 i.xc6+!? bxc6 10 .•.
...
1 48
'i'xd4, when Black should probably opt for 10 . . . 'i'xd4 1 1 ..txd4 i.a6= . After 9 lZ'lxd4, the game Fichtl Uhlmann, Zinnowitz 1 966 went 9 . . . lZ'le5 (9 . . . �xc3+ 10 bxc3 lZ'le5=) 10 i.e2 (10 0-0 lZ'lxf3+ 11 'tlfxf3 i.xc3 12 bxc3 0-0 13 c4 'iIIc 7 1 4 ii.g5 lZ'ld7 15 'i'e2 a6 16 ':'adl lZ'lc5 with a solid po sitional advantage, Kallio-Kristj an sson, Gausdal 2003) 10 . . . lZ'ld5 1 1 �d2 lZ'lxc3 12 bxc3 �e7 1 3 0-0 0-0 14 f4 (14 l:!e l 'i'd5!?:j:) 14 ... lZ'ld7 1 5 i.d3 lZ'lc5 16 'i'f3 i.d7 1 7 liae l lZ'lxd3 (17 ... 'i'a5 is a good alternative, but Black wants the two bishops) 18 cxd3 'iWc7=l= . Obviously White can't be worse after 4 ..td3, but Black can make it complicated. 8.6 4 exd5 This is another version of the Ex change Variation (Chapter 5). One of the fIrst lines played against the Winawer in the early part of the cen tury, it has been somewhat revived by new and aggressive ideas. In our main line (Black's most ambitious strategy), White can get dangerous play by winning the two bishops at an early stage. The price he pays for this is surrender of the light squares, no tably c4 and e4. 4 ... exd5 Now White has two main moves: 8.61 5 'iWf3 8.62 5 i.d3 (a) 5 a3 ii.xc3+ 6 bxc3 lZ'le7 7 lZ'lf3 (7 i.. d 3 i.. f5=) 7 . . . 0-0 (or 7 . . .lZ'lbc6 8 .i.d3 i.. g 4!) 8 i.. d 3 .i.f5 9 0-0 lZ'lbc6 10 �f4 Treybal-Nimzowitsch, Carlsbad 1923; 10 ... lZ'la5:j: - Nimzowitsch, or instead 10 . . . 'i'd7; (b) 5 lZ'lf3 lZ'le7 (a snippet from his tory is 5 . . . lZ'lc6 6 ..td3?! .i.g4:j: Lasker Botvinnik, Moscow 1936; 5 . . . ..tg4 6 h3
Winawer Variation: Fourth Move Alternatives �h5 is also good) 6 �d3 tbbc6 7 h3 �f5 8 �xf5 tbxf5 9 0-0 i.xc3 10 bxc3 0-0 1 1 'iWd3 tbd6:j: (light squares) Mannheimer-Nimzowitsch, Frankfurt 1930.
squares and good compensation) 12 i.g5 (12 tbg3?? �g4!) 12 . . . �d7 13 tbf4 g6 14 'ii'h4 tbf5 1 5 �f2 h6 16 .i.f6 0-0 1 7 i.d3 �d6:j:. 8.62 5 �d3
8.61 5 'iff3 Used by Larsen in the old days, this prevents doubled pawns after a capture on c3 and intends 5 . . . tbf6 6 ..ig5. But the White queen can be a target. 5 J/fe7+ 5 . . . tbc6 is also fine, e.g., 6 i.b5 tbge7 7 �f4 0-0 8 0-0-0 i.e6 9 tbge2 a6 10 i.d3 �d7 1 1 h3 (1 1 'iWh5 tbf5 12 'iVh3 g6:j:) 1 1 . . .�xc3! 12 bxc3 (12 tbxc3 tbxd4 13 �xh7+ �xh7 14 ':'xd4 c5) 12 . . . tba5 13 g4 tbc4+ Ghizdavu-Kap engut, Orebro 1 966. 6 tbge2 (a) 6 iLe3?! tbf6 (6 . . . tbc6 7 tbge2 is the main line) 7 iLd3 (7 a3 i.g4 8 �g3 i.d6 9 �4 i.f4 10 tbxd5 tbxd5 1 1 �xg4 0-0+) 7 . . . c5! 8 �f1 i.xc3 9 bxc3 c4 10 ii.f5 ii.xf5 1 1 'iWxf5 0-0+ Mestro vic-Maric, Kraljevo 1967; (b) 6 �e3 is passive: 6 ... tbc6 7 tbf3 (7 ii.b5 iLf5 8 'iWxe7+ tbgxe7 9 �f4 i.xc2 10 ii.xc7 ':c8:j: Saharov-Antosh in, Sochi 1 966) 7 . . . �f5 8 i.d3 �xd3 9 cxd3 0-0-0 10 0-0 Ajala-I.Farago, Har rachov 1967, and 1O . . . �xe3! 1 1 ii.xe3 tbge7 12 a3 ii.d6 13 tbb5 a6 14 tbxd6+ :xd6:j: ; the knights are better than the bishop in this kind of position. 6 tbc6 7 ii.e3 Minev gives 7 �xd5 tbf6 8 �c4 .ie6 9 'ifd3 0-0-0 10 �e3. Then best seems 10 . . . tbg4, e.g. 1 1 a3 tbce5! 12 'lWe4 f5 13 'iWf4 tbg6:j: . 7 tbf6 8 a3 i.xc3+ 9 bxc3 tbe4 with good play. Biyiasis-J.Watson, Vancouver 1977 went 10 �h5 i.e6 1 1 f3 tbd6 (1 1 . . .tbf6 12 'iVh4 �f5 1 3 �g5 0-0-0 14 ii.xf6 gxf6 with a lot of ..
Developing logically. Now Black has a choice of 5 valid moves! I'll make two of them main lines: 8.62 1 5 tbc6 8.622 5 ...tbf6 •••
Also playable are 5 . . . c5 and 5 . . . c6, whereas 5 . . . tbe7 is similar to (and often transposes into) 5 . . . tbc6. 8.621 5 ... tbc6 'The most precise equalizing line' Psakhis.
••.
.••
1 49
Play the French 6 a3 Only this sharp move challenges Black: (a) 6 ttJf3 i.g4 7 .i.e2 (7 .i.e3 ttJf6 8 a3 J..xc3+ 9 bxc3 ttJe4:J:) 7 . . . ttJge7 8 0-0 Wid7 9 iLg5 f6 10 .te3 0-0-0 11 h3 .i.e6!? 12 a3 .txc3 13 bxc3 g5! and Black's attack was well underway in G.Phillips-Barsov, Metz 2002; (b) 6 ttJge2 ttJge7 (6 . . . "Yib.4!? 7 J.. e 3 ttJge7 8 iVd2 h6 9 a3 .txc3 10 Wixc3 J.. f5 1 1 .tf4 .txd3 12 Yi'xd3 0-0-0= Czerniak-Ivkov, Eersel 1966) 7 0-0 (7 J.. g 5 f6 8 .tf4 .tg4=) 7 . . . .tg4 (or 7 . . . i.f5=) 8 h3 (8 f3 .i.f5 9 a3 i.xc3 10 bxc3 'iWd7 11 c4 .i.xd3 12 'iWxd3 dxc4 1 3 'iWxc4 0-0 14 .i.b2 ttJd5 15 'iWd3 b5! 16 llfe1 ttJa5!? with doubled-edged play, Sepman-Shchukin, St Peters burg 2000) 8 . . . .i.h5 9 a3 .i.d6 10 ttJb5 0-0 1 1 c3 lIe8 12 Yi'c2 J.. g6 1 3 ttJxd6 'iWxd6 14 .i.f4 'iWd7= O.Bernstein Bronstein, Paris 1954. 6 ... .i.xc3+ 6 . . . .ta5 has been used by Apicella and Uhlmann and is undoubtedly playable. The capture is thematic. 7 bxc3
7 ... .i.e6!? I'm going to recommend this slightly offbeat but currently ac cepted move which maintains flexibil ity (sometimes bishops or queens are 1 50
better developed before knights, as on move 3!). Then 'i!Vh5 is met by . . . ttJf6, and Black can also play a quicker . . . 0-0-0 in several lines. 7 . . .ttJf6 has also been used to achieve equal chances . More interesting is the older 7 . . . ttJge7, which is still unresolved. Here's a fairly up-to-date example, with thanks to Neil McDonald: 8 'i!Vh5 .te6 9 ttJf3 'itd7 10 ttJg5 0-0-0 1 1 0-0 (1 1 ttJxf7? 'ire8; 11 ttJxe6 'itxe6+ 12 .te3 g6=) 1 1 . . . .tg4! 12 'itxf7 h6 13 f3 lIdf8 14 fxg4 (14 'itxg7 .tf5! 1 5 ttJf7 :hg8 16 ttJe5 .l:.xg7 1 7 ttJxd7 lIfg8! 18 .i.xf5 llxg2+ 1 9 �h1 ttJxf5 20 ttJf6 118g7 2 1 ttJxd5 l:i.xc2 with an unclear position - Knaak) 14 . . J::txf7 15 ttJxf7 l:i.e8 1 6 g5 hxg5 (16 . . . h5!) 17 .i.xg5 �b8 18 lIab 1 �a8 19 libel? (19 �h1 - McDonald) 19 . . . a6 20 ttJe5 ttJxe5 2 1 dxe5 'iWc6+ Brendel-Yusupov, Stock holm 2002. 8 'iWf3 (a) 8 lIb 1 gives Black time to make a useful move: 8 . . . b6 (8 . . . ttJa5!?; 8 ... 'iWe7 9 ttJe2 0-0-0 10 0-0 ttJf6 1 1 :e1 :he8 1 2 ttJf4 'iWd6 1 3 ttJxe6 fxe6, with an unclear knights-v-bishops imbalance) 9 ttJe2 'iWd7 10 0-0 ttJge7 1 1 ttJf4 .i.f5 12 ttJh5 0-0 13 h3 (13 i.h6? i.g4!+) Nataf-Rustemov, Stock holm 2002; and although the game was shortly drawn after 13 .. .'it>h8, a more interesting move seems to be 13 . . . .i.g6!?, e.g., 14 'iWf3 ttJa5! 1 5 �e1 ttJc4!? (15 ... 11ae8 keeps more pieces on the board) 16 .i.xc4 dxc4 1 7 'iVf6!? i.xh5 1 8 'iWxe7 'iWxe7 1 9 .l:txe7 c6 20 i.f4 l:.fe8 2 1 :be 1 f6= ; (b) 8 ttJf3 'iWd7 9 ttJg5 0-0-0 10 0-0 .tf5=; (c) 8 "Yib.5 ttJf6 (8 ... 'iVe7 9 ttJe2 0-0-0=) 9 'i!Vh4 ttJe4 10 'iixd8+ l:!.xd8 1 1 ttJe2 0-0 1 2 .tf4 ttJd6= . 8 ...'iVd7
Winawer Variation: Fourth Move Alternatives Or S . . . lLlge7 9 lLle2 �d7 10 lLlg3!? i.. g 4 11 �e3 0-0 12 0-0 .l:1.aeS Haus mann-S.Gross, Litomysl 1997; 13 f3! .te6= . 9 lLle2 9 Ilb l 0-0-0 (9 ... lLla5!? is as usual a reasonable option) 10 lLle2 lLlf6!? 1 1 i.g5 (1 1 0-0 lLle4!) 1 1 . . .�g4 12 �f4 (12 �g3 l1heS) 12 . . . lLlh5 13 �d2 f6 14 i.e3 g5 - compare the note to White's 1 1th. 9 0-0-0 10 0-0 10 'iVg3 f6 helps Black's kingside attack unless White moves quickly by 1 1 lLlf4 g5 (1 1 . . .lLlh6) 12 lLlxe6 �xe6+ 13 i.. e 3 lLlge7 14 0-0 h5 15 f4 h4 with unclear prospects. 10 lLlf6 1 1 h3 1 1 �g5 i.g4 12 'ifif4 lLlh5 13 'ifid2 f6 14 .te3 g5! intending . . . lLlg7 and . . . lLlf5 or . . . h5. 11 lLle4!? 12 c4! 12 i.xe4 dxe4 1 3 'ifixe4 .:i.heS with good light square pressure. Mer 12 c4, Pinol-Bartel, Balatonlelle 2002 went 12 .. J:theS! 13 cxd5? (13 i.. e 3! f5 14 cxd5 i.. x d5 1 5 c4 i.. f7 16 I1ab l g5 17 l:i.fe l h5 is obscure) 13 . . . ..txd5 14 c4?? (14 ft5 ! lLlf6 1 5 'ifig5 h6 16 �g3 g5! with initiative) 14 . . . lLlc3! 15 �f4 (15 cxd5 lLlxe2+ 1 6 �xe2 lLlxd4) 1 5 . . . lLlxe2+?! (15 . . . ii.e4!-+) 1 6 i.xe2 ':xe2 17 cxd5 'ifixd5+ .
'ifie2+ i.. e 6 S i.h4 g5! 9 i.g3 c5!) 7 ... 0-0 S a3?! (S lLlge2 transposes to 6 . . . 0-0) S . . . l:i.eS+ 9 lLlge2 ..txc3+ 10 bxc3 c5! 1 1 dxc5 lLlbd7 12 0-0 lLlxc5 1 3 lLld4 g 5 ! 14 ii.g3 lLlce4=1= .
...
.•.
...
8.622 5 ... lLlf6 This has been played a lot recently and looks perfectly fine: 6 ..tg5 The most direct move. 6 lLlf3 comes out rather poorly after 6 .. :ilie7+! 7 'iVe2 (7 i.e3 lLle4!=I=) 7 . . . 'iVxe2+ S �xe2 0-0 and Black may even be slightly better, e.g. , 9 i.f4 l:i.eS+ 10 �d2 c5. 6 0-0 Also effective is the immediate 6 . . . h6 (often transposing) 7 i.h4 (7 ...
7 lLlge2 Consistent with White's set-up . The options are worse: (a) 7 a3?! i.. x c3+ S bxc3 �eS+! 9 lLle2 (9 'ifie2 'ific6 10 'iVd2 .l:!.eS+ l 1 lLle2 lLle4 12 ..txe4 l:i.xe4 13 0-0 f6=1=) 9 . . . lLle4 10 i.d2 (10 i.f4 lLlxc3 11 'ifid2 lLlxe2 12 i.xe2 'i'c6=1=) 10 . . . lLld7 1 1 0-0 lLlb6 12 lLlg3 (12 i.. f4 'iVc6=) 12 . . . �c6 1 3 'iVh5?! f5 14 i.xe4 dxe4 1 5 f3 e3!-+ Druckenthaner-Kindermann, Austri an Teams 2002; (b) 7 'i'd2? c5! is already better for Black: S a3 (S dxc5 d4 9 0-0-0 dxc3 10 i.xh7+ �xh7 11 �xdS l:i.xdS 12 l:i.xdS cxb2+ 13 �xb2 lLlbd7+ ; S lLlf3 'iVeS+! 9 �e2 c4+) S . . . i.a5 9 lLlf3 'i'eS+ 10 i.. e 3 c4 11 b4 cxd3 12 bxa5 dxc2 13 'iVxc2 lLlc6 14 0-0 .tg4! 1 5 lLld2 lLlxa5+ . 7 h6 8 i.h4 S .tf4 c5!? 9 dxc5 lLlc6 10 0-0 i.xc5 l 1 lLla4 i.d6 12 'iVd2 i.e6= . 8 c5!? S . . . c6 is solid, e.g. , 9 0-0 lLlbd7 10 a3 i.d6 11 i.g3 lLlb6 12 i.xd6 'iVxd6= Rechel-Eingorn, Metz 1 995. Mter S . . . c5, Skripchenko-Barua, Groningen 1997 continued 9 dxc5 ...
...
1 51
Play the French liJc6 10 0-0 .txc5 1 1 liJxd5!? g5 12 liJxf6+ 'i'xf6 13 .tg3 'i'xb2 14 l::tb 1 'iWf6= � - Y2. 8.7 4 'iWd3
White covers e4 and would like to transfer the queen to pressure the kingside. In addition, he can prepare 0-0-0 by, say, �d2. 4 . . liJ e 7 White's main idea is 4 . . . dxe4 5 'i'xe4 liJf6 6 'iVh4, which may be fine for Black but isn't worth contesting. The text recognizes that the queen is not well placed on d3, and thus doesn't need to be forced to a better spot. Another solution is 4 . . . liJc6 5 e5 (5 liJf3 liJf6 6 e5 liJe4 7 a3?! .txc3+ 8 bxc3 f6! 9 exf6 Jerez Perez-Romero Holmes, Cala Galdana 200 1 , and al though 9 . . . gxf6!? was fascinating and roughly balanced, 9 . . . 'iWxf6 10 iLe2 0-0 would have favoured Black, who can develop by . . . b6, . . . �b7 and . . . e5, while other plans such as ... a5/ .. .ta6 or ... .td7-e8-g6/h5 are also possible) 5 . . .f6 6 a3 ..ta5 7 b4 .tb6 8 liJf3 fxe5 9 dxe5 liJge7 1 0 ..tb2 (10 .tg5 h6 1 1 .th4 0-0 1 2 iLe2 'i'e8 1 3 b 5 liJf5! 14 bxc6 liJxh4=t Pascual Arevalo-Romero Holmes, Spain 1 990) 10 . . . 0-0 (10 . . . a6! 11 liJa4 ..ta7=i=) 11 liJa4 �d7 12 liJxb6 .
1 52
axb6= intending . . . liJg6, Muench Reimer, corr 1 998. 5 iLd2 Thematically preparing 0-0-0. Oth ers: (a) 5 liJge2 is the popular move.
Now Black can play 5 . . . liJbc6 (re sembling 4 liJe2 liJc6), 5 . . . b6, or: (a1) 5 . . . liJd7 6 a3 iLa5 (or 6 . . . dxe4 7 'i'xe4 liJf6=) 7 b4 iLb6 (White has stopped . . . c5, but now has to face . . . e5) 8 liJg3 a5 9 l:.b 1 axb4 10 axb4 0-0 1 1 .te2 (1 1 e5 f6 12 f4 fxe5 1 3 fxe5 liJc6 14 liJce2 'i'h4 intending . . . liJdxe5) 1 1 . . .e5! 12 liJxd5 liJxd5 1 3 exd5 exd4 14 0 - 0 liJe5=t Chevallier Luce, Torcy 1991; (a2) 5 ... c5 6 .tg5 (harmless is 6 dxc5 .txc5 7 exd5 exd5 8 .te3 ..txe3 9 'i'xe3 0-0 10 0-0-0 liJbc6 with good prospects on the queenside) 6 . . .f6 7 .td2 liJbc6!? (or 7 . . . 0-0!?, e.g., 8 a3 cxd4 9 axb4 dxc3 10 iLxc3 e5=) 8 a3 .txc3 9 .txc3 Schmittdiel-Jolles, Groningen 1 990. To counteract the bishops, Black should now win the centre by 9 . . . cxd4 10 liJxd4 e5 1 1 liJxc6 bxc6= ; (b) 5 .tg5 0-0 (5 . . . f6 is also fine) 6 liJf3 .txc3+ (6 . . . f6 7 .td2 b6 8 O-O-O;!; Vaganian; but here 7 . . . c5! improves, with good play) 7 bxc3 f6 8 ..td2 b6 9 'iWe3 .tb7 10 .td3 (10 e5!? liJf5 1 1 'iVf4
-
Winawer Variation: Fourth Move Alternatives ltJc6 - Vaganian) 10 . . . dxe4 1 1 �xe4 ltJf5! 12 'iVe2 ..txe4 13 'i¥xe4 "iWd5:j: Smirin-Vaganian, Naberezhnye Chel ny 1 988.
5 ... c5 5 . . . b6 has been played in several games, but a more common order for this idea is 5 . . . 0-0 and: (a) 6 0-0-0 c5 (or 6 . . . b6 7 ltJh3 �a6 Y2 - Y2 Meszaros-Lahlum, Kecskemet 2003, due to 8 'i¥f3 ..txfl 9 l1hxfl �xc3 10 ..txc3 dxe4 1 1 'iVxe4 'iWd5= ; 9 . . . ltJbc6!?) 7 dxc5 ltJa6!? 8 a3 (8 'iWf3 ltJxc5 9 exd5 exd5 10 ltJxd5 ..txd2+ 1 1 l::txd2 ltJxd5 unclear) 8 . . . ..txc5 9 ..te3 'iVa5= Khemelnitsky-Shabalov, Phila delphia 1993; (b) 6 a3 ..txc3 7 i..x c3 b6! 8 0-0-0 (8 ltJf3 a5 9 0-0-0 �a6=) 8 ... ..ta6 9 'lWf3 (9 'iVe3 ..txfl 10 lixfl a5! 1 1 ltJf3 ltJbc6 12 ..td2 b5!= Jerez Perez-Moskalenko, Barbera 1 999) 9 . . . ..txfl 1 0 .l:f.xfl ltJbc6 1 1 ltJe2 a5 12 a4?! 'iVd7! 13 ltJf4 dxe4 14 'i¥xe4 ltJb4!+ 0- 1 Levitt-Nogueiras, Bled 2002. 6 dxc5 ltJbc6 7 0-0-0 0-0
7 . . . �xc5 is probably a more accu rate move order. 8 ltJf3 White could try 8 exd5 exd5 9 ltJa4 �f5 10 'iVb3 when 10 . . . 'iVa5! 11 ..txb4 ltJxb4 12 c3 ltJa6 is about equal. 8 ... ..txc5 9 ..te3
9 ... �xe3+ More ambitious and at least equal ising is 9 . . . 'iVb6 10 ..txc5 'iVxc5, since 1 1 exd5 exd5 12 ltJd4! (12 ltJxd5 ltJxd5 13 'i¥xd5 'iVxf2 prepares . . . ..te6 or . . . ..tg4) of Ribeiro-Russek Libni, Mar inga 1991 can be met by 12 . . . ..te6! with the idea 1 3 ltJxe6 fxe6 1 4 f3 ltJf5. 1 0 'iVxe3 'iVa5 11 ..tb51? a61? Simpler is 1 1 . . . ..td7= . 1 2 ..ta41? 12 i.. xc6! bxc6 13 ltJd2 ! intending ltJb3 was worth thinking about. 12 ... b5 13 ..tb3 b4 14 ltJa4 Navinsek-Cebalo, Pula 2002. Al though Black went on to win after 14 . . . ..td7 1 5 'i¥c5=, he might have done better by playing 14 . . . dxe4 1 5 'iVxe4 ltJd5.
1 53
Chapter Nine Winawer Variation : Fifth M ove Alternatives
1 e4 e6 2 d4 d5 3 ltJc3 i.b4 4 e5 c5 Our repertoire choice for Black is the traditional and natural 4 . . . c5, at tacking the centre. White's primary response to 4 . . . c5 is 5 a3, so that after 5 . . . i.xc3 6 bxc3, White's threatened d4 point is fortified and he has the bishop pair. White has several other 5th moves, and like the 4th-move op tions of the last chapter, one needs to know how to answer them. I have tried to stick to previous solutions, but with multiple answers and a lengthy new analysis of 5 i.d2, which has provoked many emails and ques tions from readers who are not satis fied about what I presented in the last edition. In this position White has: 9.1 9.2 9.3 9.4
5 5 5 5
ltJf3 i.d2 dxc5 'fig4
9.1 5 ltJf3 This simple development isn't used much, mainly because White's centre isn't well supported. It is nevertheless perfectly sound. 1 54
5 ... cxd4 This capture is an independent course. It is valid but not necessarily best because Black has some handy transpositional solutions, for exam ple: (a) Mter 5 . . . ltJe7, 6 a3 ii.xc3 trans poses to the main line of Chapters 1 1 and 12, and 6 dxc5 is the 5 dxc5 ltJe7 line below. A unique line is 5 . . . ltJe7 6 a3 i.a5!? 7 dxc5, when 7 . . . ..txc3+ 8 bxc3 ltJd 7 is a normal-looking 5 dxc5 line but with a3 in; (b) 5 . . . ltJc6 is the most natural move, allowing some transpositions like 6 dxc5 (transposing to 5 dxc5) and 6 i.d2 (see 5 i.d2). White some-
Winawer Variation: Fifth Move Alternatives times plays the harmless 6 i.b5 (6 a3 i.. xc3+ 7 bxc3 tiJe7 and 7 .. :iIi'a5 are also main-line Winawers) 6 . . . tiJge7 (6 . . . .td7 7 .txc6 .txc6= has also been played) 7 0-0 .td7 (threatening . . . .txc3 and . . . tiJxe5) 8 .txc6 (8 tiJe2 tiJxe5 9 Axd7+ tiJxd7 10 c3 .ta5 1 1 dxc5 e5!:j:) 8 . . . tiJxc6 9 dxc5 Axc5 and White hasn't done anything except surrender the bishop pair. 6 'iVxd4 6 Ab5+ Ad7 is again unimpressive, and 6 tiJxd4 has several answers, the most popular being the double attack on c3 and e5 by 6 . . . �c7: 7 .tb5+ .td7 (or 7 . . . tiJdn) 8 O-O! (a gambit that Black has declined in several games) 8 . . . tiJc6!? 9 'iVg4!? (9 Axc6 bxc6 10 'ifg4 .txc3 can lead to 11 �xg7 .txd4 12 �xh8 0-0-0 with an obvious ad vantage or 1 1 bxc3 �xe5 12 i.a3 h5!? 1 3 'i'f3 'fif6!? 14 �d3 tiJe7+ Zavanelli Livie, corr 1 999) 9 . . . .txc3! (9 . . . tiJxd4!? 10 .txd7+ �xd7) 10 .txc6 (10 'i'xg7 Axd4! 1 1 'iVxh8 0-0-0+ ; 10 bxc3 'iYxe5 1 1 .tf4 �f6+) 10 . . . .txc6 (10 . . . bxc6 transposes to the note to move 9) 1 1 'i'xg7? OBut 1 1 bxc3 'i'xe5 offers no compensation after 12 .tf4 'i'f6 13 .l:!.fe l h5 14 �g3 h4 1 5 'i'g4 tiJh6) 1 1 . . . .txd4 12 'ifxh8 0-0-0 13 'i'xh7 .txe5 14 c3 tiJf6+ 15 'i'h4 tiJe4 16 i.e3 .l:!.h8 17 i.h6 i.g7 0 - 1 De Boer Schaaps, corr 1 987. 6 ...tiJc6 7 i.. b 5 7 'iVg4?! tiJge7 is a position from 5 dxc5 below, but with White's c5 pawn gone! 7 . 'iYa5 8 Ad2 8 0-0 tiJge7 (8 . . . i.. xc3 9 i.. xc6+ bxc6=) 9 �g4 .txc3 10 .txc6+ bxc6 1 1 bxc3 Aa6 12 .l:1 e l �xc3:j: was seen in Arpi-Harikrishna, India 1999, with the idea of meeting 1 3 'i'xg7 with 1 3 . . . .l:!.g8 14 'i'f6 tiJf5+. 8 ... tiJge 7 .
.
9 'ifg4 The alternatives are equal but also dull, which might encourage Black in the direction of 5 . . . tiJe7 or 5 . . . tiJc6: (a) 9 a3 i..x c3 10 Axc6+ tiJxc6 1 1 'ifxc3 'ifxc3 1 2 i.. x c3 0-0=, e.g., 1 3 0-0 a5!? (13 ... i.. d 7 14 l:tadl b6 1 5 :fel l:tac8=) 14 l:tadl Ad7 1 5 l:tfel b5= ; (b) 9 .txc6+ bxc6 10 0-0 tiJf5 1 1 'iVf4 Aa6 12 l:tfdl 0-0 1 3 a3 .txc3 14 Axc3 'ifc7= intending . . . c5, and 15 g4?! tiJe7 is just weakening; (c) 9 0-0 0-0 (9 . . . i.. xc3 !? 10 i.. xc6+ tiJxc6 1 1 'iYxc3 'iYxc3 12 Axc3 Ad7 1 3 Ad4= Hodgson- Saether, Stavanger 1 989) 10 'iYd3 �c7 is logical, e.g., 1 1 Axc6 'i¥xc6= or 1 1 a3 Aa5!? 1 2 b4 Ab6 1 3 tiJa4 tiJg6= . 9 ... 0-0 10 i.d3?! More pointed but also more com mittal than 10 i.xc6 bxc6 11 a3 i.xc3 12 i.xc3 'i'c7 13 0-0 c5= . 10 ... d41? 10 . . . tiJg6! looks good: 11 0-0 (1 1 Axg6 fxg6!) 1 1 . . .tiJcxe5 12 tiJxe5 tiJxe5 13 .txh7+ �xh7 14 'iVh5+ �g8 1 5 'iVxe5 f6:j: . 1 1 O-O?I (a) 11 tiJb l tiJxe5 12 i.xh7+ �xh7 1 3 'i'h5+ 'iit g8 14 'i'xe5 'fixe5+ 1 5 tiJxe5 ii.c5:j: with the bishop pair; (b) 1 1 'iVe4! tiJg6 12 tiJxd4 tiJcxe5 1 3 i.e2= .
1 55
Play the French 1 1 ... lLlf5 1 1 . . . dxc3 draws after 12 iLxh7+ 'ifi>xh7 13 'ii'h 5+ 'ifi>g8 14 lLlg5 %1d8= . After 1 l . . . lLlf5, Hodgson-Ravi, Lon don 1987 continued 12 lLle4 �xd2 1 3 lLlexd2 lLlxe5 14 lLlxe5 'i¥xe5 1 5 lLlf3 'ilVc5 16 life l .l:.d8!? 17 l:Iadl! ? (17 �xf5 exf5 1 8 'iVxd4! 'iVc7 19 'i'h4 �e6=), and here Myers suggests 17 . . . h6 18 .txf5 (?!) 18 . . . exf5 19 'i¥f4, but now instead of his 19 . . . i.. e 6, sim ply 19 . . . 'i!fxc2 20 lld2 (20 l:1.xd4 i.. e6) 20 . . . 'ilVc5 2 1 lLlxd4 i.. d 7 100ks good. 9.2 5 iLd2
This developing move prevents the doubling of White's c-pawns and pre pares lLlb5, from where the knight can infiltrate to d6 or simply cover d4. After a spurt of great popularity, Black seems to have found good de fensive resources. 5 i.. d2 is neverthe less critical and I give two solutions: 9.21 5 ...lLlh6 9.22 5 ..lLle7 .
9.21 5 ... lLlh6 I find this decentralising move, only just mentioned in PTF2, to hold forth the promise of original and complex play into what is a fairly thoroughly analysed variation. 1 56
At the potential cost of allowing White to weaken his pawn structure, Black brings flexibility into his range of strategies, e.g., if White plays lLlb5d6, Black can challenge that piece via . . .f6 and . . . lLlf7. By not playing . . . lLle7, Black's queen is free to move to the kingside, and can recapture on . . . f6 if necessary. Even the move . . . lLlg4 can enter the picture (e.g. after dxc5), and one should note that White's stan dard sacrifice iLxh7+ with lLlg5+ and 'iWh5 is blocked by the knight. The most obvious negative is that White can at some point play .txh6, shatter ing Black's kingside pawn structure. As in other French variations with . . . lLlh6, the disadvantage to so doing is that White forfeits his dark squared bishop and can easily become weak on the dark squares as a result (White's centre is constructed on dark squares for example). This normally wouldn't quite balance out Black's structural loss except that it will have cost White two moves to make this capture: �d2 and then .txh6. Let's see how this works out in practice: 9.2 1 1 9.21 2 9.2 1 3 9.214
6 6 6 6
lLlb5 i..xh6 a3 .td3
Winawer Variation: Fifth Move Alternatives Of these, 6 iLd3 is considered the most important, perhaps incorrectly. In addition to these, White has three alternatives worthy of consideration: (a) 6 f4 is less effective than after 6 . . . ltJe7: 6 . . . cxd4 (or 6 . . . 0-0 7 ltJf3 cxd4 8 ltJb5 iLc5 9 b4 iLb6 10 iLd3 !? f6 1 1 'iYe2 ltJf7 1 2 ltJbxd4 Fernandez Ro mero-Del Barrio Gomez, Vila Real 200 1 ; and Black should have played 12 .. .fxe5! 13 fxe5 i.xd4 14 ltJxd4 'i!Vh4+ 15 'itf2 'itxf2+ 16 �xf2 ltJxe5++) 7 ltJb5 .tc5 8 b4 iLe7 (available be cause e7 isn't occupied by a knight) 9 ltJf3 ltJc6 10 J:.b l ltJf5 1 1 iLd3 ltJh4!? (1 1 . . .a6! 12 .txf5 axb5 1 3 .td3 l:txa2) 12 0-0 (12 ltJxh4 iLxh4+ 13 g3 iLe7) 12 . . . ltJxf3+ 1 3 'i!kxf3 a6 1 4 ltJd6+ iLxd6 15 exd6 Willemze-Barsov, Vlissingen 2000; 1 5 . . . \l!kxd6! + ; (b) 6 ltJ f3 ltJc6 ( 6. . . cxd4 7 ltJb5 iLxd2+ 8 \l!kxd2 0-0=) 7 a3 (7 iLd3 transposes to 6 iLd3) 7 . . . iLxc3!? 8 iLxc3 cxd4 transposes to 6 a3; (c) 6 dxc5
can transpose into other lines, but there are several unique ideas: (cl) A safe solution is 6 . . . ltJd7 and 7 ltJf3 ltJg4! or 7 iLb5 "ilc7 8 ltJf3 ltJg4 9 0-0 .txc5= ; (c2) 6 . . . .i.xc5!? 7 iLd3 (7 ltJf3?! ltJg4 8 iLb5+ .i.d7 9 0-0 iLxb5 10 ltJxb5 \l!kb6=i= ; 7 i.. xh6 gxh6 8 'iWd2 ltJc6 with
the idea 9 �xh6 �xf2+! 10 �xf2 \l!kb6+) 7 . . . ltJc6 (7 . . . 'i!Vh4!?) 8 'i¥h5 (8 i.. xh6 gxh6 9 "iVh5 �g5 10 'iWxg5 hxg5 1 1 ltJb5 'it>e7=i=) 8 . . . ltJf5 9 .txf5?! exf5 10 ltJf3 d4 1 1 ltJa4 g6 12 "iVh3 .i.f8!=i=; (c3) 6 . . . ltJc6!? and now: (c3 1) 7 i..b 5 (not the ideal square for this bishop) 7 . . . 0-0; this threatens . . . ltJxe5 and a plausible continuation is 8 .i.xc6 bxc6 9 iLxh6 gxh6 10 \l!kd4 .l:!.b8 1 1 ltJge2 f6 12 a3 iLa5 13 b4 .i.c7 14 'iWg4+ (14 exf6 e5=) 14 . . . �h8 1 5 n 3 fxe5 16 'iWxh6 a5!?=; (c32) White can also lure the knight to g4 by 7 ltJf3 ltJg4 (7 . . . .txc5) 8 .tb5 iLxc5! 9 0-0 \l!kc7 10 'ii'e 2 .td7 unclear, e.g. 11 .l:!.ael 0-0-0 12 h3 ltJh6 1 3 .i.xh6 gxh6 14 �d2 J:.dg8 1 5 'iWxh6 .l:!.g6 16 'ii'f4 l:!.hg8 1 7 ltJh4 J:.6g7 with compensation for the pawn, planning moves like .. .f5 and . . . iLe7. 9.2 1 1 6 ltJb5 iLxd2+ 7 "ilxd2 0-0 This is similar to 5 . . . ltJe7 6 ltJb5, but the knight on h6 has its advan tages: 8 dxc5 (a) 8 f4 has at least two answers: (al) 8 . . . a6! 9 ltJd6 cxd4 10 ltJf3 ltJc6 (10 . . . f6!=i= is more accurate) 11 .td3 (1 1 ltJxd4 f6) 1 1 . . .f6 (suggested by De la Villa Garcia) gives Black the ad vantage, e.g., 12 0-0 ltJg4!+ with the idea 13 �e2 ltJe3 14 Zif2 'iVb6; (a2) 8 ... ltJc6 9 ltJf3 f6 (9 ... a6!?) 10 c3 cxd4 11 cxd4 fxe5 12 fxe5 ltJf5= 13 i.. d 3 'i!Vb6 14 0-0 i.. d 7 1 5 g4? ltJfxd4 16 ltJfxd4 ltJxe5 1 7 g2 ltJxd3 18 'iYxd3 e5-+ Perelshteyn-Shaked, Blooming ton 1 997; (b) 8 ltJf3?! a6 9 ltJd6? cxd4 10 i.. d 3 ltJc6 1 1 ltJxc8 .l:!.xc8 12 0-0 f6 13 .l:!.ae l ltJf7 14 exf6 'itxf6-+ Leiber-Shaked, Berlin 1997. 8 ...ltJc6 9 ltJf3 b6!? A very common theme, sacrificing 157
Play the French a pawn for activity. 10 'i'c3 10 cxb6 'i¥xb6 1 1 O-O-O?! tbg4 12 tbbd4 tbgxe5 13 tbxc6 tbxc6+ A.Hunt S.Williams, Witley 1 999. 1 0 ....i.d7 11 i.. d 3 Worse is 11 tbd6 !? bxc5 12 li:Jb7 'i'b6 13 tbxc5 li:Jg4! 14 li:Jxd7 'i'xf2+ 15 Wdl l:!.fcS! - Kholmov. 1 1 ... bxc5 12 'ii'x c5 tbg41+ Kholmov-Nikitin, Moscow 1995. 9.212 6 i.. x h6 gxh6 7 a3 7 li:Jf3 tbc6, e.g., S i.b5 i.xc3+ 9 bxc3 1i'a5 10 i.xc6+ bxc6 1 1 1i'd2 i.. a 6 with plenty of play. 7 ... i.a51? The most interesting move, keep ing the bishop that will attack White's dark squares. Black can also play 7 . . . i.xc3+ S bxc3 1i'a5 9 'it'd2 li:Jc6 10 tbe2!? (10 tbf3 i.d7 1 1 i.d3 0-0-0 is thematic, but also plausible is 10 . . . b6, e.g., 11 i.d3 i.a6 12 O-O!? i.. x d3 13 cxd3 cxd4 14 1i'xh6 dxc3 1 5 1iVf6 ':'fS 16 tbg5 li:Jd4! 1 7 li:Jxh7 Ii.g8 18 tbg5 :fS 19 li:Jh7=) Frittchle-J.Watson, Los Angeles 1999, and most incisive was 10 . . . b6!?, e.g., 1 1 dxc5 li:Jxe5!? 12 1i'xh6 1i'xc5 13 1i'f6 li:Jg6 14 li:Jf4 .l:.gS.
S dxc5 (a) 8 'ii'g4 cxd4 9 'i'xd4 i.d7 10 li:Jf3 li:Jc6= ; 1 58
(b) 8 tbf3 'i'b6! is surprisingly strong, e.g. 9 l:!.b l (9 b4 cxb4 10 tbb5 i.d7 11 tbd6+ rj;; e 7 12 i.e2 bxa3+ 13 rj;;fl i..b 4) 9 . . . tbc6 10 b4 (10 dxc5 i.xc3+ 1 1 bxc3 1i'xc5) 10 . . . cxb4 1 1 tbb5 bxa3+ 12 c 3 0-0 1 3 'iVd2 a 2 14 l:al rj;; g 7 1 5 'iVf4 'iVdS 16 rj;; d2 f6+ . 8 ... .i.xc3+ 9 bxc3 tb d 7
10 .i.b5 10 'iVd4 'iVa5 or 10 li:Jf3 1i'a5 11 1i'd2 tbxc5 12 c4 'ii'xd2+ 13 tbxd2 dxc4 14 tbxc4 i.d7 leave weaknesses on both sides. 10 ...'i'a5 11 'iVd3 a61? 1 2 .i.xd7+ i.xd7 and Black seems to have good chances based upon the position of White's king: 13 a4 (13 tbe2 'iVxc5 14 tbg3 l:tcS; 13 'iVe3 !tg8! 14 g3 d4! 15 1iVxd4 i.. c 6 16 f3 lidS+) 13 .. J�cS 14 1i'g3 ':'xc5 15 tbe2 i..b 5 16 l:ta3 ':'c4!? 17 'iVg7 .l:.fS IS Wd2 .l:.e4 19 tbg3 .l:.d4+ 20 'iit c l .l:.xa4. 9.2 1 3 6 a3 i.. x c3 6 . . . .i.a5!? 7 dxc5 .i.c7?! (7 . . . tbd7 8 b4 i..c 7) S f4? b6! 9 cxb6 .i.xb6 10 tbf3 tbg4 11 i..b 5+ .i.d7 12 .i.xd7+ tbxd7 13 'i'e2 i..f2 + 14 Wfl .i.b6 gave Black plenty of compensation in D .Fernan dez-J.Watson, Chicago 1 997; but 8 li:Jf3! was strong since Black can't get enough pressure on e5.
Winawer Variation: Fifth Move Alternatives 7 �xc3 7 bxc3 tDc6 8 tDf3 tDf5 9 �d3 'ikc7 10 O-O?! c4 1 1 .i.e2 .i.d7 12 tDel 0-0-0 1 3 g3 f6 14 f4 �df8 1 5 tDg2 h6= Lo bron-Hug, Beersheva 1985; an early game !
7 .. b6 (a) 7 . . .tDc6 is the move analogous to 5 . . . tDe7 lines except that White has a new idea after 8 tDf3 cxd4 9 tDxd4 tDxe5: 10 'iVh5!? (10 tDxe6 �xe6 1 1 �xe5 0-0= 12 .i.e2 - versus . . . tDg4 12 . . . �g5 1 3 �g3 tDf5 14 0-0 ll.ac8 in tending 1 5 c3 d4! 16 cxd4 ll.fd8) 1O . . . tDeg4!? 1 1 �b5+ 'it>f8 12 f3 tDf6 1 3 'ire5 .i.d 7 14 � d 3 tDe 8 1 5 0-0 'iWb 6 with the idea . . .f6, when White has definite compensation for the pawn, but how much isn't clear; (b) 7 . . . cxd4 8 �xd4 (8 �xd4!? tDc6 9 �f4 and a wild idea is 9 . . . g5!? 10 �f3! d4 1 1 0-0-0 tDf5 unclear) 8 . . . tDc6 9 tDf3 0-0 (9 . . . tDf5 transposes to 5 . . . tDe7) 10 �c5!? (10 i.c3 'i'b6 1 1 .i.d3 tDf5! 12 �xf5 !? exf5 1 3 �xd5 �e6 with compensation) 1O . . . J:te8 1 1 .i.d3 f6 12 �e2 tDf7 (a handy move!) 13 .i.b5 (13 exf6 e5!) 1 3 . . . .i.d7 14 i.xc6 �xc6 1 5 i.d4 .l:tc8= Quast B. Schmidt, Bundesliga 1 996. 8 b4 This is similar to the 5 . . . tDe7 6 a3 lines. .
(a) 8 tDf3 O-O! is better than in the analogous 5 . . . tDe7 Iine because after 9 �d3, 9 . . . ..ta6= can be played without .i.xh7+ being decisive. (b) 8 .i.b5+ �d7 9 �d3 tDc6 10 tDf3 'iVc7!? with the idea 1 1 O-O? cxd4 12 tDxd4 tDxe5 13 tDb5 'iWb8. After 8 b4, A.Martin-Shaked, Schwarzach 1997 continued 8 . . . �c7 9 dxc5 (9 tDf3 cxb4 10 .i.xb4 .i.a6!?) 9 . . . bxc5 10 tDf3?! cxb4 1 1 ..txb4 a5! 12 .i.d2 tDg4! 1 3 ..tb5+ ..td7 14 "iWe2 .i.xb5 1 5 �xb5+ tDd7 16 0-0 0-0 and the e pawn falls. 9.214 6 .i.d3 You realise that a lot of lower play ers are following theory closely when this rather unnatural response to what is already a lesser-known line is chosen more frequently than any other. Why? Because Leko used it in the only 5 . . . tDh6 game involving a world-class player. In fact, Pedersen's 3 tDc3 book deals with no other move. That aside, the move does pose prob lems for Black because it activates an important piece. But it also allows a variety of responses that break up White's centre. 6 tDc6 6 . . .cxd4 7 tDb5 .i.xd2+ 8 "iWxd2 0-0 9 tDxd4 has been considered slightly better for White. Black could still try 9 . . .f6 10 tDgf3 tDf7 intending 1 1 exf6 "iWxf6 and . . . e5, so maybe this should be looked at. 7 tDf3 Now Black has a wide range of moves, several of them yielding satis factory play. 7 f6 This seems neither better nor worse than other solutions: (a) A very simple continuation is 7 . . . tDxd4 8 tDxd4 cxd4 9 tDb5 (an al•••
•••
1 59
Play the French ternative is 9 ttJe2 iixd2+ 10 'iVxd2 0-0 1 1 f4 f6) 9 . . . iixd2+ 10 'iVxd2 0-0 intending ... f6. Maybe I'm missing something because this doesn't seem to have been tried, but I can't see anything wrong with it, e.g., 1 1 ttJxd4 (1 1 0-0 f6 12 exf6 'iVxf6 13 f4? iid7) 1 1 . . .f6 12 ttJf3 (12 f4 fxe5 1 3 fxe5 'iVh4+ 14 g3 'iVh5) 12 . . . fxe5 13 ttJxe5 'iVf6 14 'iVe2 ttJf7 1 5 ttJxf7 Mxf7= ; (b) 7 . . . c4!? is more controversial and could be anti-positional, except that Black gets central play so quickly: 8 .ltfl ! (8 iixh6?! gxh6 9 iie2 f6 10 exf6 �xf6 11 'iVd2 iidn ; 8 iie2 ttJf5 9 ttJb l iie7! 10 c3, and now R.Perez-Arencibia, Santa Clara 2002 went 10 . . . b5!? 11 0-0 h5 12 b3 iid7 1 3 a 4 ttJa5 1 4 bxc4 bxc4 15 ii c 1 ttJb3 16 Ma2 ttJxc1 1 7 �xc1 Mb8:j:) 8 ... ttJf5 9 ttJe2 iie7 10 c3 0-0 1 1 ttJg3 (1 1 g3 f6 12 exf6 iixf6:j: ; the main idea is . . . e5) 1 1 . . .f6 12 exf6 (12 ttJxf5? exf5 13 exf6 iixf6:j: - Kuzmin) 1 2 . . . iixf6 1 3 ttJxf5 exf5 14 iie2 Me8 15 0-0 iid7!? (15 ... Me4!) 1 6 g3 h6:j: H.Hernandez Marcel, Havana 1999. 8 exf6 'iVxf6 9 a3 9 dxc5 is called 'critical' by Peder sen. Then 9 . . . 0-0 10 0-0 iixc5 looks at least equal for Black. The position is reminiscent of the line in Chapter 1 5 with 2 ttJf3 d 5 3 ttJc3 ttJf6 4 e5 ttJfd7 5 d4 c5 6 dxc5 ttJc6 7 iif4 iixc5 followed by . . .f6, with several of Black's pieces on superior squares in the position before us. Sample lines: 1 1 iig5 (1 1 ttJg5 ttJf5 12 g4 ttJe5! 1 3 gxf5 ttJxd3 14 cxd3 'iVxf5 threatens ... h6, trapping the knight, and 15 ttJe2 h6 16 ttJg3 �xd3 17 ttJh3 e5 wins at least a third pawn with a nice attack) 1 1 . . .�f7 12 �d2 (12 ttJa4?! iid6 1 3 c4?! 'iVh5!) 12 ... ttJf5 13 Mae l ! h6 (13 ... �d7!?) 14 .ltf4 ttJh4! 1 5 ttJxh4 'iVxf4 1 6 'iVxf4 �xf4 1 7 ttJg6 Mf6:j: and Black develops 1 60
by . . . iid 7 -e8. 9 ...iixc3 10 bxc3 10 iixc3 cxd4 11 ttJxd4 0-0 12 0-0 ttJxd4 1 3 i.xd4 e5. 10 ... c4!? 10 ... ttJf7 is very solid, preparing . . . e5, and 10 . . . 0-0 1 1 i.g5 'iVf7 should also be considered. 1 1 iig5 �f7 12 iixh6!? 12 iie2 ttJf5 1 3 0-0 0-0 intending 14 'iVd2?! ttJd6! . 12 ... cxd3!? 12 . . . gxh6 looks okay, followed by ... 0-0-0, . . . iid7, and perhaps . . . iie8g6. Mter 12 .. cxd3, De la Villa Garcia Al Modiahki, Ubeda 1 998 went 1 3 ii e 3 dxc2 14 'iVxc2 0 - 0 1 5 0 - 0 ii d 7 16 a4 (slow) 16 ... b6 1 7 ttJe5? (17 Mfel) 17 . . . ttJxe5 18 dxe5 Mfc8:j: . 9.22 5 ...ttJe7 This is the traditional line, where Black develops naturally. White usu ally chooses one of: 9.221 6 ttJb5 9.222 6 a3 9.223 6 f4 Alternatively: (a) 6 dxc5 ttJbc6 7 'iVg4 0-0 trans poses to 5 'iVg4. Black can also play 6 . . . 0-0 or 6 . . . ttJbc6 7 f4 0-0 8 ttJf3 f6, e.g., 9 iid3 iixc5 10 �e2 ttJb4 1 1 0-0-0 ttJxd3+ 12 'iVxd3 �d7 1 3 g4 ttJc6?! (13 .. .f5 ! 14 h3 'iVb6) 14 ttJxd5?! (14 exf6 ttJb4 is unclear) 14 . . . iie8 1 5 ttJc3 �xd3 16 cxd3 ttJb4 17 ttJel? (17 d4:j:) 17 . . . .ltc6 18 ttJe4 ttJxd3+ 19 ttJxd3 �xe4 and Black went on to win in Areklett-S.Bjerke, Asker 2003; (b) 6 ttJf3 cxd4 7 ttJb5 �xd2+ 8 'iVxd2 0-0 transposes to 6 ttJb5 �xd2+ 7 'iVxd2 0-0 8 ttJf3 cxd4. Black also has 6 . . . ttJbc6 7 dxc5 (7 ttJb5 i.xd2+ 8
Winawer Variation: Fifth Move Alternatives 'ii'xd2 0-0 - see 6 lLlb5) 7 . . . lLlg6!? (7 . . . 0-0 S .i.d3 f6=) S .i.b5 i..xc5 9 0-0 .id7 1 0 l:i.el a6 11 i.. d 3 'i¥b6! 12 'ilVe2 lLld4! 13 lLlxd4 i.. xd4+ 14 i.. e 3? i.. xe3 15 .i.xg6 Holms-Tosti, COIT 1 994, and among several winning moves is 1 5 . . . hxg6 16 fxe3 'ii'xb2 1 7 'ii' d3 l:i.cS- + .
.i.g2 i.. e S!?) 16 ... i.. e S 1 7 i.. d 3 lLlb4; Omari-Golz, Leipzig 1 960 . 9.22 1 1 8 c3
9.22 1 6 lLlb5 .txd2+ 7 'ii'x d2 0-0
The main line. White's immediate concern is to resolve the central situation, and there is a wide choice of means: 9.22 1 1 8 c3 9.22 12 8 f4 9.22 1 3 8 dxc5 (a) S lLld6? cxd4 9 lLlf3 lLlbc6 (threatening . . . lLle5) 10 lLlxcS l:i.xcS 1 1 .i.d3 f6+ ; (b ) S lLlf3 a 6 ( S. . . cxd4 9 lLlbxd4 lLlbc6 can also arise via 6 lLlf3 cxd4 7 lLlb5 etc. The main line goes 10 .i.d3 lLlxd4 11 lLlxd4 f6 12 exf6 J:[xf6 13 'ilVe3 'i¥b6 14 0-0-0 lLlc6 15 c3 i.. d 7 16 llhe l e5! 17 lLlxc6 bxc6= Kupper Weinzetl, Brno 1991) 9 lLlc3 (9 lLld6 cxd4 10 lLlxd4 lLlbc6 1 1 f4 f6 trans poses to 6 f4) 9 . . . cxd4!? 10 lLlxd4 lLlbc6 1 1 f4 lLlxd4 12 'iVxd4 lLlc6 13 'ii' d 2 f6 1 4 exf6 'ii'xf6 1 5 g3 i.. d 7 16 O-O-O?! (16
8 ... lLlbc61? This is not the only move: (a) S . . . f6 9 exf6 llxf6 10 dxc5 lLld7 11 b4 a5 12 f4!? (12 lLlf3 b6 13 cxb6 'iVxb6 14 .te2 axb4 15 cxb4 i.. a 6=) 12 . . . b6! 13 cxb6 'iYxb6 with dynamic play, Martinez-Arencibia, Cuba 1 9S5; (b) S . . . a6 9 lLla3 (9 lLld6 cxd4 10 cxd4 lLlbc6 11 0-0-0 f6 12 f4 'ilVb6 1 3 lLlf3 i.. d 7 with the idea . . . lLlcS) 9 . . . lLlbc6 10 f4 (10 lLlf3 f6=) 1O . . . cxd4 1 1 cxd4 lLlf5 12 lLlf3 f6 is equal. 9 f4 9 lLlf3 a6! 10 lLla3 (10 lLld6!? cxd4 1 1 cxd4 f6 12 lLlxcS J:[xcS 13 exf6 l:txf6 1 4 .id3 and now interesting is 14 . . . lLlg6! 15 .txg6 J:[xg6 16 0-0 'ii'f6 1 7 'iith l J:[g4! IS J:[adl J:[f4 1 9 'ii' d 3 g5!? 20 h3 h5 with a nice attack) 10 . . . cxd4 1 1 cxd4 f6 12 exf6 l:.xf6 13 lLlc2 'ii'd 6 14 0-0-0 (14 .i.e2 e5 15 dxe5 lLlxe5;) 14 . . . .i.d7 15 .i.d3 l:.afS 16 l:tdel .i.eS 17 .l:ihfl lLlg6!+ Strikovic-Ivkov, Cet inje 1 99 1 . 9 ... cxd4 Or 9 . . . a6!? 10 lLld6 cxd4 1 1 cxd4 f6= . 10 cxd4 10 lLlxd4 f6 1 1 exf6 J:[xf6 12 lLlxc6 1 61
Play the French ltJxc6 13 0-0-0 .td7 14 ltJf3 .teS= with the idea of . . . .tg6, Manolov-Spasov, Elenite 1992. 10 ...ltJf5 u ltJf3 f6 12 Jt.d3 12 .te2 a6 1 3 ltJc3 fxe5 14 dxe5 �6!' 12 ... a6 1 3 .txf5 axb5! 14 i.. d 3 'iVa5!:j: Lilienthal-Mikenas, corr 1 9 5 1 .
9.22 1 3 S dxc5 White's idea is to overprotect e5 and clear d4 for a knight. 8 .. ltJd7 .
9.2212 8 f4 A thematic strengthening of the centre, but Black gets great counter play:
8 ... ltJbc6 This seems easier than the well known line S . . . a6 9 ltJd6 cxd4 10 ltJf3 (10 ltJxcS 'iVxcS 1 1 ltJf3 ltJbc6 12 0-0-0 ltJf5:j:). Now both 10 . . . ltJbc6 11 ltJxd4 f6 12 ltJxcs 'i!VxcS 13 exf6 l::txf6 14 0-0-0 and 10 .. .f6 11 exf6 l:txf6 12 ltJxcS 'i'xcs 13 0-0-0 ltJbc6 14 ltJxd4 reach the same position, slightly better for Black after 14 . . . ltJxd4 1 5 'i!Vxd4 ltJf5 16 'iVe5 'i!Vc5:j:. 9 ltJf3 9 dxc5 f6 transposes to S dxc5, an other nice line for Black. 9 ... a6 10 ltJd6 Now Black gets the advantage by 10 . . . ltJxd4! 1 1 ltJxd4 cxd4 12 'iVxd4 f6 1 3 exf6 :xf6 14 ltJxcS l:.xcS, e.g., 1 5 c3 ltJc6 16 'iVd2 d4 17 i.d3 dxc3 IS bxc3 e5+ . 1 62
A good alternative is S . . . ltJbc6 9 f4 f6, when Remmelt Otten analyses 10 exf6 l:txf6 11 ltJf3 b6 12 cxb6 'i!Vxb6 1 3 c 3 libS 14 b 3 ? ltJg6 1 5 g 3 e5 16 ltJxe5 ltJgxe5 17 fxe5 .tg4. He says this is winning for Black, and it looks strong after IS ltJd4 I!eS 19 ltJxc6 'itxc6 20 'i'd4 l:tf5 21 .td3 1:tfxe5+ 22 �d2 i.f3+ with the idea 23 I!hf1? �6+ 24 'iVf4 'itxh2+. 9 f4 (a) 9 'iVc3 a6 10 ltJd6 'iVc7 and: (al) 1 1 b4? b6; (a2) 1 1 ltJf3 'i!Vxc5 12 'i!Vxc5 ltJxc5= ; (a3) 1 1 0-0-0 'i!Vxc5 12 �xc5 ltJxc5 13 ltJf3 ltJc6 14 c4 dxc4 15 .txc4 b5 16 .td3 Holschuh-Lahlum, corr 1 996-97, and Black could play 16 . . . ltJb4! 17 .tb l i.. b 7!? I S ltJxb7 ltJxb7, bring a rook to the c-file followed by . . . ltJd5 and . . . ltJc5; (a4) 11 f4 'i!Vxc5 12 'i!Vxc5 ltJxc5 13 ltJf3 f6= Tradardi-Ottavi, Rome 1990. (b) 9 ltJf3 ltJxc5 10 .td3 (10 0-0-0 ltJe4 11 'iVe l .td7 12 ltJbd4 ltJc6 1 3 h4 f5 1 4 g3 'i!Vc7:j: Ciolac-Zysk, Vienna 1990) 1O . . . ltJc6 (1O . . . �6!:j:) 1 1 0-0 f6 12 I!fe l .tdn V.Knox-Levitt, British Ch 1993.
Winawer Variation: Fifth Move Alternatives 9 .. ltJxc5 10 ltJd4 1!fb6 Also fine is 10 ... ltJc6 1 1 ltJgf3 f6, e.g., 12 i.. d 3? ltJxd4 1 3 ltJxd4 fxe5 14 fxe5 �4+ 1 5 g3 "iWg4+ Moen Djurhuus, Norway 1 989. 11 0-0-0 .id 7 12 ltJgf3 Or 12 Wb 1 l:tac8 13 .l:.c1 ltJe4 14 "iWe3 ltJc6= . On 12 ltJgf3, Sanz Alonzo- Sion Cas tro, Salamanca 1990 saw 12 . . . .l:.fc8 13 "iWe3 ltJa4! 14 VWb3 (14 "iWa3 llc7! - No gueiras) 14 . . . 'i'c5 ! 15 i.. d 3 (15 Wb 1 b5!= or 1 5 'iixb7 .ic6 16 ltJxc6 ltJxc6 Nogueiras) 1 5 . . . b5 16 a3 l:tcb8 1 7 "iWb4 a5+ . .
9.222 6 a3 i.. x c3 7 .ixc3
Here Black can choose between a drawish line (which might deter players of White from this line) or one with mutual chances: 7 ...ltJbc6 Fully equalising, but with no pros pects. 7 . . . b6 is often played by stronger players and gives more win ning chances for both sides, e.g., (a) 8 b4 "iWc7 9 ltJf3 cxb4!? 10 .ixb4 a5 1 1 iLd2 is the old main line, con sidered equal but unbalanced, e.g., 1 1 . . .0-0 12 .id3 iLa6= J.Friedman Gulko, Philadelphia 1993; (b) 8 ltJf3 is solid, e.g., 8 ... iLa6 (8 . . . 0-0 9 .id3 ltJg6!? has ideas of both
. . .ltJf4 and . . . i.. a 6 when White has already moved the king's bishop) 9 .ixa6 ltJxa6 10 0·0 0-0 1 1 b4 ltJb8!? 12 iVd3 cxd4 1 3 ltJxd4 ltJbc6 14 ltJe2!? (14 ltJxc6 ltJxc6 is not so easy because Black has the c-file) 14 . . . l:tc8 15 f4 h5 16 i.e 1 ltJf5 17 i.. f2 ltJce7 1 8 c3 .l:.c4 with active pieces, Cabrilo-Lputian, Cetinje 1991; (c) 8 iVg4 0-0 9 ltJf3 i.. a 6 10 iLxa6 ltJxa6 1 1 iLd2 cxd4 ( 1 1 .. :�c7!?) 12 iLg5!? "iWc7 13 .ixe7 "iWxe7= De la Villa Garcia-Bareev, Leon 1995; (d) 8 iLb5+ i.. d 7 9 i.. d 3 ltJbc6 10 ltJf3 (10 f4!? cxd4 11 i.d2 0-0 12 ltJf3 f6 13 iVe2 Spassky-Mohr, Bundesliga 1988; this doesn't look very convinc ing after, e.g., 13 . . . iLe8) 10 . . . ltJg6! (or 1O . . . cxd4 1 1 iLxd4 ltJxd4 12 ltJxd4 ltJg6=) 1 1 i.xg6 fxg6 12 0-0 0-0 1 3 dxc5 bxc5 14 b4 "iWb6!+ Leko-Bareev, Wijk aan Zee 1995. 8 ltJf3 cxd4 9 ltJxd4 9 iLxd4 is also a little dull: 9 . . . ltJf5 (or 9 . . . ltJxd4 10 'ilVxd4 ltJc6 1 1 'ilVg4 0-0 1 2 i.. d 3 f5= Balashov-Svistunov, Pinsk 1993; 12 . . .f6 is perhaps more precise, but allows 13 "iWh5 and on the correct 13 . . . g6, 14 iLxg6 with a draw) 10 c3 iLd7 1 1 i.. e 2 .l:.c8!? (l 1 . . .ltJfxd4 12 cxd4 "iWb6=) 12 0-0 0-0 13 'ilVd2!? ltJfxd4 14 cxd4 ltJa5!= Savon-Hort, Skopje 1968. 9 . ltJxe5! Otherwise Black is slightly worse. 10 ltJxe6 iLxe6 11 i.. x e5 0-0 1 2 iLd3 12 iLe2 ltJc6 1 3 i.. g 3 'iYf6 14 c3 d4! Moles & Wicker. 12 ltJc6 13 i.g3 13 iLc3 d4 14 i.. d2 ltJe5! 15 iLxh7+? Wxh7 16 "iWh5+ Wg8 17 "iWxe5 .l:.e8 18 "iWg3 iLc4+ 19 Wd1 .l:.c8 with a huge attack as in the game Ljubojevic Nogueiras, Wijk aan Zee 1987. 1 3 ..."iWf6 14 .l:.b l 14 O-O!? "iWxb2 1 5 llb 1 "iWxa3 16 .
.
•••
1 63
Play the French l:t.xb7 is an important try to stir things up, but offers insufficient play for the pawn: 16 . . . :fe8!? (or 16 . . . 'iIi'c5 ! 1 7 'i!i'h 5 h 6 1 8 Wh1 liJb4 19 �e2 a5 2 0 .l:r. c 7 'iib 6 2 1 c3 Hector-Sorensen, Ber lin 1993; and now 2 1 . . .liJa6! with the idea 22 �xa6 'ili'xa6 or 22 l:te7 .:tfe8 looks good) 1 7 'i!i'h5 (17 .l:r.e 1? ':'e7!+ was Tringov-Uhlmann, Skopje 1972) 17 . . . h6 18 f4 l:t.e7 19 l:t.b3? (19 ':'xe7 'ili'xe7 20 f5 �d7 2 1 f6 Vi'e3+ 22 Wh1 'ili'g5) 19 . . . 'iIi'c5+ 20 �f2 Vi'd6 2 1 f5 ii.d7+ as in the game Obukhovsky Hasin, USSR 1973. 14 ... �f5 This position offers little for either player, and also possible is 14 . . . g6!?, e.g. 15 0-0 d4 16 �e4 (16 Vi'd2) 16 . . . llac8 and . . . IUd8. 1 5 0-0 llfe8 Or 15 . . . i.. x d3 16 �xd3 :ad8= Gue don-Naumkin, Cappelle la Grande 1993. 1 6 ii.xf5 Vi'xf5 1 7 'iWd2 As usual this is drawn, but instead of the difficult 17 . . . .l:i.e6?! of Hector Rowson, New York 1999, easier is 17 . . . h6 18 l:tfe1 d4 Y2 - Y2 Raese-Fran ke, Bundesliga 1992. 9.223 6 f4 liJf5l? Somewhat ambitious. Others tries equalise quickly: (a) 6 . . . 0-0 7 liJf3 (7 dxc5 f6! 8 liJf3 liJc6) 7 . . . liJbc6 8 dxc5 f6 transposes to 6 dxc5 liJbc6 7 f4. Black can also play 7 .. .f6 8 exf6 :xf6 9 dxc5 i.. xc5 10 �d3 liJbc6 11 'ili'e2 i.d7 12 0-0-0 liJb4 1 3 'litb 1 W.Watson-Knaak, Kecskemet 1987, and now 13 . . . liJxd3! 14 cxd3 (14 'iWxd3 i.. e 8!) 14 ... b5 intending the move . . . b4; (b) 6 . . . cxd4 7 liJb5 i.xd2+ 8 'ili'xd2 0-0 has been played in a number of contests and White has gotten no where, e.g., 9 liJxd4 liJbc6 10 liJgf3 1 64
liJxd4 (or 1O . . .f6 1 1 exf6 l:1xf6 12 ..td3 liJxd4 1 3 liJxd4 'iWb6 1 4 c3 ii.d7 15 0-0 liJc6=) 11 liJxd4 �d7 12 ii.d3 'iib 6 1 3 liJf3?! ( 1 3 c 3 liJc6=) 1 3 . . . i.b5 (instead 13 . . . 'iIi'xb2!? 14 0-0 iVb6+ 15 'lith1 f6 would be a reasonable winning try) 14 ii.xb5 'i¥xb5 1 5 Wf2 !? .l:r.ac8 16 .l:r.hc 1 h6!? Y2 - Y2 Gullaksen-0gaard, Oslo 1998. 7 liJf3 Mistaken is the try 7 liJb5? a6! 8 i..xb4 cxb4 9 liJd6+ liJxd6 10 exd6 1IVxd6+ Santo Roman-Raicevic, Ath ens 1992. 7 ...liJc6 Mter 7 ... cxd4 8 liJb5, 8 ... liJe3 9 'iWc1 �c5 10 b4 �e7 looks worth trying, as does 8 . . . �c5 9 b4 �e7 (9 . . . �b6 10 �d3 �d7 1 1 g4 a6!?) 10 liJbxd4 liJxd4 11 liJxd4 liJc6 12 liJxc6 bxc6 13 ii.d3 .l:r.b8 14 'iii'g4 g6 - McDonald. 8 dxc5 �xc5 9 ..td3 liJh41? 10 'ili'e2 McDonald gives 10 liJxh4 'ili'xh4+ 1 1 g3 'i!i'h3 12 i.. f1 �6, when 13 f5 �e3 or 13 �f3 liJd4 14 'ili'd3 'ili'h5 1 5 �g2 �d7 100ks fme. 10 ... liJxf3+ 11 'ili'xf3 liJb4 1 2 0-0-0 ii.d7 = Hebden-McDonald, British Ch 1989. In spite of its popularity, 5 i.d2 is a rather easy line to meet. 9.3 5 dxc5
Winawer Variation: Fifth Move Alternatives This cutting-edge move emphasizes piece play and is very dangerous. Of course the danger is on both sides! I will follow a main (very main) line, giving some options on moves 6 and 7, because Black can get these posi tions almost by force and secure good play thereby. 5 lLJc6 In the last edition I gave 5 . . . lLJe7 emphasising ideas with . . . lLJd7 and . . . lLJxc5. Those are fine, but I'd rather devote space to the more forcing line that follows. A couple of thoughts on other moves: (a) 5 ... 'iVc7 6 lLJf3 lLJc6 transposes to 6 . . . 'iVc7 below, and here 6 . . . lLJd7 is also playable; (b) 5 . . . d4 6 a3 .i.a5 7 b4 dxc3 8 �g4 lLJe7 is underinvestigated and worth looking into. 6 lLJf3 •••
6 'i'g4 lLJge7 transposes to 5 'iVg4, and 6 .i.f4? d4 7 a3 i.a5 8 b4 lLJxb4! 9 axb4 i.xb4 10 iLb5+ �d7 1 1 'iYxd4 iLxc3+ 12 'iYxc3 iLxb5 is good for Black. 6 lLJge7 Two potentially significant alterna tives: (a) The logical move 6 . . . 'iVc7 has been looked at more since the previ ous edition: ...
(a1) 7 .i.d2 lLJxe5 (or 7 . . . i.xc5) 8 �b5+ iLdn ; (a2) 7 .i.b5 lLJge7 (7 . . . i.d7 threatens . . . lLJxe5 and seems to equalise: 8 iLxc6 .i.xc3+ 9 bxc3 i.xc6 10 0-0 lLJe7) 8 0-0 0-0 9 lLJa4!? (9 .l:[e1 i.d7) 9 ... lLJxe5 10 lLJxe5 'iYxe5 1 1 c3 i.a5 12 f4!? 'iff6 1 3 i.e3 lLJf5 14 .i.d4 lLJxd4 ( 1 4 . . . �e7!) 1 5 cxd4?! ( 1 5 'iVxd4 �e7) 1 5 . . . a 6 16 �e2 i.d7 17 I:tc1 Gallagher- Carton, Lon don 1986; 17 . . . i.xa4! 18 �xa4 i.d2:j:; (a3) 7 i.f4 lLJge7 8 iLe2 (After 8 a3!? �xc3+ 9 bxc3, Black can choose be tween 9 . . . f6, 9 . . . �a5, and 9 . . . lLJg6 10 i.g3 h5!? with the idea 11 h4 lLJgxe5 12 lLJxe5 lLJxe5, which is now possible because White does not have 'iVh5; 8 �d3 has been played, but should come up short after 8 . . . d4 9 a3 iLa5 10 b4 lLJxb4 11 axb4 �xb4 12 �b5+ i.d7 13 �xd4 �xc3+ 14 'i¥xc3 i.. xb5 1 5 lLJd4 i.a6:j: ; 8 'iVd3 lLJg6!? 9 iLg3 .i.xc5) 8 . . . f6 9 exf6 'i¥xf4 10 fxe7 iLxc5 and Black's centre and two bishops compensate for the time it takes to recapture on e7; (b) 6 . . . d4 may be better than its reputation due to the line 7 a3 �a5 8 b4 dxc3 9 bxa5 'i'xd1 + 10 'it'xd1 lLJge7 (10 . . . a6!?) 1 1 i..b 5 (1 1 i.d3 0-0 12 'it'e2 lLJxa5 13 l:td1 lLJd5 14 i.e4 i.d7 1 5 i.xd5 Soltis-Mednis, Cleveland 1975; perhaps this is better for Black: there 1 65
Play the French could follow 15 . . . exd5 16 l:txd5 .tf5!? 1 7 .tf4 i.. xc2 18 l:i.cl .te4:j: ; 1 1 a6 bxa6 12 i.. d 3 i..b 7 13 l:tb l 0-0-0=) 1 1 . . . a6 12 i.. xc6+ ltJxc6 1 3 �e3 (13 ..t?e2 i.. d 7 14 �d3 ltJxa5 1 5 �xc3 i.c6 should be fme, e.g. , 16 i.. e 3 i.. xf3 1 7 gxf3 ltJc6 18 f4 0-0-0=) 1 3 . . . ii.d7 14 l:i.b l ltJxa5 1 5 .id4 .tc6 1 6 ii.xc3 ..ixf3+ 1 7 gxf3 ltJc6 18 ..t?e2 0-0-0= Palliser-Knott, Scar borough 200 l . 7 ii.d3 (a) 7 ii.d2 transposes to 5 ..id2 ltJe7 6 ltJf3 ltJbc6 7 dxc5; (b) 7 a3 .txc3+ 8 bxc3 ltJg6!? (8 . . :�a5 9 �d2 �xc5 10 a4 0-0 1 1 ii.a3 �a5=) 9 .tb5 'i!Va5 (9 . . . .i.d7 threatens . . .ltJxe5 and . . . 'i!Va5, so 10 .txc6 i.. xc6 11 0-0 f6!? 12 'i!Vd4 fxe5 1 3 ltJxe5 ltJxe5 14 �xe5 'i!Vf6 might fol low) 10 .txc6+ bxc6 11 'i'd4 ..ia6 12 h4 'i!Vb5!? (12 . . . .l:tb8 13 h5 ltJe7 and Black has a few more prospects) 1 3 c 4 ! i¥a4!? (13 . . . 'iYxc4! 14 'i!Vxc4 i.xc4 15 h5 ltJf8 16 h6 g6 and the opposite coloured bishops probably lead to a draw) 14 i¥c3 i.xc4 15 h5 ltJe7 Tran Quoc Dung-Szymanski, Artek 1 999; and 16 h6! gxh6 1 7 .txh6 ltJf5 would have been somewhat in White's fa vour. 7 d4 This begins a long forced sequence which ends with mutual chances. 7 . . . i.. xc5 has been played here with some success, but I don't fully trust it and probably White gets a small theoretical edge. More interesting and not fully tested is 7 . . . i¥c7 and: (a) 8 .tf4 d4 9 a3 dxc3 (9 . . . i.. xc3+ 10 bxc3 ltJd5:j:) 10 axb4 cxb2 11 lIb l ltJxb4 (1 1 . . .ltJd5 12 ..id2 ltJdxb4:j:) 12 lhb2 ltJxd3+ 1 3 'iVxd3 ltJg6 14 ..ig3 'i'xc5+ Paci-Taddei, Nancy 2003; (b) 8 0-0 is the most critical move, gambiting a pawn: 8 . . . ..ixc3 9 bxc3 ltJxe5 10 ltJxe5 (10 .tf4 ltJxf3+ 1 1 ...
1 66
'iVxf3 e5 12 i.b5+ ..t?f8 1 3 .tg3 'iVxc5 14 i.. a 4 e4:j: Smagacz-Morkisz, Slask 1996) 10 .. :ii'xe5 1 1 'i!Vg4 .id7!? (1 1 . . .0-0!, since 12 .i.xh7+ ..t?xh7 13 'iVh4+ 'it>g8 14 'iix e7 'iVxc3 1 5 i.. f4 b6! favours Black, but of course White doesn't have to play this way; . . .£6 will be a useful move in any case) 12 1:tb l ..ic6 1 3 'i'd4! Graf-Duebon, corr 1998, and 13 . . .f6 14 f4 'ilc7 15 l:i.el �f7 would have created some sort of balance. 8 a3 ii.a5 9 b4 ltJxb4 10 axb4 ..txb4
A position that has been debated for many years. I think that Black at least equalises and probably does more than that in several lines. He will also tend to pose more difficulties for his opponent than he himself has to suffer. This calls into question the value of 5 dxc5, both with 6 ltJf317 .td3, and in general. But the theoreti cal density of these lines is staggering. 1 1 0-0 Almost automatic, but occasionally other moves are tried: (a) 11 i..b 5+ i.. d 7 12 'i!Vxd4 .ixc3+ 13 'iYxc3 i.. xb5:j: 14 ltJd4?! (14 �g5 'iVd5:j:) 14 . . . ..ta6+ according to older sources. Then 15 c6?! 'iVd5! 16 cxb7 'i'xe5+ 17 i.. e 3 i..xb7 18 0-0 0-0 leaves Black on top; (b) 1 1 ltJd2 !? .txc3 12 l:i.b l i¥d5
Winawer Variation: Fifth Move Alternatives (12 . . . 0-0 13 0-0 .i.xd2 14 ii.xd2 iVc7!?) 1 3 0-0 i.xd2 14 .i.xd2 resembles the main line, but Black hasn't had to spend a tempo on . . . h6: 14 . . . ii.d7 15 l:1.e l ii.c6 16 .i.e4 'i'd7 1 7 l:1.b3 Upton Harley, England 1996, and here sim ply 17 . . . .i.xe4 18 l:txe4 'iVd5 and . . . 0-0 leaves White with awful pawns. 1 l ... .i.xc3 12 l:tbl
17 .i.f4 h5?! 18 liJe4 0-0-0 19 lLld6+ liJxd6 20 cxd6 'iVa5 21 f3 Y:z - Y:z Palkovi F.Portisch, Zalakaros 2000) 16 . . . gxf5! (16 . . . exf5 1 7 e6 'iVd5 18 exf7+ 'it'd7 19 f3;t was eventually drawn in Bokros Hertneck, Austria 200 1) 17 'iVg7 'it'd7! and Black is better, according to the post mortem by Shaked and Galla gher an example being 18 1Llxf7 (18 ' lLlxh7 �c7 1 9 'iYxf7+ 'iYd7) 18 ... 'iVg8 19 'iVxg8 .l:.hxg8, for example, 20 f3 .l:!.g7 21 lLld6 �ag8 22 g3 h5 23 .i.f4 h4 24 �f2 hxg3+ 25 hxg3 l:1.h8 26 l!thl l:tgh7 27 l:1.xh7+ l:1.xh7- + . 9.31 12 ...'iVc7
Now there are two main moves: 9.31 12 ...'iYc7 9.32 12 ... h6 But what is possibly a very good third one is 12 . . . .i.d7!? Of course White has a lot of ideas to play with and the 5 dxc5 experts don't seem worried, so I'm pretty sure that the following doesn't tell the whole story: 13 liJg5 .i.c6 14 'i'h5 ('±' according to Pachman back in 1968! McDonald calls 14 lLle4 'critical'; if nothing else, 14 . . . .i.xe4 15 .i.xe4 �b8 100ks safe and sound. He gives 14 . . . 'i'd5 15 liJd6+ 1i?f8 16 .i.e4 'i'xe5 1 7 f4 with various complications, although I think that Black is okay in those lines too) 14 . . . g6 (14 . . . liJg6 15 liJxe6! fxe6 16 ii.xg6+) 1 5 'i¥h6 (15 'iYh3 'i'c7 16 i.. f4 a5 1 7 ii.c4 lLlf5 Y:z - Y:z Gallagher Shaked, Cannes 1997; Black is bet ter) 15 . . . 1L1f5 16 i.. xf5 (16 'i'h3 'i'c7f.
This has been played just a few times but I like it. 1 3 liJg5!? The only move used by strong players: White wants to play lLle4-d6. But the alternatives are also note worthy: (a) 13 lLlxd4!? and now 1 3 . . . i.. xd4 14 .i.b5+ .i.d7 15 'i¥xd4 �xb5 16 ':'xb5 l!td8 17 1lVb2 l:1.d7 yields some advan tage, as does 13 . . . i.. d 7, but 13 . . . a6?! 14 ii.b5+! axb5 15 lLlxb5 'i'xe5 16 'iYf3 is less clear; (b) 13 i.. d2 lLld5 is solid; after 13 . . . 'iVxc5 14 ii.xc3 dxc3 15 ii.b5+ lLlc6 16 ii.xc6+ 'iYxc6 17 'iVd3 D .Myers Katz, corr 1 983, easiest is 17 . . . b6!+ 1 67
Play the French with . . . �b7 to follow; (c) 13 'i'e2 i.. d 7 (13 . . . h6 14 tDd2 �xd2 15 i.. xd2 i.. d 7 16 'ite4 i.. c 6 1 7 'i'xd4 0-0 18 'i'd6 l:I.fc8 i s unclear; but 13 . . . tDc6 100ks promising) 14 tDg5 i.. c 6 15 tDe4 (15 'i!fu5 g6; 15 tDxh7 tDd5 16 tDg5 .t.b4!) 15 . . . �xe4 16 �xe4 tDc6� ; (d) 1 3 �f4!? hasn't been tried but I think it is the most challenging move other than 1 3 tDg5: 1 3 . . . tDg6 14 ..txg6 hxg6 (14 . . . fxg6 15 tDxd4 0-0 isn't bad either) 15 tDxd4 (15 tDg5 f5!) 1 5 . . . 'iWxc5 16 tDb5 �xe5 17 ..txe5 'iWxe5. 1 3 .. :i¥xe5 14 'i¥h5 g6 15 'i¥h6 tDf5 1 6 �xf5 gxf5
1 7 i.. f4 The crazy 1 7 tDxf7!? might be used if the main line 17 i.. f4 isn't satisfac tory: 17 . . /tixf7 18 'i'h5+ �e7 (18 . . . �f8 19 i.. h 6+ �e7=) 19 .t.g5+ �d7! (19 ... Wf8=) 20 'iWf7+ Wc6 21 ..tf6 'iic 7 22 'i'xc7+ �xc7 23 .t.xh8 �c6 24 l:t.fdl Wxc5 and it's hard to believe that Black isn't better with his bishops and a-pawn, but White can go after the queen's pawn by 25 l:I.d3 b 5 26 l:I.bdl �b7 27 �xd4+ �xd4 28 l1xd4 �d5 and Black has slightly better chances due to his ideal piece place ment, but those will be hard to con vert. 1 7 ...'iWxc5! 1 68
1 7 . . .'iWxf4? 1 8 'i'g7 d3 19 'i'xf7+ �d8 20 .l:txb 7 �xb7 1-0 Gallagher Atanu, Sangli 2000. 18 'iWh5 "iJle7
1 9 �e5 I'm not sure if 1 9 ..td6 has been analysed, but Black seems okay after 1 9 . . . 'i'f6!? 20 i.. e 5 'i'xe5 2 1 'i'xf7+ �d8 22 tDf3 'iVd5 23 "iJlf6+ �c7 24 'i'xh8 b5! 25 'iVxh7+ �b6 and 26 IUdl �b7 with good chances or 26 'itg7 a5 27 l1fdl i..b 7 28 tDxd4 lId8! 29 11xb5+ 'i'xb5 30 'iWc7+! �xc7 31 liJxb5+ �b6 32 liJxc3 11xd1+ 33 liJxdl �e4� . 19 ... h6 20 tDxe6 i.. x e6 2 1 �xh8 and 'White is better but Black is clearly still in the game' - Gallagher. This is a fascinating position pitting Black's bishops and passed a-pawn against White's extra exchange and potential dark square play. I doubt that White is any better at all. Unfor tunately, it took me 38 moves and numerous subvariations to arrive at equality, so I'll just get you started: 2 1 . . .�c4 22 l1fcl (22 'iYxf5 i.. xfl 23 �xf1 'i'e6! 24 'i'xe6+ fxe6 25 l:t.xb7 a5�) 22 . . . 'iYe6! (the a-pawn is a factor in all of these positions; worse is 22 . . . �d2 23 l:tdl �e2 24 'iWxf5! �xdl 25 l:I.xdl �c3 26 'i'f3 l1b8 27 �f6 'i'e6 28 g3;!;) 23 .l:[xb7 (23 'ii'f3 'i'e4 24 'i'xe4+ fxe4 25 l:ixb7 a5) 23 .. :i\¥e4!?
Winawer Variation: Fifth Move Alternatives (Black can already draw by 23 . . . 0-0-0 24 .l:.cb l �a2 25 .l:1.b8+ rl;c7 26 l18b 7+ ct>c8=) 24 'i'if3 'i'ixf3 25 gxf3 -td5 26 .l:.c7 iLxf3 27 ..tf6 a5 and my analysis has 28 l:tb 1 ! (and not 28 l1e7+) lead ing to a dead draw. 9.32 12 ... h6 '!' Gallagher. The point is to stop lLlg5-e4.
gests that he should probably have continued by 19 . . . lLlf5 ! , since 20 -txf5? lIxd2 21 'iWxg7 'i'ixf2+ mates . 1 5 ... �c6 16 ltb4 16 ..tb5?! 'i'ic8 (16 . . . 0-0!?i) 1 7 ..txc6+ lLlxc6 18 'iWb 1 0 - 0 was drawn in Rubery-Harley, London 1996; McDonald analyses 1 9 .l:1.e1 .l:i.b8 20 J:b3 'iVc7 21 'iVa I ! ? lIxb3 22 cxb3 ltb8i .
13 lLld2 Much better than 13 ct:Jh4 'iVd5 ! 14 f4 ..td7 1 5 f5 exf5 16 �xf5 Zinser Ackerman, corr 1964; 16 . . . �xf5! 17 lLlxf5 ct:Jxf5 1 8 ':'xf5 0-0-0 19 'ii'g4 'iVe6+ - Moles & Wicker. 13 ...ii.xd2 Black prevents lLle4-d6. 14 �xd2 -td7! There is a mass of analysis on 14 . . . iVc7. I have no idea why, but the ory says that White comes out with a small edge! 15 .l:.xb7 Poorer is 1 5 'iVg4 �c6 16 ltb4 'iVa5!? (or 16 ... iVd5 1 7 ltxd4 iVxe5 1 8 lIel 'ii'f6 19 ..tc3, and here instead of 19 . . . iVg5? 20 h4! , I recommended 19 . . . 0-0, intending 21 Ji.a1 'i'ig5 or 20 h4 lLlf5) 17 'ili'xd4 .l:i.d8 18 �g4 (18 l:.a1 .l:.xd4 19 lIxa5 Ibb4 20 .llxb4 a6i) 18 . . .iVxc5 1 9 l:[e1 Y2-Y2 Lima-McDon ald, Hastings 1988; McDonald sug-
1 6 ... a5! Actually, I don't really know if the fantastically convoluted lines after 16 . . .'i!id5 1 7 'i'g4 .l:1.d8 18 lle1 g5 !? or even other orders beginning with 16 . . . iVd5 are playable; but I only have space to present one alternative and the experts in this system seem to indicate that the advance 16 . . . a5 is best. 17 .l:.b6 'ii'd 5 Black's idea has been to lure the rook away from its attack on d4; in addition, the vital a-pawn gets roll ing. Nevertheless, the tradeoff is double-edged: for one thing, the move lIxc6 at the right moment can be dev astating. 18 'iVg4 'iVxe5 19 .l:1.el! 19 �fb 1 is no longer considered correct, although it's not completely clear: (a) 19 . . . 0-0?! 20 J.f4 'i'f6 was drawn 1 69
Play the French in one game after 21 'ilVg3, but win ning the exchange was better: 21 J.. d 6 J.. d 5 (2 1 . . .l:i.fc8? 22 �xe7 Wixe7 23 l:.xc6 ':'xc6 24 'iWe4) 22 f4! (threaten ing �e5) 22 . . . lDg6 23 .i.xfB lDxfB 24 c6± ; (b) 19 . . . g5!? is plausible, but per haps unnecessary, because: (c) 19 . . . h5! seems to solve all Black's problems after 20 "iWh4 f6!? (20 . . . "iWxc5! looks good, or perhaps 20 . . . "iWd5!?) 21 i.. x a5? (2 1 "iWh3! intends �c4, so 2 1 . . .'ilVd5 22 liel �f7 might follow with complex play) 2 1 . . .'iVxc5+ 22 .l::txc6?? lDxc6- + Van Mechelen -Meessen, Charleroi 200 l . 1 9 ...'iVxc5 This time 19 . . . h5 might run into 20 ':'xe5 hxg4 21 ':'g5!? a4 (2 1 . . .g6 22 lixg4 e5 23 c3!? f5 24 ':'xc6!) 22 l:.xg4 a3 23 Iib l .l:!.h5 24 l1xd4, but that's certainly not clear, so 19 . . . h5 deserves to be looked at. 20 ':'xc6! lDxc6 20 . . . 'ilVxc6? 21 i.. e 4 'iVc8 22 Wixg7 J:.g8 23 'iVxd4 and the bishops are a terror, with �h7 a threat. 21 'ilVxg7 'ltle7 Or 2 1 . . JH8 22 i..xh6 'ltle7. 22 �g6 l:.hf8
23 . . . 'ilVc3, Steenbekkers did some long analysis that indicates that 23 . . . a4 leads to a draw, particularly in the line 24 h4 a3 25 i.xf7 l:.xf7 26 J.. g 5+ e8 27 ltxe6+ lDe7 28 Wig8+ .l::tfB 29 'i¥g6+ 'ltld7 30 .l:i.xe7+ 'iWxe7 31 J.. x e7 'ltlxe7 and the a-pawn forces White into perpetual check. The whole line with 5 . . . lDc6 should serve Black well. Apart from the 3 choices on move 12, the options on moves 6 and 7 give one plenty of ideas to play with. 9.4 5 'ilg4 This aggressive queen sortie mobi lises White's forces quickly at the in evitable cost of letting his central structure be compromised. 5 ... lDe7
In this position White has two at tacking ideas: 9.41 6 dxc5 9.42 6 lDf3
23 �xh6 Gallagher-Steenbekkers, Cappelle la Grande 2002. Here instead of 1 70
Others are less effective: (a) 6 a3?! 'iWa5! 7 iLd2!? (7 axb4 'iWxal 8 dl cxd4 9 lDb5 O-O!+ Jansa Korchnoi, Luhacovice 1 969, due to 10 lDc7 lDa6! l 1 lDxa8 �d7; 7 lDge2 cxd4 8 axb4 "iWxal 9 lDb5 O-O! 10 lDc7 lDa6 1 1 lDxa8 i.. d 7 12
Winawer Variation: Fifth Move Alternatives 'i!Va4+ 14 �d2 li:Jxb4 0- 1 EngholmFreyer, corr 1 968) 7 . . . cxd4 8 axb4 'i!Vxal + 9 li:Jdl li:Jbc6 (9 . . . li:Jf5 ! - Moles) 10 li:Jf3 �d7 1 1 'i!Vxg7 l:tg8 12 'i!Vxh7 a6; - Uhlmann; (b) 6 'i!Vxg7 ':g8 7 'i!Vh6 (7 'i!Vxh7? cxd4 8 a3 'i!Va5-+ , e.g., 9 lIb 1 dxc3 10 axb4 'iWa2) 7 ... cxd4 8 a3 iLxc3+ 9 bxc3 iIIc 7 10 li:Je2 (after 10 �dl, Tisdall suggests 10 . . . li:Jd7! and I think that's the way to go. This position can be compared to the main-line Winawer with 7 'i!Vg4 'i!Vc7 8 'iWxg7 ':g8 9 'iWxh7. In the position before us, crucially, White hasn't picked up the pawn on h7) 10 . . . dxc3 1 1 f4 ii.d7 12 :b l li:Jbc6 13 li:Jg3 (13 'iWxh7 d4 14 g3 0-0-0 15 iIId 3 J.Edwards-Mayo, corr 1986, and 15 . . . li:Ja5 'preserves all the chances' Edwards) 1 3 . . . 0-0-0 14 ii.d3 li:Jd4 1 5 'i!Vh5 (15 i.. e 3? l:txg3! 16 hxg3 li:Jdf5+; 15 0-0 li:Jdf5 ! 16 li:Jxf5 li:Jxf5 1 7 �xf5 'i!Vc5+! 1 8 �hl exf5+ - Uhlmann) 1 5 . . :iic 5!+ Sigurjonsson-Uhlmann, Hastings 1975/6.
i.. x c3+ 9 bxc3 bxc6 intending . . . i.. a6) 8 �d2 (8 ii.xc6+ bxc6 9 i.. d2 li:Jf5 10 li:Jge2 h5 11 'iif4 iIIxc5 12 0-0 i.. a 6= Yudasin-Lputian, Simferpol 1988) 8 . . . 0-0 9 ii.xc6 li:Jxc6 10 li:Jf3 iIIx c5 (10 .. .f5! 1 1 exf6 lIxf6;) 1 1 0-0 f5 12 'iWf4 d4= Hebden-Raicevic, Vrnjacka Banja 1989; (b) 7 li:Jf3? d4! 8 ii.b5 'iVa5 (8 . . . Wf8 is also promising) 9 iLxc6+ bxc6 10 iIIx d4 (10 iIIxg7 ':g8 11 iIIxh7 i.. xc3+ 12 Wdl l:tb8 1 3 li:Jg5 i..xb2-+) 1O . . . li:Jf5 1 1 'iWc4 (1 1 'iWe4 i.. xc3+ 12 bxc3 'iWxc3+ 13 li:Jd2 I1b8!-+) 1 1 . . .i.. a 6 12 'iVb3 'iVb5 13 Wdl 'iVxc5 14 i.. d 2 Ba nas-Prandstetter, Czechoslovakia 1978; 14 . . :iWxf2 ! 1 5 'i'xb4 'ikxg2 16 l:tel iIIxf3+ 1 7 �c1 0-0-0+ . 7 ... 0-0 Also played is 7 . . . li:Jg6, which con tinues to score well and should be looked into if Black desires an alter native.
9.41 6 dxc5 This can transpose from and is re lated to 5 dxc5. White attacks the bishop on b4. 6 ...li:Jbc6
Now White has: 9.4 1 1 s li:Jf3 9.412 S i.. d 3
7 ii.d2 (a) 7 ii.b5 'i!Va5 (or 7 . . . 0-0 8 i.. xc6
Mter 8 0-0-0, 8 . . . b6!? has been played recently (8 .. .f5 9 exf6 l:txf6 10 i.. d 3 'iWf8 11 li:Jf3 i.. xc5 is given by Pedersen; it looks at least equal) 9 li:Jf3 (9 cxb6 'iWxb6!? with open queen1 71
Play the French side lines; Pedersen mentions the move 9 . . . axb6) 9 . . . bxc5 10 iLd3 ltJg6 1 1 'iWh5 and now Pedersen suggests 1 1 . . .'iWe8!, one idea being 12 ltJb5 c4! 1 3 ltJc7 �xd2+ 14 1:.xd2 'fie7 1 5 ltJxa8 cxd3 etc. 9.4 1 1 8 ltJf3 f5
9 'fih4 (a) 9 'iWh5 .lid7 (or 9 . . . d4) 10 a3 .lixc5 1 1 �d3 ltJd4!;; (b) 9 exf6 lixf6 1 0 0·0-0 e5 1 1 'iWh5 1:.f5! 12 'iWh4 �xc3! 13 bxc3 (13 .i.xc3 lH4) 13 . . . 'fia5-+ Pietzsch-Uhlmann, Germany 1963; (c) 9 'ifg3 ltJg6 10 .td3 d4 11 ltJb5 �xc5= . 9 ... d4 Or 9 . . . ltJg6! 10 'ifxd8 (10 'iWh3? ltJgxe5 1 1 ltJxe5 ltJxe5 12 'ife3 ltJc6+ Popovych -Mednis, US Ch 1972) 10 . . . 1:.xd8 1 1 ltJa4 .lidn. 10 ltJe2?! ii.xd2+ 1 1 ltJxd2 'iVd5!; Estrin-Rittner, corr 1966, which continued 12 ltJb3 ltJg6 1 3 'iVg3?! (13 'fih3 - Schwarz; 1 3 ... a5!?) 1 3 . . .f4 14 'fif3 'fixe5 1 5 0-0-0 1\Vg5 ! ; . 9.412 8 � d 3 .lixc3!? A move investigated by Tal Shaked and me. A couple of alternatives are: (a) 8 . . . �xc5!? 9 ltJf3 ltJg6 10 'fih5 ltJb4 gets rid of the bishop but keeps 1 72
White's centre intact, Van Perlo Stigar, corr 1993; (b) 8 .. .f5!? 9 exf6 ltxf6 10 'i¥h5! g6 (1O . . . h6 1 1 g4! is very dangerous) 1 1 'i¥h4 was difficult for Black in Stren zewilk-J.Watson, Philadelphia 1 996, but 1 1 . . .ltJf5! intending 12 .lixf5 lixf5 would have been equal. 9 iLxc3 9 bxc3 ltJxe5 10 �xh7+ c;i;>xh7 1 1 'iWh5+ c;i;>g8 12 'iVxe5 f6;. 9 ... d4 1 0 0-0-0
White's most obvious attempt to keep the initiative. Others in this po sition are: (a) 10 �d2 ltJxe5 1 1 �xh7+ 'it>xh7 12 'iWh5+ 'it>g8 13 'iVxe5 f6 (a common theme; 13 . . . ltJc6!?) 14 'iVd6! e5 15 c3! ltJf5 16 'iVxd8 .l:r.xd8= ; (b) 10 1:.dl dxc3!? 1 1 .txh7+ 'it>xh7 12 ':'xd8 .l:r.xd8 13 'fih3+ (the alterna tive 13 bxc3 l:td5 14 ltJf3 <;t>g8 favours Black) 13 . . . 'it>g8 14 'iVxc3 ltJd5 is at least equal. 10 ... dxc3! 1 1 .txh7+ Iit>xh7 12 :xd8 cxb2+ 13 c;i;>xb2 ':'xd8 I think that this favours Black, al though it's close. He has the ideas of . . . ltJf5-h6 . . . . ltJg6, and/or . . . .td7-e8 for defence, if necessary. On the positive side, there is an attack via . . . .l:r.d5 and . . . b6. Here are some of the possible lines:
Winawer Variation: Fifth Move Alternatives 14 liJf3 �g8 15 'ilYh5
9.42 6 liJf3
15 ...liJg6!? This may or may not be the best move, but it sets up a tough defence. 15 . . . liJf5 is also promising, e.g. , 16 g4 (16 liJg5 is met by 16 . . . liJh6 1 7 liJe4 .l:td5 or 16 . . . liJxe5!? 17 'iHh7+ �f8 18 lIel liJg6) 16 . . . liJh6 17 g5 liJf5 18 g6 fxg6 19 'ii'xg6 b6 (19 . . . .i.d7!?) 20 liJg5 (20 l:i.el liJce7 21 'iVh5 bxc5; 20 .l:.gl bxc5) 20 . . . liJxe5 21 �h7+ 'iitfS 22 l:tel (22 f4 liJc4+ 23 'it?c3 liJce3) 22 ... liJf7 23 liJxe6+ .i.xe6 24 .l:r.xe6 .l:td5! 25 c6 :c8� . These are just sample analyses, but I think that Black stands well. 16 liJg5! 16 l:i.el l::t d 5, and one of many pos sibilities would be 17 liJg5 liJcxe5 18 1Vh7+ 'it>fS 1 9 h4 b6 20 h5 liJc4+ 2 1 �c3 l::tx g5 22 hxg6 I1.xc5 23 'i¥h8+ 'it>e7 24 'ii'x g7 liJe5+ 25 'iitb 3 liJxg6- + . Mter 16 liJg5, play can go 16 . . . liJcxe5 1 7 'iVh7+ WfS 18 h4 liJg4! 19 h5 (19 'i¥h5 l::t d 4! 20 f3 liJe3 21 g4 .l:td2-+) 19 . . . liJf6 20 hxg6 liJxh7 2 1 gxh7 'it>e7 22 h8'i' .l:r.xh8 23 l:i.xh8 b6�. This is all exciting stuff, and it's hard not to get the impression that Black is somewhat better throughout. The 5 'iVg4 liJe7 6 dxc5 line is clearly a seri ous threat, but in the end it seems that White has to be careful not to come out worse.
Here White tries to hold the d4 square and develop quickly, but Black no longer has to worry about blitzkrieg attacks. 6 ... cxd4 This move has become increasingly popular. For reasons of space I won't analyse 6 . . . liJbc6 from the last edition again, but it might be good to men tion that 7 dxc5 transposes to the last section and that 7 a3 needn't be an swered by 7 . . . .txc3+ with a Positional Winawer main line, but allows 7 . . . i.. a 5! , when after 8 'i'xg7?! J:.g8 9 'ii'xh7 cxd4 10 b4 I quoted a game with 1O . . . liJxb4 1 1 axb4 .i.xb4 leading to equal and unclear play, but an in teresting alternative was seen in Skripchenko-B. Socko, Cappelle la Grande 2000: 10 . . . i.. c 7!? 1 1 liJb5 a6 12 liJbxd4 (12 liJxc7+ �xc7 threatens . . . liJxe5 and . . . liJb4, e.g., 13 .i.b2 liJxe5 14 liJxd4 liJc4 15 .txc4 �xc4 McDonald) 12 . . . liJxd4 13 liJxd4 .i.xe5 14 c3 (McDonald gives both 14 .i.b2 "ikc7 15 0-0-0 i.. x d4 16 l:i.xd4 e5 1 7 .l:td2 .i.f5 and 14 'i'd3 i.. f6 intending . . . e5 and . . . .tf5) 14 . . . "ikc7 15 .id2 .i.f6! 16 liJe2 e5 17 liJg3? e4! 18 'i¥h6 i.. g 5 ! - + . 7 liJxd4 liJg6 7 .. :flc7 seems fine. As far as I 1 73
Play the French know, the main move 8 .tb5+ hasn't been answered by 8 . . . ltJd7!? since Grigorov-Lukov, Bulgaria 1976. Nev ertheless, after 9 0-0 .txc3 (9 . . . 0-0 10 ltJxe6 fxe6 11 1Wxb4 ltJxe5= - Minev) 10 bxc3, Minev suggests 10 . . . 0-0!? when one idea is 1 1 .i.a3 ltJxe5:j:.
8 i.. d 3 Others may not even equalise: (a) 8 ltJf3?! i.xc3+ 9 bxc3 'iWcn ; (b) 8 .tb5+ i.. d 7 9 0 - 0 0 - 0 (still bet ter looks 9 . . . i.xc3! 10 bxc3 'iic 7 1 1 .td2 0-0+) 1 0 i.xd7 (10 ltJxe6?? .txc3 ! - + ; 10 i.. d 3 ltJxe5 1 1 iLxh7+ 'it>xh7 12 ft5+ 'ito>g8 13 'ii'x e5 ltJc6:j: Murey-Sigurjonsson, Brighton 1 983) 10 . . . 'ii'x d7 1 1 ltJf3 ltJc6 12 1Wh5 f6 1 3 exf6 l:lxf6 14 .i.g5 l:!.f5:j: Fernando Garcia-Matamoros, Seville 1 992; or here 14 . . . .txc3! 15 bxc3 ':'f5+; (c) 8 .i.d2 0-0 9 ltJf3 ltJc6 10 0-0-0 f5 ! 1 1 exf6 �xf6 12 'ito>b 1 e5 13 �g3 ltJf4:j: Movsesian- Shaked, Zagan 1997. 8 . . 0-0 9 �xg6 If White tries to maintain e5 by 9 .
1 74
ltJf3, Black has 9 . . . .i.xc3+ 10 bxc3 f5 ! with the idea 1 1 'iVh5 'WIc7 or 1 1 exf6 'i'xf6. 9 ... fxg6 By this recapture, Black activates his rook, but also prevents problems based upon ltJf3-g5. 10 0-0 10 .td2 .te7 (10 . . . .tc5!?) 1 1 0-0 ltJc6 12 ltJf3 'i!kb6 (12 . . . .:.f5!?:j: - Kinsman) 13 ltJd1 .td7 14 ltJe3 l:lad8 15 b3 i.. c 8 16 ':'ad1 'i'c7 17 'ii'g 3 b5! 18 .tc1? d4! 19 ltJg4 h5 Y2 - Y2 Murey-Kinsman, Paris 1996; Black is practically win ning after 20 ltJf6+ J..xf6 2 1 exf6 'iWxg3, but he wanted to secure a norm. 10 ... Sl.e7 11 i.e3 11 ltJf3 ltJc6 12 I:i.e1 .td7 1 3 ltJe2 m6 (threatening . . . ltJxe5) 14 i.. e 3?! (14 ltJf4 'ifc5 1 5 c3 ':'f5+) 14 . . . 'iWxb2 15 l:tab 1 Hebden-Matamoros, London 1987, and easiest was 15 . . . �xa2! or even 1 5 . . . ltJxe5! . 1 1 ltJc6 12 ltJxc6 bxc6 1 3 ltJa4 1Wc7 14 'iWd4 ':'f5! 15 f4 White might try 1 5 g4, but even 15 . . . .:.f7 16 f4 ':'b8! favours Black: 1 7 c3?! c5!+. After 1 5 f4, Murey-LFarago, See feld 2002 went 15 . . . g5! 16 g4 (16 fxg5 'WIxe5 17 I:i.xf5 'i'xf5 18 l:lf1 e5! Farago, although 18 . . . 'iVxc2 seems easier) 16 . . . c5! 17 'iVc3 (17 ltJxc5 ':'xf4!) 17 . . ..:.xf4! 18 ':'xf4 (18 .txf4 gxf4 19 .l::i.xf4 iLb7 and . . . d4 and . . . 1Wc6 are coming - Farago) 18 . . . gxf4 and Black stood clearly better. •••
Chapter Ten Winawer Variation : Main Li n e with 7 "iYg4
1 e4 e6 2 d4 d5 3 liJc3 �b4 4 e5 c5 5 a3 �xc3+ 6 bxc3 liJe7 7 1Wg4 This is the quintessential Winawer variation and the subject of ceaseless testing since Alexander's win over Botvinnik in 1 946. It is as important today as ever, with many leading grandmasters using it as their anti Winawer weapon. With 7 'YWg4, White is trying to exploit Black's lack of a dark square defender and provoke weaknesses on the kingside. Given his own structural disadvantage (the doubled c-pawns), the first player needs to utilise his bishop pair; his cramping e5 pawn makes the king side a natural area to proceed on. As in many modern openings, however, both players (and especially Black) will frequently play on both sides of the board. What's more, whether and how White or Black should open the game varies from case to case. In the variations presented in this book (with 7 liJf3 and 7 'iVg4 0-0), some kind of central action tends to be de sirable for Black ( . . . f6 or . . . f5 and/or . . . cxd4), to open lines for his pieces and establish aggressive posts for his knights. If he closes the position by
both . . . c4 and .. .f5, the bishops will come into their own. That is, White will be able to rearrange his forces at his leisure and then break through (e.g., by g4) . In addition, White can sometimes benefit from the line opening exf6 (in response to .. .f6 or . . . f5) and, less often, the move dxc5 can open lines for early rook and bishop action. Proper piece and pawn play is terribly complex and often paradoxical in this variation, which in general can only be understood by example. 7 0-0 A new system for this edition. I don't believe for a moment that 7 . . . 'YWc7 deserves the negative and unsupported criticism that commen tators regularly churn out (and I've looked at every new development for some years now) . But addressing the enormous bulk of new theory would now take up too much of this book and anyway it's fun to present some thing fresh. The ideas behind 7 . . . 0-0 (apart from protecting g7, which is an important function!) will again be come clear by example, and every thing said above applies. I have ...
1 75
Play the French drawn very heavily upon Kinder mann and Dirr's definitive work Franzoesisch Winawer for material and will refer to its authors' contribu tions by 'K&D', or as 'Kindermann' for cases in which I saw his analysis separately. After 7 . . . 0-0 White has: 10.1 S lLlf3 10.2 S �d3
Nothing else is dangerous, e.g., (a) 8 h4?! f5 9 exf6 l:1xf6 10 i.. g5 lIfl 1 1 lLlf3 leaves Black effectively a tempo up on 8 lLlf3 f5 9 exf6 .l:!.xf6 10 i.. g 5 lIfl, a respectable line for Black. Thus 1 1 . . .'iVa5 (or 1 1 . . .lLlbc6) 12 .i.d2 lLlbc6 13 lLle5 lLlxe5 14 dxe5 lLlc6 15 f4 'iVa4!+ threatening . . . lLlxe5 and . . . "iWxc2, Durao-Ivanov, Cappelle la Grande 1995; (b) 8 'i'h3 ttJbc6 9 .i.d3 ttJg6 10 ttJf3 c4 11 .i.xg6 fxg6 12 0-0 .i.d7. White has a normal 8 .i.d3 ttJbc6 line but he is committed to 0-0 and has his queen less than ideally placed on h3.
that analysis is included in this sec tion. I will show one entertaining so lution to 8 ttJf3; it is also important because it solves a couple of the variations that could also arise after 8 .i.d3. S . . lLlbc6 8 . . .f5 9 exf6 l:1xf6 is also well known and considered satisfactory. Then a crazy line that seems to be holding up is 10 i.. g 5 ttJd7!?, based upon 1 1 i..xf6? ttJxf6 1 2 'iVf4 "iWa5 and the c pawn can't be defended. The solid and respectable 10 . . . .l:!.fl should also equalise according to K&D's book. 9 .i.d3 f5 10 exf6 (a) 10 'fill 3 contains some poison. The idea is not to open lines for Black on the kingside, at the same time de veloping and in some cases trying to get moves like g4 and l:1g1 in. Black's best-known answer is 10 . . .'iWa5 1 1 i.. d 2 'iYa4, threatening . . . c4 and . . . �xc2. .
10.1 S ttJf3
This natural move has lost in popu larity as effective remedies have been found for Black, but in several cases it transposes into a 8 .i.d3 line and 1 76
(al) Mter 12 �dl ?!, K&D give 12 . . . cxd4!? 1 3 cxd4 ttJxd4 14 "iWh4 ttJec6 1 5 i.. e 3 ttJb5! with advantage, or 12 . . . i.. d 7 13 g4 c4! 14 i.. e 2 i.. e 8! ; (a2) 12 dxc5 i s the safest try: 12 . . . ttJg6 13 'i'g3 'iVg4 14 0-0 (14 lIb l h5!? 1 5 i.. g 5 ttJgxe5 16 ttJxe5 ttJxe5 1 7 lib4 'i'xg3 1 8 hxg3 g6 - Mitkov) 14 . . . �xg3 1 5 hxg3 ttJcxe5 16 lLlxe5
Winawer Variation: Main Line with 7 'ii'g4 ttJxe5 1 7 Itfel ttJxd3 18 cxd3 �d7 19 .l:.ab l b6!= Mitkov-Kasimdzhanov, Istanbul 2000; (a3) 12 lIgl!? (an ambitious pawn sacrifice) 12 . . . c4 13 i.e2 'iWxc2 14 Itc l 'i!Vb2 1 5 g4 ttJa5 16 i.dl ttJb3 1 7 litc2 'iWxa3 18 .ih6 (18 ttJg5!? h6 19 ttJf3 ttJxd2 20 'iitxd2 fxg4 2 1 l:!.xg4 Velimi rovic-Drasko, Niksic 1996, and K&D now improve upon 2 1 . . .ttJf5?! with 2 1 . . .'iit h 8! , e.g., 22 l:!.c l b5 23 i.c2 b4 24 ':xg7 'i'xc3+ 25 �e2 'iitxg7 26 l::I g l + �f7 27 ifu5+ ttJg6!-+) 1 8 . . . ttJg6 (pol gar gives 18 . . . gxh6 19 'iWxh6 ttJg6 as leading to perpetual check) 19 gxf5 l::I xf5 20 �e3 (K&D give 20 ttJh4 ttJxd4! 21 ttJxg6 ttJxc2+ 22 'iitfl 'iVa I, with a small advantage after great complications) 20 . . . i.d7 21 ttJg5 ttJfB 22 f4 J.Polgar-Shaked, Hoogeveen 1998, and now 22 . . . 'ifa l ! (various) when Black stands better; (b) 10 'ii'g 3 'iWa5 1 1 i.d2 b6! (there are other solutions, e.g., 1 1 . . .'iWa4) 12 0-0 'i'a4 13 l::Ife l (13 h3? c4 14 i.e2 f4!+ and . . . 'iVc2) 13 . . . i.d7 14 l:ta2 Y2 - Y2 Spassky-Ehlvest, Belfort 1988; K&D suggest 14 . . . cxd4 15 cxd4 ttJxd4 16 .ib4 ttJxf3+ 1 7 'iWxf3 .l:.f7:j:. 10 ..liixf6 ..
l l .ig5 Natural and critical, but 11 'i'h5!? is also important: 1 1 . . .h6 (1 1 . . .g6 12 'ifg5
'iffB is also played) 12 0-0 (12 ttJe5 ttJxe5 13 dxe5 I!fB 14 0-0 c4 15 �e2 .idn intending . . . �e8-g6, Vitolins Zlotnik, Nabereznye Chelny 1 988) 12 . . . i.d7! (12 . . . c4 13 �e2 'i¥fB!? 14 .id2 i.d7 15 'i'g4 ttJg6= Gallagher-Farago, Hastings 1990) 13 ttJe5 (13 l::I e l c4 14 i.e2 i.e8 1 5 'i'h3 'iVa5 16 i.d2 �g6+ Lau-Vladimirov, Moscow 1989; Black doubles rooks, plays . . . �e4 and at tacks) 13 . . . �e8 14 'ii'g4 c4!? (14 . . . 'i'a5:j:) 15 .te2 ttJg6 (15 .. :ii'a5!? - K&D) 16 f4 unclear - K&D; then Black could play 16 . . . ttJcxe5 17 fxe5 l::Ixfl + 18 �xfl i.f7:j:. 1 l e5! ...
This surprising move commits Black to a material investment, but destroys White's centre and brings all the pieces into play. 12 'ii'g3? ! This has been played most but in the end it doesn't work out welL (a) 12 'i'h4 simplifies and is proba bly the safest move, even if Black retains some edge: 12 . . . e4 13 .txf6 gxf6 14 'ii'xf6 exd3 (14 . . . 'i'fB 1 5 'ii'xfB+ 'iit xfB 16 dxc5! exf3 17 gxf3 �f5 18 :b l b6 19 .txf5 ttJxf5 20 cxb6 axb6 21 .l:.xb6 ttJce7= - K&D; White has 4 ex tra pawns, but 6 weak ones!) 15 cxd3 cxd4! 16 ttJxd4 ttJxd4 17 'ii'xd4 ttJc6 and instead of 19 'ii'g3+ �h8 20 0-0 1 77
Play the French i.. f5 (20 . . . dxc3!? - K&D) of Vitolins Dokhoian, Porz 1991, K&D suggest 18 'iVe3 d4 19 cxd4 i.. f5 20 0-0 'i'xd4 unclear, although it's hard not to pre fer Black's pieces; (b) 12 i.. xh7+ 'itxh7 13 'iVh5+ 'iit g 8 14 i.. xf6 gxf6 15 dxe5 'iVf8 is a posi tion that has arisen many times.
16 exf6 (16 0-0 'iilf7 1 7 'iilxf7+ 'iitxf7 18 exf6 'iit xf6+ ; 1 6 c4 d4 [here 16 . . . dxc4!? is a good try] 17 exf6 Wixf6 18 Wixc5 d3 leads to unclear complica tions) 16 . . . Wixf6 17 0-0 (1 7 Wig5+ Wig7 18 0-0-0 iLg4 1 9 l1d3 :l.f8 20 'i'xg7+ rJitxg7 Enders- Hiibner, Bundesliga 1 999) 17 . . . Wixc3!? (17 . . . i.. f5 18 l:tael rJitg7 19 li'lh4 :l.h8 20 li'lxf5+ Wixf5 un clear - K&D; I like Black in all these positions, if only mildly) 18 li'lg5 Wig7 19 .l::!. ael �f5 20 .l:.e3 'iil g6 21 Wih4 i.xc2 22 g4!? - K&D ; but the attack may not fully compensate White for his investment. 12 .l::!. x f3! Actually, this 'sacrifice' is forced in view of White's threats. It is nonethe less a good example of Black's dy namic chances in this line. 13 gxf3 13 'iilxf3 e4 14 ..txe4 dxe4 15 'iVxe4 'i'd5! 16 'i'xd5+ li'lxd5 and Black's pieces are too active, e.g., 17 dxc5 i.f5! 18 0-0-0 li'lxc3 19 .l::!. de l �c8 20
i.. e 3 li'la5 etc. 13 c4 ...
14 �e2 At least this keeps the bishops. Others: (a) 14 �xe7?! 'iVxe7 15 i.e2 exd4 hardly improves, e.g., 16 �f1 �f5+ Abramovic-Dokhoian, Belgrade 1 988; (b) 1 4 i.xh7+?! 'itxh7 15 :l.gl 'iVf8 16 Wih4+
...
1 78
This leads to fascinating and in-
Winawer Variation: Main Line with 7 'ii'g 4 structive play with respect to typical exchange-down positions ill the French pitting bishops versus knights. As far as 1 can see, Black always gets an advantage. 1 will point out a few problems with existing the ory: (a) 15 0-0 (the most direct move, but perhaps not best) 15 . . . i.f5! 16 cxd4! (the main line in theory is 16 l:tfe 1 dxc3, when White has no useful tactics, e.g., 1 7 i.xc4? dxc4 1 8 l:txe7 tiJxe7 19 l:te1 and 19 . . . h6 or 19 . . . 'i\Yb8!? 20 l:txe7 'iWxg3+ 2 1 fxg3 .txc2 22 l:txb7 i.. d 3 and Black's pawns were too strong in Szalanczy-Kindermann, Dortmund 1 992) 16 . . . �xc2! (I found this move over the board, fortunately missing theory's 16 . . . tiJxd4, when White has two moves. The books give 17 .td1 �xc2 with a large advantage, although 18 'ii'g 4! then approximately equalises; my opponent was going to play still more actively by 17 l:tfe 1 !?, for example, 1 7 . . . tiJxc2 1 8 i.xc4! tiJxe1 19 l:txe1 dxc4 20 l:txe7 with near equality due to the opposite-coloured bishops, a mistake being 20 . . . 'iIi'dl +?! 21 Iit?g2 .td3? 22 :xg7+! Iit?h8 23 h4!, since 23 . . . lit?xg7 24 'iWe5+ leads to mate) 1 7 l:.fe1 tiJf5! 18 'iWg4 (18 i.xd8 tiJxg3 19 fxg3 ':'xd8 and the d-pawn falls) 18 . . . 'iVfB- + 19 .tfl tiJfxd4 20 .tg2 .tf5 21 'i!ig3 tiJc2 and Black was clearly winning in Rensch-J.Watson, Chicago 2003; (b) 1 5 cxd4?! tiJxd4 1 6 .td1 (16 'i!ie5 'iVa5+! 17 Iit?d1 'iWc3 18 l:tc1 .tf5-+ Kindermann) 16 . . . 'ii'a 5+ 1 7 .td2 c3 18 .te3 tiJdf5+ - Efimov; (c) 15 .tf6! 'i!ifB 16 .txd4 is the sol idest position available to White, even if he has to play very accurately after 16 . . . tiJf5 17 'ifg5! (17 'i!if4 has been played twice, but Black has a solid advantage after 1 7 . . . tiJfxd4 18
'i!ixf8+ 'it>xf8 19 cxd4 tiJxd4 and 20 .td1 c3! 21 0-0 .tf5 or 20 Iit?d2 .tf5 2 1 .t d 1 l:.e8 22 l:ta2 o f Nguyen Anh Dung-Galyas, Budapest 2000, when apart from the game's 22 . . . b5, 22 ... tiJb5! enforces ... c3 with a win ning position, unless White opts for 23 c3 d4! 24 cxd4 c3+ 25 'it>c1 tiJxd4 etc.) 1 7 . . . tiJfxd4?! (17 . . . .te6! develops quickly and keeps options open; it favours Black somewhat after 1 8 ':' d 1 ? 'ili'xa3 o r 1 9 'iVd2!? l:.e8! o r 18 i.. e 3 d4! 1 9 cxd4 tiJfxd4 20 .td1 l:.e8 and moves like . . . tiJxf3+ or . . . c3) 18 cxd4 i.f5 1 9 c3 l:.e8 20 l:!.a2! l:te6 2 1 h 4 o f Maliutin-Piskov, Moscow 1 989, which Kindermann justly assesses as unclear. 15 i.. d 2 15 0-0 tiJf5 16 'iVg2 exd4 is worse still. 1 5 ...tiJf5
16 'iWg5 Not much better is the rather mis erable-looking 16 'i!ig2 when Black still has a big advantage after 16 . . . exd4 1 7 cxd4 c3 1 8 .te3 tiJfxd4 19 i.. xd4 tiJxd4 20 i.. d 3 .tf5! 2 1 .txf5 l:.e8+. 1 6 . exd4 17 cxd4 c31 18 i.e3 tiJcxd4 1 9 .txd4 Kindermann analyses 19 .td1 .td7 20 0-0 'ue8+ or 20 . . . l:.f8+. ..
1 79
Play the French 19 ... lLlxd4 20 11g1 g6 21 'iWe5 'iWc5! 22 'ife8+ 'iff8 23 'iYxf8+ 23 I:txg6+ hxg6 24 'iWxg6+ 'iWg7 25 'iVe8+ c;t>h7. 23 ...c;t>xf8; Roth-Kindermann, Vienna 1996; the game concluded 24 c;t>d1 .i.f5 25 ]:tc1 lie8 26 l:i.e1 b5 (zugzwang!) 27 .i.d3 .i.xd3 28 .l:r.xe8+ �xe8 29 cxd3 c2+ 30 �e1 a5 31 f4 b4 32 axb4 axb4 33 c;t>d2 lLlb3+ 34 �xc2 lLlxc1 35 �xc 1 c;t>d7 0- 1 .
Here 9 lLlf3 f5 transposes back to (10. 1) 8 lLlf3, but there is one more serious option with the line 9 �g5 'i!Va5 10 lLle2
10.2 8 .i.d3
This has been the main move for a number of years (bishops before knights!). By attacking h7, White would like to get Black to make a weakening or at least committal move on the kingside, after which he can return to normal development. I'll examine two solutions: 10.21 8 lLlbc6 10.22 8 ... f5 ..•
10.21 8 lLlbc6 My preferred solution: it is both sound and offers plenty of chances. Again we look at two moves: •..
10.2 1 1 9 iVh.4 10.212 9 iVh.5
1 80
which leads to typically complex play which has never been fully re solved. Black has various ideas, but these two look the best: (a) 1O . . .cxd4 11 f4! dxc3 had a bad reputation but appears okay: 1 2 0-0 lLlg6 13 iVh5 (13 11f3? lLlcxe5! 14 fxe5 lLlxe5 15 �xh7+ c;t>xh7 16 'i'h5+ �g8 and Atlas gives 17 �f6! lLlxf3+ 18 gxf3, but misses 1 8 ... 'iVb6+ 1 9 h1 e5-+) 13 . . . 'i!Vc5+ 14 h1 lLlce7 1 5 ]:tf3 f6!? 16 exf6 gxf6 17 l:1h3 J:l.f7 18 i.xf6 .l:r.xf6 19 'iWxh7+ c;t>f8 20 11g3 'iif2 2 1 h4!? (2 1 'iVh6+ f7 22 'iVh7+=; 2 1 lLlg1 'ilVxf4 22 lUI 'iYxf1 23 �xf1 11xf1 24 h4 �e8 25 h5 lLlf8 or 25 . . .lLlh4 Shaposh nikov-Riazantsev, Moscow 2000; I don't believe that White has enough for his material deficit) . After 21 h4 (Hertneck-Uhlmann, Bundesliga 1992), K&D offer 2 1 . . .e5!, which is equal after 22 h5 exf4! (K&D give the move 22 . . .lLlxf4, but it leaves the h pawn too strong) 23 �xg6 fxg3 24 'iVh6+ �g8 25 'iVh7+ �f8 26 'iVh6+ etc.; (b) 10 . . . lLlg6 1 1 0-0 ( 1 1 h4?! lLlxd4!) 1 1 . . .'i!Va4 12 f4 c4 1 3 �xg6 fxg6 14 .l::t a 2 i.d7 1 5 h4
Winawer Variation: Main Line with 7 "fiIg4 9 .. tbg6 Black has two other moves that look satisfactory: (a) 9 . . . tbf5!? 10 'i!Vxd8 llxd8 1 1 g4 (1 1 Ji.xf5?! exf5 12 tbe2 b6! 13 ..tg5 .l:[d7 14 0-0-0 h6 15 iLe3 ..ta6:j: S.Hall Ludevid, email 1 993) 1 1 . . .tbfe7 1 2 �g5 (12 tbf3;!; - K&D, but I like 12 . . . l:tfS! , preparing . . f6 preceded in some cases by . . . c4, e.g., 13 dxc5 tbg6 14 ..txg6?! fxg6 with excellent play. White has better, of course, but Black seems okay in all lines) 12 . . . l:td7!? (better seems either 12 ... �d7 or 12 ... .l:[fS 1 3 tbf3 c4 14 iLe2 �d7 un clear; Black can always play for . . . b 5/ . . . a5/ . . . b4, but he can also play for . . .f6, e.g., 15 h4 h6 16 .i.e3 'iith 7 1 7 g5 !? tbf5 18 gxh6 gxh6=) 1 3 tb f3 ':'c7 (a clever reorganisation, but perhaps too exotic) 14 h4!? (14 0-0 �d7 1 5 .l:tab U) 14 . . . iLd7 1 5 h 5 h 6 16 �f4 and Black has made no progress although he eventually won in Resika- E.Lund, Budapest 2002; (b) 9 . . . f5!? is a suggestion of Kin dermann, whose main line (mixed with just a few of my notes) runs 10 exf6 (10 tbf3 "fiIa5 [or just 10 ... .td7] 1 1 �d2 c4!? 12 �e2 ..td7) 1 0 . . . tbf5 .
.
15 . . .:f'5 16 h5 (16 tbg3 'iVa5! 17 tbxf5 gxf5 18 'iVg3 �e8! 19 iLh6 "fic7 with good play for the exchange, I.A1masi Galyas, Budapest 2000) 16 . . . gxh5 17 'iWxh5 h6 18 g4 m 19 'iitg2! iLe8 20 "i¥h3 hxg5 2 1 f5 is unclear and hard to assess, Koch-Apicella, Chambery 1994. White has a dangerous attack but I'm skeptical of its ultimate worth. 10.2 1 1 9 'i'h4 This is relatively rare and not ter ribly ambitious, but it's a sound move and important to know, particularly since the queens will come off and the second player needs be careful not to slip into passivity. The basics of a queenless middle game are not other wise explored in this chapter, so I'll give them a detailed look here.
with: (a) 1 1 .txf5 exf5 and the f-pawn is pinned, while 12 �g5 (12 dxc5 ':'xf6 1 81
Play the French 13 liJe2 'ilVa5 14 0-0 'ilVxc5 is unclear according to K&D) 12 . . . 'iWa5 13 'iWg3 (13 liJe2?! liJxd4!) 13 . . . f4! 14 Jixf4 :xf6 1 5 liJe2 liJxd4! 16 liJxd4 cxd4 is no fun for White; (b) 1 1 £'7+ �xf7 (to keep an eye on e5) 12 'i!Vxd8 (12 �5+ 'it>g8:j: - K&D) 12 . . . .l:!.xd8 13 liJf3 (13 dxc5 e5 unclear; 13 g4 liJfe7 14 liJf3, and instead of K&D's 14 . . . �g8, 14 . . . e5! 15 liJxe5+ liJxe5 16 dxe5 .ltxg4 seems very prom ising; 13 Jig5 �e8 14 dxc5 e5 15 l:.b l h6 16 .i.d2 :e7 17 liJe2 �e6 18 0-0 l:.c8 unclear) 1 3 ... cxd4 14 �xf5 exf5 1 5 cxd4 :e8+ 16 �dl 'unclear'. Con sidering the importance of maintain ing control of e5 and d4 and the dy namic necessities of opposite-coloured bishops, one is tempted to play 16 .. .f4!?, to answer 17 �xf4 with 17 . . . �g4 and 17 h3 with 17 . . . iLf5. But simply 16 . . . h6 to prepare . . . .l:!.e4 also looks fine. 10 'iWxd8 l%.xd8
must be at least equal; (b) 1 1 �e3 .i.d7 (1 1 . . .liJh4!? is an interesting HiArcs suggestion, when 12 g4 [12 �f1 liJf5] 12 . . .f6 13 f4?! fxe5 14 fxe5 l:.f8:j: is one of many ideas) 12 h4 liJge7 1 3 g4 cxd4 14 cxd4 liJa5!? 1 5 h5 (15 �g5!? l:te8 16 h5 'is little bit more accurate, but doesn't change estimation of position as equal' Psakhis) 15 . . . a6! (15 . . . �a4 1 6 .ltg5 lId7 17 h6 g6 was strategically com plex in Bezgodov-Riazantsev, St Pe tersburg 2000; K&D like 18 f4!) 16 ..tg5 rLe8 17 h6 g6 18 a4 (18 f4 ..tb5) 18 . . . l:.ac8! 19 liJf3 liJc4, intending . . . b5 and on axb5, . . . .ltxb5. This is an ab breviated analysis of K&D's; they consider the position unclear. 1l :f8 12 liJf3 ...
12 f6 Or 12 . . . c4!? 13 i.e2 f6 14 exf6 gxf6 1 5 iLh6 Sutovsky-Drasko, Valle d'Aosta 2002; since White is going to play 0-0-0, Black should defend his d pawn by 1 5 . . . l:td8, e.g., 16 0-0-0 b5 to protect c4 - 17 h4 e5 18 dxe5 fxe5 19 �g5 lld6 20 h5 liJh8= with . . . liJ£'7 and perhaps . . . ..te6 next . 1 3 exf6 gxf6 14 ..th6 .l:te8 This was the logical course in Mit kov-Gongora Montes, Merida 2002. After 1 5 0-0-0 the natural and sound move was 15 . . . e5, and likewise after ...
..t g5 The most aggressive move and most testing, but it will lose time to .. .f6. (a) 11 liJf3 cxd4 12 cxd4 f6! is ana lysed by Kindermann. Black should be fine, e.g. , his main line runs 13 h4 fxe5 14 �xg6 e4!? 15 ..txh7+ �xh7 16 liJg5+ �g8 1 7 ..te3 e5, when Black U
1 82
Winawer Variation: Main Line with 7 "iig4 the game's 15 . . . i.. d 7 16 h4, 16 . . . e5 would have made sense, e.g., 1 7 dxe5 (17 c4?! cxd4 18 cxd5 0,ce7 19 i.. c4 0,f5 !) 17 . . . 0,cxe5 18 0,xe5 0,xe5 19 i.. e 2 .te6 20 .th5 l:ted8 and White's bishops are balanced by his weak pawns. 10.212 9 1\i'h5 0,g6 10 0,{3 1\i'c7
This introduces the Hertneck Var iation, which has proven viable for over a decade now. The queen de fends on the second rank against White's immediate threats and pre pares . . . c4, after which Black hopes to gain his usual kingside counterplay. I should mention that recently 10 . . . 'ii'e 8!? has been played a few times, with several themes that are also found after 10 . . :�c7. One might find this a fun area for exploration. After 10 . . . 'iVc7, with apologies, I must split the material a final time: 10.2 121 11 0-0 10.2122 11 i.. e 3
However, others are both impor tant and instructive: (a) 1 1 0,g5 is natural, but now we see why the queen is on c7: 1 1 . . .h6 12 0,xf7 'iVxf7 1 3 'ii'xg6 (13 i.. xg6?? 'ii'xf2+ 14 �d1 cxd4-+) 13 . . . 'ii'xg6 (not
13 . . . 'i'xf2+?? 14 �d1 +-) 14 �xg6 cxd4 1 5 cxd4 0,xd4 and: (a 1) 16 i.. e 3 0,f5 17 i.. c 5 I:!.d8 18 h4 (18 i.. xf5 exf5= due to the opposite coloured bishops; 18 0-0 b6 19 i.b4 i.. a 6 20 i.. xf5 exf5 21 .l:i.fd1 �f7=) 18 . . . b6 19 i..b 4 i.. d 7 20 0-0-0 a5 2 1 i.. c 3 l:tdc8= Baklan-Dgebuadze, Cap pelle la Grande 2002; (a2) 16 �d3 .td7 17 .tb2 0,f5 (17 . . . l:tf4 1 8 �c1 l:.f7 1 9 �b2 .l:i.f4 20 .tc1 Y2 - Y2 Gelfand-Hubner, Wijk aan Zee 1 992) 18 g3 a6 19 �d2 i..b 5 20 .txf5 Y2 - Y2 Vogt-Hertneck, Bad Lau terberg 199 1 ; (b) 1 1 h 4 is the hot move now, but it seems quite manageable: 1 1 . . .cxd4 (1 1 . . .c4? 12 0,g5 h6 13 i..xg6 fxg6 14 'ii'xg6 hxg5 1 5 hxg5 0,xe5! 16 'Ml5 m'5 Skorchenko-Potkin, Ufa 2000; and here 16 'ii'h 8+! r;i;;f7 1 7 .l:i.g7 looks strong) 12 r;i;; d 1 (12 0,g5? h6 13 i..xg6 fxg6 14 'ii'xg6 hxg5 15 hxg5 1\i'xe5+ 16 'it> d1 'i'f5) 12 . . . dxc3 1 3 .l:i.h3 (13 0,g5 h6 14 f4 0,ce7!+ - Kindermann) 13 . . . b6!? (one of two good moves: true, 13 . . . 0,ce7?! 14 0,g5 h6 15 0,f3 f5 16 i..xh6 ! led t o a dangerous attack in Rowson-McDonald, Edinburgh 2003; but Rowson's 1 3 . . . f6! looks very good, with 14 0,g5 fxg5 15 hxg5 �f7 or 14 exf6 e5 1 5 fxg7 .l:i.f6! - McDonald) 14 0,g5! (14 l:.g3?! 0,ce7! 1 5 0,g5 h6+ with the idea 16 0,xf7?? .l:.xf7 17 ..txg6 0,xg6 18 'ii'xg6 l:Ixf2-+) 14 . . . h6 15 0,f3 (15 0,xf7?? 'ii'xf7 16 .l:i.f3 0,f4!) 1 5 . . . 0,cxe5! 16 0,xe5 'iYxe5 1 7 .txg6 'iVd4+ 18 i.. d 3 'i'xf2 19 "iVe2 1\i'gl + 20 'ii'e 1 'ii'xe 1 + 2 1 r;i;;xe 1 e5 22 :g3 Wh8:j: - K&D ; Black's central pawns are awesome; (c) 1 1 a4 c4 (1 1 . . .cxd4 12 0-0 dxc3 is also playable) 12 .txg6 fxg6 resem bles the main line but White's bishop won't get to a3 so 1 1 a4 is a bit of a waste of time: 1 3 'i!Vg4 (13 �3 'iVa5 1 83
Play the French 14 �d2 i.. d 7= is normal, but easier may be 13 . . . i.. d 7!?, since 14 .i.a3 "iVa5 15 �d2 1:[£4 threatens d4 and 16 "iVg3 l:.f5 is very comfortable) 14 .td2 .i.d7 15 'ii'g4 1:[f5 (15 .. .'�je7 was balanced in J.Polgar-Gdanski, Budapest 1993) 16 h4 l:.af8 'unclear' according to K&D; compare the main lines; (d) Similar is 1 1 .i.d2 c4 12 .txg6 fxg6 13 'ieg4 i.. d 7 14 h4 lH5 15 ltJg5 l:.af8 16 h5 (16 f4 'ieb6 - Koponen) 16 . . . ltJxe5!? (16 . . . gxh5! 17 'it'xh5 h6 18 g4 l:.xg5 1 9 �xg5 �e8 20 n2 and 20 . . . hxg5= or even the promising ex change sacrifice 20 . . . .i.g6!?) 17 'i\Vg3 .l:.xf2 18 dxe5 1:[2f5 19 hxg6 "iix e5+ 20 'iexe5 1:[xe5+ 21 �dl Wiesner-Geb hardt, Hessen 1992, and Black had to play 2 1 . . .hxg6! with a messy equality.
g3 (16 ltJg5 h6 17 'iee8+ l:.f8 1 8 ltJxe6 l:txe8 19 ltJxf4 l:.e4! 20 g3 g5 - K&D; this is at least equal) 16 . . . 'ied6 1 7 'it'g5 .td7= Oll-Dolmatov, Groningen 1993; (c) 14 .lidl i.. d 7 1 5 :le I l:.afS 16 1:[b l h6:j: 1 7 �hl?! ltJce7! and White was completely tied down in Kamsky Yusupov, Belgrade 199 I . 12. . .fxg6 1 3 'ieh3 1 3 'ii'g4 is similar to the 11 .te3 main line, but after 1 3 . . . .i.d7!, White's best plan in that line is h4-h5, which has now lost its point without the rook on h I . The idea h4-h5 is also without value in the line 13 . . . 'ii'f7 !? 14 ltJg5 'it'e8 .
10.2 121 1 1 0-0 c4 12 �xg6 Not 12 ltJg5?? h6 13 ltJxf7 cxd3. 12 �e2 has been played many times but simply allows Black to free his game and gain activity after 12 .. .f6 (':j:' Knaak) 1 3 exf6 (13 a4 .i.d7 1 4 exf6 l:.xf6 1 5 ltJg5 h6 16 ltJh3 'iea5+ Shetty L.Ravi, Indian Ch 1 994) 13 . . J:txf6 13 h61? Two other moves are simpler: (a) 13 . . . 'it'a5!? stops any a4 and .ta3 plan, e.g., 14 i.. d2 .lid7!? (14 . . . ltJe7 or 14 . . . "iVa4!? 1 5 11a2 i.. d 7!?=, which relies upon the sequence 16 ltJg5?! h6 1 7 ltJxe6 'iVa5! 18 i..xh6 ]:tf7! 19 f4 'iVb6:j:) 15 ltJg5 h6 16 ltJf3 l:.f7 17 'ieg4 g5=; (b) 1 3 . . . .lid7 is straightforward: 14 a4 'it'a5 1 5 .lid2 l:tf5 1 6 ltJh4 1:!.f7:j: . 14 a4! 14 'ieg4 'ii'f7 15 ltJel!? (15 a4! 'i!Vf5 16 'i!Vxf5 gxf5 1 7 h4 b6 is equal) 15 . . . �d7 16 'iee2? (16 f4 'i!Vf5 or 16 . . . g5 !? 17 1:[f3 'it'g6 18 fxg5 :xf3 19 gxf3 h5! 20 'ieg2 ltJe7 ! = ; the pawn is irrelevant) 16 . . . g5 17 llb l b6 18 .i.d2 ...
(a) 14 g3 i.. d 7 (14 . . . h6) 1 5 i.. e 3?! (15 �g5 - Hertneck; 1 5 . . . 1:[f5!:j:) 1 5 . . . l:laf8 16 �g5 J.Polgar-Hertneck, Cologne TV 1991; and ECO gives 16 . . . i.. e 8!+ ; (b) 14 �g5 ltJf4! 1 5 �xf4 'ii'xf4 16 1 84
Winawer Variation: Main Line with 7 'iYg4 0,e7 19 :dl 'i'g6 20 �cl .l:[f7 2 1 f3 .l:[af8+ R.Rodriguez-De la Villa Gar cia, Alcobendas 1993. 14 ... g5? Loosening. Black's play should be pretty easy here, e.g., 14 .. .'1Wf7 15 i.. a 3 .l:[d8 16 i.d6 b6 17 'iVg4 'iVf5=. The opposite-coloured bishops are particu larly drawish in these positions. 15 'iVg4! Not 15 �a3?! :f4!, but 15 'iVg3!? preparing h4 is interesting: 1 5 . . . 0,e7! 16 i.a3 0,f5 1 7 \i!Vg4 l:tf7= 18 h4? 0,xh4 19 0,xh4 .l:[f4 - Koponen. 15 ... \i!Vf7 16 h4! Better than 16 i.a3 'i'f5=. 1 6 ... gxh4 1 7 i.xh6 Ashley-J.Watson, Seattle 2003. Now White will win a pawn with a nominal edge, but the opposite bish ops and his own weak pawn structure will make a draw the correct result. The game continued 17 . . . 0,e7 (or 17 . . . h3 18 g3 0,e7;!;) 18 0,xh4 0,f5 19 �cl ! 0,xh4 20 'ilxh4 'iVg6 2 1 i.. a 3 l:i.e8 22 l:.a2 i.. d 7 23 i.d6 'iYf5 (intending . . . <J:tf7 and . . . .l:.h8) 24 f4 'iVh7! 25 \i!Vg5 'iWf5 with an eventual draw after the exchange of queens. But Black's al ternatives on moves 13 and 14 would have avoided this discomfort.
Protecting f2 and thus threatening 0,g5. There are over 200 games with this move in my largest database! 1 l ... c4 1 1 . . .0,ce7 is also played; one area to look at is the line 12 h4 i.d7 1 3 'iWg4!? f5 14 'iVh3 ! , which has recently won two top-level games, including what I think is the stem game Nataf Pelletier, Cap d'Agde 2002. 12 �xg6 (a) 12 0,g5? h6 13 0,xfl cxd3 14 0,xh6+ gxh6 15 'iYxg6+ 'iYg7+ ; (b) 12 �e2 strongly resembles 12 0-0 c4 1 3 �e2. One example: 12 ... f6 1 3 h4!? ..td7 14 g3 0,ce7!? 1 5 �g4 f5 (15 . . . fxe5! 16 h5 exd4 1 7 hxg6 dxe3 18 gxh7+ �h8 19 fxe3 e5+) 16 'iVh3 h6 1 7 h 5 0,h8 1 8 0,h4 i.. a 4:j: Minasian Apicella, Paris 1 994. 1 2 .fxg6 1 3 'i'g4 13 'iVh3 is like 1 1 0-0 c4 12 �xg6 fxg6 13 'ilVh3 above except that the bishop on e3 is poorly placed. True, h4-h5 would have more point, but the queen on h3 is in the way and White's play is slow. 1 3 ...�d7 ..
10.2122 1 1 i.. e 3
This may not be better but it is less worked out and contains a different set of dynamic possibilities than the more popular 13 . . . 'iVf7. In that line most top-level games recently seem to 1 85
Play the French demonstrate that Black (with very accurate move orders) can sacrifice the exchange and/or block the posi tion to draw! That's not bad if you're willing to study and keep up, but of course it could change as theory charges into the 20+ move range. Here's a couple of exemplary recent games that illustrate the exchange sacrifice theme, one that is ubiqui tous in modern French practice: 14 ttJg5 'i!Ve8 1 5 h4 h6 16 ttJh3 b5
sentially did nothing for the next 20 moves and a draw was agreed. Fasci nating and possibly even useful as a weapon, but this is not what we're looking for as Black. 14 h4 14 ttJg5 h6 1 5 ttJh3 ttJe7 16 a4 l:tf5 17 ttJf4 'it>h7 18 h4 '!:!'af8= Stefansson Hjartarson, Iceland 1 994. 14 lif5 1 5 h5 1 5 ttJg5 has been played often but with little effect:
1 7 'i!Ve2 (an amazing game was 17 ttJf4 ttJe7 18 'it>d2 a5 19 'i!Vh3 �d7 20 l:tab l 'i!Vf7 2 1 f3 .l:.ab8 22 ttJe2 ttJc6 23 .l:i.b2 b4! 24 axb4 axb4 25 cxb4 c3+ 26 ttJxc3 ttJa5! 27 'it>e2 ttJc4 28 l:!.b3 ttJxe3 29 'it>xe3 'i!Vf4+ 30 'it>d3 l:tfc8 31 'iVg4 J::[xc3+! 32 'it>xc3 l:tc8+ 33 'it>b2 'iVd2 34 .l:i.cl 'it>h7 Anand-Yusupov, Bastia 2002; White is the exchange and two pawns up(!), but apart from the struc tural issues moves like . . . h5 and . . . �a4 are in the air; the game was eventually drawn) 17 . . . a5 18 g4 �d7 19 h5 gxh5 20 J::1. g 1 ! ttJe7! 21 gxh5 'it>h7 22 "iii g4 ttJf5 23 'it>d2 "ile7 24 'iVg6+ 'it>h8 25 ttJf4 ttJh4 26 'i!Vg4 .l:txf4! (again! the light squares are often full compensation in such positions) 27 i.. xf4 .l:i.f8 28 lIhl ttJf5 29 i.. e 3 ..te8 30 f4 i.. f7 3 1 �f2 lIb8 32 'i!Vh3 'it>h7 Leko Z.Almasi, Budapest 2003. Black es-
15 . . . l:taf8 (another good order is 15 .. :iVa5 16 i.. d 2 .!:!.af8, e.g., 17 0-0 ttJd8! 18 "ile2 i.. a 4 1 9 .l:i.acl l:te8 20 g4 l:tff8 2 1 'it>g2 ttJf7= Kovacevic-Sparic, Yugoslavia 1 994; recently Dirr tried 15 . . . h6 16 ttJh3 ..te8 17 'iVe2 l:tf8 18 g4 with a position resembling the 13 . . . 'iVf7 lines above, although I like the bishop on e8 better; the game went 18 . . . ttJe7 19 'i¥d2 l:tc8!? 20 ttJf4 'iVd7 2 1 'it>e2 l:tc6! 22 J::[ a gl l:tb6 Ni jboer-Dirr, Godesberg 2002 with complications in which Black looks better, although the game was drawn; 22 . . . l:ta6! was probably more accurate) 16 h5 (16 'iVe2 h6 17 ttJh3 'i¥d8 18 g3 b5 19 'i!Vd2 'i!Ve7 20 'it>e2 a5 with good play, David-Enders, Godes berg 2000) 16 . . . 'iVa5 (or 16 . . . h6 !? 1 7 hxg6 'i!Va5 18 0 - 0 "iVxc3 1 9 ttJh7 lId8=) 17 �d2 h6 18 ttJh3 (18 ttJf3 g5=i= -
1 86
...
Winawer Variation: Main Line with 7 "iIIg4 K&D) 18 . . . gxh5 19 lig3 �hn Ling nau-Kipper, Germany 1 998. 1 5 ... gxh5 1 6 Ibh5 'uxh5!? This was originally queried, but I feel that it's a good solution. Never theless, Black has an alternative in 16 . . . l:.af8, which is dicey but may suf fice theoretically. Let me just men tion a few ideas: (a) 1 7 liJg5 �a5! 18 i.d2 .l:i.f4! 19 'iWh3 h6! 20 l1xh6! gxh6 2 1 �xh6 114f7 ! , and according to theory (ex tended by K&D), the game will be drawn or end in an unclear position on move 35!; (h) 1 7 �d2!?
(h I) 1 7 ... lIxh5 18 "iWxh5 �e8 19 �3 l::txf3 ! ? 20 "iWxf3 �g6 is a famous type of exchange sacrifice played sev eral times right here. A brilliant con ception, but nevertheless I don't fully trust this idea and in any case only White can win. Concretely, Black must deal with lines like 21 �h3 "iWd7 22 .l:i.b l liJd8 23 �g5 liJf7 24 �h4 b6 25 lIb4 liJh6 26 g4 "iWf7 27 'iWg3 �h8 28 f3 'iYe8 29 a4 with nagging pres sure, Makovsky-Raidna, email 2000; (h2) Surprisingly I haven't seen 17 . . . .Jie8! mentioned, but I like it: 18 ,Uh2 .Jig6!? (18 ... 'iVb6!? was played successfully by Hillarp Persson and is probably okay but needs more analy-
sis) . Then play might go: 19 "iII g3 !? (19 liJh4 iLh5= 20 "illh 3?! lIxf2+ 2 1 .Jixf2 lIxf2+ 22 \it>el lIxc2 23 liJf3! .Jig6 24 Wxe6+ Wf7 25 lic8+ lie8;) 19 . . . .l:i.xf3 20 gxf3 'iVb6 ! ; (c) 1 7 .l:i.h3 .Jie8 (17 . . . Wa5 18 �d2 .Jie8 transposes) . Black's defence is not easy after 1 8 liJg5 (18 �d2? ! h5 1 9 �h4 J:!.xf3 20 gxf3 liJe7 2 1 a4 liJf5 22 �g5 �e7 with a total bockade and awful white pawns, Rowson-Barsov, York 2000) 18 . . . �a5 19 �d2 and Black must again play an exchange sacrifice 19 . . J:txg5 20 �xg5 iLg6 with compensation, but how much is not clear. It's better to leave one pair of rooks on if possible, so this position might be okay for Black. Neverthe less, 16 . . . lIxh5 100ks easier to me. 1 7 �xh5
1 7 ... .Jie8!? As far as I know, this move hasn't been played or suggested; in fact it has been thought to be a serious mis take. The normal continuation is 17 .. JH8, but this often depends upon the forementioned exchange sacrifice, e.g., 18 �3 lIxf3!? 19 Wxf3 .Jie8 20 �3 'iWd7;!; ; compare the last note. White has begun to win this type of position. 1 8 �h3 K&D say that White has a clear
187
Play the French advantage now. But I doubt that there's anything wrong with 18 . ..tiJd8! Black plans to play . . . .i.g6 next (without having to sacrifice an ex change!). Here's some analysis:
1 9 ltJg5 This is almost certainly the reason that no one has tried 17 . . . i.e8, but it backfIres. White has better alterna tives: (a) 19 .ltg5!? i.. fl 20 �d2! (20 i.. x d8 'ii'xd8 21 �d2 'iWe7=i=) 20 . . . 'iWd7!? (to prepare . . . i.. g6) 2 1 l:i.hl .i.g6 22 ltJh4 'iWfl= ; (b) 1 9 �d2 also equalises, if only because the tactics work out to a draw: 19 . . . i.. g6 20 l1hl (20 ltJh4 'iWfl! , although the opposite-coloured bish0ps will leave few prospects) 20 . . . 'ii'b 6!? (or again 20 . . . 1IVfl; Black's plan might be expansion on the queenside; but 21 ltJg5!? 'iWf5 leaves only an optical advantage, and 2 1 . . :tWe8!? 22 ltJxh7 'ii'a 4! only draws after 23 ltJf6+! <3ifl 24 1IVh8! 'ii'xc2+ 25 <3iel with the idea 25 ... gxf6 26 exf6! 'iVb 1+) 2 1 ltJh4 i.xc2 !? 22 �xc2 'ilVb3+ 23 �d2 1IVb2+ 24 �dl 'ii'xc3!? (24 ... 'iib 3+=) 25 ltJf3 'ii'd 3+ 26 �el h6 (26 ... 'iVb l +=) 27 i.xh6 gxh6 28 1IVg4+ �f8 29 l:txh6 'itb l += . 19 ... h6 20 ltJxe6
1 88
20 ltJf3 .i.g6 ! + . 20 ...'i¥a5! White may still survive this posi tion, but he's lost the chance for full equality: 2 1 .i.xh6!? 2 1 <3id2? i.d7 wins for Black, but 21 ltJxd8! 1IVxc3+ 22 �e2 'ii'x c2+ 23 <3if1 l1xd8 24 11Ve6+ seems to minimize the white disadvantage: 24 . . . �h7 (24 . . . i.fl 25 'ii'e 7 l::t e 8 26 'ii'xb7 'ii'c 3=i=) 25 1IVe7 1IVb2 26 :el 'itb6 27 e6! and 27 . . . 1IVd6=i= will be hard to win, but Black must avoid 27 . . . i.. g6? 28 i.. xh6! . 21 . . .ltJxe6 22 'ii'x e6+ i.. f7 23 'ii'g4 'ii'x e3+ 24 �e 2 24 �dl 'ii'x al + 25 �e2 i.. h 5! 26 'ii'xh5 'ii'x d4 will win. 24 .. .'i'xe2+ 25 �el 1IVh7 26 .i. e l 'iVhl + 27 �d2 1IVfl !+ Black is threatening both ... 'iWxf2+ and . . . c3+ and will gain a large ad vantage. If this analysis is correct, Black can be happy with the most critical lines of the Hertneck Variation. 10.22 8 ... f5
Rather than supply a complicated repertoire with this move, I'll recom mend particular byways to simplify one's use of it. In particular, we'll use favourable transpositions from the
Winawer Variation: Main Line with 7 'iig4 8 . . . ltJbc6 variation to help us. g exf6 (a) 9 'i!ih3?! b6 (or 9 . . :iVa5 10 ..td2 ltJbc6 1 1 ltJf3 'iia 4=, transposing to 8 ltJf3 ltJbc6 9 �d3 f5 10 'i'h3 above) 10 ltJf3 'ii'd 7 (to protect e6 before playing . . . ..ta6) 1 1 0-0 �a6 12 a4 ..txd3 13 cxd3 cxd4 14 ltJxd4 ltJbc6 1 5 ltJxc6 ltJxc6 16 �a3 l:tfc8 17 iLd6 ltJd8f in tending . . . ltJf7 - analysis by Kinder mann and Dirr; (b) 9 'iig 3 ltJbc6 10 ltJf3 transposes to 8 ltJf3 ltJbc6 9 i.. d 3 f5 10 'iig 3. 9 l1xf6
Black although not resignable, e.g., 1 8 h4 (18 l:tgl? l:taf8 19 ltJxg5 ltJxe2 20 'ii'xe2 e5!) 18 . . . gxh4 19 .i:l.gl ltJg6 20 ltJxh4 ltJce7+ . lO l:!.f7 ...
•••
lO �g5 10 'i!ih5 h6 11 ltJf3 (1 1 g4!? can be wild: 1 1 . . .ltJbc6 12 g5 g6! 13 'ilVxh6 ':'f7 threatens . . . l:th7, and 14 .i.xg6 .t:1.g7 15 .i.d3 e5! 16 dxe5 ltJxe5 1 7 'i'h4 ltJf5 18 'ilVf4 ltJg6 1 9 'iid2 ltJgh4+ is analysis by Tischbierek; K&D continue 20 iLe2 b6 2 1 ltJf3 .i.b7 22 ltJxh4 ltJxh4 23 l:tgl 'ii'e 7 with attack; it's also worth pointing out that 15 . . . c4 16 i.e2 ':'h7 17 'ilVg6 l:.g7 was a draw) 1 1 . . .c4!? (1 1 . . .ltJbc6 transposes to the 8 . . . ltJbc6 line again) 12 1i.. e 2 ltJbc6 13 i.e3 iLd7 14 g4?! 'iia 5 1 5 1i.. d 2 ltJg6 16 g5 0 - 1 Zelcic-Dolmatov, Istanbul 2003. Unless it's a typo, White's res ignation seems to be based on 16 . . .ltJf4! 17 'iig4 (17 1i.. xf4? 'iWxc3+) 17 . . . hxg5, which is very good for
1 1 'iYh5 (a) 1 1 ltJf3 e5!? (1 1 . . .ltJbc6 is a well known line satisfactory for Black, but the theory on it is immense) 12 iLxh7+
1 89
Play the French small advantage) 21 .l:.ab l b6 22 iLe4 iLbn Schreber-Knaus, email 2000. 1 1...g6
of these replies look good: (a) 13 . . . liJec6 14 liJf3, and 14 . . . liJd7! is a new and paradoxical idea.
12 "iid 1 12 "iih4 "iia 5! 13 i.. d2 (13 liJe2? liJbc6 threatens . . . c4, but then 14 dxc5 e5+ is not what White wants) 13 . . . liJbc6 14 liJf3 and a simple solution is 1 4 . . . c4 1 5 i.. e2 liJf5 16 'iWg5 i.. d 7 with the idea 1 7 h4 .l:.af8! 18 h 5 h6!+. 12 ..."iia 5
Although blocking off his own pieces, Black insists upon enforcing . . . e5: 1 5 0-0 (15 h4 e5 16 dxe5 liJdxe5 17 liJxe5 liJxe5 18 h5 i.. g4! :j: ; 15 i.. e2 tries to prevent . . . e5 by indirectly covering d5, but after 1 5 . . . liJf6, the threat of . . . liJe4 practically forces White into 16 i..xf6 l:txf6, when 1 7 O-O? i s met b y 1 7 . . . cxd4 and otherwise White is a bit stuck for a move; he might therefore try 17 dxc5 'iVxc5 18 0-0 e5 19 .l:.ab l , but 19 ... a6! , stopping l:tb5, seems to retain some advantage) 15 . . . cxd4! (15 . . . e5 16 c4!? 1lVxd2 17 liJxd2 e4 1 8 cxd5 exd3 19 dxc6 bxc6 20 dxc5;t) 16 cxd4 'iVxd2 17 i.. xd2 e5 (17 . . . 'u'xf3 !? 18 gxf3 liJxd4 19 'it>g2 liJe5 20 f4 liJxd3 21 cxd3 liJb3 22 l:1.a2 iLd7 ought to give Black enough for the exchange: a pawn and the opportu nity to exploit White's isolated and weak pawns) 18 liJxe5 (18 liJg5!?) 18 . . . liJdxe5 1 9 dxe5 liJxe5 20 .l::i.fe l (20 iLe2 iLf5=) 20 . . . liJxd3 2 1 1:1e8+ 1:1f8 22 .l:.xf8+ 'it>xf8 23 cxd3 iLf5 24 llcl .l:.c8 25 l:txc8+ Y2 - Y2 Leko-Ivanchuk, Mon aco (rapid) 2002; (b) Kindermann analyses 1 3 ... c4! , which has apparently not yet re ceived tests but seems one of Black's
Kindermann's move, easier and more fun than the also-satisfactory line with 12 . . . liJbc6 1 3 liJf3 'iVf8. 1 3 ..id2 Of course 13 liJe2?? loses to 13 . . . c4. But 13 1lVd2 is a logical positional at tempt that has caused Black some real problems in the past. It still hasn't been fully worked out; but both 1 90
Winawer Variation: Main Line with 7 'i!Vg4 most interesting responses to 1 3 'i!Vd2: 14 i.. e 2 l2Jf5 15 h4!? (15 l2Jf3? l2Jd6! 16 l2Je5 l2Je4+ ; 15 i.. f4 l2Jc6 and 16 l2Jf3? l2Jfxd4! or 16 l2Jh3 e5!? 1 7 dxe5 l2Jh4 18 0-0 i..xh3 19 gxh3 lie8� ; 1 5 Ji.f3 l2Jc6 16 l2Je2 e5 17 dxe5 l2Jxe5 18 i..x d5 l2Jd3+ 1 9 cxd3 'iWxd5 unclear all from Kindermann) 1 5 . . . l2Jc6 (maybe Black could try 15 . . . h6!? 16 Ji.f4 l2Jc6 unclear, again threatening . . . l2Jfxd4; then play might go 1 7 g3! �d7 1 8 l2Jf3 l:taf8 19 0-0 b5!) 1 6 h5 e5 1 7 hxg6 hxg6 18 dxe5 l2Jxe5 19 l2Jf3 l2Jxf3+ 20 i.xf3 Ji.e6= . 13 ... l2Jbc6 14 l2Jf3 'iVc7! Preparing . . . e5. 14 . . . c4 closes the centre; it has been played with vary ing success, but theoretically favours White slightly.
15 0-0 (a) 15 l2Jg5 lif6 16 dxc5 (16 f4? cxd4+) 16 . . . e5 17 c4 e4 transposes to 'b'; (b) An important recent game went 15 dxc5 e5 16 l2Jg5 l:tf8 17 c4 e4 18 cxd5! (18 SLe2 'i!Ve5=) 1 8 ... exd3 1 9 d6 'iVd7 20 dxe7 'i!Vxe7+ 2 1 i.. e 3 Pelletier Lutz, Biel 2003; and Kindermann suggests 2 1 . . .lif5 ! 22 l2Jf3 (22 h4 h6 23 l2Jf3 Ji.e6) 22 . . . i.. e 6 23 0-0 i.. d 5 24 l2Jd2 lih5! 25 liel 'i!Vh4 26 l2Jel and 26 . . . dxc2 27 'iWxc2 l2Je5 100ks best, intending 28 f4 l2Jc4. 1 5 ... e5
1 6 l2Jg5 16 l2Jxe5 l2Jxe5 17 dxe5 'i!Vxe5 18 c4 Ji.f5 (18 . . . d4 also looks equal) 19 :el (19 i..xf5 l2Jxf5 20 l:tel 'i!Vd4= Stefans son-Potkin, Pardubice 2000) 19 . . . 'i!Vf6 20 'iWf3 Ji.xd3 2 1 'iWxf6 l:txf6 22 cxd3 l2Jf5 23 lie5 dxc4 24 dxc4 l:td6= Lutz Kindermann, Porz 1 997. 1 6 ...:'f8 1 7 c4 exd4! 18 l:i.el 18 cxd5 l2Je5 (18 . . .l2Jxd5 19 i.. c 4 l2Jce7 looks reasonable) 19 SLe4 l2Jg4! 20 g3 (20 f4 l2Jf6!) 20 . . . l2Jf6 21 i.. f4 'iWd7 (2 1 . . .'iWd8 22 d6 l2Jed5 23 i.. e 5! with attack - Kindermann, although simplification via 23 . . . l2Jxe4 24 l2Jxe4 'i!Ve8!? 22 c4 h6 looks safe enough) 22 c4 h6 23 l2Je6 l2Jxe4 24 l2Jxf8 Iifo>xf8 25 Ji.xh6+ 1ifo>f7 'unclear' - Kindermann. 18 ...Ji.f5 19 cxd5 l2Jxd5 20 i.. c 4 liad8 White has two bishops and will try to exploit the weakened kingside dark squares. Black has space and is centralised, so his prospects are good. 21 'i'f3 'i'd6 22 'iWb3!? A dynamic balance arises after 22 i..f4 'i!Vd7 23 g4!? h6! 24 gxf5 hxg5 25 Ji.xg5 l:i.xf5=. After 22 'iWb3, Hracek Kindermann, Bad Homburg 1997 went 22 . . . b6 23 l2Je4 i..xe4 24 ':'xe4 l2Ja5 25 Ji.xa5 bxa5 26 lib l Iifo>g7 27 'i!Vh3 lid7 (27 ... l2Jb6! 28 lie6 'i!Vd7 was worth a try) 28 l:tbel Y2 - Y2 .
1 91
Chapter Eleven Winawer Variation : Positional Lines
1 e4 e6 2 d 4 d5 3 t'Llc3 .i.b4 4 e5 c5 5 a3 �xc3+ 6 bxc3 t'Lle7 In this chapter, White chooses one of several positional continuations: 1 1 . 1 7 t'Llf3 1 1.2 7 a4 1 1 .3 7 h4
In contrast to our experience with 7 'iVg4, White keeps his pieces close to the centre and doesn't immediately try to get at Black's king (even with Black's king on the queenside, 7 h4 introduces a positional attack via h5h6). To complement his centre, White has numerous ways to attack on both wings, e.g., h4-h5-h6, or a4 with .i.a3 and a5, or even dxc5 with t'Lld4, hit ting both sides of the board. For his part, Black can put pressure down the c-file, blockade the queenside by . . . .t.d7-a4 or . . . 'fia5-a4 (with or with out . . . c4) , and at some point he usu ally attacks White's centre by . . . f6. The extreme flexibility of both sides' strategies has produced fascinating strategic battles over many years. Arguably, the Positional Winawer has been one of the most intellectu1 92
ally fruitful openings not just of the French Defence but of modern chess. That said, we have probably seen less of an evolution of these lines in the past few years than since the time of Botvinnik. A decline in top-level in terest has been partly due to a resur gence of 7 'fig4, but also to the im pression that the positional lines no longer offer as much challenge for Black. His strategies and concrete defensive schemes look very good at the moment, although this can of course change at any time. Aside from the above main moves above, White has these rare ones: (a) 7 f4 cuts off the c l bishop but is playable. One answer is 7 . . . 'fia5 (or 7 . . . t'Llbc6 8 t'Llf3 t'Llf5) 8 'i1id2 (8 .i.d2 t'Llbc6 9 t'Llf3 cxd4 10 cxd4 'i1ia4 1 1 .i.c3 b6= intending . . . .i.a6 - Korchnoi) 8 . . . .i.d7 (8 . . . b6= intending . . . .i.a6) 9 a4 t'Llbc6, e.g., 10 t'Llf3 cxd4 1 1 cxd4 'itxd2+ 12 'it>xd2 t'Llf5 13 .i.b2 h5 14 a5 (versus . . . t'Lla5) 14 . . . a6 with queenside pressure after . . . l:tc8, . . . t'Lla7, . . . �b5 etc . ; (b) 7 .i.d3!? is a bit committal but has been used by Short and should be taken seriously:
Winawer Variation: Positional Lines (h I) 7 . . . ttJbc6 8 'iVg4 (8 ttJf3 trans poses to 7 ttJf3 ttJbc6 8 i.d3) 8 . . . 'iVa5 (8 . . . c4 9 i.e2 0-0 10 h4 f6! - Dvoret sky; then ECO gives 10 ttJf3 f5!? 1 1 'iVg3=) 9 i.d2 c 4 10 i.e2 0-0 (10 . . . ttJf5 1 1 ttJf3 i.d7 12 ttJg5 h6 1 3 'iVh5 ,S,f8 14 ttJh7 and 14 ... ,S,h8= or 1 4 ... ,S,g8 1 5 g4 ttJfe7 unclear, Velimirovic-Timman, Saraj evo 1 984) 1 1 h4 f6 12 f4 (other wise Black breaks down the centre) 12 . . . 'iVa4 13 i.d1 'iVb5 (threatening . . . 'iVb2) 14 i.el 'iVa5 1 5 i.d2 and in Short-Shulman, Dhaka 1999 Black repeated by 1 5 . . . 'iVb5 16 i.c 1 'iVa5 etc.; but 15 . . . 'iVb6 would have been very interesting, since . . . 'iVb2 is still a threat and now 16 i.c1? fails to 16 . . . fxe5 1 7 fxe5 ttJxe5!; (h2) 7 . . . 'iVa5 8 i.d2 ttJbc6 (or 8 . . . c4 9 i.fl 'iVa4) 9 'iVg4!? c4 10 i.e2 0-0 1 1 ttJf3 'iVa4 1 2 's'a2 f6 1 3 exf6 's'xf6=; (h3) 7 ... 'iVc7 8 ,S,b 1 ttJbc6 9 'iVg4 Short-Poulton, Birmingham 2002, and as pointed out in CBM, Black should play simply 9 . . . 0-0! 10 'iVh5 (10 ttJf3 f5 11 exf6 ,S,xf6 12 i.g5 e5!:(:) 1O . . . ttJg6 1 1 ttJf3 c4 12 i.xg6 fxg6, practically a tempo up on the main line of our 7 'iVg4 0-0 chapter, since 's'b 1 looks wasted; (c) 7 ttJe2!? 'iVc7! (my suggestion, stopping ttJf4) 8 ,S,b 1 (8 g3 b6! 9 i.g2 i.a6 10 0-0 ttJbc6:(:) 8 . . . b6 (8 . . . ttJd7 and . . . ttJb6 is promising, or simply 8 . . . 0-0) 9 ttJg3?! cxd4 10 cxd4 'iVc3+ 1 1 i.d2?! (a strange pawn sacrifice) 1 1 . . . 'iVxd4 12 f4 Spassky-Bruch, France 1994; and 12 . . . 0-0 is obvious and strong, with the idea . . . i.a6. 1 1 . 1 7 ttJf3 The obvious move and still the most popular. 7 . .ttJbc6 For this edition I give a different move order, in part just to infuse .
some originality into the proposed lines and in part to more comfortably meet h4-h5 systems .
A look at the books will show that 7 . . . ttJbc6 is not the most common con tinuation, perhaps because it doesn't put any immediate questions to White (7 . . . 'iVa5 attacks c3, for exam ple, and 7 . . . i.d 7 has the idea of . . . i.a4) . It's very difficult to say in concrete, demonstrable terms what the advantages and disadvantages are of the various possible move or ders here. For example, the com monly played moves 7 . . . 'iVa5, 7 . . . i.d7, 7 . . . 'iVc7, and 7 . . . b6 (even 7 . . . h6 has been played at the top levels) can all be met by 3 to 5 legitimate strategies, each of which in turn can be an swered by a new set of possible ideas that don't resolve themselves until well into the middlegame. This may seem true of many openings, but the difference here is the vague nature of the positions that result. Practically every type of strategic decision taken by either side is good in some posi tions and bad in very similar ones. On top of that, both sides have nu merous possibilities for transposition (as well as whether to allow transpo sitions) . It's interesting, for example, to see some general comments out of 1 93
Play the French two recent books covermg the Win awer by first rate authors who under stand these positions in depth. Neil McDonald finds himself 'rather puz zled as to the reason for' 7 a4, a move played by Smyslov, Fischer, Spassky and many others, and then offers as a possible justification for the move a rare manoeuvre that was employed by none of them! Pedersen on the other hand says that he prefers 7 a4 and considers 7 lLlf3 the easiest move to equalise against (suggesting 7 . . . b6). Without trying to resolve the general question, I do like 7 . . . b6 (a handy system to have at the ready), but it's not necessarily better than other solutions. Probably any of Black's 7th moves are ultimately playable (including 7 . . . i.d7, given in the last edition) . But none of them escape the necessity for solving origi nal strategic problems over the board that aren't found (yet) in the theoreti cal books. The important thing is to be comfortable with the most common types of positions that arise and to be willing to compete in strategically dense battles. That said, there are certain specific positional and tactical traps that must be avoided, as with any opening. After 7 . . . lLlbc6, we look at:
choose the latter move. I will try to include all the main h4-h5 systems under 7 h4, although the ideas are so popular and flexible that irregular ones sometimes pop up in other sec tions. As usual there are some odds and ends which need to be considered be fore we look at the main lines: (a) 8 g3 is rather slow and has sev eral answers, e.g., 8 .. :�¥a5 (or 8 . . . 0-0 9 i.g2 f6=) 9 i.d2 'iVa4 (a traditional manoeuvre; or 9 . . . cxd4 10 cxd4 'iVa4) 10 'iVb l (10 dxc5?! lLlxe5! 1 1 lLlxe5 'iVe4+ 12 'iVe2 'iVxhl 13 'iVb5+ �f8) 10 . . . c4 1 1 i.g2 i.d7 12 0-0 (12 'iVxb7?? J:i:b8 13 'iVc7 'iVa6! 14 0-0 J:i:b6 15 'iVd6 lLlf5 16 'iVc5 J:i:b5) 12 . . . 0-0-0 13 lLlg5 J:i:df8 14 f4 f6 15 lLlf3 fxe5 16 fxe5 h6 with good attacking chances since Black's king is secure; (b) S J:i:b l 'iVa5 (the alternative S . 'iVc7 is also a natural continuation) 9 J:i:b5 is supposed to be good, but af ter 9 . . . 'iVc7! (the right square for the queen anyway) 10 J:i:xc5? (10 i.d3?! cxd4 11 cxd4 lLlxd4!) 10 . . . a6 and . . . b6 follows. B.Socko-Bartel, Warsaw 2001 went 9 'iVd2 c4 (9 . . . 0-0 keeps other options open) 10 h4, and Black should just play the natural 10 . . . 0-0 1 1 h5 h6 12 i.e2 f6 13 exf6 J:i:xf6= ; (c) S dxc5 leaves the valuable e pawn up for grabs: S . . 'iVa5 (S . . . lLlg6!? 9 �b5 'iVa5 10 i.xc6+ bxc6 11 'iVd4 i.a6 unclear) 9 .lid2 lLlg6 10 c4 'iVxc5 1 1 cxd5 exd5 12 i.e3 'iVa5+ 13 i.d2 'iVcn . . .
.
1 1. 1 1 8 i.d3 1 1 .12 8 i.e2
Two very important moves, 8 h4 and 8 a4, are dealt with in the next two sections by transposition via 7 a4 (1 1 .2) and 7 h4 (1 1 . 3) . But the trans position involving 8 h4 is not a simple matter. If Black had chosen 7 . . . i.d7 or 7 . . :¥ilc7, for example, the move 8 h4 poses more difficulties than after 7 . . . lLlbc6, which is one good reason to 1 94
1 1 . 1 1 8 i.d3 This formerly rare move is being looked at more seriously these days. It can transpose to various other lines, but in this section we look at independent ideas involving the gain of tempo via . . . c4.
Winawer Variation: Positional Lines
S JWa5 Also fine is 8 . . . c4 9 li.f1 (rerouting the bishop rather than playing i.e2/0-0/l:tel/i.f1 and then g3) 9 .. JWa5 (or 9 . . . 0-0 10 g3 h6 1 1 li.g2 f6 12 exf6 ':'xf6=) 10 'iWd2 li.d7 1 1 g3 h6 12 i.h3 0-0-0 13 0-0 Rizsonkov-Papp, Szeged 1998; and 13 . . . l:!.df8 intending . . .f6 or .. .f5 with . . . g5 was one idea; but com pare the main line. ..
9 ..id2 (a) 9 0-0 is generally not played be cause it allows the draw after 9 . . . c4 10 i.d3 'iWxc3 1 1 .i.d2 'i!Vb2 12 I!b 1 \\!Vxa3 1 3 .l:.al etc. But Korchnoi and others have played for the win as Black by 9 . . . c4 10 i.. e 2 iLd7!? (10 . . . 'iWa4!? is another way to go, e.g., 1 1 lLlg5 h6 12 lLlh3 �d7 1 3 f4!? 0-0-0 14 iLf3 f6 1 5 'iie 2 N.Rogers-Edelman,
USA 1987, and now 15 . . . fxe5 16 fxe5 g5 looks right, intending either . . . i.. e 8-g6 or . . JldfS and . . . lLlf5) 1 1 a4 lLlc8 (Rogers and Edelman had an other game with 1 1 . . .f6 12 'iWd2 fxe5 1 3 lLlxe5 lLlxe5 14 dxe5 0-0 15 �a3 l:tf7=) 12 'iVd2 lLlb6 13 \\!Vg5 (or per haps 13 'i¥f4!? 'i'xc3, e.g., 14 �a3 \\!Vxc2!? 15 i.. d l 'i'f5 16 'tIid2 f6 17 i.. c 2 \\!Vh5 18 exf6 gxf6 19 a5 lLlc8 20 \\!Vf4 lLl8e7 2 1 \\!Vxf6 l:tfS 22 'iWg7 0-0-0 23 \\!Vxh7 \\!Vxh7 24 .txh7 l:th8 25 i.. c 2 .l:tdg8 is extremely complicated, pro bably dynamically equal) 13 . . . .l:.g8 14 l:r.a3 lLlxa4 (14 . . . h6 15 \\!Vh5 lLlxa4 16 g3 b5 17 i.. d2 'i'd8 18 l:tb l 'iWe7 19 l:r.aal a6 20 lLlh4= Van der Sterren Korchnoi, Wijk aan Zee 1978, as sessed as equal by Short) 1 5 iLd2 b5 1 6 1Ha l 'iid 8 1 7 'iWh5 h6 1 8 lLlh4 \\!Ven Lovholt-Hoidahl, corr 2000; (b) 9 'iWd2 is tame: 9 . . . c4 10 i.. e 2 'iWa4 (10 . . . 0-0 11 0-0 f6 12 exf6 .l:.xf6 is safe and solid) 1 1 h4 (an instructive manoeuvre occurred in Posohov-Ku rochkin, Kharkov 2002: 1 1 0-0 iLd7 12 lLlel 0-0-0 13 f4 f6 14 i.. g4!? f5 15 iLdl l:!.dg8 16 lithl h6 1 7 lLlf3 lLld8! 18 l:I.b l lLlf7 19 lLlgl g6 20 .i.e2 .J:!.g7 and Black followed with . . . l:thg8 and . . . g5 with strong pressure) 1 1 . . .i.. d 7 12 h5 h6 13 lLlh4 lLlf5!? 14 lLlxf5 exf5 1 5 l:!.gl?! lLle7 16 g4 fxg4 1 7 ..ixg4 i.xg4 1 8 11xg4 lLlf5 19 \\!Ve2 \\!Vd7 20 l:!.f4 'iVe6 21 'iWf3 g6+ as in Smyslov-Botvinnik, USSR Ch 1 940. 9 c4 10 i.. f1 The slow 10 i.. e 2 hasn't achieved much: 1O . . . .td7 (10 . . . 0-0!? with the idea . . .f6 is also played) and: (a) 1 1 0-0 f6 (1 1 . . .0-0-0 12 lLlg5 .l:i.df8 is complex but also satisfactory, e.g., 13 f4 f5 14 exf6 gxf6 15 lLlh3 lLlf5 16 i.. g4 lLlce7= Ljubisavljevic-Ugrino vic, Italy 1970; but Black also had 16 . . . lLld6 1 7 lLlf2 f5 ! 18 .th3 lLle4=i=) 12 ...
1 95
Play the French ':el (12 exf6 gxf6 13 liJh4 0-0-0=) 12 . . . fxe5 13 dxe5 (13 liJxe5 liJxe5 14 dxe5 0-0=) 13 . . . 0-0-0 (13 . . . 0-0 14 �f1 :f5 15 g3 llafB 16 .l:te3 ':5f7 17 �g2 "iic 7 De Firmian-Gulko, Malmo 2001) 1 4 liJd4!? liJxe5 1 5 f4 liJ5c6 1 6 i.. g4 liJxd4 17 cxd4 Fillion-Rozentalis, Quebec 200 1 ; and best was 17 . . . c3 intending . . . liJf5; (b) 11 liJg5 f5 !? (l 1 . .h6 12 liJh3 0-0 1 3 liJf4 f6=) 12 exf6 gxf6 1 3 i..h 5+ liJg6 with the idea 14 liJxh7?! �f7 1 5 "iig4 1Iag8 !:j:.
exf6 gxf6 1 5 'ife2 �b8 16 l:[fb l �de8 17 i.. g2 "iia 4 (It was time for 17 . . . e5:j:) 18 l:ta2 l:te7 1 9 :ab2 i.. e 8= . 1 1 . 1 2 8 i.. e 2
This modest move, less dynamic and threatening than 8 h4 or systems with i.. d 3, is still an important one that reflects the desire to prop up the centre quickly via 0-0 and moves like lIe l . 8 Jia5 Black tries to tie White to defence of c3, and can also prepare the block ading move . . . 'i!i'a4 in some situations. This move order has always been considered a little odd because White doesn't have to defend the pawn on c3 immediately and apparently gets more leeway to develop. One can also argue that Black's deployment is less effective when White hasn't yet com mitted to a4. But Black gains new possibilities of blockade and White often plays a4 at some point anyway. It's a little surprising that so few players have made the simple devel oping move 8 . . . 0-0!? when White has played the relatively passive i.. e 2 and not the attacking i.. d 3. This directly supports Black's standard idea of playing . . .f6. It's true that this forfeits the option of . . . 0-0-0 and can expose . .
1 0 ... i.. d 7 1 0 ... 0-0 11 g3 f6 12 i.. h 3 (12 exf6 .l:.xf6 1 3 �g2 i.. d 7 1 4 0-0 :afB) 12 . . . fxe5 13 liJxe5 liJxe5 14 dxe5 Ri cardi-J.Watson, Ubeda 1 999, and now 14 . . . liJf5!? 15 0-0 'iKc5 was best, e.g., 16 .txf5 :xf5 1 7 i.. e 3 'iVa5 1 8 i.. d 4 �d7:j: . 1 1 g3 f6 Or 1 1 . . .f5= or 1 1 . . .0-0-0, e.g., 12 i.. h 3 (12 liJg5 .l:.df8 1 3 i..h 3 h6 1 4 liJf3 f6) 12 ... f5 1 3 exf6 gxf6 1 4 0-0 e5 15 i.. x d7+ l:txd7 16 liJh4!? liJg6! 1 7 liJf5 h5 and at any rate Black chances are not inferior. Mter 1 1 . . .f6, G.Garcia-J.Watson, Philadelphia 1998 continued 12 ..th3!? (simpler was 12 exf6 gxf6 1 3 �g2 0-0-0 14 0 - 0 liJf5=) 12 . . . 0-0-0 (or 12 . . . fxe5 1 3 dxe5 0-0=) 1 3 0-0 liJg6 14
1 96
Winawer Variation: Positional Lines Black to an early kingside attack, but perhaps that attack is no worse than after 7 �g4 O-O? Just for fun, and to get used to the themes, here are a couple of ideas: (a) 9 .td3 threatening iLxh7+ is the move given by what little theory there is (e.g., Minev), although it does waste time, e.g., 9 . . .f6 10 exf6 (10 iLf4? fxe5 11 i.. x e5 ttJxe5 12 ttJxe5? 'i'a5+ with the idea 13 'i'd2 cxd4 14 cxd4 'i'xd2+ 1 5 c;t>xd2 .l:t.xf2+) 10 ... .l::!. xf6 1 1 i.g5 l:1.f7 12 0-0 (12 dxc5 e5) 12 . . . c4 13 i.e2 h6! (13 .. :i'd6!?) 14 .th4 (14 .te3 'i'c7 15 'ifd2 .td7) 1 4 ... 'i'a5 15 'ifd2 .td7 16 .l::!. fe l ttJf5; (b) 9 O-O! f6!? (probably better is 9 . . . 'i'a5 with a normal position, but also possible is 9 . . . h6!? 10 .l::!. e l f6 1 1 exf6 .l::!.xf6) 1 0 exf6 llxf6 1 1 .tg5 .l::!.f7 12 .l::!. e l 'iWd6 1 3 'iVd2 .td7, probably ;1; . Presumably White has a move order that leads to advantage in these lines, but his edge would be a small one and can well be compared to the play that follows.
9 O-O! The reason for the decline in the . . . ttJbc6/ . . . 'i'a5 move order. White can also defend the pawn by: (a) 9 �d2. This is usually played to prepare a4 and .ta3, but here Black hasn't committed to . . . .td 7 (see 1 1 . 3),
so a good answer is 9 . . . b6! with the idea . . . .ta6, e.g., 10 dxc5 (10 a4 i.a6 1 1 .txa6 'i'xa6 12 dxc5 ttJg6!? 13 cxb6 axb6 14 'i'e3 0-0 with great pressure) 1O . . . bxc5 1 1 c4 (1 1 0-0 .ta6 12 c4 0-0) 1 1 . . :�Wxd2+ (1 1 . . .'i'a4) 12 .txd2 ttJg6 13 cxd5 exd5 14 c3! (14 .tb5 .td7 1 5 .txc6 .txc6�) 14 . . . c4!? ( 1 4 . . . ttJgxe5 1 5 ttJxe5 ttJxe5 1 6 .te3) 15 i.. e 3 0-0 1 6 .l::!. d l i.. e 6� ; (b) 9 .td2 0-0 (two other moves of interest are 9 . . . h6!? and 9 . . :�'a4, the latter of which answers both 10 .l::!.b l and 10 'i'b l by 1O . . . c4, and if 10 0-0, Black contests the light squares with 10 . . . b6; lastly, the older 9 . . . .td7 10 0-0 c4 with ideas of ... 0-0-0 and . . .f6 is still respectable, but requires loads of study)
This position arises often and bears some analysis: (b l) 10 c4 'i'c7 weakens White's centre, e.g., 1 1 cxd5 exd5 12 c3?! .tg4 13 .te3 'i'a5 14 'iWd2 ttJf5; (b2) One of the problems with an early . . . 0-0 tends to be 10 i.d3 (threatening i.xh7+) , but here Black can defend by 10 . . . ttJg6 (or 10 . . . h6 intending 1 1 0-0 c4 12 .te2 f6) 1 1 ttJg5 (1 1 c4? 'i'c7; 1 1 0-0 c4 1 2 iLe2 f6) 1 1 . . .h6 12 ttJh3 c4 13 i.. e 2 f6 14 iLh5 ttJge7 15 exf6 llxf6 with active pieces; . . . .td7-e8 is a good follow-up; 197
Play the French (b3) 10 O-O! c4!? (10 . . . h6 1 1 c4 "ilc7 12 cxd5 exd5 13 dxc5 is better for White than in the last two examples, but Black should be able to defend after 1 3 . . .'!t:Jxe5 14 liJxe5 "ili'xe5 15 .l:tel �c7 1 6 .td3 .te6;!;) and the play can be very instructive: 1 1 liJg5 (1 1 a4 f6 12 exf6 l:txf6 1 3 i.g5 l:tf8 14 'iWd2 liJf5= intending 15 l:tfel?! liJd6! and . . . liJe4) 1 1 . . . h6 12 liJh3 f6 13 exf6 (13 f4 i.. d 7 planning ... .te8-g6 at some point, or just 13 . . . fxe5 14 fxe5 lixfl+ 1 5 i..xfl .td7= and . . .lU8) 1 3 . . .lilxf6 1 4 f4 (14 "ili'e l e5 1 5 dxe5 liJxe5 1 6 liJf4 ..lif5;; 14 liJf4 can be answered by 14 . . . ..lid7= , but more interesting is 14 . . . e5!? 15 dxe5 liJxe5 16 i.e3 ..lid7 17 'i'd2 l:taf8 when neither side has a clear plan) 14 . . . .td7 1 5 i.g4 �af8 1 6 'iVe2 liJf5 (equal, but not necessarily best; 16 . . . liJc8!? 1 7 liJf2 liJd6 is one idea, but most attractive is 16 . . . liJd8!?, a multipurpose move that defends b7 and e6 and frees the bishop to go to a4 or to the kingside, e.g., 17 a4 .te8 18 l:tfb l .tg6 19 11b5 'iVa6) 1 7 liJf2 (17 i.xf5? exf5 and Black takes the e-file) 1 7 . . . 'iWa4 (or 1 7 . . . .te8 1 8 i..f3 liJh4) 1 8 l:Ifc l liJd6 1 9 ..th5?! h7 and White i s out o f effec tive ideas, since 20 liJg4 is answered by . . ..:.f5 and 20 g4?? by 20 . . . g6. 9 ... b6
1 98
A move that has barely been tried or investigated, despite this position having arisen time and again. I think that it is logical to contest the light squares. Normally 9 . . . .td7 or 9 . . . c4 are played. Of course, Black cannot grab the pawn in view of 9 . . . "ilxc3?? 10 iLd2 Vib2 1 1 l:tb 1 'i'xa3 12 l:tb3 'ilVa2 13 'i'c l and lIa3. 10 a4 White tries to give c4 added effect. With the queen out on a5, Black's game hangs on one tempo, but I don't think there's any real danger: (a) 10 iLd2 .ta6 1 1 c4 (1 1 a4 i.xe2 12 'i'xe2 0-0) 1 1 . . .'iWa4 12 cxd5 exd5! (12 ... .txe2 1 3 'i'xe2 exd5 14 dxc5 bxc5 1 5 c4 "ili'xc4 16 "ili'xc4 dxc4 1 7 :fc l l:td8 is close to equality) 13 dxc5 bxc5 14 c4 (14 .txa6 "ili'xa6 15 l:tel 0-0=) 1 4 . . . "iIi'xdl 15 ftfxdl 0-0 16 .td3 �ab8=; (b) 10 c4 cxd4 (1O ... dxc4!? 11 dxc5 .ta6! 12 cxb6 axb6 also looks good, but the d-pawn capture is more forc ing) 1 1 cxd5 (1 1 liJxd4 liJxd4 12 'i'xd4 ..ta6 13 "ili'g4 'i'c3! 14 ..te3 0-0+) 1 1 . . .'i'xd5 12 ..lib2 ..tb7 13 ..txd4 l:td8 (13 . . . liJxd4 14 "ili'xd4 'iVxd4 15 liJxd4 0-0;) 14 .tb2 'i'c5 15 .td3 0-0; . 10 .....ta6 1 1 i.. xa6 1 1 .td2 i.xe2 12 'i'xe2 0-0 13 c4?! "ili'a6; . 1 1 .. JWxa6 12 dxc5 bxc5 1 3 i.. a 3 This is Babula-Trichkov, Czecho slovakia 1992. Black should retain a small advantage after 1 3 . . . 'i'a5, e.g., 1 4 c4 l:td8 1 5 cxd5 liJxd5 (16 'i'el 'iWxel 1 7 I1fxe l liJdb4!). More direct if riskier would be 1 3 . . . 'i'xa4!? 14 i.. xc5 'iWc4 15 i.d4?! (better 15 .txe7! when 15 .. .'�xe7!? 16 "ili'd3 ! 'i'xd3 17 cxd3 lihb8 gives only a tiny advantage at best, while 15 . . . liJxe7 16 'iWd3 'i'c5!? 1 7 c4! equalises) 1 5 ... 0-0 16 liJd2!? (16 "ili'd3? liJxd4 1 7 cxd4 :tfc8+) 16 ... "ilb5
Winawer Variation: Positional Lines 17 l:.bl (17 �g4 �b2 ! hits c2 and pre pares .. .'�Jf5) 17 . . . �a6 18 l'Llb3 (18 �g4 �a4! 1 9 l'Llb3 l'Llxe5 20 �g3 l'Ll5g6) 18 . . .l:.fc8; with the idea . . . l'Llf5 and meeting 1 9 l'Llc5? by 19 . . . 'iYc4! (20 �d3 l'Llxe5) . Play in the manner o f 9 . . . b6 i s ex perimental and will produce original situations with creative possibilities. 1 1.2 7 a4 This used to be the main line of the Positional Winawer, both preparing i.. a 3 and preventing a blockade by either . . . i.. d 7-a4 or . . . �a5-a4. It also provides a potential response to . . . b6 systems, namely a5. 7 . �a5 ..
Black plays the most forcing move in order to make White commit either the queen or bishop to d2. Thus:
Another solid solution is 8 . . .b6!? preparing ... i.. a 6, and: (a) 9 l:.b l i.. a6 10 i..b 5+ i.. xb5 1 1 axb5 a6 1 2 bxa6 l'Llbc6 13 l'Lle2 �xa6 14 0-0 0-0; Maciej a-Gdanski, Poland 1992; (b) 9 i.. a 3 �xa4!? (playable, but risky; Black can play 9 . . . i.. a 6! 10 i..b 4 cxb4 11 cxb4, when the queen is trapped but 1 1 . . .�xfl 12 bxa5 i.. xg2 1 3 f3 i.. xhl 14 axb6 axb6 1 5 'it>f2 l'Llbc6 16 l'Lle2 i..xf3 17 'it>xf3 0-0; fol lows; material is even but Black is better coordinated and . . .f6 is a po tential move) 10 l'Llf3 "iVc6 1 1 �g5 l'Llg6 12 dxc5 bxc5 13 h4 h6 14 'i'g3 0-0 15 �d3 i.. a 6!+ (still with danger), I.Almasi-Blauert, Budapest 1993; (c) 9 i..b 5+ i.. d 7 10 i.. d 3 l'Llbc6 (here 10 . . . �xa4!? is a suggestion of Peder sen) 1 1 l'Llf3 f6 (1 1 . . . 0-0 12 0-0 f6 1 3 exf6 l:i.xf6=) 12 O-O? ( 1 2 exf6 gxf6 1 3 0 - 0 c 4 14 �e2 0-0-0=) 12 . . .fxe5 1 3 dxe5 0-0; Borkowski-Drasko, Polani ca Zdroj 1 988; (d) 9 l'Llf3 i.. a 6 10 i.. d 3 (10 i.. xa6 'i'xa6 1 1 �e2 �xe2+ 12 'iitxe2 l'Lld7 1 3 a5 b 5 ! ? 14 i.. a 3 a6; Pelitov-Uhlmann, Szombathely 1966) 1O . . . i.xd3 1 1 cxd3 l'Llbc6 (1 1 . . . l'Lld7 12 0-0 .uc8) 12 dxc5 bxc5 13 0-0 0-0 14 i.. a 3 .ufb8= Wies mann-Holzke, Bundesliga 1990. 9 l'Llf3 f6!?
11.21 8 'iVd2 11.22 8 i.. d 2 11.21 8 �d2 White plans to put his bishop on a3. This allows an endgame after . . .cxd4/cxd4, when White hopes that possession of the two bishops will se cure him an advantage. Practice has not confirmed this. 8... l'Llbc6
1 99
Play the French Korchnoi's move, given for some variety. I have covered the main move 9 . . . Jl.d7 in some detail in previ ous editions and can only assure the reader that nothing has happened to change its assessment of full equality. 10 �d3? This looks aggressive but doesn't work out. Black holds his own versus others: (a) 10 .ta3 fxe5 and: (a l) 1 1 dxe5? 'iYxa4! 12 i.e2 b6 1 3 c 4 tbd8!? ( 1 3 . . .tb a 5 also opens a re treat for the queen while controlling c4, so 14 cxd5 exd5 15 0-0 tbc4! be comes possible) 14 O-O?! (14 cxd5 exd5 15 c4;) 14 . . . 'iVd7 15 tbg5 h6 16 tbh3 0-0 17 .l:.fdl tbdc6 18 .tb2 .i.b7+ Kui jpers-Korchnoi, Wijk aan Zee 1971; (a2) 11 .i.xc5!? e4 12 tbgl (useless is 12 .i.b4 Wic7 13 tbgl O-O! 14 f3 e5 1 5 fxe4 tbxb4 16 cxb4 dxe4 1 7 tbe2 tbd5!+)
White has actually played this re peatedly, which says something about the favourable status of 9 . . .f6! The first player exerts some dark square pressure but he is behind in devel opment and outmanned in the centre: 12 . . . 'iYc7 (or 12 . . . 0-0 13 i.b5 l:1f7 14 tbe2 a6 15 i.xc6 tbxc6 16 tbg3 i.d7 1 7 'ilfe3 'iVc7+ Zedtler-Schwertel, email 1998) 13 �e2 b6 14 i.a3 0-0 15 c4 :f6 200
16 .l:!.b l tbf5 1 7 c3 tba5! 18 c5 Ji.d7+ 19 tbh3 e3! 20 fxe3 tbg3 ! 2 1 hxg3 "i¥xg3+ 22 cl Brunner-Arnason, Gausdal 1 988; here Black missed 24 . . . YWe4! when White can resign; (b) 10 exf6 gxf6 and: (h I) 1 1 .tb5 .td7 12 .ta3 (12 0-0 cxd4 1 3 tbxd4!? e5 14 tbb3 'iVc7=F Zude Mueller, Germany 1 989) 12 . . . cxd4 13 cxd4 l:i.g8 14 .l:tgl !? "i¥xd2+ 1 5 e3 i.c6; with control of the light squares, Houtsonen-Es kola, Finland 1992; (b2) 11 i.a3 cxd4!? ( 1 1 . . .c4) 12 cxd4 "i¥xd2+ 13 'iio>xd2
13 . . .tbf5 ('=' - Uhlmann; another idea is 13 . . . .i.d7 14 .i.d3 c,t>f7 1 5 .i.d6 tba5 16 l:the l l:tac8=) 14 c3 .l:iog8 15 g3 liJa5 16 l:tb l Pezdirc Praznik-Zelic, corr 1978; play on both sides has been logical, but now instead of 16 . . . c2 h6= ; (b3) 1 1 dxc5 e5 is a standard struc ture in these positions. The chances are equal, e.g., 12 c4! (12 .te2!? .te6 13 0-0 0-0-0 14 .l:i.e l l:thg8 15
Winawer Variation: Positional Lines J:!.b5 a6 1 7 .u.xa5 Jtxc2 is a remarkable position which seems in some sort of balance) 16 ,U,b6 0-0 1 7 'it'dl ii.g4 1 8 h 3 ii.h5 19 g 4 Jtg6= Kuijpers-Pytel, Wijk aan Zee 1974. 10 ...fxe5 11 dxe5 0-0 1 1 . . .c4 12 Jte2 'tJg6 13 0-0 0-0, and disaster followed 14 i.. a 3? .l:!.f5 1 5 g4 .l::[f4 16 h3 h6 17 'iit>h l .l:!.f7 18 Jtdl 'tJcxe5 1 9 'tJxe5 'tJxe5 20 f4 11Hxa4 0 - 1 C.Carlson-Fawbush, corr 1 9 7 1 . 12 0-0 'u'xf3! These days this is almost a routine exchange sacrifice, although with the bishop and rook out of play and the queen over on the queenside, it still takes some nerve to play it! 13 gxf3 c4 14 �e2 'tJxe5
Black has a pawn for the exchange and White has six isolated pawns! 15 Jta3 15 f4 'tJ5g6 16 Jth5 i.. d 7!?, and a possible line is 1 7 i.. a 3 11Hc7 1 8 �xe7 'tJxe7 19 Jtg4 ,U,f8 20 'u'fe l ':f6+, when White is at loss for a move. After 1 5 �a3, Felgaer-Korchnoi, Bled 2002 continued 1 5 . . . 'tJf5 16 f4 'tJg6 1 7 Jtg4 'tJgh4 18 Jtb4 11Hc7 1 9 a5 11Hf7 20 f3 Jtd7 21 'u'ael 11Hg6 22 'it'hl Jtc6 23 11Hf2 d4 24 �xf5 'tJxf5 25 cxd4 Jtd5 26 c3 �5 27 .l:!.e5 �6 28 .l:!.e4 �3 29 .l:!.gl b6 30 11Hg2 �5 31 a6 ':c8 32 Jtd6 g6 33 �e5 'tJh4 0 - 1 .
1 1.22 8 Jtd2 White abandons the idea of Jta3 (although �c1-a3 can occur later), but keeps the kingside dark squares covered and gives solid support to the weak c3 square. 8 ...'tJbc6 9 'tJf3 (a) 9 11Hg4 compares poorly with the main 7 11Hg4 line because the move a4 is practically wasted; play might go 9 . . . 0-0 (a cute game went 9 . . . Jtd7 10 11Hxg7 0-0-0 1 1 11Hxf7? �6! 12 .td3 ,U,df8 1 3 �5 c4 14 ii.f1 'tJxd4! 1 5 0-0-0 'tJxc2! Rebej-Spitzenberger, corr 1965, due to 16 'iit> xc2 �3+ 1 7 'iit> c 1 Jtxa4) 10 'tJ f3 £6 (or 10 . . . f5 1 1 �5 �6!? 1 2 dxc5 11Hcn) 11 exf6 ,U,xf6 12 �5 Jtd7 1 3 i.. d 3 h6 14 0-0 c4 1 5 Jte2 Jte8 1 6 �4 ii.g6:'1= ; (b) 9 'tJe2 11Hc7. Now that the bishop can't come to a3, this is an ideal spot for the queen. It also ties White to defence of d4. Myreng-Glueck, St Paul 1982 went 10 �f4 'tJg6 1 1 .tg3 11Ha5 12 11Hd2 f6= (or 12 . . . 0-0! and . . . £6) . 9 ...Jt d7
Here White normally plays: 1 1 .221 10 i..b 5 1 1.222 10 �e2
None of the many other moves is as critical:
10th
201
Play the French (a) 10 ttJg5?! intends 'i!fh5, but abandons the centre: 10 . . . h6 1 1 'i'h5 (1 1 ttJh3? 'iVc7! hits d4: 12 i.c 1 ! ? cxd4 13 cxd4 ttJb4! 14 c4 dxc4 15 i.e2 ttJd3+! 16 .i.xd3 cxd3 17 'iVxd3 .i.c6+ Milton-J.Watson, San Diego 1986; 12 'ikg4 ttJf5 1 3 .i.d3 cxd4 1 4 cxd4 ttJxe5!+ - Suetin) 1 1 . . .g6 12 'i!fh3 cxd4 1 3 cxd4 'ikb6! 14 c3 'ikb2 15 .l:[d1 ttJa5+ ; (b) 10 'ikb 1 'i'c7 1 1 �d3 c4 12 �e2 f6 13 �f4?! ttJg6 14 JLg3 fxe5 15 dxe5 'ika5 16 O-O!? Schoneberg-Tischbierek, Leipzig 1 986, and easiest was 1 6 . . . b6!+; (c) 10 c4 'fIc7 11 cxd5 exd5 (1 1 . . .ttJxd5=) 12 dxc5?! (12 c3 0-0 in tending . . . cxd4 and . . . i.g4) 12 . . . 0-0 1 3 i.e2 ttJxe5 14 0 - 0 ttJxf3+ 1 5 i.xf3 'ikxc5:j: Dubinin-Boleslavsky, USSR 1 947; (d) 1 0 h4 doesn't mix well with a4: 10 . . .f6 (or 10 . . . 0-0-0! 11 h5 h6) 1 1 h5 fxe5 12 ttJxe5 (12 h6 gxh6 13 ttJxe5 ttJxe5 14 dxe5 0-0-0 15 l1xh6 'iVc7 and . . .ttJf5) 12 . . . ttJxe5 13 dxe5 'iVc7 14 f4 0-0-0 1 5 'iVg4 h6!= Pirttimaki-Farago, Helsinki 1983; (e) 10 l1b 1
best, but 1 5 . . . ttJg6 gives Black no problems, e.g., 16 c4!? ttJc6 17 'i'a1 d4!:j:) 1 1 i.e2 (1 1 i.b5 is well met by 1 1 . . .a6 intending 12 �e2 f6 or 12 Axc6 ttJxc6; compare 10 i.b5) 1 1 . . .f6 12 �f4?! ttJg6 13 .i.g3 fxe5 14 i.b5? cxd4 15 cxd4 'i'a5+ 16 ttJd2 exd4 1 7 i.. d 6 'i'd8 18 h 4 'iVf6+ Hirsig-Jurek, Bad Ragaz 1993; (f) 1 0 g3 0-0-0 (safest looks 10 ... 0-0 11 �h3?! [1 1 i.g2 f6=] 1 1 . . .f6 12 exf6 .l:.xf6:j:) 1 1 i.h3 (1 1 .tg2 f6 12 c4 'fIa6!? 1 3 cxd5 ttJxd5= with many options for both sides) 1 1 . . .f5 ! 12 .i.g2 (12 exf6 gxf6 1 3 0-0 e5 14 JLxd7+ .l::i.x d7 1 5 c4 'fIc7+ Doda-Raicevic, Banovici 1979) 12 . . . h6 13 h4 "ilc7 14 .i.c1 (14 0-0 ttJa5 15 dxc5 ttJc4 16 l:leU) 14 . . . cxd4 1 5 cxd4 ttJb4 16 �d2 �e8!? 1 7 � a 3 ttJbc6 18 Ad6 l:lxd6! 19 exd6 'fIxd6 20 �c1 (20 c3 f4 21 g4 h5 22 g5 �g6) 20 .. .f4! with a strong attack, Kavalek-Hort, Montreal 1979; (g) 10 i.. d 3 is a tempo down on some 10 .i.e2 lines after 1O ... c4 1 1 i.e2 f6 1 2 0-0 fxe5 1 3 ttJxe5 (13 dxe5 O-O! 14 l:te1 ttJg6 1 5 �f1 l:1.f5) 1 3 . . . ttJxe5 14 dxe5 0-0 1 5 f4 .i.e8 16 Ag4 'iVb6+ 1 7 �h1 i.. g6 18 'ikb 1 'fIc6! 1 9 'iVb5 'i'c8! 20 l:ta2 �e4+ S.Nikolic Ivkov, Sarajevo 1967. 1 1.221 10 i.b5
10 . . . 'iVc7 (10 . . . 0-0-0!? 1 1 l:1.b5 'i'xa4 with the idea 12 :xc5? b6 13 l:1.b5 ttJxe5!; better is 12 'ii'b 1 ttJa5 1 3 .l:.b2 ! 'iVc6 14 l:1.a2 'ilc7 Enders-Piersig, Germany 1985, and now 15 dxc5 is 202
Winawer Variation: Positional Lines White tries to provoke . . . a6, hoping that it will be a weakness. 10 ... a6 Anyway! I'm not really sure that any logical move isn't good here, for example, 10 . . . c4 1 1 0-0 f6. But a par ticularly sound alternative is 10 . . . 'ilic7 1 1 0-0 0-0
with these continuations: (a) 12 i.. d 3 f6 13 exf6 lIxf6 14 i.. g5 (14 dxc5!? lIaf8 1 5 c4 lIxf3! 16 gxf3 tDe5 with plenty of compensation) 14 . . . .l:!.ff8 (14 . . J::tfl! 15 lIe1 c4 16 i.. fl lIaf8) 1 5 .l:.e1 ! c4 1 6 i.. f1 tDg6 (16 . . . tDf5!? would anticipate 17 tDe5 tDxe5 18 .l:!.xe5 tDd6! intending . . . tDfl or . . . tDe4) 17 g3 Roiz Baztan-Carrasco Martinez, Spain 1994; now 17 . . . h6! 18 .tc1 .l:tf6= ; (b) 12 ..tc1 b6 1 3 �a3 tDa5! 14 dxc5 (14 i.. x d7 'iVxd7 1 5 dxc5 l:tfc8! 16 cxb6 axb6 - Knaak) 14 . . . .txb 5 1 5 cxb6 axb6 16 axb5 llfc8 1 7 .tb4 tDc4= Nunn-Yusupov, Belgrade 1 99 1 ; (c) 12 dxc5 tDxe5 1 3 .tf4 tDxf3+ 1 4 'i'xf3 'i'c8 1 5 i.. d 6 'ue8 1 6 .td3 tDf5 17 ..txf5 Y2-Y2 Zelcic-Psakhis, Bled 2002; (d) 12 lIe 1 , the most flexible move. Black has two good responses: (d1) 12 . . . b6 13 i.. d 3 h6 14 'iVc 1 c4 15 .tfl f6 16 g3 fxe5 17 tDxe5 tDxe5 18 �xe5 tDc6 (or 18 . . .tDg6 1 9 lIe3 !? e5:j: as in the game Hellsten-Wiedenkeller,
Borlange 1992) 19 :e1 e5 20 .tg2 e4 2 1 'iVd1?! (2 1 i.. xh6! gxh6 22 'iixh6 i.f5 23 g4 ..th7!? 24 'i'g5+ �h8 25 "tWxd5 lIf4!? unclear) 2 1 . . . .l::[ a e8 22 f3 exf3 23 lIxe8 i.. x e8 24 ..txf3 'ilifl 25 i.g2 1ii'f2 + 26 �h1 ..tfl+ and White's weak pawns hurt him in Kinder mann-Zugzwang, Lippstadt 1 998; (d2) 12 . . . h6 13 ..tf4 (13 ..tc1 a6 14 i.. xc6 i.. xc6 intending 15 ..ta3 ..txa4 16 ..txc5 ..tb5=) 13 . . . tDg6 14 .tg3 tDce7 15 .td3 c4 16 i..xg6 (16 .te2 tDf5 1 7 'i'd2 tDxg3 1 8 hxg3 f6 19 exf6 .l:!.xf6 20 tDh2! lIafS 21 f4 tDe7 22 tDg4 :f5= Nunn-Yusupov, Belgrade 1991) 16 . . .fxg6! :j: 1 7 h4 lIfl 18 1\i'b 1 �h7 19 "tWb4 tDf5 20 i.h2 .l::[ af8+ Byrne-Vag anian, Moscow 1975. 1 1 .txc6 Probably best, in view of: (a) 1 1 i.. e 2 (extremely popular, on the theory that . . . a6 would be a weakness by comparison with 10 .te2, but it hasn't proven so) 1 1 . . .f6 12 c4 'ilic7 1 3 cxd5 (13 exf6 gxf6 1 4 c3 Morris-Klein, New York 1991; 14 ... dxc4 1 5 i.. xc4 cxd4 16 cxd4 tDxd4!?) 13 . . . tDxd5 14 c4 tDde7 (14 . . . tDdb4!?) 15 exf6 gxf6
1 6 dxc5 (16 d5!? exd5 1 7 cxd5 tDxd5 18 'ii'b 3 ..te6:j: 19 0-0 1ii'g 7?! 20 .tc4 0-0-0 21 l:i.fe 1 lIhg8 22 g3 Wlfl un clear, Chandler-Kummerow, Solingen 203
Play the French 2000; Black can probably improve) 16 . . . 0-0-0 17 �c3 e5 18 'iWd6 �f5!? (18 ... liJf5 looks more accurate: 1 9 'iYxc7+ rJilxc7 with compensation, Tis chbierek-Poldauf, Germany 1 9 9 1 ; compare 10 ite2) 19 'iWxc7+ ( 1 9 'i!Vxf6!? l:thf8 20 'iVh6 liJg6 - Pelletier) 19 . . .'it>xc7 20 ':'cl (20 0-0 �d3 Pelletier) 20 . . . liJg6!? 2 1 g3 liJf8;!; in tending . . . liJe6xc5, Short-Pelletier, Leon 200 1 ; (b) 1 1 i.d3 c 4 12 ite2 f6 1 3 O-O?! (13 exf6 gxf6 14 0-0 e5=) Dineley Levitt, Swansea 1995, and instead of 13 . . . 'i!Vc7!?, the standard 13 . . . fxe5 ! was good: 14 liJxe5 liJxe5 15 dxe5 0-0 1 6 'i!Vb l ! 'i!Vc7 1 7 f4 liJf5�; (c) 11 c4!? 'iWc7 12 cxd5 liJxd5 (12 . . . axb5? 13 d6!) 13 itxc6 'ii'xc6 (13 . . . itxc6! with a healthy advantage, e.g., 14 0-0 lId8!? 15 c3 liJe7 16 ite3 O-O�) 14 0-0 cxd4 (14 . . . 0-0 looks a lit tle better) 15 liJxd4 'i!Vc7= Casella Shulman, Connecticut 2002.
1 l ... liJxc6 Here both recaptures look fine as long as Black is careful; thus 1 1 . . .i.xc6 and: (a) 12 liJg5 h6 (12 . . . 0-0 13 'iYh5 h6) 13 'i!Vh5 g6 14 'iWh3 'iWc7 1 5 O-O? cxd4 16 cxd4 �xa4+ Berzinsh-Kahn, Ber lin 1994; (b) 12 c4?! 'iVc7 1 3 cxd5 itxd5 204
(13 . . . liJxd5 is also good) 14 0-0 l:td8� C.Lopez-Celis, Buenos Aires 1996 (or 14 . . . cxd4 15 itb4 liJf5;); (c) 1 2 0-0 h6 (12 ... 'iVc7!? 1 3 dxc5 'iVa5 ! is a cute idea that seems to equalise immediately) 1 3 dxc5 (13 'iWcl c4 14 'iVa3 'iVc7 1 5 J:[fel O-O-O!? 16 h4 ':'dg8 1 7 a5 'iWd8 1 8 'iVc l !? liJf5= Suetin-Misiano, Biel 1997; 13 'i!Vel 'iVc7 14 dxc5 'i'd7! 1 5 liJd4 �xa4 1 6 f4 Ji.c6!? 1 7 'iWg3 liJf5! 1 8 'i!Vh3 g6 1 9 g4 Y2 - Y2 Gashimov-Guseinov, Baku 2000) 13 . . .'iWxc5 14 liJd4 0-0 15 'iVg4 'it>h7 16 lIfe 1 i.d7 1 7 J:[ab l 'iWc7 18 lIe3 Kovalov-Tischbierek, Bundesliga 1991, and apart from 18 . . . liJf5 1 9 liJxf5 exf5, which was drawish, Black might try 18 . . . liJg6 19 liJf3 (19 f4 �xa4) 19 . . .f6 20 exf6 l:.xf6; . 1 2 0-0 12 liJg5?! h6 13 'iWh5 O-O!? (13 . . . .:.f8 1 4 liJf3 0-0-0 1 5 0-0 c4 16 lIa2 Sher zer-Filatov, Philadelphia 1993; and now 16 . . . liJe7! =, tying White to the a pawn) 14 liJf3! (14 liJh3 �e8 15 'iVg4 f5 !) 14 . . . i.e8 15 'iWg4 f5 16 exf6 lIxf6 1 7 c4 'i!Vb6 1 8 cxd5 exd5 19 dxc5 'iWxc5�. 12 ... 0-0 12 . . . 'iWc7 13 lIel liJa5 14 dxc5 liJc4 equalised in EI Taher-Shulman, Par dubice 1999. 1 3 J:[el 'i!Vc7 14 dxc5 Black answers just about any move with . . . liJa5. 14 ... liJa5 1 5 ite3 15 liJd4 liJc4!? (this is natural, but 15 . . . 'iWxc5 might cover the kingside better, e.g., 16 'iVg4?! f5) 16 itg5! l1ae8? (16 . . . 'iWxc5 17 'iVg4 'it>h8 with an unclear White attack; Black threat ens . . . liJxe5!) 17 liJb3! liJxe5 18 Ji.f4 f6 1 9 'iVd4± , winning on the dark squares, Wedberg- Spielmann, Stock holm 2002. Mter 1 5 i.e3, Cabrilo-Shulman,
Winawer Variation: Positional Lines Belgrade 1998 went 1 5 . . . lDc4 1 6 'iVb l ! ? 1:.ab8?! (16 . . . lDxe3! 1 7 .l:i.xe3 �ab8 would have quashed worries on the kingside) 17 �d4 b6 18 cxb6 lDxb6 19 'iVdl lDxa4 20 lIe3 'iVb7 2 1 lDd2 h 6 � - � . Again White has some real attacking chances to compensate for his weaknesses, and 22 c4 is also an option. 1 1.222 10 Ji.e2 This is still the most challenging move. 10 ... f6
A central attack that has discour aged White from 10 i.e2, and yet many things remain unclear. In my opinion, a related move order that has been tried recently is very signifi cant: 10 . . . 0-0-0! ('!' for reasons given below) 1 1 0-0 f6, to reduce White's options in the main lines. For exam ple, 1 1 c4 'fiIc7 12 cxd5 lDxd5 13 c4 lDde7 transposes to a main line but avoids ideas like dxc5 and a5 (14 dxc5? at once allows 1 4 ... lDxe5). An independent test of this move order would be 10 . . . 0-0-0 1 1 lDg5, e.g., 1 1 . . . .l:t.dfS 12 i.h5 (12 f4 f6 1 3 lDf3 fxe5 14 fxe5 h6 15 0-0 g5 with good coun terplay) 12 . . . g6 (12 . . . .te8!? also looks okay) 13 Ji.g4?! h5! 14 i.e2 (14 i.h3 'i'b6! with the idea 15 ii.e3 'iVb2 or 1 5
dxc5 'iYxc5:j:) 14 . . :i!i'c7 15 .tc1 lDf5 16 Ji. a3 cxd4! 1 7 .ixf8 .l:i. xf8 18 cxd4 lDfxd4:;: . 1 1 c4 This is aggressive but loosening. Here are typical examples of other moves: (a) 11 0-0 c4 (or 1 1 . . .0-0-0, e.g. , 12 'iVc l ! ? 'iVc7 13 "ili'a3 c4 14 l:1.fel lDg6 1 5 exf6 gxf6 16 i.h6 lIhg8 1 7 'iWc l e5:j: intending . . . .i.f5-e4 Marfia-J.Watson, Colombus 1 977) 12 l:tel fxe5 13 dxe5 0-0 14 Ji.fl ':f5! 1 5 g3 l:.af8:;: Pein Plaskett, British Ch 1 987; (h) 11 exf6 gxf6 and 12 c4 will transpose. Others: (h I) 12 0-0 c4 (a creative strategic battle followed 12 . . . 0-0-0 13 'iVcl c4 14 .l:t.el l:.de8!? [14 . . . e5] 1 5 .ifl lDd8 1 6 �f4 lDf5 1 7 .l:i.b l lDf7! 1 8 l:tb4 lD7d6 1 9 g3 Ricardi -Dominguez Gonzalez, Malaga 1999; then a good move was 19 . . . h5! = and the knights coordinate to defend and attack) 13 lDh4 0-0-0 (a typical pawn sacrifice is 13 . . . lDg6!? 14 i. h5 0-0-0 1 5 i.xg6 hxg6 16 lDxg6 l:.hg8 1 7 lDf4 e5 18 lDe2 l:tg7 19 lDg3 l:.h8 with attack) 14 g3 (14 Ji.h5 lDg6!) 1 4 ... e5 (14 ... lDg6!:j:) 15 f4 exd4 16 cxd4 'iVc7 17 c3 lDf5 18 lDg2 lDa5 19 i.f3 lDb3 20 .l:i.a2 'iVa5= Spassky-Shaked, Hoogeveen 1 998; (b2) 12 lDh4!? 0-0-0 13 .ih5 'ii'c 7! 14 i.e3?! lDg6! 1 5 .ixg6 hxg6 16 lDxg6 .l:i.xh2 1 7 l:.gl .l:i.h7:;: Gold-J.Watson, New York 1 978; (c) 11 .l:.b l ! ? is Wedberg's active move, important to know but no longer feared: 1 1 . . :iWc7! (safest) 12 .i.f4 (12 exf6 gxf6 1 3 dxc5 e5 with typical central play: 14 c4 .ie6 1 5 cxd5 Ji.xd5 16 0 - 0 0-0-0 1 7 'ii'c l l:!.hg8 18 g3 lDf5 19 c3 Poretti-Weber, email 2000; and by 19 . . . i.e4!? 20 l:tb5 �d3! 2 1 i.xd3 lIxd3 the light squares and White's shattered pawns must give 205
Play the French Black at least a satisfactory game) 12 . . . liJg6 1 3 .i.g3 fxe5
(c 1) 14 i.b5 cxd4 15 cxd4 'ii' a 5+ 1 6 'it>f1 0 - 0 1 7 .txc6 .txc6 1 8 liJg5!? (18 liJxe5 liJxe5 1 9 .txe5 1\Vxa4) 1 8 ... h6 19 liJxe6 lif6 20 liJc5 exd4 2 1 "it'xd4 b6-+ and ... 'i¥xa4. Weinzettl-Dueckstein, Nendeln 1 986; (c2) 14 0-0 cxd4 (14 .. :ii'c8!? seems good enough: 15 dxe5 0-0 and 16 c4 liJce7! 17 'i'd2 i.c6:j: Maki-Ogaard, Oslo 1983 or 16 'iVd2 liJce7:j:) 15 cxd4 0-0 16 i.b5 (16 dxe5 liJf4 17 c4!? d4!:j:) 16 . . . liJf4 (a safe alternative is 16 . . :iWc8 17 dxe5 liJce7) 17 i.xc6 (17 liJxe5 liJxe5 1 8 dxe5 ii.xb5 19 lixb5 Karell D.Myers, corr 1 990; 19 . . . .l:.f7! - Myers) 1 7 . . . .txc6 18 liJxe5 ii.xa4 19 11b4 (a much-tested position: 19 'ii' g4 h5! wins; 19 'i'd2 liJh5 is strong; 19 .l:.b2 .l:f.ac8!+ with the idea 20 'i'd2 'i'c3) 19 .. :Wxc2! (This seems clearer than 19 . . . .tc6 20 liJd3 1\Va5 2 1 .txf4 .l:.xf4 22 liJxf4 'i'xb4 23 liJxe6 l:!.e8:j: Ivanov Hanauer, corr 1 996) 20 .l:i.xa4 (20 'ii' g4 h5) 20 . . . liJe2+ 21 �h1 �xd1 22 .l:.xd1 liJc3+ 23 lida 1 liJxa4 24 lixa4 lifc8 25 h3 a5 26 liJd7 b5 27 .l::i. a 1 lic2 28 liJc5 a4-+ Bickel-Hund, corr 1983. Notice that in all of these lines Black plays . . . 0-0-0, indicating that the order 10 . . . 0-0-0 (instead of 10 . . . f6) probably won't create insoluble prob206
lems in these situations. 1l ...'ii' c 7 12 exf6 12 cxd5 liJxd5 13 c4 is more accu rate, preventing . . . liJf4. 12 ... gxf6 1 3 cxd5 13 .i.c3 O-O-O! 14 dxc5 d4! and 1 5 i.xd4 i. e8 o r 1 5 liJxd4 i.e8+ . 13 ... liJxd5
14 c4 (a) 14 dxc5 is perhaps the best: 14 . . . 0-0-0 (14 . . . liJf4? ! 15 0-0 liJxe2+ 16 'ii'x e2 e5 1 7 a5! i.g4 1 8 i.c3± ; 14 . . . e5 15 a5 a6 16 c4 liJde7 1 7 1lfb 1 ! ? 0-0-0 18 0-0 i.g4 1 9 .l:.d1 i.xf3 20 .i.xf3 liJd4 2 1 .l:la3;!;) 1 5 a5! (15 0-0 e5 1 6 c4 liJdb4! 17 i.xb4 liJxb4 18 'ilfd6 11hf8 19 'ii'x c7+ �xc7 20 lifd1 liJa6+ McCrory J.Watson, Denver 1976) 15 . . . a6! (15 . . . .te8!? 16 a6 liJc3 17 axb7+ �b8 18 i.xc3! :xd1 + 19 .l:.xdl±) 16 c4 (or 16 0-0 e5 1 7 .l:.b 1 ..tg4 18 liJh4) 16 . . . liJde7 1 7 i.c3 e5 18 O-O;!; . Black has in fact done well from this kind of position, and these lines are certainly playable, but I don't trust them to fully equalise. This might be a good reason to play 10 . . . 0-0-0, which avoids such play; (b) 14 c3 0-0-0 15 0-0 ':'hg8 16 .l:.e1 e5!? (or 16 ... liJf4 1 7 i.xf4 'ii'xf4:j: Mann-Flugge, corr 1983) 1 7 c4 (17 dxe5 .th3 unclear - Korchnoi) 17 . . . i.h3! 18 i.. f1 liJb6 (18 . . . liJf4!? 19
Winawer Variation: Positional Lines ..txf4 exf4 20 d5 ..tg4! has occurred in two games and equalises) 1 9 d5 tDxc4! 20 dxc6 'iWxc6 2 1 g3 .txfl 22 l:txfl e4; 23 'iVc2 (23 'iVb3?! 'iWd5 24 .:tac 1 Spassky-Korchnoi, Belgrade 1977; 24 . . . tDe5! 25 lhc5+ 'iWxc5 26 ]:tc 1 'iVxcl + 27 .txc1 tDxf3++) 23 . . . "iYd5 24 .tf4 exf3 25 :fc1? (25 ]:tac 1 tDe5=) 25 . . . tDe5 26 'iVf5+ :d7 27 "iYxf6 tDd3+ Mazi-Farago, Bled 1992. 14 ... tDde7 Black can also play 1 4 . . .tDdb4 and 14 . . . tDf4! (e.g., 15 ..txf4 �xf4 16 d5 tDb4! 17 dxe6 ..txe6 18 0-0 ..td7! 19 g3 'iVc7 20 'ilYb3 0-0-0; Spraggett-J.Wat son, Columbus 1 977), but White can avoid both by the 12 cxd5 move order.
15 .te3 (a) 15 d5!? exd5 16 cxd5 tDxd5 17 'i'c2 Ratsch-Raderer, corr 1 982; 17 . . . tDcb4!? 18 'iVe4+ �d8! 19 l:1dl .l:.e8 20 �c4 �c8 and White is a pawn down and tied up (2 1 O-O?? tDb6); (b) 1 5 dxc5!? is similar to the last note. Black has done well in practice, but the a5 idea is still irritating: 15 . . . 0-0-0 16 a5 a6 (16 . . . l:thg8 1 7 0-0 e5 1 8 a6!) 1 7 0-0 (17 .i.c3 e5 1 8 'iVb3 lihg8 [18 . . . tDf5 1 9 'ib6!] 1 9 g3 ..tg4 20 'ib2 tDf5 ! with good play and the idea 21 h3? tDxg3 ! ! - + 0- 1 Mauro-Pieretti, corr 1 999, due to 22 fxg3 e4!-+) 1 7 ... e5 18 'iVb 1 ! ? .tg4 1 9 ':d1 f5 20
.i.c3;!; . 15 ... 0-0-0 16 d5 1 6 0-0 .i.e8! 1 7 d5 .l:.f8 18 'iVc2 exd5 19 �xh7 .tg6+ - Korchnoi . 16 ... exd5 17 j,xf6 17 cxd5 �e6 18 dxe6?! (18 j,xf6 .i.xd5 1 9 'iVc 1 ':hf8;) 18 . . . .l:.xd1 + 1 9 .l:.xd1 'iVf4! 20 0 - 0 .l:.g8+ Ekstrom Belyavsky, Stockholm 1 986. 1 7 ... lihg8 17 . . . l:thf8 has also been played with the idea 1 8 ..th4 �a5+ ! ' 18 exd5 And now 18 . . . ..te6! 19 0-0 .i.xd5! (improving upon two games with 1 9 . . . l:1xd5?! 20 'ilVc2) 20 �c 1? (20 'iVc2?? 'i'f4! ; best is 20 'i'd3 l:i.g6 2 1 .i.xe7 'iVxen) 2 0 . . . l:tdf8!+ 2 1 'iVc3? 'i'f4 22 ..txe7 ':xg2+! 23 �xg2 �g4+ 24 �h1 :xf3 25 'i'xc5 .i.e4- + . 1 1 .3 7 h4 An ultra-modern way of playing it: White will grab space on the kingside and worry about his development later. He also preserves some leeway as to where to put his pieces, depend ing upon how Black replies. Before turning to h4 proper, an important move order that I have put here to group similar ideas (which can of course transpose) is 7 tDf3 tDbc6 8 h4 (the less common 7 h4 tDbc6 8 tDf3 transposes) . In both move orders, White waits to see how Black re sponds, e.g., after 7 h4, he may think that 7 . . . tDbc6 requires 8 tDf3, whereas after 7 . . . .i.d 7 he might play 8 h5. Even after 7 ... �c7 White can forge ahead with 8 h5!? cxd4 9 cxd4 'iVc3+ 10 .i.d2 'iWxd4 n tDf3 'iVe4+ 12 .te2, a gambit whose reputation is not bad. Then White threatens l:th4, but he can also play for 0-0 and .i.d3. Like wise, after 7 tDf3, White may plan to answer 7 . . . .i.d7 or 7 . . . 'iVc7 with 8 a4
20 7
Play the French (for example), but prefer to answer 7 . . .'�Jbc6 with 8 h4, which is the posi tion we are looking at (see the dia gram).
I will present two unusual solu tions: (a) As far as I know, 8 .. .f6!? is not suggested in the major sources, yet it seems fully playable. Mter 9 h5 (9 exf6 gxf6 is inconsistent and must be fine for Black; now both 9 iLb5 'iVa5 and 9 .i.e2 fxe5 10 dxe5 'iVc7 are also unattractive) 9 . . .fxe5, play might con tinue: (al) 10 h6 exd4!? (10 . . . g6 1 1 dxe5 'iWa5 - among others - 12 .id2 .id7 is also possible) 1 1 hxg7 l:tg8 12 cxd4 (12 1:1.xh7? ! , and 12 . . . e5 is playable, e.g., 13 cxd4 cxd4 1 4 ltJg5 .tf5 1 5 l:1.h8 'it>d7!; but 12 . . . 'iWa5 ! threatens to get castled, while 13 ltJg5 e5! 14 'iWh5+ 'it>d8+ threatens . . . 'iWxc3+, . . ..tf5 etc.) 1 2 . . . cxd4 (or 12 . . . .l:r.xg7 1 3 .ih6 .l:r.g4!? intending 14 dxc5 e5) 1 3 ltJxd4 (13 l:[xh7 e5) 1 3 . . JIxg7 14 "i!kh5+ lIf7 15 ltJf3 ltJg6 16 ltJg5 'ii'f6 ! ; (a2) 10 ltJxe5 ltJxe5 1 1 dxe5 'ii'c 7 12 f4 iLd7 (12 ... ltJf5 1 3 Ith3!?) 1 3 iLe2 0-0-0 14 0-0 g6! 15 iLe3 (15 h6? ltJf5) 15 . . . Ithg8 (15 . . . gxh5! 16 �xh5 ltJf5+) 16 .i.g4 gxh5 1 7 iLh3 ! l:i.g3+ Fontaine Bergez, Clichy 1 998; (b) 8 . . .'iWa5 9 �d2 f6!? 208
I like this very direct move. Just about everything has been played in this popular position, among them 9 . . . iLd7, 9 . . . c4, 9 . . . 'ii'c 7, 9 . . . b6, and 9 . . . cxd4 10 cxd4 'iWa4 (transposing to one of our main lines below after 1 1 h5!?) . Mer 9 . . .f6!? then: (b l) 10 c4 'iWc7 1 1 cxd5 ltJxd5 12 c4 ltJde7 13 exf6 gxf6 14 dxc5 e5 15 .tc3 Reis Luis-Dias, Lisbon 1 999, and now 15 . . . .i.e6! 16 'iWd6 'iWc8! intending . . . 0-0 and . . . l:i.d8 or a variety of ideas such as . . . ltJg6-f4, or . . . iLg4 and . . . ltJf5d4; 10 h5 fxe5 1 1 h6 gxh6! 12 ltJxe5 ltJxe5 13 dxe5 iLd7 14 Itxh6 0-0-0 1 5 "i!kh5 ltJg6 1 6 J:!.xh7?! Itxh7 1 7 'iWxh7 ltJxe5 18 'ii' g 7 ltJc6 19 iLd3 c4 20 .i.g6 e5+ Pulkkinen -Manninen, Finnish Teams 1989; White's pieces are tied up and his king has nowhere to go. (b2) 10 exf6 gxf6 1 1 ltJh2 ! (an in genious try, planning simply ltJg4 to pressure Black's kingside; 1 1 a4 'iic 7 '!+' with the idea . . . e5 is given by Shi rov, and 1 1 . . .�d7! is a good alterna tive) 1 1 . . .'ii' a 4! 12 l:tb l c4 13 'iWh5+ �d8 14 ltJg4 (14 Itcl Itf8!? 15 ltJg4 'iVxa3!? 16 "i!kh6!? ltJf5 1 7 'ii'xh7 iLd7=) Shirov-Plaskett, Reykjavik 1992; and now instead of 14 . . . e5?, Shirov gives 14 . . :i!i'xc2 ! 1 5 Itcl 'ili'e4+ 16 .te3 'ili'f5! 1 7 'iWxf5 ltJxf5 1 8 ltJxf6 ltJxe3 19 fxe3 We7 20 ltJg4 b5+ (perhaps 20 . . . h5 21
Winawer Variation: Positional Lines �h2 e5:j: is better) . Overall this is a fun line that seems dynamically bal anced. Returning to 7 h4:
7 ...lDbc6 8 h5 �a5 8 . . . h6 is another standard answer, when again there are some unique sequences like: 9 �g4 lDf5 19 �d3 0-0 11 lDe2 c4 12 SLxf5 exf5 13 �g3 �h7=, and 9 lDf3 SLd7!? 10 �d3 'Wic7 1 1 0-0 c4 12 �e2 f5 ! , avoiding 12 .. .£6 13 �f4 ! . 9 �d2 And now I recommend two sys tems, beginning with: 1 1 . 3 1 9 ... SL d7 1 1.32 9 ... cxd4 1 1 . 3 1 9 ...�d7 Black emphasises quick develop ment. 10 lDf3 10 h6 is the more common order: 10 . . .gxh6 1 1 lDf3 0-0-0 etc. Here 1 1 l:[xh6 is possible, but then 1 1 . . .�c7!? 12 lDf3 lDf5 13 l:Ih3 cxd4 14 g4 dxc3:j: is good. 1 1 l:Ib l ! ? allows a draw by 1 1 . . . �xa3 12 l:Ial �2 1 3 l:Ib l �a3, and Black can also play 1 1 . . .0-0-0, because 12 l:tb5?! �a4! 1 3 l:Ixc5 b6 wins the exchange (unless 1 4 l:Ib5 �xd4 1 5 l:Ib2 ttJdc6+ follows) .
10 ... 0-0-0
1 1 h6 1 1 l:Ib l h6!? 12 l:Ib5 �xa3 13 � 1 can lead to 1 3 . . . c 4 14 l:Ixb7 �a6 1 5 l:Ib2 �c7 16 l:Ia2 l:Ib8 1 7 �al �5 1 8 � c 1 � 1 = Wang Zili-Reefat, Calcutta 200 1 ; or to the crazy 13 . . . ttJxd4!?, when 1 4 cxd4 �xb5 1 5 SLxb5 a6 16 �d3 lDc6 1 7 dxc5 �xc5 is hard to as sess; .. .£6, . . . b5, and . . . ttJa5-c4 are all possibilities; and 14 ttJxd4! SLxb5 1 5 ttJxb5 �a4 with the idea 16 ttJd6+ l:Ixd6 17 exd6 ttJf5 is also highly un clear. 1 1...gxh6! 12 i.d3 (a) 12 l:Ixh6 l:Idg8! has done well (12 . . . ttJg8!? resembled our main line in Berndt-Wegener, Graz 2002, al though White has many more op tions), e.g., 1 3 g3 (13 l:Ib l �c7 14 l:Ih4 ttJf5 15 l:If4 f6 16 exf6 e5! 1 7 dxe5 ttJxe5 18 ttJxe5 'Wixe5+ 1 9 �e2 �xf6+ Kotter-Kindermann, Germany 2001) 1 3 ... ttJf5 14 l:lh5 i.e8 1 5 i.d3 f6 16 �xf5 �xh5 1 7 �xe6+ �b8 1 8 i.xg8 l:lxg8 19 exf6 l:If8+ Vujadinovic Karabalis, email 1999; (b) 12 a4 ttJg8! has historical value: 13 SLb5 lDce7? (13 . . . c4= or 13 . . . �c7=) 14 dxc5 a6 15 c4 �c7 16 cxd5 ttJxd5 17 c4± Drimer-Uhlmann, Raach 1 969; a game that no one seemed to notice for 30 years !
209
Play the French 12 ... c4 13 �e2 liJ g8!
Another brilliant Uhlmann idea (we saw the precursor above), retreat ing for positional reasons. With an obvious move like 13 . . . liJf5 or 1 3 . . . .l:thg8, Black won't get in the break . . .£6, which is the move needed to contest the key central area. 14 a4 14 'it>£1 wasn't much better after 14 . . .£6 1 5 �e l fxe5 16 liJxe5 (16 dxe5!? l:.f8 1 7 g3 Tischbierek Uhlmann, Baden-Baden 1992; and now 17 .. :ilc7 gives Black the advan tage - Uhlmann, with the idea 1 8 �f4 liJce7 1 9 liJd4 liJg6) 1 6 . . . liJxe5 1 7 dxe5 liJ e 7 18 .i.xh6 .:I.hg8 19 i.f3 .i.e8! 20 a4 i.g6+ Short-Psakhis, Isle of Man 1999. 14 .. .lU8 14 . . . f6 may be just as good or better in view of White's next idea, but with 14 . . . .:.f8, Black hopes to gain a tempo by not ceding h6 yet. 15 'ilVcl ! ? f6 16 �a3 ':'f7 17 i.f4 liJge7 18 exf6 ':'xf6 1 9 .i.xh6 19 .i.e5!? wasn't much of an im provement in Sedlak-Kristjansson, Budapest 2002: 19 . . . liJxe5 20 dxe5?! (20 liJxe5 liJc6 2 1 liJxc6 i.xc6=i=) 20 . .1:1.£5 (or 20 . . . .l:.g6! 21 liJh4 .l:.g7 22 liJf3 liJf5) 2 1 l1xh6 liJg6 22 �4 �c7 23 �d6; and although 23 . . . 1!i'xd6 .
21 0
should have given Black the edge, 23 . . ..l:.hf8! (or 23 . . . .l:.g8 24 l:txh7 liJxe5) 24 'iYxc7+ �xc7 25 .l:txh7 liJxe5 would have secured Black a permanent structural advantage. Mter 19 .i.xh6, Hector-Hillarp Per sson, York 1 999 continued 19 . . . .l:.g8 20 'it>fl liJf5 2 1 i.d2 .:I.fg6 22 g3 e5 23 .u.xh 7?! (but 2 3 liJxe5 liJxe5 24 dxe5 .:I.xg3! or 23 dxe5 ltxg3 24 fxg3 liJxg3+ 25 �f2 'ilb6+ leaves Black much bet ter - Pedersen) 23 . . . e4 24 liJe5 liJxe5 25 dxe5 e3! 26 i.xe3 liJxe3+ 27 fxe3 .l:.xg3 28 c,t>f2 'i!Vb6 29 :£1 l:.xe3 30 a5 l:.g2+! 31 'it>xg2 'iYg6+ 0- 1 . 11.32 9 . . . cxd4 10 cxd4 �a4
A sound second system for Black. I will deal with the main line in less detail than 9 . . . i.d7. l l liJf3 (a) 1 1 c3 is better played when the move liJf3 isn't in, because in this kind of position White wants to play liJe2, defending c3 and leaving open the moves f4 and liJf4. Black has no real trouble after 1 1 . . .'i¥xd1+ 12 Ihdl h6 13 g4!? (13 �d3 i.d7 14 .l:.b l b6 1 5 liJe2 liJ a 5 16 g4 liJc4= Dowden-Wang, Wanganui 2003) 13 . . . .i.d7 14 f4 .u.c8 15 ..td3 liJa5 16 �b l b6= Shaked Remlinger, Los Angeles 1993; (b) 11 iLc3 b6 is easy for Black, e.g.,
Winawer Variation: Positional Lines 12 h6 gxh6 1 3 'ii' d 3 (13 tZ'lf3 �a6 14 �xa6 'ifxa6 1 5 'ii'e 2 'ii'xe2+ 1 6 <Jtxe2 ':'c8 17 <Jtd3 J:!.g8 18 ':'agl .l:1.g6= Pritchett-Rogers, Blackpool 1 988) 13 . . . a5 14 �d2 tZ'lf5 15 .l:r.b l!? tZ'lcxd4 16 <Jtdl Short-Korchnoi, Wijk aan Zee 1987, and now 16 . . . i.a6! ; . 1 l . . .tZ'lxd4
12 i.d3 This is a gambit made famous by Kasparov. I don't think that it pre vents Black from equalising, how ever, and he can sometimes do better than that. 12 ...tZ'lec6 The main move by far, but what I suggested in the last edition still de serves tests: 12 . . . tZ'lef5!? 13 .l:!.b l (13 h6 gxh6 14 g4!? and 14 . . . tZ'le7 unclear or even 14 . . . tZ'lxf3+ 15 'iYxf3 tZ'ld4 16 �f6 ':'g8, e.g., 17 lIxh6 tZ'lxc2+ 18 �xc2 'iWxc2 19 .i.b4 'ii'e4+ with perpetual check; 13 �b4 tZ'lxf3+ 14 'ilVxf3 a5) 13 . . . tZ'lc6 (13 . . . a5!?) 14 <Jtf1 b6 15 'iYe2 ! i. d 7 1 6 h 6 Dominguez-Harikrishna, Oropesa del Mar 1 999, and 16 . . . tZ'lxh6!? 1 7 �xh6 gxh6 should give Black at least equal chances. 13 <Jtn tZ'lf5 14 <Jtgl An attempted improvement upon 14 i.. xf5 exf5 15 h6 J:tg8 16 �g5 �e6= Hellers-Gulko, Bie1 1993.
14 ...'ilVg4 15 'ilVe2 f6 16 h6! fxe5 1 7 hxg7 lig8 1 8 J:txh7 e 4 1 9 tZ'lg5 'iYxe2 20 �xe2 1hg7 2 1 l:th8+ '?' - McDonald, who gives instead 2 1 lhg7 tZ'lxg7 22 c4 'with sustained pressure'. Fair enough, but is Black actually worse? I doubt it, for exam ple, 22 . . . We7 23 ':'dl (23 .i.c3 tZ'lf5 24 J:tdl d4 25 .i.b2 e3) 23 . . . tZ'lf5 24 .i.cl ! tZ'lcd4 25 i.g4 dxc4! 2 6 tZ'lxe4 e 5 2 7 i.g5+ <Jtf7 (with the idea 2 8 �f6? tZ'lg3) , and it's hard to choose between White's bishop pair and Black's out post and extra pawn. A verdict of equality seems fair.
2 1 .. .';ir>e7!? McDonald annotates: '2 1 . . .<Jtd7! heading for the queenside. Black seems clearly better, e.g., 22 f4 - to stop the threat of 22 . . . e3! cutting off the defence of the knight - 22 . . . Wc7! 23 tZ'lxe6+?! .i.xe6 24 l:.xa8', and here he likes 25 . . . tZ'lh4, but I think that simply 24 . . . tZ'lfd4! gives a clear edge. Mter 2 1 . . .'it>e7, Fedorov-Gulko, Las Vegas USA 1 999 went 22 f4, and Fe dorov suggests 22 . . . e5! 23 .i.g4 e3 24 �el tZ'ld6 ! ; it would be odd if Black weren't at least equal in that case. Thus the 9 . . .cxd4 and 1O . . . 'iWa4 line looks like another good solution to 7 h4.
21 1
Chapter Twelve Winawer Variation : Black Plays 6 'iVc 7 . . .
This chapter was guest written by Hans Olav Lahlum, an expert in and great enthusiast of the 6 . . :i8c7 Winawer. It describes a variation that was included in the second edition but not originally intended for this one. Since Black has the options of the main 6 . . . ltJe7 Winawer lines and the Classical 3 . . . ltJf6 lines, I had decided to forego an update of the early queen development. However, players from all over the world have expressed interest in the 6 . . . 'i!Vc7 repertoire from Play the French 2, so when Hans Olav offered to write up a survey of the 6 . . . 'iVc7 system using those lines I was happy to agree. He not only has extensive experience in the lines recommended, but has a sharp critical eye. The following is the result of his work, with editing and a few suggestions on my part. 1 e4 e6 2 d4 d5 3 ltJc3 �b4 4 e5 c5 5 a3 i.xc3+ 6 bxc3 'i!Vc7 First popularized in the 1940s by World Champion Mikhail Botvinnik and then in the 1960s by Edmar Mednis and Jan Hein Donner, 6 . . . 'i!Vc7 might now be characterised as a
21 2
'Scandinavian' variation, since it has been played by a number of young players from Norway and Sweden in recent years. Among others, Danish GM Curt Hansen, Swedish 1M Emanuel Berg, and Norwegian GMs Rune Djurhuus and Leif Erlend Johannessen have all been successful with 6 . . :iWc7 in numerous games. Black's main point when playing 6 . . . 'iVc7 is to avoid the critical 6 . . . ltJe7 variations involving 7 'ilVg4 followed by 8 'i!Vxg7, because now 7 1\Vg4 f5 protects the g7-pawn with the queen and also wins a tempo on White's queen. For his part, Black hopes to exert pressure down the c-file and immediately threatens 7 . . . cxd4 8 cxd4 'iVc3+. When he plays .. .f5 Black naturally loses the thematic French idea of . . .f6 with play along the f-file. By delaying . . . ltJe7, Black might also gain time for exchanging his troublesome light-squared bishop by . . . b6 and . . . �a6, which incidentally deprives White of his bishop pair and removes his best kingside attacker. On the minus side, of course, he delays the development of his kingside. In addition, Black has
Winawer Variation: Black Plays 6. :ilc 7 .
exchanged his dark-squared bishop and can be vulnerable on the kingside to attacking strokes like ttJg5 and/or 'i¥h5. He will often find it effective to castle long or leave his king in the centre in order to launch a kingside pawn storm. Thus White has to take into account each of these possible king placements. Black's king's position will also influence whether he closes the queenside with . . . c4 or opens it up with . . . cxd4; often Black remains flexible by postponing both decisions as long as possible. Piece exchanges will usually favour Black, since White has better possibilities for a kingside attack while Black's better structure will favour him in the endgame. Finally it's worth mentioning that because the centre commonly remains closed, flank attacks are critical for both players. White's main alternatives are:
ttJbc6 (8 . . b6 intending 9 . . . .lta6 is also sound) 9 ii.e2 ii.d7 10 0-0 0-0 1 1 ttJh4 A.Geller-Peterson, USSR 1 967, after which Korchnoi gives 1 1 . . .ttJg6 12 ttJxg6 hxg6= . 1 1 . . .f6!? tries for more and is probably better; (c) 7 h4?! is too ambitious: 7 . . . cxd4!? (or 7 . . . ttJe7) 8 cxd4 'iVc3+ 9 �d2 'ii'x d4 10 ttJf3 'iVb6 1 1 �d3 ttJc6 12 h5 f6 13 h6 g6 14 exf6 ttJxf6 1 5 .l:r.b l 'iVc7 and Black consolidated in Van Delft-L.E.Johannessen, Guarap uava 1995. .
12.1 7 'if g4 f5
12.1 7 'ii'g4 (going for an immediate kingside attack) 12.2 7 ttJf3 (settling for piece devel opment)
Now White has: Other moves are played much less: (a) 7 ttJe2?! leaves the knight on a less active square and deprives the £I -bishop of its natural development. Then 7 . . . ttJe7 intending 8 . . . b6 is normal, but directly 7 . . . b6!?, to get in . . . �a6 as soon as possible, might be even better. One example is 8 g3 ii.a6 9 �g2 ttJc6 (9 . . . ii.xe2 and 9 . . . cxd4 are also attractive) 10 O-O? cxd4 1 1 cxd4 �xe2 12 'iVxe2 ttJxd4-+ Raddatz-Leer Salvesen, Hamburg 1999; (b) 7 �b2?! also seems too slow to challenge Black: 7 . . . ttJe7 (7 . . . b6 is also interesting, as 8 'ifg4?! f5 9 �b5+?! 'it>f8! followed by . . . c4 and . . . a6 traps White's bishop at b5 - Bjerke) 8 ttJf3
1 2 . 1 1 8 'ifg3 12.12 8 'iVh5+
8 exf6?! ttJxf6 9 'iVg3 has been played by GM Alexei Shirov, but that's the most positive thing that can be said about it: 9 . . . 'iVa5 ! 10 .ltd2 0-0 (,:1=' - Pedersen) has the idea 1 1 c4? 'ifxd2+! 1 2 'it>xd2 ttJe4+ - + . Instead, 1 1 'iWh4 'iWa4! 12 l:.a2?! (12 ttJf3 ttJc6=i=) 12 ... e5! was - + in Mork Djurhuus, Bergen 2000. 1 2 . 1 1 8 'i/g3 This might be called the traditional main line, keeping up pressure
21 3
Play the French against g7 and hoping for a kingside attack. 8 cxd4 Safer than 8 . . . lDe7!? - compare PTF2. 9 cxd4 lDe7 9 . . . 'ii'xc2?! 10 �d2 Wlic7 11 lIc l lDc6 12 .i.b5 gave too much pressure for a pawn in Layton-Kimber, corr 1965. ...
10 .i. d2 Practically forcing Black to commit his king. 10 "WIxg7?? l:I.g8 1 1 Wlixh7 "WIc3+ loses, but after 10 i.. d2 Black has to reckon with 1 1 Wixg7. One alternative is particularly important: (a) 10 h4?! 'ii'xc2!? 1 1 .i.d2 lDbc6 12 lDe2 0-0 1 3 h5?! f4! is already better for Black, Kniazer-Wrichselbaumer, Amsterdam 1 954; (b) 10 i.. e 2!? probably works tactically after 10 . . . "WIxc2 1 1 i.. d 2; but Black should play as in the main line by 10 . . . 0-0 1 1 lDh3 b6 12 lDf4 Wlid7 1 3 0 - 0 i.. a 6; (c) 10 c3 0-0 11 lDh3 (1 1 lDe2!? transposes into 10 lDe2 0-0 11 c3; 1 1 i.. e 2 b 6 intending 1 2. . . i.a6 - Moles; 11 h4 b6 1 2 h5 i.a6 1 3 i.xa6 lDxa6 14 lDh3 l:I.ac8= Firnhaber-Kaiser, corr 1987) 1 1 . . .lDg6!? 12 �d2 b6 13 �d3 i.. a 6 1 4 �xa6 lDxa6 1 5 0-0 .l:!.fc8 16 lDg5 lDf8= Pfeifer-Kacirek, Klatovy 1997;
21 4
(d) 1 0 lDe2!? 0-0 1 1 c3 has scored very well in practical play; this idea might become important because ultimately 10 �d2 seems fine for Black. By omitting .i.d3 White hopes to save a tempo on the main line, as he did after 1 1 . . . b6 12 lDf4 iVd7 13 h4 i.a6 14 i.xa6 lDxa6 15 h5 in Laursen B.Lundberg, corr 1986 and after 1 3 a 4 i. a 6 14 i.b5+!? �xb5 1 5 axb5 a 5 16 � a 3 l:I.e8 1 7 h 4 li a 7 18 h5 in Ernst-Nordahl, Gausdal 1994. Of course neither line gives White a decisive advantage. Instead, Black can play 1 1 . . .lDbc6 (with the idea . . . lDa5), and now: (dl) 1 2 lDf4 lDa5 1 3 l:Ib l lDc4 14 h4 (or just 14 i.e2 intending 0-0) of Doggers-Westerman, Utrecht 2000 gave White a pleasant advantage. (d2) 12 a4 lDa5 13 i.a3 lDc4 14 .i.xe7 Wlixe7 1 5 lDf4 with an edge, Kruppa-Feigin, Nikolaev 1995; (d3) What is probably the most critical line was tested in Sax Lahlum, Hamburg 2002, which went 12 h4!? lDa5 1 3 i.g5 (Tiemann's 13 l:tb l i.d7 14 h5 .i.a4 1 5 i.. g5 is also interesting) 13 . . . .l:tf7 14 lDf4 (14 h5!?) 14 . . . i.. d 7 (or 14 . . . Wlib6 1 5 i.. e 2 lDc4 16 .i.h5 g6 1 7 i.e2± Pavlovic-Dgebuadze, Ubeda 1997; this attacking manoe uvre is worth noting, especially with a Black rook on f7: White spends two tempi to force . . . g6 because h5 will then be very powerful) 1 5 i.e2 ILc8 16 l:.h3?! (16 i.h5 g6 1 7 i.e2 'i!kxc3 and here White has some compensation) 16 . . . 'i!kb6 1 7 .l:tcl?! Wlib2 1 8 l:I.dl, when 18 . . . i.a4! would have forced White to give up an exchange for insufficient compensation. Although 10 lDe2!? can hardly refute this variation, it seems at the moment to be White's best bid for an advantage and is the only move
Winawer Variation: Black Plays 6. . . 'iVc 7 against which Black has not (yet) demonstrated how to equalise. 10 ... 0-0
1 1 i.d3 Mter the seldom-played 11 tiJh3!?, 1 1 . . :ii'xc2?! 12 Itc 1 gives White at least adequate compensation, but 1 1 . . .tiJbc6 12 c3 tiJg6 13 1t.e2 1t.d7 14 0-0 tiJce7 was solid and at worst ;!; in A. Fischer-S.Meyer, Hessen 1 992. Also natural is 1 1 . . .tiJg6!? planning 12 i.e2 tiJbc6 and otherwise ... b6 and . . . i.a6. 11 b6 12 tiJ e2 12 h4?! 1t.a6 was Ernst-Djurhuus, Gausdal 1 994, and Ernst gives 13 tiJe2 (13 h5? i.xd3 1 4 cxd3 tiJbc6 1 5 tiJe2 f4! 16 'iVg4 tiJxd4+) 1 3 . . . .i.xd3 14 cxd3 tiJbc6 1 5 l:Ic1 'iVd7= . 12 i.a6 1 3 tiJf4 (a) 1 3 i.b4?! Itc8 14 tiJf4 'iVd7 might transpose to 14 i.b4?! , which seems better for Black. An attractive alter native is 13 . . . 1t.xd3 14 cxd3 tiJbc6: 1 5 1t.d6 'iVd7 16 0 - 0 Itf7 (or 16 . . . Itfc8) 1 7 1t.xe7 (White avoids a bad bishop versus good knight ending; not 17 a4? f4!+ Gabrielsen-Djurhuus, Oslo 1 994) 17 . . . l:Ixe7 18 lUc1 l:Ic8 19 'iVe3 (19 Itc3? tiJxd4! 20 lIxc8+ 'iVxc8 21 tiJxd4 'iVc3 Schleiffer-Just, COIT 1 987) 19 . . :iVe8 20 .l:ta2 Itec7 21 Itac2 'iVd7 22 f3 tiJd8:j: due to Black's better pawn .•.
...
structure, Koch-Schnicke, corr 1990; (b) The slow 1 3 O-O?! was Gullak sen-Djurhuus, Alta 1996: 13 . . . ..txd3 14 'iWxd3 Itc8 15 l:Ifc1 'iVc4 16 'iWxc4 Itxc4 17 a4 tiJbc6 18 c3 tiJa5+ . This example illustrates that White has to do something in the middlegame so as not to end up in a worse ending. 1 3 ...'iVd7
14 h4 (a) 14 .i.b4?! Itf7? 15 tiJh5! with an attack was the old main line, which now seems refuted by GM Grigory Kaidanov's 14 . . . l:Ic8!, coolly pointing out that Black's rook is not needed to defend the kingside: 15 'iVg5 (15 tiJh5 tiJg6 or 15 . . . i.xd3 followed by the move 16 . . . tiJbc6; 15 'iVh4?! tiJec6 [or 15 . . . i.xd3 16 cxd3 tiJec6] 16 .l:!.d1?! J.Berg-Djurhuus, Alta 1 996, and I see nothing wrong with 16 . . . tiJxd4! ; 15 h4 i.xd3 16 1!Vxd3 l:lc4 17 i.d2 tiJbc6 1 8 tiJe2 Itc8 with the initiative, Zaw Htun-Villamayor, Vung Tau 2000; 1 5 0-0 i.xd3 1 6 1!Vxd3 Itc4 with coun terplay as in the main line) 15 . . . tiJec6! 16 tiJh5 'iVf7 17 l:Id1?! (17 i.xa6 tiJxa6 18 c3 tiJaxb4 19 axb4 h6 20 'iVh4 a5! opens up the queenside to Black's advantage - Watson) 17 . . . tiJxd4! 1 8 tiJf6+ �h8 19 tiJxh7 tiJxc2+-+ Shaked Kaidanov, San Francisco 1 995; (b) 1 4 0-0 i.xd3 1 5 cxd3 (15 'iWxd3
21 5
Play the French liJbc6 16 h4 .l:t.ac8 17 h5 liJd8 18 g3 .l::t c 4 19 'iit g2 liJec6= Leko-C.Hansen, Groningen 1995) 1 5 . . . liJbc6 (15 . . . .l:t.c8) 16 .te3 Wh8 (not necessary because Black can always meet liJh5 by . . . liJg6; 16 .. .l':tfc8!?) 17 .l:t.ac1 Mednis Foldi, Varna 1958, with a position similar to the main line. Moles' 17 .. J::tac8 18 h4 liJd8 looks = since White has to watch out for queenside; (c) 14 liJh5?! liJg6 1 5 i.xa6 liJxa6 1 6 h 4 threatens 1 7 liJf6+ gxf6 1 8 h5 with some attack, but then 16 . . . 'iith 8! preparing 17 . . . .l:t.ac8, 17 .. JWa4 or 17 .. .f4 - Watson; (d) 14 .i.xa6 liJxa6 15 'iWd3 (15 h4 .l::t£7 ?! 16 h5 .:rc8 17 c3 .l:t.c4 18 h6 g6 1 9 liJe2 liJb8 20 .i.g5 'iVc8= a s i n Ze zulkin-Matlak, Czech Republic 2003. 1 5 ... .l:i.ac8 intending 1 6 c3 .l:t.c4 looks more pointed. In general, . . . .l:t.£7 should be played only when it is absolutely necessary to defend the kingside) 15 . . . liJb8 16 h4 .l:t.c8 is the main line. Instead, Reshevsky-Bot vinnik, Moscow 1 948 went 16 . . . liJbc6 17 .l::th 3 (17 h5 .l:t.ac8 18 'iit e 2 liJd8 1 9 .l:i.acl O .Karlsson-Hylen, corr 1973, and again 19 ... .l:.c4=i= seems best) 19 . . . .l:i.ac8 1 8 llg3 'iith8?! (18 . . . liJd8; compare 15 . . .liJbc6 in the main line) 19 h5 .l:i.£7 20 h6 g6 2 1 .l:t.c1 .l::tff8 = .
14 ... i.xd3 15 'iVxd3
21 6
This loses a tempo for the kingside attack, but 15 cxd3?! is seldom played as it destroys White's pawn structure and gives Black a queenside pawn majority, e.g., 1 5 . . . liJbc6 16 i.e3 liJa5 1 7 h5 .l:t.ac8 1 8 h6 g6 19 0-0 'iith 8 20 .l:i.ab 1 liJg8 21 .i.d2 liJc6 22 liJe2 liJd8 23 .l:i.fc 1 liJ£7= Smyslov-Botvinnik, Moscow 1957.
15 . . .l::t c 8! I consider this better than the natural and often-played 1 5 . . . liJbc6?! (this position can also be reached after 14 .i.xa6 liJxa6 15 'i'd3 liJb8 1 6 h 4 liJbc6) because Black can increase the pressure against d4 by playing . . . .l:t.c4 before . . . liJc6. One prominent example with 15 . . . liJbc6?! was 16 .l:lh3 .l:i.ac8 17 .l:i.g3 .l::t£7 ?! (Kasparov gives 17 . . . liJd8! 18 liJh5 liJg6 as unclear; one line is 19 liJxg7 c;t>xg7 20 h5 liJ£7 2 1 c;t>e2 'iVd8 2 2 .l:t.hl llh8 23 c3 'iWe7) 1 8 h5?! (18 'iitf1 ! liJd8 19 c;t> g l;!; - Kaspa roy) liJd8 19 c3 .l::tf8?! (19 . . . 'iWa4=i= Kasparov) 20 c;t>f1 .l::t c 4 2 1 c;t>gl liJ£7 22 a4 .l:i.fc8 with balanced chances, Kasparov-Short, Novgorod 1997. Short later preferred 1 5 ... .l::t c 8. 16 .l:t.h3 16 h5 .l:t.c4 17 h6 g6 18 'iWg3 liJbc6 19 c3 'iit h 8 20 'i!Vh4 .l:t.f8 21 0-0 liJa5 22 .:rfb 1 liJg8= Teichmeister-Halldorsson, corr 2000. The primitive h5-h6 sel.
Winawer Variation: Black Plays 6. �c 7 ..
dom gives White anything unless he can infiltrate the dark squares after . . . g6. 16 .. .lIc4 17 �g3 17 ltJh5?! (threatening Ji.h6) 1 7 . . . 'iVe8?! (17 . . . ltJbc6!? intending 1 8 i.h6? ':'xd4 o r 1 8 . . . ltJbc6!? with the idea 1 9 ltJxe6?! 'iVc8) 1 8 ltJf4 1!Vd7?! Sebastian-Lahlum, Hamburg 2001 transposes back to the main line. 17 ...ltJbc6 18 c3 18 ltJe2?! �h8 1 9 h5 ':'c8 20 �f1 ltJd8 2 1 c3 ltJf7 22 'iit g l ltJc6 23 �hl ltJa5 24 ':'gl (not a convincing manoeuvre) 24 . . . ':'a4 25 l:i.h3 ltJc4 26 i.f4, and now both ways of taking the a-pawn would have been =+= , Michiels Wintzer, Amsterdam 200 1 . Black's queenside play advanced steadily while White failed to demonstrate anything on the kingside. 1 8 ... l:tfS The more subtle 18 . . . �h8 also has done well in several games, e.g., 19 1!Ve2 ltJa5 20 1!Vdl l:tc6 2 1 �f1 ltJc4 22 Ji.cl b 5 23 ltJh5!? (23 ltJd3 'iVe8 with about even chances - Psakhis) 23 . . . ltJg6 24 ltJxg7 �xg7 25 h5 .l:.h8 26 Ji.g5 �f8 27 Ji.f6 l:tg8 28 hxg6 ':'xg6 29 l:.h3 lIa6 Svidler- Cu.Hansen, Esbjerg 2000; best was 30 1!i'h5 �g8 3 1 'iVxg6 hxg6 32 l:i.h8+ r:J;f7 33 l:th7+!= Psakhis. 19 r:J;n 1 9 'iit e 2?! ':'f7 20 ltJh5 ltJg6 2 1 l:thl?! 'iWe7! 22 l::t xg6 hxg6 2 3 ltJf4 of D.Hersvik-Lahlum, Bergen 2003 was a creative kingside attack, but 23 . . . b 5 ! planning 24 ltJxg6 �xa3 would have given White insufficient compensation for the exchange. 19 l:tf7 20 liel 20 r:J;gl ltJa5 2 1 a4 'iVc6 22 l:ta3 �f8 23 � 1 ltJb7 Y:! - Y:! Teran Alvarez Short, Lanzarote 2003. Black would have to run risks to try to win. ...
20 ...'iWc8 This is the game Konstantino polsky-Olle Smith, corr 1955, given as =+= by Watson. That might be over blown, but Black is in control. The game went 2 1 'i¥f3?! ltJa5 22 n5? ltJb3 23 .i.e3 lixc3 24 l:i.dl ltJa5 25 �6?! ltJg6! 26 l:txg6 hxg6 27 �g5 ltJc4-+ , but for unknown reasons Black forced a draw a few moves later. As shown above, the old main line with 8 � g3 often ends with a balance of power: White has chances on the kingside and Black on the queenside, but both have difficulties making anything out of their position because they have to keep some pieces defending the other wing. If Black watches out for dramatic kingside breakthroughs, he will in any case avoiding losing. This main line is demanding for both players but seems sound for Black; hence White's most critical continuation might be the still rarely-played 10 ltJe2!? 12.12 8 'iVh5+
This might be called the modern main line. White is using tempi with his queen to provoke a weakening of Black's kingside. 8 .. g6 .
21 7
Play the French This is an important decision: Black weakens his kingside and the dark squares, but wins tempi for active counterplay on the queenside. More sound but less dynamic is 8 . . :iVf7, when White has three alternatives: (a) 9 'i!Vxf7+ �xf7 10 ttJf3 (10 .l:.b l cxd4 1 1 cxd4 ttJe7 12 .td2 b6 1 3 ttJf3 h6 intending 14 . . . .ta6 didn't give White any lasting advantage in Jakic-Haba, Bibinje 200 1 ; 10 dxc5!? ttJd7 11 c4 Sutovsky-Komarov, Vond roll 1996, and 1 1 . . .ttJxe5 13 cxd5 exd5 seems ;l; ; 10 c4 cxd4 1 1 cxd5 exd5, and now Uhlmann gave 12 ttJf3 ttJc6 1 3 .tb2 ttJge7 14 0-0-0 h6, but Gullak sen's 12 f4! safeguards e5 and is at least a ;I; after 12 . . . ttJc6 13 ttJf3) 1O . . . cxd4 1 1 cxd4 (1 1 ttJxd4?! .td7 12 f4 ttJe7 1 3 .l:.b l b6 14 c4 .l:.c8! was = or even :j: in Kindermann-Yusupov, Mu nich 1990) 1 1 . . .h6 12 c4 ttJe7 13 c5 (otherwise 13 . . . dxc4 followed by b6) 13 . . . ttJbc6= Vogt-Bodo Schmidt, Bun desliga 1993; (b) LBerg gives 8 . . :�'f7 as "?!" due to 9 .tb5+! .td7 10 1li'xf7+ �xf7 1 1 .l:.b 1 ± . White might very well be ;l;, but I am not sure about 1 1 . . . .txb5 12 .l:.xb5 b6 1 3 dxc5 ttJd7 1 4 cxb6 axb6 1 5 ttJ f3 h 6 with compensation. Black can also limit White's advantage by 9 . . . ttJc6 10 1li'xf7+ �xf7 1 1 ttJf3 ttJge7 12 .l:.b l h6, planning . . . .td7 or even . . . a6. Thus White has no clear road to an advantage after 9 .tb5+; in practice he only has a very slight plus score with a high drawing rate. Black plays without much margin for error, however, and must be careful not to open the position for White's bishop paIr; (c) 9 1li'dl is playable: 9 . . . b6 10 a4!? (10 ttJf3 .ta6 11 dxc5 .txfl 12 .l:.xf1 bxc5 1 3 c4 ttJe7= Csabo Horvath21 8
Uhlmann, Dresden 1988) 1O . . . .ta6 1 1 .txa6 ttJxa6 1 2 ttJh3 h6 (or 1 2 . . . ttJe7, as 1 3 ttJg5 1li'g6 is not really a threat) 13 ttJf4 ttJe7 14 1li'e2 c4 1 5 iLa3 (15 a5!? .l:.b8) 1 5 ... ttJc6, at best ;1; , Gullak sen-Lahlum, Copenhagen 1 998. 9 1li'dl This seems paradoxical, but although other moves might be underrated White has no better square for the queen: (a) 9 1li'h3 ttJc6 (9 . . . cxd4!? 10 cxd4 1li'xc2 planning . . . ttJc6, but White has compensation) 10 ttJf3 (10 ttJe2 .td7 11 g4?! fxg4 12 1li'xg4 0-0-0 was at worst = in Fletzer- Casaldi, Venice 1948) .td7 1 1 .te2 0-0-0 (1 1 . . .ttJge7; 11 . . .h6) 12 0-0 h6 13 a4 ttJge7 14 .ta3 g5 15 ttJd2 ttJg6 unclear, Boleslavsky Bondarevsky, USSR 1954; (b) 9 1li'f3 .td7!? (9 . . . cxd4!? 10 cxd4 1li'xc2 has scored well in a few games, e.g., 1 1 ttJe2 ttJc6 12 h4 h6 13 1li'e3 ttJge7 and Black converted her extra pawn, Polovnikova-Zielinska, Zagan 1997) 10 h4 ttJc6 1 1 ttJe2 0-0-0 12 h5 ttJge7 1 3 .tg5 .l:.de8= Westerinen D.Madsen, Gausdal 1 99 1 .
9 iLd7 (a) 9 . . . cxd4?! 10 cxd4 1li'c3+ 11 .td2 1li'xd4 1 2 ttJf3 might be better than it looks, but White gets (too) much development for a pawn; ...
Winawer Variation: Black Plays 6. . . 'iVc 7 (b) 9 . . . lLlc6!? was given by Watson and still seems sound after 10 a4 cxd4 1 1 cxd4 lLlb4 or 1 0 lLlf3 Jl.d7 1 1 dxc5 h6! 12 i:tb l lLlge7= from Spassky-Portisch, Mexico 1980. More critical than 10 a4 is 10 h4!? lLlge7? (Fyllingen gives 10 . . . h6! ; then 1 1 lLlf3 lLlge7 is complex and at worst ;1;) 1 1 h 5 Jl.d7 12 hxg6 lLlxg6 1 3 .i.g5 (13 lLlf3 might be even better) 1 3 ... cxd4 14 cxd4 'iWa5+ 15 'iVd2 'iYxd2+ 16 'it>xd2 lLlxd4 1 7 �f6 0-0 1 8 i:tb l Gullaksen-Lahlum, Porsgrunn 1 999.
10 lLlf3 (a) Against 10 h4, 1 O . . . h6 is most natural, but Black also has 1O . . . cxd4!?, when Watson's analysis goes 1 1 cxd4 (1 1 h5? dxc3 with the idea 12 hxg6 hxg6 1 3 i:txh8 'i¥xe5+) 1 1 . . . 'i'c3+ 12 Jl.d2 'i'xd4 13 lLlf3 (13 lLle2 'i'c4 14 h4 lLle7+ Kreiman Shaked, US Junior 1996) 13 .. :iVe4+ 14 .i.e2?! Jib5 + ; (b ) 1 0 a4 cxd4 (Watson suggests both 10 . . . h6 planning . . . lLlge7 and 10 . . . lLlc6 with the idea l 1 lLlf3 cxd4 12 cxd4 lLlb4 or 11 .i.a3?! cxd4 12 cxd4 iVa5+) 1 1 cxd4 'i¥c3+ 12 �d2 'i'xd4 1 3 lLlf3 'iVe4+ 14 .i.e2 lLlc6 1 5 0 - 0 lLld4 yielding only just compensation, T.Ptihtz-Dhlmann, Dresden 1 988. 1 0 ... Jia4!? This original but flexible bishop
move establishes pressure against c2. Black has two principle strategies: he can either play . . . c4, castle long, and attack on the kingside; or castle short or leave the king in the centre to attack down the c-fIJ.e. Instead, 10 . . . cxd4 1 1 cxd4 il.a4 can be played but Black can keep that option open.
l l i:tb 1 ! ? This is 'most critical and probably best' according to E .Berg. I will use it as the main line here, although five other alternatives have been played and three of them might be as good as 1 1 i:tb l : (a) 1 1 c4?! lLle7 12 �e2 ( 1 2 dxc5 lLld7 13 �e2 lLlxc5 14 0-0 0-0-0 1 5 il.d2?! lLle4 16 cxd5 i:txd5 1 7 'iVcl lLlxd2 1 8 lLlxd2 �xe5+ Ariel-Matlak, Groningen 1998) 12 . . . lLlbc6 13 0-0 h6 14 cxd5 lLlxd5 1 5 'iVd3? (15 dxc5 ! , but Black still has an active position after 15 . . . 0-0-0 and seems fine) 15 . . . cxd4 16 "iVc4?! b5! 1 7 'ilfd3 'i!fb6 1 8 lLlh4 'it>f7 1 9 l:I. e l 'i'c5 20 'ilfg3 lLlce7 0 - 1 Richard son-Kujala, corr 1993; (b) 11 i.. d 3 cxd4! (or Watson's 1 1 . . .lLld7, e.g., 12 0-0 c4!? 13 i.. e 2 h6 14 g4!? fxg4 15 lLlh4 lLlf8 16 i.. xg4 0-0-0 17 �hl g5, Anka-Koczka, Buda pest 1991) 12 cxd4 'i'c3+ 13 i.. d2 'iVxd3! 14 cxd3 �xdl 15 'it>xdl lLlc6 16 i:tb l (16 h 4 h 6 ! 1 7 �e2 lLlge7 1 8 a4
21 9
Play the French 'iitf8 19 .l:thb l .l:.b8 20 a5 <:Jitg7 21 l:!.b2 b6!� Stefansson- Short, Reykjavik 2000) 16 . . . b6 1 7 I1cl <:Jitd7 18 h4 h6 19 We2 ttJge7 20 l:!.c2 I1ac8 2 1 l:thc l ttJd8� J.Polgar-Morovic, Buenos Aires 1992; (c) 11 �d2 !? is also critical: 1 1 . . . ttJd7 (1 1 . . .cxd4? 12 cxd4! 'iVxc2?? 13 ..tb5+! or 12 . . . SLxc2? 13 ..tb5+ ttJd7 14 l:tc l lic8 1 5 'ilVe2 'ifd8 16 ttJg5! ; 1 1 . . .ttJc6 12 .i.d3 h 6 13 h 4 ttJge7 14 'iVcl 0-0-0 1 5 l:th3 l:!.h7 16 <:Jitfl l:!.dh8 17 'iit g l c4= Zhao-E.Berg, Athens 200 1) 12 .i.d3 ttJe7 (12 . . . h6 - Tie mann) 1 3 0-0 h6 14 h4 <:Jitf7!? 1 5 l:!.b l .l:Iac8 1 6 l:tb2 ttJb6 1 7 'ifc l c:J;g7 18 �e3 c4� F.Andersson-A.Olsson, Hallsta hammar 2002; (d) 11 l:ta2 !? overprotects c2 and removes . . . 'iVc3+ threats. On the mi nus side, the rook is less active at a2 than on b l , e.g., 1 1 . . .ttJd7 (or 1 1 . . .h6, but not 1 1 . . .ttJe7? 12 �h6! ttJd7 1 3 ..t d 3 cxd4 14 cxd4 'ifc3+ 1 5 'ifd2 ! Gullaksen-Lahlum, Asker 2000)
(dl) 12 h4 h6 1 3 i.e3 (13 �f4? ttJgf6! - planning 14 . . . ttJe4 - 14 �xh6 l:!.xh6 1 5 exf6 ttJxf6 1 6 i.. d 3 c4 1 7 'ilVc l 'ilVg7 18 �e2 ttJe4 19 'i!Yb2 g5 !+ Stokke L.E.Johannessen, Bergen 2002) 0-0-0 14 'ilVc l <:Jitb8 1 5 l:th3 l:tc8 16 �d3 'iit a 8 17 'iitfl c4 (closing off the queenside and planning . . . g5) 18 �e2 ttJb6 19 ttJgl 'i1Vg7 20 l:tg3 ttJe7 2 1 ttJh3 220
l:[cg8=/� Gullaksen-Leer Salvesen, Asker 2000; (d2) Critical is 12 j.,d3 ! : 12 . . . h6 (12 . . . c4?! 1 3 j.,e2 h6 Gdanski-Matlak, Cappelle la Grande 1 999, 14 ttJh4 ttJe7 15 g4 - E.Berg; 12 . . . cxd4? 1 3 cxd4 'iVc3+?! 14 We2! - Watson, or here 13 . . . i.b5 14 O-O! �xd3 15 cxd3±) 13 0-0 c4 14 ..te2 0-0-0 (14 . . . ttJe7 E.Berg) 1 5 h4 (15 g4!? fxg4 16 ttJh4) 15 . . . ttJe7 16 g3 ttJc6 17 c:J;g2 �dg8 18 lthl ttJd8 19 .i.e3 ttJf7 with a promising position Wedberg-E .Berg, Sweden 2002; (e) 11 dxc5!? tries to open up the queenside; it has been heavily debated among young Polish stars recently:
1 1 . . .ttJd7 (1 1 . . .'iVxc5!? 12 i.d2 ttJc6 13 i.d3 0-0-0 14 0-0 h6 15 'i!Vb l g5 16 ttJd4 'ilVe7 1 7 ttJxc6 ..txc6 looks solid, Dworakowska-Kadziolka, Brzeg 2001) 12 'ilVd4 (12 j.,e3?! ttJxc5 13 .l:.b l ttJe4 14 'ilVd4 SLd7 15 'i!Vb4 b6 16 c4 ttJe7 17 SLe2 l:!.c8� Carlsen-A. Olsson, Stock holm 2003) 12 . . . ttJxc5 and: (e l) 13 �d3!? h6 1 4 0-0 (or 14 h4 b6 15 0-0 0-0-0 16 i.f4 'iVh7 17 lIfb l ttJe7 18 'iitfl ttJc6 1 9 'iVe3 ttJe4 and Black evetually won in Dworakowska M.Socko, Ostrow 2002) 14 . . . ttJe7 (14 . . . b6!?) 15 SLe3 b6 16 'iVh4;!; Chandler-Kinsman, England 1 998;
Winawer Variation: Black Plays 6. . . iVc 7 (e2) 1 3 �e3 b6 14 �d3 h6 1 5 0-0 (Berg gives 1 5 h4 !:De7= and 15 'iVh4!? !:Dxd3+ 16 cxd3 iVxc3+ 1 7 �e2 �b5 18 iVd4 'with compensation') 15 ... g5 16 iVb4 �d7 17 c4 !:De7 1 8 a4 0-0 19 h3= Macieja-E.Berg, Pardubice 2002.
1 l ... !:Dd7 White's point is 1 1 . . .cxd4?! 12 l:tb4! (E.Berg) or even 12 !:Dxd4!? iVxc3+ 13 iVd2 iVxd2+ 14 i.xd2 with a strong initiative. 12 �e2 White later turned to 12 dxc5!? opening queenside lines: 12 ... !:Dxc5 13 .te3 b6 (13 ... !:De4? 14 iVd4 !:Dxc3? 15 i.d2 ! !:Dxb l 16 iVxa4+ �d8 1 7 iVb4 !:Dxa3 1 8 iVf8+ -+ Groszpeter-Atanu, Paks 2000) 14 iVd4 !:De7 (14 . . . .txc2? 15 i.b5+ �e7 16 l:tb2 i.e4 17 !:Dh4) 15 i.d3 h6 16 O-O?! (16 'iWh4!? !:De4 1 7 0-0 0-0-0 looks =) 16 ... g5! 1 7 c4 0-0 18 .l:tfe l f4 19 i.d2 (19 cxd5? fxe3 20 d6 iVd7 21 dxe7 .l::txf3!-+ - E.Berg) 19 . . . !:Dxd3 20 cxd3 !:Df5 2 1 iVc3 dxc4 22 l:te4 iVd7!:j: Wedberg- E.Berg, Malmo 2002. 12 ... h6 1 3 dxc5!? 1 3 0-0 c4?! (Black should be fine if he remains flexible with 13 . . . !:De7 Watson) 14 !:Dh4! !:De7 1 5 f4 0-0-0 16 i.e3 iVa5 1 7 iV d2;!; Velimirovic-Moro vic, Vrsac 1985. After 1 3 dxc5!?, Timman-Shaked,
Merillville 1997 went 1 3 . . . !:Dxc5 14 l:tb4 i.d7 1 5 0-0 !:De4 16 i.d3 !? !:De7 17 !:Dd4 (Timman mentions 1 7 c4, but 1 7 . . . a5 intending . . . dxc4 and 1 7 . . . !:Dc6 are good) 17 . . . !:Dxc3 18 iVe l !:Da2 1 9 l:tb3 !:Dxcl 20 iVxc l , and now 20 . . . b6?! 21 �b5 �fl 22 a4 l:tad8 23 l:tel a6 24 i.xd7 l:txd7 25 iVb2 l:tb8 26 l:th3 !:Dg8 27 l:tc3 gave White a strong initiative. Critical is 20 . . . iVxe5! (as suggested by Timman), e.g., 21 !:Df3! (2 1 c3?! iVc7! 22 f4 0-0 23 iVe3 [or 23 l:tel b6] 23 . . . l:tac8 24 !:Dxe6 i.xe6 25 iVxe6+ l:tfl) 2 1 . . .iVf6 (or 2 1 . . .iVd6 22 iVb2 0-0 23 l:txb7 l:tab8 - Lie; 2 1 . . .iVc7? 22 iVb2±, since 22 . . . 0-0 23 l:txb7 iVd6? 24 llVe5 ! wins - Tallaksen) 22 l:txb7 i.c6 23 i.b5 i.xb5. White still has to prove that he has compensation. Thus we can conclude (for now) that in the modern main line with 7 iVg4 f5 8 iVh5+ g6 9 iVdl i.d7 10 !:Df3 �a4!? White has many possibilities that have not been tested properly on a top level but he also lacks a clear way to any advantage. Black, having a flexible position, should be fine as long as he is not afraid to play the typical closed positions; those demand much patience and accuracy from both players. White's bishop pair is not much of an advantage and Black's better pawn structure often gives him the last laugh (if there is one) . 12.2 7 !:Df3 b6!? 7 ... !:De7 transposes a Winawer with 6 . . . !:De7 7 !:Df3 iVc7. With 7 . . . b6, Black claims that it is worth some lost tempi to exchange the light-squared bishops by 8 . . . i.a6. Recommended by Watson in the last edition, this line still seems satisfactory for Black. The similar 6 . . . !:De7 7 !:Df3 b6 has remain ed popular in recent years, and it is
221
Play the French not clear that . . . tDe7 is more useful than . . . flic7.
8 a4 This flank attack tries for a5 before Black can complete his development. (a) 8 iLd3?! falls in with Black's plans: 8 . . . .ia6 9 0-0 .ixd3 (9 . . . tDe7 10 dxc5 bxc5 11 l:te1 .ixd3 12 cxd3 0-0 1 3 .ig5 tDbc6 1 4 'ilVa4 h6 1 5 iLe3 tDxe5 16 tDxe5 'i'xe5 1 7 .ixc5 'ilVc7= De Firmian-E.Berg, Gausdal 2003) 10 cxd3 (10 "iVxd3 might be better, but 10 . . . tDe7 retained the advantage after 1 1 a4 h6 12 iLa3 0-0 1 3 tDh4 tDbc6 Blazek-Heyken, Pardubice 1996, and 1 1 c4 dxc4 12 'i'xc4 tDd7 13 dxc5 0-0 14 'i'e2 tDxc5 1 5 a4 lHd8 Sakhatova Couche, Cappelle la Grande 1995) 1O . . . tDe7. Here Black is equal, e.g., 1 1 a4 tDd7 12 .i.a3 0-0 1 3 :c1 1He8 14 c4 llac8 15 cxd5 tDxd5 Kaminski-Gdan ski, Warsaw 1995, or 1 1 dxc5 bxc5 12 a4 0-0 1 3 .ig5 h6 14 .ixe7 Y2 - Y2 Biolek-Matlak, Czech Republic 1994, or 11 l:te1 tDbc6 12 'i'a4 0-0 13 dxc5 bxc5 14 'ilVg4 'i'a5 Ketola-Lehtivaara, Finland 1 996; (b) 8 h4?! appears overambitious: 8 . . . tDe7 9 h5 h6 10 .id3 .ia6 11 0-0 .ixd3 12 'i'xd3 tDd7 1 3 a4 l:tc8 14 iLa3 0-0 and Black is better, Riemsdijk Matlak, Moscow 1 994; (c) A more popular and disturbing
222
attack on Black's strategy is 8 .i.b5+!? .id7 9 .id3 (9 a4 .ixb5 transposes into 9 i.b5+), when 9 . . . tDc6 was discussed in the second edition, but here we concentrate upon 9 . . . iLa4, as recommended by Watson. It has been much discussed since. Black will sometimes play . . . c4 followed by . . . 0-0-0, with the idea of attacking the centre via . . .f6. Of course the problem with this strategy is that regardless of the king position White can take command on the kingside, as Black's pieces can be bound up on the queenside. So Black often pursues the alternative strategy of playing . . . cxd4 for a direct attack down the c-file.
(c1) 10 h4? ! tDe7 1 1 h5 h6 12 lih4 tDd7 13 dxc5 b5! 14 'iVe2 a6 15 l:tg4 �f8 1 6 l:tb 1 ':'c8+ Wedberg-Djurhuus, Munkebo 1 998 due to Black's better structure; (c2) 10 dxc5!? is a new and critical idea from the German GM Roland Schmaltz, avoiding the c4 blockade and keeping the queenside open. Schmaltz-Leer Salvesen, Amsterdam 2000, continued 10 . . . bxc5 (10 . . . 'i'xc5?! 1 1 0-0 'i'xc3 12 l:tb 1 tDc6 13 'ilVe2 'i'c5 14 .i.e3 with too much initiative for the pawn, Trescher-Lahlum, Kecske met 2003; but L.E.Johannessen's 10 . . . b5 (!) planning . . . tDd7 and . . . tDxc5
Winawer Variation: Black Plays 6. . . 'iWc 7 may be best - compare Wedberg Djurhuus above) 1 1 c4 liJe7? ! (better is 1 1 . . . d4 12 0-0 liJc6 or 1 1 . . . dxc4 12 i.xc4 liJbc6) 12 0-0 (12 cxd5 might be more accurate) 12 . . . liJbc6 13 cxd5 liJxd5 14 :tel Ii.d8 15 'iYe2 liJd4 16 liJxd4 cxd4 17 'iWe4 'iVc5 18 c4 liJe7 19 'iVg4 0-0 with some initiative for White; (c3) 10 liJg5!? (suggested by Rich ard Beale and Watson) is worth ex amining; White sacrifices his centre and development, but may get good attacking chances; (c4) 10 0-0 and: (c41) Here Watson recommended 10 . . . c4!? 1 1 �e2 liJe7 with the point of 12 liJh4 liJg6 ! ; then 1 3 i.g5 h6 14 liJxg6 fxg6 1 5 i.h4 liJc6 16 f4 0-0 1 7 g 4 liJ e 7 18 'iWd2 l:i.f7 1 9 Ii. f2 Ii.afB 20 �f1 'iWd8 with the idea of . . . g5 and . . . liJg6! was = equal in the game Djurhuus-L.E.Johannessen, Oslo 1999. White also has 12 liJe1 (12 g3 h6 13 liJh4 liJbc6 14 f4 g6 1 5 g4 h5 1 6 gxh5 0-0-0= A. Horvath-L. E.Johannessen, Oropesa del Mar 1 998; Watson also gave 12 liJg5 h6 13 liJh3 liJbc6 14 f4 0-0-0=) 12 . . . liJd7 (12 . . .f6 1 3 �h5+ liJg6 - Watson, but 1 4 exf6 gxf6 1 5 f4! is critical) and Black retains the possibility of castling on either wing; (c42) 10 . . . liJd7 1 1 liJg5 ( 1 1 Ii.e1 liJe7 planning . . . h6 and . . . 0-0-0 - Watson) 1 1 . . .c4 (1 1 . . .h6 - Watson) and 12 .i.e2 h6 or 12 illi 5 liJfB 1 3 i.e2 i.xc2; (c43) 1O . . . liJe7 11 h4 (1 1 �e1 h6 12 h4 liJd7 1 3 .l::1.b 1 a6 soon gave Black an advantage in Walek-Matlak, Pardu bice 1 996, while 12 l:i.b 1 liJd7 1 3 �e3 a6 1 4 liJh4 b5 15 l:i.b2 O-O-O!? 1 6 dxc5 liJc6 was Bruzon - L.E.Johannessen, Linares 2002) 1 1 . . .h6 12 h5 liJd7 13 liJh4 with the idea of f4-f5 . This is Timman's recent aggressive plan, but it seems that Black can successfully
control the f5-square and get coun terplay: 13 . . . 0-0-0 14 f4 c4 15 �e2 g6 16 hxg5 fxg6 17 .i.g4 liJf8 18 �f3 liJf5= Rudolf-Enders, Schwerin 1997. 8 .i.a6 9 iLxa6 Misplacing Black's knight at a6 exchanging bishops is since unavoidable anyway. 9 i..b 5+?! is less threatening after 9 . . . .i.xb5 10 axb5 a5! (Moles' 10 . . . liJe7 1 1 0-0 a5 12 dxc5 bxc5 13 c4 liJd7! is fully satisfactory) 1 1 c4!? (1 1 bxa6 gives White nothing here, e.g., 1 1 . . .Ii.xa6 12 l:i.xa6 liJxa6 1 3 0 - 0 liJ e 7 14 'iWe2 liJb8 1 5 iL a 3 0-0, and 16 'iVb5?! liJd7 17 Ii.b 1 l:i.a8:j: Elseth Lahlum, Porsgrunn 2003, or 16 c4!? dxc4 17 'ilVxc4 Ii.c8 18 dxc5 bxc5 19 l:i.b 1 liJd7= R.Smith-A.Dowden, New Zealand 2001) 1 1 . . . dxc4 12 liJg5 (12 'ilVe2 liJe7 13 'ilVxc4 liJd7 14 0-0 l:i.c8=) 12 . . . 'iYb7! (Epelman's move) 1 3 f3 liJe7 1 4 liJe4 0-0 1 5 iLa3 liJd7 16 liJd6, and now Watson's 1 7 . . . 'ilVd5 !? seems + . 9 liJxa6 ...
...
10 �e2 (a) 10 0-0 liJe7 1 1 liJg5 (1 1 'iVe2 liJb8 12 c4 dxc4 13 dxc5 0-0 14 'iVxc4 liJd7 15 �a3 Ii.fc8 16 .l:!.fe1 liJxc5= Tim mermans-Jetzl, Graz 1997) 1 1 . . .h6 12 liJh3 Ree-Bouwmeester, Amsterdam 1 968; 12 . . . cxd4 13 cxd4 l:i.c8:j: Watson; (b) 10 c4 liJe7 1 1 cxd5 liJxd5 12 0-0
223
Play the French cxd4 13 lDxd4 0-0 (13 . . . 'iVxe5? 14 .l::1. e l lDc 3 1 5 'ifd2) 14 llel lDab4 and Black's active knights gave him at least equality in the game Golubev Haba, Chemnitz 1 998; (c) 10 Vid3 (White's queen is usually worse placed here than on e2 as Black can get in . . . c4 or . . . lDxe5 with tempo in some lines) 1O . . . lDb8 (10 . . . c4!?) 1 1 0-0 lDd7 (1 1 . . .lDe7 12 a5 bxa5 13 .ta3 lDd7 14 dxc5 lDc6 15 c4 lDdxe5! 16 lDxe5 lDxe5 17 'iVg3 0-0 1 8 .i.b2 f6 1 9 i.c3 'i!Vxc5 20 cxd5 l:tac8 2 1 .i.xe5 fxe5 2 2 dxe6 li1.fe8 23 Via3 'i!Vxa3 24 lha3 l:tc5 Y2-Y2 Bareiss-De Waard, corr 1998; Moles also gives 17 'i!Vd4?! lDxc4 18 'iVxg7 0-0-0) 12 a5 (12 i.a3 lDe7 1 3 lDd2 0-0 1 4 f4 .l:.fe8 15 lDf3 h6 16 1:tfc1 I:i.ac8=i= due to the c-file, Zoughani-L.E.Johannessen, Patras 1 999) 12 . . . lIc8!? 1 3 axb6 axb6 14 dxc5 bxc5 15 lIel lDe7 16 h4 Suetin Donner, Havana 1 968, when Moles' 16 . . . h6! looks =i= as White has no kingside attack; (d) 10 a5!? is more critical, the obvious answer being 1O . . . bxa5 with play similar to the main line, e.g., 1 1 'i!Vd3 lDb8 12 'i!ib5+ lDd7 1 3 i.a3 lDe7 14 i.xc5 'iVc6 15 'iVxc6 lDxc6 16 .td6 f6= A.Horvath-E.Berg, Oropesa del Mar 1 999. Others are 10 . . :iYh7!? in tending 1 1 axb6 axb6 12 'i!Ve2 lDe7 1 3 'i!ib5+ lDc6 - Watson; and the more ambitious 10 . . . cxd4!? 1 1 cxd4 ( 1 1 axb6 'fixc3+ 12 i.d2 'iVc4 intending 1 3 b7 .l::[b 8 14 'i!Ve2 .l::1. xb7!) 1 1 . . .'iVc4 12 axb6 axb6, at least = . 10 lDb8 Preparing to reactivate the knight to a more active square at c6 or d7. 10 . . . c4?! is less flexible but playable. 1 1 a5!? The only critical try; Black is fme if given time to consolidate with 1 1 0-0 lDe7 and: ...
224
(a) 12 c4 dxc4 1 3 dxc5 bxc5 14 'i!Vxc4 lDd7 1 5 .l::[ e l 0-0 16 i.g5 lDg6 17 'i!Ve4 lIab8 was complex but probably about balanced in Bakre-E.Berg, Athens 200 1 ; (b) 12 dxc5 bxc5 1 3 c 4 d4!? 14 lDg5 lDbc6 15 'ilVh5 lDg6 with mutual chan ces, Bobras-B. Socko, Warsaw 2002; (c) 12 i.a3 lDd7 1 3 c4 0-0 14 .l:.fe l .l::[fe8 1 5 dxc5?! lDxc5=i= E l Faher Djurhuus, Cairo 1 999; (d) 12 a5 bxa5 1 3 i.a3 lDd7 14 dxc5!? a6 1 5 c4 lDxc5 16 cxd5 lDxd5 17 lDg5 0-0 18 i.xc5 'i!Vxc5= Arutun ian-E.Berg, Pardubice 2002. Compare 12 0-0 lDe7 in the main line below. 1 1 bxa5 Watson's 1 1 . . .lDd7 12 axb6 lDxb6 planning . . . lDc4 Iooks = . But 1 1 . . .bxa5 is sharper and as far as I can see works well for Black. ...
12 �a3 Consequent. 12 0-0 lDe7 13 i.a3 lDd7 14 c4 (14 dxc5!? a6! 1 5 c4 lDxc5 16 cxd5 lDxd5 17 lDg5 0-0 18 i.xc5 'iVxc5 19 'i!Vd3 lDf6 20 exf6 Y2 - Y2 was the game Sarbok-E .Berg, Gausdal 2003) 14 . . . 0-0 1 5 cxd5 lDxd5= Ivkov Donner, Bamberg 1968. 12 lDd7!? 12 . . . lDe7!? 1 3 i.xc5 lDd7 14 'iWb5 .l::[b 8 1 5 'iWa4 'i!Vc6 16 'i!Vxa5 (16 Vixc6 lDxc6 17 .i.d6= Carlson-Davis, corr ...
Winawer Variation: Black Plays 6 . �c 7 .
.
1 976) 16 . . . ttJxc5 'unclear' - Moles; this is solid but less ambitious. 13 dxc5 ttJe7
14 c6! Otherwise 14 . . . ttJec6! , when Black is in control. White needs to open the a3-fS diagonal to get an attack. 14_.:iWxc6 15 0-0 'ilYxc3 1 6 11fdl 1 6 il.d6?! ttJf5 1 7 "i!fb5 ttJxd6 IS exd6 'ii'c 5 1 9 'ii'xc5 ttJxc5 20 l1xa5 ttJe4 2 1 lHa1 ttJxd6 22 l:i.xa7 l1xa7 23 1:txa7 0-0+ Benassi-Ooorebeek, corr 1997. 1 6 ...ttJc6 1 7 .i.d6 Otherwise Black can close off the diagonal a3-fS with 1 7 . . . ttJb4 and then castle. 1 7 ... 'ilYc4 1 8 'ii'e 3 'iVe4 lS . . . h6!? to prevent 19 "iYg5 is an untested move which might work. 19 ft3 1 9 "iYc3 'iVc4 20 'ire3 'i!Ve4 2 1 'i!Vc3 "iYc4= was von Bahr-Lahlum, Gothen burg 2002, when White played on with 22 'i!Vd2?! f6!? (22 . . . h6!) 23 l:i.a3!? ttJcxe5 24 ttJxe5 ttJxe5 25 'i!Vxa5 �c6 26 SLxe5 fxe5 27 l1c3 �d6 28 "i!fb5+ 'i'd7 29 'iVc5 l:1.f8:j: . 19 .. _ttJb6 20 c4! 'i!Vxc4 2 1 'i!Va3 21 'ilVb2?! ':'cS 22 'i!Vd2 h6 23 lldc 1 "iYa6 24 'iVf4 ttJ e7 25 l:i.xc8+ ttJbxc8 2 6 'i'g4 0 -0 27 l:t c1? ttJxd6 0 - 1 Scheske N. Stewart, Wittlich 1997. 21 .. .vWa6
22 ':'acl Thus far the famous game Tal Donner, Wijk aan Zee 1 968, which for decades frightened Black players away from this variation. White stor med through in classical Tal fashion after 22 . . . .l:tcS?! 23 ttJd2 ! (planning ttJb3-c5) 23 .. .f6? (Moles gives 23 . . . ttJd4! 24 l:i.xc8+ 'i!Vxc8 25 �xa5 'itd7, and 25 . . . 'i!Vc2 !? as suggested by Evans looks even better. Moles also gives 24 �hl 'it>d7, when 25 St.c5 !? ttJe2 26 l1b 1 might be best, but White's compensation looks airy. But this is just an extra possibility, as Black can improve earlier on.) 24 exf6 gxf6 25 'i!Vf3 �d7 26 'ii'xf6 .l:the8 27 ttJe4! . However, Black keeps his pawns without losing his king after 22 . . . ttJc4! 23 �a4 l:i.c8 - Shamkovich, who gives 24 ttJd4? ttJb2! 25 l:txc6 l:i.xc6-+ . Instead 24 ':'c2 f6! 25 l1dc 1 'it> f7 was tested in two correspondence games of the Nordic Cup 1997, Pelle Larsen proved it insufficent for White after 26 h4? �hd8 27 l:.c3 �g8 0 - 1 Kontuilainen-P.Larsen and 2 6 exf6 gxf6 27 i.. g 3 .u.he8 28 .u.xc4? (desper ation, as Black was about to evacuate his king) 28 . . . dxc4 29 .u.xc4 ttJe7 30 �c2 �xc4 3 1 ttJe5+ fxe5 32 "iYxh7+ 'it>fS 0 - 1 Szalai-P.Larsen. 24 ttJd2!? might be best, but still Black held the
225
Play the French advantage all the way through Seydoux-Dieu, corr 1998: 24 . . . ttJxd6 25 exd6 'iVb7 26 ttJb3 0-0 27 ttJc5 'i\Va8 28 d 7 flcd8 29 ttJxe6 fxe6 30 Itxc6 'ilVb7 31 �xe6 'iYxd7 32 'i\Vxd7 Itxd7 33 Ita6 d4 34 1ha5 d3 � - K Many sources give the Tal-Donner game as the way to play for White against 6 . . . 'iYc7 7 ttJf3 b6, but without quoting Moles' analysis or testing whether the pawn sacrifices were correct. It seems they are not. Although the last word almost certainly hasn't been written about these sharp variations, correct play from both sides seems to result in a
226
black advantage. White undoubtedly gets some attacking chances, but 2 or 3 pawns is a lot to invest and the queenside is a risky area for White to operate in since Black has natural advantages on that side of the board. Hence after 7 ttJf3 b6 White will have to settle for one of the equalizing lines or try to explore some new territory, probably by risking a double-edged dxc5!? at some stage. I have devoted a great deal of time to study of this particularly demand ing variation, and my conclusion is that Black is fine from a theoretical point of view.
Chapter Thirteen Classical Variation : 4 e5
1 e 4 e 6 2 d4 d 5 3 liJc3 liJf6 This is the Classical Variation of the French Defence. It is similar in pawn structure to the Advance Varia tion as well as those lines of the Tar rasch variations that include the moves . . .liJf6 and e5. 4 e5 In the next chapter we consider 4 �g5. 4 .td3 is seldom played due to 4 . . . c5 and: (a) 5 exd5 cxd4 (5 . . . exd5=) 6 �b5+ �d7 7 �xd7+ (7 �xd4 �xb5 8 liJxb5 liJxd5 9 liJe2 liJc6 10 �a4 �c5!:j:; 1 0 . . . a6 1 1 liJbd4 liJb6 12 liJxc6 liJxa4 13 liJxd8 lIxd8= - Steinitz) 7 .. :ihd7 (7 . . . liJbxd7!? 8 dxe6 dxc3 9 exd7+ �xdn) 8 dxe6 'i'xe6+ (8 .. .fxe6!? 9 liJce2 liJc6 10 liJf3 .i.c5:j:) 9 liJce2 liJc6 10 liJf3 .tb4+ 1 1 .i.d2 0-0-0= Lasker Marshall, New York (2) 1907; (b) 5 liJf3 cxd4 (5 ... liJc6=) 6 liJxd4 e5 (6 . . . .tb4!?; 6 . . . liJc6 7 �b5 .i.d7 8 exd5 exd5 9 0-0 �e7 10 �e3 0-0= Lasker Bogoljubow, Zurich 1934) 7 liJf3 d4 8 liJe2 �g4, just a shade better for Black after 9 0-0 (9 liJxe5?? 1li'a5+) 9 . . . liJc6. 4 ...liJfd7 Now White has:
1 3 . 1 5 liJf3 13.2 5 liJce2 13.3 5 f4
The only other interesting move is 5 'i'g4!? c5 (5 . . . b6!?) and the centre is breaking up so White must try to get liJb5-d6 in, or perhaps dxc5 at the right moment: 6 liJf3 (6 liJb5 cxd4 7 �f4 1li'a5+ 8 �d2 �6:j:) 6 . . . cxd4 7 liJb5 (7 1li'xd4 liJc6 8 �f4 a6! intending . . . 1li'c7) 7 . . . liJc6 8 liJd6+ �xd6 9 1li'xg7 �xe5 10 liJxe5 'iWf6 1 1 'iWxf6 liJxf6 12 liJxc6?! (12 .i.f4! and 12 . . .'�e7 1 3 �g3 liJe4= or 12 . . . liJd7 13 .i.b5 liJcxe5 1 4 �xe5 'ug8 1 5 g 3 a 6 16 �xd7+ .i.xd7 17 �xd4 <3;; e 7=) 12 . . . bxc6 13 �f4 c5! 14 �e5 <3;; e 7 1 5 b4 cxb4 16 �xd4 'ug8 17 �c5+ <3;; d 7 18 �xb4 a5 19 �a3 �a6:j: . 1 3 . 1 5 liJf3 Aiming for piece play instead of pawn expansion; it is played fre quently but considered fairly harm less. 5 ... c5 6 liJe2 Now we have reached the position from 5 liJce2 c5 6 liJf3. 6 dxc5 trans poses to 1 e4 e6 2 liJf3 d5 3 liJc3 liJf6 4
22 7
Play the French e5 liJfd7 5 d4 c5 6 dxc5, which is in the Odds and Ends chapter via 2 liJf3 d5 3 liJc3. 6 ... liJ c6 6 . . . cxd4 7 liJexd4 liJc6 is also rea sonable, e.g., 8 liJxc6 bxc6 9 i.. d 3 liJc5 10 .te2 .i.e7 11 0-0 0-0 12 i.. e 3 llb8= Jimenez Chacon-Hamdouchi, Dos Hermanas 1 998. 7 c3 cxd4 7 . . . .te7 is a flexible option: 8 a3! (8 liJg3?! "iVb6 9 .te2 cxd4 10 cxd4 i.b4+ 1 1 'it>fl f6;; 8 liJf4 cxd4 9 cxd4 �6 10 .i.e2? Yudkovsky-Coello, Cannes 1997, and Black should play 10 . . . g5! 11 liJh5 g4 12 liJg5 i..b 4+ 1 3 'iifi> f1 'ili'xd4+) 8 . . . a5 9 liJf4 �6 10 h4 a4= . 8 cxd4 f6
9 liJf4 9 exf6 liJxf6 10 liJc3 i.d6 is com fortable for Black (compare a 3 liJd2 liJf6 Tarrasch; here White is commit ted too early to liJc3) : (a) 1 1 i..b 5 0-0 12 0-0 'WIc7 (or 12 . . . �6!) 1 3 h3 .i.d7 14 l:tel 11ae8 1 5 i.. e 3 a 6 16 i.. f1 Lau-Gleizerov, Dres den 1994, and here 16 . . . e5! was strong; (b) 1 1 i.. e 2 0-0 12 0-0 h6 (12 . . . �6!) 1 3 h3 !? i.d7 14 i.. e 3 Hort-Knaak, Bundesliga 1996; now 14 . . . i.. e 8! 1 5 l:t e l .i.g6 was promising; (c) 1 1 i.. d 3 0-0 12 i.. g5 'i'e8!?
228
(12 . . . �6!) 13 'i'd2? liJh5? (13 . . . e5!;) 14 liJe2? (14 'iVc2; 14 0-0 It:lf4) 14 ... e5 1 5 liJxe5 ( 1 5 dxe5? .i.b4 16 liJc3 l:1.xf3! 1 7 gxf3 d 4 18 f4 dxc3 1 9 bxc3 liJxe5! 20 fxe5 'ili'xe5+ - + McDonald) 1 5 . . . liJxd4 16 liJxd4 'ili'xe5+ 17 i.e3 liJf4+ Nijboer Glek, Wijk aan Zee 1999. 9 ....i.h4+ 10 .i.d2 'ifle 7
1 1 .i.xh4 1 1 exf6?! liJxf6; 12 i.d3 0-0 1 3 0-0 liJe4 14 .te3 .i.d6 1 5 g3 'iflf6 1 6 .i.xe4 dxe4 17 liJd2 .txf4 18 gxf4 i.d7! 19 liJxe4 'i'g6+ 20 liJg3 liJe7! 21 lIc1 i.c6 with light square domination, Tate Shulman, Sioux Falls 200 l . After 1 1 i.. xb4, Shirov-Ivanchuk, Tilburg 1993 continued 1 1 . . . 'i'xb4+ 12 'i'd2 'iifi> e 7! 1 3 exf6+ gxf6 14 'i'xb4+ liJxb4 15 'iifi> d2 liJb6 16 a3 liJc6 17 .i.b5 .i.d7!? (17 . . . liJa5! 18 l:tae l liJac4+ 19 'iitc l 'iitf7= with the idea . . . a6) 18 .i.xc6 i..xc6 19 llhe l i.. d 7 20 b3 'it>d6 2 1 11e3 and White had a small advantage. As a whole, however, 5 liJf3 is rather easy to play against. 13.2 5 liJce2 This is considered the optimal or der to get to the main f4 lines, but that is not so clear. Before we get to how Black can throw a wrench (spanner!) in the works, let's take a close look at 5 f4. That is the move
Classical Variation: 4 e5 used in the next section (13.3) , but I'm going discuss it here fIrst for early deviations in the context of tLlce2 sys tems. The move order issues exam ined here and in the next few notes are of great importance and will re pay study: 5 f4 c5 6 tLlf3 (6 dxc5 and other non-transpositional moves not leading to tLlce2 systems are given in 1 3 . 3) 6 . . . tLlc6 7 tLle2
This can also arise after 5 tLlce2 c5 6 f4 tLlc6 7 tLlf3, as well as from 3 tLld2 tLlf6 4 e5 tLlfd7 5 f4 c5 6 tLldf3 tLlc6 7 tLle2. Although it is appearing regu larly, books tend to either treat lightly or even ignore this position. Black has several ways to respond. 7 .. :iVb6 returns the game to normal channels after 8 c3. And 7 . . . b5!? has been played quite a bit recently, al though it seems more logical once White has played g3. A problem for Black, not necessarily insoluble, is 8 a3 l:.b8 9 c3 a5 (9 . . . b4 10 dxc5! bxc3 1 1 b4 a5 12 tLled4) 10 dxc5 tLlxc5 1 1 ..te3!? with the idea 1 1 . . .tLle4? 1 2 tLlg3±. But it's not so easy for White. Two other nontranspositional moves are playable, with one that I fInd particu larly attractive: (a) 7 . . . cxd4; McDonald thinks that this is dubious, yet it has a long his-
tory. Perhaps White is a little better, but Black has many possibilities, for example, 8 tLlexd4 ..tc5 (or 8 . . . tLlxd4 9 tLlxd4 'iVb6 10 c3 ..tc5 1 1 i.e2 0-0) 9 c3 'iVb6 10 b4!? (10 ..te2 O-O!? 1 1 1i.d3! g6 12 i.c2 tLlxd4 13 cxd4 i.b4+ 14 Wf2 f6;!;) 10 . . . tLlxd4 1 1 tLlxd4 i.xd4 12 �xd4 'iVxd4 13 cxd4 tLlb6 14 1i.d3 (14 a4 i.d7 1 5 a5 tLlc4=) 14 ... ..td7 1 5 ..td2 l:1.c8 (15 . . . a6 16 0-0 ..tb5!?) 16 �e2 a6 17 .l:.hcl 0-0 18 g4 tLla4 with the idea . . . ..tb5, was Niedermaier-B. Schneider, Bundesliga 1 986. Neither side played perfectly, of course, but at least we see some ideas; (b) 7 . . . ..te7!? is an interesting wait ing move, not committing to . . . �6, ... cxd4, or ... b5. I quite like it.
White has to show his cards: (b 1) 8 ..te3 �6 forces White to lose a crucial tempo defending his b-pawn, which isn't so easy - compare main lines; (b2) 8 g3 �6 gives White the op tion of entering a main line by means of 9 c3 (also from the 3 tLld2 tLlf6 Tar rasch Variation), but with g3 already in, which isn't always his plan in the 5 tLlce2 lines. An independent and logical plan for Black would be 8 . . . 0-0 9 ..tg2 b6!? intending 10 0-0 1i.a6. In fact, even the immediate 8 . . . b6 is worth thinking about; 229
Play the French (b3) S c3 is the key move. It allows Black to transpose into a main line by S . . . 'i!fb6 or he can go his own way by S . . . O-O, e.g., 9 a3 (9 g3 allows another known position after 9 . . . 'iWb6, but Black can also play a standard anti-f4 plan by 9 . . . f5 and queenside expan sion; or choose an aggressive strategy by 9 . . . a5!? 10 i.g2 b5!? having in mind . . . �a6 and . . . b4) 9 . . . a5 10 g3 cxd4 (alternatives are 10 . . . b6!? and 10 . . . a4!? 1 1 �h3) 1 1 cxd4 ( 1 1 liJexd4 liJc5= intends moves like . . . �d7 and . . . a4) 1 1 . . .f6!? (1 1 . . .f5 12 iLg2?! a4 1 3 0 - 0 liJb6 wins key queenside squares) 12 �h3 fxe5 13 dxe5 (13 iLxe6+ 'iith S is messy, but the fIrst point is 14 dxe5?! liJdxe5!) 1 3 . . . 'i!fb6!? (13 . . . liJc5! with the idea . . . a4) 1 4 liJc3? (for better or worse, correct was the risky 14 �xe6+ �hS 1 5 iLxd5) 14 ... liJdxe5!? (14 ... liJc5 ties White down completely, since 15 liJd4 liJxd4 16 Wi'xd4 'iVa6 threatens . . . liJb3 and . . . liJd3+) 15 fxe5 liJxe5 16 i.g2 iLd7 1 7 lifl liJc4 with a strong attack for the piece. Bologan M. Gurevich, Cap d'Agde 2002 went IS Ilb l �f6 19 Wi'd3 l:!acS 20 liJg5 iLxg5 21 iLxg5 .l:.xfl + 22 1Lxf1 lifS 23 Wi'e2 �f5! 24 iLf4 g5 25 liJxd5 exd5 26 i.bS lifS 0- 1 .
5 . . .c 5 6 c3 6 liJf3 transposes into 5 liJf3 c5 6
230
liJe2 above. 6 f4! may well be the best move order, because of the issues raised in the note to 6 . . . liJc6 below.
Now 6 . . . liJc6 7 liJf3 is the last note, but 7 c3! will transpose to our main line without having to deal with the 7 . . . f6 variation below. And 6 . . . cxd4 7 liJxd4 liJc6 S liJgf3 also transposes to the last note, but not to Black's best line. Nevertheless, I think that Black has other good opportunities to devi ate at this point. Here are a some ideas: (a) 6 . . :vib6 7 liJf3 (7 c3 iLe7 S liJf3 0-0 transposes) 7 . . . iLe7 is a sort of waiting game: S c3 0-0 with the idea 9 dxc5 liJxc5 10 liJed4 liJc6 1 1 1Le2 liJe4= ; (b) 6 . . . iLe7 is a similar idea, when White might try 7 liJf3 0-0 S c3 (S �e3 'i!fb6), leading to S . . .f6!? (or S . . . 'i!fb6) 9 �e3 fxe5 10 fxe5 'i!fb6 1 1 'iWd2 liJc6 with mixed prospects, e.g., 12 0-0-0 (12 liJf4? cxd4 1 3 liJxd4 l1xf4! 14 iLxf4 liJxd4 15 'iVxd4 'iVxb2 16 lidl 'iVxa2+) 12 . . . 'iVa5 13 'iitb l cxd4 14 cxd4 liJb6! 1 5 'iVxa5 liJxa5 16 b3 iLd7= intending . . . liJc6 and . . . a5-a4; (c) 6 ... 'iVa5+!? 7 c3 (7 iLd2 'i!fb6 S i.c3 liJc6 9 liJf3 cxd4!? 10 liJexd4 i.b4 is promising) 7 . . . b5!? (risky but per haps okay; 7 . . . liJc6 is of course an op tion) S liJf3 (S dxc5 b4 9 a3 i.xc5 10
Classical Variation: 4 e5 ii.d2 'iWb6=) 8 . . . b4 9 ii.d2 ttJc6 10 cxb4 cxb4 with the idea 1 1 a3!? i.e7! 12 axb4 ttJxb4 etc. The interesting thing is that if any of these moves is satis factory for Black, White has no order that gets him to the main lines by force without legitimate options for Black. 6 ...ttJ c6 Assuming that Black has achieved this exact position after all the move order jockeying implicit in the last two notes, he actually has a signifi cant move that seems to equalise at this point: 6 . . . cxd4 7 cxd4 f6 ! :
10 lIc 1 0-0 11 ii.xb4 'iVxb4+ 12 'iVd2 'iVe7 unclear - Atalik; 10 "iVh5+ g6 1 1 ttJxg6 i.xd2+ 1 2 Wxd2 'iYxb2+ 1 3 �e3 'iVxa1 14 ttJxh8+ <;f;?d8 and Black is at least okay) 10 . . . 'iVxb4+ 1 1 'ii d2 'iixd2+ 12 �xd2 �e7
13 exf6+ (the most commonly played move; 13 ttJf3 fxe5 14 ttJxe5 ttJxe5 1 5 dxe5 ttJc6 16 ttJd3 i.d7 [or 16 . . . b6 1 7 :c1 ii.b7=] 1 7 f4 i.e8 18 l::t c 1 l::tfB 1 9 g3 is ;1; according to the ory, but then 19 . . . ii.g6!, e.g. , 20 ttJc5 l::tfc8!? 21 ttJxb 7 [2 1 ii.b5 ttJb4!] 2 1 . . .ttJb4! and White loses the c-file) 13 . . . gxf6 14 l::t e 1 (14 ttJf3?! Wd6; 14 ttJge2 ttJc6 1 5 ttJd3 e5) 14 . . . ttJb6 and now two important moves: (c 1) 15 i.d3?! allows Black to force things in the centre by 15 . . . ttJc6 16 ttJf3 <;f;?d6! 1 7 ttJh5 l::tf8! (17 ... e5 18 ttJxf6 e 4 1 9 ttJg5 �c7 is unclear) 18 i.xh7 (18 l::t e2 e5 1 9 ttJxf6 e4 20 ttJxe4+ dxe4 2 1 i.. xe4 i.. g 4=i=) 18 . . . e5 (18 . . . .l:th8!? is messy) 19 dxe5+ ( 19 ttJg3 e 4 20 ttJh4 ttJxd4 2 1 f3 :h8! 22 ttJhf5+ ii.xf5 23 ttJxf5+ �e5 and Black's king was a monster in Bez godov-Sakaev, Russia Ch 1999) 19 . . .fxe5 20 Wc1 (20 l::thf1 �g4+ ; 20 ttJg3 i.. g4!+) 20 . . . i.g4 21 ttJg3 ii.xf3 22 gxf3 ttJd4- + Arakhamia Grant-Gleiz erov, Port Erin 200 1 ; (c2) 1 5 ttJf3 <;f;?d6 (15 . . . ttJc6 16 .i.b5 '
(a) Black's development i s fast af ter 8 exf6 ttJxf6 9 ttJf3 .i.d6 (9 . . . ttJc6 10 ttJc3 transposes to 13. 1 , the note to 9 ttJf4, which was fme for Black) 10 g3 ttJc6 1 1 .i.g2 0-0 12 0-0 'iWb6 13 I1b 1 ii.d7 14 ii.f4 ttJe4 with good activity in Bologan-Glek, Bundesliga 1992; (b) 8 f4 fxe5 9 dxe5 (9 fxe5 ii.b4+ 10 Wf2 0-0+ 11 ttJf3 ttJc6 with initiative and ideas of sacrificing on e5) 9 . . . ttJc6 10 ttJf3 .i.b4+ 1 1 ttJc3 (1 1 i.d2 ttJc5) 1 1 . . .ttJc5 12 i.e3 Shirov-Ivanchuk, Tilburg 1993, and a straightforward path was 12 . . . ttJe4 13 'iVc2 (13 :c1 1i'a5 14 ii.d2 0-0) 1 3 .. .'iVa5 14 l::tc 1 'iYxa2+ ; (c) 8 ttJf4 ii.b4+ 9 ii.d2 "iib 6 10 i.xb4 (10 exf6 ttJxf6 11 'iVa4+ ttJc6=i= ;
'
231
Play the French i.. d 7 17 �xc6 bxc6 18 l:te2 favoured White in Anand-Bareev, Shenyang 2000) and here White must fmd something better than transposing into the previous note by 16 liJh5 (16 i..b 5 a6 1 7 i.. d 3 liJc6 is no improve ment) 16 .. JH8 17 i.. d 3 liJc6. 7 f4 'ilkb6 For 7 . . . i.e7, compare 6 f4 i.e7, al though delaying . . . liJc6 there allowed a queen on b6 to cover e6 for the .. .f6 break. 8 liJf3 f6
Finally the main line position of the 5 liJce2 variation has been reached. Black wants to blast open the centre, sacrificially if necessary, whereas White knows that if he can hold on to his space advantage with out compromising his pawns or creat ing weaknesses, he must stand bet ter. A particularly fascinating thing about this position, which has been around for years, is that White has begun to neglect his development en tirely in order to expand on the flanks. Thus one will see a3 and b4 as well as h4-h5 in order to clamp down on both wings. In the end I believe that Black, with superior develop ment, has sufficient resources to achieve dynamic equality. In any case, the play will seldom become 232
drawish! At this point there is a vi tally important fork in the road: 13.21 9 g3 13.22 9 a3
(a) 9 'iVb3 is normally answered by 9 . . . cxd4 10 "iVxb6 liJxb6=; a fun alter native is 9 . . . 'i'c7 (what's the queen doing on b3?) : 10 g3 cxd4 (1O . . . fxe5 1 1 fxe5 cxd4 1 2 cxd4 .i.b4+ 1 3 .i.d2 i.xd2+ 14 liJxd2 0-0) 1 1 cxd4 �b4+ 12 'ittf2 0-0 13 'itt g2 i.. a 5 (13 . . .fxe5!? 14 fxe5 i.e7 1 5 iLe3 liJb6:j: 16 liJf4 liJa5 1 7 1!i'dl liJac4) 14 i.e3 i.b6 15 lIcl 1!kb8 1 6 exf6 liJxf6 17 liJe5 i.d7 18 liJxd7 liJxd7= B.Lalic-Dizdar, Porec 1998; (b) 9 liJg3? is better after the move a3: 9 . . . cxd4 10 cxd4 i.b4+ 1 1 'it>e2 (this is the only move) 1 1 . . .g5! 12 exf6 g4 1 3 n+ �e7 14 liJd2 liJxd4+ 1 5 'it>el 1!i'c7; (c) 9 exf6 liJxf6 10 g3 i.d6 11 i.g2 0-0 1 2 0-0 cxd4 1 3 cxd4 e5!? (13 . . . i.d7 14 b3 liJe4:j:) 14 fxe5 liJxe5 15 liJxe5 i.xe5 16 liith l i.. c 7 17 liJc3= Aseev Pesiakov, St Petersburg 1 999; (d) 9 h4 is playable but probably a move too early: 9 . . . cxd4 (9 . . . i.. e 7 10 h5? fxe5 11 fxe5 cxd4 12 cxd4 0-0+) 10 cxd4 fxe5 11 fxe5 .i.b4+ 12 liJc3 0-0 1 3 a3 i.e7 14 i.e2? 1Ixf3 ! 1 5 gxf3 'iVxd4+ Mahdi-Weinzettl, Mureck 200 1 . 13.21 9 g 3 cxd4 and the play splits a last time: 13.2 1 1 10 liJexd4 13.212 10 cxd4 13.2 1 1 10 liJexd4 This has become the key line of late, following some White improve ments.
Classical Variation: 4 e5
10 ... fxe5 The only solution that I see work ing here apart from my main line is the immediate 10 . . . ttJxd4 1 1 cxd4 �b4+ (without . . .fxe5) 12 �d2 (12 Wf2 0-0 13 �e3 fxe5 1 4 fxe5 �e7 15 b3 J:tf7 16 h4 K.Schmidt-Merker, corr 1987; and the same idea of 16 . . .ttJf8 1 7 �d3 iLd 7 and . . . iLe8 or . . . iLb5 works fine) 12 .. .fxe5 13 fxe5 0-0. This sidesteps the Wf2-g2 idea and equal ises effortlessly after 14 iLxb4 'ii'xb4+ 15 'i!fd2 'i!fxd2+ 16 ttJxd2 ttJb8! 17 iLe2 ttJc6 18 ttJf3 iLd7, e.g., 19 0-0 iLe8 20 J:tadl h6 21 h3 iLg6. 11 fxe5 11 ttJxe6 is normally queried al though none of the given solutions is clear: 1 1 . . .ttJf6 (1 1 . . .ttJe7?! 12 ttJxf8 J:txf8 1 3 'i!fe2 ! ; 1 1 . . .e4 12 ttJfd4! ttJf6 1 3 a 4 ! and 13 . . . � d 7 ends i n equality, but more interesting is 13 . . . ttJxd4 14 ttJxd4 iLc5 1 5 �b5+ iLd7 16 �xd7+ Wxd7=) 12 fxe5 (12 ttJxf8 �g4 13 �e2 �xf3 14 iLxf3 J:txf8=) 12 . . . iLxe6! 13 exf6 0-0-0 with good compensation, e.g., 14 'i!fe2 J:te8! 15 iLe3 iLc5 16 �xc5 'i!fxc5 and White has problems: 17 fxg7 J:thg8+ 18 0-0-0 �g4 19 'i!fd3 Strobel-Piasecki, Baden 1993; now 19 . . . 'i!ff2 ! + . 1 1 . . .ttJ c5! A move played by Luther that
gives Black his full share of the play. With other moves Black won't equal Ise: (a) 1 1 . . .ttJdxe5? 12 ttJxe5 ttJxe5 1 3 'i!fh5+; (b) 1 1 . . .ttJxd4 12 cxd4 �b4+ 13 Wf2 0-0 14 Wg2 �e7 15 iLd3 J:tf7 16 h4 ttJf8 17 ttJg5 iLxg5 18 hxg5 g6 1 9 l:th4;!; o r better, a s occurred in two games; (c) 1 1 . . .�c5? 12 ttJxe6 ttJdxe5 1 3 ttJxc5! 'i!fxc5 14 ttJxe5 ttJxe5 1 5 'i!fh5+ ttJg6 1 6 'i!fe2+!± with iLe3 and �e2 Pedersen.
12 iLh3 I don't know of other games, so here's some analysis: (a) 12 ttJxc6 bxc6 13 'i!fc2 (13 ttJd4 g6! 1 4 l:tb l [14 iLe2 iLg7] 14 . . . iLg7 1 5 �f4 0-0 1 6 �e2 ttJe4 1 7 0 - 0 c5 18 ttJf3 �bTi=) 1 3 . . . iLe7! 14 iLe3 0-0 1 5 iLe2 �a6 16 b4 ttJd3+ 17 Wd2 c5! ; (b) 12 �b5 �e7 1 3 0 -0 0 -0 14 iLg5 h6 15 �xe7 ttJxe7 16 a4 a6 17 a5 'i!fc7=. 1 2 ...iL e7! 12 ... 'i!fa6!? 13 �f1 'i!fb6= draws. 13 0-0 0-0 14 'i!fe2 14 Whl ttJxd4 1 5 'i!fxd4 'i!fcTi= . 14 ...ttJxd4 15 ttJ xd4 Or 15 cxd4 ttJe4� . After 15 ttJxd4, Atlas-Luther, Moerbisch 200 1 went 15 . . . l:txf1 + 16 Wxf1 iLd 7 17 Wg2 Wh8
233
Play the French (17 . . . l:If8!? 18 �e3 a5=) 18 �e3 'ikc7 19 �gl .l:tf8 20 l:Iel with equality. 13.212 10 cxd4 This move is so associated with miniature disasters for White that some people are surprised to find that it's theoretically playable, if only equal. I won't spend too much time showing brilliancies by Black. 10 ... .tb4+ I think that this move order IS more flexible and therefore preferable to the exchange of f- for e-pawn first. But often they transpose and there are some nuances: 10 . . . fxe5
(a) 1 1 dxe5? ii.b4+ 12 i.d2 'ike3 ! is winning: 13 ii.g2 (13 'iib 3! 'ike4!) 1 3 . . . 0-0 14 �fl? (14 a3 liJc5! ; 14 'ikb3! 'ikxb3 1 5 axb3 liJc5+) 1 4 ... i.xd2 1 5 'ikxd2 'ikxd2 16 liJxd2 liJdxe5- + ; (b) 1 1 fxe5 .tb4+ 12 liJc3 0 - 0 1 3 �f4 (this is the good defensive move that White doesn't have if Black omits the exchange on e5; compare the main line; not 13 a3? l:Ixf3 !+) 13 . . . �e7 (threatening b2 and also . . . g5 in many cases) 14 'ilVd2? (14 �h3? 'iWxb2 15 ..txe6+ �h8 16 ':c1 �xf4! 17 gxf4 liJxd4-+ - Sisniega; 1 4 liJa4 is equal after 14 . . . 'iWa5+ 15 liJc3 'iVb6 16 liJa4 'ika5+= etc. ; for 14 a3!, compare the mam line; White has some advan234
tage) 14 . . . g5! 15 liJxg5 �xg5 16 �xg5 liJxd4 17 �g2 liJxe5-+ Dolmatov Bareev, Elista 1997. 1 1 liJc3 11 �d2 has lost countless games after 1 1 . . .fxe5 12 fxe5 0-0 13 �g2 liJdxe5! 14 dxe5 liJxe5, e.g., 15 liJed4 (15 liJf4 'iWe3+ 16 'i£tfl ii.xd2-+) 1 5 . . .liJd3+ (15 . . .liJc4 also wins) 1 6 'iit e2 liJxb2 17 'ikb3 'iWa6+ 18 'iitf2 i.xd2 19 'ikxb2 'iWd3-+ was Hamann-Uhlmann, Halle 1963. 1 1 ... 0-0
Now there's no i.f4 defence, so White's task is much harder. 12 a3! (a) 12 exf6 liJxf6 13 �g2 (13 a3= Bondarevski; 13 . . . i.e7!i=) 13 . . . liJe4 14 'iWd3 e5! 1 5 0 -0 (15 fxe5 �g4+) 15 . . . liJxc3 16 bxc3 e4- + Grischuk Mraz, corr 1 986; (b) 12 ii.h3? is often played, but is only good after 12 a3 i.e7. Here it has lost numerous contests in the manner of 12 . . .fxe5 13 i.xe6+ 'iti>h8 14 fxe5 liJdxe5! 1 5 ii.xd5 liJxf3+ 16 ..txf3 liJxd4-+ Chevalier-Lempereur, Paris 1 994. 12 ... ii.e7 Possible is 12 ... ..txc3+ 13 bxc3 fxe5 14 fxe5 'ikc7 intending . . . liJb6 with light square pressure. 1 3 -th3!
Classical Variation: 4 e5 Without i.f4, the threat of 1 3 . . . fxe5 14 fxe5? 11xf3 is difficult to meet: (a) 13 ..td3?! fxe5 14 dxe5 (14 fxe5 liJxd4 1 5 liJxd4 liJxe5 with more than enough compensation, even after the mild 16 l:f.fl ! i.. d 7! 17 i..b 1 i.. c 5=i=) 14 . . . liJc5 1 5 i.. c 2 .l:!.d8=i= ; (b) 1 3 liJa4 'fiIc7 14 i.. d 3 fxe5 (14 . . . a6!?) 1 5 fxe5? :xf3! 16 'fiIxf3 liJxd4 1 7 'fiIdl (or 1 7 'ilVe3 liJxe5!) 1 7 ... ttJxe5 18 0-0 liJxd3 19 �xd3 e5+ ; (c) 1 3 exf6 liJxf6 14 i.. d 3 a6 1 5 i.c2 i.. d 7 16 0-0 i.. e 8!=i= with the idea of . . . i..h 5. 13 ... fxe5 1 3 . . . �h8!? isn't bad, e.g., 14 liJa4 (14 ..txe6? liJdxe5!; 14 0-0) 14 . . . 'fiIa6! 1 5 ..tf1 b5 16 liJc3 fxe5 1 7 i..xb5 'fiIb6 18 ii.xc6 (18 fxe5 l:!.xf3 ! 19 i.xc6 liJxe5! ! 20 i.xa8 liJd3+ 21 \t>e2 'ilVxd4+ Knaak) 18 . . . 'ilVxc6 19 fxe5 ..ta6!, was Sznapik-Knaak, Bratislava 1983; Black has a small edge after 20 ii.f4 .l:!.ab8. 14 i.. x e6+ 14 fxe5? l:i.xf3 15 'i'xf3 liJxd4- + . 14 . . .'�h8 1 5 liJxd5 15 dxe5 liJdxe5. 1 5 :iWd8 16 .txd7 16 liJxe7 'ilxe7 1 7 d5 e4 18 liJg5 liJc5! 19 ..txc8 (19 O-O! i.. xe6 20 dxe6 liJxe6 21 liJxe4 liJxf4= - Yudasin and Ivanov; or 19 . . . liJxe6) 19 . . . liJd3+ 20 rJtfl .l:!.axc8 21 dxc6 'fiIxg5 22 cxb7 .l:!.b8 23 rJtg2 �f6 24 �e2 ':'xb7 unclear, Lukin-Se.Ivanov, USSR 1984. 1 6 ... i.xd7 1 7 dxe5 i.g4 With activity and the light squares, Black is probably a little better de spite the two pawn deficit. Smagin Dimitrov, Prilep 1992 continued 18 0-0 i.c5+ 19 i.e3 i..x e3+ 20 liJxe3 'Wb6 21 'iVd2 ..txf3 22 l:f.xf3 liJd4 23 l:f.ff1 (23 :tafl liJxf3+ 24 .l:!.xf3 unclear - Smagin) 23 . . . liJb3 24 'ilf2 liJxal 25 ':xa 1 l:f.ae8 26 l:f.c1 g5 ! =i= . ••
13.22 9 a3 i.e7
10 h4 The modern move. Alternatives: (a) 10 b4 cxd4 and: (a1) 11 liJexd4?! fxe5! 12 fxe5?! (12 liJxe6! liJf6 planning 1 3 liJxg7+ rJtfS 14 b 5 liJe4!? 1 5 'ilVc2 liJd8 1 6 liJh5 'Ml6+ Pedersen) 12 . . . liJdxe5 13 liJxe5 liJxe5 14 'Ml5+ liJg6 15 i.. d 3 0-0 16 i..xg6 hxg6 17 'ilVxg6 i.. f6 18 i.. e 3 e5 0- 1 F.Meyer-Pedersen, Germany 1 999; (a2) 11 cxd4 0-0 12 .l:!.b 1 a5 (12 . . . liJa5!? 1 3 liJc3 liJc4 is unclear; 12 . . . .l:!.f7!? 13 'ilVd3 liJf8= with the idea . .. i.. d 7, . . . l:f.c8) 13 b5 a4 was Shirov Gurevich, Munich 1993; 14 exf6 liJxf6 1 5 liJc3 liJd8 16 i.d3 liJf7 unclear (Shi rov) ; (b) 10 liJg3 is easily met by 10 . . . 0-0 11 i.d3 fxe5 12 fxe5 (12 dxe5 c4 13 i.. c 2 liJc5=i=) 12 . . . cxd4 1 3 cxd4 g6! =i= and 14 . . . liJxd4 or 14 . . . .l:!.xf3 next, since i.xh7+Mh5 ideas are eliminated. 10 ... 0-0 1 1 :th3! The star move for this variation. 1 1 b 4 cxd4 12 cxd4 a 5 (12 . . . .l:tf7=) 1 3 b5!? (13 bxa5 �xa5+ 1 4 i.. d 2 'Wb6=i=) 13 . . . 'ilxb5!? (13 . . . liJd8=) 14 liJc3 m6 1 5 l:f.b 1 'ild8 16 ..td3 gave White some attack in Bauer-Bareev, Enghien les Bains 2003; 16 . . . h6 17 liJb5 .l:!.a6!? (17 . . .f5) 1 8 i.c2 f5 19 g4 liJb6 20 gxf5 exf5=i=.
235
Play the French 1 l ... a5 A legitimate alternative is 1 l . . . cxd4 12 cxd4 liJa5 13 b4 liJc4 14 liJc3 (14 liJg3 a5 1 5 �d3 f5) 1 4 ... a5 1 5 b 5 .l:!.f7! with the standard plan . . .lIf7, ... liJfS, . . . �d7; I used to play this way in the Tarrasch 3 . . . liJf6 variation with 5 f4. 12 b3 'ii'c 7
1 3 liJegl ! An ingenious retreat: White unde velops his pieces just to hold the cen tre and play prophylactically. A pretty game followed 13 'iVc2?! b6! 1 4 il. d 2 � a6 1 5 a4?! .l:!.ac8 16 f5? fxe5 1 7 fxe6 e 4! 1 8 exd7 'ii'x d7 (winning) 19 liJe5 liJxe5 20 dxe5 'iWxh3! 2 1 gxh3 �xh4+ 22 �d1 J:lxf1 + 23 .tel ':xe 1+ 0- 1 Jamrich-H.Schneider Zinner, Bu dapest 200 l . 1 3 ... a4 Anand recommends 1 3 . . .b6, which he thinks is fine based upon the mis taken combinative try 14 il.d3 (14 i.e3 .ta6 1 5 i.xa6 I!xa6= - Anand) 14 . . . i.a6 15 i.xh7+? �xh7 16 liJg5+ fxg5 17 hxg5+ �g8 18 n5 (18 g6 :f5-+) 18 . . . i.xg5! 19 fxg5 lIf1+ 20 �d2 J:lafS 21 'iih 7+ �f7- + . 14 b 4 fxe5 1 5 fxe5 liJdxe5!? 1 6 dxe5 liJxe5 17 liJxe5 'ii'x e5+ 18 'iVe2 �xh4+!? Generally condemned. Black could also play lS . . . 'ii'c 7 followed by . . . e5
236
(hitting h3) and/or . . . i.f6. This isn't easy to assess because Black has the centre and White's king is exposed. Of course it's only two pawns and if White can develop he's winning. Oddly enough, HiArcs already likes Black. There might follow something like 19 �g5!? (19 liJf3 cxb4 or 19 . . . e5; 19 'ii'h 5 cxb4 20 axb4 g6 2 1 'ii'h6 e5 22 .l:!.g3 .:txf1 +! 23 �xf1 'ii'c4+ 24 liJe2 il.g4+) 1 9 . . .i.xg5 20 hxg5 e5 21 :f3 (2 1 lie3 e4 22 :d1 cxb4 23 cxb4 �e6 24 liJh3 lIadS and . . . 'ii'e 5 with a strong attack) 2 1 . . .i.f5 22 'ii'f2 !? cxb4 23 cxb4 e4 24 iVg3 .l:[acS! . This shows that the assumption that White was better throughout this game should at least be questioned.
19 �dl Not 1 9 �xh4? 'iWg3+. 19 .. :i'f6? After this Black is doomed. There have been various statements (with little analysis) about the ending after 19 . . . 'iVxe2+! 20 i.xe2 i.f2 (actually 20 . . . .tf6 has some good points, e.g., 21 bxc5 :a 5 22 i.e3 i.xc3 23 lib 1 d4 or 23 . . . e5; maybe this is better?) . Posi tions so deep into the opening seldom determine its validity, and Black has had better options on the last few moves, so I won't take too serious a look here. But to me chances seem
Classical Variation: 4 e5 about balanced. Since 2 1 It'lf3? e5 looks bad, White's has 21 l:thl (Kavalek) 2 1 . . .e5 (or 2 1 . . .cxb4) 22 bxc5 l:ta5! ; or Anand's 21 i.. e 3 e5 22 �xf2 llxf2 23 lIg3 and 'White has a comfortable edge in the ending'. This strikes me as unclear, one plausible continuation being 23 . . . b6 24 WeI (24 It'lh3 llf6 25 l:1.g5 l:tf5 26 i..b 5 l:ta 7!) 24 . . . .l:i.f6 (24 .. .1:H4!?) 25 l:te3 .l:.e6 and White may be slightly better but this would have to be proven. After 19 . . .'ilVf6, Anand-Shirov, Te hran 2000 continued 20 It'lf3! (Kavalek analyses the alternative 20 l:tf3 'fie7 21 i.e3 d4 to White's advan tage, but 2 1 . . .b6! is better) 20 . . . 'i'xc3 (there isn't anything better) 2 1 .i.b2 'iWb3+ 22 'iitc l e5 23 l:txh4 (23 It'ld2! .i.xh3 24 It'lxb3±) 23 . . .�f5 24 'iVdl e4 25 'ilVxb3 axb3 26 It'ld2 e3 27 It'lf3 (27 It'lxb3±) 27 . . . l:tae8 28 'iit d l c4 29 i.e2 and White's two extra pieces decide. 13.3 5 f4 c5
6 lt'lf3 Here we look at lines not involving It'lce2. 6 dxc5 It'lc6 7 a3 is an old line: 7 . . . ..txc5 8 'fig4 0-0 9 It'lf3 It'ld4 10 �d3 f5 1 1 'iVh3 It'lxf3+ 12 'iWxf3 �b6 1 3 It'le2 It'lc5; Tarrasch-Spielmann, Nur emberg 1906) 7 'ilVg4 0-0 8 .i.d3 f5 9
'iVh3 i.xgl ! (this is an original idea; 9 . . . lt'lc6=) 10 lIxgl It'lc5 1 1 i.. d2 It'lc6!? 12 It'lb5?! 'ilVb6 1 3 0-0-0 .i.d7 14 It'ld6 It'la4! 15 �b5 It'ld4 16 i.e3 It'le2+! 1 7 .i.xe2 'ilVxb2+ 1 8 'it>d2 'iWb4+ 19 'iit c l It'lc3+ Fischer-Benko, Curacao 1962. 6 ...lt'lc6
6 . . . cxd4 7 It'lxd4 �b4!? is definitely worthy of consideration; I don't know of any examples. 7 .i. e3 For 7 It'le2, see 13.2 5 It'lce2. Others are barely touched upon (if at all) in the books: (a) 7 �d3?! cxd4 8 It'le2 iYb6 9 0-0 f6! 10 exf6 It'lxf6 11 Whl .i.c5 12 a3 a5 1 3 b3 0-0 14 �b2 g6? (14 . . . lt'lg4! 1 5 'ilVe l e 5 ! 16 It'lxe5 It'lcxe5 1 7 fxe5 l:txf1 + 1 8 'ilVxfl It'lxe5+) 1 5 'ii'e l .i.d7? ! (15 . . . lt'lh5;) 16 �4 unclear, Abdel Aziem-Peng Zhaoqin, Cairo 2002; (b) 7 a3? meets the fate of any slow move: 7 . . . cxd4 8 It'lxd4 It'ldxe5! with the idea 9 fxe5 'ii'h 4+; (c) 7 .i.b5 a6!? (7 . . . cxd4 8 It'lxd4 It'lxd4 9 'ii'x d4 a6 10 .i.xd7+ .i.xd7 1 1 0-0 llc8 1 2 .i.e3 .i.c5=) 8 .i.xc6 bxc6 9 0-0 (9 It'le2 cxd4 10 It'lfxd4 c5 1 1 It'lf3 �e7 12 0-0 It'lb6!? 1 3 c3 0-0 14 'ilVc2 a5!; Varga-Lehmann, Hungary 1 999) 9 . . . cxd4 10 It'lxd4 c5 1 1 lt'lf3 i.e7; Me gibow-Cotton, Concord 1995; the bishop pair counts for something; 23 7
Play the French (d) 7 dxc5 i.xc5 8 �d3 a6!? (8 . . . f6! 9 exf6 liJxf6 10 'iVe2 0-0 1 1 liJe5 'i!Vc7:j:) 9 'iWe2 'iWc7 10 i.d2 b5 1 1 a3 l:i.b8= Bole slavsky-Pachman, Saltjobadenl948. 7 cxd4 This exchange is in line with my philosophy for a second system: to give the reader a solid and extremely well-established variation that will always be playable. S liJxd4 �c5 ...
9 'iW d2 Preparing 0-0-0. Nothing else is considered sharp enough: (a) 9 g3 'i!i'b6 10 liJa4 'i!Va5+ 1 1 c3 i.xd4 12 i.xd4 liJxd4 13 'iWxd4 b6= 1 4 'iVb4!? 'iVxb4 1 5 cxb4 f6 1 6 �b5? fxe5 17 fxe5 0-0+ Antal-Hoang Thanh Trang, Budapest 2002; (b) 9 �e2 'i!i'b6!? (usually arrived at by 8 . . . 'iWb6 9 �e2 �c5) 10 liJa4 'iWa5+ 1 1 c3 liJxd4 12 i.. x d4 i.. xd4 13 'i!Vxd4 0-0 14 0-0 b6 15 i.dl �a6!? 16 b4 'iVb5 1 7 �f3 'iWc4= Vera-Borges Mateos, Havana 1987; (c) 9 liJa4? i.. xd4 10 i.xd4 liJdxe5! 11 i.. e 2 (11 i.. x e5 liJxe5 12 fxe5 'iWh4+ 13 g3 'i!Vxa4+) 1 1 . . .liJg6 12 0-0 liJxd4 13 'iWxd4 0-0+ Redmon-Marechal, Glenalmond 1996; (d) 9 i.b5 0-0 10 liJxc6 bxc6 1 1 i.xc5 liJxc5 1 2 �xc6 (or 1 2 'iWd4 'iVb6 13 b4 cxb5 14 bxc5 'i!Va5= Suetin)
238
12 . . . l:f.b8 13 'i!Vd4 'iVa5= Laplaza-Roge mont, corr 1992. 9 0-0 It's an indication of Black's flexibil ity in this line that he has still an other option here in 9 . . . �xd4 10 i.. xd4 liJxd4 1 1 'iWxd4 'iVb6 ...
This plan has been around for years and is still popular right now among the chess elite; in particular, Korchnoi and Bareev have champi oned it. The basic idea is that in the ending, a kingside advance by White tends to be harmless, so that despite White's better bishop Black will have time to develop rapidly on the queen side using the open c-file to level the chances. The interesting endgame it produces tends to lead to a draw with accurate play by Black; but that can be a problem for both sides. Mter all, White will also have limited pros pects against a strong player or one who knows the ideas well. Thus 9 . . . i.. xd4 trades safety for complexity. From the diagram, searching for genuine winning chances, White has been offering a pawn by 12 'iWd2!? 'ii'xb2 (12 . . . liJc5 13 0-0-0 i.. d 7 is also playable) 13 l:f.b l 'i!Va3 14 liJb5 'iVxa2 15 liJd6+ <:J;; e 7 but for the moment Bareev has demonstrated in two top level games that Black's two pawns
Classical Variation: 4 e5 are as important as White's attack. After 9 . . . 0-0, White can choose be tween two strategies: 1 3 . 3 1 10 g3 13.32 10 0-0-0
The alternative 10 i.e2 gives Black the choice of playing for a simplified game by 1O . . . tl)dxe5!? 1 1 tl)xc6 (11 fxe5 tl)xd4 with the idea 12 i.xd4 'iWh4+) 1 1 . . . i.xe3 12 tl)xd8 1i.xd2+ 13 <Ji>xd2 tl)c4+! 14 i.xc4 dxc4 1 5 tl)xf7 ftxf7 16 'it>e3 b6 1 7 .l:lhd1 i.a6!=, in tending . . . e5 and on fxe5, . . . lIe8. Or he can play for more with the 'normal' 10 . . . a6 1 1 0-0 tl)xd4 12 i.xd4 'iYb6 13 i.xc5 'iWxc5+ 14 Wh1 b5 1 5 i.d3 tl)b6 16 b3 i.b7= Rayner-C.Daly, Debrecen 1 992, when 17 tl)e2?! was answered by 1 7 . . . tl)c4!:t:. 1 3 . 3 1 10 g3 This is a semi-waiting move in the sense that White wants to see what Black is doing before committing his king's bishop or king. White's bishop may go to g2 or h3, a drawback being the loss of control over c4.
10 ... a6 The direct plan with 10 . . . i.xd4 1 1 i.xd4 tl)xd4 12 'ii'x d4 tl)b8! intending . . . tl)c6 is also satisfactory. Black can
play along the c-file after . . . i.d7 and .. JIc8 and may also think about . . . f6, for example, 13 0-0-0 tl)c6 14 'iWf2 i.d7 15 'it>b 1 'ilVa5 16 1i.d3 1Uc8 17 'ilVe1 (17 a3!? b 5 !?) 1 7 . . . tl)b4 1 8 a3 tl)xd3 19 ftxd3 llc4 20 'ilVd2 l:.ac8 21 tl)e2 'ilVxd2 22 llxd2 l:te4 23 tl)c3 Anand-Shirov, Leon 200 1 ; and although Black shouldn't have serious problems with other moves, it seems to me that 23 . . . fte3 ! 24 l:thd1 (24 tl)e2 f6) 24 . . . .l:f.c4 is a good set-up. Black may play for . . .f6 after, say, . . . 'it>f8 and/or . . . i.c6.
11 tl)ce2 White logically reinforces d4. Nev ertheless, it takes pressure off of d5 and e4. Options are: (a) 1 1 .tg2 tl)xd4 (1 1 . . .tl)a5!? 12 b3 b5=) 12 i.xd4 b5 1 3 tl)e2 'iWc7 14 i.xc5 tl)xc5 15 tl)d4 i.b7 16 0-0 tl)e4= Kuc zynski-Gunnarsson, Ohrid 200 1 ; when Black gets this move in he's normally okay; (b) 1 1 i.h3 tl)xd4 12 i.xd4 �c7 (both 12 . . . b5 13 0-0 'i!Vb6 14 tl)e2 b4 and Leitao's more active suggestion 13 . . . a5!? 14 tl)xb5 i.a6 15 a4 i.xb5 16 axb5 'i!Vb6 are unclear) 1 3 0-0-0 (13 i.xc5 tl)xc5 14 0-0 b5 1 5 'iVd4 b4! 16 'ilVxb4 llb8 1 7 'ilVa3 'itb6 1 8 'it>h1 d4!=) 13 ... b5 14 l:i.he 1 Paehtz-Lomineishvili, Leon 200 1 ; and 14 . . . b4! 1 5 tl)a4!? i.xd4 16 'ilVxd4 a5 17 f5 fte8 18 fxe6
239
Play the French fxe6 would be equal. 1 1 .. JWb6 In this particular instance, I think 1 1 . . .YWe7 makes sense, since after 12 i.. h 3 Black's defence of e6 gives him leeway to contest the light squares:
12 .. .f5 !?, a unique idea that I like. It challenges White to capture en passant and lose central control for the sake of free play; otherwise, any plan with g4 grants Black f-file play: 13 0-0 (13 exf6 liJxf6 threatens . . . liJe4, and Black gets easy play after 14 J.g2 e5 15 liJxc6 bxc6 16 �xc5 'iVxc5 1 7 fxe5 liJg4; 1 3 0-0-0 liJxd4 14 liJxd4 liJb6 15 b3?! il.d7 with the speedier attack) 13 . . . liJxd4 14 liJxd4 and a sample line is 14 . . . il.a7 (with the idea . . . liJc5) 15 .l:i.ac 1 ! liJc5 16 c4 liJe4 1 7 'ii'e 2 i.d7 1 8 .l:i.fdl i.. xd4 1 9 i.xd4 dxc4 20 'ii'xc4 .l:i.ac8 21 'ifb3 i.. c 6 22 i.b6 i.. d 5= . 1 2 c3 12 0-0-0 f6 ! . 1 2 ... a5 Here too 12 .. .f6 looks quite reason able, since 13 exf6 liJxf6 14 i.. g2 i.. d 7 discourages 1 5 0-0 due to 15 . . . liJg4. After 12 . . . a5, Ermolaev-Sumets, Lviv 2002 continued 13 i.. h 3 a4 14 'iilf2 'i'id8 1 5 'iil g2 i..b 6 16 b4 i.. c 7!? 1 7 liJb5 il.b8 18 liJed4 liJb6 19 'ii' d 3 liJc4 20 �f2 J.d7= . 240
13.32 10 0-0-0 a6
Preparing . . . b5 with an advance on the queenside. My first instinct in this type of position was that White, with more central space and good pieces, probably stands slightly bet ter. But White's goals on the kingside are hard to achieve and Black has White's king as a target for his queenside expansion. Moreover, in some cases a well-timed simplifica tion will equalise on the spot. Statis tically, in fact, Black has done re markably well from this position in terms of both raw results and per formance rating. 11 h4
A critical point. With this move White demonstrates his intention to attack on the kingside, perhaps via
Classical Variation: 4 e5 l:th3, a move that incidentally defends along the third rank. White can choose also to play positionally, which usually entails strengthening the d4 point and/or neutralising Black's ideas on the queenside, as shown by: (a) 1 1 'i1tb l (this will often trans pose to 1 1 h4 lines, but sometimes White switches plans) 1 1 . . .ttJxd4 12 �xd4
12 . . . b5 (after 12 . . . 'iVc7, it's impor tant to note that 13 h4 b5 transposes to one of the main lines below with 1 1 h4 ttJxd4 1 2 oltxd4 'iVc7 1 3 ..ti?b l b5; instead White can play 1 3 'i!Vf2 b5 14 i.d3 b4 1 5 ttJe2 a5 16 'i!fh4 g6 1 7 .l:thfl �a6 Zahariev-LIvanisevic, Chania 2000; then Black is way ahead of the lines with 1 1 1\Vf2 and can follow up by . . . .l:tfcS and even . . . ttJf8 if needed ChessPublishing) 13 'iWe3 (13 h4 transposes to 1 1 h4 ttJxd4 12 ii.xd4 b5 13 ..ti?bl) 13 . . . 'iVc7 (13 . . . 'i!Ve7 could be considered - Pedersen) 14 ii.d3 �xd4 (14 . . . b4 15 ft3!) 15 'iVxd4 i.b7! (This is a fashionable treatment of such lines; Black wants to develop quickly and tie White's pieces to e4 by placing a knight on c5 before pushing his queenside pawns. Still, after 1 5 . . :�c5 16 ttJe2 b4 17 l:the l l:teS 18 'i!Vxc5!? ttJxc5 1 9 ttJd4 Libiszewski-Buhmann, Pula 2003, 19 . . . i.d7! doesn't look bad)
16 l:the l ttJc5 1 7 ttJe2 .l:tacS!? ( 1 7 . . .l:UcS looks better, because a later . . . b4 and . . . a5 will be more effective) I S 'iie 3 ttJe4 1 9 ttJd4 (19 �xe4 dxe4 empowers the bishop, e.g., 20 .l:td2 f6! 2 1 exf6 l:txf6 22 :edl �d5=) 1 9 . . . 'iie 7!? (19 . . . b4!?) 20 ttJf3!? (I pre fer White a bit here, because Black isn't well set up for . . . b4 and . . . a5) 20 . . . 'i!Vb4 2 1 ..ti?al "fIe7 22 ttJg5 d4!? 23 'i!fh3 ttJxg5 24 fxg5 g6 25 'iig 4;!; J.Polgar-Luther, Ohrid 200 1 . White intends h4-h5, although things are still complex and Black went on to WID;
(h) 1 1 'i!Vf2 was recently played by Kramnik. The idea is to launch a kingside attack by i.d3 and 'i!Vh4. Here are two responses: (h I) 1 1 . . .�b4!? meets the threat 12 ttJxc6 and tries to force White's knight to a worse square. It may well equal ise, e.g., 12 ttJce2 (uninspiring is 12 ttJxc6 bxc6 13 ttJe2 .l:r.bS= ; and a mis take is 12 ttJa4? 'iWa5) 12 . . . ttJa5 (head ing for c4; 12 . . . 'iia 5!?) 13 c3 i.e7 1 4 ttJg3 b 5 1 5 i.. d 3 �b7 16 'iWc2 h 6 1 7 ..ti?b l ttJc4 18 i. f2 ttJc5= . These moves are hardly forced, but the general idea is that White has a difficult time attacking; (b2) 1 1 . . .ttJxd4 12 i.xd4 should pro bably be played in a non-standard
241
Play the French manner, because otherwise White remains somewhat better: 12 . . . 'iVc7 (for example, 12 . . . 'iVe7 would keep an eye on h4 and should be considered; then the move . . . f6 is also appropriate in some lines) 13 �d3 iLxd4!? (13 ... b5?! 14 'iWh4 h6 1 5 tiJe2 f6? ! 16 �g4!± was the game Kramnik-Radja bov, Linares 2003; 1 3 ... f6 !? might be playable, with the idea 1 4 exf6 l::txf6! 1 5 f5 !? b5) 14 �xd4 tiJc5 (14 . . . b5) 1 5 f5 ! tiJxd3+ 1 6 'u'xd3 with a nice advan tage; this is the sort of position White wants; (c) Nijboer has specialised in the relatively unexplored 1 1 tiJb3. A cou ple of his recent games went 1 1 . . .b6 12 tiJe2 a5!? (12 ... 'iWe7 1 3 tiJed4 tiJxd4 14 tiJxd4 f6 15 exf6 tiJxf6 16 �d3 iLb7 1 7 tiJf3 Nijboer-Arizmendi Martinez, Bled 2002; and Nijboer recommends 17 . . . tiJh5 with complications, appar ently dynamically equal ones) 1 3 tiJed4 tiJxd4 14 tiJxd4 �a6 1 5 'it>b 1 WIIc 7 16 �d3 (16 �xa6 'u'xa6 1 7 g4!?) 16 . . . a4 1 7 'tIYe2 iLxd3 18 cxd3 Nijboer Stellwagen, Wijk aan Zee 2003; and simply 18 . . . a3 was equal.
1 l tiJxd4 1 1 . . :iWc7 has led to very similar games, and is better played later (see below) . One early example went 12 ,U,h3 (12 h5!;!; or even 12 g4 tiJxd4 13
�xd4 b5 14 g5!?;!;) 12 . . . tiJxd4 1 3 .txd4 b5= 14 WIIe 3!? i.b7 15 i.. d 3 f6! 16 exf6 l::txf6� Mokry-Bareev, Trnava 1 989. 12 .txd4 b5
13 'u'h3!? This has been very popular, but re cently White has been looking at other courses as well: (a) 13 h5 looks scary, but gives Black a critical tempo for his own attack, e.g. , 13 . . . b4 (13 . . . Vic7 is often arrived at by a move order with an earlier . . . Wllc 7; when White commits to h6 it usually doesn't achieve much, e.g. , 14 h6 g6 15 'it>b 1 and 15 . . . �b7 16 .txc5 tiJxc5= o r 1 5 ... .tb 7 16 .txc5 tiJxc5=)
...
242
This leads to a typical and major decision. Does White want to block the queenside and simplify, or to con-
Classical Variation: 4 e5 centrate upon the centre at the cost of allowing Black's queenside pawns to advance? We have: (al) 14 liJe2 1 5 'ii'e 3 'ii'c 7 16 'iitb l ( 1 6 i.. xc5 liJxc5 1 7 liJd4 �a6 1 8 f5 McDonald; 1 8 . . . liJe4! with the idea 19 fxe6 'i'xe5 or 1 9 'ii'f4 exf5=) 1 6 ... �a6 17 i.. xc5 liJxc5 18 liJg3?! l:tfc8 19 l:!.c1 a4! 20 i.. xa6 .l:txa6 and both . . . a3 and . . . b3 are serious worries for White, Topalov-Morozevich, Sarajevo 1 999; (a2) 14 liJa4 i.xd4 1 5 'i'xd4 a5 (15 .. .f6!? 16 h6 g6 and Black's king seems secure while White's centre is becoming loose) 16 �b5 (to stop . . . i.a6; this same manoeuvre is tried in the main line) 16 . . . 'u'b8 17 i.. d 3 and a recent game Olenin-Zvj aginsev, Togliatti 2003 went 17 . . . i..b 7 18 'i£;>b l i.c6 1 9 liJc5 i..b 5 20 liJxd7 'iWxd7 2 1 g4!? l:tfc8 22 f5 i.. xd3 23 cxd3? b3! 24 a3 ':'c2=1=; (b) 1 3 i..x c5 liJxc5 1 4 'ii d4 'ii'c 7 1 5 a 3 (15 f5 i..b 7 16 f6?! - committing too quickly - 1 6 . . . gxf6 1 7 exf6 'i£;>h8=1= A.lvanov-Glek, USSR 1987) 15 . . . i.. d 7 16 f5 :fc8! 17 f6?! (the same problem) 17 . . . gxf6 (17 . . . liJe4!) 18 exf6 Apicella M.Gurevich, Clichy 200 1 ; 18 . . . liJe4 ! ; (c) 1 3 'iitb l !? i s a subtle move order that has recently come into promi nence. 1 3 . . . 'ii'c 7
An important position
IS
reached by other orders, e.g., 11 �b l liJxd4 12 i.xd4 'i'c7 13 h4 b5. Now both of White's moves are critical: (cl) 14 l:th3 (Anand gives this '?!') 14 . . . b4!? (as usual, 14 . . . �b7 is an op tion, especially in view of 15 h5 b4 16 liJ a4 i.xd4 1 7 'iix d4 'ii'a 5 1 8 b3 i.c6!=I=) 15 liJa4 i..x d4 16 'ii'xd4 l:!.b8!? (or 16 . . . a5 1 7 c3 i.. a 6!) 1 7 l:!.e3 a5 18 f5?! (18 h5 i.b7 1 9 f5 libc8 20 .i.d3 �a6! with initiative) Anand-Buhmann, Stuttgart 2002; and Anand gives 18 . . . exf5! 1 9 'ii'x d5 i.b7! 20 'iWxd7 l:tbd8 21 e6 l:!.xd7 22 exd7 i.. c 6 23 liJc5 l:!.d8 24 �c4 �f8 25 l:tdel i.. e 4 26 liJxe4 fxe4 27 �b5 f5=1= ; (c2) 1 4 h 5 ! b4 1 5 liJa4 i.. xd4 16 'ii'xd4 a5 (16 . . .f6!) 1 7 h 6 (17 i..b 5 lib8 18 h6 g6 19 c4 bxc3 20 liJxc3 f6!=) 1 7 ... g6 1 8 l:!.c1 ! 'ii'c 6 19 b3 i.a6 20 �xa6 'ii'xa6 2 1 c3 bxc3 22 ':'xc3;!; Khamatgaleev-Kosic, Patras 2002. 1 3 b4 One last time 13 . . . 'i'c7 14 'i£;>b l transposes, this time to note 'c' 13 �b l above. An important alternative, currently popular, is 13 . . . i.b7 Al though move order is certainly impor tant, this may be one of those varia tions where a knowledge of basic ideas (like the difference between . . . b4 and . . . i.b7 lines) is more helpful that memorising exact orders. ...
often 243
Play the French Here are some examples of play af ter 1 3 . . . -tb7: (a) 14 f5?! is premature: 1 4 ... exf5 1 5 tiJxd5 �xd5 1 6 �xc5 tiJxc5 1 7 'i!Vxd5 'i!Vb6 1 8 'i!Vd6 'i!Va7! 19 .l:!. f3 ttJe6 20 'i!Va3 f4 21 .l:!.d6?! 'i!Vgl 22 'i!Vd3 l:i.ac8 and everything is going Black's way, Van der Weide-Stellwagen, Leeuwar den 200 1 ; (b) 14 .l:!.g3 !? b4 1 5 tiJa4 �xd4 1 6 'i!Vxd4 'iWa5 1 7 b3 i.c6 1 8 tiJb2 ( 1 8 f5 !? - Pedersen) 18 ... ttJc5 19 �d3 l1fd8:j: was Dutreeuw-M. Gurevich, Brussels 1995; (c) 14 g4 (a popular move) 1 4 ... h4 1 5 tiJe2 a5 1 6 g5 'i!Vb6 1 7 'iitb l ..ta6 18 l:te3 (18 h5, and instead of 1 8 . . . .l:!.ac8?! 19 g6! Coco-Daconto, corr 1997, Black should play 18 . . . .l:!.fc8!, as follows . . . ) 18 . . . .l:!.fc8! 1 9 �g2 (19 h5 a4 20 g6 hxg6 21 hxg6 fxg6 22 .l:!.g3 tiJf8)
1 9 . . . ..txe2! 20 .l:!.xe2 a4 2 1 f5 b3 22 cxb3 axb3 23 a3 �xd4 24 'iWxd4 l:tc5!? 25 h5 .l:tac8 26 �d3 ':c2+ Borriss Korchnoi, Panormo 200 1 ; (d) 14 'iitb l?! has been played a lot, but Black's attack is better than usual, e.g., 14 . . . b4 15 ttJe2 a5 16 'iYe3 Wic7 17 .l:tg3!? SLa6:j: 18 .ixc5? tiJxc5 1 9 f5 f6?! (19 . . . ttJe4!-+) 20 fxe6? (20 exf6 I!xf6 2 1 ttJf4 d4!) 20 . . . fxe5 2 1 l:tf3 d4-+ Aagaard-Stellwagen, Wijk aan Zee 200 1 . 244
14 tiJa4 14 tiJe2 is rather slow in the face of 14 . . . a5 15 Yi'e3 (15 h5 SLa6 16 h6 g6 17 g4 'i!Vb6 18 'iitb l a4 19 .l:!.e3?! ..tc4 20 c3?! SLxa2+! 2 1 'iitxa2 b3+ 22 'iitb l a3 with a huge attack, Adriano-C.Jones, corr 2003) 15 . . . 'i!Vc7. Then 16 �xc5?! tiJxc5 17 tiJd4 a4 18 'iitb l a3! 1 9 b3 �a6 20 i.xa6 .l:!.xa6 favoured Black, J.Polgar-Shirov, Prague 1 999. 14 .....txd4 1 5 'iYxd4 a5 Black will simplify and White's at tack is not yet underway. There are two exciting alternatives, one fully satisfactory: (a) 1 5 . . . f6 16 'i!Vxb4 fxe5 1 7 'iWd6! 'iWf6 1 8 f5 ! ; this is Kasparov's ingen ious move. A playable line is that given by Ron Langeveld in ChessPub (with some additions from me) : 18 . . . .l:!.e8! 19 .l:!.b3 (19 fxe6 'iYf4+ 20 'iitb l 'iVxa4 21 exd7 �xd7 22 .l:!.a3! 'iWg4 23 Wixd5+ �e6;!;; 19 'iYc6 .l:!.b8 20 fxe6?! 'iWf4+ 2 1 Wb l tiJf6 22 .l:!.f3 'i!Vg4+ ; 19 tiJc3!? Wh8! 20 fxe6 'i!Vh6+ 2 1 1ifi?b l tiJf6 22 g4!? tiJxg4 23 tiJxd5 SLxe6:j:) 19 . . .'iWxh4 20 fxe6 'i!Vxa4 21 ':'b4 (2 1 'iWxd5 tiJc5! !) 2 1 . . .'i!Vxb4!? 22 'iWxb4 tiJf6 with a small edge for White; (b) 15 . . . 'iVa5! seems fine for Black, the main line being 16 b3 i.b7!? (or 16 . . . f6! 1 7 exf6 ttJxf6=) 1 7 c3! (17 l:i.g3 �c6 18 tiJb2 tiJc5!) 17 . . . .l:!.fc8 18 Wb2 bxc3+ 19 .l:!.xc3 l:i.xc3 20 'iWxc3 Nijboer Luther, Leuwaarden 1992; and after 20 . . . 'iWd8 21 I!cl?! (2 1 g3! l:i.c8 with a tiny edge for White) 2 1 . . . .l:!.c8 22 'i¥b4, Langeveld suggests 22 . . . .l:!.xc l ! 23 Wxcl ..tc6! 24 �xa6 'iYxh4+ . 16 ..tb51? This is considered best, to prevent 16 . . . ..ta6. 16 c4!? �b7! is easier than 16 . . . bxc3 1 7 ILxc3!? 'iVxh4 18 g3 'iWd8 19 1ifi?b l with some play for the pawn, Nijboer-Korchnoi, Arnhem 1999. 16 ....l:!. b8
Classical Variation: 4 e5 1 7 . . .'iWc7 equalises in both theory and practice, e.g., 18 l:tel 'iic 6 19 b3 .ta6 20 �xa6 'ii'xa6 21 h5 (2 1 11g3!? c;.t;;>h 8 Pedersen) 2 1 . . .l::tfc 8 22 f5 'iWc6 23 l:te2 It'lc5= Sedlak-Antic, Subotica 2000. 1 8 exf6 18 b3?! fxe5 19 fxe5 'iIIc 7 20 l:tel 'i'c6 2 1 h5 £La6; Garofalo-Latronico, email 2000. 1 8 'iixf6 1 9 'i!i'xf6 ..•
1 7 �d3! White's loss of tempo is supposed to be worth prevention of . . . �a6. The alternatives lack punch, e.g., 1 7 �e2 .tb7 18 'it>b l .tc6 19 It'lc5 �b5 (19 . . . lt'lxc5 20 'iWxc5 l:tc8 21 'iWd4 f6=) 20 �d3 It'lxc5 2 1 'iWxc5 'iWd7= De la Riva Aguado-Glek, Saint Vincent 1999. 17 f6! ...
1 9 lt'lxf6 A recent test of 17 .. .f6 by Shirov went 19 . . . 11xf6 20 l::t e 3 l:txf4 2 1 llxe6 It'lf6 22 lid6 �g4 23 I:tel �h5 24 It'lb6 �f7 25 It'ld7 It'lxd7 26 l:lxd7 l:txh4 27 litee7?! (27 l::t a 7=) 27 . . . lIfB 28 i.f5 l1h6 29 .tg4 l:tf6; Langheinrich-Shirov, Cologne 2003. After 1 9 . . . lt'lxf6, Pedersen's analy sis goes 20 .l:lf3 It'le4 2 1 .txe4 dxe4= . White has nothing after 22 l!e3 (al though not 22 l:tffl? £La6 or 22 l:tf2 e5!) 22 .. J:txf4 23 g3 ltf8 24 1:txe4 l::tb 5 ! = . Black's idea is to cover c5 and e5 while preparing . . . ltd5 or . . . e5, freeing his bishop . ...
A rare but effective solution in this well-known position, recommended by Pedersen. It also seems that
245
Chapter Fourteen Classical Variation : 4 �g5 (4 . . . dxe4 5 ctJxe4 �e7 )
1 e4 e 6 2 d4 d5 3 lbc3 lbf6 4 SL g5 dxe4 5 lbxe4 �e7 This system, called the Burn Variation, is comparable to the Ru binstein Variation (3 . . . dxe4 4 lbxe4 lbbd7 and 5 . . . lbgf6) except that the opposition of White's bishop on g5 and Black's on e7 will almost inevita bly lead to an exchange. 5 . . . lbbd7 is also popular at this point, instead of 5 . . . iLe7. The disadvantage of any . . . dxe4 system is that it leaves White with space and better control of the centre. The advantage of the varia tion we see here is that Black will normally get the two bishops and is constantly threatening to equalise in the centre with . . . c5 or . . . e5. Although essentially a solid line, it tends to be more double-edged than the other two . . . dxe4 systems, and thus enjoys great popularity in contemporary play. 6 iLxf6 It's remarkable that the two main (and best) recent books covering these lines, by Jacobs and Pedersen, don't mention natural alternatives at this point, nor many deviations from main lines in the moves to follow. Through
246
the years, White has tried a number of other approaches, some quite often: (a) 6 .i.d3? is a mistake that was made by several top-class players in the early part of the 20th century, but not established as weak until later: 6 . . . lbxe4 7 SLxe7 lbxf2!? (or 7 . . . li'xe7! 8 .txe4 'iWb4+ 9 c3 'i'xb2 10 lbe2 c6 1 1 0-0 lbd7 1 2 .i.c2 'i'a3 1 3 'i'd3 lbf6 and White lacked compensation, Leon hardt-Swiderski, Vienna 1908) 8 SLxd8 lbxdl 9 i.. xc7 lbxb2 10 .te2 lba4! 1 1 SLd6 Barasz-Balla, Gyor 1906, and here 1 1 . . .iLd7! 12 :tb l .tc6 was good; (b) 6 lbg3
and:
Classical Variation: 4 iLg5 (4. . . dxe4 5 tDxe4 iLe 7) (b l) After 6 . . . 0-0, 7 ttJf3 is natural and likely to be met by . . . ttJbd7. In stead, Djurhuus-U.Andersson, Malmo 1995 took an interesting positional course after 7 J.d3 c5 8 dxc5 ttJbd7 9 �xf6 ttJxf6!? 10 'iVe2 �a5+ 1 1 c3 �xc5 12 ttJf3 �d7 1 3 0-0 .l:tad8 14 .l:tfel �c8!? 1 5 l:tadl 'iVc7 1 6 ttJe4 1Id5!?=; (b2) 6 ... c5 7 dxc5 (7 ttJf3 ttJbd7 8 �c4 0-0 9 0-0 a6! 10 dxc5 ttJxc5 1 1 'iWe2?! b5 12 iLb3 ttJxb3 1 3 axb3 'Wic7 14 lUe l �b7+ Van Gompel-De Bock, Vlissingen 1 995) 7 . . . 'Wia5+ 8 'Wid2 'ii'xc5 9 ttJf3 ttJbd7 10 0-0-0 b6 1 1 �e3 '§d5 12 'Wixd5 ttJxd5= Gesos-Van den Doel, Agios Kyrikos 2000; (c) 6 ttJxf6+ allows Black to develop quickly and achieve an equalizing central break. An older example: 6 . . . i.. xf6 7 i.xf6 'iVxf6 8 c3 0-0 9 i.d3 ttJc6 10 ttJf3 e5 1 1 dxe5 ttJxe5 12 ttJxe5 'iVxe5+ 13 'iVe2 '§xe2+ 14 i.xe2 i.e6 15 0-0 lIfd8 16 l:tfdl 1:i.xd1+ 17 �xdl l:td8=i= Thomas-Tartakower, Hastings 1 945 - see also note (a) to White's seventh move. (d) 6 ttJc3 has been tried surpris ingly often. A simple solution is 6 . . . 0-0 7 ttJf3 ttJbd7 8 �e2 c5 9 0-0 b6 10 'iVd2 i.b 7 1 1 l:tadl ttJd5= Botvin nik-Rauzer, Odessa 1 929. 6 �xf6 ...
There are roughly 3000 games
from this position in my database, with players of all strengths repre sented. Recently it has become an extremely popular opening at the highest levels. Of the leading con temporary players, Bareev and Mik hail Gurevich have employed it most consistently over the years, but Shi rov, Khalifman, Korchnoi and Short have been exponents, and a number of others such as Kramnik, Dreev, Ehlvest, Psakhis and (many others) have played it from time to time. 7 ttJf3 Again, others have often been seen: (a) 7 ttJxf6+ 'Wixf6 8 ttJf3 0-0 (8 . . . c5 also equalises) and now: (al) 9 YWd2 is a simplifying line that offers few positive prospects, e.g., 9 . . . lId8 10 0-0-0 c5 1 1 'iVc3 b6 12 �b5 iLa6!? 13 'iWc4 iLxb5 14 'iWxb5 ttJd7 15 'iYe2 l:!.ac8=i= � - � Anand-Ba reev, Paris 1992; (a2) 9 c3 ttJd7 (or 9 ... b6 10 g3 ttJd7 11 �g2 e5! 12 ttJxe5 ttJxe5 13 dxe5 'iYxe5+ 14 'iYe2 '§xe2+ 1 5 �xe2 iLa6H was Szamoskozi-Jakab, Paks 1 995) 10 iLd3 (10 iLe2 e5 1 1 dxe5 ttJxe5 12 ttJxe5 'iVxe5= Capablanca Alekhine, New York 1 927) 10 . . . e5 1 1 0-0 exd4 1 2 ttJxd4!? ttJe5 1 3 �e4 l:td8 1 4 'iYc2 g6 15 l:tfdl � - � Oll-Kinder mann, Debrecen 1 990; (a3) 9 �d3 ttJc6!? (or 9 . . . c5 10 0-0 cxd4 1 1 ttJxd4 l:td8= Gashimov-Rad jabov, Baku 1998) 10 c3 e5 1 1 dxe5 ttJxe5 12 ttJxe5 'ii'x e5+ 1 3 'iVe2 'ii'x e2+ 14 Iifilxe2 .l:.e8+ 15 �d2 i.e6= Yurchenko-Moiseenko, Kaluga 1 996; (b) 7 c3 ttJd 7 and: (b l) 8 f4!? 0-0 9 ttJf3 b6 10 �d3 �b7 1 1 0-0 c5 12 dxc5 ttJxc5 13 ttJxc5 bxc5= Boll-Dreev, Tilburg 1993; (b2) 8 'iVc2 e5 9 dxe5?! (9 O-O-O! Bareev, when 9 . . . exd4 10 ttJxf6+ ttJxf6 1 1 lIxd4 'ii'e 7 1 2 ttJf3 0-0 looks equal) 24 7
Play the French 9 . . . liJxe5 10 f4 (Knaak gives the inter esting line 10 i..b 5+ c6 1 1 :dl 'iIIe 7 12 liJd6+ 'iio>f8 1 3 i.. e2 i.. e 6�, which seems true, e.g., 14 b3 l:ld8 15 liJe4 i.f5 !) 1O . . . liJg6 1 1 g3 0-0 12 i.. d 3 (12 i.g2 l:te8) 12 . . . 'i!fd5 and Black is slightly better: 1 3 a3? (13 liJxf6+? gxf6-+ with the idea 1 4 0-0-0 'ikxhl 1 5 i.e4 j.f5 ! ; 1 3 liJe2!�) 1 3 . . . liJxf4! 14 liJxf6+ gxf6 1 5 i.xh7+ 'iio> g 7-+ Topalov-Bareev, Lin ares 1994; (b3) 8 liJf3 0-0 9 'itc2 (9 i.e2 e5 10 d5 Ae7 and . . . f5) 9 . . . e5= 1 0 0-0-0 (10 d5 i.e7 11 0-0-0 has arisen more than once and 1 1 . . .f5 !? is curiously un tried:, e.g., 12 d6! i.xd6 13 liJxd6 cxd6 14 llxd6 e4! 15 liJd4 "fie7 16 l:te6 'ikg5+ 17 'it>b l liJb6 with satisfactory play) 10 . . . exd4 1 1 liJxd4 (1 1 liJxf6+? ! 'ii'xf6 12 l:rxd4 liJc5 and Black is al ready better, Ljubojevic-Dreev, Lina res 1 995) 1 1 . . .i.. xd4!? 12 l:lxd4 'i!fe7 1 3 h 4 liJ e 5 14 liJg5 g6= Glek-Lputian, Dortmund 1992. 7 ...liJ d7
I use this move as the first reper toire line because it prepares the ideas of . . . e5 and . . . c5 without delay. Nevertheless, 7 . . . 0-0 is flexible with respect to minor piece development and is in fact more popular. This has turned out to be a solid position for Black. Here's a mini248
repertoire of critical ideas:
(a) 8 c3 liJd7 9 'iVc2 (9 i.d3 e5) 9 . . . e5 transposes to 7 c3; (b) 8 "fid3 (to target h7 and play 0-0-0 quickly) 8 . . . liJd7 9 0-0-0 b6 10 h4 �b7 11 liJfg5 (1 1 liJeg5 g6 1 2 'i!fe3 h6 13 liJe4 i.. g 7 14 i.. d 3 c5!, or here M.Gurevich's 12 . . . c5!? 13 dxc5 'ilVe7) 1 1 . . .i.. xe4!? 12 'ilVxe4 Axg5+ 1 3 hxg5 "tWxg5+ 14 'it>b l liJf6 15 'ikc6 Milos Shirov, Las Vegas 1999, when White had only limited compensation for the pawn; (c) 8 Ac4; as Pedersen points out, many of the top players think that this is the drawback to 7 . . . 0-0, be cause that move doesn't contribute towards stopping the d5 break (as . . . b6 does) or enforce . . . c5, (as . . . liJd7 does) . But I have grown fonder of 7 ... 0-0 since I realised that Black can now switch plans by 8 . . . liJc6!? 9 c3 e5 Then the only challenging move is 10 d5 , when both moves o f Black's moves are valid: (c l) 10 . . . liJa5 1 1 i.d3 b6 12 h4, and Jacobs suggests 12 . . . liJb7! intending . . . liJd6, e.g., 13 liJxf6+ 'ikxf6 14 'ilVc2 g6 15 h5 liJc5=; (c2) 1 0 . . . liJe7!? 11 liJxf6+ gxf6 has done reasonably well for Black, the current main line going 12 liJh4 liJg6 13 'iWh5 'iIId 7! (intending .. .'ilVg4) 14 h3
Classical Variation: 4 JJ..g 5 (4. . . dxe4 5 1:fJxe4 i.e 7) 'iWa4! 1 5 b3 'i'a5 1 6 0-0 Shirov V.Akopian, Kallithea 2002; and here Shirov suggests 16 . . . �g7 17 b4 'iWa3 18 JJ..b 3 a5 with complications; (d) 8 'if d2 is the most common move, when 8 . . . liJd7 transposes to our 7 . . . liJd7 main line, or Black can play 8 . . . il.e7 9 0-0-0. In the latter case, 9 . . . 'iWd5!? (9 . . . liJd7 again transposes to 7 . . . liJd7 8 'iVd2 0-0 9 0-0-0 J.. e 7 below) 10 liJc3 'iWa5 is independent: 1 1 a3!? (11 liJe5 JJ..b 4 12 liJc4 JJ.. xc3 1 3 liJxa5 il.xd2+ 14 l:Ixd2 b6 15 liJc4 JJ..b 7= was the game Ehlvest-Khalifman, Japfa 2000) 1 1 . . .liJd7 and:
(dl) 12 i.c4 c6 13 l:the l b5 14 JJ..b 3 Landa-Baklan, Halkidiki 2002; and Baklan suggests 14 . . . b4 15 liJa2 .l:tb8 16 �b l .l:td8 'with counterplay'; (d2) 12 'it>b l 'iWb6 (or 12 . . . c6!? Knaak; both sides have numerous options here) 13 'i'e3 liJf6 14 liJe5 Po nomariov-Ivanchuk, and here Knaak suggests the simple 14 . . . liJd5! 1 5 1\Vf3 (15 liJxd5 exd5 frees Black's bishops) 15 . . . liJxc3+ 16 'i'xc3 c5, which should be fine for Black, e.g., 17 dxc5 (17 JJ.. e 2 cxd4 1 8 ':'xd4 i.c5=) 1 7 ... 'iWxc5 1 8 'i'xc5 i.xc5 1 9 f3 ! (19 f4 b6 20 i.b5 i.bTi=) 19 . . . a6! 20 il.e2 b5 2 1 liJd7 i.xd7 22 l:txd7 .l:tfd8 23 .l:thdl i.b6= . I have dwelled upon 7 . . . 0-0 because it is a good alternative for anyone
who mistrusts something in the 7 . . . liJd7 Iines. 8 'i'd2 A very important choice. 8 'iVd2 is perhaps the best move since it leaves Black few ways to exploit his early . . . liJd7 and usually merges with a line that could arise after 7 . . . 0-0. There are many other moves, however, sev eral of them leading to highly critical play: (a) 8 'i'e2 !? 0-0 9 0-0-0 b6 10 h4!? i.b7 11 h5 c5! (breaking up the centre based upon a tactic) 12 h6 g6 1 3 dxc5?! (13 li'lxf6+ 'iVxf6 14 l:th3 cxd4 15 li'lxd4=) 13 . . . J..x e4 14 1\Vxe4 li'lxc5 15 'iib 4 (15 l:txd8 li'lxe4+) 15 .. JWc7 16 JJ.. e 2 J.Polgar-Bareev, Cannes 200 1 , and Black launched a n attack aided by the opposite-coloured bishops: 16 ... b 5 ! 1 7 'it>b l a5! 1 8 'i'g4 li'la4 1 9 li'ld4 .l:tfc8 2 0 :d2?! (already there is no good defence, but better was 20 c3 b4 21 il.b5 bxc3! 22 JJ.. xa4 l:iab8+) 20 . . . liJxb2! 21 liJxb5 .l:tab8! 22 'it>cl (22 li'lxc7 li'ld3+) 22 . . . 'iVe5- + ; (b) 8 il.c4 is by far the most impor tant option to 8 'iVd2 because White wants to support d5 in some varia tions, especially after . . . c5. Now 8 . . . 0-0 transposes to certain passive 7 . . . 0-0 lines that are unappealing. But Black has the alternative 8 . . . a6! :
249
Play the French A dynamic and ambitious idea on which M.Gurevich has taken out a patent. Black wants to develop his bishop aggressively even at the cost of some looseness: Play normally con tinues 9 Wie2 (9 a4 0-0 10 Wid2 b6 1 1 0-0-0 b5! - McDonald; 9 0-0 b 5 10 �d3 �b7=, e.g., 11 c3 �e7!? 12 a4 O-O! 13 axb5 axb5 14 l:txa8 Wixa8 1 5 �xb5 �xe4 1 6 �xd7 l:td8 1 7 �b5 c5+; the key in these lines is to stay dy namic) 9 . . . b 5 (9 . . . 0-0 is safer: 10 0-0-0 b 5 11 �b3 �b7 12 d5 exd5 1 3 �xd5 c6 14 �b3 Wic7 15 l:the 1 Lutz-Dizdar, Frohnleiten 2002; and here 15 . . . c5 is similar to our main line, e.g., 16 �d5 �xd5 1 7 l:txd5 �e7= ; no better was 1 5 ct::J d6 ct::J c 5= Anand-Radjabov, Dubai (rapid) 2002)
Now there's a significant division: (b 1) 10 �d3 �b7 1 1 ct::Jxf6+ Wixf6 12 �e4 i.xe4 1 3 Wixe4 0-0 14 0-0 Wie7 1 5 a 4 ct::Jb 6 16 axb5 axb5 1 7 Wic6 b4 '12 - '12 Timman-M.Gurevich, Wijk aan Zee 2002; Black can soon play . . . Wid6; (b2) 10 �b3 0-0 1 1 0-0-0 (1 1 0-0 c5 12 dxc5 �b7 13 l:tfd1 �xe4 14 Wixe4 ct::J xc5= Van den Doel-Komarov, Cler mont-Ferrand 2003; 1 1 l:!.d1 �b7 12 0-0 c5 13 c3 '12 - '12 Milos-Bareev, Shen yang 2000) 1 1 . . .i.b7 12 d5!? exd5 1 3 �xd5 c 6 14 �b3 Wic7 1 5 l:!.he 1 c5 16 �d5 �xd5 1 7 l:!.xd5 �e7 1 8 'it>b 1 Van 250
den Doel-Mullon, Clermont-Ferrand 2003; and Black might have kept up some pressure by 18 . . . l:tfe8! ; (b3) 10 �d5!? l:!.b8 1 1 0-0-0 0-0 (1 1 . . .l:!.b6!?) 12 �c6! l:!.b6 1 3 d5 exd5 14 i.xd5 c6 15 i.b3 c5! (15 . . . a5!? 16 c3 a4 17 �c2 Wic7;l;) 16 �d5 (16 ct::Jxc5? i.xb2+! and 17 'it>xb2 Wif6+ or 17 'it>b 1 i.f6 18 ct::J xd7 �xd7 19 ct::J e 5 i.xe5 20 Wixe5 Wic8+ Goloshchapov-M.Gurevich, Wijk aan Zee 2001) 1 6 . . . Wic7 1 7 l:!.he 1 �d8! (with the idea . . . ct::J f6) 1 8 ct::J e g5 (18 Wie3 c 4 19 ct::Jfg5 �xg5 2 0 Wixg5 ct::J f6 2 1 g 3 '12 - '12 Shirov M. Gurevich, New Delhi 2001) 18 . . . i.xg5+ 1 9 ct::J x g5 ct::Jf6 20 Wie5?! '12 '12 Svidler-M. Gurevich, Esbjerg 2000; in fact, 20 . . . Wixe5 21 l:!.xe5 ct::Jx d5 22 l:!.exd5 �b7 23 l:!.xc5 i.xg2! is quite appealing. When the world's leading players make no progress against a variation, it's a good indication that Black's play is sound; (c) The natural 8 �d3 can be an swered with 8 . . . c5!, when 9 dxc5 (9 ct::J d6+? 'it>e7 will actually win material for Black; 9 ct::J xf6+ Wixf6 10 0-0 cxd4 1 1 l:!.e1 0-0 12 ct::J xd4 l:!.d8 with an easy game, Kindermann-Nikolaczuk, Dort mund 1986) 9 . . . ct::J xc5! was Alekhine's idea in 1927! If 10 ct::Jx c5 (10 ct::J xf6+ Wixf6 1 1 O-O?! Wixb2+ Yates-Alekhine, Kecskemet 1927), 10 . . . Wia5+ 1 1 c3 Wixc5 frees Black's game and keeps the two bishops; (d) 8 ct::Jxf6+ is tame: (d1) 8 . . . ct::Jxf6 9 �b5+ (9 ct::J e 5 0-0 10 i.d3 c5! 11 dxc5 Wia5+ 12 c3 Wixc5 1 3 Wie2 b 6 = Standke-Krueger, Germany 1995) 9 . . . i.d7 10 �xd7+ Wixd7 1 1 Wie2 0-0 12 0-0 l:!.fd8= Efimov-M. Gurevich, Saint Vincent 2003; (d2) 8 ... Wixf6 9 Wid2 (trying to an ticipate and perhaps clamp down on . . . c5 and . . . e5; 9 �c4 0-0 10 Wie2 a6!? 11 0-0-0 b5 12 �d3 �b7= ; 9 i.d3 c5
Classical Variation: 4 i.g5 (4. . . dxe4 5 ttJxe4 Ji.e 7) 10 0-0 cxd4 1 1 J:te1 ltJc5 12 i.b5+ i.d7 1 3 i.xd7+ 'h - 'h Sigurjonsson-Pach man, Munich 1 979) 9 . . . c5! (9 . . . 0-0 10 'i!Ve3!) 10 0-0-0 cxd4 11 'i'xd4 'itxd4 12 l:.xd4 e5 13 l:r.d2 f6= . We return to the main line after 8 'iVd2:
8 0-0 In my database this main line posi tion, also arising from 7 . . . 0-0 8 'iWd2 ltJd 7, has scored 51 % for Black with only a miniscule edge in performance rating for White. It continues to offer reasonably active play while keeping risks to a minimum. It's definitely worth it to briefly compare a couple of the older moves in this position, one of which difficult yet instructive, the other of which is doing well again: (a) 8 . . . i.e7, to retain the bishop pair, has had a poor reputation ever since Fischer-Benko, Curacao 1962, but Benko's recent notes illustrate how even a passive-looking and un derdeveloped position can be playable when one has the bishop pair: 9 0-0-0 ltJf6 10 i.d3 0-0 11 ltJxf6+ Ji.xf6 12 'iVf4 c5 (queried, but Black needs ac tivity) 13 dxc5 'ita5 14 'iWc4 i.e7 15 h4 'iWxc5 16 'i!Ve4 f5 1 7 'iWe2 b5! 18 ltJg5 (18 i.xb5 l:.b8 19 Ji.c4 Ji.f6 20 i.b3 a5 is unclear - Benko) 18 . . ...tf6! 19 ...
ltJxe6!? (19 'i!Vh5 h6 20 'i!Vg6 hxg5 2 1 hxg5 i.xb2+ 2 2 �xb2 'i&'b4+ 23 'iit> c 1 'itf4+ 24 ,U,d2 'iVd4 25 'i!Vh7+ 'iit> f7 is given as unclear by Benko) 19 . . . i.xe6? (19 . . . Ji.xb2+! 20 �xb2 'iVb4+ 21 c1 'ii'a 3+ 22 �d2 'i'a5+ with perpetual check - Benko) 20 'iVxe6+ �h8 2 1 'iit>b 1 'i'xf2 2 2 'i'xf5 'itxf5 23 i.xf5;!;, a game Fischer went on to win after further mistakes by Black. But the opening was not to blame; (b) 8 . . . b6 is another old move, quite popular in the early days of the French Defence. It scored overwhelm ingly for White up to about 1985, but has served Black well since and re mains a viable alternative to 8 . . . 0-0: (b 1) 9 i.b5 0-0 10 i.c6 l:.b8 1 1 0-0-0 i.b7 12 i.xb7 (12 d5!? ltJc5 13 'iff4 ltJxe4 14 'i!Vxe4 i.xc6 1 5 dxc6 'ifc8 16 ltJe5 'i!Va6 1 7 �b 1 i.xe5 18 'i!Vxe5 'iWc4=) 12 . . . l:.xb7 1 3 d5 ltJc5 14 ltJxf6+ 'iWxf6 1 5 l:.he1 lId8 16 'i'e3 �bb8 1 7 dxe6 ltJxe6= Balcerak-Atalik, Cap pelle la Grande 2003; (b2) 9 0-0-0 i.b7 10 'iWf4 (10 i.d3 i.e7 11 'iWf4 ltJf6 12 h4 i.xe4 13 i.xe4 ltJxe4 14 'iWxe4 "YWd5 1 5 l:1.he 1 (15 'i'xd5 exd5 16 l:.he 1 f6=) 1 5 . . . 'i'xe4 16 lixe4 Solovjov-Grishina, St Petersburg 200 1 ; and best was 16 . . . .l:.d8! 17 c4 c6 18 ltJe5 l:r.c8=) 10 . . . 'i!Ve7!? 1 1 i.c4 O-O-O!? (1 1 . . .0-0!=) 12 .l:i.he 1 ltJf8 13 g3 ltJg6 14 'ii'e 3;!; Vallejo Pons-Korchnoi, Bie1 2002. 9 0-0-0 9 'ii'c 3, similar to Kasparov's idea 9 0-0-0 i.e7 10 'i!Vc3, hasn't been tried as far as I know. Again Black can break up the centre, e.g., 9 . . . c5! 10 ltJxf6+ (10 ltJxc5 ltJxc5 1 1 'ii'xc5 b6 12 'i'c6 l:.b8 13 "YWa4 i.b7 14 l:td1 i.xf3 15 gxf3 'ii' d 5:j:) 1O ... "YWxf6 11 0-0-0 l:r.d8 12 i.d3 b6 (12 ... cxd4 is also playable) 1 3 i.e4 .l:i.b8 14 l:r.he 1 i.b7 ( 1 4 . . . h6=) 1 5 i.xb7 .l:i.xb7= . 251
Play the French 9 0-0-0 introduces the main posi tion for this chapter. Black has a choice: 14.1 9 ... �e7 14.2 9 ... b6 14.1 9 ... �e7
9 ... .i.e7 has been a favourite weapon of Bareev's for years. Black courageously retains his two bishops at the cost of time. This does strengthen the support for . . . c5 and makes way for . . . ltJf6; but White also gains in that the time spent makes it more likely that a kingside attack will yield results. Thus this may be seen as a somewhat riskier and more ambitious strategy than developing move 9 . . . b6 of the next section. Al though it occasionally suffers a tacti cal reverse, the two bishops often have their say and Black actually has an excellent performance rating with 9 . . . �e7. 10 �c4 I make this the main line because of its importance rather than its fre quency of use. Black's problem now is that the move . . . c5 is generally an swered by d5. In fact, a complete so lution to 10 �c4 could be taken as a general sign of the system's worth. 252
Nevertheless, of the alternatives that follow, 10 �d3 and 10 'iWc3 are par ticularly important to take into ac count: (a) 10 'iWf4!? c5!? 1 1 dxc5 'iWa5 12 'it>b l ltJxc5= ; (b) 10 'it>b l b6!? (or 10 . . . c5 1 1 dxc5 'iWc7) 1 1 .tc4 ltJf6 12 ltJxf6+ �xf6 1 3 'iWe3 (13 d5 b5 1 4 �b3 c5! and now 1 5 dxe6 c 4 o r 1 5 dxc6 'ilfb6=) 1 3 . . . i.b7 14 l:the l Wile7 1 5 g4!? liad8 16 ltJe5 c5= Stefansson-Bareev, Yerevan 1996; (c) 10 i.d3 is a solid move that aims for an attack that Black musn't underestimate: (c l) 10 . . . b6 1 1 ltJeg5 ! h6? ! (1 1 . . .ltJf6 is solid and natural) 12 �h7+ 'it>h8 1 3 i.e4 ! ! hxg5 ( 1 3 . . . �xg5! 14 ltJxg5 l:tb8 and 15 ltJf3 ltJf6 16 .i.d3 �b7= or 1 5 h4 ltJf6= - Polgar) 14 g4! ( 1 4 �xa8 g4 and . . . .i.g5 if the knight moves) 14 . . . .l:tb8 15 h4 g6 (I doubt that there's a sufficient defence, e.g., 15 . . . gxh4 1 6 g 5 f5 1 7 Wilf4! looks very strong) 1 6 hxg5+ 'it>g7 1 7 Wilf4! i.b7 1 8 l:th7+! �xh7 19 'i!Vh2+ 'it>g8 20 l:thl �xg5+ 2 1 ltJxg5 'ii'x g5+ 22 f4 'iWxf4+ 23 'iVxf4 �xe4 24 'ii'x e4 1-0 J.Polgar-Berkes, Budapest 2003. A superb game, whatever its theoretical significance; (c2) 10 . . . c5! 11 dxc5 'iWc7 1 2 'ilfe2 (12 l:the l ltJxc5 13 ltJxc5 �xc5=) 12 ... b6!? (12 ... ltJxc5=) 1 3 c6!? 'iWxc6 (13 ... ltJb8!) 14 ltJeg5 i.xg5+ 1 5 ltJxg5 ltJf6= Fress inet-Radjabov, Halkidiki 2002. An other critical sequence here would be 1 1 ltJxc5 ltJxc5 12 dxc5 Wild5 13 �b l 'ii'xc5 14 h4 a5!? intending . . . a4 and/or . . . b5; (d) 10 'iWc3!? is Kasparov's inven tion, designed to meet Bareev's fa vourite 9 . . . �e7 when the two met in Sarajevo 2000. It tries to clamp down on both . . . c5 and . . . e5. Although this poses serious problems, Black has a couple of ideas that seem satisfactory.
Classical Variation: 4 i.g5 (4. . . dxe4 5 tDxe4 i.e 7) For example, since White has no threats, a move like 1 0 . . . a5 is plausi ble, with ideas of expansion on the queenside as well as . . . lLlb6-d5. Othewise: (dl) 10 . . . lLlf6 was Bareev's re sponse: 1 1 lLlxf6+ i..xf6 12 i.d3 �d6 13
17 a3 (17 'i£tb l l:tfc8 18 lId2 f6 1 9 lLld3 g5 ! , restricting the knights and preparing the simple minority attack with . . . b5-b4, . . . a5-a4) 1 7 . . . i.b3 1 8 l:td2 litfc8 1 9 c3 b5 20 lLld3 a5 with a powerful minority attack, for exam ple, 21 'i£tb l b4 22 axb4 axb4 23 lLlxb4 .txb4 24 cxb4 lIa8! and wins, since . . J:tca7 follows. Returning to 10 i.c4:
lO a6!? An aggressive move resembling several other lines in this chapter, but retaining the two bishops. (a) 10 ... lLlf6 11 lLlxf6 12 .l:Ihe l fa vours White due to his space and bet ter development. This sort of position is hardly that bad, however, because White's major break with d5 would tend to free both of Black's bishops. In Yermolinsky-Bareev, Lucerne 1997, the play became sharp instead after 12 . . . l:1b8!? 13 lLle5 b5! 14 i.d3! (14 lLlc6 'fid6 15 lLlxb8 bxc4 threatens . . . c5) 14 . . . lIb6!? (14 . . . i.b7) 15 'fie3 and White was still somewhat better; (h) An alternative is 10 . . . c5, which is double-edged yet holding up well: (h I) 1 1 dxc5 1Wc7 12 l:1he l lLlxc5 1 3 lLlxc5 'fixc5 14 i.d3 i.f6 1 5 'fie2 g 6 16 h4 l:t d8 1 7 i..b 5 i.xb2+ Y2 - Y2 Almasi Dizdar, Makarska Tucepi 1995; (h2) 11 d5!? exd5 12 i.xd5 'fic7 ...
253
Play the French with the idea . . . lDb6 - Timman; one would think that White's centralised pieces and control of d5 were vital, but the two bishops seem to hold their own. (b3) the widely-recommended move is 1 1 SLb5 ! , but 1 1 . . :�c7! 12 SLxd7 SLxd7 looks fine, e.g. 13 lDxc5 SLc6 1 4 lDd3 'uac8! 1 5 lDde5 SL d5 with com pensation, or 13 dxc5?! SLc6� with the idea 14 lDd6 b6!+; (b4) 11 'uhel cxd4 12 lDxd4 (12 'ilVxd4 lDb6 1 3 SLb3 'ili'xd4 1 4 'uxd4! SLd7 1 5 lDe5 SLc6 16 lDxc6 bxc6 1 7 c3 c5 1 8 'uddl c4=) 1 2 . . . lDb6 1 3 i.b3 SLd7 1 4 'ilVf4 a5 1 5 'ili'g3!? (this is probably too ambitious, but 15 a3 a4 16 SLa2 'ua5 ! covers a lot of key squares) 15 . . . a4 16 SLxe6 fxe6 17 lDxe6 SLxe6 18 .l:lxd8 SLxd8 (I like Black's bishop pair) 19 lDd6 SLd5 (McDonald analy ses 19 . . . SLxa2!?, which seems very good for Black after 20 b3 axb3 2 1 cxb3 h6, e.g., 2 2 lDxb 7 i.g5+ 23 �dl l1a3-+) 20 'ue8 Fressinet-Radjabov, Pamplona 200 1, and here Radjabov gives 20 . . . i.c7 with advantage. 1 1 i.b3 1 1 'uhe l b5 12 SLd3 SLb7 equalises easily, e.g., 13 �b l i.d5! (13 . . . c5=) 1 4 'ilVf4 c5. 1 1 ... a5! Suddenly changing plans based upon the bishop's position. Black threatens . . . a4. Also good enough is 1 1 . . .c6: 12 'ilVf4 b5 13 h4 (13 lDe5 'ilVc7) 13 . . . .l:.a7 14 lDeg5 lDf6= Khalifman Bareev, Belgrade 1993. 12 SLc4 Alternatives are 12 a3 a4 13 i.a2 lDf6, and 12 a4 lDb6!?, targeting a4. 12 ... c6 13 b 1 'ilVc7 14 h4!? Y2 - Y2 Anand-Bareev, Wijk aan Zee 2003. Here 14 . . . b5 1 5 SLd3 SLb7 could follow, when Black is well enough developed to resist any attack. 254
14.2 9 ... b6 10 SLc4 10 'ilVf4 SLb7 1 1 h4 c5 12 SLb5!? cxd4 13 SLxd7 SLxe4 14 'ilVxe4 'ilVxd7 15 g4 (15 lDxd4 'ilVa4 16 �b l ,Ufd8 17 c3 l:[ac8�) 15 . . . g6 16 g5 i.g7 17 h5 l:[fd8 18 hxg6 hxg6 19 'i!Vh4 'ilVd5 Gallagher Eliet, Clermont-Ferrand 2003, and White seems to lack any compensa tion whatsoever. 10 ... SLb 7!?
Black should certainly consider Dreev's slow move 1O . . . c6 to stabilise the centre and prevent d5. Then if White tries to eliminate Black's bishop pair by 1 1 lDd6 'ilVc7 12 lDxc8 'ufxc8, he has only encouraged Black to attack on the queenside. Better is 1 1 lDxf6+ 'ilVxf6 12 .l::i.he l h6 13 'ilVe3 SLb7 with . . . c5 shortly. 1 1 d5!? An ambitious move, and the only try for a serious advantage. Others allow Black to consolidate: (a) 1 1 l:lhe l SLd5!? (probably okay, but 1 1 . . .i.xe4! 12 .l:!.xe4 c6 would be safe and solid; . . . b5, . . . a5, and . . . lDb6 could follow) 12 SLd3 (12 SLxd5 exd5 13 lDg3 Vallejo Pons-Radjabov, Leon 200 1 ; 13 . . . g6!= 14 lDe5 lDxe5 15 dxe5 SLg7 1 6 'ilVxd5 'ilVg5+ 1 7 �b l .l:.ad8= ; 12 'ilVd3 c6 1 3 SLb3 SLxb3 14 'ilVxb3 'ilVc7=) 12 . . . c5!? 13 c4 i.b7!? (or 13 . . . SLc6, when 14 lDeg5 h6! looks fine
Classical Variation: 4 .i.g5 (4. . . dxe4 5 tbxe4 iL.e 7) and otherwise the bishop is slightly better placed) 14 dxc5 bxc5 1 5 ttJxf6+ and White got a small but lasting edge after 15 . . . 'iVxf6?! in Hubner Short, Novi Sad 1990; but there seems nothing wrong with the active 1 5 . . . ttJxf6! intending . . . �xf3 or . . :ifb6, e.g., 16 ttJe5 (16 'ii'f4 'ii'a 5; 16 ttJg5 'ii'c 7) 16 . . . 'ii'e 7 17 g4!? l:.fd8 18 'ii'c 2 g6 19 g5 (19 f4 .l:td4!) 19 . . . ttJd7= ; (b) 1 1 'iVf4 iL.d5 (1 1 . . .'iVe7!?) 12 �d3 iL.e7!? (12 . . . c5! with the idea 13 c4 i.c6 14 dxc5 bxc5:j:) 13 c4 .i.b7 14 g4!? ttJf6 1 5 ttJxf6+ �xf6 16 .i.e4 iL.xe4 17 'ikxe4 'fie7 1 8 h4 'i'b4! 19 'iVc2 'iVd6 20 'iVe4 'ilb4= Bologan-M.Gurevich, Bel fort 1998. 1 l ... b5!? Also satisfactory is 1 1 . . .e5 (Hub ner) 12 g4!? (12 .l:the l a6!=) 12 . . . a6 13 g5 �e7 - McDonald. 12 i.b3 12 dxe6? bxc4+ ; 12 .i.xb5 exd5 13 ttJxf6+ ttJxf6 14 .l:thel c5. 12 ... c5! The 'I' is partly for ingenuity. Safer options are 12 . . . ttJb6 1 3 ttJxf6+ 'iVxf6 14 dxe6 fxe6= and 12 . . . exd5 1 3 i.xd5 c6 14 �b3 c5 15 i.d5 (15 'ii'x d7 .i.xe4 16 'ii'xb5 'ii'c 7) 15 . . . :b8 unclear M.Gurevich; compare the game. 13 ttJd6! White loses after 1 3 dxe6? .i.xe4 14 exf7+ �h8 1 5 'iVxd7 'fixd7 16 l::t x d7 c4 and lacks enough compensation after 13 ttJxf6+?! ttJxf6 14 dxe6 i.xf3 15 exf7+ Wh8 16 'iYxd8 l:1axd8 1 7 ':xd8 Ihd8 18 gxf3 c4. 13 ....i.xd5 14 iL.xd5 exd5 15 'iVxd5 ttJb6! 16 'iVe4 16 'fixc5? ttJa4 is awful, as is 16 'iVc6? i.xb2+ 17 �xb2 'iVf6+. But a serious alternative is 16 'fif5!? as played in Almasi-Tukmakov, Pula 200 1 . Because of the number of ear lier paths to equality, I'm not going to
analyse this deeply but I think that it's equal. Here's one line of many: 16 . . . g6 (16 . . . ttJa4 also seems fine, e.g., 1 7 ttJe5 'ike7 1 8 .l:the l litad8! 19 ttJc6 'ikxd6 20 .l:hd6 i.xb2+ 21 �b l lixd6:j:) 17 'iVf4 and now easiest was 17 . . . ttJa4! 1 8 c3 (18 ttJc4 bxc4 1 9 .l:txd8 �xb2+ 20 �b l .l:taxd8 21 'ikxc4 l:.b8=) 18 . . . �g7 19 l:.he l (19 ttJe5 'iVc7) 19 . . . 'iVa5= .
16 ... ttJa4 16 . . . 'iVd7!? is also fully equal. Mter 16 . . . ttJa4, Van den Doel-M. Gurevich, Hoogeveen 1999 went 17 ttJe5 'ilVb6?! (17 . . . 'iVa5! 1 8 f4 c4 1 9 �b l ttJxb2! 20 �xb2 'ilVb4+=) 18 'iVd5 (18 ttJdxf7? 'fie6-+) 18 . . . :ad8 19 f4! (19 ttJexf7? �xb2+ 20 �d2 'iVa5+ -+) 19 . . . c4 20 �b l 'iVc7 2 1 .l:the l (2 1 ttJxb5 .l:txd5 22 ttJxc7 J:[xdl + 23 .l:txdl .l:tb8) with a small edge for White.
In this chapter we have seen just about every possible type of idea stemming from the 4 . . . dxe4 and 6 . . . .i.xf6 system. There are a lot of options for Black at almost every move, so this can be a particularly flexible and practical system. Both sides can steer away from tactical play if they wish, so one should be willing to play somewhat quieter po sitions than are produced, for exam ple, by the Winawer Variation. 255
Chapter Fifteen Odds and Ends
1 e4 e6 I n this chapter I will briefly discuss White's irregular second and third moves. I've had relatively little feed back on these moves so I will mainly redo what was in the second edition in order to conserve space for the rest of the book. One advantage of the French Defence is that most of White's moves other than 2 d4 are harmless. Of what follows, only 2 liJf3 d5 3 liJc3 presents a serious challenge to Black. The main White third move after 1 e4 e6 2 d4 d5 is 3 .i.e3 (a gam bit) . I will organise as follows: 1 5 . 1 2 b3 15.2 2 liJf3 15.3 2 d4 d5 3 Sl. e3
It's worth mentioning that Black generally has two or more good an swers in addition to the one I've pre sented. Here are some rare second moves: (a) 2 e5 c5 (2 . . . d6 3 exd6 i..x d6 4 d4 liJc6 5 liJf3 liJge7 6 .i.d3 e5= ; 2 . . . d5 3 liJf3 c5 4 b4 is given under the order 2 liJf3 d5 3 e5 etc.) 3 f4 liJc6 4 liJf3 liJh6 (4 . . . d6=) 5 g3 (Steinitz-Mason, Vienna 256
1 892) 5 . . .liJf5 6 .i.g2 d6= ; (b) 2 Sl.b5 c6 3 Sl.a4 d5 4 'iVe2 'iWa5! 5 Sl.b3 liJf6 6 e5 liJfd7 7 c3 c5 8 d3 liJc6 9 f4 liJd4! 10 'iWdl liJxb3 1 1 'iWxb3 c4! 12 dxc4 liJc5:j: Wahls-Vaisser, Berlin 1 988; (c) 2 f4 d5 3 e5 (3 liJc3, and 3 . . . d4 4 liJe2 c5= or 3 . . . dxe4 4 liJxe4 c5 5 liJf3 liJc6 6 .i.b5 i.. d 7 7 'ife2 liJh6 8 0-0 .i.e7 9 d3 0-0=) 3 . . . c5 4 liJf3 liJc6 5 c3 (5 c4 d4 6 d3 liJh6=) 5 . . . d4 (or 5 . . . liJh6 in tending 6 d4 'iWb6) 6 d3 liJh6= Weiss Maroczy, Budapest 1895; (d) 2 c4 d5 (2 ... c5 is a Sicilian De fence) 3 exd5 (3 cxd5 exd5 4 'iVa4+ .i.d7 5 'iWb3 liJc6! 6 'iVxd5 liJf6 7 'iVc4 .i.e6 8 'iVa4 Sl.c5+ - ECO) 3 . . . exd5 4 cxd5 (4 d4 is an Exchange Variation) 4 . . . liJf6 5 liJc3 (5 ..tb5+ liJbd7 6 liJc3 .i.e7 7 liJge2 0-0 8 0-0 a6 9 .i.xd7 'iWxd7 10 liJf4 b5 11 'iWf3 Sl.b7 12 a3 a5+ Grob-Johner, Zurich 1 941) 5 . . . liJxd5 6 i.. c 4 liJb4! 7 liJf3 liJd3+ 8 �fl i.. e 7 9 'iVe2 liJxc l 10 lIxcl O-O:j: Velimirovic-P.Nikolic, Novi Sad 1 984; (e) 2 g3 d5 3 liJc3 (3 ..tg2 c5= or 3 . . . dxe4 4 liJc3 i.. d 7!= and . . . ..tc6, or 3 . . . ttJf6 4 e5 liJfd7 5 d4 c5 6 c3 'iVb6 7 liJf3 liJc6 8 0-0 ..te7= intending . . . 0-0, . . . f6 - ECO) 3 . . . dxe4 (3 ... liJf6 4 e5
Odds and Ends lLlfd7 5 f4 c5 6 lLlf3 lLlc6 7 .tg2, and 7 . . . JJ.. e 7 or 7 . . . a6 8 a4 JJ.. e 7 9 0-0 0-0 10 d3 lLld4 11 lLle2 lLlxf3+ 12 JJ.. xf3 f6= Vavra-Totsky, Leonid 1995) 4 lLlxe4 lLlf6 (4 . . . JJ.. d 7 5 lLlf3 JJ.. c 6 6 'ii'e 2 lLlf6= Bartsch-Reefschlaeger, Bundesliga 1981) 5 lLlxf6+ 'iVxf6 6 JJ.. g2 JJ.. c 5 7 'i!i'f3 'iVe7 ! 8 lLle2 e5=i= Suttles-Uhlmann, Palma de Mallorca 1970; (f) 2 lLle2 d5 3 exd5 exd5 4 lLlg3 lLlf6 (4 . . . 'it'h4!? - ECO) 5 d4 .td6 6 .td3 0-0 7 0-0 lIe8= Tartakower-Fine, Kemeri 1937; (g) 2 lLlc3 d5 tranposes to the main line after 3 d4, or to 2 lLlf3 below after 3 lLlf3, or to note 'e' after 3 g3. 1 5 . 1 2 b3 A safe move, but b3 and e4 do not mix very well.
.tf5= Sandmann-Rausch, Bundesliga 1989) 6 JJ.. e 2 0-0 7 0-0 .tf5 8 .:te l c5 9 d4 lLlc6 10 lLlbd2 lIc8=i= (c-file and ac tivity) Castro-Petrosian, Bie1 1976. (b) 4 e5 lLlfd7 5 f4 c5 6 lLlf3 (6 lLlc3 lLlc6 7 lLlf3 a6 8 g3 b5= Tartakower Alekhine, Kemeri 1937) 6 . . . b6!? (or 6 . . . JJ.. e 7 7 �e2 lLlc6=) 7 d4 JJ.. a 6 8 �xa6 lLlxa6 9 'i!i'e2 lLlb4 10 0-0 JJ.. e 7 1 1 h1 .l:1.c8 1 2 lLla3 0-0 1 3 c 3 lLlc6 14 lLlc2 cxd4 1 5 lLlfxd4 lLlc5 16 .:tf3 lLlxd4 17 cxd4 lLle4 and Black dominates the light squares, E .Hummel-Archan gelsky, Wijk aan Zee 2000. 3 .tb2 JJ..b 7 4 lLlc3 The e-pawn is awkward to defend: 4 'i!i'e2 lLlf6 5 lLlc3 c5 with the idea . . . lLlc6 is equal, as is 4 d3 d5. 4 ... c5 Or 4 . . . d5 5 exd5 exd5 6 d4 lLlf6= . 5 lLlf3 lLlc6 6 d4 cxd4 7 lLlxd4 lLlxd4 8 'iVxd4 lLlf6 Black shouldn't delay his develop ment with moves like . . . a6 and . . . d6. 9 f3?! 9 ii.e2 JJ.. c 5 10 'ii'd 3 0-0= . 9 .tc5 10 'ii'd 2 0-0 l l lLla4 JJ.. e 7 Bury-J.Watson, Leominster 1977. Black has the idea 12 0-0-0 lLlxe4! and is better developed. •••
15.2 2 lLlf3 d5 and now: 2 ... b6 Trying to force White to defend e4. Perfectly good is the normal 2 . . . d5 3 .tb2 (3 exd5 exd5 simply frees Black's pieces) 3 . . .lLlf6 (3 . . . dxe4 is fine, but there's a fair amount of theory) and: (a) 4 exd5 exd5 (4 . . . 'ii'x d5 5 lLlc3 'iVa5=) 5 lLlf3 (5 'iVe2+ JJ.. e 6 6 'i!fb5+ lLlbd7 7 'i!fxb7 .tc5! with strong com pensation, one point being 8 .txf6 'iVxf6! 9 Vi'xa8+ rJite7 10 'iVxh8 'iVxf2+ 1 1 d1 'iVxf1 mate) 5 . . . .te7 (5 . . . .tc5 6 d4 .tb4+! 7 c3 .td6 8 JJ.. e 2 0-0 9 0-0
15.21 3 lLlc3 1 5.22 3 e5 15.21 3 lLlc3 lLlf6 3 . . . d4 is also fully satisfactory. 4 e5 4 "iVe2 JJ.. e 7 achieves nothing, and 4 exd5 exd5 5 d4 is an inferior type of Exchange Variation: 5 . . . JJ..b 4 (5 . . . c6 and; 5 . . . .te7 are equal) 6 a3 JJ.. xc3+ 7 bxc3 .tg4 (7 . . . 0-0! 8 JJ.. e 2 .:te8 avoids 8 'iVe2+) 8 .te2 (8 'i!i'e2+! 'i!i'e7 9 'ii'x e7+
257
Play the French rJ;;;x e7 10 ltJe5 SLf5 1 1 SLd3!) 8 . . . 0-0 9 0-0 ltJe4 (or 9 . . . ltJbd7!?+ with the idea . . . ltJb6) 10 SLb2 ltJd7 1 1 c4 Bibiloni Bosco, Buenos Aires 1993; and very strong was 1 1 . . .dxc4! 12 SLxc4 (12 c3 ltJb6) 12 . . . SLxf3 13 gxf3 (13 'iVxf3 ltJd2) 13 . . . ltJd6 14 SLb3 "iVh4+ . 4 ...ltJfd7 5 d4 c5
Black can also wait for White to commit by 5 . . . SLe7, e.g., 6 i.d3 (6 ltJe2 c5 7 c3 ltJc6 8 1tJf4 - Minev, but 8 . . . cxd4 9 cxd4 'iWb6 threatens . . . g5 and then . . . SLb4+) 6 . . . c5 7 0-0 (7 dxc5 ltJxc5=) 7 . . . ltJc6 8 SLe3? (8 dxc5 ltJxc5=) 8 . . . cxd4 9 SLxd4 ltJxd4 10 ltJxd4 ltJxe5 1 1 SLb5+ i.d7 12 Itel ltJc6+ Fleger Sermier, Bie1 1 993. 6 dxc5 (a) 6 ltJe2 transposes to Chapter 13, 5 ltJf3; (b) 6 i.b5 has more than one an swer, but the most important is 6 . . . ltJc6! 7 0-0 cxd4 8 ltJxd4 ltJdxe5! (8 ... 'iVc7 9 SLf4!) 9 f4 (9 SLf4 SLd7 10 Itel ltJg6+ ; 9 Itel SLd6! 10 f4! ltJd7 1 1 ltJxd5 ltJxd4 1 2 'iVxd4 0-0 1 3 ltJe3 'iVc7+ ; 9 'iVh5?! SLd6 10 SLg5 'iVc7+) 9 . . . a6! 10 SLa4 ltJd7! 1 1 f5 (1 1 ltJxd5?? ltJxd4 12 'iVxd4 b5-+) 1 1 . . .ltJxd4 12 SLxd7+ SLxd7 13 'iVxd4 .l:[c8+ was D . Schneider-J.Watson, Chicago 1992 . 6 ... ltJc6 7 i.f4 7 ..tg5 i.e7 8 ..txe7 'iVxe7 9 SLb5 258
ltJxc5 10 0-0 0-0 11 Itel SLd7= Ge lashvili-Radj abov, Halkidiki 2002 . 7 ... SLxc5 8 ..td3 f6 9 exf6 White cannot maintain the pawn on e5: (a) 9 SLb5? fxe5 10 ltJxe5 (10 SLxe5 O-O! 1 1 SLg3 d4! 12 SLxc6 dxc3 1 3 SLxd7 cxb2 14 Itb l ..tb4+ 1 5 rJ;;; e 2 SLxd7-+) 10 . . . ltJdxe5 11 SLxe5 SLxf2+! 12 rJ;;;xf2 0-0+ and . . . ltJxe5; (b) 9 'iVe2 fxe5 10 ltJxe5 ltJdxe5 1 1 ..txe5 ltJxe5 1 2 'iVxe5 0-0 1 3 0-0 '(i'f6 (or 1 3 . . . 'iVh4!? 14 g3 '(i'd4) 14 'iVxf6 .l:[xf6+ with two bishops and the mo bile centre, e.g., 15 g3 ..td7 16 rJ;;; g2 g5! 17 Itael Itaf8 18 f3 ..tc6+ . 9 ... ltJxf6 10 0-0 0-0
A key position for this variation. White's control of e5 and the e-file are balanced by Black's central majority and f-file pressure. Unquestionably White has the harder time of it in practice. We have: 15.2 1 1 1 1 ..tg3 15.212 11 ltJe5
Others: (a) 1 1 'iVe2 ltJh5! (or 1 1 . . . ..td7 12 Itadl a6= 13 ltJa4?! SLa7 14 c4?? e5!+ 1 5 cxd5 ltJd4! 1 6 ltJxd4 SLxd4-+ Rigo Gulko, Rome 1988) 12 SLg5 ltJf4! 1 3 'iVd2 'iVc7 14 ltJb5 'iVb8 1 5 c4? (15 SLxf4
Odds and Ends l:txf4 16 l1ael �dn) 1 5 . . . l2Jxd3 16 'it'xd3 l2Je5 1 7 l2Jxe5 'it'xe5 18 iLh4 a6! 1 9 l2Jc3 l£'f4+ Larsen- Spassky, Stock holm 1 969; (b) 11 l2Je2 "ile7 (or 1 1 . . .l2Jh5= 12 �g5 "ild6) 12 l2Je5 l2Jxe5 1 3 �xe5 l2Jg4 14 �g3 l£'f6! hitting b2 and f2 and giving White problems; (c) 11 'it'd2 'it'e8 (objectively 1 1 . . .a6 is probably as good e.g., 12 l:.ae l l2Jh5 13 �g5 l£'c7 14 �h4 l2Jf4 15 iLg3 .i.d6 16 �xf4 �xf4 17 'it'dl .i.dn Rigo H.Meyer, Delmenhorst 1986) 12 ':ael (12 l2Jb5 l2Jh5 1 3 iLd6 [13 l2Jc7? 'iVd8] 13 . . . .i.xd6 14 l2Jxd6 "ile7 15 l2Jxc8 l:.axc8 with the ideas . . . e5, . . . l2Jb4 Nunn) 12 . . . �d7 1 3 a3!? (13 l2Je5= Nunn; 13 . . .l2Jxe5 14 �xe5 "ilh5;) 13 . . . 'iVh5 14 b4 �b6 15 .i.d6!? l2Jg4! with a powerful attack, Rigo-Klinger, Vienna 1986; (d) 11 l2Ja4 �d6 12 .i.xd6 "ilxd6 13 c4 .td 7 14 l2Jc3 l:.ac8= . 15.2 1 1 1 1 �g3 This is very natural and considered a safe option. But remarkably it seems almost impossible for White to even equalise. 11 ... l2Jh5! 12 �h4 (a) 1 2 'iVe2 l2Jxg3 13 hxg3 �d7 (13 . . . 'it'f6!?;) 14 l2Ja4 .i.d6 15 c4 l2Jb4! 16 l2Jc3 l:.c8; ; (b) 12 .te5? l2Jxe5 1 3 l2Jxe5 l2Jf4 14 l2Je2 "ilg5 1 5 l2Jxf4 "ilxe5 1 6 g3 'iVg5 17 'it'c l e5!+. (c) 12 �h4 'it'c7! 1 3 l2Jb5 'it'd7! 14 l2Jg5 (14 l:.el l2Jf4 15 l2Jg5 h6! 16 l2Jh7 ':f7+) 14 . . . g6 15 c3 (15 l:tel l2Jf4 16 .tg3 e5+) 15 . . . e5 (15 . . .l2Jf4 16 �c2 a6 17 l2Ja3 b5;) 16 �c2 l:.f4!? 17 .i.g3 (17 g3 Iixh4! 1 8 gxh4 l2Jf4+) 17 ... l2Jxg3 18 hxg3 l:.g4 1 9 l2Je6!? (19 l2Jf3 e4!? 20 l2Jfd4 a6 2 1 l2Ja3 'iVd6+) 19 ... "ilxe6 20 l2Jc7 'it'd6! 2 1 l2Jxa8 e4+ 22 b4 'it'xg3 23 'iVxd5+ �h8 24 "ilxe4 l2Je5-+ .
15.212 1 1 l2Je5 .td7
The traditional move. 1 analysed 1 1 . . . "ile8!? in the last edition, and A.Bennett-J.Watson, Seattle 2003 took an interesting course after 1 1 . . . a6!? (I like this move) 12 'iVe2 "ile8 (12 . . . l2Jxe5 13 .i.xe5 b5!? 14 l1ae l .l:.a7!? 1 5 'it'hl l:.af7! 16 f4 �b7 and . . . d4) 1 3 l2Jxc6 (13 l2Ja4?? l2Jd4) 13 . . . bxc6 14 l2Ja4 .i.a7 1 5 �d6, and although Black eventually mobilised his central pawns and won after 15 . . . .:f7= , 1 might have tried my original intention 1 5 . . . c5!? 16 l2Jxc5 (16 b3 c4) 16 . . . e5! 17 i.. xf8 (17 l2Jxa6 e4 18 �b5 'it'g6!?) 17 . . . 'it'xf8 18 l2Jxa6 'it'd6 with a fascinating attack. 1 2 l2Jxc6 There are several options, none impressive: (a) 12 a3 a6 (12 . . . l2Jxe5 1 3 .i.xe5 'iVb6 14 'iVe2 �d6;) 13 l2Jxc6 .i.xc6 14 'iVe2 'iVe7! 1 5 l:.ael l:.ae8 16 .te5 l2Jd7! 1 7 "ilh5 l:.f5 ! ! 1 8 �xf5 exf5 1 9 �hl (19 .i.g3 'iYxel) 19 ... l2Jxe5 20 f4 1£'f7+ ; (b) 12 'iYf3 !? can b e answered by 12 . . . .td4! 13 l2Jxc6 .i.xc6 14 l:.ael e5! 1 5 �xe5 .txe5 16 llxe5 d4 1 7 �c4+ �h8 1 8 l2Jd5 ':c8! 19 'iVb3 b5- + ; (c) 12 'iWe2 'it'e7 1 3 I:[ae l l:.ae8 14 a3 (14 i.. g 3 is recommended by Baker; then some interesting play can follow 14 . . . a6!? 15 a3 b 5 ! , e.g. , 16 l2Jdl l2Jxe5 259
Play the French 17 iLxe5 �c6 18 c3 ltJe4!? 19 ltJe3 'ilVg5!? 20 f4 'ifg6) 14 . . . a6 15 'it>h1 (15 ..tg3 ltJxe5 16 'iWxe5!? �c6 1 7 b4 iLa7 with the idea . . .�b8 and a small ad vantage, Drozdov-Hoang Than Trang, Budapest 1 994) 15 . . . �d4! 16 ltJxd7 'ilVxd7 17 iLd2 e5:j: Ljubojevic-Petro sian, Las Palmas 1973. 12 ... �xc6 13 'ilVe2
1 3 .. JWe 7 Or 13 . . . ltJe4!? 14 g3 (14 ..te3?! �xe3 15 'i!lxe3!? d4 16 'ife2 dxc3 17 �xe4 cxb2 18 I1ab 1 'iWf6 19 c4 lIab8!+ Bel lon-Speelman, Amsterdam 1978) 14 . . . ltJxf2 15 I1xf2 �xf2+ 16 'itxf2 Vogt-Farago, Kecskemet 1979, and now 16 . . . g5! 17 ii'xe6+ 'itg7 unclear Vogt; but this seems to favour Black, e.g., 18 'i!lg4 �h8!? 19 'i!fh5 ii'e7 20 %:tel 'ilVg7 with advantage. 1 4 l:[ae l I1ae8 15 �e5!? 1 5 �g3 a6 16 a3 'iVf7!? 1 7 b4 �d4 18 �e5 �xe5 19 'i!lxe5 ltJd7 20 'iVg3 e5:j: Spassky-Petrosian, Moscow 1966. 15 ...ltJd7 1 6 ltJb5 16 'i!lh5 g6! 17 ..txg6 ltJxe5 18 iLxe8 �f5! 19 I1xe5 %:txh5 20 iLxh5 iLd6 2 1 l:.e2 'i!lh4 22 h 3 ii'xh5 23 %:txe6 iLb4-+ . 1 6 ... a6!? 16 . . . ltJxe5!? 17 'iVxe5 %:txf2! 18 l:[xf2 �xf2+ 19 �xf2 .txb5 20 �xb5 ii'c5+ 21 I1e3 'ilVxb5:j:, but Black probably can't win this.
260
After 16 . . . a6, Shilov-Ulibin, Bar linek 2001 went 17 ltJd4 ltJxe5 1 8 'ilVxe5 iL d 7 1 9 f4 �d6 20 iVh5? (20 'ilVe3 e5 21 fxe5 i.xe5 22 'ith1 !=) 20 . . . g6 2 1 'ii' g4 e5! 22 'ilVg3 and one win was Bangiev's 22 . . . e4 23 .te2 'iff6. 15.22 3 e5 c5 4 b4!? This is the French Wing Gambit, trying to win the centre by diverting Black's c-pawn. It has never attracted much attention from strong players and in my opinion it is suspect. Some passive alternatives: (a) 4 d3 ltJc6 5 c3 (5 i.e2 Vi'c7 6 �f4 f6) 5 . . .f6 6 d4!? and Black is a tempo ahead of the Advance Variation: 6 . . . i.. d 7!? (6 . . . fxe5; 6 . . . 'iVb6!) 7 g3 I1c8 8 �g2 cxd4 9 cxd4 �b4+ 10 .i.d2 'ii'b 6 11 a3? i.xd2+ 12 ii'xd2 ltJa5-+ Hamadto-Rahman, Dubai 1 986; (b) 4 g3 ltJc6 5 �g2 g5 !? 6 h3 ..tg7 7 Vi'e2 h6 (7 . . . 'ilVc7!:j:) 8 0-0 'i!lc7 9 .l:!.e1 ltJge7 10 ltJa3 a6 11 c3 c4! 12 b3 b5:j: with the idea ... ltJg6 Gatica-Levitt, New York 1 994. 4 ... cxb4
5 a3 5 d4 is also played: (a) 5 . . . i.d7 (Harding gives this '!') 6 a3 (6 i.d3 has several answers, e.g., 6 ... 'iVb6 intending ... �b5 or 6 ... ltJe7 7
Odds and Ends a3 lLlbc6 8 axb4 lLlxb4:j: Barendregt Bronstein, Hamburg 1965) 6 .. :iWa5 !? 7 ii.d3 i.b5 ('This is a hard line for White to play' - Harding) 8 0-0 (8 axb4? 'ii'xal 9 i.xb5+ lLlc6 10 i.d2 'ii'b 2 11 c3 a5+ Day-Hubner, Jerusa lem 1967) 8 . . . i.xd3 9 'ii'x d3 'ii'a 6! 10 'ii'b 3 lLlc6 1 1 'ii'b 2 'ii'b 6!? 12 .l:.dl Ved der-Poulton, Gausdal 1990; and here 12 . . .f6 was one of several moves, still a pawn up; (b) 5 . . . lLlc6 6 a3 and: (h I) 6 . . . bxa3 7 c3 lLlge7 (7 .. .f6 8 i.d3 'ii'c 7) 8 i.xa3 lLlg6!? 9 i.xf8 (9 h4 i.e7 10 h5 lLlh4 1 1 h6 lLlxf3+ 12 'i¥xf3 g6:j:) 9 . . . lLlxfB 10 .itd3 f6 1 1 lLla3 fxe5 12 lLlxe5 lLlxe5 13 dxe5 lLlg6!? 1 4 'i¥h5 0-0 15 0-0 (15 .itxg6 hxg6 16 'iWxg6 'ii'c 7) 15 . . . i.d7 16 f4 .l:.c8+ Bosch Blees, Dieren 1 988; (h2) 6 .. .f6 7 axb4, and 7 . . . i.xb4+ 8 c3 i.a5 intending . . . i.c7 or 7 . . . fxe5 8 lLlxe5 lLlxe5 9 dxe5 'ii'c 7 (9 . . . i.xb4+ 10 c3 i.c5+) 10 'iWd4 lLle7+ .
5 ...lLlc6 5 . . . d4!? was given in the last edi tion but is too worked out now. 5 . . . lLlh6!? is a clever idea played in C.B.Baker-Harding, corr 1 986: 6 axb4 i..xb4 7 c3 .ite7 (the point: . . . lLlh6 de veloped, but left e7 free for the bishop) 8 d4 lLlf5 9 lLla3 (Harding pre fers 9 i.d3) 9 . . . .td7!? (9 . . . lLlc6 10 i.. d3
f6!+) 10 i.d3 lLlc6 1 1 g4 lLlh4 12 lLlxh4 i.. x a3?! (12 . . . i..xh4:j:) 13 i.xa3 'i¥xh4 14 .l:.b l with some compensation. 6 axb4 i.xb4 7 c3 i.e7 7 ... i.a5 8 d4 f6 is also good, but re quires more accurate play, as Hard ing has shown. 8 d4
8 ... f6! Or the creative 8 ... lLlh6! 9 .txh6!? (9 .itd3 lLlf5) 9 . . . gxh6 10 i.d3 f6! 1 1 Wc2 fxe5 12 dxe5 'ii'c 7+ 1 3 0-0 i.c5! 14 lla4 i.d7 1 5 i.g6+ e7! 16 .txh7 lLlxe5+ Lanzani -Vezzosi, Chianciano Terme 1 990. 9 i.. d 3 fxe5 10 lLlxe5 10 dxe5 'ii'c 7 11 0-0 lLlxe5 12 lLlxe5 'i'xe5 13 .tb5+ f7!+ Hamed-Garma, Novi Sad 1 990. Mter 10 lLlxe5, Reindermann-Glek, Groningen 1992 continued 10 . . . lLlf6 1 1 i.g5 lLlxe5 12 dxe5 lLle4! 1 3 i.xe7 'i'xe7 14 i.. xe4 dxe4 1 5 'iVd4 0-0 16 'iix e4 (16 :xa7 llxa7 1 7 'iWxa7 'iVg5+ was the game Buturin-Glek, Belgorod 1989) 16 . . . i.. d 7! 17 'iWe3 (17 'iVxb7 'iVc5 !) 17 . . . i..b 5 18 lLld2 'i'c7 19 f4 .l:.ad8+ intending . . . .l:.d3. 15.3 2 d4 d5 3 i.e3 This is called the 'Alapin-Diemer Gambit' by Harding. The idea after 3 . . . dxe4 is to get into a sort of Black-
261
Play the French mar-Diemer Gambit (1 d4 d5 2 e4 dxe4 3 f3, or 2 liJc3 liJf6 3 e4 dxe4 4 f3) , normally playing f3 early on. The drawback to this strategy is that the bishop on e3 would be poorly placed in a Blackmar-Diemer (an opening of marginal soundness anyway) . Never theless the line has its defenders and White with the extra tempo can get away with more in the opening than Black. The only other serious 3rd move options are: (a) 3 .td3 (An innocuous move that seeks to avoid risk) and: (al) 3 . . . c5 4 c3 liJc6 (4 . . . cxd4 5 exd5 exd5 6 cxd4 liJc6 7 i.e3 .tb4+ 8 liJc3 liJge7 9 a3 i.xc3+ 10 bxc3 .tf5= Ma roja-Kovacevic, Zadar 2000) 5 liJe2 cxd4 6 cxd4 liJb4!? 7 .tb5+ i.d7 8 i.xd7+ 'iVxd7 9 e5 liJe7 10 liJbc3 liJf5 1 1 a3 liJc6 (a sort of Advance Varia tion with the light-squared bishops off) 12 b4 i.e7 13 0-0 'uc8 14 1i¥d3 f6 1 5 g4 liJh4 16 f4 f5 ! 1 7 h3 fxg4 1 8 hxg4 h5! 1 9 gxh5 liJf5+ Bluvshtein Degraeve, Montreal 2002; (a2) 3 . . . dxe4 4 �xe4 liJf6
and now: (a2 1) 5 i.g5 c5! 6 �xf6 (6 c3 cxd4 7 cxd4 'iVb6! 8 iLxf6 �xb2 9 liJd2 gxf6) 6 . . . 'iVxf6 7 liJf3 liJc6 8 c3 cxd4 9 i.xc6+ bxc6 10 cxd4 l:!.b8 1 1 'i!i'd2 c5=i=; (a22) 5 .td3 c5! 6 liJf3 (6 dxc5 .txc5 262
7 liJf3 liJc6 8 0-0 0-0 9 liJc3 liJb4!=i= Rivera-P.Hummel, Oropesa del Mar 1998) 6 . . . liJc6 7 c3 �e7 (7 . . . cxd4=) 8 dxc5 .txc5= Tartakower-Trifunovic, Saltsjobaden 1 948; (a23) 5 �f3! (at least this is active) 5 . . . c5 6 liJe2 (6 c3 liJc6 7 liJe2 cxd4 8 cxd4 .te7 9 liJbc3 0-0 10 0-0 'iVb6=) 6 ... liJc6 7 .te3 cxd4 (7 . . . e5!? 8 i.xc6+ bxc6 9 dxe5 'i!i'xdl + 10 �xdl liJg4= Suechting-Alapin, Barmen 1 905) 8 .txc6+?! (8 liJxd4 liJe5!?) 8 . . . bxc6 9 'i!i'xd4 1i¥xd4 10 i.xd4 c5! 1 1 .txf6 gxf6=i= with open lines and two bishops that compensate for Black's weak nesses: 12 liJbc3 ,Ug8 13 liJe4 �e7 14 0- 0 �b7 1 5 f3 0-0-0 16 'uadl f5! 1 7 liJ4g3 c4=i= Bezgodov-Hanley, Hoogev een 2002; (b) 3 c4?! dxe4 is a dubious gambit, e.g. , 4 liJc3 liJf6 (or 4 . . .f5 5 f3 liJf6 6 fxe4 fxe4=i=) 5 i.g5 (5 f3 c5 6 d5 exd5 7 cxd5 exf3 8 liJxf3 iLd6+ - ECO) 5 . . . iLe7 6 f3 liJc6! 7 d5 (7 liJge2 exf3+) 7 . . .exd5 8 i.xf6 (8 cxd5 liJxd5+) 8 ... iLxf6 9 cxd5 liJd4 10 liJxe4 0-0+ ; (c) 3 f3 c5 (3 . . . liJf6 4 e5 liJfd7 5 f4 c5=) 4 iLe3 liJc6 5 c3 dxe4 6 fxe4 liJf6 7 e5 liJd5 8 .tf2 cxd4 9 cxd4 Stanka Lalic, Oberwart 200 1 ; and one path was 9 . . . iLb4+ 10 liJd2 f6 1 1 exf6 liJxf6 12 liJgf3 (12 a3 iLxd2+ 13 �xd2 liJe4+) 12 . . . liJe4+ . 3 ... dxe4 4 liJd2! 4 f3 liJh6!? threatens . . . liJf5 (or 4 . . . liJf6 5 liJc3 and 5 . . . liJd5 6 liJxd5 exd5 7 1i¥d2 iLd6+ or 5 . . . .tb4 6 a3 liJd5 ! - Minev) 5 'i!i'd2 (5 i.xh6 �4+ is good for Black; 5 fxe4? 'iVh4+ 6 �f2 'i!i'xe4++) 5 . . . liJf5 6 .tf2 c5!=i= with the idea 7 dxc5? 'i!i'xd2+ 8 'iitxd2 e3+! 9 i..x e3 liJxe3 10 'iitxe3 .txc5++ . 4 ...liJf6 5 c3! (a) 5 f3?! liJd5! 6 1i¥e2 liJxe3 7 'iWxe3 exf3 8 liJgxf3 i.e7 or 8 . . . g6, and White lacks compensation;
Odds and Ends (b) 5 .i.c4 ttJc6 6 ttJe2 e5 7 c3 exd 4 8 ttJxd4 ttJe5 9 'i!Vb3 .td6+ Edvardson J.Watson, Gausdal 1980. 5 ...b61?
This has been the most frequent move; also reasonable is getting the
bishop pair by 5 . . . ttJd5 6 ttJxe4 ttJxe3 7 fxe3 g6= . 6 g3 White doesn't get his pawn back af ter 6 'i!Vc2 �b7 7 ttJe2 'i!Vd5 8 g3 �d6 9 �g2 'i!Vf5;. 6 ... �b 7 7 �g2 c51? 7 . . . ttJbd7 8 'i!Vc2 .te7 9 ttJxe4 ttJxe4 10 �xe4 �xe4 1 1 'i!Vxe4 0-0= Mieses Mason, Monte Carlo 1903. After 7 ... c5, play might go 8 dxc5! i.xc5 9 J..xc5 bxc5 10 'iVc2 'i!Vd5 1 1 ttJe2 ttJbd7 1 2 c4 'iYe5 1 3 0-0-0 llb8! (targeting b2 and preparing ... e3; 13 ... 0-0 1 4 ttJc3=) 14 ttJc3 e3! 1 5 ttJde4 .i.xe4 16 ttJxe4 ttJxe4 17 .i.xe4 ttJf6! 18 .i.c6+ c:j;; e 7 and Black is still a pawn up with open lines, e.g., 19 l:thel Wic7 20 i.f3 exf2 21 'iYxf2 l:thd8; .
263
Index of Variations
1 Advance Variation: Introduction 1 e4 e6 2 d4 d5 3 e5 c5
1 . 1 4 dxc5 1 0 1 . 2 4 'iWg4 1 4 1 . 3 4 liJf3 1 8 2 Advance Variation: 5 ...j" d 7 1 e4 e6 2 d4 d5 3 e5 c5 4 c3 liJc6 5 liJf3 �d7
264
Index of Variations 2.1 6 dxc5 23 2.2 6 a3 26 2.3 6 i.. e 2 30 2 . 3 1 6 . . . tiJge7 30 2.3 1 1 7 0-0 30 2 . 3 1 2 7 tiJa3 33 2.32 6 . . . f6 35 3 Advance Variation: 5 .. .1Wb6 1 e4 e6 2 d4 d5 3 e5 c5 4 c3 tiJc6 5 tiJf3 �6
3 . 1 6 a3 tiJh6 7 b4 cxd4 38 3 . 1 1 8 .txh6 39 3 . 1 2 8 cxd4 tiJf5 41 3 . 1 2 1 9 i.. e 3 41 3 . 12 2 9 i..b 2 44 3.2 6 .i.e2 tiJh6 4 7
3.21 7 b 3 4 7 3.22 7 .txh6 49 3.3 6 .td3 cxd4 7 cxd4 .ltd7 8 0-0 tiJxd4 52 3 . 3 1 9 tiJg5 53 3.32 9 tiJxd4 54
265
Play the French 4 King's Indian Attack 1 e4 e6 4 . 1 2 d3 d5 3 i2ld2 5 7 4. 1 1 3 . . . i2lf6 4 i2lgf3 58 4. 1 1 1 4 . . . i2lc6 59 4. 1 1 2 4 . . . i.c5 61 4. 1 1 3 4 . . . b6 63 4 . 1 2 3 . . . c5 4 i2lgf3 i2lc6 5 g3 i.d6 6 i.d2 i2lge7 7 0-0 0-0 65 4 . 1 2 1 8 .l:!.el 66 4.122 8 i2lh4 68 4.2 2 �e2 68
4.2 1 2 . . . i2lf6 69 4.22 2 . . . c5 71 5 Exchange Variation 1 e4 e6 2 d4 d5 3 exd5 exd5
5 . 1 4 c4 73 5.2 4 i2lf3 75 5 . 2 1 4 . . . i.g4 75 5.22 4 . . . i.d6 77 5 . 3 4 i.d3 79
266
Index of Variations 6 Tarrasch Variation: Introduction and 3 ... c5 1 e4 e6 2 d 4 d5 3 l2Jd2 c5
6 . 1 4 l2Jgf3 85 6.2 4 exd5 �xd5 5 l2Jgf3 cxd4 6 i.. c 4 'iVd6 7 0-0 l2Jf6 8 l2Jb3 l2Jc6 9 l2Jxd4 l2Jxd4 10 l2Jxd4 a6 8 7
6 . 2 1 1 1 i.b3 92 6.22 1 1 c3 93 6.23 1 1 b3 94 6.24 1 1 l:.el 'iic 7 12 .tb3 i.. d6 1 3 l2Jf5 SLxh2+ 14 �hl 0-0 1 5 l2Jxg7 .litd8 16 �f3 �xg7 1 7 .th6+ �g6 18 c3 e5 19 i.c2+ e4 95
6.241 20 .txe4+ 99 6.242 20 g3 1 00 6.243 20 l:i.xe4 1 01
267
Play the French 7 Tarrasch Variation: 3 . . . .i.e7 1 e4 e6 2 d4 d5 3 ttJd2 i.. e 7
7 . 1 4 c3 1 04 7.2 4 e5 c5 5 'i'ig4 1 05 7.21 5 . . . g6 1 0 7 7.22 5 . . .';i?f8 1 08 7.3 4 ttJgf3 1 1 1 7.4 4 i.. d 3 c 5 5 dxc5 ttJf6 6 'ii' e 2 1 1 6 7.41 6 . . . 0-0 1 1 7 7.42 6 . . . ttJc6 1 1 9 8 Winawer Variation: Fourth Move Alternatives 1 e4 e6 2 d4 d5 3 ttJc3 i..b 4
8.1 4 a3 i.xc3+ 5 bxc3 dxe4 1 23 8. 1 1 6 f3 1 25 8. 1 1 1 6 . . . ttJd7 1 25 8. 1 1 2 6 . . . e 5 1 26 8 . 12 6 'i'ig4 ttJf6 7 'i!i'xg7 Itg8 8 'iYh6 ttJbd7 1 26 8 . 1 2 1 9 f3 1 2 7 8 . 1 2 2 9 ttJh3 1 28 8.123 9 ttJe2 1 29 8.2 4 'i'g4 1 31 8.3 4 i.. d 2 dxe4 5 'i'ig4 ttJf6 6 'ii'x g7 l:i.g8 7 'i'h6 1 33 8 . 3 1 7 . . . 'i'ixd4 0-0-0 1 34 268
Index of Variations 8.3 1 1 8 . . . lDg4 1 35 8.312 8 . . . .tf8 1 36 8.32 7 J�g6 1 39 8.4 4 lDe2 1 41 . .
8.41 4 . . . lDc6 1 41 8.4 1 1 5 e5 1 42 8.4 1 2 5 a3 ..ta5 1 43 8.4 1 2 1 6 e5 1 43 8.4 122 6 b4 1 44 8.4 1 2 3 6 ..te3 1 45 8.42 4 . . . dxe4 1 4 7 8 . 5 4 .i.d3 1 48 8.6 4 exd5 exd5 1 50 8.61 5 '(Wf3 1 51 8.62 5 .td3 1 51 8.621 5 . . . lDc6 1 51 8.622 5 . . . lDf6 1 53 8.7 4 'i!Vd3 1 54 9 Winawer Variation: Fifth Move Alternatives 1 e4 e6 2 d4 d5 3 lDc3 ..tb4 4 e5 c5
9 . 1 5 lDf3 1 56 9.2 5 ..td2 1 58 9.21 5 . . . lDh6 1 58
269
Play the French 9.2 1 1 6 tiJb5 1 59 9.212 6 .ixh6 1 60 9.2 1 3 6 a3 1 60 9.214 6 .id3 1 61 9.22 5 . . . tiJe7 1 62 9.221 6 tiJb5 .ixd2+ 7 "iVxd2 0-0 1 63 9.22 1 1 8 c3 1 63 9.22 1 2 8 f4 1 64 9.22 1 3 8 dxc5 1 64 9.222 6 a3 1 65 9.223 6 f4 1 66 9.3 5 dxc5 tiJc6 6 tiJf3 tiJge7 7 .id3 d4 8 a3 .ia5 9 b4 tiJxb4 10 axb4 .ixb4 1 1 0-0 .ixc3 1 2 l:rb 1 1 66 9 . 3 1 1 2 . . . "iVc7 1 69 9.32 1 2 . . . h6 1 71 9.4 5 "iVg4 tiJe7 1 72 9.41 6 dxc5 tiJbc6 7 .id2 0-0 1 73 9.4 1 1 8 tiJf3 1 74 9.4 1 2 8 .id3 1 74 9.42 6 tiJf3 1 75 1 0 Winawer Variation: Main Line with 7 "iVg4 1 e4 e6 2 d4 d5 3 tiJc3 .ib4 4 e5 c5 5 a3 .ixc3+ 6 bxc3 ti'Je7 7 "iVg4 0-0
1 0 . 1 8 ti'Jf3 1 78 10.2 8 .id3 1 82 1 0 . 2 1 8 . . . ti'Jbc6 1 82 10.2 1 1 9 v,wh4 1 83 10 . 2 1 2 9 v.lVh5 ti'Jg6 1 0 ti'Jf3 "iVc7 1 85 1 0 . 2 1 2 1 1 1 0-0 1 86 10.2122 1 1 .ie3 1 8 7 10.22 8 . . . f5 1 90
2 70
Index of Variations 1 1 Winawer Variation: Positional Lines 1 e4 e6 2 d4 d5 3 ltJc3 �b4 4 e5 c5 5 a3 .txc3+ 6 bxc3 ltJe7
1 1 . 1 7 ltJf3 ltJbc6 1 95 1 1 . 1 1 8 �d3 1 96 1 1 . 1 2 8 �e2 1 98 1 1 .2 7 a4 'iVa5 201 1 1 . 2 1 8 'i!id2 201 1 1 .22 8 �d2 ltJbc6 9 ltJf3 �d7 203 1 1 .221 10 �b5 204 1 1 .222 10 �e2 20 7 1 1 . 3 7 h4 ltJbc6 8 h5 'i!ia5 9 �d2 209 1 1 . 3 1 9 . . . .td7 21 1 1 1 .32 9 . . . cxd4 21 2 1 2 Winawer Variation: Black Plays 6".'iVc7 1 e4 e6 2 d4 d5 3 ltJc3 �b4 4 e5 c5 5 a3 �xc3+ 6 bxc3 Wic7
1 2 . 1 7 'iVg4 f5 21 5 1 2 . 1 1 8 Wig3 21 5 1 2 . 1 2 8 Wih5+ 21 9 12.2 7 ttJf3 223
2 71
Play the French 1 3 Classical Variation: 4 e 5 1 e4 e6 2 d 4 d5 3 ltJc3 ltJf6 4 e5 ltJfd7
1 3 . 1 5 ltJf3 229 13.2 5 ltJce2 c5 6 c3 ltJc6 7 f4 'fib6 s ltJf3 f6 230 1 3 . 2 1 9 g3 cxd4 234 1 3 .2 1 1 10 ltJexd4 234 1 3 . 2 1 2 10 cxd4 236 1 3.22 9 a3 23 7 1 3 . 3 5 f4 c5 6 ltJf3 ltJc6 7 1Le3 cxd4 S ltJxd4 �c5 9 'iVd2 0-0 239 1 3 . 3 1 10 g3 241 1 3. 32 1 0 0-0-0 242 14 Classical Variation: 4 1Lg5 (4 ... dxe4 5 ltJxe4 �e7) 1 e4 e6 2 d4 d5 3 ltJc3 ltJf6 4 �g5 dxe4 5 ltJxe4 1Le7 6 i.xf6 .txf6 7 ltJf3 ltJd7 S 'ifd2 0-0 9 0-0-0 14.1 9 1Le7 254 14.2 9 b6 256 ...
. . .
1 5 Odds and Ends 1 e4 e6 1 5 . 1 2 b3 259 1 5 . 2 2 ltJf3 d5 1 5 . 2 1 3 ltJc3 ltJf6 4 e5 ltJfd7 5 d4 c5 6 dxc5 ltJc6 7 .i.f4 .i.xc5 S .lid3 f6 9 exf6 ltJxf6 1 0 0-0 0-0 259 1 5 .2 1 1 1 1 .tg3 261 1 5 . 2 1 2 I 1 ltJe5 261 1 5.22 3 e5 c5 4 b4 262 1 5 . 3 2 d4 d5 3 �e3 263
2 72