Beating the Sicilian 2

JOHNNUNN

B.T.Batsford Ltd, London

First published 1990 ©John Nunn 1990

ISBN 0 7134 6445 3 A CIP catalogue record for this book is available from the British Library

All rights reserved. No part of this book may be reproduced, by any means, without the prior permission of the publisher

Typeset by Latimer Trend Co Ltd, Plymouth, Devon and printed in Great Britain by Dotesios (Printers) Ltd, Trowbridge, Wilts for the publishers, B. T. Batsford Ltd, 4 Fitzhardinge Street, London WIH OAH

A BATSFORD CHESS BOOK Adviser: R. D. Keene GM, OBE

Technical Editor: lan Kingston

Contents

Preface V

Najdorf Variation

2 Scheveningen Variation 19

3 Classical Variation 40

4 Pelikan Variation 62

5 Dragon Variation 77

6 Kan Variation 94

7 Maroczy Bind 110

8 Taimanov System 125

9 Sicilian Four Knights 134

1 0 Lowenthal Variation 140

l l Pin Variation 145

12 2 . . . <2:lf6 149

1 3 2 . . . a6 155

14 Unusual Lines 159

Index of Variations 168

Preface

This second edition of Beating the Sicilian does not need such a lengthy introduction, as the con­cept of a 'repertoire' book has become a familiar one. However I will take time to detail some changes which have been made to the original Beating the Sicilian. The earlier work was based on the premise that a book like The N aj­dorf for the Tournament Player provides a surfeit of information for the average White player, who only wants to know one good line to play against the Najdorf. It therefore developed a complete repertoire for White against the Sicilian, covered in sufficient de­tail for most players to meet any line of the Sicilian with confi­dence. To my surprise I found that the book became popular not only amongst the intended audience of club players, but also at the Grandmaster level.

There were faults in Beating the Sicilian. I recommended some unusual sidelines, but this recom­mendation was self-defeating, because many of these lines were suddenly played in Grandmaster games. At this level the reasons why the lines were in fact sidelines became abundantly clear. Thus the book contributed to its own

rapid dating. This time I have adopted a different policy. In every variation I have recom­mended one of the current main lines against that variation. Such main lines cannot be refuted, although the whims of fashion may lead to them becoming more or less popular as the years go by. In this way the lines recom­mended in the book should re­main valid for years to come (un­til the third edition?) and readers may be confident that the effort put into studying them will earn a long-term reward. One side-effect of this policy is that the analysis has become considerably more detailed. You cannot learn the main line of the Yugoslav Attack without studying a fair number of variations, but as we only need to examine Black alternatives the work required is not too onerous. In any case the growth of opening theory at all levels of chess has forced players to devote more ef­fort to the study of opening theory in order to maintain (or increase) their level of success. Unfortunately this has led to scrapping the policy of offering alternative White choices such as 6 a4 against the Najdorf. The book would have grown too big

vi Preface

(and expensive) if these had been left in. In any case, these lines have not changed greatly since the first book, so much of the old material is still valid.

I have maintained the arrange­ment of material from the earlier book. The biggest changes in the repertoire are the abandonment of the Classical Scheveningen (now only the Keres attack is recommended), the switch to 9 J..xf6 in the Pelikan, adopting the main line of the Yugoslav Attack against the Dragon, and moving to 5 � b5 against the Taimanov. Needless to say, all chapters have been thoroughly updated, even if the chosen variation is unchanged. All material up to October 1 989 has been consulted, including /nformator 47, N ew in Chess yearbook 1 2 and Chess Base magazine 13.

In some cases, usually at a fairly late stage of the analysis, I have given more than one move for White where there is genuine doubt as to which is the best alter­native. I have devoted special at­tention to a few relatively unusual lines which in some cases have a 'cult' following. Examples of this are the Pin Variation (Chapter 1 1 ) and 2 . . . �f6 (Chapter 12). These lines are probably not very good for Black, but they are tricky and

are sometimes used as a surprise weapon. All too often a White player, when confronted by one of these unusual lines, will decide to play safe and not enter the main theoretical paths. In doing this he often throws away any chance to gain the advantage. A short sur­vey of the critical lines is enough to enable one to counter such unusual lines, but I have given a fairly detailed coverage for the sake of completeness.

One problem with the division of material into games is that while it is convenient for the reader who starts at page l it is not so suitable if the book is to be used as a reference work. Finding which game contains a particular line is not so easy, but an unusually detailed index should help with this problem. I have again included plenty of diagrams at critical moments to provide further assistance.

Finally I would like to thank all those readers of Beating the Sici­lian who wrote with comments, suggestions and games. Grand­masters are usually only too will­ing to offer their opinions, and it is useful to get balancing feedback from the club players who provide the main readership of a book such as this.

1 Najdorf Variation

Of all the lines in the Sicilian which Black can adopt, the Naj­dorf has developed the largest body of theory. Whole books have been written on mere sub­variations of the Najdorf, for ex­ample the Polugayevsky variation and the infamous Poisoned Pawn. Devising a counter which is viable in tournament play, while at the same time necessitating relatively slight book knowledge, has proved especially hard. The Naj­dorf starts with the moves I e4 c5 2 �f3 d6 3 d4 cxd4 4 �xd4 �f6 5 �c3 a6. Black's first aim is to play . . . e5 without allowing the reply ll. b5 + , while the secondary point is to prepare queenside expansion by b5. Some of White's sys­tems against the Najdorf are spe­cifically aimed at preventing e5, while others allow Black to play this move in the hope of exploiting the backward d-pawn later. 6 ll.g5 and 6ll.e2 are the most common replies, but as both lead to reams of analysis I have settled on a less familiar system which offers good attacking chances while retaining an ele­ment of solidity. This system, based on 6 f4, has become more popular since the first edition of this book was published and

games by players such as Belyavsky, Dolmatov and A. Sokolov have shed new light on several lines. In games 1 -4 we investigate the replies most com­monly encountered in practice. The main division is between those lines involving an early e5 and those in which Black de­lays this move or omits it entirely. We will postpone consideration of the former lines until games 3 and 4, and concentrate first on the alternatives to e5.

1 B

Game I F. Olafsson-Sax Novi Sad 1976

1 e4 c5 2 �f3 d6 3 d4 cxd4 4 �xd4 �f6 5 �c3 a6 6 f4 (1)

2 Naj dorf Variation

6 f4 is a flexible move; White gives little away regarding his piece deployment, and waits for Black's reply before deciding where to put his bishops.

6 <E)bd7 This move is designed to re­

serve the option of playing e5 under more favourable circum­stances if White should develop his pieces to unsuitable squares. 6

'Wfc7 is game 2, but there are other playable moves:

( l ) 6 ... <E)c6 7 <E)xc6 bxc6 8 e5 <E)d7 (8 . . . dxe5 9 'Wfxd8 + '\t;xd8 1 0 fxe5 <E)d5 1 1 jj_d2 is good for White) 9 jJ_c4!? dxe5 1 0 0-0 e6 1 1 f5 jj_c5+ 12 fph1 with good at­tacking chances.

(2) 6 .. . g6 7 jJ_d3 and after a subsequent . . . 'Wfc7 or <E) bd7 there will probably be a transposi­tion into lines considered in game 2.

(3) 6 e6 (after this White may transpose into various lines of the Scheveningen, but since these lines do not form part of the repertoire recommended in this book, we suggest an independent alternative which promises good chances for White) 7 'W/f3 (White's advantage over similar lines is that his bishops are not committed, so he can force through g4-g5 very quickly) and now:

(3a) 7 ... �bd7 8 g4 e5 9 <E)f5 is dangerous for Black, for example 9 g6 (9 . . . exf4?! 10 jj_xf4 <E)e5 1 1 jJ_xe5 dxe5 1 2 g5 <E�d7 1 3 jJ_c4 gives White a tremendous attack)

1 0 g5 gxf5 1 1 exf5 and Black is under heavy pressure.

(3b) 7 ... 'Wfc7 8 g4 b5 9 g5 �fd7 (9 b4 1 0 <E)cb5 axb5 1 1 gxf6 gxf6 1 2 jJ_xb5+ jJ_d7 1 3 jj_d3 gave White a good game in Lau-Schuh, West German Ch. 1 987) 1 0 a3 jJ_b7 1 1 jj_e3 (1 1 jJ_g2 g6 12 'W/f2 <E�c6 1 3 <E)de2 h6 1 4 f5 was also promising in Smyslov­Kamsky, New York Open 1989) <E)c6 1 2 jj_h3 b4 13 axb4 <E)xb4 1 4 0-0 <E)c5 15 madl g6 1 6 .!!d2 jJ_e7 17 K df2 with advantage to White, Timman-Hjartarson, Bel­fort World Cup 1988.

(3c) 7 . .. 'W/b6 8 <E)b3 'Wfc7 9 g4 b5 1 0 jj_d3 jj_b7 1 1 g5 <E�fd7 1 2 jj_e3 <E)c6 1 3 'W/h3 b4 (1 3 <E�c5 1 4 0-0 <E)b4 1 5 l!adl <E)bxd3 16 cxd3 <E)xb3 17 axb3 'W!d7 18 f5 exf5 19 jj_d4! was good for White in Ulybin-Labunsky, USSR 1 987) 1 4 <E)e2 g6 1 5 0-0 -0 jJ_g7 1 6 K hf l 0-0-0 17 f5 <E)ce5 18 'lt;bl with a slight plus for White, Uly­bin-Magerramov, Uzgorod 1988.

7 jJ_e2 (2) This position frequently arises

via the move order 6 jJ_e2 <E)bd7

(instead of the more common 6 e5) 7 f4.

7 eS This move is most common, the

following lines being somewhat unpleasant for Black:

( 1) 7 . . . g6 (an attempt to reach positions akin to the Dragon, but here White can exploit an interest­ing tactical resource) 8 g4 Qlc5 (8

. . h6 9 f5 Qlc5 10 J.,f3 e5 1 1 <El b3 gxf5 12 exf5 e4 13 J.,g2 Qlxb3 14 axb3 )!!g8 15 h3 J.,xf5 16 'i$'d4 J.,e6 was unclear in Belyavsky­Ljubojevic, Bugojno 1984, but this line is certainly risky for Black) 9 Q�b3! b6 (not 9 . . . Qlfxe4? 10 Qlxe4 '2:!xe4 1 1 �d4 Qlf6 12 g5 winning, nor 9 . . . Qlxb3 10 axb3 J.,g7 1 1 g5 Qld7 12 J.,e3 with a clear plus for White, Liberzon-Savon, Sukhumi 1972) 10 g5 Qlfd7, Arnason-Tringov, Plovdiv 1 986, and now 1 1 J.,f3 J.. b7 12 �e2 supporting e4 and preparing J.,d2 and 0-0-0 gives White a promising position.

(2) 7 Qlc5 8 J..f3 *b6 9 '2:! b3 (once again this move, put­ting the question to the c5 knight, gives White the advantage) Qlxb3 10 axb3 g6 11 e5 dxe5 1 2 fxe5 Qld7 13 Qld5 'i'ird8 14 J.,g5, Gips­Iis-Quinteros, Olot 1973 and White stands very well since 14 . . <El xe5? loses to 15 <El f6 + exf6 16 tij'xd8 + f!'xd8 1 7 J.,xf6 +

(3) 7 . . . �b6 8 Qlb3 g6 9 litd3 (preparing jle3, when Black is driven back in confusion) �c7 1 0 g4 Qlc5 1 1 Qlxc5 *xc5 1 2 j_e3 �aS 1 3 b4! with advantage,

Naj dorf Variation 3

Torre-Quinteros, Leningrad IZ 1973.

(4) 7 . . . bS (I suggested this move in 1982 but a few months later found a good reply; 8 Qld5! J.,b7 (9 Qlc6 was the threat and 8

Qlxd5 9 exd5 gives White a superb outpost at c6) 9 Qlxf6 + '2:!xf6 (or else White has a positio­nal advantage) 1 0 e5 dxe5 l l fxe5 followed by e6, when Black has problems developing his kingside p1eces.

8 Qlf5 Inexperienced White players

sometimes try 8 <£) b3 but after the reply 8 b5 White should be thinking about equalizing! Natu­ral moves like 8 <£) b3 quite often turn out badly in the Najdorf, which is one reason why it is so popular with Black players.

8 Qlc5 9 Qlg3

White's knight manoeuvre solidly defends the sensitive e4 square and he is now ready to complete his development by 0-0 and J..e3. If Black does nothing to stop this then White's position will be very promising, so his next move is directed against both the developing moves mentioned above.

9 �b6 Or 9 j_d7 10 0-0 j_e7 11

j_e3 ( 1 1 a4 i s also possible, when 11 g6? 12 J.,e3 J..c6 1 3 fxe5! '2:!fxe4 1 4 '2:!cxe4 Qlxe4 15 e6! was very good for White in Kiprich­nikov-Petkevich, USSR 1975) � c8 1 2 fxe5 dxe5 13 <£) d5 j_ e6!

4 N aj dorf Variation

1 4 <E:Jxf6+ (or 1 4 tt:Jxe7 �xe7 1 5 Jtg5 .:E:lcd7 1 6 c3 with an edge for White) Jtxf6 1 5 �cl Jte7 with a small plus for White, Tsesh­kovsky-Tarjan, Riga 1 979.

10 Kb1! (3) It may seem to be a major con­

cession to play such a move merely in order to achieve j_e3, and for a long time 1 0 f5 was played in order to delay the pain­ful decision to waste time with l!l bl . However, it is now recog­nized that White should retain the option of opening up the position by fxe5 and so the immediate .1. b l is thought best today. The situation in this line is typical of opening variations in which Black plays very actively. If he can keep his initiative going then every­thing turns out well for him, but if he falters and has to start moving backwards the result is usually disastrous. Which of these alter­natives actually occurs frequently depends on obscure tactical points, as happens here.

3 B

10 Jtd7 10 JJ..e7 sets a neat trap

which Karpov once fell into-11 Jte3?! exf4 1 2 j_xf4 .:E:lcxe4 1 3 .:E:lgxe4 .:E:lxe4 1 4 .:E:lxe4 �b4+ and Black wins a pawn, although White probably has sufficient compensation to draw. The best line against 1 0 j_e7 is 1 1 fxe5 dxe5 1 2 j_e3 �c6 (12 �c7 1 3 0-0 0-0 1 4 <ffh l b5 15 � xf6! Jtxf6 1 6 .:E:ld5 �c6 1 7 .:E:lxf6+ gxf6 1 8 �el is also dangerous for Black) 1 3 0-0 0-0 1 4 .:E:ld5 .:E:lxd5 1 5 exd5 �c7 1 6 b4 .:E:la4 1 7 .:E:le4 and White's powerful queenside majority gives him a substantial advantage, Vogt-Fernandez, Halle 1978 .

10 . . . exf4 1 1 j_xf4 .:E:le6 is a recent idea for Black, which has had success in two games. After 12 itd2 .:E:ld4 13 itd3 (or 1 3 .:E:ld5 .:E:lxd5 1 4 exd5 Jte7 1 5 JJ..c3 .:E:lf5 with rough equality), Kinder­mann-King, Dortmund 1988, continued 1 3 Jte7 1 4 .:E:lh5 0-0 15 .:E:lxf6+ Jtxf6 1 6 .:E:ld5 �d8 1 7 0-0 j_e5 1 8 �h5 j_e6 19 j_g5 f6 20 Jte3 b5 with equality, while Vogt-H. Olafsson, Thessaloniki 01. 1988 went 1 3 . . �c5 1 4 .:E:ld5 .:E:lxd5 1 5 exd5 �xd5 1 6 0-0 .:E:lc6 1 7 j_e4 �c5 + 1 8 <ffh l .:E:le5 19 .:E:lf5 J..xf5 20 K xf5 �c7 and this too ended in a draw. The second line looks risky for Black, but King's handling of the position affords White few chances for an advantage. However White has an interesting option on move 13, namely 13 .:E:lh5!? .:E:lxh5 1 4 j_xh5 j_e6 (or else .:E:ld5) 1 5 0-0!? To castle into a double check appears

suicidal, yet this is the only way to make j_e3 into a threat! After 15 . . . g6 16 J,.e3 J,.g7 17 <E)a4 'fl!a7 18 c3 or 15 . . . � xc2 + 16 fihl IE}e3 17 'fl!a4 + White is clearly better, so Black should probably play the safe 15 <E)c6+ 16 * h 1 g6, although even here 17 jtg4! looks at least slightly better for White.

1 1 fxe5 dxe5 12 j_e3

Here, too, this is White's best. The threat of b4 forces Black to move his queen, and since he must try to keep up the momentum it is natural to attack the e4 pawn. Everything then hinges on whether White can successfully gambit this pawn.

12 'fl!c6 13 0-0 0-0-0

Black would like to play 13 <E) cxe4 14 <21 cxe4 <21 xe4 15 J. f3 jtc5 but White wins material by 1 6 �xe4 j_xe3+ 1 7 ff h l 0-0 18 <E)c3 'fl!c8 1 9 IE}d5 J..c5 20 b4 j_d6 21 <E)b6. If Black can't take the e-pawn then his whole plan looks suspicious, but he could have minimized his disadvantage by 1 3 !it, d8! As played his king is subjected to a devastating attack.

14 b4! White forces Black to take the

poisoned pawn. This would also have been the reply to 13 . . . j_e7.

14 �cxe4 1 4 . . . <E)e6 1 5 <E)d5 threatening

c4 and b5 is also horrible. 15 IE}cxe4 IE}xe4 16 <E)xe4 'flfxe4 (4)

Najdorf Variation 5

17 j_b6 K e8? Now Black gets mated. The

best practical chance was to jetti­son the exchange by 1 7 . . . J..e7, although 18 J..f3 'fl!g6 1 9 jtxd8 � xd8 20 'f!ld5 'f!lb6+ 21 fihl f6 22 'f!lf7 attacking the vulnerable kingside pawns should win for White in the long run.

18 J.. f3 'W!g6 19 jtxb7 + '*b8

19 '\ffxb7 20 'fl!xd7+ fixb6 2 1 'flfxe8 jtc5 + is impossible because White takes the bishop with check.

20 J..e3 The simplest. Material is level

and White's attack against the naked Black king must be deci­sive.

20 j_b5 21 jtf3 'W!d6

21 . . . jtxf l loses to 22 'f!ld7. 22 'fl!e1 f5

22 jtxfl 23 'fl!xfl followed by b5 or H dl wins.

23 c4 24 b5 25 'fl!a5 26 .!il(xb5 +

jt xc4 axb5 e4 Resigns

6 Naj dorf Variation

26 jlxb5 27 tyxb5 + *c7 28 H c l + *d8 29 j_b6+ *e7 30 jlc5 wins.

Game 2 Nunn-Griinfeld

England-Israel Telex Match 1981

1 e4 c5 2 �f3 d6 3 d4 cxd4 4 � xd4 �f6 5 �c3 a6 6 f4 tyc7

If Black wishes to delay e5 (or even dispense with it alto­gether) then this is probably the most reliable way to go about it. Black avoids the tactical problems resulting from a quick e5 by White and can continue his de-velopment by g6, jlg7, � bd7 and maybe b5 and 0-0 as well.

7 �f3 (5) This is more accurate than 7

j)_d3, when 7 e5 8 �f3 b5 transposes into a relatively com­fortable line for Black.

5 B

7 �bd7 After this we reach a standard

position which can arise by a wide

range of move orders. The main question is whether Black can exploit White's early �f3 by play­ing 7 . . . e6. The analysis runs 7

e6 8 j_d3 and now: (I) 8 . . . b5 9 e5! dxe5 (9 . . . b4

10 �e4 �xe4 1 1 jlxe4 d5 12 Ad3 is slightly better for White, while Sax-Guerra, Dubai 01. 1986 con­tinued 9 . . . �fd7 10 0-0 �c6 !I f!h l .f/J_e7 1 2 tyel 0-0 1 3 tyg3 f5 1 4 exf6 �xf6 1 5 j_d2 with advan­tage to White) 10 fxe5 �fd7 (1 0

�g4 1 1 tye2 j_b7 12 jle4 also gives White an edge) Il 0-0 �c6 1 2 J_f4 .f/J...b7 1 3 *hi �c5 1 4 �e4 �xe4 15 jlxe4 �b4, Reeh­Schulz, West German Ch. 1 987, and now 1 6 �g5!? gives White dangerous attacking chances.

(2) 8 . . . <2lc6 9 0-0 and now: (2a) 9 .f/J...e7 10 tyel and

Black has a range of possibilities. The passive 10 . . . �d7 11 �g3 0-0 1 2 f!hl �e8?! 13 e5! �b4 14 f5! gave White a strong attack in Hazai-Karolyi, Hungary Ch. 1 986, while 10 . . . 0-0 is met by 11 e5 dxe5 12 fxe5 �d7 1 3 J_f4 �c5 1 4 ,.g3 �xd3 1 5 cxd3 *h8 16 �e4 and White i s better accord­ing to Sax and Hazai. Perhaps the best line is 10 . . . �b4 1 1 e5 �xd3 1 2 cxd3 �d5 1 3 .:£)xd5 exd5 14 tyg3 0-0 1 5 *hi dxe5 16 fxe5 jlf5 1 7 �d4 j)_g6 1 8 .f/J_d2 with a level position, Sax-de Firmian, New York Open 1 987.

(2b) 9 . . . b5 10 'l!!te l j_b7 I l *hi jle7 1 2 e5! dxe5 1 3 fxe5 ®d7 1 4 .f/J_f4 �c5 1 5 �e4! (as in line I above, except that the

moves �e1 and . . . .fi.e7 have been added) <£)xe4 ( 15 <£)xd3 1 6 cxd3 0-0 17 �g3 'f!h8 1 8 <£)f6! �d8 19 <£)g5 is very good for White) 16 .fi.xe4 h6?! ( 16 . <E:lb4 17 .fi.xb7 �xb7 1 8 �g3 g6 may be better, although 19 .!i.h6 cuts Black off from the kingside ), Sax­A. Sokolov, Reykjavik 1988, and now 1 7 a4 b4 18 �f 2 gives White good attacking prospects.

(3) 8 . . . <£)bd7 9 0-0 11.e7 (9 b5 may be met by the simple 10 �e1 , or even by 1 0 e5!? dxe5 1 1 fxe5 <E:lxe5 12 <E:lxe5 �xe5 1 3 �f3 and now 13 . . . lilt b8?! 14 .!i.f4 'i'ifc5 + 15 \tihl Ji.b7 1 6 .fi.e4! <E:lxe4 1 7 .fi.xb8 f5 18 li;l\ ae 1 was good for White in Nicevski-Mar­kiewicz, Dembica 1 987, so Black should have played 13 ... � a7 14 .!i.f4 �h5 15 'i'lfg3 with an unclear position) 1 0 �el 0-0 11 e5! <E:le8 12 'ltg3 <£)c5 13 .fi.e3 1l_d7?! 14 .fi.xh7+ ! \tixh7 15 <E:lg5 + \tig8 ( 15 .fi.xg5 1 6 fxg5 �h8 17 /i;il,xf7 is unpleasant) 16 '/!l!i'h4 .fi.xg5 17 fxg5 Ji.c6, Wedberg-1 onescu, Berlin 1988, and after 1 8 exd6 <E:lxd6 (not 1 8 . . . �xd6 19 /i;il,ad1 �e7 20 'l!!fb4) 19 .fi.xc5 <£)f5 Black does not have enough for the pawn.

8 1l_d3 g6 8 . e5 9 a4 transposes to game

3, while 8 e6 is line 3 in the above analysis.

9 0-0 White's strategy in this line is

rather crude. He intends a straightforward attacking build­up on the kingside by *e1-h4, f5,

Naj dorf Variation 7

.!i. h6, and <E:l g5. Of course Black is also playing moves while all this is going on but if he continues naively with his development without taking specific counter­measures he can easily fall victim to White's attack.

9 .!i.g7 10 �e1 (6)

10 0-0 This move is probably already

an inaccuracy. The alternatives are:

( l ) 10 ... <E:lc5 1 1 e5 dxe5 1 2 fxe5 <£)fd7 1 3 .fi.f4 <£)e6 1 4 .fi.g3 <E:lb6 (after 1 4 0-0 1 5 fi'hl Black has to find an answer to <E:ld5) 1 5 a4 .!i.d7 1 6 a5 �c8 1 7 <£)e4 and White's initiative proved decisive in Sax-Minic, Rovinj­Zagreb 1 975.

(2) 10 . . . b5 (probably the best move, aiming to complete Black's development before he gives White a target to attack by castling) 1 1 e5 (attacking moves like � h4 serve no function while Black's king is still in the centre) dxe5 1 2 fxe5 <£) g4 1 3 e6 fxe6 14 �h4 with an unclear position. For the pawn White has some

8 N ajdorf Var iation

initiative and Black has problems finding a refuge for his king. White also has the option of opening lines on the queenside by a timely a4, and in practice Black will not have an easy defensive task ahead of him.

1 1 .-h4 b5 12 f5

At one time White invariably played * h 1 before proceeding with his attack. This type of con­solidating move is often a symp­tom of chess laziness, in that White does not want to be both­ered with calculating the conse­quences of Black's queen check in every variation and so simply rules it out, even though it may cost him a vital tempo.

12 jj_b7? Black continues to play normal

Sicilian moves without realizing how critical his position has become. The point is that after White's fxg6 Black does not want to play hxg6 when �g5 gives White a permanent mating threat on h7. However, the recapture . . . fxg6 invites White's knight to come in at e6 and Black's . . . jj_b7 removes a vital defence from this square. 1 2 . . . �c5 was essential, when 1 3 jj_h6 b4 may enable Black to defend.

13 fxg6 In a game Velikov-Valenti,

Pernik 1979, White played 13 j}_e3 (laziness again) when Black missed his second chance to play

�c5 and lost after 1 3 b4? 1 4 �d5! jlxd5 1 5 exd5 �xd5 1 6

fxg6 hxg6 1 7 �g5 �5f6 1 8 gf3 with a crushing attack.

13 fxg6 After 1 3 . . . hxg6 14 �g5 Black

cannot move either knight since �h5 is met by g4. White can

just build up by � f3 and � afl to eliminate the defensive knights at f6.

14 .f)g5 �c5 ( 7) Too late! 1 4 �b6+ (1 4

�h5 1 5 J..e3 is good for White) 1 5 \ffh l �h5 was best, but even then 16 j}_ d2 threatening both �e6 and .:2)d5 gives White a pro­mising attack.

15 );! xf6! � xf6 16 �xh7 + 'if1fB 17 j}_e3

White's material investment is very slight for such a strong at­tack. The main threat is 18 .:2)d5 jj_xd5 19 exd5 attacking g6 and preparing b4 followed by the occupation of e6 by White's knight.

17 .:2)xd3 1 7 . . . e5 18 .:2\ d5 j}_ xd5 1 9 exd5

e4 20 .il.. e2 � e8 21 b4 followed by .f)e6 + is also winning.

18 cxd3 'ltd7 Black cannot meet the threat of

�d5 by 18 e6 since 19 �xe6 + ! li( xe6 20 � f l + <lieS (20

� f6 21 il! xf6 + wins the queen) 21 'lt'g8 + is decisive.

19 .!2ld5 j}_xd5 20 exd5 'lt'f5 21 2£le6 + � xe6 22 dxe6 'lt'xe6 23 j}_h6 Resigns

23 J..xh6 24 'ith8 + "'ffl 25 ii f l + wins everything.

1 2 3 4 5 6

Game 3 Nunn-Cserna Lugano 1984

e4 c5 �f3 d6 d4 cxd4 .!2lxd4 .!2lf6 .!2lc3 a6 f4 e5

Black's most popular move. After all, a6 was designed to prepare . . . e5 and the determined Najdorf player will generally play

e5 unless it is absolutely im­possible.

7 .!2lf3 'lt'c7 At one time it was held that

Black should prevent the active development of White's bishop at c4 and so this move was almost universal. But more recently 7 �bd7 has become the most popu­lar move. We consider this in game 4.

8 a4 The alternative is 8 J..d3. The

continuations after 8 a4 and 8 J..d3 are rather similar, but there

Naj dorf Variation 9

are some important differences. Firstly 8 a4 expends a tempo, but this is not especially serious since White can often omit * h 1 (after 8 j_d3 b5 White usually has to play *hi since Black's check on b6 gives him extra defensive possibi­lities). More significantly, a4 re­serves the c4 square for White's use (by 'lt'e2 and J..c4 or .!2ld2 and j_c4) and in some lines the added pressure White can exert on f7 by these manoeuvres improves his chances considerably.

8 B

8 .!2lbd7 9 J..d3 (8)

9 g6 Black's main decision is

whether the f8 bishop should go to e7 or g7. There are two other lines, one in which Black commits himself to . . . j}_e7 at once and one in which he postpones the decision:

(1) 9 • . . J..e7 10 0-0 0-0 1 1 .!2lh4 g6 (this is usually unavoidable in lines where Black plays . . . j_e7, for example 1 1 J..d8?! 1 2 .!2lf5 g6 1 3 <2\h6+ fig7 14 f5 b6 15 g4 with a dangerous attack, Sax-

10 Najdorf Var iation

Bukic, Vrbas 1980) 12 f5 d5!? 13 exd5 e4 (13 �xd5? 1 4 �xd5 �c5+ 1 5 �e3 �xd5 16 fxg6 hxg6, Cramling-Gallagher, Oak­ham 1984, and now White could have won by 1 7 �g4!, e.g. 17 . . . �f6 1 8 �xg6 j_xg4 1 9 �xe7 + 'fih8 20 <2)xd5 <2lxd5 2 1 j_h6, 17 . . . j_c5 18 <2lxg6 J..xe3+ 19 fih l "ii\'e6 20 J..f5, 17 . . . �b6 18 "ii\'g3 and White has threats at b6 and g6, or finally 17 . . . J..xh4 1 8 "ii\'xh4 followed by j_c4 and j_h6, and in every case White has a winning position) 1 4 j_e2 j_d6 15

better. This i s an example of the advantage of having c4 available for the bishop.

10 0-0 11 �e1 12 fxe5 13 �h4 14 J..h6 15 �g5 (9)

�g7 0-0 dxe5 b6 j_b7

This position is the natural re­sult of White's blunt play. Although its evaluation has fluc­tuated over the years, the scales have recently tipped in White's favour. It is very rarely seen today because Black players steer well clear of it.

g3 b6, Sax-Andersson, London 1980, and now 16 <2)g2 is best, followed by j_h6 and �e3 sup­porting the pawns at d5 and f5, 9 when White should have the ad- 8 vantage.

(2) 9 . . . b6 (the problem with this delaying move is that Black may have trouble getting castled) 1 0 0-0 j_b7 11 �e1 g6 ( 11 . . . J..e7 1 2 fihl 0-0 1 3 �h4 g6 1 4 fxe5 dxe5 15 j_h6 H e8 1 6 �f5! J..f8 1 7 �xf8 � xf8 18 �e3 is slightly better for White, Szna­pik-Ostermeyer, Oslo 1 983) 1 2 fxe5 dxe5 1 3 �h4 J..e7 (1 3 J..g7 14 J..h6 0-0 transposes to the main line) 14 j_g5 h6 1 5 f1h1 (15 �d2? J..c5 + and 1 6 �h5) f1f8 (or 15 0-0-0 16 j_e3 with an automatic attack against Black's weakened queenside) 16 �d2! f1g7 1 7 j_e3 �c5 18 j_c4 � af8 19 �f2 j_d8 20 j_xc5 "ii\'xc5 21 'i!J\,1'xc5 bxc5 22 a5 H. Ranta­nen-Nunn, Helsinki 198 1, although White is distinctly

15 };i(fc8 The only other reasonable

move is 15 "ii!'d6 (15 ... �h5? 16 j_xg7 *xg7 17 Hxf7+ and 15 • • • 1ii! ae8 16 g4! are to be a voided, while 15 . . . l;:l: fe8 1 6 .�xg7 fixg7 17 �xf7! 'ifi>xf7 18 �xh7+ 'lt'e6 19 �xf6 + ! fixf6 20 };i(fl + 'l!'e6 2 1 . .l\'l.c4 + *d6 22 liild l + f1c6 23 J..d5 + *c5 24 j_xb7 Resigns was Rantanen-Morris, Gausdal 1978), but 1 6 Hadl causes serious problems. After 1 6 . . . �h5 White

plays 18 j_xg7 'f5xg7 19 J..e2 �c5 + 20 'I;> hi, when 20 . . . �hf6 loses to 2 1 � xd 7 and 20 . . . '21 df6 loses to 2 1 j_xh5 �xh5 22 � xf7 + Other 16th moves are almost as bad.

16 '15h1 Black intends to meet moves

such as 16 g4 and 16 gad! by 'ii'Jc5 + and 'ii'Jf8, when the rook on c8 prevents J..c4 and Black successfully defends. Un­fortunately White has a tactical idea which cuts across Black's plan to bring his queen to f8.

16 �d6 17 j_xg7 *xg7 18 �xti!

This hardly counts as a sacri­fice, since White immediately gains three pawns for the piece, while Black's king is left floating around in the middle of the board.

18 *xti 19 •xh7 + *e6 20 •xg6 ( 10)

There was even a second good line in 20 �xf6+ �xf6 21 •xb7 since the attempt to liquidate by 2 1 •c7 allows 22 j_xa6.

10 B

Naj dorf Variation 11

20 ftte7 21 � ad1

As is so often the case, tt ts better to spend time cutting off the king's escape route than to give pointless checks which only serve to drive the king into safety.

21 �h7 22 �g3 *e7

Black cannot play 22 because of the check at c4.

23 �d5 + jtxd5 24 exd5 �h6?

li;i!.g8

This error allows White to force the king onto the back rank, cut­ting off both Black rooks from the kingside. However, even the best line 24 e4 is good for White after 25 �del H g8 26 \Wc7! 'ii!Jg6 27 g3 � ac8 28 d6 + *e6 29 �b7, and with the fall of the e4 pawn White's rooks can get to grips with the Black king.

25 d6 + *d8 26 J..f5! � a7

It is hardly surpnsmg that Black has no reasonable move. The immediate threat is 27 j_xd7 *xd7 28 �xe5.

27 j_xd7 28 it xf6

Game 4 Nunn-King

� xd7 Resigns

Bundesliga 1986/7

1 e4 c5 2 �f3 d6 3 d4 cxd4 4 �xd4 �f6 5 �c3 a6 6 f4 e5 7 �f3 �bd7

12 Najdorf Variation

In my view this is Black's best response to 6 f4. Although White's bishop can now be de­veloped more actively at c4, Black saves a vital tempo by missing out

'f!lc7 and this gives him good chances to equalize. In fact cur­rent theory suggests that White's best plan is to ignore the option to play �c4, and to proceed with his normal development by �d3. Admittedly Black benefits from missing out -wtc7, but it is not clear that the alternatives to '$c7 fully equalize.

8 a4 White cannot do without this as

8 �c4 allows 8 . . . b5 9 �d5 )i:i( b8 1 0 �g5 (after 10 fxe5 dxe5 1 1 jtg5 �b4 Black was slightly better in Hort-Andersson, Wijk aan Zee 1 979) �xd5 1 1 wxd5 'itf/e7 1 2 0-0 h6 with equality, Korchnoi-Hort, Zurich 1 984, while 8 jtd3 allows 8 . . b5, just the line White is trying to avoid.

8 �e7 If Black relents by 8 . 'l!!fc7,

we reach game 3. 8 d5 is a speculative recent idea, but if White replies 9 exd5 (better than 9 fxe5, as played in Ulybin-Odeev, USSR 1 989) e4 1 0 �e5! 11.. b4 1 1 �c4 �b6 1 2 jtb3, he gains the advantage after 12 . • . �fxd5 1 3 a5 �xc3 14 �xd8+ wxd8 15 �xf7+ 'l!/e7 1 8 bxc3 �xc3+ 17 fl>f2 or 1 2 . . • �bxd5 13 0-0 �xc3 1 4 bxc3 0-0 15 *d4. This analysis has not been tested in practice, but it looks good.

9 �d3 0-0 10 0-0 (11)

10 �c5 It is risky to take the pawn by

10 exf4 1 1 Axf4 'l!!fb6+ 1 2 'itrh 1 li(xb2 1 3 'l!!le l , for example 1 3 'l!l!l'b4 ( 1 3 �c5 1 4 liil b l �xd3 15 cxd3 'l!!fc2 16 d4 l;[e8 1 7 �f2! 'l!!ld3 1 8 Jiilc l Af8 1 9 Jge2 with advantage to White, Ciocal­tea-Danner, Timisoara 1 982) 14 � bl 'l!!fc5 1 5 �d5 �xd5 1 6 exd5 Af6 ( 1 6 .ztd8 1 7 c4 !E:lf6 1 8 jte3 'f<!tc7 1 9 jtd4 gave White a dangerous attack in Ledermann­Lau, Ramat-Hasharon 1 982) 1 7 c4 (or 1 7 '/!l;l'e4 g6 1 8 1J..h611..g7 1 9 Axg7 'l!/xg7 20 �g5 '2\e5 2 1 -Wth4 h6 22 �e4 'l\l'!'xd5 with a double­edged position, Korolev-Lipiridi, corr. 1 984) 'l¥fc7 (a draw was agreed here in Hazai-Marin, Warsaw 1 987) 1 8 'l!!lg3! �e5 1 9 jtg5 Jtxg5 20 jtxh7 + ! \t1xh7 2 1 !2Jxg5+ wh6 (2 1 wg8 22 • h4 Ji:!! e8 23 lll, be I also gives White a winning attack) 22 �h4+ 'l!/g6 23 .i, b3 f5 24 lf,!g3 '2:)g4 25 � xg4 fxg4 26 � xf8 'l!!fe7 27 � f7 �e8 28 �h7 + *xg5 29

�xg7 + *f6 30 �h6+ flf5 3 1 �g5 + *e4 32 �e7+ 1-0, Vogt­Womacka, E. Germany 1989.

1 1 *h1 This is not the only reasonable

move, but judging by recent re­sults it is the most dangerous for Black.

1 1 d5 This is the tactical justification

of Black's play, but there are quieter alternatives:

(I) 1 1 ... .:2\ xd3 12 cxd3 �a5 13 'i!lfel exf4 14 l£)d5 �d8 15 l£)xf4 . .lll..d7 16 j_d2 was good for White, Mateo-Byrne, New York 1986.

(2) 11 . . . Yiic7 12 �el ( 12) (threatening 13 fxe5 dxe5 14 �g3, when Black has no natural way to defend the e5 pawn) and now:

12 B

(2a) 12 .. .. �d7 13 fxe5 dxe5 14 'ii1'g3 .M,ae8 15 �xe5 j_d6 16 �d4, but Black's compensation looks inadequate in view of White's control of d5.

(2b) 12 ... j_e6 13 f5 .lll_d7 14 g4!? ( 14 j_g5 J_c6 15 jlxf6 J_xf6 16 g4 is also slightly better for White) J_c6 ( 14 l£)xg4 loses to 15 l£)d5 �d8 16 Jggl .:2lf6 17

Najdorf Variation 13

J.h6 l£)e8 18 X xg7 + l£)xg7 19 litg3 J..f6 20 li!gl with a winning attack) 15 g5 .:2lh5 (retreating to d7 gives White a completely free hand on the kingside, e.g. 15 l£)fd7 16 f6 j_d8 17 lith4 l£\e6 18 .:2ld5 and wins) 16 f6 j_d8 (not 16

gxf6? 17 lit h4 winning) 17 �h4 g6 18 j_e3 (White's kingside attack has come to a temporary halt, so the time has come to bring the remaining pieces into play. There is a positional threat of 19 j_xc5 dxc5 20 j_c4, followed by the occupation of d5. Thus the c5 knight must move, but 18 . . . l£\e6 19 l£)d5 J_xd5 20 exd5 l£\ef4 2 1 J..xf4 exf4 22 J..f5! followed by Jl_g4 is good for White, hence Black's next move) l£)xd3 19 cxd3 *h8 20 l£\e2 d5? (a very natural move aiming to weaken e4 and activate the c6 bishop against the White king. However, it also weakens the important e5 pawn and this turns out to balance White's weak spot at e4. 20 ... jld7, intending lilfc2, is refuted by 2 1 � fc I when 2 1 �a5 surprisingly loses to 22 Xc4! and Black has no defence to 23 b4; the best defence is 20 ... litd7! aiming for counterplay by attack­ing a4, when White should con­tinue 2 1 �g3 jlxa4 22 l£\xh5 gxh5 23 lilfxh5 jl b5 24 M a3 and Black still has to find a defence to the threat of .Mgl-g4-h4) 2 1 l£\g3 dxe4 (Black cannot play 2 1 litd6 because of 22 d4!, when both 22 . . .' dxe4 23 l£\xe5 and 22 .. . exd4 23 e5 followed by Jtxd4

14 Najdorf Var iation

leave the c6-hl diagonal blocked by a Black pawn) 22 dxe4 \'td6 23 liil, ad I ! 'it b4 (23 \'te6 24 _§_ c5 M. g8 25 J. d6 leads to the loss of e5) 24 .:t:lxe5 Q)xg3 + 25 hxg3 \'txe4 + (after 25 J.xe4 + 26 *gl £lb6 27 J.xb6 �xb6 + 28 � f2 Black cannot meet the threats of �xe4 and \'th6) 26 �xe4 ..l\1xe4 + 27 * h2 (surpris­ingly the exchange of queens does not stop the attack. White's im­mediate threat is 28 J.c5 and the lines 27 . . . Ji:i:c8 28 �d7 *g8 29 .:t:lg4! heading for h6 and 27 . . . 'fi'g8 28 j}_c5 �e8 29 .:t:lxf7! ffxf7 30 1itd7+ 'fjle6 3 1 f7 M.f8 32 �d4 � xf7 33 I, d6 + both win for White) j}_c7 28 ..l\1c5 .�xe5 29 Axf8 lii\xf8 30 �fe 1 .�c2 3 1 !ltd2 Axa4 (Black cannot get two pawns for the exchange since 3 1

J..xg3 + 32 *xg3 Axa4 33 'fJ e7 threatens both � xb7 and M.xf7) 32 �xe5 h6 33 gxh6 'fi'h7 34 g4 *xh6 35 g5 + 'li' h7 36 .1i\ e4 jlc6 37 1ith4 + *g8 38 'fi'g3 Resigns, Nunn-Portisch, Brussels 1986.

(2c) 12 . . . exf4 13 Axf4 M. e8 (after 13 j}_e6 14 .:t:ld4 �b6 15 .�.e3 .:t:lg4 16 j}_g1 .:t:le5 17 .:t:lf5 �.xf5 18 .:t:ld5 �d8 19 exf5 Af6 20 £l.e2 .:t:led7 21 .:t:lxf6+ .:t:lxf6 22 j}_f3 White's two bishops gave him the advantage in Short-Gal­lagher, British Ch. 1987) 14 .:t:ld4 J..d7 15 ..l\1g5 '$d8, Hazai­Novikov, Camaguey 1987, and now White should have taken the chance to activate his bishop by 16 �c4!, pointing it at the sensi-

tive square f7. The key tactical line 16 .:t:lxa4 17 'jJ xa4 J..xa4 ( 17 b5 18 Q)cxb5 axb5 19 'fJ xa8 �xa8 20 J..xb5 is good for White) 18 .:t:lxa4 b5 19 .:t:lc6 �c7 20 .:t:lxe7 + �xe7 2 1 J..xf6 gxf6 22 .:2l b6 bxc4 23 .:t:ld5 �e5 24 .:t:lxf6 + 'fjlg7 25 �h4 h6 (25 M. h8 26 �g4 + ) 26 M.f5! turns out well for White since 26 �xb2 loses to 27 � g5 + !

(3) 1 1 . . . exf4 12 J_xf4 ..l\1d7!? 13 �e2 'jJc8 14 a5 lii(e8 15 J..e3 :tf8 16 ..l\1d4! (intending .:t:lg5) and now:

(3a) 16 . . . jle7 17 b4 .:t:le6 18 j}_b6, 16 . . . h6 17 .:t:ld2 .:t:lg4 18 Ac4 and 16 . . . .:t:le6 17 J..b6 �e7 18 .:t:lh4 g6 19 �ae1 are clearly good for White.

(3b) 16 . . . J..g4 17 'iire3 Ah5 18 j}_xf6 �xf6 19 .:t:ld5 �d8 20 .:t:ld4 j}_g6 2 1 .:t:lf5 ll\e5 22 b4! .:t:lxe4 23 Axe4 ,ll\c4 24 �ae 1 ,ll\cxe4 25 'i'lrxe4 'fJ xe4 26 .Ill xe4 was very promising for White in Ulybin­Pigusov, Pavlodar 1987.

(3c) 16 . . . .:t:lfxe4 (Ulybin's recommendation in Informator, but it appears to have a tactical flaw) 17 .:t:lxe4 .:t:lxe4 18 J_xe4 £1. b5 19 c4! (Ulybin only con­sidered 19 'l'lre3, which leads to a draw) ,ll\xc4 (or 19 J._xc4 20 �c2 J..xfl 2 1 Axh7 + *h8 22 �f5 with a winning attack) 20 �d3 '!Jc5 2 1 Axh7+ *h8 22 �b3 .�xfl 23 j}_xc5 dxc5 24 ifi(xfl *xh7 25 �xf7 with a large ad­vantage for White.

12 .:t:lxe5 ( 13) 12 Q�fxe4

Black has two important alter­natives:

(I) 12 . . . l£l cxe4 13 it xe4 dxe4 14 'l{!te2 and now:

( la) 14 . . . itf5 1 5 g4 jtc8 16 !!:dl �e8 1 7 g5 l£)d7 18 l£)c4 e3 (18 . . . l£)c5 19 b4 l£)e6 20 l£)d5 and 18 . . . b6 19 J..e3 Ab7 20 l£)d5 are also good for White) and now there are two favourable lines for White, either 19 itxe3 b5 20 axb5 itb7+ 21 fl>gl axb5, Kengis-Loginov, Pavlodar 1 987, and now 22 1: xa8 J..xa8 23 l£)xb5 �c8 24 .:E�cd6 �c6 25 fl>f2 leaves Black with inadequate play for the two pawns, or the simple 19 �xe3.

( l b) 14 . . . �d4 1 5 11dl 'tlltb4 16 a5 J..d8 17 �a4 �e7 is Logi­nov's suggestion in Informator, but now 18 b3! appears good for White.

(2) 12 . . . dxe4 13 j_e2 (Black gains time, but the pawn on e4 obstructs Black's pieces) 'tlltc7 ( 13

'tlltxdl 1 4 .i,xdl j_e6 15 j_e3 Xfd8 1 6 g4 g6 1 7 g5 l£ld5 18 �xd5 j_xd5 19 b3 l£)e6 20 .:E�c4 J_xc4 21 J..xc4 jtc5 22 j_xc5 l£l xc5 23 * g2 .1, ac8 24 '11> f2 was

Najdorf Variation 15

marginally better for White in Kindermann--de Firmian, Biel II 1986, but 19 b4 .:E�e6 20 c4 itc6 21 l£)xc6 be 22 !! xd8 !! xd8 23 c5 looks more dangerous) 14 j_e3 (White may also play 1 4 'tllte l first, so as to meet 13 l£)e6 by 14 it d 1 attacking e4) b6 15 'tllte 1 J..b7 16 t�tg3 and now:

(2a) Belyavsky-Chandler, Vienna 1986 continued 16 l£)e6?! (with the idea of exchang­ing bishops by j_c5, but this plan fails tactically) 17 11 ad 1 (not 1 7 f5 J.. d6!, but now f5 is a serious threat) Ac5? (Black's position was uncomfortable in any case) 1 8 f5! Ad6 (the point is that 18 jtxe3 1 9 fxe6 fxe6 loses to 20 l!xf6! ;gxf6 2 1 �d7) 19 !!xd6 'tlltxd6 20 fxe6 and White won.

(2b) Black tried 16 ... � ad8 in Psakhis-Balashov, Irkutsk 1986, but 1 7 !! ad 1 l£)cd7 18 J..d4 l£)xe5 19 fxe5 l£)d7 20 b3! (20 J..xa6! was even stronger) followed by jtc4 gave White a decisive attack. Black should have played for exchanges by 1 7 !! xd 1 18 )!! xd 1 !! d8, but White is still slightly better.

13 J..xe4 dxe4 14 l£)d5 ( 14)

Better than 14 jte3, played in Belyavsky-Portisch, Tilburg 1986, when 1 4 f6 15 J_xc5 jtxc5 1 6 �xe4 'tlltxd 1 17 .1, axd 1 fxe5 1 8 l£l xc5 J..g4 1 9 11 de 1 exf4 led to equality. The idea of 1 4 �d5 i s to eliminate the e7 bishop; Black's remaining bishop will be obstructed by the e4 pawn, while

16 Najdorf Variation

White's can become active along the b2-g7 diagonal.

14 B

Or: 14 jle6

(I) 14 . . . j1d6 15 l£)c4! and Black has immediate difficulties since the only natural developing move 15 jL e6 loses a piece to 1 6 l£)xd6. Otherwise White can proceed with l£)db6, or b4 fol­lowed by 1:1. b2.

(2) 14 . . . f5 15 b41£)d7 16j;j_e3! is good for White, when both 16

1£\ f6 and 16 1£\ xe5 1 7 fxe5 jle6 lose material to l£)xe7 + and jlc5.

(3) 14 . . . f6! 15 l£)xe7 + fi(xe7 16 l£)c4 .�e6 1 7 L2\e3 f5 leads to a typical position. White continues with moves such as b3, 1:1. b2 and )llte l when White's bishop is more effective than Black's. However the opposite coloured bishops will exert a drawish tendency, particu­larly if Black can exchange knights by .£Jd7-f6-d5.

15 l£)xe7 + fi(xe7 16 f5 f6

1 6 llil,ad8 1 7 �g4 j;j_c8 ( 17 . . . f6 18 fxe6 fxe5 1 9 j;j_g5 llil, xfl +

20 �xfl lif8 2 1 �xf8 + �xf8 22 h3 is very good for White) 1 8 fi(g3 (not 18 f6 �xf6!) f6 19 l£)g4 leads to a position much like the game, except that Black's bishop is on c8 instead of f7. Although Black is exerting pressure on f5 from c8, in my view the bishop will be needed for defending the kingside and therefore the game continuation is better.

17 l£\g4 17 fxe6 fxe5 1 8 J:l.e3 l£)xe6 19

'l!l{d5 may give White a minute advantage, but the move played is much more interesting. The reply 17 jL c4 is ineffective after 1 8 � f4 and the bishop will soon be driven away by b3 in any case.

17 J:l.f7 18 '1$'e1 ?!

I decided that it was time to start developing my queenside pieces, but I should have spent just one more tempo improving my position by 1 8 a5! It looks strange to put a pawn on a black square when White's plan is to imprison Black's bishop using the pawns on c2, b3, a4, e4 and f5, but it is very useful to have the option of attacking the knight on c5. Not only may White push it away by b4 at a later stage, but by prevent­ing .. . b6 White can also attack it by jla3. The extra possibility of winning the e-pawn by getting rid of this knight (e.g. after a5, b3, 1:1. b2, tl' h4 and � ae I) would have made Black's position even more uncomfortable .

18 aS!

Black seizes on the mistake and secures the c5 square.

19 b3 lilfd8 This is not very logical, since in

a few moves Black decides that he needs a rook on f8 to support f6, so he could have saved time by 19

b6 followed by ill( ad8--d6. 20 Ab2 Ji d6 21 -rtg3 <1Jh8 22 J!lael b6 23 tth4 � f8 24 a e3 (15)

White tries a little trick; per­haps Black won't notice the threat of 25 ll h3 jtg8 26 �e5.

24 ttd8 Unfortunately he does! Now

White has to decide on a plan. At first sight his position looks very promising: since both f6 and h7 are weak, the knight on c5 cannot move without losing the e-pawn and Black has no real counter­play. However it is hard to find a concrete way White can improve his position, since almost all his pieces are already on their best possible squares. Black's potential counterplay down the d-file limits

Najdorf Variation 1 7

the action of White's rooks and although White can hope to win the e-pawn by a timely �f2xe4, this would not be enough to win the game. I could only conceive of two possible ideas, namely �f2-h3-f4 coupled with ll h3 to aim at the weak g6 square, or g4--g5 to open up the long diagonal. If White wants to play g4 the knight must move, so it seems natural to play �f2, which keeps both plans open. White therefore aims for the optimum arrangement of rook on e3 and knight on f2, with the bishop on c3 to prevent counter­play by �. d2. Black tries to prevent White arriving at this set­up.

25 <IJgl A voiding immediate back rank

problems and future long diag­onal troubles after g4. White does not need to play jtc3 straight away because 25 liil. d2 loses to 26 �xf6 jtg8 27 �h5!

25 jtg8 Now, however, 26 )(ll. d2 is a

threat. 26 jtc3

If Black does nothing White will be able to play 27 .:E}f2, so he deflects the rook away from e3.

26 � dl 27 � eel � d6

If correctly followed up this is a perfectly good defence. 27 gxel 28 �xel �d6 29 �e3 ,l4d8 30 �d l �e7 31 lii(d4! .!.!xd4 32 jtxd4 is still not completely equal, since if the Black queen leaves the defence of the e-pawn,

18 Najdorf Variation

.:&.xc5 and �xe4 wins it, while otherwise White intends '/11!\'f4-b8. Thus Black's choice was probably correct.

28 .£)£2 Sometimes an oversight is the

best chance to win! It seemed to me that this was the moment to set up the position with QJf2 and Jii\ e3, but it fails for tactical rea­sons. However, there seems to be no other winning attempt.

28 '&l'c8! Since Black obviously cannot

play �xf5 because of QJxe4, I decided to continue with my plan. There was nothing better in any case.

29 Jii\e3 After having made this move I

suddenly noticed that Black could play 29 'i'i'\'xf5 30 QJxe4 �e6! with an immediate draw, but the confident way I had made the move evidently led my opponent to accept that the pawn was invul­nerable.

29 � d5? 30 g4!

White not only achieves the position he has been aiming for, but does so in a very favourable form since the rooks on d5 and f8 are both vulnerable to the ma­noeuvre <2:1 h3-f4-g6.

30 �c7 Black could perhaps have

offered more resistance, but the twin possibilities of QJh3-f4 and g5 make this position very un­pleasant.

31 QJh3 QJd3 After 31 J..f7 32 QJf4 gd6,

the continuation 33 QJg6 + J..xg6 34 fxg6 h6 35 g5 blows up Black's position.

32 cxd3 33 QJf4 34 QJxd5 35 �el

�xc3 g5 J.,xd5

Thanks to the earlier * g I Black has no real counterplay and he is soon forced to give up.

35 �d4 36 dxe4 J.. xe4 37 �c3 �d5 38 �c4 Resigns

2 Scheveningen Variation

This line is popular with many of the world's leading players, including Kasparov, and so one would hardly expect there to be a clear way for White to obtain an advantage. The characteristic feature of the Scheveningen is Black's pawn centre at d6 and e6 covering all the central squares on Black's 4th rank. Thus Black avoids the slight weakness at d5 inherent in the Najdorf and Dra­gon systems. Behind the cover of his modest but solid pawn centre Black intends to complete his de­velopment in peace. The most common move order for Black to adopt if he is aiming for a Sche­veningen is I e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 �f6 5 �c3 d6, but Black can invert his second and fifth moves in this line.

The amount of theory in the Scheveningen has increased dra­matically since the first edition of the book, so it is no longer feas­ible to cover two major systems within the confines of a relatively short book. Therefore I have de­cided to concentrate on the Keres Attack, which is currently con­sidered a more critical test of the Scheveningen than the Classical System. The Keres Attack starts I e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4

�f6 5 �c3 d6 6 g4. This kingside pawn push aims firstly to drive the knight from f6, thereby making it harder for Black to break open the centre by . d5, and secondly to gain space on the kingside and dissuade Black from castling there. Although White often obtains good attacking chances with this system he must not neg­lect his development since Black can often break the position open by d5 and even if this loses a pawn White can find his own king stuck in the centre. In other words, a balance must be struck between furthering White's own kingside ambitions and restrain­ing Black in other sectors of the board.

There are two major options for Black after 6 g4. He may either prevent the further advance of the pawn by 6 h6, as covered in games 5, 6, and 7, or he may continue his own plans and allow the knight to be driven back to d7. In this case Black may choose 6 .. a6, 6 JJ...e7 or 6 . . . �c6. The specific choice of sixth move may not be very important because there are many transposi­tions. In game 8 we deal with lines specific to 6 �c6 (i.e. those involving an early 'Wfc7), and

20 Scheveningen Variation

the other lines may be found in game 9. Apart from these two major options, there is a third possibility for Black, namely to counterattack by 6 e5. With this move Black loses a tempo, but he hopes to prove that White's g4 has only served to weaken his kingside. This move is a lot better than its reputation, and it has recently seen a surge of popularity; it is covered in game 8.

Game 5 Nunrt-Bischoff Lugano 1986

1 e4 2 �f3 3 d4 4 �xd4 5 �c3 6 g4

c5 e6 cxd4 �f6 d6 h6 (16)

Black avoids having his knight driven away from f6. For a long time White almost played 7 g5 hxg5 8 j_xg5, Anatoly Karpov being one supporter of White's point of view. Although this con­tinuation gives White a lead in

development it has defects, not the least being that Black's rook on h8 is activated and presses down on White's weak h-pawn. Now 7 h4 is considered to give White better chances than 7 g5 and is currently the most popular line. White wants to continue with Itgl and g5, driving away the knight after all. Black can either aim for counterplay by preparing

d5 (game 6), or he can aim to nip White's kingside play in the bud with a timely . . . h5 (game 7).

7 h4 a6!? Although this is a natural

move, it has only become popular recently. The main variation is 7

�c6, and this will be exa­mined in games 6 and 7, but there is one other important possibility, namely 7 j_e7 (7 e5 8 �f5! jle6 9 g5 �xe4 10 �xg7 + j_xg7 l l �xe4 d5 1 2 gxh6 jlxh6 13 jlxh6 �xh6 14 'Wf"d2 �xh4 15 j_ b5 + � c6 1 6 0-0-0 was very good for White in Stanciu-Vegh, Ulan Bator 1986) 8 � f3 ( 17) and now:

(1) 8 . . . �fd7 9 itg3 �c6 1 0 j_e3 a6 1 1 0-0-0 'Wfc7, Ljubojevic-

Timman, Brussels SWIFT 1 986, and now Ljubojevic recommends 12 J..e2 as slightly better for White.

(2) 8 . . . g6!? 9 g5 hxg5 1 0 j_xg5 a6 l l 0-0-0 e5 1 2 �de2 J,.g4 1 3 titg3 <E:�bd7 14 f3 j_e6 15 J,.h3 J..xh3 16 K xh3 !! c8 1 7 f4 was unclear in De Wit-0 11, Groningen 1984/5. This interesting idea does not seem to have been repeated. Perhaps 1 4 f4 is better, hoping to prove that the exposed position of the g4 bishop is a liability.

(3) 8 . . . h5 9 gxh5 and now: (3a) 9 <E:�c6 1 0 j_b5 (an

attempt to exploit Black's move order; 10 <E:�xc6 bxc6 l l J,.g5 may be better, when Black has to prove that he has something better than l l <El xh5 transpos­ing to line 3b) j_d7 l l j_xc6 bxc6 1 2 e5 <E:�d5?! (12 . . . dxe5 1 3 <E:�xc6 j_xc6 1 4 titxc6+ 'ltff8 15 h6 gxh6 1 6 j_ d2 � b8 17 0-0-0 tit b6 1 8 titxb6 K xb6 is equal according to Ljubojevic) 1 3 exd6 j_xd6 1 4 J.. g5 tit b6 15 0-0-0 j_ e5 16 <El xd5 cxd5 17 c3 with an edge for White, Ljubojevic-Timman, Bugojno 1 986.

(3b) 9 <E:�xh5 1 0 J,.g5 (1 0 j_e3!? is an interesting untested idea, offering the h-pawn in return for a quick attack with 0-0-0) <E:�c6 l l <E:�xc6!? bxc6 1 2 0-0-0 j_xg5 + ?! (accepting the sacrifice turns out to be too risky; Black should develop by 1 2 . . . K b8) 1 3 hxg5 titxg5+ 1 4 'ltfbl f1Je7 (not 1 4 d5? 15 exd5 cxd5 1 6 <E:�xd5 exd5 1 7 H xd5 nor 14 . titc5? 1 5

Scheveningen Variation 21

e5! and White wins in both cases, while 14 tite5 15 j_ e2 g6 1 6 tite3 intending f4 gives White a dangerous initiative) 1 5 J..e2 g6 1 6 �xd6! f'lxd6 17 titxf7! (some­what surprisingly Black has no defence) a5 18 �dl + 'ltfe5 19 j_xh5 �xh5 20 f4+ titxf4 21 titg7 + Resigns, Sobura-Pienia­zek, Poland 1988.

8 J..g2 White abandons his plan to

play ii£ gl and g5 because after 8 Kgl d5 9 exd5 <E:�xd5 1 0 <E:�xd5 'l!ltxd5 l l J..g2 titc4 12 c3 j_e7 1 3 g5 <E:�d7 14 tite2 titxe2+ 1 5 'ltfxe2 <El b6 White had no advantage in Karpov-Kindermann, Vienna 1986.

8 <E:�c6 Or: ( I ) 8 . . . g6 9 g5 hxg5 10 J..xg5

j_e7 l l *d2 e5 1 2 <E:�de2 j_e6 13 0-0-0 .ebd7 1 4 f4 tita5 (or 14 titc7?! 1 5 fxe5! dxe5 16 <E:�d5 J..xd5 17 exd5 !I c8 1 8 K hfl ! with advantage for White, Ghinda­Bonsch, Halle 1 987) 1 5 '\tfbl <E:�b6 16 b3 with some advantage for White, Gufeld-Georgadze, USSR 1 981 .

(2) 8 . . . d5 9 exd5 <E:�xd5 10 <E�xd5 exd5 is given as unclear by ECO. However in distinction to 8 I;Igl d5, White's rook is still defending the h-pawn, so White might consider 9 e5 �fd7 10 f4, when 1 0 . . j_e7 l l h5 and 10 . . . h5 l l gxh5 look good for White, so the critical reply is probably 10

tit b6. 8 d5 9 g5 is also interesting.

22 Scheveningen Variation

9 g5 hxg5 10 hxg5 � xh1 + 1 1 J,.xh1 <21d7 ( 18)

If Black attacks the g5 pawn White continues 11 <Elxd4 12 'ljxd4 <Elh7 13 e5! <Elxg5 (13 dxe5 14 '!jh4 traps the knight) 14 exd6 j_xd6 15 ftlxg7 (15 j_xg5 'ljxg5 16 �xd6 �gl + is less clear) with a tremendous attack for no material loss.

12 J..g2 This move was the result of

lengthy thought, but even so it wasn't the best. White has very few constructive moves apart from g6 and 12 g6 <Elxd4 13 gxf7+ fJ'xf7 14 �xd4 '!jh4 (this represents the main advantage of 7 . . . a6 over 7 J,.e7, since in the corre­sponding position with the bishop on e7, White's g6 is much more effective) 1 5 J..g2 <Ele5 is obscure. White might be able to claim a slight plus after 1 6 J,.e3 'ljg4 1 7 * f l , but both kings are unhappily placed and I wanted to find some­thing clearer. 12 j}_e3 is bad after 1 2 <Elde5 threatening <21c4 (1 3 'lje2 <Elxd4 loses a pawn). Thus the only direct alternative to

12 g6 is 12 f4, but I was reluctant to create a huge empty space around my king. However after the next move White finds himself committed to f4 in any case, so it would have been better to play it at once. As a result of Nunn-Bischoff, subsequent White players took my advice and played the more accurate 1 2 f4!, with the continuation 12 . . . '!jb6 (or 12 . . . �xd4 1 3 �xd4 'ljb6 l 4'1jxb6 �xb6 15 a4!? J..d7 16 a5 'Zlc8 17 Ji..e3 Ji..c6 18 0-0-0! fJ'd7 1 9 J,.f3 �e7 20 ii..g4 with a clear plus for White, Ghinda-Vogt, Halle 1987) 1 3 <Elde2 g6 14 b3 'ljc5 (14 �c7 15 J..b2 b5 16 'Wt'd2 J..b7 17 �dl 0-0-0 18 <Ele3 J..e7 1 9 0-0-0 �b6 20 fjbl fJ'b8 21 <2:Jcl was also a little better for White in Griinfeld-Bischoff, Munich 1987) 15 �d2 b5 1 6 j}_b2 J..b7 17 0-0-0 0-0-0 18 fJ'bl 'ljf2!? (or 18 j_e7 19 �cl! f6 20 gxf6 'E)xf6 21 '-Z\d3 with an edge) 19 �el! j}_e7 20 'E)dl 'tll!'c5 21 <2:Je3 and White has a small but permanent advan­tage, Short-Kindermann, Dort­mund 1986. 12 Ji..g2 and 1 2 f4 lead to similar positions, so it is well worth studying Nunn-Bischoff even if you intend to play 12 f4.

12 g6! The point of 1 2 J,.g2 is that

White improves the position of his bishop (particularly in the g6 line given above) while Black lacks useful moves. 12 <21de5 allows 1 3 f4 with gain of tempo (13 'E)c4 14 b3 ltb6 1 5 <2\ce2), 12 . 'tll!'c7 (or J..e7) allows 13 g6

and 1 2 . �b6 1 3 �b3 leads to a loss of time after a subsequent j_e3. Black's move is the best, cutting out g6 by White and again posing the question as to how White can improve his position.

13 f4 Now the defect of 12 j_g2 is

revealed. In the analysis of 12 f4 we saw that Black generally plays

g6 in any case, while White's j_g2 is usually not necessary. Therefore Black gains a tempo, although in this type of position an extra move is not especially valuable. Now that e5 is denied to Black's knights, White threatens simply .Jte3 so Black's rely is more or less forced.

13 �b6 14 �de2

White can only complete his development by �d2 (or 'i'i1d3), b3, lit b2 and 0-0-0, so before playing f4 I had to make sure that Black couldn't use the four free tempi to harass White's central­ized king.

14 �c5 Black settles for finishing his

own development by b5 and . . .§tb7. There wasn't much

choice, as if the queen quits the b(r.gl diagonal White can play j)_e3.

15 'l!ll'd3 It would have been slightly

better to play �d2 as the position of the queen gives Black a tactical opportunity in a few moves.

15 b5 16 b3

Scheveningen Variation 23

16 j)_ e3 is countered by 16 �b4!

16 jJ_b7 17 j)_b2 l! c8

A provocative move. Black de­cides to keep his king in the centre to help c-file counterplay. After 1 7

. 0-0-0 18 0-0-0 White has a very slight advantage as in the ex­amples given after 12 f4.

18 0-0-0 � b4 Black goes in for tactics and

given his choice last move there was no other consistent continua­tion, for otherwise he has no compensation for the long-term handicap of a poor king position.

19 �d2 �xc2 20 lft xc2 b4 21 *b1 *f2!

This is the point of Black's combination. The immediate 21 . . bxc3 22 �xc3 leaves Black in a

very poor position as the exchanges have not created any weak spots in White's position to provide counterplay.

22 .§th1 bxc3 At this stage Black offered a

draw but although White must adopt the much less satisfactory recapture with the bishop, thus leaving e4 weak, I decided to play on.

23 .§txc3 �a7? A misjudgement. The main

merit of Black's combination is that his queen has become a nui­sance by taking up residence in the heart of White's position, the more so as White cannot contem­plate a queen exchange which

24 Scheveningen Variation

would relieve Black of any wor­ries about his king. Bischoff re­treats it to the passive square a8 in return for inconvenient but not really serious pressure on e4. He should have played 23 �c5 24 J..d4 t,;-h2! (24 . . �xe4 25 t,;-d3! loses a piece while 24 t,;-h4 25 �c3 e5 26 J..e3 looks good for White) when White has problems with his e4 pawn. 25 t,;-e3 e5 26 J.. b2 J..g7 creates a very awkward threat of . exf4, so White would prefer 25 <2)c3 t,;-xd2 26 H xd2 with equality.

24 j_b2 25 'ii!1e3 26 <2)c3

t,;-a8 <2)c5

26 <2) g3 with the idea of f5 is also promising.

26 J..g7 (19) After 26 a5 the reply 27

�d4 ties Black up by preventing J..g7.

19 w

27 <2)d5! Before playing 26 <2)c3 I had to

think very carefully about the tac­tics initiated by this move, since if White had been reduced to the passive 27 H et (27 H xd6 j_xc3 is

bad since Black wins after 28 j_xc3 j_xe4+ 29 J,.xe4 *xe4+ 30 'ii!1xe4 <2)xe4 or 28 f(xc3 j_xe4+ 29 j_xe4 *xe4+ 30 ffal *h i + 31 J..c l <2)xb3+ ) he would have no advantage.

27 J,.xb2 The most natural move. 27

exd5 (27 . . . J..xd5 28 exd5 J,.xb2 29 dxe6 wins) 28 J..xg7 �xe4 (28

dxe4 29 H xd6 is worse as the undefended �c5 prevents � d8, while 29 . . . �d3 allows H xd3) 29 J..xe4 dxe4 gives White a strong attack on the black squares and Black's queen is bur­ied on a8. However it is not easy to find a concrete way to con­tinue. 30 H xd6 li;!! d8! 31 �d4 .i, xd6 32 �xd6 t,;-d8 (or else J..f6) 33 'i!lff8 + *d7 34 �xf7 + wins a pawn, but in view of the opposite coloured bishops and Black's passed e-pawn this might not be enough to win. White should therefore prefer the slow build-up by 30 J..f6, with many defensive problems for Black.

28 �b6 j_xe4 + After 28 t,;-a7 29 �xc8

j_xc8 White can't take the bish­op, but 30 b4 picks up the knight instead, when Black has insuf­ficient compensation for the lost exchange.

29 f!'xb2 *a7 The only defence. Now 30 �c4

d5 31 'i!lfd4 (31 �d6+ f!'d7 and 31 j_xe4 dxe4 also seem satisfac­tory for Black) fails to 31 �a4 + ! 32 bxa4 H b8 + followed by �xd4 .and . . . J..xh l .

30 �xd6 �d3+ Black's moves continue to be

forced. 30 jj_xh 1 3 1 �d4! (at­tacking the � c8 and threatening �h8+ ) g b8 (3 1 . . 'g c7 32 �h8 + and mate at d8) 32 ll!fxc5 (threat !=! d7) li;li, d8 (32 ll!fe7 33 �d7 � a8 34 �e5 wins) 33 �e5 gives White a decisive attack, e.g. 33 . . �xd6/'iite7/�b8 34 'l¥t'h8+ or 33 .i4. b8 34 � xe6 + fxe6 35 �h8+

31 lfj a3 3 1 � xd3 is met by 31 jj_xd3.

31 �c6! I had overlooked this ingenious

defence when I played 26 �c3. Other moves lose quickly:

( 1 ) 31 . . . .i4, c5 32 !,ii xd3 �a5+ (32 . �xb6 33 j}_xe4) 33 \11b2 jj_xd3 (33 jj_xh l 34 �d4) 34 jj_c6+ '1id8 (or else a knight check wins the queen) 35 �d4 + 'flc7 36 �a8+ '1i b8 37 �d8 mate.

(2) 31 . . . "i!fc7 32 �xd3 jj_xd3 33 �xc8 �xc8 34 �xd3 �c l + 35 '1ib4 \Wxhl 36 �xa6 with a won ending.

(3) 31 . . . Axh1 32 ll!fh3 � b8 33 �xe6+ fxe6 34 \Wh8+ win­ning the queen.

32 �xc6 The best, as 32 �d5 �xe3 33

�f6 + is a draw and 32 !=! xd3 (32 �xd3 Axd3 33 Axc6+ lfj f8 is fine for Black) jj_xh I 33 �c4 is about equal.

32 �e7 + (20) 33 \WC5!

This move was a visible shock for my opponent. 33 J!i!, c5 J..xh 1

Scheveningen Variation 25

34 b4 �xc5 35 \Wxc5 is not so good as Black can avoid the exchange of queens by 35 "itd8.

33 �xc5+ 33 �xc5 34 � c8 + \Wd8 35

K xd8 + lfjxd8 transposes except for the position of Black's king, which makes no real difference.

34 I bc5 �xc5 35 b4!

White is a1mmg for a good knight v bad bishop ending, but he must avoid 35 '1i b4 �d3+ 36 '1ic3, which allows 36 �f2! Knight v knight endings are still slightly better for White, but far less promising than the position arising in the game.

35 Axh1 36 bxc5 Ad5?

With very little time left to reach move 40 Black blunders. He should have tried 36 'lle7 (not 36 . . f6 37 gxf6 <1Jf7 38 �d7 and

j}_c6 is impossible, while even if Black's king were on d8 as in the note to Black's 33rd, 36 'llc7 37 \11b4 \11c6 38 lfjc4 followed by \11d4 and �c4-e5 wins) 37 f1b4 e5! (37 . . . f6 38 lfja5 Ab7 39 �c4

26 Scheveningen Variation

followed by fib6 should win) 38 fxe5 fie6, but even here White wins: 39 .£)c4 J..d5 40 a3! j}_xc4 (there is not much choice as 40 j}_ moves allows fia5 and 40 fie7 41 2E:J e3 and 42 fia5 wins) 41 *xc4 fixe5 42 'ltJb4! (42 a4? f5 43 gxf6 *xf6 44 fid5 'ltJe7! 45 fic6 g5 46 fib? g4 47 c6 leads to an ending of � + a- Jl, v �, which should be a draw, but Black must avoid 44 . . g5? in this line, when 45 fid6 g4 46 c6 g3 47 c7 g2 48 c8 (Q) g 1 (Q) 49 �f8 + wins his queen) fi d5 (42 f5 43 gxf6 *xf6 44 fi a5 g5 45 c6 wins) 43 a4 and Black is in zugzwang.

37 12) xd5 exd5 White wins because Black's a­

pawn has moved, while White's can still advance either one or two squares.

38 fib4 fld7 39 fic3

and Black lost on time, but 39 fic6 40 'ltJd4 a5 41 a4 and 39 flc7 40 fid4 'fic6 4 1 a3 a5 42 a4 lead to the same fatal zugzwang.

Game 6 Karpov-8passky

Tilburg 1980

1 e4 c5 2 Q)f3 e6 3 d4 cxd4 4 iZ)xd4 12)f6 5 �c3 d6 6 g4 h6 7 h4 12)c6 (21)

For 7 a6 and 7 !J_e7 see game 5.

8 :l!! g1 d5

Another idea is 8 . . . 12)d7 (for 8 h5 see game 7) 9 g5 hxg5 and

now White had a good game after both 10 J_xg5 'l!!l'b6 1 1 12)b3 a6 1 2 h5 fi!c7 13 fi!e2 b5 14 0-0-0, Tseshkovsky-Zarubin, Sochi 1981 , and 10 hxg5 12) de5 11 j}_e3 j_d7 12 !J_e2 a6 13 f4, Matulovic­Simic, Smederevo 1981 -my per­sonal preference is for 1 0 j}_xg5, since Black's knight is badly placed at d7

9 .�b5 10 exd5 11 �xd5 12 j}_e3!

j}_d7 Q)xd5 exd5

Karpov's play in this game forced a reassessment of the whole line. Hitherto White had played 12 �e2 + , with rather uninspiring results, but Karpov's idea of !J_e3, "l!lfd2 and 0-0-0 launching a direct attack (even if this means giving up the h4 pawn) strengthens White's play consi­derably.

12 !J_e7 12 �xh4 13 'l!!fe2 exd4 1 4

j}_xd4 + �e7 1 5 j}_xd7+ fixd7 16 j}_ e5 H d8 is not so clear, but 13 �d2 J..e7 1 4 0-0-0 followed by

�f5 gives White a strong attack as in the game. It is curious that Kasparov recommends 12 . . . •xh4 both in ECO and in his book with Nikitin on the Sche­veningen, but in neither case does he mention 1 3 itd2, even though it was given in Informator 30.

22 B

13 itd2 (22)

13 Axh4?! 13 . . . 0-0 1 4 �f5 Axf5 1 5 gxf5

fr'h7 16 0-0-0 is also very good for White, but Black has better survi­val chances after 13 . . . �xd4 14 Axd7+ 'l!ltxd7 15 •xd4 �f6 16 •b4 �e7, when he went on to draw in Marjanovic-Cebalo, Yugoslav Ch. 1982, although he needed to defend accurately until move 64 to achieve this!

14 0-0-0 1tf6 1 4 �xd4 1 5 �xd7 + 'l!ltxd7

16 �xd4 0-0 17 f4! followed by g5 and 14 0-0 15 g5! hxg5 16 �xg5 �xg5 17 I;I xg5 �xd4 1 8 .l. dg l g6 19 •xd4 both give White a winning attack.

15 �f5 � xf5 16 gxf5 a6

Black could not castle without

Scheveningen Variation 27

losing his vital h6 pawn, but now his king is permanently pinned down in the centre.

17 �xc6 + bxc6 18 �c5!

Now White only needs to rip open the d-file by c4 to finish Black off. Although Spassky launches an ingenious counterat­tack his inability to bring the h8 rook into the game enables Kar­pov to repulse the threats.

18 K b8 19 b4 H b5

Black's only chance is to elimi-nate the deadly bishop.

20 .!:! gel + 'f;>d7 21 c4 � xc5 22 bxc5 Ag5

After 22 itb8 23 cxd5 J..g5 24 � e3 �xe3 25 fxe3 •e5 26 dxc6+ fixc6 27 itd7+ White should win easily enough.

23 f4 itf6 (23) With the point that 24 fxg5

ital + 25 fr'c2 •xa2 + 26 fr'd3 •xc4 + 27 f!'e3 hxg5 gives Black four pawns and a tremendous at­tack for the rook.

24 cxd5!

28 Scheveningen Variation

Liquidating to a winning end­ing.

24 'i'lfa1 + 25 f1c2 'i'lfxa2 + 26 *d3 'i'lfxd2 +

26 'it b3 + 27 'i'lfc3 also forces the queens off.

27 X xd2 jlxf4 Although Black has two pawns

for the exchange all White's pieces are very active and Black is unable to organize himself against the advance of the c-pawn.

28 B, a2 cxd5 29 �:I xa6 h5 30 <11d4 h4 31 f1 xd5 �b8 32 f6 gxf6 33 � xf6 jlg3 34 � xti + f1d8 35 � f8 + Resigns

As 35 f1d7 36 c6 + *c7 37 K e7+ * b6 38 B, xb8+ jtxb8 39 B. b7 + ends the game.

Game 7 Nunn-Sax

Rotterdam 1989

1 e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 �f6 5 �c3 d6 6 g4 h6 7 h4 �c6 8 Jlii g1 h5

This is currently the most fashionable move. White's best reply is to take on h5 and since the recapture �xh5 leaves the knight badly placed, Black nor­mally returns it to f6. The net

effect of this is to reach a position similar to that after 6 g4 h6 7 g5 hxg5 8 jlxg5, but with White hav­ing the two extra tempi h4 and �gl . This might seem to be a great improvement, but in fact the disadvantages almost balance the advantages. White has two prob­lems: firstly the h-pawn can become weak without the defence of the rook and secondly Black's

� b6 effectively pins the f­pawn against the undefended � g l , so it is hard for White to play f4.

9 gxh5 <E)xh5 Black players have been known

to try 9 �xh5 1 0 �g5 �h8, but this is quite pointless. Black reaches the same position as in the main line, but having forfeited the right to castle kingside.

24 B

10 jt g5 (24)

10 �f6 The alternative is the immedi­

ate 10 �b6, which attempts to avoid the loss of time inherent in 10 �f6. The problem is that the knight is genuinely badly placed at h5, so Black gains

nothing by keeping it there. After 1 0 �b6 1 1 � b3 a6 1 2 jte2 g6 ( 1 2 � f6 transposes to the analysis of 1 1 . . . � b6 in the main line) 1 3 �d2 Black may play:

( l ) 13 . . . jte7 14 K g2! (White need not offer his f-pawn by cas­tling immediately) jtd7 1 5 0-0-0 I;I c8 1 6 '�i b l "i!Jic7 1 7 a3 b5 1 8 jtxb5 axb5 1 9 �xb5 �b8 20 jtxe7 �xb5 2 1 jtxd6 with a clear plus for White, Lobron-Marjano­vic, Reggio Emilia 1 985/6.

(2) 13 . . . jtd7 14 K g2! �c7 1 5 0-0-0 b 5 1 6 a3 �e5 1 7 �d4! )[i( h7 1 8 f4 � c4 1 9 jtxh5 1!l( xh5 20 � e 1 jtc6 2 1 �d5 jtxd5 2 2 exd5 e 5 23 �d3 JJ..e7 24 �d4! and again White stands well, Motwani­Roca, Dubai 01. 1 986.

(3) 13 'fltc7 (this is even worse than the lines above because White need not spend time on K g2) 14 0-0-0 b5 1 5 a3 jtd7 jtxb5! axb5 1 7 �xb5 �b8 1 8 �xd6 + jtxd6 19 �xd6 �xd6 20 K xd6 with advantage to White, Govedarica-Mokry, Trnava 1 987.

ll jte2 This flexible move, which pre­

pares a possible h5, has gained in popularity, even though White sometimes has to sacrifice his f2 pawn after �d2 and 0-0-0. In reply the immediate �b6 turns out badly because h5-h6 becomes strong, so Black nor-mally bides his time with a6.

Or: l l a6

( 1 ) l l . . . jte7 1 2 �d2 � xd4 1 3

Scheveningen Variation 29

�xd4 'itb6 14 JJ.. b5 + \t>f8 ( 1 4 . . . jtd7 1 5 J..xd7 + *xd7 1 6 �d2 is good for White) 1 5 �xb6 axb6 1 6 0-0-0 e5 was unclear i n Ljuboje­vic-Adorjan, Linares I 985, but it is hard to believe that there is no way White can exploit the weak b­pawns. Perhaps 1 6 � a4! is best.

(2) l l . . . �b6 1 2 � b3 a6 1 3 h5 �c7 (or I 3 . . . JJ..d7 14 h6 K h7 1 5 �d2 �g8 1 6 jte3 �c7 I 7 hxg7 .1. xg7 I 8 0-0-0 with a plus for White, Alzate-Frias, Dubai 01 1 986) 14 h6 �d7 I 5 hxg7 jtxg7 1 6 �d2 JJ.. f8 17 0-0-0 b5 1 8 a3 Ab 7 19 � h I K xh I 20 K xh I � ce5, Hellers-Sax, New York Open I987, and now 2 I .1, h8! �g6 22 1!1 h7 would have given White a very dangerous attack.

12 �d2 JJ..d7 Playing for b5 is a new idea,

but the critical continuation is probably he older 1 2 'ltb6 1 3 � b3 (25) and now:

( l ) 13 . . . 'ltc7 (this doesn't make much sense; Black may as well keep his queen on b6 and dare White to offer his f-pawn) 14 h5 � xh5 ( 14 . . . b5 15 a3 jt b7 I 6

30 Scheveningen Variation

O-O-O b4 1 7 axb4 � xb4 1 8 trd4 d5 1 9 h6! !;l xh6 20 jtxh6 e5 21 .zt f4! won for White in Luthar-Bonsch, East German Ch. 1 989) 1 5 � hi g6 1 6 .ztxh5 gxh5 1 7 "i!i!(e2 b5 1 8 ». xh5 K xh5 1 9 trxh5 .zt b7 20 0-0-0 b4 2 1 � e2 � c8! 22 � d2 <2Je5 23 <2Jg3! aS 24 f4 with a clear plus for White, A. Rodriguez­Douven, Amsterdam 1 987.

(2) 13 . . . _ztd7 14 h5 <2J xh5 (Black should take everything on offer; the passive 14 0-0-0 1 5 h6 J� h7 1 6 0-0-0! !fJ_e7 1 7 !fJ_ e3 trc7 1 8 !;l xg7 � xg7 1 9 hxg7 � g8 20 J� g l <2J e5 2 1 .ztd4 <2Jg6 22 lit h6 !fJ_ c6 23 !fJ_ d3 was very good

20 exd6 <2J e5 2 1 <2Jd4 trg4 22 .zte7 <2Jc4 23 trd3 <2Je5 24 tre3 � c4 25 trd3 �e5 there was only a draw by repetition. It seems to me that White's best chance is to follow the Rodriguez-Grooten game, but instead of 21 � c5 play 2 1 <2Je4! white intends a combination of <2J f6 and <2J bc5, while after 2 1

0-0-0 22 � f6 � dh8 23 � xd7 *xd7 24 � c5 + * c8 25 _ztf3 White has dangerous threats.

l3 0-0-0 b5 14 � xc6!

The immediate 1 4 tre3 is met by 14 trb6.

14 JJ..xc6 for White in Korolev-Agzamov, USSR 1 983) 1 5 � h 1 g6 1 6 0-0-0 trxf2 (once again the crucial 26 move; 1 6 trc7 1 7 !J;j_xh5 gxh5 B 1 8 tre2! JJ..e7 1 9 JJ..xe7 <2J xe7 20 � xh5 � xh5 2 1 trxh5 <2Jg6 was played in Tseshkovsky-Mokry, Trnava 1 986, and now 22 trh2!

15 fte3 (26)

was promising for White) 1 7 e5! trf5! ( 1 7 � xe5 18 <2Je4 trf5 1 9 tre3! JJ..c6 20 <2J bd2! gives White a crushing attack) and the evalu­ation of the whole line depends critically on this position. In A. Rodriguez--Grooten, Dieren 1 987, White played 18 exd6 trxg5 19 trxg5 JJ..h6 20 trxh6 � xh6 2 1 <2Jc5 <2Je5! 2 2 <2J 3e4 0-0-0 23 <2J xd7! *xd7 24 � h3 .)I e8, and although Rodriguez gives 25 b4! as unclear, in my opinion this line is not convincing. Therefore in Chandler-Hellers, Thessaloniki 01. 1 988 White tried 18 !J;j_xh5, but after 1 8 . . . � xh5 19 � xh5 gxh5

White has the unpleasant threats of 1 6 e5 and 1 6 <2Jd5, while after 1 5 tra5 White can afford to take time out for 1 6 * b l because 1 6 b 4 may be met by 1 7 <2Jd5 <2Jxd5 1 8 exd5 jtxd5 1 9 � xd5 trxd5 20 JJ..f3 . Black is therefore forced into the unpala-table 1 5 trc7.

15 16 <2J d5 17 exd5

trc7 JJ..xd5 e5

After 1 7 g c8 White replies 1 8 c3, and Black has achieved nothing positive, while giving away the chance to castle queen­side. 1 7 . . . e5 is better, but even so White's lead in development and Black's exposed king give him dangerous attacking chances.

18 *bl?! Chess laziness. Of course * b I

is a desirable move, but by giving Black a free tempo White's attack loses much of its momentum. The immediate 1 8 f4! was correct, when White has a clear advan­tage. Now by accurate defence Black survives the immediate cri­sis.

18 .:2)h7 Eliminating the g5 bishop

makes it easier to flee with the king, should that prove necessary, and ultimately the opposite-col­oured bishops might provide a drawing mechanism.

19 f4 .:2\xgS Not 1 9 f6 20 fxe5 dxe5 2 1

J.,h5 + 'fr'd8 2 2 d 6 J.,xd6 23 g xd6 + '!ifxd6 24 g d l '!ifxdl + 25 J.,xd l fxg5 26 '!ifxe5 with an excellent position for White.

20 K xg5 A difficult choice, as although

White may win a pawn by 20 fxe5 (20 hxg5 g6 21 f5!? is probably also slightly better for White) dxe5 2 1 !! xg5 0-0-0! 22 li xe5 j}_ d6, Black completes his de­velopment and the opposite­coloured bishops become an im­portant factor.

20 K c8

Scheveningen Variation 31

Now 20 . . . 0-0-0 is bad because of 2 1 a4!, so Black must adopt a different defensive plan.

21 c3 '!ifc5 Of course this is only possible

when White had not exchanged on e5. Black gains time to reorga­nize his defences.

22 '!ifg3 exf4 23 '!ifxf4 K c7! (27)

An excellent move. Black's rook covers the vulnerable square f7 and when it arrives at e7 the attack on the bishop will gain more time for Black.

24 a4! White's only chance to make

something of his waning initiative is to create a new target on the queenside.

24 25 26

Not 26 27

'1l e7 j}_d3 g6 axb5 axb5

J.,h6? 27 '!itf6. '!itd4

The ending now represents the best winning chance for Whit�. Although White's gain of a pawn is only temporary, the passed b­pawn combined with an exposed

32 Scheveningen Variation

Black king gives White a nagging advantage.

27 'ltxd4 28 Jl_xb5 + ftd8 29 !;l xd4 K e1 + 30 ft a2 J..e7 31 � gg4 l;l h1

The h-pawn is doomed, so White switches to harassing Black's king. In this the opposite­coloured bishops prove a big help.

32 .1, a4 !:[ 8xh4 33 Ill xh4 i: xh4

33 . . . J..xh4 34 H a7 is worse, since White threatens the f-pawn directly and the d-pawn indirectly via .ii! d7 +

34 � a8 + ? A careless check which drives

the Black king to a better square. White should have cut the king off by 34 !:[ a7! (threat � d7 + ) � e4 (after 34 g5 35 l:l d7 + '*e8 36 i: xd6 + ftf8 37 i: c6 the d-pawn becomes dangerous) 35 jtc6! with some winning chances because Black cannot challenge White to a pawn race (35 . . . g5 36 b4 g4 37 b5 g3 38 b6 wins because the mate threat gains a tempo).

34 '*c7 35 J..c6

35 .l, a7 + flb6 36 H xe7 fj xb5 37 )ll xf7 '\tr'c5 is an easy draw.

35 .l, e4! White is effectively a tempo

down over the above line since after 36 l!l( a7 + '*b8 White must waste time with his rook. This tempo makes all the difference and Black can now draw comfor­tably.

36 b4 jtf6 37 H a7 + .b8 38 H xf7 j_xc3 39 b5 H b4 40 ffa3 .ill b1 41 l! b7 + fjc8 42 b6 j_d4 43 K c7 + .b8 44 !:[ b7 +

Draw

Game 8 Karpov-Dorfman USSR Ch. 1976

1 e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 �f6 5 �c3 d6 6 g4 �c6 (28)

I must confess to have taken some liberties with the move ord­er of Karpov-Dorfman, which actually continued 6 . . . Jte7. We transpose back in a few moves, but this move order makes it easier to explain the proposed repertoire. Of the sixth move alternatives, we only consider 6 • . . e5 here; the others may be found in game 9. After 6 . e5 (played by Murei and Suba) 7 j_b5 + j_d7 8 j_xd7 + '/txd7 (8

�bxd7 9 � f5 is awful for Black) 9 � f5 h5 10 f3 hxg4 1 1 fxg4 K h3!? 1 2 'lte2 (better than 1 2 �e3 'ltc6 1 3 '£)ed5 �bd7 1 4 �e2, Howeii-Suba, London 1 988, and now 1 4 �xd5 1 5 � xd5 � b6 is unclear) 'ltc6 ( 1 2

� xc3?! 1 3 bxc3 'ltc6 1 4 0-0 '*xe4 1 5 �xe4 <2lxe4 1 6 &J[ e 1

L£J c5 1 7 Jl. a3 '£l ba6 1 8 I: ad I Jf!i d8 19 h4 g6 20 '£J e3 is good for White, Wahls-Mainka, Luxem­bourg 1 989) 1 3 0-0 '£l bd7 14 g5 � xc3, Tisdaii-Suba, Preston 1 989, and now 1 5 '£J xg7 + Jl.xg7 16 gxf6 � xc2 ( 1 6 '£l xf6 I 7 bxc3 tyxe4 1 8 tyd3 d5 1 9 tyg3 is promising) 1 7 fxg7 *e7 ( 1 7 0-0-0 1 8 trg4) 1 8 Jl.g5 + ! f6 1 9 tyf3 i s very good for White.

7 g5 '£Jd7 8 h4 a6

Or 8 '£Jxd4 9 tyxd4 L£Je5 1 0 Jl.e2 '£Jc6 1 1 �d3 Jl.e7 (or 1 1 a6 1 2 Jtf4 trc7 1 3 0-0-0 '£Je5 14 '*d4 Jl. d7 15 h5 with strong pres­sure for White, Lutikov-Malich, Leipzig 1 977) 1 2 Jtf4 (were it not for this move, exposing the weak­ness of d6, Black's scheme would be viable-this is one of the few situations in the Sicilian where a direct attack on d6 works) 0-0 1 3 0-0-0 e5 1 4 Jl. e3 Jl. e6 1 5 <21 d5 tya5 16 a3 Jtxd5 1 7 tyxd5 tyxd5 1 8 1:: xd5 with the type of ending Sicilian players have nightmares about, Nunn-Jansa, Dortmund 1 979.

9 Jte3 'f!lc7

Scheveningen Variation 33

The repertoire proposed in this book involves meeting lines with­out . . . tyc7 by the new idea Jl. c4 (see next game), while if Black plays . . . "W!fc7 preventing Jtc4 then White replies with tre2 and 0-0-0. Unfortunately White can­not guarantee to play Jl.c4 in every line, because if White plays Jl.c4 without preparing it by Jl.e3 then the reply <2!de5 is very awkward.

10 tre2 Karpov's move is very logical

in that it prepares queenside cas­tling as quickly as possible, while the fl bishop and h 1 rook are left at home since it is not yet clear which square is best for these pieces. At e2 the queen sets up tactical chances down the e-file and avoids attack from a !Jiack knight arriving at c4.

10 Jl.e7 We are now back in Karpov­

Dorfman. 1 1 0-0-0 b5

Tactical ideas for White are already in the air, for example 1 2 '£Jf5 exf5 1 3 '£Jd5 trd8 14 exf5, but although this is quite good for White Black can improve by 1 2 '£Jf5 b4!

12 '£lxc6 trxc6 13 Jl. d4 b4 ( 29)

Black forces White to sacrifice on d5, but this move was itself virtually forced as 1 3 0-0 14 Kg I gives White a crushing at­tack, e.g. 1 4 b4 1 5 <21d5 exd5 1 6 exd5 'f!l xd5 1 7 trxe7 'f!lxa2 1 8 g6 hxg6 1 9 � xg6 and wins.

34 Scheveningen Variation

14 � d5 15 j_ xg7 16 exd5 17 J..f6

exd5 � g8 'W/c7

The position of White's bishop at fl is shown up as a defect since 1 7 !! e l l£l e5 1 8 j_xe5 dxe5 19 f4 exf4 achieves nothing when d6 is impossible. If the other rook could come to e l Black would be finished.

17 '£)e5 18 J..xe5 dxe5 19 f4

Now White wins a third pawn for the piece since 1 9 . . . e4 fails to 20 d6 j_xd6 2 1 'Wfxe4 + Black's king must remain stuck in the centre so one must consider Kar­pov's sacrifice correct, although in the subsequent play Black's resourceful defence almost saves the game.

19 J.f5 20 J.h3

White takes time out to neutral­ize Black's counterplay as 20 fxe5 at once allows the unclear 20 � c8 21 � h2 'l!lfa5.

20 J.xh3

21 li:i( xh3 ![ c8 22 fxe5

After this Black activates his queen and Karpov is obliged to play with extreme accuracy to maintain his advantage. In his notes Karpov suggested 22 b3 to prevent the following manoeuvre.

22 �c4! 23 � dd3 .f4 +

23 . �xa2 24 d6 (threat d7 + ) g c4 (24 g c5 25 �f2 and 26 dxe7) 25 dxe7 � a l + 26 * d2 �xb2 27 g d8 + fijxe7 28 � d7 + fj>xd7 29 �xc4 and 23 g xg5 24 hxg5 'itxa2 25 d6 j_xg5 + 26 lii,\ he3 !i( c4 27 'itg2 are good for White.

24 *b1 � c4! The rook follows the queen's

path with the aim of causing White some problems on the back rank.

25 d6 � e4 26 � he3 Ji! xe3 27 � xe3 �xh4 (30)

If Black attempts to save his bishop by 27 J.d8 (27 Af8 28 'l!lfxa6 i s even worse) he is crushed after 28 Ji! f3 'it g4 29 e6

fxe6 30 d7 + winning the queen, so he quite rightly decides to grab as many pawns as he can while White is taking his bishop.

28 trf3! Naturally not 28 dxe7 at once

since Black exchanges queens by 28 trh1 + White's advantage lies in the insecure black king, which causes trouble even when Black restores material equality.

28 'ltxg5 28 jlxg5 29 e6 fxe6 30

� xe6 + fJ'd8 (30 . . . fJ'd7 3 1 'lt f7 + ) 3 1 'ltc6! and 28 � xg5 29 'ltc6 + fJ'f8 30 dxe7 + fgxe7 31 a3 win for White.

29 K el 29 'ltc6 + fJ'f8 30 dxe7 + trxe7

3 1 'lth6 + � g7 is a little better for White and this may in fact be his best line.

29 'ltg2? 29 . . . 'ltg4 was better, when it

is far from clear if White can do more than draw.

30 trf5 K g6 31 K fl 'ltd5 32 dxe7 fj'xe7

Material equality is re-estab­lished but Black's king position makes his defensive task difficult. Detailed analysis of this position would take us too far afield, but Black does not seem to have any real improvements hereafter and the task of defending both his king and his pawns soon over­stretches his forces.

33 'ltf4 34 '!th4 + 35 'ltxh7

aS fj'e8 'ltf3

Scheveningen Variation 35

36 'lth8 + *e7 37 �h4 + fj'e8 38 �c4 'ltb7 39 b3

One of the most impressive features of this game is the way Karpov managed to conduct all the tactical operations with a vul­nerable black rank. Many players, through nervousness or laziness, would have wasted a tempo on b3 earlier, and this might well have cost the game (note that although b3 was a good idea at move 22, the point was to prevent the 'ltc4-f4 manoeuvre rather than to give the king air).

31 w

39 � e6 (31)

40 K gl?! Perhaps Karpov assumed that

the exposed king must succumb quickly to the combined attack of White's queen and rook, but the task is much more difficult than appears at first. I suspect that if Karpov had realized this he would not have been so hasty in giving back the pawn, because he could have waited for a more favourable moment.

36 Scheveningen Variation

40 J� xe5 41 K g8 + *e7 42 trh4 + *d7 43 tyf6! H e7 44 tyf5 + *d6 45 trxa5 H e5

45 tye4 would have lasted longer, but the result is not in doubt.

32 B

46 47 48 49 50

tyd8 + * b2! � f8 trc8 + tyc4 +

'fle6 f6 trg7 'fld5 Resigns

Game 9 Nunn-Thorsteins

Lugano 1986

1 e4 c5 2 �f3 e6 3 d4 cxd4 4 � xd4 �f6 5 �c3 d6 6 g4 (32)

6 jLe7 We also need to consider those

lines which are special to 6 . . a6. After 6 . . . a6 7 g5 <E)fd7 8 h4 b5 (8

�c6 9 jLe3 will lead to game 8 or to the main line of this game) 9

a3 A b7 (9 �b6 is less accurate since after I 0 h5 White may meet 10 jLe7 with the dangerous piece sacrifice 1 1 tr g4 e5 1 2 .:2! f5 g6 1 3 hxg6 fxg6 14 jLe3! gxf5 1 5 exf5; in Nunn-Walden, Notting­ham 1 983, the continuation 1 0 . �8d7 1 1 � h3 �c5 1 2 g6 f6 1 3 � g3 e5 1 4 �c6 trc7 1 5 �b4 was good for White) 10 Ae3 ( 1 0 h5 at once is not so good since 1 0 §Le7 awkwardly attacks the g5 pawn) � b6 (10 . . . �c6 1 1 tre2 �de5 1 2 0-0-0 �c4 1 3 <E) xc6 jLxc6 14 f4 �a5?! 1 5 �d5! gave White a crushing attack in Alex­ander-Lundholm, corr. 1 970--1 , while 10 . . . �c5 1 1 tyg4!? � bd7 12 0-0-0 � e5 1 3 tyg2 �c4 14 Axc4 bxc4 15 h5 intending g6 is good for White according to Boudy) 1 1 h5 �d7 1 2 � h3! (the discovery of this move led to a reassessment of many lines in the Keres Attack-the point is to defend the rook in preparation for g6) Black may play:

( I ) 12 . . . �e5 1 3 g6 hxg6 14 hxg6 K xh3 15 gxf7 + �xf7 (Black avoids the displacement of his king) 1 6 A xh3 and Black's weak pawn at e6 gives White a good game, Torre-Vogt, Polanica Zdroj 1 977.

(2) 12 . . . d5 1 3 g6 e5 reaches a complex position. In two games White played 14 �f5, and after 14 . . . hxg6 15 hxg6 � xh3 16 gxf7 + *xf7 1 7 Axh3 .:2l f6 1 8 Axb6 �xb6 1 9 �xd5 A xd5 20 exd5 � d8 2 1 � e3 A c5 22 A e6 + , Yakovic-Espig, Leipzig 1 986 and

14 . . . d4 1 5 h6 hxg6 1 6 hxg7 JJ..xg7 1 7 <2\ xg7 + 'ffj e7 1 8 JJ..xd4! exd4 19 'it xd4 K xh3 20 0-0-0! , Fogarasi-Espig, Budapest Open 1 987, White had a clear advan­tage. However, as Lukacs and Hazai point out, the improvement 14 . . . <2\f6! leads to a totally unclear position. I therefore suggest 14 exd5!? exd4 ( 1 4 . . . <2\ xd5 1 5 <2\xd5 JJ..xd5 1 6 <2\xb5) 1 5 JJ..xd4 with an enormous at­tack, e.g. 1 5 . . <2\xd5 1 6 <2\ xd5 JJ..xd5 17 )J:! e3 + JJ..e6 18 'ltf3! winning.

(3) 12 . . . JJ..e7 1 3 g6 JJ..f6 14 tyg4 'lte7 15 gxf7 + flxf7 16 f4 � hc8 ( 1 6 � ac8 1 7 f5 e5 1 8 <2\ e6 � hg8 1 9 h6! g6 20 � g3 is also good for White) 1 7 f5 <2\e5 1 8 fxe6 + fi f8 1 9 t)'g2 � xc3! 20 bxc3 <2\a4 2 1 lli, g3 ! <2\xc3 22 il_d3 'Wfc7 23 e7 + ! turned out well for White in Fernandez Garcia-D. Cramling, Barcelona 1 986.

(4) 12 . . . <2\c5 3 g6 'Wfe7 14 gxf7 + 'Wfxf7 15 'W/g4 JJ..c8 1 6 e5! dxe5 17 � f3 'ltb7 18 <2\de2! i#fc7 19 0-0-0 JJ..d7 20 JJ..g5 aS? 21 <2\g3! with a massive advantage for White, ldelstein-Barash, Israel 1 987.

7 g5 <2\fd7 8 h4 <2\c6 9 JJ..e3 0-0

If Black plays 9 a6 White may again reply 10 il_c4, with play similar to that in the main line. Black has one more possi­bility, namely 9 <2\ b6 10 f4 and now:

( I ) 10 . . . 0-0 1 1 )!lJtf3 d5 1 2

Scheveningen Variation 3 7

0-0-0 JJ..d7?! 1 3 exd5 exd5 14 JJ..d3! JJ..c5 15 <2\xc6 JJ..xe3 + 16 trxe3 bxc6 1 7 h5 � e8 1 8 trd4 'Wfe7 1 9 h6 life3 + 20 tyxe3 � xe3 21 hxg7 with a decisive advantage for White, Lanka-Strautinsh, corr. 1 986.

(2) 10 . . . d5 1 1 JJ..b5 ( 1 1 e5!? is possible) JJ..d7 12 exd5 exd5 1 3 'W! f3 JJ..b4 1 4 0-0 JJ..xc3 1 5 bxc3 <2\xd4 1 6 JJ..xd7 + 'ltxd7 1 7 JJ..xd4 0-0 1 8 f5 <2\c4, Tseitlin-Lukin, Leningrad 1 987, and now 1 9 � fl! intending h5--h6 should be good for White.

10 JJ..c4!? (33)

An intriguing new idea. White plans a similar strategy to that in the Velimirovic Attack, which arises after 1 e4 c5 2 �f3 <2\c6 3 d4 cxd4 4 <2\xd4 <2\ f6 5 Qlc3 Q�c6 6 JJ..c4 e6 7 JJ..e3 JJ..e7 8 'i/!i'e2 0-0 9 0-0-0, one typical variation being 9 . . . a6 1 0 il_b3 fltc7 1 1 g4 <2\xd4 1 2 Ji;!t xd4 b5 1 3 g5 <2\d7 14 h4, with an obvious similarity to Nunn-Thorsteins. One major dif­ference stands out. In the Velimir­ovic Attack White cannot play g4

38 Scheveningen Variation

under the most favourable pos­sible circumstances; he must either waste time preparing g4 by l l .1, hg I , or he must reconcile himself to the relatively unfavour­able recapture with the rook on d4 (the reason being that l l g4 <2}xd4 1 2 J_xd4 fails to 1 2 . e5). In the Keres Attack position White has already achieved g4-g5 without playing K hg 1 , and there is no reason why he should not take on d4 with his bishop. There­fore if Black plays just as in the Ve1imirovic Attack, he will end up with an unfavourable version of that system, and White's kingside breakthrough by h5 and g6 will have added force. On the other hand White is two moves away from castling, so the question is whether Black can do something quickly while White's king is still in the centre. It is worth adding that 1 0 'ii!i'e2 is perfectly playable here, just as in Karpov-Dorfman above, but having played 0-0 instead of �c7 improves Black's prospects.

10 <2}xd4 Perhaps the most natural at­

tempt to exploit White's king position is to open the centre by 10 <2} b6 1 1 J_ b3 d5, but after 1 2 �e2 it is not clear what Black's

next move is. 12 . . . e5 loses a pawn and 12 . . . <2}xd4 1 3 J_xd4 dxe4 14 0-0-0 followed by <2}xe4 prepares a sacrifice on f6. Maybe 12 . . . jtb4!? is best, with an unclear position.

1 1 �xd4?!

Inconsistent. White should stick to his idea and play 1 1 J_xd4. I rejected this because of 1 1 d5 12 exd5 exd5 13 <2} xd5 .!!K e8, but a little more thought would have shown that after 1 4 <2}e3 ! Black has no real compensa­tion for the pawn. Indeed, after 1 4

�c7 (intending <2} e5) White can go over to the attack himself by 1 5 J_xf7 + *xf7 1 6 �h5 + \fl f8 1 7 J_xg7 + *xg7 1 8 <2} f5 + with a large advantage.

1 1 a6?! Black misses his chance. 1 1

<2}e5 would have forced White's bishop to abandon the active b3-g8 diagonal and after 1 2 J_ e2 Qlc6 1 3 �d2 a6 the position would have been roughly level.

12 0-0-0 Now White can set up the type

of position he is aiming for. 12 b5 13 J_ b3 <2} c5 14 f4

Defending g5 m preparation for h5-h6.

14 �aS Black intends to meet the ad­

vance of the h-pawn by e5, relying on a tactical point to se­cure the e5 square, but unfortu­nately White has a tactical counterpoint!

15 h5 b4 16 h6 e5 (34)

White cannot take twice on e5 because of the discovered attack along the 5th rank, but by means of a queen sacrifice White can convert his lead in development

into a crushing attack against Black's king.

17 � d5! � xb3 + Black must remove one attack­

ing piece. The immediate capture of the queen leads to mate after 1 7

exd4 1 8 <i)xe7 + \trh8 1 9 hxg7 + ltixg7 20 J..xd4 + f6 2 1 gxf6 + J1l; xf6 2 2 H dg l + \ftf8 23 Jiit g8 + ltixe7 24 H xh7 + H f7 25 a xf7.

18 axb3 j}_xg5 Black decides to decline the

offered queen. After 1 8 . . . exd4 1 9 <E::J xe7 + ltih8 20 j_xd4 f6 (20

� g8 2 1 hxg7 + H xg7 22 Ii!, xh7 + mates) 21 g6! Black is strangely helpless against the threat of 22 hxg7 + ltixg7 23 Ii!, xh7 mate, for example 2 1 H g8 (21 . . • gxh6 22 K xh6, 21 . . . hxg6 22 <E::Jxg6 + ltig8 23 h7 + ltif7 24 <E::Jxf8 and 21 . . . >�tal + 22 ltid2 'tltxd l + 23 ltixd l j}_g4 + 24 ltid2 hxg6 25 <E::Jxg6 + \t'h7 26 <E::J xf8 + � xf8 27 hxg7 + 'iJtjxg7 28

Scheveningen Variation 39

K g l are no better) 22 hxg7 + f;jxg7 (22 K xg7 23 :;: xh7 + K xh7 24 j}_xf6 + K g7 25 J;� h l + mates) 23 K xh7 + '1Jf8 24 X f7 + 'lle8 25 <E::Jxg8 j}_e6 26 <E::Jxf6 + 'lld8 27 .I f8 + and in return for the queen White wins almost all Black's pieces.

19 fxe5 White must not repeat the offer

because after 1 9 fxg5 exd4 20 <E::Je7 + \t'h8 he cannot play 2 1 J..xd4 due to 2 1 >�txg5 + .

19 J..xe3 + After 1 9 dxe5 20 >�txe5

J..xe3 + 21 fJ b l White will at the very least win Black's queen.

20 >�txe3 g6 Or else hxg7 wins instantly.

21 >�tg5 f6 Black must jettison material to

meet the threats of 22 �e7 + and 22 >�tf6.

22 <E::Je7 + Even stronger than taking on

f6. After 22 'llh8 23 � xg6 + hxg6 24 >�txg6 K a7 (or 24 H g8 25 >�txf6 + \t'h7 26 >�tf7 + '1Jh8 27 h7 >�t a l + 28 'lld2 ll. g2 + 29 ltie3) 25 exf6 there is no defence to the threat of f!/g7 + . Black has only one alter­native, but then it is White who can exploit a line-up on the 5th rank.

22 23 e6 +

*ti Resigns

3 Classical Variation

In the first edition of this book I christened the line 1 e4 c5 2 � f3 �c6 3 d4 cxd4 4 � xd4 � f6 5 �c3 d6, which can also occur with the move order 2 . . . d6 and 5 . . . �c6, with the name 'Classical Variation' This nomenclature seems to have caught on, so I will keep the name in this edition. The line I am recommending against the Classical is 6 jlg5, called the Richter-Rauzer Attack even though the treatment used today doesn't st!em to owe anything to Richter. This line is very common in practice, so there is a large body of theory. In general I will keep to the main lines in the proposed repertoire, but where there are interesting sidelines I will give them a brief mention.

The idea of 6 jlg5 as it is played today is based on a quick '{lfd2 and 0-0-0, exerting pressure down the d-file and restraining Black from active play in the centre. Black's most solid reply is the natural 6 e6 7 ttd2 jle7 8 0-0-0 0-0, but despite its solid appearance it can often lead to sharp tactical play. This is covered in game 10. Sometimes Black players postpone 0-0 so as to delay exposing the king to a possible pawn storm. The line 6

. e6 7 tl!l'd2 a6 8 0-0-0 jtd7 is the subject of game 1 1 while 6 . . . e6 7 �d2 a6 0-0-0 h6 appears in game 12 . Finally some players have experimented with the omission of 6 e6, not fearing the dou­bled pawns resulting from jlxf6, and the most popular of these ideas, 6 jld7, forms the basis of game 13 . Unusual lines involv­ing e6 are dealt with in game 1 0, while the others are in game 1 3 .

Game 10 Anand-lnkiov Calcutta 1986

1 e4 c5 2 �f3 d6 3 d4 cxd4 4 �xd4 �f6 5 �c3 �c6 6 .zlg5 e6

Other moves are considered in game 1 3 .

7 �d2 (35) J..e7 7 a6 appears in games 1 1

and 1 2. There are two other alter­natives:

( I ) 7 . . . h6 (7 j_ d7? 8 � db5 is just a mistake) 8 j_ xf6 gxf6 (8

�xf6 9 �db5 and 1 0 0-0-0 wins the d6 pawn) 9 0-0-0 (playing for 0-0 is also slightly better for

White, e.g. 9 .1. d 1 a6 1 0 j}_ e2 h5 1 1 0-0 J..d7 1 2 � b3 Y!fc7 1 3 f;h1 0-0-0 1 4 f4, or 9 J..e2 h5 1 0 0-0 a6 1 1 f; h l ! J..d7 1 2 f4 Y!fc7 1 3 .l. f3! � xd4 1 4 Y!fxd4 j_e7 1 5 I;� d 1 h4 1 6 K fd3 K d8, van der Wie1-J. Piket, Leiden 1 986, and now 1 7 J..f3 gives White an edge) a6 1 0 f4 j_d7 1 1 j}_e2 h5 ( 1 1 Y!tb6 1 2 j_h5 Y!fxd4 1 3 Y!fxd4 �xd4 1 4 .l, xd4 .H g8 1 5 g3 j_e7 1 6 .i. fl j_c6 1 7 f5 .A g5 1 8 j_e2 with ad­vantage, Bondarevsky-Botvin­nik, USSR Ch. 1 95 1 ) 1 2 f; b 1 Y!fc7 (or 1 2 . . . Y!lb6 1 3 � b3 0-0-0 1 4 K hfl �a5 1 5 B. f3 �xb3 1 6 axb3 f;b8 1 7 �a4 Y!fa7 1 8 f5 with a clear plus for White, Keres­Botvinnik, Moscow 1 956) 1 3 .A hfl 0-0-0 ( 1 3 J..e7 14 K f3 �xd4 1 5 Y!fxd4 Y!fc5 1 6 Y!td2 j_c6 1 7 � e3 Y!fa5 1 8 a3 H d8 1 9 j_c4 gives White an edge, Liberzon­Botvinnik, USSR 1 967) 14 � b3 f; b8 1 5 .1. 0 J..e7 16 J,! h3 h4 1 7 Y!fe1 and Black has not yet equa­lized, Vasyukov-Shamkovich, Dubna 1 973.

(2) 7 . . . �xd4 8 Y!fxd4 j_d7 9 0-0-0Y!fa5 1 0 J..d2 a6 ( 1 0 . . . Y!fc5 1 1 Y!fxc5 dxc5 12 �b5 is good for

Classical Variation 41

White) 1 1 J..e2! J..c6 ( 1 1 Y!fc5 is still bad, this time because after 1 2 Y!fxc5 dxc5 1 3 e5 �d5 14 �xd5 exd5 15 J..f3 j_c6 1 6 j_a5! White prevents . . . .1. d8) 1 2 f3 Y!fc7 1 3 g4 j_e7 1 4 g5 <E}h5 1 5 Y!fg 1 ! (not 1 5 f4 at once because of 1 5 . . . e5) and White's threat of f4 proved very hard to meet, Tal­Sosonko, Wijk aan Zee 1 982.

8 0-0-0 0-0 Or 8 �xd4 9 Y!fxd4 0-0 1 0

e 5 dxe5 1 1 Y!fxe5 ( 36) and now:

36 B

( 1 ) 1 1 . . . �d7 is bad after 1 2 J..xe7 Y!fxe7 1 3 Y!fc7!

(2) 11 . . . Y!fb6 12 j}_e3 �g4 gives White a favourable ending after 1 3 J..xb6 <E}xe5 1 4 j_c7 �g4 1 5 �g3 <2:l f6 1 6 j_b5! a6 1 7 j}_e2 b5 1 8 J.. f3 K a7 1 9 j_d6, as in Vasyukov-Boleslavsky, USSR 1 957.

(3) 11 . . . J..d7 12 h4 � c8 ( 1 2 Y!fe8 transposes t o line 4) 1 3

lil h 3 (White's rook can come to d3 or g3) Y!fc7 ( 1 3 . l;;!!, c5 14 �e3 �c8 1 5 M g3 \t'h8 1 6 \t'b1 fol­lowed by the push of the h-pawn gives White a dangerous attack) 1 4 �xc7 � xc7 1 5 � b5 J..xb5 1 6

42 Classical Variation

j_xb5 and White's two bishops give him an edge in this ending.

(4) 11 . . . iJ}te8 1 2 h4 j_d7 ( 1 2 a6?! 1 3 <2:le4 <2:l d5 1 4 J..xe7

<2:lxe7 1 5 h5 <2:1c6 1 6 iJ}tg3 iJ}te7 1 7 h6 g6, Hubner-Timman, Belfort 1 988, and now 1 8 j_ c4 is very good for White) 1 3 � h3 j_c6 14 � g3 itb8 15 llte3 H d8, Tal­Timman, Brussels SWIFT 1 988, and now 16 J..d3 ! j_d6 1 7 J..xf6 j_xg3 1 8 <2:le2! gxf6 1 9 <2:lxg3 *f8 20 ith6 + *e7 2 1 � e l � d5 22 <2:lf5 + Ji xf5 23 J..xf5 gives White the advantage according to Tal.

The position after 8 0-0 is one of the most important in the whole Sicilian Defence and de­spite decades of practical ex­perience no definite assessment can be given. Although 9 <2:1 b3 was often played in the late fifties and early sixties, it fell into disuse and 9 f4, which has always been regarded as the main line, became virtually universal. However 9 <2:1 b3 has been regaining popular­ity and now rivals 9 f4 for the distinction of being considered the 'main line'.

9 <2:lb3 (37) Notice that 9 j_xf6? is bad since

Black can play 9 jlxf6 1 0 <21xc6 bxc6 1 1 iJ}txd6 itb6 when the threats to b2 and f2 are more than enough compensation for the pawn.

The main point of 9 <2:1 b3 is that it unveils an attack against the d6 pawn and so prepares j_xf6. Black has four main methods of countering White's plan. The first

is to sacrifice the d-pawn, for ex­ample by 9 h6 1 0 jlxf6 j_xf6, but a number of games prove that White can gain an advantage. A second plan is to allow jtxf6 and then recapture with the pawn. This leaves Black's king some­what exposed, but White has no immediate method of launching an attack. More serious is that Black's central pawns are inflex­ible and White may have enough time to start a kingside pawn storm. The assessment of this line depends on the speed of the re­spective attacks. Black's third plan is to counter White's pres­sure on d6 directly by playing 9

a5 I 0 a4 d5 (moreover the immediate 9 d5 is just about possible). The final plan is to counterattack f2 by 9 iJ}tb6, gaining enough time to defend d6 by � d8. Apart from the first plan all these lines are playable.

9 a5 Apart from the two major alter­

natives of 9 a6 and 9 it b6, there are a number of less com­mon ideas:

( 1 ) 9 . . . dS 1 0 J..xf6 J_xf6 1 1 exd5 jlxc3 ( 1 1 � b4 1 2 a3 /2l xd5 1 3 �xd5 exd5 14 �xd5 is slightly better for White since although Black has the two bis­hops it is not easy for him to avoid the exchange of queens) 1 2 �xc3 exd5 and now:

( l a) 13 g3 j_g4 14 � d2 �e7 1 5 Ag2 lr;! fe8 1 6 'iti' b l � ac8 1 7 �c5! 121 b4? ( 17 J.. f5 is better) 1 8 �xe7 � xe7 1 9 � c l with the superior ending for White, Psak­his-Aseev, Sevastopol 1 986.

(1 b) 13 �d4 should give White a small but safe advantage.

( l e) 13 jibS!? �g5 + and now both 14 �d2 ytxg2 1 5 !i[ hg l and 1 4 f} b l d4 1 5 'i!lfc5 ytxg2 give White an attack in return for the pawn.

(2) 9 . . . � aS 10 fJ b l � xb3 1 1 cxb3 a6 1 2 f4 b5 1 3 Axf6 gxf6 14 j_d3 \9ih8 15 f5 b4 1 6 <E:� e2 e5 1 7 jlc4, Anand-Mateo, Dubai 01. 1 986, is worth mentioning because it is a perfect example of what Black should avoid. His king position has been weakened without any compensating queen­side attack and Black has played

e5 at a moment when White can reply jlc4 to gain control of d5.

(3) 9 . . . h6 (it now seems estab­lished that this line is good for White) 10 Axf6 J..xf6 1 1 �xd6 jlxc3 ( 1 1 �b6 12 �c5 �c7 1 3 g 3 JJ...e7 14 yte3 a6 1 5 f4 b 5 1 6 lJ...g2 �e5 1 7 yte2 was also good for White in Marjanovic-Barlov, Yugoslav Ch. 1 985) 1 2 bxc3 �h4

Classical Variation 43

1 3 g3 �f6 (not 1 3 -�xe4? 14 jld3 and j_h7 +) 14 ·�c5 e5 1 5 JJ...c4 jle6 (or 1 5 J_g4 1 6 hl[ d6 �g5 + 1 7 \fib2 J..f3 1 8 � e l � ac8 1 9 h4 �g4 20 �e3 � fd8 2 1 � xd8 + .l! xd8 22 _!d5 JJ...g2 23 !! g 1 �f3 24 g4 with advantage to White, Lobron-Kunsztowicz, Bad N euenahr 1 984) 1 6 JL xe6 �xe6 1 7 � d6 �h3 1 8 '/l!(e3 � fd8 19 � d5 and now both 19 . . . .I deS 20 � hd 1 � c7 2 1 f4 exf4 22 '!Wxf4 � e7 23 <E:�c5, Chandler­Torre, London 1 984 and 19 . . . �g2 20 � hd l �xh2 2i �c5, Klovan-Tal, Jurmala 1 983, were good for White.

We now move on the major lines:

(4) 9 . . . a6 10 J..xf6 and now: (4a) 10 . . . JJ...xf6 (this is

dubious) 1 1 �xd6 Ji_xc3 (or 1 1 ·�b6 1 2 f4 J..e7 1 3 'i'l!{d2 a5 14

a4 � d8 1 5 i/..d3 with advantage to White, Shaposhnikov-Boles­lavsky, USSR 1950; as usual 1 2

j_xc3 1 3 bxc3 �e3 + 14 ffb2 �xe4 loses to 1 5 Ji..d3 and 1 6 Ji..xh7 + ) 1 2 bxc3 (the position is the same as after 9 h6, except that Black has played a6 in­stead; the verdict is unchanged) � f6 ( 1 2 �h4 1 3 g3 �[6 14 �c5 e5 15 JJ...c4 �.g4 1 6 g d6 �g5 + 1 7 f4 and White was clearly better in Ivanovic-Popo­vic, Novi Sad 1 984) 1 3 i'l!i'g3 e5 14 �c4 j_ e6 15 jlxe6 �xe6 1 6 g d5 and again White had won the opening battle, Benjamin-Chris­tiansen, USA 1 984.

(4b) 10 . . . gxf6 1 1 �h6

44 Classical Variation

(although this line had not been played very often, in my opinion it offers White the best chances; the plan is a general kingside pawn advance by g4, f4, h4 and g5) * h8 1 2 � h5 (Black is now virtually forced to lose time with his queen because he must free the f8 rook to defend h7 by ii;! g8-g7) �e8 ( 1 2 � g8?, Ernst­Chandler, London 1 988, and now l 3 'l!lfxf7! '!i!, g6 1 4 f4 j_d7 1 5 � d3 e5 1 6 ii;! g3 ! j_e8 1 7 'i'�!!'d5 exf4 1 8 Jilt xg6 hxg6 1 9 'iii!'d2 g 5 2 0 <2:ld4 �xd4 2 1 �xd4 is good for White) 1 3 f4 b5 (in view of the note to White's 1 5th move it might be more accurate to play 1 3 . . . ii g8) 14 jj_d3 li( g8 1 5 g4 (Tisdall points out the possibility of 1 5 �d5, based on the tactical point 1 5 exd5 1 6 exd5 '!i!, g7 1 7 dxc6 J..g4 1 8 �d5 .Rxd l 1 9 ,ii;i xd l with fan­tastic compensation for White's slight material deficit; of course Black should play 1 5 J..b7 1 6 <2:lxe7 �xe7 and i t i s far from certain that the exchange of minor pieces favours White, especially as he has spent two tempi achieving it) '14. g7 ( 1 5 b4 1 6 e5 clears e4 for the knight with gain of tempo) 1 6 h4 b4 1 7 <2:le2 a5 1 8 g5 a4 1 9 <2:l bd4 and now:

(4b l ) 19 . . . <2:lxd4 20 <2:lxd4 j_d7 2 1 gxf6 ji_xf6 22 e5 dxe5 23 � hg l �g8, Arnason-Inkiov, Plovdiv 1 986, and now Inkiov gives 24 <2:lf3! ji_e8 (24 . . . exf4 25 � xg7 J..xg7 26 <2:lg5 wins) 25 j_ e4 (25 fxe5 looks even better to me) as good for White.

(4b2) 19 . . . b3!? 20 axb3 (20 cxb3 axb3 2 1 a3 is also possible, when 2 1 <2:l xd4 22 <2:lxd4 'lta4 is unclear) axb3 21 <2:lxb3 ji_b7 22 <E)c3 � b4 23 � hg l 'ltc6 24 � g3 � b6 25 'lte2 d5 was played in Psakhis-Kotronias, Dortmund 1 989, and now the best move is 26 'lte3!, when Kotronias gives 26

<2:lxd3 + 27 H xd3 'i'l!l'xe3 + 28 � dxe3 ji_ d6 29 e5 fxe5 30 fxe5 J.. e7 intending h6 as unclear. However after 3 1 <2:l b5! I doubt if Black has enough for the pawn, e.g. 3 1 h6 32 gxh6 � xg3 33 '14. xg3 ji_xh4 34 K g7 intending �d6.

(5) 9 . . . �b6 (the counterat­tack against f2 nullifies White's threat to take on f6, so Black gets time to support his d-pawn by H d8; White's usual reaction has been to start a kingside pawn storm, but he must be careful because too many pawn moves might encourage Black to open up the centre by d5) 10 f3 (38) and now:

(5a) 10 . . . a6 1 1 h4! (this is the move which has put 1 0 a6 out of favour) .f;l d8 1 2 h5 'ltc7 (when

the pawn reaches hS Black faces a tricky problem-is he willing to let his black squares be weakened by allowing h6, or should be play

h6 himself, even though this makes the subsequent advances of White's g-pawn very strong? After 1 2 h6 1 3 J.. e3 'fjc7 14 �f2 �d7 1S g4 �ceS 1 6 � gl bS 1 7 gS White had a very dangerous attack in Hellers--J. Piket, Amsterdam 11 1 98S) 1 3 g4 (or 1 3 h6 g6 1 4 �f4 �e8 l S J.. xe7 -wtxe7 1 6 -wte3 bS 1 7 11 e2 � b8, Marti­novic-Popovic, Yugoslav Ch. 1 986, and now 18 � d2 followed by � hd l is slightly better for White according to Martinovic) bS 14 j_e3 �d7 1 S gS � ceS 1 6 g6! b4 ( 1 6 fxg6 1 7 f4! �c4 1 8 J..xc4 bxc4 1 9 Z!dS! exdS 20 �xdS + '\fih8 2 1 hxg6 � f6 22 � xh7 + �xh7 23 Ji h l wins) 1 7 gxf7 + 'l¥ xf7 1 8 �dS! exdS 1 9 '*'xdS + '\fif8 ( 1 9 'itre8 20 h6 g6 2 1 f4 �g4 22 j_c4 is also good for White) 20 'l!lfxa8! R_b7 2 1 'l!lfa7 with advantage to White since 2 1

�c6 i s met b y 22 �d4, Serper-Brodsky, USSR 1 986.

(Sb) 10 . . . Zii, d8 1 1 * b l (a useful semi-waiting move; the re­ply 1 1 dS is bad for tactical reasons, so Black normally plays 1 1 a6, when White can switch to his kingside attack plan with-out allowing dS) and now:

(Sb1 ) ll . . . d5?! 12 �xf6 j_xf6 (Black should not play 1 2 . . . dxe4? because of 1 3 j_xe7! Zil xd2 14 �xd2! when 14 �xe7 1 5 � c4 Y¥tc7 1 6 � b5 wins the queen,

Classical Variation 45

so Anand-Benjamin, Wijk aan Zee 1 989 continued 14 exf3 and now 15 �c4 fxg2 1 6 J.. xg2 ftfc7 1 7 §td6 �d8 1 8 jtg3 -wte7 19 � he I is crushing for White) 1 3 exdS j_xc3 ( 1 3 aS 1 4 � a4 "*a7 1 5 d6 b6 1 6 �e3! lil!. b8 1 7 .Z1 b 5 was very good for White in Mokry-Conquest, Gausdal 1 989) 14 �xc3 exd5 ( 1 4 . . . � b4 is worse as 15 d6! � xd6 16 Ac4 makes it hard for Black to com­plete his development) 1 5 'i!!fcS followed by j_b5, with advantage to White.

(Sb2) ll . . . a6 1 2 j_e3 'f#c7 1 3 'i!i!i'f2 (thanks to the threat of j_b6 Black has no time for . . . dS) �d7 14 h4 bS with a position similar to line Sa above. Admittedly White has spent a move on '\fi b ! , but this is certainly not a waste of tempo, and while White's chances are not quite as good as in line Sa he has fair attacking chances. In Sax­Wilder, Lugano 1 989, White adopted the rather strange plan l S h5 J.. b7 16 �g3 J.. f8 1 7 J..gS � e8 and now Black is at least equal, but the simple l S g4 is better.

10 a4 This appears to be the best res­

ponse to the advance of the a­pawn.

10 d5 l l jlb5 ( 39)

This move was introduced by Tal in his game against Sisniega in the 1 98S Taxco Jnterzonal, although Vitolinsh was appar­ently the originator.

46 Classical Variation

39 B

Or: 1 l Q)b4

( 1 ) 1 1 . . . Q)xe4 1 2 Q)xe4 dxe4 1 3 �xd8 ..§..xd8 1 4 _§_xd8 Q)xd8 1 5 Q) c5 f5 (or 1 5 b6 16 Q) xe4 Jtb7 1 7 � he l jj_d5 18 0 � c8 1 9 Q)c3 jj_a8 20 jj_ d 7 � c7 21 �b5 � c5 22 � d6 !2Jc6 23 Q)c3 with a slight plus for White, Rohde­Joshi, USA 1 986) 16 l:!\ d6 '!fl>f7 1 7 � hd l f!/e7 1 8 i.:..d7 (Tal-Sis­niega, Taxco 1 985) and now 18 . . . jj_xd7 was just slightly better for White according to Tal. In the game 18 . . . � ti 1 9 Q)xe6 Jtxd7 20 JZJc7 .R.xa4 21 Ji)xa8 led to a quick White win .

(2) 1 l . . . dxe4 1 2 'ltxd8 Jtxd8 1 3 lii( he 1 Q)a7 (or 13 h6 14 J.. xf6 Axf6 1 5 Q) xe4 with an edge for White) 14 _§_c4 h6 1 5 Jtxf6 gxf6 1 6 Q)xe4 f5 1 7 Q)d6 Jtc7 18 g3 b6? ( 1 8 H d8 1 9 Q)b5 Q)xb5 20 Axb5 would have been slightly better for White according to Tal) 1 9 Q)xf5! and White won in the famous game Tal-Korchnoi, Montpellier 1 985 .

(3) 11 Q) a7 and now there are two tempting lines for White:

(3a) 12 Jte2 Ad7 ( 12 . . . b5 1 3

exd5 bxa4 14 d 6 axb3 1 5 dxe7 'Wfxe7 1 6 cxb3 was unclear in Oll­Ryskin, USSR 1 987, but 1 3 J.. xf6 J.. xf6 14 <EJxb5 is better according to Oil) 1 3 J.. xf6 j}_xf6 14 exd5 _txc3 1 5 'f!txc3 _txa4 1 6 dxe6 'f!te7 1 7 exf7 + * h8 (Black has some initiative for the two pawns, but not nearly enough) 1 8 _tc4 � ac8 19 � he l 'f!tg5 + 20 'f!td2 'f!tg6 21 _te6 � c7 22 'f!td3 'i!lt'xg2 23 'f!td6 and Black's counterat­tack has collapsed, Gelfand-Rys­kin, Minsk 1 986.

(3b) 12 Jtxf6 Jtxf6 13 exd5 _txc3 14 11f'xc3 Q)xb5 1 5 axb5 a4?! ( 1 5 exd5 16 Q)d4 ttb6 is rela­tively best, but still good for White after 1 7 H he l ) 1 6 dxe6! 'l!l!fg5 + (the point is that 1 6 . . axb3 1 7 � xd8 � a 1 + 1 8 'i91d2 )j xd8 + 1 9 \tie2 !! xh 1 fails to 20 llrc7 )j f8 2 1 e7 � e8 22 'f!td8) 1 7 'i!it'd2 'f!tf6 1 8 Q)d4 and Black has very little for his minus pawn, Hoffman-Timoshchenko, Buda­pest Open 1 989.

(4) 1 1 . . . J.. b4 (an untested suggestion by Tal) 1 2 exd5 exd5 1 3 'f!tf4 with an edge for White.

12 � bel (40) Or 12 e5 Q)d7 1 3 J.. xe7 ttxe7

14 f4 (in this French Defence type of position, the exchange of black-squared bishops theoreti­cally favours White, but with the kings castled on opposite sides of the board the game is more likely to be decided by the speed of the respective attacks rather than by the endgame advantage of the better bishop) Q) c5 (after 14 . . .

b6 1 5 � he l � c5 1 6 �d4 A d7 1 7 * b l � ac8? 1 8 g4 � fd8 1 9 f5 �e4 20 �xe4 dxe4 2 1 c3 �d3 22 A xd3 exd3 23 * f4! A xa4 24 � xd3 White had an excellent position in Balashov-Khalifman, Minsk 1 986, but 1 7 � fc8 intending . Axb5 was better) 1 5 <E'Jxc5 *xc5 1 6 h4 (the idea is to bring the rook to g3, not only helping the kingside attack, but also providing useful defence along the third rank) b6 and now 1 7 h5 was quite unclear in Kinder­mann-Felsberger, Vienna 1 986. Perhaps 1 7 � h3 was more accur­ate because in some lines the h­pawn plays no important role on h5, but in any case the position is very double-edged.

40 B

12 dxe4 Or: (I) 12 . . . h6 and now: ( l a) 13 Axf6 A xf6 14 exd5

exd5 1 5 <E'Jxd5 j'tg4 1 6 f3 ltg5 1 7 <E'Je7 + ! ( 1 7 <E'Je3 Jt f5 1 8 <E'Jd4 j'tg6 19 g3 j'txe3 20 �xe3 � c8 gave Black enough for the pawn in 011-Temirbaev, Kuibyshev 1 986) f/!xe7 ( 1 7 . . . j'txe7 1 8

Classical Variation 47

)ll!Jxd8 � fxd8 19 � xd8 + iii( xd8 20 il, xe7 jie6 2 1 <E'Jd2 gives Black nothing for the pawn) 1 8 iii( xe7 Jtxe7 19 i'ite2 j'tg5 + 20 fl>b1 A f5 2 1 <E'Jd4 jtg6 22 '\t1al ! with excellent winning chances for White, Oll-Kha1ifman, USSR 1987.

(I b) 13 exd5!? exd5 ( 1 3 hxg5 14 d6) 14 '!ioi'e3 Ae6 1 5 JJ..f4 iJ, c8 (Winsnes-Khalifman, Groningen 1 985/6) and now 1 6 * b l (intend­ing -2]d4) -2]h5 1 7 . .iLe5 -2] c6 1 8 Jtxc6 bxc6 19 A.d4 i s good for White according to Donaldson.

( l e) 13 Ji.e3 �c7 14 1i. f4 e5 1 5 exd5 exf4 1 6 d6 gives White an edge, Andrijevic-Kapetanovic, Yugoslavia 1 988.

(2) 12 . . . Jtd7 (perhaps the most solid move) 1 3 e5 ( 1 3 exd5 .iLxb5 14 d6 'ilfxd6 1 5 �xd6 jJ_xd6 1 6 £ xd6 j'tc6 gives White an edge, as does 13 Ji.xd7 'li\'xd7 14 e5 �e8 15 j'txe7 ljjxe7 16 f4 tz:Jc7 1 7 -2Jd4 .:t:Jc6 1 8 �db5 .:t:J xb5 19 -2] xb5 � ac8 even though a draw was agreed here in Rachels-0. Gurevich, Boston 1 988) �e8 14 h4 <E'Jc7 1 5 -2]d4 -2]c6 1 6 . .iLxe7 �xe7 1 7 �g5 and again White has a small advantage, Wang Zili­D. Gurevich, Belgrade, 1988.

13 )litxd8 White has a speculative alterna­

tive in 1 3 � xe4 -2]xe4 14 �xd8 Ji.xg5 + 1 5 "Wfxg5 <E'Jxg5 1 6 h4 e5 ( 1 6 . . . h6 1 7 hxg5 hxg5 1 8 £ d6 offers White reasonable play for the pawn as Black's development is very difficult) 1 7 hxg5 1i.f5 1 8 � d 2 !! fc8 ( 1 8 . . f6 was more

48 Classical Variation

cautious) 1 9 c3 Ae6 20 <E) a l B, c5 2 1 J..d7! with some advantage to White, Vitolinsh-Inkiov, Junnala 1 985.

13 � xd8 1 3 . . . A xd8 1 4 <E)xe4 is good

for White after 1 4 . . . <E)xe4 1 5 A xd8 <E) xf2 1 6 B, d2 or 14 . . . Ae7 1 5 <2\xf6 + J..xf6 1 6 j}_xf6 gxf6 1 7 )!lt d6.

14 <E)xe4 <E)bd5 15 c4! <2\c7

Or 1 5 <E) b4 1 6 B. xd8 + �xd8 1 7 l! d 1 Ae7 1 8 <E)d6 with a clear plus.

41 B

16 ll xd8 + 17 M d1 18 <E)xf6 + 19 Ae3 20 axb5 (41)

� xd8 jte7 gxf6 <E)xb5

The liquidation has left White with a clear advantage. Black still has problems developing his pieces and White's queenside majority is ominously near to creating a passed pawn. True, Black has the two bishops, but White can always force an exchange (e.g. by �c5) if they show signs of becoming active.

20 f5 Black would like to play 20 . . .

e5 to free his white-squared bishop, but after 2 1 �c5 A xc5 22 <2\ xc5 j}_g4 23 l! d5 l, c8 24 b3 Black's queenside pawns are in big trouble (24 . . . K c7 25 b6 K e7 26 <E)e4). Black therefore prepares to duck the exchange of bishops after �c5 by �g5 + .

21 <E) c5 e5 In lnformator Anand gives 2 1

f4 22 jtd4 f6 (22 e5? 23 jtxe5) as only slightly better for White. I find this assessment unduly modest, since after 23 <E)a4 e5 24 <E) b6 g b8 25 <E) d5! J..d8 (25 *f7 26 <E)xe7 ff/xe7 27 jtc5 + *e8 28 B, d6 is also unpleasant) 26 � a7 l;li a8 27 jtb6 Black is in serious trouble.

22 <E)d7 f4 23 �b6 f6

Practically the only legal move! Black intends *f7-e6.

24 �c7 The threat of <21 b6 forces liqui­

dation into a winning rook and pawn ending.

24 ll._xd7 25 � xd7 ll._c5 26 11._ d6 11._ xd6

After 26 �xf2 27 lii xb7 White's two connected passed pawns roll forwards.

27 l;1 xd6 � c8 Black goes for counterplay, but

he cannot repair the fundamental defects of his position. After 27

� f8 28 *c2 e4 29 M e6 f5 30 *c3 White's active king will de­cide the issue.

28 b3 a4 29 rffb2 axb3 30 \txb3 \tf7

Black offers a pawn to get his own majority moving, but the two connected passed pawns outweigh anything Black can do with his e­pawn.

31 H d7 + \te6 32 H xb7 e4 33 H a7 e3 34 fxe3 fxe3 35 H a2 H d8

The immediate 35 \te5 is also met by 36 fgc3, and if 36 fge4 then 37 b6.

36 fgc3 *e5 Black could have quite reason-

ably resigned instead. 37 b6 *e4 38 b7 H d3 + 39 rffb4 .i, d2 40 b8 (�) .l, xa2 41 �b7 + *d3 42 �d5 + *e2 43 c5 Resigns

Game 1 1 Ernst-Popovic Subotica 1987

1 e4 c5 2 <2l f3 <2l c6 3 d4 cxd4 4 <2l xd4 <2l f6 5 �c3 d6 6 jj_g5 e6 7 'i(1d2 a6 8 0-0-0 ( 42) Jl.d7

8 . . . h6 is examined in game 1 2. Other ideas are dubious, e.g. 8 . . .

Classical Variation 49

!fJ..e7 (8 �b6 9 <2lb3 !fJ_d7 l O !fJ_e2 'W/c7 l l f4 h6 1 2 !fJ_xf6 gxf6 1 3 !fJ.. h5 is good for White, Pan­chenko--Csom, Las Palmas 1 978) 9 f4 file? (9 �xd4 10 '*'xd4 'lit a5 l l e5 dxe5 12 fxe5 <2l d5 1 3 JJ..xe7 �xe7 l 4 JJ..d3 .:t:Jc6 1 5 'W/h4 �xe5 16 <2le4 f6 17 H hfl with a very dangerous attack for White, Adler-Bannik, USSR 1 978) and now:

( I ) 10 fgb1 JJ..d7 ( 1 0 0-0 l l !fJ_e2 H d8 1 2 jj_f3 h6 l 3 h4 <2lxd4 14 �xd4 b5 1 5 iYf2 !fJ..b7 1 6 g4 was good for White in Kavalek­Larsen, Montreal 1 979) 1 1 <2l f3 H d8 1 2 JJ..d3 b5 1 3 � he 1 b4 14 <2le2 a5 15 <2l g3 0-0 1 6 e5 <2ld5 17 \\lfe2 was unclear in Przewoznik­Bielczyk, Katowice 1 986.

(2) 10 JJ..e2 <2lxd4 1 1 �xd4 b5 12 e5 dxe5 1 3 fxe5 � d5 14 .JJ.. xe7 <2lxc3 1 5 J.. f3 ! <2l xd 1 1 6 J..d6! and White went on to win quickly in Tal-Larsen, Montreal 1 979 after 1 6 �c4 1 7 iYb6! <2l f2 1 8 jj_c6 + J..d7 1 9 .JJ.. xd7 + *xd7 20 �b7 + *d8 2 1 '/!i!1"xa8 + �c8 22 �a7 Resigns.

(3) 10 1L xf6 gxf6 l l g3 .il_d7 12 f5 � xd4 13 � xd4 £ c8?! ( 1 3 . . .

50 Classical Variation

0-0 and 1 3 b5 are possible improvements) 14 * b l b5 1 5 �d2 �c5 1 6 j_d3 h5? 1 7 fxe(l fxe6 1 8 <i:Je2 intending <i:l f4, Short-Larsen, London 1 986, and White has the advantage.

9 f4 b5 Black has two major alterna-

tives, 9 h6 and 9 !P..e7: ( l ) 9 . . . h6 (9 Jl c8 10 <i:l f3

�a5 1 1 * b l b 5 1 2 e 5 b4 1 3 exf6 bxc3 14 fxg7 !P..xg7 1 5 �xd6 '!!A c7 16 <i:Je5 A xeS 1 7 fxe5 1:! g8 18 h4 is good for White, Yanofsky­Olafsson, Dallas 1957) 1 0 !P..h4 ( 43) and now:

43 B

( l a) 10 • • • !P..e7 (10 . . . <i:Jxd4 1 1 �xd4 !P..c6 1 2 !P..c4 is clearly good for White, while 10 . . . K c8 1 1 <i:lf3 �a5 is similar to the note to Black's 9th move) 1 1 .:E�f3! b5 1 2 e5 ( 1 2 !P..xf6 !P..xf6 1 3 �xd6 )i,i( a7 14 e5 !P..e7 1 5 �d3! -a5 1 6 * b 1 I:t c7 1 7 �e3! 0-0 1 8 �e4 was also good for White in Geor­gadze-Makarychev, Nikolaev 1 983) b4 1 3 exf6 bxc3 14 �xc3 gxf6 1 5 f5 and White stands well, Thiemann-Reynolds, corr. 1966.

( l b) 10 . • . <i:Jxe4 1 1 �e1 .:E�f6

( l l g5 1 2 <i:Jxe4 gxh4 1 3 "Wt"c3) 1 2 <E:� f5 "Wt"a5 (12 • . • "Wt"b8 1 3 !P..xf6 gxf6 14 <i:Je4, 12 • • . !P..e7 1 3 <E) xd6 + \fi'f8 1 4 <i:Jxb7 "Wt"c7 1 5 �d2 and 1 2 . • . "Wt"c7 1 3 !P..xf6 gxf6 14 <i:Jd5 "Wt"d8 1 5 "Wt"e3 ! are all very pleasant for White) 1 3 <E) xd6 + J.xd6 14 .M xd6 0-0-0 (14

. • • �c7 1 5 lilt d2 is unsatisfactory after 15 . . • 0-0-0 1 6 "Wt"f2 <E:�e7 1 7 .!id3 !P..c6 1 8 f5 e 5 1 9 � hd l or 15

• . • "Wfxf4 1 6 J_e2 <i:Je4 1 7 <i:Jxe4 "Wfxe4 1 8 "Wtf2, Gligoric-Barden, Bognor Regis 1 957, while 14 . . . <i:Je7 1 5 K d 1 .:E�g6 1 6 <i:Je4! "Wtxe1 17 <i:Jd6 + 'ltle7 18 J.xe l <i:Jd5 1 9 <i:Jx7 <i:Jxf4 20 g3 <i:Jg6 21 !P..g2 and 14 . • . <i:Jb4 1 5 a3 <i:J bd5 1 6 �e5 !P..c6 17 !P..c4! are both very good good for White) 1 5 K d l ! �c7 (15 . . . <i:Je7? 16 <i:Jd5 wins, 15 . . . g5 1 6 fxg5 hxg5 1 7 !P..g3 gives White very strong pressure on the dark squares and 15 • • . e5 1 6 fxe5 K he8 1 7 jj_g3 <i:J xe5 18 <i:J b5 ll_g4 1 9 K xd8 + K xd8 20 <i:Ja7 + * b8 2 1 <i:Jc6 + bxc6 22 �xe5 + is a very good ending for White) 1 6 "Wt f2 <i:Je7 1 7 jj_ d3 jj_c6 1 8 f5 e5 1 9 M he l <i:Jed5 20 <i:Jxd5 K xd5 (20

_ztxd5 2 1 "Wta7) 2 1 �g3 e4 22 "Wfxc7 + \fi'xc7 23 ll_xf6 exd3 24 !P..xg7 K hd8 25 .zte5 + with good winning chances for White, Spassky-Rabar, Goteborg 1 955.

( l e) 10 . • . g5 1 1 fxg5 <i:Jg4 and now:

( l c l ) 12 <i:lf3 hxg5 ( 1 2 jj_e7 1 3 g6! J. xh4 1 4 gxf7 + \fi'xfi 1 5 "Wtf4 + is good for White, or 14 . . . *f8 1 5 "Wtxd6 + jj_ e7 1 6 t�txd7 <2:Jf2 1 7 "Wtxb7 <i:J xd 1 1 8 "Wfxc6 and

White should win) 1 3 jj_g3 jle7 14 jle2 �ge5 and now:

( l c l l ) 15 'fl'bl b5 ( 1 5 f6 1 6 h4! gxh4 1 7 �xh4 �a5 1 8 � f5! � xh l 19 � xh l exf5 20 .M h8 + jlf8 2 1 jlxe5! �xe5 22 �xd6 is good for White, Mokry-Banas, Trnava 1 986) 1 6 a3 M b8 1 7 �a2 a5 1 8 �cl �xf3 1 9 gxf3 e5? ( 1 9 . �e5! is unclear) 20 .il..f2 jj_e6 2 1 jj_e3 with an edge for White, Riemersma-A. Rodriguez, Dieren 1 987.

( l c l 2) 15 � hfl f6 ( 1 5 '*'a5 16 'fl' b l f6 17 �e l b5 18 jj_d3 0-0-0 19 �d2 \fJb7 20 h3 * a8 2 1 jl f2 was a little better for White, Vogt-Barczay, Zalakaros 1 987) 1 6 � xe5 �xe5 1 7 jlxe5 dxe5 18 � f3 �c7 1 9 � d3 .M d8 20 h3 jlc8 2 1 l! xd8 + �xd8 22 'i!ire3 ·ii!i'a5 (22 'ilfc7 preventing jj_c4 ap­pears better) 23 jj_c4 �c5 24 'il!fe2 jj_d8, Vogt-Wirius, Zalakaros 1 987, and now 25 \fJ b l gives White an edge.

( l c2) 12 � xc6 (this seems the better choice) jlxc6 1 3 jj_e2 � e5 1 4 g3 �g6 1 5 \fi b ! h5 ( 1 5 jj_e7?! 1 6 gxh6 jj_xh4 1 7 gxh4 �xh4 18 �xd6 was good for White in Marjanovic-Popovic, Belgrade 1 987) with a further branch:

( l c2 1 ) 16 �e3 jle7 (better than 16 . . . jj_g7 1 7 �d5! exd5 1 8 exd5 + '*e7 1 9 �xe7 + *xe7 20 dxc6 bxc6 21 c3 with an edge for White, Chandler-Bellin, Com­monwealth Ch. 1 985, or 16 . . . '*e7 1 7 � hfl jj_g7 1 8 � f2 *f8 1 9 � dfl , Mainka-Popovic, Dort-

Classical Variation 51

mund 1 988 and White is better) 1 7 � hfl �c7 1 8 �f2 �e5 1 9 h3 0-0-0 20 g4 hxg4 21 jj_xg4 �xg4 22 hxg4 � h7 is unclear, Jansa­Banas, CSSR Ch. 1986.

( l c22) 16 .M hfl ( 1 6 lil, dfl !? intending 17 �d I attacking h5 is an interesting idea) �c7 ( 1 6 jj_g7!? is better) 1 7 e5! 0-0-0 1 8 J.,d3 � xh4 1 9 gxh4 J..g7 20 *f2! and White stands well, Tsesh­kovsky-Fahnenschmidt, Baden­Baden 1 988.

( I d) 10 . . . b5 1 1 jj_xf6 �xf6!? 12 e5 ( 12 j}_xb5 axb5 1 3 �dxb5 �d8 14 �xd6 + jlxd6 1 5 trxd6 �e7 16 e5 �xd6 1 7 ii:i!. xd6 *e7 is unconvincing) dxe5 1 3 �dxb5 �d8 14 �d6 + jlxd6 1 5 �xd6 exf4 16 �e4! ft(e7 1 7 ft(c7 ii:il. a7 1 8 .£Jd6 + \fi f8 1 9 �b6 and White IS

at least slightly better, Sobura­Berebora, Poland 1 988.

(2) 9 . . . jle7 1 0 �f3 b5 (other moves are inconsistent, e.g. I 0 tllfc7 1 1 e5 dxe5 1 2 fxe5 � d5 1 3 � xd5 exd5 14 jlxe7 � xe7 1 5 jj_d3 0-0 1 6 �g5 �c6 1 7 trh5 or 10 h6 1 1 jj_xf6 gxf6 1 2 f5 trc7 1 3 \fi b ) 0-0-0 14 jlc4 with a clear plus for White in both cases) 1 1 e5 b4 (not 1 1 dxe5 1 2 fxe5 b4 1 3 exf6 bxc3 1 4 ft(xd7 + and White wins) 12 exf6 bxc3 1 3 Yllfxc3 gxf6 14 jj_h4 (44) and now:

(2a) 14 . . . aS ( 1 4 tllfa5 1 5 jlxf6 � b4 1 6 jj_c4 H c8 1 7 a3 is good for White) 1 5 \fi b ! �b4 ( 1 5

!! b 8 1 6 g4 � b4 I 7 a 3 .M c8 1 8 tll!b3 � d5 1 9 K xd5 with an excel­lent position for White, Gligoric­Conrady, Dublin 1 957) 1 6 a3 K c8

52 Classical Variation

44 B

1 7 '\tb3 �d5 1 8 l;l xd5 exd5 and now either 1 9 � d4 or 1 9 '\txd5, with very good compensation for the exchange.

(2b) 14 . . . d5 1 5 * b l with a further branch:

(2b I ) 15 . . . a5 1 6 J.. b5 M c8 1 7 � d4 �xd4 1 8 J_xd7 + '\txd7 1 9 '\txd4 K g8 20 g 3 .- b s 2 1 M be l is good for White.

(2b2) 15 . . . �b4 1 6 �d4 'll!!fa5 ( 1 6 . . . K c8 1 7 'itb3 '\ta5 1 8 J..e l ! J..a4 1 9 '\ta3 with a clear plus for White, Matanovic-Jansa, Lugano 1 968) 1 7 a3 �c6 1 8 '\tg3 �xd4 1 9 �t xd4 K b8 20 K d3 *f8 2 1 J..e2 with a small plus for White according to Lukin.

(2b3) 15 . . . � a5 1 6 f5 l, c8 1 7 'itd2 '\tc7 1 8 fxe6 fxe6 1 9 j_d3! �c4 20 J..xc4 '\txc4 21 K he l K g8 (2 1 0-0 is most simply met by 22 g4!) 22 h3 K b8 23 *a I h5 24 g3 with a small advantage for White, Tseshkovshy-Lukin, USSR 1 982.

10 J_xf6 gxf6 l l � xc6 J..xc6

White's chances lie in the fact that Black's king has no really safe spot, and in an attack against

the weak e6 square by f5, j_d3 and �e2-f4. Black must be care­ful about playing . . . b4; he may gain time by driving the knight away, but he may also allow the bishop to become active at c4. In general the exchange of queens favours Black, so White must make sure he avoids this.

12 'ite1 (45) Probably the most accurate

move. Black's immediate am­bition is to bring his queen to an active square, so White often plays 1 2 *e3 to prevent . . . '\tb6, but after 'ite7-a7 White has to waste time avoiding the queen swap. 12 lill'e I also prevents 1 2 . . '!'lfb6 because of 1 3 �d5, but it also tucks the queen away from the unwelcome attentions of her opposite number.

12 J..e7 Or: ( I ) 12 • . . b4 1 3 �d5 a5, A.

Ivanov-Ermolinsky, USSR 1 98 1 , and now 1 4 M d4! intending .!11 c4 is good for White.

(2) 12 . . . '!'lfa5 1 3 * b l 0-0-0 14 j_d3 \t¥ b8 15 '!'lfh4 '!'lfc7 1 6 f5 was

promising for White in Kuzmin­Tukmakov, USSR eh. 1 977.

(3) 12 . • . jte7 1 3 j_d3 J..g7 ( 1 3 j_b7 14 \t;' b l 0-0-0 1 5 a4 b4 1 6

<2\a2 a 5 1 7 c3 d5 1 8 cxb4 axb4 1 9 exd5 K xd5 20 jte3 jtd7 2 1 li;l c l + \t;'b8 2 2 J_ e4 with some advantage for White, Spraggett­Mednis, Lugano 1 985) 14 jtg3 (I prefer 14 fi' b l intending f5) llii g8 1 5 � he I * ffi 1 6 'Wf h4 f5 l 7 jtxe7 + fi' xe7 1 8 exf5 J..xc3 1 9 bxc3 K xg2 with equality, Kud­rin-Christiansen, USA Ch. 1 986.

13 j_d3 jtb6 14 \t;'b1

Safer than 14 'Wf g3 b4 1 5 <21 e2 \t;'d7 1 6 f5 e5 1 7 jtf3 jtc5, Menc­inger-Damljanovic, Bled 1 984, with an unclear position.

14 b4 The alternative is 14 h5 1 5

f5 b4 (after 1 5 jtc5 White i s a little better after the solid 1 6 K fl !?, but 1 6 �g3 fl'd7 1 7 �h3 K ag8 18 K he1 � g5, Hellers-van der Wiel, Amsterdam 1 986, is also possible, when 1 9 g3 gives White an edge) 1 6 <2\e2 e5 1 7 jtg3 ( 1 7 <2\g3 a5 1 8 �e2 h4 1 9 <2\fl was unclear in Watson-Yrjola, Kecs­kemet 1 988) J..ffi? 1 8 'W/f3! and White stands well, A. lvanov­Anikaev, USSR 1 979. ECO suggests 1 7 . . . d5, but 1 8 'W/f3 looks good in this case too.

15 <2\e2 aS 16 f5!

More active than 1 6 'W/g3, as played hitherto. White threatens 1 7 fxe6 fxe6 1 8 <2'!f4, so Black's reply is virtually forced.

Classical Variation 53

16 e5 17 <2\g3 ( 46)

With the simple plan of J..c4-d5.

46 B

17 *c5 18 ,.e2 a4

There is nothing more Black can do to prevent White's bishop manoeuvre, so his only chance is to aim for queenside counterplay

19 J..c4 0-0 Black needs his other rook if

this queenside play is to genuinely worry White, but there is an obvious danger to his king after

0-0. 20 <2\h5 * h8 21 Jii hfl a3?

Overlooking that White has an immediate mating threat, but even the best line 2 1 K fc8 22 b3 K a7 23 � f3 axb3 24 J..xb3 is very unpleasant for Black.

22 � f3 White can afford to abandon

the queenside since he has a forced win on the other flank.

22 axb2 23 H h3

54 Classical Variation

Intending 24 <E:� xf6 Jlxf6 25 'i4J'h5.

23 � g8 24 � xf6 Resigns

After 24 ilxf6 25 !l xh7 + fj' xh7 26 'i4J'h5 + fl g7 27 'i4fxf7 + flh6 (27 flh8 28 'i4J'h5 + flg7 29 'i4J'g6 + and mate next move) 28 'i4fxf6 + flh7 29 'i4fh4 + flg7 30 'i4J'g5 + the rook is the first of many Black pieces to disappear.

Game 12 Short-Ljubojevic

Amsterdam (Euwe Memorial) 1988

1 e4 c5 2 �f3 d6 3 d4 cxd4 4 �xd4 �f6 5 �c3 <E:�c6 6 ilg5 e6 7 'i4J'd2 a6 8 0-0-0 h6

Black once again aims to further his queenside ambitions by . . . ild7 and b5 before committing his king, but in this line he first of all forces White to decide where to put his bishop.

9 ile3 (47) This move is currently thought

best. White intends to play f4 and develop his king's bishop to e2 or d3. If Black castles kingside then White can launch a direct attack by h3 and g4-g5.

9 ild7 Black has a number of possible

moves, but the basic rule is 'f4 against anything' The alterna­tives are:

(I) 9 • • . 'i4fc7 (9 .:2:J g4 10 .£)xc6 bxc6 1 1 ilc5 is good for White) 10 f4 ,ile7 1 1 Jle2 Jld7 (or 1 1 .:2:Ja5 1 2 e5! dxe5 1 3 fxe5 �xe5 14 _R f4 �c5 1 5 .:2:Ja4 'i4fd5 16 '!ti b 1 ! with a very strong attack for the sacrificed pawn) 1 2 .:2:1 b3 � a5 13 .:2:Jxa5 'i4fxa5 14 flb1 ilc6 1 5 il f3 'i4fc7 16 � he 1 (the im­mediate 1 6 g4 may be even better) lii( c8 1 7 g4 with advantage to White, Chiburdanidze-Lanka, USSR 1 980.

(2) 9 . . . ;Le7 10 f4 .:2:J xd4 1 1 ;Lxd4 b5 1 2 ile2 transposes to line 3 .

(3) 9 . . . .:2:Jxd4 1 0 ;Lxd4 b5 1 1 f4 .ile7 (11 • . . b4 1 2 Axf6 'i4fxf6 1 3 .:2:Je2 � b8 1 4 .:2:ld4 !;K b6 1 5 ilc4, Ta1-Radu1ov, Malta 01. 1 980 and 11 . . . _Rb7 12 Jlxf6 gxf6 1 3 ild3 'i4J'b6 1 4 fJ b 1 ;Le7 1 5 f5 e5 1 6 ;Le2 followed by ilf3, Jansa-Spassov, Sochi 1 980, were both slightly better for White) 1 2 .ile2 (a promising pawn sacrifice) b4 ( 1 2 ilb7 1 3 A f3 b4 1 4 .il xf6 .il xf6 1 5 .:2:1 e2 'i4r a 5 1 6 a 3 is good for White) 13 .:2:Ja4 .:2:Jxe4 1 4 'i4fe3 �f6 1 5 _Rf3 and now:

(3a) 15 . . . d5 1 6 fl b 1 (this has

been played in practice, but in fact 1 6 g4 0-0 1 7 � b6 !![ b8 1 8 * b I may be more accurate, transpos­ing into Short-A. Rodriguez be­low but without allowing Black so much choice) 0-0 ( 1 6 A d7 1 7 � b6 1i b8 1 8 g4 A b5 1 9 h4 * ffl 20 g5 �e8 2 1 f5 was also danger­ous for Black in Chandler-Kos­ten, Hastings 1 988/9) 1 7 � b6 1i b8 1 8 g4 Ad6?! ( 1 8 �d7 1 9 � xd7 jtxd7 20 h4 jtf6 2 1 g5 Axd4 22 Jiil xd4 was better, even though White still has a danger­ous attack) 19 g5 hxg5 20 fxg5 �d7 21 �xc8 li( xc8 22 g6 with a clear plus for White, Short-A. Rodriguez, Subotica 1 987.

(3b) 15 . . . � b8 1 6 Aa7 Ad7! (not 16 . . . � b5 1 7 jt b6 �d7 1 8 � c5 and White wins, Balashov­Tukmakov, Sverdlovsk 1 987) 1 7 � b6 jtb5 1 8 \fi b ! !? (after 1 8 g4 0-0 1 9 Axb8 �xb8 20 g5 hxg5 2 1 fxg5 jtd8! 22 gxf6 �xb6 2 3 '*txb6 jtxb6 Black had sufficient com­pensation in Khalifman-Ionov, USSR 1 988) 0-0 1 9 f5 f/fc7 (the only move) 20 jtxb8 li xb8 2 1 � a8! �d8! (not 2 1 . . . �c8? 22 'fr!!a7 � ffl 23 fxe6 fxe6 24 <Elc7 and the knight excapes, nor 2 1 'fr!/d7 2 2 fxe6 fxe6 2 3 Rhel d 5 24 � b6, followed by taking on e6) 22 'fr!!a7! (after 22 fxe6 ;ii xa8 23 exf7 + \fixf7 24 �xa8 'i#xa8 25 Jii he ! �b7 Black's active minor pieces are at least as valuable as White's rooks) d5! (22 . . . exf5 23 � he ! �e4 24 �xe4 fxe4 25 ;a xe4 J..f6 26 :t::Jc7 and 22 e5 23 � he I! 1L d7 24 �c7 JLxf5 25

Classical Variation 55

� xa6 are good for White) 23 � c7 JLc6 (threatening 24 . . . � b7, and if 24 �xa6 then 24 ;a a8 wins) 24 fxe6 � b7 25 <Elxd5! � xa7 (not 25 �xd5 26 exf7 + fifE 27 �xa6 � c7 28 '*d3 :!i d7 29 �h7 � f6 30 f/fh8 + *xf7 3 1 �xd8 and wins) 26 �xf6 + JLxf6 27 � xd8 + JLxd8 28 exf7 + \fixf7 29 JLxc6 (normally an extra pawn offers some winning chances in a rook and opposite bishop ending, but White must overcome the problem of his inactive king) Af6 30 � d l ;ii c7 3 1 JLf3 JLe7, Nunn­van der Wiel, Lucerne 1 989, and now 32 g d5! � c5 33 1i d3 leads to an ending in which Nigel Short believes White has significant win­ning chances. The plan is a3 or c3, followed by advancing the king to attack Black's remaining queen­side pawn. Certainly Black will be tortured for a long time.

10 f4 b5 Against other moves White

adopts the same general plan of JLd3, * b l , and then a kingside pawn advance, but he has to be careful against 1 0 JLe7, because 1 1 jtd3 allows the awk­ward 1 1 <Elg4! Therefore the best answer to 1 0 jte7 is 1 1 h3 b5 1 2 JLd3, transposing into the note to White's 1 2th move.

11 JLd3 JLe7 ( 48) Two alternatives are 1 1 . . . � c8

1 2 fi b ) � a5?! 1 3 e5! b4! 14 � ce2 dxe5 1 5 fxe5 �d5 1 6 liii, hfl �c4 1 7 J..xc4 � xc4 18 <El f4 � xe3 19 \'¥(xe3 Jl.c5 20 � g6! , Hazai­Szabo, Hungary 1 983, and 1 1 . . .

56 Classical Variation

.:£)xd4 1 2 J..xd4 b4 1 3 �e2 lta5 14 J.. xf6 gxf6 1 5 * b 1 , Psakhis­lvanovic, Sochi 1 979, with a clear plus for White in both cases. Against other moves, such as 1 1

lta5 o r 1 1 . . . ltc7, White proceeds with his plan by 1 2 * b l .

12 *b1 12 * b I has been played a

number of times recently, but there is a strong argument for the immediate 1 2 h3, e.g. 1 2 . . . <2)xd4 1 3 Jl..xd4 J..c6 ( 1 3 . . . b4 1 4 �e2 a5 1 5 g4 j_c6 1 6 <2) g3 d5 1 7 "i!ite2 "tl!l' b8 18 � hfl b3 19 cxb3 a4 20 * b l axb3 21 a3 gives White a modest advantage, Murei­Lobron, Lyons 1 988) 14 "tl!l'e3 ( 1 4 M del 0-0 1 5 * b l .:£) d 7 1 6 g4 was slightly better for White in Timoshchenko-Tukmakov, USSR Ch. 1 978) b4 1 5 <2)e2 fife? 1 6 e5 dxe5 1 7 j_xe5 ftfb7 1 8 f5 <2)d5 1 9 ftfg3 with dangerous threats for White, Timosh­chenko-Sirov, Moscow GMA 1 989.

12 b4 This move has been criticised,

but it isn't clear that the alterna-

tives are better, for example 12 . . . flfc7 1 3 h3 <2) a5 ( 1 3 . . . <2)xd4 14 J..xd4 J..c6 i s probably the most sensible, but even here White has an edge) 1 4 g b4 1 5 <2)ce2 <2) c4 1 6 j_xc4 fifxc4 1 7 <2)g3 a5 1 8 � hfl lta6 1 9 g5 with advantage, Hodg­son-Csom, Tel Aviv 1 988, or 12 . • . <2) xd4 1 3 j_xd4 b4 14 <2)e2 ftfb8 1 5 <2)g3 ( 1 5 h3 is more con­sistent) 0-0?! ( 1 5 e5 is better) 1 6 e5! <2)d5 1 7 <2)e4 � d8 1 8 ftff2 dxe5 1 9 fxe5 J.. e8 20 � hfl and White is better, Hazai-Lobron, Rotterdam 1 988.

13 <2) ce2 0-0 14 h3

Black's problem is that White has an automatic attack by h3, g4, <2)g3 and g5 while Black has to struggle to create any counter­chances at all. It is curious that Black's troubles stem from the apparently innocuous 8 . . . h6, which in this type of position creates a fatal kingside weakness.

14 flfc7 15 g4 ltb7 16 <2)g3 <2) xd4 17 J..xd4 j_c6

All Black has achieved is to create a threat to e4, which delays White's attack by precisely one move.

18 � het lit fe8? This doesn't help the belea­

guered kingside. The last chance was to play 1 8 <2) d7 intending

e5 to block the deadly long diagonal, but in this case Short gives the line 1 9 g5 hxg5 20 lit g 1 (threat <2)h5) e5 2 1 <2) f5 lit fe8 22

fxe5 <2!xe5 23 .,;}_ xe5 dxe5 24 k! xg5! j_xg5 25 �xg5 f6 26 J..c4 + fi'ffi 27 �h5 and mate at h8.

19 g5 hxg5 20 fxg5 <2!d7 ( 49)

After this we are treated to a king-hunt in the style of the 1 9th century. 20 � h7 was objec­tively better (anything is better than being mated), but 2 1 h4 loc­king the knight out of play is very good for White.

49 w

21 j_xg7! *xg7 22 <2!h5 + \tlg6

There is no choice as 22 * g8 23 g6 fxg6 24 "�th6, 22 . . . \tlf8 23 g6 J..f6 24 <21xf6 <2!xf6 25 � f1 \tle7 26 �g5 and 22 . • . \tlh8 23 g6 J.. ffi 24 H g l fxg6 25 � xg6 <2!e5 26 � h6 + ! all lead to disas-ter.

23 e5 + \tlxh5 24 t!'f4 j}_xg5

All Black's moves are forced. 25 "�txf7 + \l;'h4 26 �h7 + fi>·g3

The main problem when play­ing such positions with White is trying to keep a broad grin off

Classical Variation 57

your face. The main problem when playing such positions with Black is to avoid looking at the broad grin on your opponent's face.

27 �h5 \tlh2 Or 27 H g8 28 �g4 + \tlh2

29 'ltgl + flxh3 30 A fl + flh4 (30 . . . j_g2 3 1 K d3 + and 32 Axg2) 3 1 �h2 + \tlg4 32 .i;i d4 + \tlf5 3 3 �h7 + H g6 34 J.h3 mate is a nice line given by Short.

28 'flfxg5 In fact White could have forced

mate by 28 *e2 + \tlxh3 (28 j_g2 29 ;i;i h l + 1trg3 30 �g4 + lff f2 3 1 li:!( hfl + Axfl 32 lli!. xfl + lffe3 33 )!!; e l + 1f!>f2 34 �g l + wf3 35 i1H1 mate) 29 �h5 + J. h4 30 H e3 + l;l, f3 (30 . . . ltrg2 3 1 "�txh4 or 30 . . . J. f3 3 1 H h 1 + 1trg2 32 �xh4) 3 1 � h 1 + wg2 32 *xh4 etc., but there is nothing wrong with winning Black's queen (and having his king on h2).

28 29 � d2 + 30 jtf4+ 31 J.e4 32 'i!lj'xe4

H g8 jtg2 Ji;i g3 �xe4 Resigns

Game 13 Kupreichik-Kuzmin

Minsk 1982 1 e4 c5 2 <2!f3 �c6 3 d4 cxd4 4 <2!xd4 <2!f6 5 <2!c3 d6 6 Ag5 (50)

58 Classical Variation

50 B

6 jtd7 This is the most popular alter-

native to 6 e6, but there are other moves:

(I) 6 . . . g6 (this appears to be an underrated move) 7 j'txf6 exf6 8 j_c4 (if Stoica's recommenda­tion below is effective, White should prefer 8 j_b5 jtd7 9 0-0 J..g7 10 � de2 with a slight plus) J..g7 (after this Black must sacri­fice a pawn, but 8 A.e7 9 �d2 followed by 0-0-0 is depressing for Black) 9 <2)db5 0-0 1 0 �xd6 f5 l l 0-0-0 �a5 ( l l �g5 + 1 2 f4 �xg2 1 3 e5 is given as good for White by theory, but Stoica suggests 1 3 �g4!, threatening both . . . 't'fxf4 + and . . <2) xe5; at the moment I can't see a good reply for White) 1 2 flfc7 a6 (the lines 12 . . . j_xc3 1 3 bxc3 'lit'a4 14 <2)d6, 12 . . . �b4 1 3 <2)d6 J..xc3 14 bxc3 fifxc3 15 J..xf7 + liih8 1 6 'fi b l and 1 2 . . . fxe4 1 3 �xa5 <2)xa5 14 j_d5 j_h6 + 1 5 liib l are all good for White) 1 3 'iifxa5 <2)xa5 14 � c7 � a7 1 5 j_b3 j_xc3 1 6 bxc3 fxe4 (Kholmov-Cher­nikov, USSR 1 982) and now 1 7

<2)d5! is good for White according to Kholmov.

(2) 6 . . . �aS j'txf6 gxf6 8 jtb5 A.d7 9 <2) b3 �c7 10 <2)d5 �d8 1 1 �h5 e6 1 2 <2)e3 a6 1 3 J..e2 �c7 14 0-0-0 J,.e7 1 5 'fi b 1 0-0-0 16 f4 was good for White in S. Nikolic­Gufeld, Kislovodsk 1 968.

(3) 6 . . . a6 7 �d2 <2)xd4 (other moves transpose to the main var­iations) 8 �xd4 e5 9 �a4 + (White has no trouble keeping a slight advantage by 9 �d3 J..e6 lO 0-0-0 !;! c8 1 1 <2)d5 J..xd5 1 2 A.xf6, but with �a4 + he is play­ing for more) jtd7 10 �b3 b5 1 1 j'txf6 gxf6 1 2 jte2 j_e6 1 3 <2)d5 j'th6 14 a4 with some advantage for White, Marjanovic-Stoica, Istanbul 1 988.

(4) 6 . . . �b6 (this has become popular recently) 7 <2) b3 (7 j_e3!? is interesting, e.g. 7 �xb2 8 'i:Jdb5 �b4 9 jtd2 �c5! l O J..e2!? �b6 1 1 .!!! b1 <2)e5? 1 2 j_e3 �a5 1 3 !;! b3 g6 14 g a3 i!!td8 1 5 J..xa7 <2)ed7 1 6 f4! and White stands well, Balashov-Petrienko, Voro­nezh 1 987, but 1 1 . . . �d8 was the critical test) e6 and now:

(4a) 8 j_f4 <2)e5 9 j_e3 �c7 l O f4 <2)g6 1 1 �f3 J..d7 1 2 J..d3 j_e7 13 0-0-0 j'tc6 14 <2)d4 was good for White in Greenfield-Schren­zel, Israel 1 983, but 8 e5 may be better.

(4b) 8 �d2 a6 9 0-0-0 jte7 (9 . . . jtd7 l O f4 'fllc7 1 1 j_e2 b5 1 2 J..xf6 gxf6 1 3 J..h5 J..h6 14 g hfl was a little better for White, Kho1mov-Petrienko, USSR 1 980, while 9 . . . �c7 l O '* b l b5 1 1

�f4! '2!d7 1 2 .§txb5 axb5 1 3 '2lxb5 itb8 14 l£lx.d6 + Axd6 1 5 'ltxd6 �xd6 1 6 l!l. xd6 l£lce5 was unclear in van der Wiel-Sprag­gett, Wijk aan Zee 1 985) 10 * b 1

Classical Variation 59

when Black continues 7 �xd4 8 'l!ltxd4 *'a5. The move Ae2 is well motivated since if Black exchanges on d4 he has lost a tempo over the usual line, so he has to change his plan. ( 1 0 h4 is a promising alternative

which anticipates Black's 0-0) 'ltc7 1 1 f4 Ad7?! (Black allows a 51 central breakthrough) 1 2 e5 dxe5 B 1 3 fxe5 '2!d5 ( 1 3 l£lxe5 14 Af4 Ac6 1 5 tte1 is very good for White) 14 '2lxd5 exd5 1 5 Axe7 � xe7 16 'ltg5 Af5 1 7 '2!d4 Ag6 18 Ad3 and White has a danger-ous lead in development, lvano­vic-Popovic, Vinkovci 1 982.

(4c) 8 Ad3 (aiming for 0-0 has been the most popular reply, and it has also scored well) Ae7 9 0-0 a6 1 0 a4 ( 1 0 lfi h 1 'ltc7 is also possible, e.g. 11 a4 b6?! 1 2 f4 Ab7 1 3 f5 �e5?! 14 �d4, Kinder­mann-Liberzon, Beersheva 1 984 or 1 1 f4 Ad7 1 2 "�te2! h6 1 3 Ah4 g5?! 1 4 jlg3, Byrne-J. Benjamin, USA Ch. 1 984 with advantage to White in both cases although Black's play was doubtful in these examples) � a5 1 1 �xa5 ( 1 1 Ae3 �c7 1 2 � xa5 'ltxa5 1 3 h3 Ad7 14 f4 Jl;!i, c8 15 itd2 Ac6 1 6 f5 e5 1 7 b4 'f/!tc7 1 8 b5 gave White some advantage, Grunfeld-lvanov, Toronto 1 984) *xa5 1 2 Ad2 ith5 1 3 � e l �g4 14 h3 �e5 1 5 Ae2 �h4 1 6 'l!ltd 1 g5 1 7 f4 gxf4 1 8 � xf4 �g3 1 9 Ag4 �g6 20 �e2 �h4 2 1 g3 and White wins, Iva­novic-Piachetka, Stara Pazova 1 988.

7 Ae2 (51) White usually plays 7 itd2

7 a6 A flexible reply, but 7 . . . �a5

may be better. This and other options:

( I ) 7 . . . �b6 8 � db5 (threa­tening 9 Axf6 and 1 0 �d5, while at the same time preventing e6) 1i c8 9 0-0 a6 10 il, xf6 gxf6 1 1 �d5 !i1d8 1 2 �bc3 e6 1 3 �e3 and White's knights proved well­placed in Vogt-Mascarinas, Pola­nica Zdroj 1977.

(2) 7 . . . e6 8 '2!db5 �b8 9 a4 Ae7 1 0 �d2 a6 1 1 � a3 fijc7 1 2 � d 1 � d8 1 3 �c4 jlc8 1 4 Ae3 'l!lt b8 1 5 Q) b6 with an edge for White, Spassky-Hort, Moscow 1 97 1 .

(3) 7 . . . 'iii'aS 8 .�xf6 gxf6 9 0-0!? and now:

3a) 9 . . . 'l!ltg5 10 '2! f5! � g8 ( 1 0 Axf5 1 1 f4 'l!ltg6 1 2 A h 5 is

good for White) 1 1 '2!g3 with an edge for White.

60 Classical Variation

(3b) 9 . . . �e5 (too ambitious) 10 <2\f3 �c5 ( 1 0 � a5 is better even though this is an admission that Black's last move was a mis­take) 1 1 <E\d5 � c8 1 2 c3 a6 1 3 �d4 .fJ..g7 14 b4 >WJa7 1 5 JJ..h5 with advantage to White, Stoica­Kotronias, Istanbul 1 988.

(3c) 9 . . . � xd4 10 >Wfxd4 >WJc5 (this is Black's best; 10 � c8 1 1 <2\d5 �c5 1 2 �d2 'Wfxc2 1 3 >WJe3 �c4 14 'l!ll\'f4 gives White a strong initiative in return for the pawn) and now Stoica gives 1 1 '/l!\'xc5 dxc5 12 ..§..c4 as slightly better for White. I cannot see any White advantage after 1 2 e6, so I prefer 1 2 <2\d5 � c8 ( 1 2 0-0-0 1 3 il_h5 and now 13 . . . il_e6 14 <2\f4 and 13 . . . Qe8 1 4 Ji;!, ad l are good for White) 1 3 � ad l and White probably does have an edge. White can also consider 1 1 'lt"d3 intending *h I and f4.

8 .fJ..xf6 gxf6 9 0-0

There is a second possibility which, like the main line, is based on the move � f5, namely 9 l£![5 'lt"a5 1 0 0-0 l:l: c8 1 1 �d5 'Wfd8 1 2 �de3! � e5 1 3 f4 �g6 14 >WJd3! h5 15 l:! ad l b5 16 a4 and White is better, Sznapik-Hawelko, Poland 1 984.

9 'lt"b6 9 e6 10 'lfl h l JJ..e7 1 1 f4

�xd4 1 2 >WJxd4 'lt"a5 1 3 M ad l is clearly good for White, Geller­Hort, Palma de Mallorca 1 970.

10 �f5! (52) Many players would have auto­

matically retreated the knight to

b3 but Kupreichik realizes that he can afford to give up his b-pawn. Although Black can organize e6 to expel the knight we have already seen in Vogt-Mascarinas above that knights on c3 and e3 can be well placed.

52 B

10 0-0-0?! Black prepares e6 but it

turns out that this is too slow. He should have tried 10 . . . e6 ( 1 0 �xb2? at once fails to 1 1 � d5 and 1 2 U, b l ) l l l£Jxd6 + il_xd6 1 2 'Wfxd6 >WJxb2 1 3 � fdl and now:

( 1 ) 13 . . . � e5 14 �d5! exd5 1 5 >WJxf6 <E)c4 ( 1 5 M g8 1 6 B, ab l >WJc3 1 7 K xd5 loses at once) 1 6 c3 K f8 ( 1 6 � g8 1 7 JJ..h5) 1 7 � ab l 'Wfxe2 1 8 � e l followed by exd5 + forcing Black to play JJ..e6, when White continues dxe6 with a winning attack.

(2) 13 . . . 0-0-0 14 �d5! exd5 1 5 exd5 'lt"c3 ( 1 5 '*'e5 1 6 dxc6 '*'xd6 1 7 � xd6 _§_xc6 1 8 l:! xf6 wins a pawn) 1 6 dxc6 'i:i;i'xc6 1 7 >WJf4 followed by J..f3 when Black has problems with his exposed king and his weak f-pawns.

(3) 13 . . . � d8! 14 t�rg3 (the

sacrifice 14 l;l ab l '*xc3 1 5 l;l xb7 flops after 1 5 jt c8 16 g e7 + *f8) *a3 1 5 .zl_h5 and White still has some pressure although far less than in the game.

l1 a4 1 1 <E)d5 *a7 only leaves White

with the problem of meeting e6.

l 1 In the Sicilian, Black can nor­

mally only contemplate castling queenside when White has also played 0-0-0, since in a race between attacks on opposite wings the missing Black c-pawn gives White a large head start.

12 a5 'Wfc7 13 <E) a4

Black's . . . <E) b4 has stopped <E)d5 by White but there are other ways to reach b6.

13 e6 1 3 'Wfxa5 14 c3 <E)c6 1 5 b4

'Wic7 1 6 'W/b3 gives White two free tempi and an open a-file for his attack.

14 <E)d4 jtxa4 White could not prevent this

exchange by playing 14 <£) b6 + * b8 1 5 <E)d4 as then 1 5 e5 followed by 'Wfxc2 would con­fuse the issue.

Classical Variation 61

15 M xa4 d5 Black's only chance is to find

some counterplay quickly, or else he will be crushed by c3 followed by b4-b5.

16 c3 <E)c6 Naturally not 16 dxe4 1 7

cxb4 e 5 when 18 � b 1 threatening M c 1 wins.

17 exd5 M xd5 18 .ztf3 )ii( d6

18 liii xa5 19 <E)xc6 � xa4 20 �xa4 bxc6 21 .ztxc6 is one of those positions in which the op­posite-coloured bishops increase the strength of an attack to alarm­ing proportions.

19 lill c4 lli!, g8? Black overlooks the threat. 1 9 . e 5 was necessary but even then

20 �xc6 )ii( xd 1 21 � xd 1 bxc6 22 £:I xc6 gives White an ending with an extra pawn and the better posi­tion, while some players might prefer 20 .� xc6 bxc6 21 'Wf g4 + and 22 LE:\ f5 .

20 'li'Ja4 So simple; c6 collapses and with

it Black's position. 20 21 _ztxc6 22 LE:\xc6

� g5 bxc6 Resigns

4 Pelikan Variation

This line arises after 1 e4 c5 2 lE}f3 �c6 3 d4 cxd4 4 �xd4 � f6 5 <E}c3 e5. Black is willing to accept a backward d-pawn in return for active piece play and, in some variations, the two bishops. The historical background to this line is rather obscure since many players have adopted it over the years with different ideas in mind. The names of Lasker and Pelikan are associated with it, but the modern handling probably owes most to the Soviet Grandmaster Sveshnikov. We have given Pelik­an's name to the whole system with 6 e5, reserving that of Sveshnikov for the 8 b5 varia­tion, today considered the main line. It has gained many other adherents in recent years and is regarded as an excellent way to play for a win with Black, since unbalanced positions arise in almost every line. Very recently it has suffered a decline, probably more because of changing fashions than for any clear objec­tive reason.

There is a second move order by which the Pelikan can arise, namely 1 e4 c5 2 �f3 �c6 3 d4 cxd4 4 ®xd4 � f6 5 �c3 e6 (or 2 . . . e6 and 5 ®c6) 6 �db5 d6 7 jtf4 e5 8 jtg5, reaching the same

position as after 1 e4 c5 2 lE}f3 �c6 3 d4 cxd4 4 �xd4 lE}f6 5 <2lc3 e5 6 � db5 d6 7 jtg5, but in one extra move. To avoid the con­fusion of having two different move numbers in each position I will take the 5 e5 order as standard, although in practice the two move orders are equally com­mon.

Since Black is incurring stra­tegic weaknesses White's most logical (and best) lines are those in which he limits his immediate am­bitions to nullifying Black's piece play and only later turns his mind to the exploitation of his long­term advantage. Our recommen­dation for White runs 6 �db5 d6 7 ji_g5 a6 8 �a3 b5 (the less common line 8 jL e6 is con­sidered in game 14) 9 jtxf6 gxf6 1 0 �d5 f5 ( 1 0 jtg7 is an important alternative) 1 1 jtd3. In this unbalanced position Black pits his two bishops and central pawn majority against White's control of d5 and superior pawn structure. The offside knight on a3 can be an important factor, and White usually aims to bring the knight back into the game by playing c4. All the lines of the Sveshnikov are covered in game 1 5 .

53 B

1 2 3 4 5 6

Game 14 Karpov-Nunn London 1982

e4 <2\f3 d4 l£lxd4 12lc3 <2\db5

c5 l£lc6 cxd4 l2)f6 e5

(53)

6 d6 All Black's 6th move alterna­

tives give White a clear plus: ( I ) 6 . . . a6 7 <2\ d6 + jj_xd6 8

'j!jxd6 'l!je7 9 'l!jxe7 + <2\ xe7 (9 *xe7 1 0 jtg5 <2\ b4 1 1 0-0-0 is similar) 10 jj_g5 and White has undisputed control of d5.

(2) 6 . . . jj_c5 7 jj_e3!? (7 12l d6 + is also good) jtxe3 8 <2\d6 + * f8 9 fxe3 'it' b 6 1 0 12lc4 'l!jc5 1 1 'j!jd6 + 'l!jxd6 1 2 l£lxd6 and f7 is about to come under heavy attack by j_c4.

(3) 6 . . . j_b4 7 a3 jj_xc3 + 8 l£l xc3 d6 9 jtg5 h6 (9 a6 1 0 l£l d 5 is also very pleasant for White) 1 0 j_xf6 'l!jxf6 1 1 <2\ b5 and Black will lose his d-pawn for insufficient compensation.

Pelikan Variation 63

(4) 6 . . . h6 (by preventing j)_g5 Black avoids the loss of control of d5 as in line I, but the move is really just too slow) 7 <2\d6 + jtxd6 8 'l!jxd6 'l!je7 9 l2l b5 (Spassky introduced this pawn sacrifice-the older lines 9 'l!jxe7 + fjxe7 1 0 b3 and 1 0 j_e3 also give White a favourable end­ing) 'l!jxd6 (9 0-0 10 'l!jxe7 <2\ xe7 1 1 iE:l d6 or 1 0 b3 l£lxe4 1 1 'l!jxe7 l£lxe7 1 2 jla3) 1 0 l2)xd6 + *e7 1 1 �f5 + frf8 1 2 b3 d5 ( 1 2

l£lxe4 1 3 jj_a3 + *g8 and now 14 f3 or 14 iE:l d6) 1 3 jj_a3 + *g8 14 exd5 �xd5 1 5 <2\ d6 I! b8 1 6 A c4 jJ_ e6 1 7 0-0-0 and White has a very pleasant position, Spassky­Gheorghiu, Bath 1 973.

7 J..g5 a6 Black must meet the threat of

<2\ d5 so the only other move is 7 jte6, but then White does not

need to retreat his b5 knight to the bad square a3 and can gain the advantage by 8 l2)d5 � c8 (8 jtxd5 9 exd5 <2\e7 is good for White after 1 0 �f3 or 1 0 c3 a6 1 1 'i!ita4) 9 c3 a6 1 0 l£l a3 j_xd5 1 1 J_ xf6 gxf6 1 2 'l!jxd5 'l!ja5 1 3 jtc4, Jansa-Danek, CSSR Ch. 1 982, with a firm hold on d5 in both cases.

8 l2l a3 jJ_ e6 8 b5 is the line popularized

by Sveshnikov and is examined in game 1 5. Other moves are defini­tely inferior:

( 1 ) 8 . . . jte7 (Black commits the bishop too soon) 9 <2\c4 jte6 (9 . . . 12ld4 1 0 j)_xf6 J_ xf6 1 1 iE:ld5 b5 1 2 l2)cb6 � b8 13 l2)xc8 � xc8

64 Pelikan Variation

1 4 c3 �e6 l S a4! is very good for White, Averbakh-Korchnoi, semi-final USSR. Ch. 1 9SO, 9 . . . 0-0 10 �xf6 J..xf6 1 1 � xd6 doesn't give Black enough for the pawn and 9 bS 10 �xf6 gxf6 1 1 �e3 gives White a crushing bind) 1 0 J..xf6 gxf6 1 1 � e3 (thanks to Black's �e7 he cannot now dislodge the knight by �h6) '*l'd7 1 2 L£)cdS fol­lowed by �d3 and '*l'hS, once again with a total white-squared bind.

(2) 8 . . . d5 (not correct) 9 �xdS �xa3 1 0 bxa3 �a5 + 1 1 �d2 'l'!!fxd2 + 1 2 �xd2 -2J xdS 1 3 exdS �d4 14 il_d3 followed by 0-0 and f4, when White has two bishops in an open position and a moderately relevant extra pawn.

9 �c4 i;i, c8 (54) 9 . . . iJ.. e7 transposes to line 1 of

the last note, while the alternative 9 . . . -E:ld4 (9 b5 10 Jtxf6 �xf6 1 1 �e3 and 1 2 �cdS is very good for White) often leads to the knight being driven back with loss of time, e.g. 9 �d4 10 J..xf6 gxf6 ( 1 0 *xf6 1 1 � b6 .M b8 1 2

54 w

�cdS �d8 1 3 c3 leaves Black a tempo down on Karpov-Nunn) 1 1 �e3 l! c8 12 itd3 ith6 ( 1 2 . hS 1 3 0-0 h4 14 �cdS il_g7 l S c3 �c6 1 6 •f3 is excellent for White, Bronstein-Pilnik, Moscow l 9S6) 1 3 0-0 0-0 14 � cd5 and again Black's knight is expelled by c3, Dely-Flesch, Hungary Ch. 1 96S.

10 il_xf6 �xf6 With this move Black accepts a

loss of time to prevent his pawn structure being damaged but as a result he is driven into a passive position. The important alterna­tive 10 gxf6 is met by 1 1 �d3 'E:le7 (11 . . . .£ld4 1 2 �e3 trans­poses to the last note while l l . . . l! g8?! 12 0-0 J.. h6 1 3 � dS f5 14 �hS! J..f8 1 S �cb6 f4 16 �xc8 1l.g4 1 7 �xh7 l! g7 1 8 )llfxg7! 1l.xg7 19 �cb6 gave White too much for the queen in Matulovic­Arnason, Zeman 1 983) 1 2 � e3 il.h6 ( 1 2 . . . �b6 1 3 0-0! �xb2 14 �cdS il_xdS 1 S �xdS �xdS 16 exd5 �d4 17 �f3 'ftje7 1 8 a4! M c7 1 9 M fd 1 .-c3 20 I:� ab l .-as 2 1 .-e4 gave White more than enough for the pawn in Mednis­Lombardy, USA Ch. 1 978 -this type of pawn sacrifice in return for white-squared pressure and attacking chances occurs fre­quently in the 1 0 gxf6 line) 1 3 0-0 ii_xe3 1 4 fxe3 �b6 I S .-o hS ( I S .-xb2? 16 �dS j_xdS 1 7 exdS i s bad for Black) 1 6 �dS JtxdS 17 exdS � h6 18 � ab l .-as!? ( 1 8 . 'i! c7 1 9 c4 was good for White in Tseshkovsky-

Chandler, Minsk 1 982) 1 9 e4 f5 ( 1 9 'ltxa2? 20 'lte3 � h8 2 1 �b6 wins) and there i s n o doubt that White is better, although it is not clear by how much. In de Firmian-Matulovic, Vrnjacka Banja 1 983 White played 20 a3 'fJf8 and now 2 1 c4 or 2 1 exf5 is good for White, but in Kolosovs­kaia-Suspanova, corr. 1 986 Black muddied the waters by 20 a3 f4!? 2 1 c4 �g6 22 � fc 1 'ltd8 23 b4 � h4. Therefore the simple 20 exf5 may be stronger, when White's effective bishop and Black's mis­placed rook on h6 guarantee White at least a slight advantage.

1 1 �b6 It is very dangerous for White

to take the pawn by 1 1 �xd6 + 1l._xd6 1 2 �xd6 when 1 2 g d8! 13 'lllfc5 �d4 1 4 1l._d3 �g5 1 5 'fJfl � c8 1 6 '!ttb4 b5 gives Black ad­equate compensation.

1 1 � b8 12 �cd5 �d8

1 2 'i!i(rg6 1 3 ttd3 1l._e7 14 ttJc7 + \t1d8 15 �cd5 followed by 0-0-0 gives White a dangerous at­tack.

13 c3 J...e7 Attempting to develop the bish­

op more actively runs into trouble after 1 3 g6 14 �a4! 1l._xd5 1 5 �xd5 1tg7 1 6 _txa6 .ii a8 1 7 jj_xb7 � xa4 1 8 _txc6 + with a clear plus for White.

14 1l._c4 Karpov correctly steers clear of

the complications resulting from the win of a pawn by 14 �c4 0-0 1 5 �xe7 + �xe7 16 'ltxd6 �h4,

Pelikan Variation 65

and quietly consolidates his grip on d5. White's knights are rather clumsily placed but Black's possi­bilities for active play are very limited and aiming for f5 is his only constructive plan.

14 0-0 15 0-0 J..g5 16 a4 *h8

Chekhov suggests the imagina-tive 1 6 )'!tl(e8 intending 1 7 jj_d8, but after 1 7 a5 (with the queen on e8 Black can no longer meet this move by �xa5) j_d8 18 b4 \trh8 19 ,!,;\ a2 f5 20 exf5 J... xf5 2 1 � e3 White has the better chances.

17 �e2 (55) This move and the next are

evidence of Karpov's understand­ing of the position. White would like to play 1 7 a5 defending the b6 knight and thereby freeing the tangle of minor pieces, but at the moment it just allows 1 7 �xa5! 1 8 )g( xa5 .§,xd5. I t seems natural, therefore, to prepare a5 with 1 7 b4 but in Sznapik-Simic, Smederevo 1 98 1 , Black obtained active play by 1 7 b4 f5 1 8 b5 �a5 1 9 J..d3 g6 and equalized comfor­tably.

Karpov's first concern is to take the sting out of . . . f5 by preparing to answer it with exf5 and f4. For this purpose 'Wt'e2, which pins the e-pawn against the loose bishop on e6, and his next move 'fJ h l removing the king from the vul­nerable diagonal are excellent preparation. Only when Black's counterplay is completely neutra-

66 Pelikan Variation

lized does White return to the exploitation of his queenside space advantage and d5 control.

17 g6 I made use of Karpov's chess

lesson five years later in the game Nunn-Manor, London (Lloyds Bank) 1 987, which continued 1 7

a5!? 1 8 J,i ad l J\l.h6 (after 1 8 g6 1 9 jla2! Black should

avoid 19 . . • J;Je7 20 'i'Jii b5 and 19 . . . jlh6 20 Qlc4 f5 2 1 exf5 gxf5 22 J;Jdb6 with a clear plus for White, but even his best line 1 9 f5 20 exf5 gxf5 2 1 Qlc4 f4 22 f3 is slightly better for White) 1 9 \\> h l .:2:Je7?! ( 1 9 g6 was more solid, as in Karpov-Nunn) 20 e,xe7 �xb6 (20 �xe7 21 jlxe6 fxe6 22 trb5 is very bad) 2 1 .:2l f5 and Black has no compensation for his serious weaknesses. White went on to win.

18 \\>hi 19 b4?!

ji_h6

Afterwards Karpov thought that this was still too soon and that 1 9 � ad I or � ael would have been better.

19 20 exf5

f5 gxf5

21 f4 jlxd5 Black hopes for salvation in the

opposite-coloured bishops, but White's bishop has a fine outpost at d5 whereas Black's is rather useless.

22 .:2:J xd5 e4?! 22 .:2:Je7 at once is more

logical, based on the fact that White cannot win a pawn by 23 fxe5 .:2:Jxd5 24 jlxd5 dxe5 25 *xe5 + because of 25 ji_g7 26 'lte6 � f6.

23 a5 ji_g7 24 !i acl .:2:Je7 25 � fdl .:2:J xd5 26 Axd5 'ltc7 27 � c2 'lte7 28 'l!il'e3 .i, bc8 29 c4 14. c7

White has two ways of making further progress. He can either play b5 to leave Black with an isolated pawn on the queenside, which will be hard to defend when his bishop is operating only on the kingside, or he can prepare g4 to attack Black's king. For the mo­ment White is not sure which plan offers the best chances.

30 g3 � e8 31 14. g2 trf6 (56) 32 g4

This doesn't have the desired effect and it would probably have been better to try the other plan. If Black moved his forces to the queenside White could then have contemplated g4 later.

32 fxg4 33 M. xg4 'ltc3 34 [4. g3 �xb4

If Black swaps queens Karpov gives the line 34 ytxe3 35 B, xe3 .!, ce7 36 b5! J..h6 37 H fl I;!; f8 38 ,g xe4 � xe4 39 J.. xe4 .!. xf4 40 l! xf4 J..xf4 4 1 j_xb7 and wins.

35 H dgl ttb2 Stopping White's threat of

'ii!'d4. 36 ,!! g5 ytf6 37 H lg4 �at + 38 ffg2 �b2 + 39 ffh3 ,!! ce7 40 f5 �f6?

This was the sealed move (move 4 1 in the game, which started with the 2 . . . e6 move order) and, as so often happens, after a long period of difficult defence a player's relief at reaching the time control re­sults in a casual sealed move. White obviously has considerable pressure for the pawn but after 40

H f8 (Karpov also suggests 40 � a l threatening . . . 'ftl'fl + ) it

is likely that Black can draw. Black's passive queen move gives White the freedom of action he needs to mount the decisive as­sault.

41 H h5 H ffi

Pe/ikan Variation 67

42 l! gh4 h6 White cannot now play 43

.!! xh6 + J..xh6 44 !! xh6 + due to 44 'ftl'xh6 + 45 Y:llfxh6 + H h7.

43 l! g4 ,!! e5 Forced in order to meet 44 !it g6

by 44 *'xg6. 44 � gg5 H c8 (57)

There isn't much Black can do to meet the threat of ltrg4 fol­lowed by !it g6.

45 ltrg4! *h7?

Going under without a fight. I should have tried 45 !:il, xd5 46 cxd5 !it c2 although White is still winning even after this.

46 H g6 �f8 47 �g5 �xf5 +

Or else 48 H gxh6 + .fl_xh6 49 �g6 + is the end.

48 �xf5 49 H xg7 + 50 H xf5

H xf5 ffxg7 Resigns

Game 1 7 Short-sax

Candidates' Match Saint John 1988

1 e4 c5 2 <El f3 12\ c6

68 Pelikan Variation

3 d4 cxd4 4 i£)xd4 i£)f6 5 i£)c3 e5 6 i£)db5 d6 7 jlg5 a6 8 i£)a3 b5 9 jl xf6 gxf6

9 �xf6 lO �d5 �d8 1 1 c4 i£)e7 ( 1 1 b4 1 2 �a4 j_d7 1 3 i£)b5! axb5 1 4 �xa8 �xa8 1 5 J}Jc7 + wins material) 1 2 cxb5 €:lxd5 13 'l!lfxd5 ( 1 3 exd5 is also safe and good) j_e6 14 �d2 d5 1 5 bxa6 J..xa3 1 6 J.. b5 + ! gives White a clear advantage.

10 i£)d5 (58)

10 f5 This is an already an important

moment for Black. At one time 10 f5 was the only move played in

this position, but recently the alternative lO j_g7 has become popular. The idea behind this move is to delay Black's the­matic break f5, giving first priority to removing the powerful knight from d5 by t;Je7. Theory gives no clear-cut recom­mendation against this new plan, so we give a more detailed cover-

age than usual: l O J..g7 ( 1 0 J..e6 1 1 c3 J..g7 1 2 i£)c2 f5 1 3

exf5 J..xf5 1 4 i£\e3 leaves Black a tempo down over a standard line and after 14 . . . J..e6 1 5 �f3 0-0 1 6 j_d3 or 14 . . . J..g6 1 5 a4 White has good chances) and now:

( 1 ) 1 1 -tth5 ( 1 1 i£)e3 i£) e7 1 2 � h5 is the same, while 1 1 c4 0-0 1 2 cxb5 <E!d4 is very risky) i£)e7 (better than 1 1 f5 1 2 exf5 �a5 + 1 3 c3 b4 14 �g5! �xd5 1 5 �xg7 bxa3 1 6 �xh8 + fr d7 1 7 b3 with advantage to White, Sir­ias-Ochoa de Echaguen, Havana 1 984) 1 2 �e3 f5! 1 3 exf5 e4 14 0-0-0 0-0 and now:

( l a) 15 f3?! �b6 16 �g5 d5 1 7 i£1 xd 5? ( l 7 f6 'l¥t' xf6 1 8 � xf6 j_ xf6 1 9 i£)xd5 /2lxd5 20 ll( xd5 J..e6 was an improvement, but Black is at least equal) i£)xd5 1 8 !;! xd5 frh8! and Black is better, Klinger-Vaiser, Szirak 1 985.

( l b) 15 h4 ( 1 5 g4!? is untested) d5!? 16 l! h3 �d6 1 7 � g3 �e5 1 8 c3 b4 19 i£) ac2 bxc3 20 bxc3 l! b8! (not 20 . . . '1!!4'xc3? 21 K d4!, nor 20 . . . 12lxf5? 2 1 K xg7 + !) 2 1 12ld4 j_d7 2 2 K d2, Vakhnov­Shipkov, USSR 1 987 and now 22

frh8 23 12l g4 �f4 24 f6! J..xf6 25 i£) xf6 �xf6 26 �g5 \'4fxg5 27 hxg5 M b6 is roughly equal. This murky variation must be a good place to look for a White im­provement!

(2) 1 1 j_d3 i£)e7 1 2 12lxe7 (not 1 2 '2!e3 f5! 1 3 'fllfh5 d5!) 'fllfxe7 (59) with a branch:

(2a) 13 c3 f5 14 i£) c2 0-0 (14 . . .

d5?! I 5 exf5 e4 1 6 j_e2 and 14 . • . j_b7 I 5 121e3 fxe4 I 6 121 f5 �f6 I 7 ll.xe4 d5 I 8 J,.xd5 H d8 I 9 �g4 j_xd5 20 121xg7 + '11vf8 2 I 121h5 are good for White, but 14 • • . �g5 I 5 0-0 J..b7 I 6 f3 0-0 I 7 exf5 d5 was unclear in Feigeison­Malyshev, USSR I 988) I 5 0-0 d5! (not 15 . . . fxe4?! I6 J..xe4 lit b8 1 7 121 b4, while 15 . . . � b8 I 6 exf5! e4 I 7 Ji e I 1l. xf5 I 8 121 e3 1l. g6 I 9 121 d5 �e5 20 j_c2 f5 2 I f4 'jffe8 22 121c7 �c6 23 121e6 led to a slight plus for White, Dolmatov-Simic, Belgrade I 988) 1 6 exf5 e4 I 7 1l. e2 � d8 I 8 121d4 j_xd4 I 9 cxd4 j_xf5 20 �d2 j_e6 2 I f4 f5 Draw, Psak­his-Dolmatov, Klaipeda I 988.

(2b) 13 c4 f5 I4 0-0 (not I 4 •d2?! j_ b 7 I 5 exf5 e4 I 6 ll.e2, Hardicsay-Vaiser, Berlin I 988 and now I 6 e3! I 7 fxe3 j_xg2 I 8 H g i �h4 + 19 '11vd l �xh2 20 '\1vc2 j_e4 + 2I 'fj' b3 bxc4 + 22 121xc4 � b8 + 23 '\1va3 ll.f8 wins for Black) 0-0 (I4 j_b7!? is interesting, e.g. 15 cxb5 fxe4 I 6 bxa6 exd3 1 7 axb7 �xb7 I 8 �xd3 d 5 1 9 � adl and now 1 9 . . • 0-0 20 �xd5 *xb2 2 I 121c4 �xa2 22 � xe5 �xd5 23 Ji xd5 j_xe5

Pelikan Variation 69

should be a draw, while after 19 . . . �xb2 20 M bl •xa3 2 I � b8 + *e7 22 !;!; b7 + '\1ve6 23 Ji,i: b6 + 'fj'e7 White unwisely dec­lined the draw by repetition and lost in Solozhenkin-Skvortsov, Berdiansk 1 987; 1 5 t�tf3! is proba­bly the best reply) with a wide choice for White:

(2b I ) 15 cxb5?! d5 and I 5 'ti!i'e2 j_b7! I6 )1:( ad l � fd8 1 7 cxb5 ( 1 7 exf5?! bxc4 I 8 .£Jxc4 d5 1 9 121e3 e4 20 ll.b1 �f6 is also good for Black) fxe4 1 8 j_xe4 d5 I 9 bxa6 j_xa6 20 1l.d3 1l_b7! and Black has very good compensation for the pawn.

(2b2) 15 �h5!? );;t, b8 (15 . . . bxc4 1 6 .£J xc4 d5 1 7 exd5 e4 1 8 J..e2 and 15 . . . 'Wt'b7 1 6 l!( fe i d5 17 exd5! e4 I8 ll.fl j_xb2 I 9 E( ab i , Vogt-Chekhov, Berlin I 988, are good for White) I 6 exf5 e4 I 7 � ae l j_b7 and now:

(2b2 1 ) 18 � e3 bxc4 19 1l_xc4 ( 1 9 � h3 h6 20 ll.xc4 d5 2 I J..xd5 j_xd5 22 f6 �xf6 23 �xd5 li xb2 24 •xe4 !i( xa2 25 12\c4 *e6 26 '*d3 'Wfe2 27 �d5 �e6 28 'Wtd3 *e2 29 •d5 �e6 Draw, Smagin­Vaiser, Sochi 1 988) d5 20 1l_xd5 .ffi.xd5 2 I f6 �xf6 22 �xd5 l!( xb2 23 � xe4 lit d8 24 �c4 �c3 25 �xa6 •c5 26 lili, c4 •f5 was unclear in Hjartarson-Yusupov, Munich 1 988.

(2b22) 18 �g4 liii fe8 19 cxb5 ( 1 9 f3 bxc4 20 j_xe4 �f6 was also good for Black in Kuijf-Schmitt­diel, Luxembourg I 988) d5 20 bxa6 il_c6 2 I a7 !i( xb2 22 l!( c i j_d7! 2 3 j_ b5 h5! 24 f6 hxg4 25

70 Pelikan Variation

fxe7 ll_xb5 and Black is better, Bosboom-van der Wiel, Nether­lands Ch. 1 988.

(2b23) 18 cxb5 d5! 19 bxa6 ll.. c6 20 )ii e3 � xb2 21 ll_ b l l!il fb8 with a clear plus for Black, Kolotilin­Shipkov, USSR 1 984.

(2b3) 15 � et fxe4 ( 1 5 . . ll_b7 1 6 exf5 e4 1 7 �g4 � fe8 1 8 cxb5 d5 19 bxa6 j_xa6 20 j_xa6 � xa6 2 1 �b5 was good for White in Ivanovic-Ivanov, Belgrade 1988) 1 6 j_xe4 � b8 1 7 cxb5 axb5 and now 18 �b3 j_e6 1 9 j_d5 J,i fc8 20 !!! adl J..f5! 2 1 j_e4?! (21 � c l would have been equal) j_xe4 22 � xe4 �e6 gave Black some ad­vantage in Klovans-Ivanchuk, Frunze 1 988. However 18 iitd3!? is interesting ( 1 8 �c2 f5 1 9 j_ d 5 + J.. e 6 20 � b4 i s also pos­sible), when Nunn-Kosten, Bri­tish !-hour Ch. , continued 1 8 f5 1 9 ll_d5 + fth8 20 � xb5 e4 2 1 '(!ltb3 �e5 22 a4 j_d7 2 3 � e2 !( fc8 24 I;i d l � c5 25 ll_c4 f4? 26 � xd6! and wins. Slim evidence, but there are unexplored ideas in this line.

At the moment Black's chances in the I 0 ll..g7 line appear quite good, but a number of lines above, such as 1 b and 2b3, de­serve further investigation by White players.

1 1 ll.. d3 J.. e6 1 1 �g5 ( 1 1 f4 1 2 g3! is

also good for White) 1 2 g4! ftd8 1 3 gxf5 ll_xf5 14 �e3 J..e6 1 5 �d2 gives White a positional ad­vantage.

12 0-0 (60)

12 j_g7 Or 1 2 j_xd5 ( 1 2 f4?! 1 3

c4! is very good for White after 13 bxc4 14 j_c2! J..g7 15 Jj_a4

I;;i, c8 1 6 � xc4 or 13 . . . b4 14 l:!Ya4! j_d7 15 �b5) 1 3 exd5 � e7 14 �xb5 j_g7 1 5 �c3 e4 1 6 j_c4 �a5 ( 1 6 �c7 1 7 �e2! puts pressure on the a6 pawn) and now:

( 1 ) 17 jj<d2 � c8 1 8 j_b3 (not 18 .£)xe4? �b6 19 'j#g5 � g8 20 .£)d2 j_c3! 2 1 'j#xg8 + .£)xg8 22 bxc3 'j#a5 with a clear advantage for Black) j_xc3 1 9 bxc3 �xc3 20 ii((h6 \frd7 21 g ae l ;g, cg8 22 'ii!'e3 (22 g3 � g6 23 �cl � hg8 is unclear) 'j#xe3 23 g xe3 .£)g6 24 f3 .£) h4 25 ll_ a4 + 'lfd8 with a sharp but balanced ending, Malishauskas-Krasenkov, Vilnius 1 988.

(2) 17 .£)e2! ll_xb2 ( 1 7 �c5 is well met by the exchange sacri­fice 1 8 b3!) 1 8 � b l J..e5 was played in Short-van der Wiel, SWIFT Brussels 1 987. Here Short played 19 � c l and although White was still slightly better Black managed to escape from his difficulties. However White could

have tried other moves, for ex­ample 1 9 l! b7 l! c8 20 jt b3 iJ. c7 2 1 .!;i b8 + X c8 22 .!;i xc8 + .:21 xc8 23 12:Jg3 tz:Je7 24 �h5 appears good for White, as does the sim­ple 1 9 12:Jg3, and if 1 9 !:! c8 then 20 �e2.

13 �h5 f4 Not 1 3 0-0? 14 exf5 Jtxd5

1 5 f6 e4 16 fxg7 !:! e8 17 �xd5, nor 13 h6?! 14 f4! opening the position while Black's king is still in the centre.

14 c4 bxc4 After 14 . . . b4 ( 1 4 j'J_xd5 1 5

exd5 12:Je7 1 6 M ad 1 b4 1 7 tz:Jb1 12:Jg6 1 8 g3 with advantage to White, Vogt-Georgadze, Halle 1 978) 1 5 12:Jc2 X b8 White has a variety of promising ideas, since with d5 secured he can play on the queenside with b3 and a3, or on the kingside with g3. Finally Qle1 -f3-h4-f5 can be awkward. The speculative 14 0-0 1 5 cxb5 12:Jd4 1 6 Ql c2 Ql xb5 ( 1 6 .:2)xc2 1 7 Jtxc2 axb5 1 8 Jt b3 is a safe positional plus for White) is dubious after 17 a4 ( 1 7 Qlcb4 12:J d4 1 8 � c I is also promising) .:2)a7 1 8 a5 .:2)c6 1 9 b4 f5 20 .:2)b6 � a7 2 1 exf5 J..f7 22 �h3 �f6 23 j_e4 12:Je7 24 � ad ! and White keeps control of d5.

15 Jtxc4 0-0 16 l! acl (61) tz:J e1

The position after 1 6 .!;[ ac l is the key to the whole line and is critical for the assessment of Sveshnikov's variation. Black has a wide range of options:

( I ) 16 . . . 12:Jd4?! 1 7 12:Jc2! 12:J xc2

Pelikan Variation 71

( 1 7 f3 18 Qlxd4 fxg2 19 12:J f5! gxfl ( ilt) + 20 fi' xfl Jt xf5 2 1 exf5 h6 22 lli!, c3! *g5 23 �xg5 hxg5 24 12:Je7 + 'lrh7 25 M h3 + Jth6 26 f6 e4 27 Jt d5 wins for White) 18 � xc2 frh8 1 9 .!;i fc 1 M a7 20 b4 M g8 2 1 a4 with an excellent posi­tion for White, Mokry-Vodichka, Decin 1 979.

(2) 16 . . . H a7!? and now: (2a) 17 12:Jxf4!? exf4 ( 1 7

Jtxc4? 1 8 l! xc4 12:J a5 1 9 � a4 exf4 20 !:! xa5 Jtxb2 2 1 .:2)c4 is good for White) 1 8 J..xe6 .:2)e7! (not 1 8

fxe6 1 9 )i:l xc6) 1 9 Jtc4! (after 1 9 Jtd5 or 1 9 J..f5 Black can take the bishop and then play 20 j_xb2, when the sacrifice 21 M b 1 Jtxa3 2 2 � b 3 Jt c 5 2 3 � h3 doesn't work because Black can advance his f-pawn and defend the second rank) 12:Jg6 (not now 1 9 Jtxb2? 20 1! b l j_g7 2 1 1! b3 intending M h3) 20 Ji:I c2 � e7 21 ltd5 ll[ e5 22 �d1 �h4 23 12:J c4 l! h5 24 h3 l! g5 25 fiih 1 Jte5 and Black has enough for the pawn, Dvoris-Basagic, Sibenik 1 988.

(2b) 17 K fd1 � b8!? ( 1 7 <�ih8? 1 8 .:2)xf4 exf4 1 9 Jtxe6

72 Pelikan Variation

really does work because f7 hangs after l£)e7) 1 8 b3 ( 1 8 <2}xf4 exf4 1 9 jtxe6 l£)e7 is still unclear) <2}d4 1 9 <2}c2 1£\ b5 and now 20 l£)cb4! appears good for White.

(3) 16 . • . 'ltfh8 1 7 � fd l l£) d4 1 8 L2!c2 is slightly better for White.

(4) 16 . . . � b8 (the main line) 1 7 b3 (62) and now:

(4a) 17 • . . <�ih8 1 8 l£)xf4!? exf4 1 9 Jlxe6 l£) d4 20 ]tf5! l£) xf5 2 1 �xf5 ]tb2 22 l£\c4 ]txc l 23 M xc 1 is a very promtsmg exchange sacrifice, while 17 . . . Jlxd5 1 8 ]txd5 o2l b4 1 9 .!! fd l l£\xa2 20 � c6, with l£\c4 to come, leaves Black very badly placed.

(4b) 17 • . . �a5 (a recent idea) 1 8 1£\b l !? ( 1 8 l£) c2 �xa2 should lead to a draw after 19 �h4 h6! 20 l£) f6 + ]txf6 2 1 '*'xf6 ]txc4 22 �xh6 ]txfl 23 �g5 + ) 'fph8 (not 1 8 �xa2? 1 9 l£) bc3 �a5 20 ll;i a l lltc5 2 1 :l;ii, xa6) 19 l£)bc3 f5 20 l£\e7 <2}xe7 2 1 ]txe6 :�;!( f6! (not 2 1 fxe4? 22 l£)xe4 d5 23 ll! c5 lit b4 24 1£\ g5 h6 25 K c7) 22 exf5 e4! 23 llte2! (not 23 l£)xe4? .!! xe6 24 1£)g5 �xf5 and Black wins) d5, Estevez-Timoshchenko, Mana-

gua 1 988, and now 24 �d2! f3 (24 l£)xf5 25 �xd5 �xd5 26

]txd5 e3 27 l£)e4 is very good for White) reaches the critical posi­tion. Black has some compensa­tion for the pawn after 25 l£) xe4 �xd2 26 l£\xd2 fxg2 27 'fpxg2 l£\xf5, but 25 g4! appears quite unpleasant for Black (25 ]th6 26 �d4).

(4c) 17 . . . �d7 18 � fd l <�ih8 (not 18 . . . ]tg4? 19 �g5, while 18 . . . -2l d4 19 1£\ c2 1£\ xc2 20 .liii xc2 *h8 2 1 � d3 ]tg4 22 �h4 f5 23 f3 fxe4 24 fxe4 is a little better for White) 1 9 �h4 and now:

( 4c 1 ) 19 . . . f5!? 20 l£)xf4 exf4 2 1 j_xe6 ·Wfxe6 22 !! xc6 fxe4 23 � cxd6 �e8 reaching a very sharp position. It is perhaps surprising that nobody has repeated 1 9 f5!?, because i n the only game played, Stanciu-Brkovic, Pernik 1 983, the continuation 24 l£\c4 e3 25 fxe3 (not 25 � e l jtc3) fxe3 26 � e l �fl! 27 l£)xe3 iiiK be8 28 l£\c4! � e4! 29 �g3 was satisfac­tory for Black since he could now have forced a draw by 29 . ]td4 + 30 .!;it xd4 .14 xd4 3 1 'l!l!fe5 + �g7 32 lltxg7 + *xg7. It seems to me that it was not necessary to play l£\c4 at once, and White should first try to defuse the ad­vance of the e-pawn. Therefore 24 �g4 is logical, and after 24 . . . e3 simply 25 �f3 (25 �e2!?), with l£\c4 coming next move, and Black has to prove that he has enough for the pawn. In any case no other players have been con­vinced by Black's sacrifice.

(4c2) 19 Jt xd5 20 J.xd5! (stronger than the old continua­tion 20 g xd5) � b4 (after 20 . . . €)d4 2 1 l!l c4! both 21 . . . f3 22 �c2 �e2 + 23 fl fl �f4 24 gxf3 and 21 . . . f5 22 �c2 fxe4 23 �xd4 exd4 24 Jtxe4, Svesh­nikov-Vyzmanavin, Moscow 11 1 987, are good for White) and now:

(4c2 1 ) 21 l!l d2 f5 22 �c4 (22 � h3 should be met by 22 . . . �xd5 23 li;ii xd5 � b7 24 'lifd3 fxe4 25 'ittxe4 f3 rather than 22 . . . 'life7 23 exf5 e4 24 �c4 g f6 25 li;ii e l e3 26 fxe3 M h6 27 'li!'g4 J.c3 28 lit xf 4 Zil f6 29 � ed I it xd2 30 !:ll xd2 with advantage for White, Klovans-Vyzmanavin, USSR 1 987) �xd5?! (22 .1i b5 23 !:ll cd I �xd5 is better) 23 � xd5 fxe4 24 !:ll xd6 �f5 25 h3! h5 26 �e7 e3 27 fxe3 fxe3 28 li d7! and White stands well, Renet-Korch­noi, Lugano 1 988.

( 4c22) 21 Ji c3!? � xa2 (after 2 1 . . f5 22 jj, h3 1t f6 23 ii_e6! both 23 . . . 'Wfe7 24 �h5 fxe4 25 � c4 d5 26 �d6 and 23 . . . 1t xh4 24 ii_xd7 are good for White) 22 l;i!, h3 h6 23 �c4 � bd8 24 g4! and White has dangerous attacking chances since if Black exchanges on g3 the route is open for White's knight to move to f5 .

17 � fdl There is an interesting alterna­

tive in 1 7 �c7!? Jtxc4 ( 1 7 �xc7 1 8 1txe6 'ii!'b7 1 9 1tb3! is good for White since 19 . . . �xe4 loses to 20 Jt c2) 1 8 � xc4 � c8 1 9

Pelikan Variation 73

I! fc I ; perhaps 19 . . . 'lifd7 is the best reply.

17 K c8 18 �xe7 + 'lifxe7 (63)

V p to this point the game has followed Matanovic-Sax, Buenos Aires 1 978, which continued 1 9 �e2 *h8 20 1txa6 f5! with dan­gerous counterplay. Black went on to win and later it was sug­gested that 20 .I c2 followed by doubling rooks on the c-file would have been good for White. Short prefers to double rooks on the d­file with his queen still actively posted on h5.

19 .li c3! Black has no immediate

threats, so White can afford to take time out to prepare b3 fol­lowed by }i! cd3. Black must aim for f5, since this provides the only possible counterplay to off­set the backward d-pawn. The im-mediate 1 9 f5 fails to 20 exf5 .

19 'fih8 Short's innovation caused Sax

to use a lot of time over this and his next few moves, so that before long White was an hour ahead on

74 Pelikan Variation

the clock. 1 9 !K c6? 20 A xe6 � xc3 2 1 A f5 wins.

20 b3 f5 20 d5 2 1 exd5 Axd5 22

!K xd5 �xa3 fails after 23 Ad3 e4 (23 . . . h6 24 �f5 and 23 . . . f5 24 Axf5 are even worse) 24 A xe4 h6 25 � xc8 jj( xc8 26 h3. Black could have tried 20 Ad7, but then White can keep the advantage by 2 1 12J b I ! (intending ll cd3 and 12Jc3) f5 22 I! cd3 fxe4 23 !! xd6 followed by 12Jc3.

21 l!l( h3 h6 22 Axe6 �xe6 23 I! hd3 .!t cd8

Although this looks passive it is the best way to defend the d­pawn, for example 23 � c6 24 �f3 or 23 . . . fxe4 24 liil xd6 'Wte7 25 12Jc4 with a clear plus for White in both cases. There is little point leaving a rook on the c-file when it will soon be blocked by 12Jc4.

24 �e2! It is easy to win a pawn, but if

this involves allowing Black to advance his central pawns Black's powerful bishop may well provide enough compensation. One such line is 24 12Jc4 (24 �f3? �g6! 25 exf5 �xf5 followed by e4 is also bad) fxe4 25 liil xd6 I! xd6 26 liil xd6 �e7 27 a xa6 e3 28 fxe3 fxe3 and the threats to White's king practically force him to play 29 � xh6 + (29 �xe3 �c5 30 'Wt'e2 e4 threatening A d4 is awkward) A xh6 30 �xh6 + \fig8 3 1 'ltxe3, but 3 1 •f6 32 h3 �fl + 33 * h2 �f4 + is a likely

draw because the knight and queenside pawns cannot assume a stable defensive configuration. White correctly prefers to keep the bind and cash in later.

24 fxe4 24 . �g6 25 f3 d5 (25 . . . fxe4

26 '*xe4 is good for White) 26 li xd5 a xd5 27 exd5 e4 28 12! c4 e3 may seem unclear, but White's passed pawn is just as far ad­vanced as Black's and he has an extra pawn.

25 �xe4 f3 26 �c4! (64)

Much better than 26 a xf3 d5 27 'i!i!'h4 (27 liil xf8 + Axf8 attacks a3) /4 fe8 followed by the advance of the central pawns, with good compensation for the pawn.

26 li!l f4 26 fxg2 27 � xd6 � xd6 28

liil xd6 '/!tff7 29 liil d2 and 30 �e3 is also very good for White.

27 �d5 'ittg4 Black cannot avoid a miserable

ending. Although he now suc­ceeds in playing d5 his bishop remains blocked by the e5 pawn.

28 lii[ xf3 lii[ xf3

29 �xf3 �xf3 30 gxf3 d5 31 'llfl J..f6

Despite White's doubled pawn the position should be a win since Black's d-pawn is easily block­aded by White's king and the bishop will be impeded by the central pawns. Moreover White has a powerful queenside pawn majority.

32 i£jb6 d4 33 '81e2

33 .!;l c l ! was more accurate, activating the rook before Black can play j_g5. Then 34 I! c8 is a threat, and after 33 "'/g7 34 � c7 + <1Jg6 35 � d7 White gets his rook behind Black's passed pawn and follows up with *e2-d3. However the move played should also be sufficient to win.

33 j}_g5 34 i£lc4 J..f4 35 h3 � g8

Black will have to play this sooner or later, for otherwise White plays *d3 and � gl , seiz­ing the open file for his own use.

36 b4 X g2 The rook must move up the g­

file to allow the Black king to cross, and at g2 it delays White's ltrd3.

37 a4 *g7 38 b5 axb5 39 axb5 *f6 40 b6 <1Je6 41 b7 H g8 42 � b1 El 118 ( 65)

The win still requires some work. White first secures the ad-

Pelikan Variation 75

vanced pawn and moves his king up. Black is paralysed and can only adopt a waiting strategy.

43 )!i b5 '1fl>d7 44 l£la5 wc7

White's only worry is that Black might try to liquidate to a rook and pawn ending by j_g5--d8xa5. This might prove awkward to win as White would be left with only f- and h-pawns. The immediate 44 j_g5 45 � xe5 J..d8 fails to 46 lii( d5 + we6 47 lii( xd8 .

45 '1id3 '1id6 46 h4

Definitely ruling out . . . j_ g5. 46 if!Jc7 47 l!i b2 j_ h2 48 we4

White's plan is to play §l c2 + at a moment when Black must reply \fld7 to prevent � c8. Then White will seize the g-file by � cl -g l . Finally the penetration of the rook combined with the advance of the White king to c4 and b5 will decide the game. The immediate 48 El c2 + fails to 48 . . . \flb6! 49 )ii c8 *xa5, so White

76 Pelikan Variation

must blockade the e-pawn before starting his plan.

48 J..f4 Black can only delay the end.

This move covers cl to prevent � c l -g l .

49 .!i\ c2 + *d7 Or 49 * b6 50 � c8 Jii,( xb7 5 1

'-£)xb7 *xb7 5 2 �f8 (threat )jj[ xf4) j_h2 53 � f6 and wins.

50 .d3! Black is in zugzwang. The rook

cannot move because of '-£)c6, the king cannot move because of � c8 and . . h5 doesn't help.

50 J..h2 51 la cl J..f4

The same again! Black has to free g l .

52 � gl *d6 53 *c4 Resigns

The finish might be 53 . . . � f8 54 '-£) b3 (if White gets the chance to play <El b3--c5 the winning pro­cess is simplified) la b8 55 IE\c5 l;ii, h8 56 � g6 + *c7 57 *b5 d3 58 � c6 + wb8 59 .b6 mating, or 53 . . . *c7 54 � g6 la h8 55 *c5 d3 56 � g7 + *b8 57 *b6 with a similar conclusion.

5 Dragon Variation

The Dragon i s one o f the most controversial lines in the Sicilian. At various times over the years it has appeared to be in its death throes, only to be suddenly revived by the discovery of new ideas for Black. Practitioners re­gard the variation as their private property, and defend it with an almost religious fervour against the many White players who wish to commit the heresy of mating Black down the h-file. The amount of opening analysis spawned by the Dragon is notori­ous, but much of this theory is concerned with the many White alternatives, so although I am recommending one of the main lines for White the quantity of analysis is not excessive. The material is divided amongst three games. I call the continuation I e4 c5 2 � f3 d6 3 d4 cxd4 4 .:£J xd4 � f6 5 f:lc3 g6 6 .ii_e3 .ii_g7 7 f3 0-0 8 �d2 ..£Jc6 9 ,.Rc4 ,.Rd7 10 0-0-0 � c8 1 1 ,.Rb3 � e5 1 2 h4 h5 1 3 jtg5 14 c5 the 'main line' Game 1 8 deals with all deviations from the main line before Black's 1 2th move, while game 1 9 covers Black's 1 2th and 1 3th move alter­natives. The 'main line' itself is deferred until game 20.

The Dragon is characterized by

the initial moves 1 e4 c5 2 �f3 d6 3 d4 cxd4 4 .£lxd4 � f6 5 f:lc3 g6, intending to develop Black's bish­op actively at g7. If White plays quietly and castles kingside the bishop will still be useful in sup­porting Black's minority attack on the queenside, while if White aims to attack the enemy king and plays 0-0-0 the Dragon bishop comes into its own. Hosts of White players have seen their queensides disintegrate under the laser-like power of the g7 bishop, supported perhaps by �a5 and g c8 to step up the pres­sure on c3. The true Dragon player will analyse six exchange sacrifices on c3 before breakfast, and White players need to be con­stantly on the alert for combi­nations based on blowing open the long diagonal.

The critical lines are those in which White tries to deliver mate by advancing his h-pawn, when Black has two possible defences. He may ignore the advance and concentrate on developing his own attack, or he may block White by playing h5. This lat­ter approach allows a possible sacrificial breakthrough by g4, but this has to be very well-timed or it just loses material. General

78 /)ragon Variation

pn nci ples aren't much help in the I >ragon, since success or failure is determined by tactical considera­t ions.

Game 18 Kroncke-Schroder

Correspondence, 1987

1 e4 c5 2 � f3 d6 3 d4 cxd4 4 � xd4 <E)f6 5 � c3 g6 6 J._e3

If White intends castling queen­side this move is the most usual. 6 J._e2 only fits in with 0-0, since the bishop is usually better placed at fl or c4 in the more aggressive lines resulting from castling on opposite wings.

6 J._g7 The Dragon differs from many

other Sicilian systems in that Black often omits the typical move a6. Time is of particular importance in the Dragon and Black simply cannot afford the tempo spent on preparing b5, which can often be played without . . . a6 in case White castles queen­side. The idea of playing a6 and b5 before castling has been tried, but after 6 a6 (6 . . . Q)g4? loses material after 7 J. b5 + ) 7 f3 ..f} bd7 8 'i'itd2 b5 9 a4! bxa4 (9 b4 1 0 ..f}d5 is also very good for White) 10 � xa4 J.g7 l l J._e2 0-0 1 2 0-0 Qlc5 1 3 � a3 J._b7 14 � fa ! �c8 15 Ql b3 White had strong queenside pressure in Kavalek-Bilek, Sousse IZ 1967.

7 f3 (66) This is more or less forced as 7

J._c4 and 7 'if(d2 can both be met by 7 ..f}g4.

66 B

Or: 7 0-0

(I) 7 . . . a6 (7 Q� bd7 exerts no pressure on d4 so simply 8 'if(d2 followed by 0-0-0, J.c4 and J._h6 gives White a strong attack) 8 �d2 Ql bd7 (8 b5 9 a4! is similar to Kavalek-Bilek above) 9 J.h6 (9 0-0-0 is also good) J._xh6 (9 0-0 is suicidal since White's h4-h5 attack is much stronger than usual) I 0 'if(xh6 b5 1 1 0-0-0J.b7 1 2 <fi b l , Bastrikov­Hasin, USSR 1 96 1 , and Black's king is stuck in the centre .

(2) 7 . . . Q)c6 (this can trans­pose to the main line if Black castles quickly, so we only explore lines in which Black plays .�.d7 and . . . )ji( c8 before . . . 0-0) 8 �d2 J._d7 9 0-0-0 lii( c8 10 g4 and now:

(2a) 10 . . . Q)e5 1 1 h4 h5 1 2 g5 Q�h7 1 3 f4 <E)g4 and now P. Lit­tlewood-Mestel, London 1 978, continued 14 ..§..g1 0-0 1 5 'fi b ] e5

1 6 Q)f3 �e6 1 7 fxe5 2£) xe5 1 8 Qlxe5 �xe5 19 .t) b5 � e8 and Black had no problems. However White can improve by playing f5 at some point, for example 1 5 f5 or, more provocatively, 14 f5!? since although White's dark­squared bishop disappears after 1 4 Qlxe3 1 5 'lifxe3 0-0, it isn't easy to see a constructive plan for Black after 1 6 �h3.

(2b) 10 . . . 0-0 1 1 h4 and now: (2bl ) 11 . . . b5 1 2 <2)cxb5 ..'2le5

1 3 h5 2lxf3 (once Black has started sacrificing he must con­tinue) 1 4 Qlxf3 �xg4 1 5 '/Wg2 ( 1 5 � e2 Ql xe4 1 6 t4r e 1 ! would be a safe answer to Black's aggression) '!ifa5 1 6 a3 � xc2 + ! 1 7 lf!'xc2 �a4 + 1 8 ltfd2 'l\il'b3 1 9 Qlc3 'lf1xb2 + 20 lf!'d3 �xa3 with hor­rendous complications, Mestei­Christiansen, Hastings 1 978-9. This remarkable game concluded 2 1 � c l (2 1 � c l was probably better) �b4 22 �d2 I:!l c8 23 hxg6 hxg6 (23 h5!? reserving h6 for the bishop was possible) 24 � h4 � xc3 + 25 �xc3 � xe4 + 26 ltrd2 .£)d5? (26 � xf3 27 � xe4 <2l xe4 + 28 lf!'e3 �xg2 29 �xg7 'l!; xg7 30 �xg2 Q)c5 was best and the strange ending resulting looks to be better for White) 27 �xg7 'ltte3 + 28 'ltrc2 �xf3 29 �b2! e5 30 �d2 'lttc5 + 3 1 \trbl �e3 32 � c l �b6 33 '*'h2 Qlf5 34 � c8 + ltfg7 3 5 � b4 Resigns.

(2b2) 11 . . . �a5 1 2 'l!lbl ( 1 2 h5 allows 1 2 <tJ b4 with � xc3 to follow) � fd8? ( 1 2 . �xd4 1 3 j_xd4 � fd8 was better according

Dragon Variation 79

to Hiibner, who assesses the resulting position as slightly better for White) 13 .£) b3 l'lfc7 14 h5 jle6 1 5 hxg6 fxg6, Hiibner­Hort, match 1 979 and now 1 6 Ql d5! would have given White a clear plus according to Hiibner.

(2b3) 11 . . . h5 1 2 �e2 (it is by no means certain that this is White's best. Ligterink-Sosonko, Dutch Ch. 1978 went 12 gxh5 Qlxh5 13 I:!l g l lf!'h7 14 �e2 Qlxd4 1 5 jj_xd4 �h6 1 6 �e3 �g7 1 7 �d4 JLh6 Draw, but 1 4 ltf b l avoiding the draw may be good for White) ..'2le5 1 3 gxh5 �xh5 1 4 � d g I ( 14 B\ hg I i s a n alternative, but whether better or worse is not easy to decide) �c4 1 5 � xc4 (a defect of 12 �e2 is that this cap­ture will involve a loss of a tempo) � xc4 1 6 l!l g5 (threatening 1 7 � xh5 gxh5 1 8 � g l ) � c5 1 7 ..'2ld5 e6 18 I:!l xh5 exd5 19 � xd5 � xd5 20 exd5 and White was a little better in Speelman-Liu Wenzhe, China 1 98 1 .

8 ·�Wd2 �c6 Or 8 d5 (other moves meet

with common-sense replies, e.g. 8 . . . jj_e6 9 <-2l xe6 fxe6 1 0 e5 or 8 . . . a6 9 0-0-0 b5 1 0 h4) 9 e5 Qle8 (9

.£) fd7 1 0 f4 Qlb6 1 1 �e2 Qlc6, Popovic-Sax, Vrsac 1 98 1 , and now 1 2 O-O-Ojj_d7 1 3 .£) b3 looks unpleasant for Black since 1 3 e6 1 4 h4 gives White a very dan­gerous attack) 10 f4 f6 1 1 0-0-0 fxe5 ( 1 1 Qlc6 12 �f3) 1 2 fxe5 <2)c6 ( 1 2 �xe5 1 3 lt:l f3 either regains the pawn with advantage or, after 1 3 . . �xc3 1 4 '*xc3 e6

80 Dragon Variation

1 5 h4, gives White an enormous attack) 1 3 �0 jtg4 ( 1 3 e6 1 4 jth6 leaves Black with a very bad bishop) 1 4 �xd5 !;'!. xO 1 5 gxO jtxO 1 6 jtg2 jtxd l 1 7 � xd l jtxe5 1 8 Ac5 e6 1 9 �e7 + f!Jg7 20 � xc6 �xd2 + 2 1 !;'!. xd2 bxc6 22 !;'!. d7 + with a good ending for White.

6 7 B

9 Ac4 (67)

9 jtd7 This is by far the most common

in practice, but a number of other moves have been tried:

( I ) 9 . . . a6 (too slow) 1 0 jtb3 �e5 ( 1 0 � a5 1 1 h4 b5 12 h5 �xb3 13 axb3 j_d7 1 4 J.h6 Axh6 1 5 �xh6 e6 1 6 0-0-0 b4 1 7 � a2 �a5 1 8 * b l is good for White) 1 1 h4 b5 1 2 h5 �xh5 1 3 Ah6 e6 1 4 jlxg7 *xg7 1 5 g4 � f6 1 6 0-0-0 with a dangerous attack, Yurtaev-Shur, USSR 1 976.

(2) 9 . . . a5 (this allows White to block the queenside) 1 0 J..b3 jtd7 1 1 a4 �xd4 1 2 1i_xd4 1i_e6 1 3 0-0-0 �d7 1 4 * b l !;'!. fc8 1 5 � he l a a6 1 6 �e2 � ac6 1 7 �d5 �e8 18 �b5 and White is clearly

better, Hardicsay-Honfi, Buda­pest 1 977.

(3) 9 . . . �xd4 (bringing White's bishop to d4 reduces the power of the Dragon bishop on g7) 10 jtxd4 jte6 1 1 jtb3 �a5 1 2 0-0-0 It fc8 ( 1 2 jtxb3 1 3 cxb3 � fd8 14 f!Jbl s d7 1 5 g4 � ad8 16 �e2 jth8 1 7 h4 and 12 b5 1 3 f!J b l b4 14 €ld5 J..xd5 1 5 exd5 �b5 1 6 �d3 �b7 1 7 Ji:K he l a5 1 8 J..a4 give White some advantage) 1 3 ltr b l E:i, c6 ( 1 3 b5 1 4 � he t _1xb3 1 5 cxb3 b4 1 6 1i_xf6 bxc3 1 7 J..xc3 _R xc3 1 8 bxc3 � xc3 1 9 ii;; e3 Jii( ac8 20 � xc3 �xc3 2 1 �xc3 Ji;i\ xc3 22 � c 1 is a fairly typical line in this variation; White's outside passed pawn gives him a big advantage) 1 4 h4 jtxb3 1 5 cxb3 b5 and now both 1 6 a3 and 16 h5 are good for White.

(4) 9 • • • '\'ita5 may be met by 1 0 J..b3, and after 1 0 1i_d7 1 1 0-0-0 we reach the main line of this game. Note that the tempting 10 �b3 �b4 1 1 J..e2 may not be good, e.g. 1 1 . . . �a5 1 2 a3!? �xb3 1 3 cxb3 '®'xb3 14 E e l jte6 1 5 ii.d l �c4 1 6 <2ld5 1i_ xd5 1 7 � xc4 jtxc4 1 8 1i_ xa7! <2lxe4! with a satisfactory position for Black.

(5) 9 . . . <E�d7 (the most import­ant 9th move alternative) 1 0 h4 ( 68) and now:

(5a) 10 . . . h5 (a new idea of Larsen's) 1 1 J..b3 (the flexible 1 1 0-0-0 may be more accurate with this move order) <E:lc5 ( 1 1 � b6 12 0-0-0 <2la5 1 3 �d3 J..d7 14 g4 hxg4 1 5 h5 � c8 1 6 hxg6 <2lxb3 + 1 7 axb3 fxg6 1 8 e5! dxe5 1 9 �xg6

68 B

exd4 20 l;!l h7 l;!l t7 2 1 � dh l wins) 1 2 0-0-0 .£lxb3 + 1 3 axb3 �a5 14 'iti>bl jld7 15 g4 hxg4 16 h5 gxh5?? (a blunder; Sherzer suggests 1 6 .:2)xd4 1 7 jlxd4 jlxd4 1 8 �xd4 g5, but the finesse 1 7 .£ld5! '14j'd8 1 8 J,. xd4 appears good for White) 1 7 .£l xc6 bxc6 1 8 .£l d 5 �d8 1 9 J,.h6 f6 20 � xh5 cxd5 2 1 §txg7 *f7 22 l\1 xf8 yj,fxf8 23 �xd5 + Resigns, Sherzer-Lar­sen, New York Open 1 988.

(5b) 10 • . • �b6 1 1 J,.b3 .£la5 (or 1 1 .£je5 1 2 �e2) 12 ·�d3 J,.d7 1 3 h5 � c8 14 hxg6 hxg6 1 5 0-0-0 Ql bc4 and White has two promising lines:

(5b l ) 16 J,.h6 e5 ( 1 6 J,.xh6? 1 7 l;!l xh6 e5 18 ii:ll dh l '1Wg5 + 1 9 \fr b l exd4 20 .£ld5 f5 21 �xd4 �xb3 22 axb3 �d2 + 23 �xd2 ytxd2 24 l;ll, h8 + 'iti>f7 25 g l h7 + *e8 26 l;!t e7 + mates) 1 7 !JJ.. xg7 '1Jxg7 1 8 .£lde2 �g5 + 19 'iti b l b5 ( 1 9 )'!ltxg2? 20 J,.xc4 �xc4 2 1 b3 i s very good for White) 20 g3 b4? (20 � h8 21 � xh8 � xh8 22 a4! a6 23 J,.xc4 �xc4 24 b3 � a3 + 25 'ltlb2 b4 26 .£ld5 is Black's best chance, but White

Dragon Variation 81

still has some advantage) 2 1 .£ld5 J,.b5 22 f4! 'l!o/d8 23 ttf3 J,.c6 24 f5 f6 25 fxg6 j}_xd5 26 � h7 + 'ii'g8 27 'l!o/h5 Resigns, Fleck­Kozul, Cesme 1 983.

(5b2) 16 !JJ..g5 .£lxb3 + 17 cxb3 l#a5 1 8 jlxe7 �e5 1 9 'i'!\"c2! and Black has a depressing choice:

(5b2 1 ) 19 . . . � xc3 20 bxc3 l;!i c8 21 'ii' b2 �b6 22 J,.g5 �c4 + 23 'ii'a l �a3 24 �b2 � b5 25 .£jxb5 'l!o/xb5 26 .§te3! with a clear plus for White, Wedberg­Sosonko, Haninge 1 988.

(5b22) 19 . . . b5? 20 1ti' b 1 � fe8 2 1 <2!d5! )4 xc2 22 � f6 + J..xf6 23 .§txf6 wins.

(5b23) 19 . . . <21c6 20 .£lxc6 J;g xc6 21 '\ffb l � fc8 22 ii, d5! �b6 23 'i!Wd2 � xc3 24 bxc3 g xc3 25 � xd6 � xb3 + 26 axb3 ttxb3 + 27 lfi'c l •a3 + 28 'll'd l �al + 29 life ) "fio\'a4 + 30 fj>e2 J,.b5 + 3 1 '\\'f2 .§td4 + 3 2 '1flg3 J..e5 + 3 3 f4 and wins.

(5b24) 19 . . . � fe8 20 J,.xd6 b5 21 b4 tta6 22 � b3 ..te6 23 .£l c5 � xc5 24 §txc5 J,.xa2? 25 l;l: d6 Resigns, Kruppa-Savchenko, USSR 1 984.

10 0-0-0 (69) There are arguments in favour

of the alternative move order with 1 0 h4, but we shall adopt the most usual sequence in which White castles first.

10 �a5 When the Dragon was revived

in the 1 960s, this was the most popular line for Black. It frees the f8 rook to come over to c8 so that both Black rooks may participate

H2 DraKon Variation

in the attack. It is unpopular at the moment both because of very bad practical results for Black and because of the depressing theoretical situation detailed be­low. It becomes clear that the time Black spends moving his queen to a5 only serves to present White with an important tempo by � b3 in the main line. 10 . . . l;ll c8 1 1 Ji_b3 �e5 is covered in games 1 9 and 20, and we deal with the remaining lines now:

( I ) 10 . . . 'liifc7 1 1 �b3 M fc8 ( 1 1 �e5 1 2 A.h6) 1 2 h4 �e5 13 h5 �c4 1 4 A.xc4 't¥1xc4 1 5 A.h6 .:&h8 1 6 A_g5 b5 1 7 hxg6 fxg6 1 8 j).xf6 A.xf6 1 9 -t)d5 .�xd4 ( 1 9 �xa2 20 'liif h6! gives White a crushing attack) 20 �xe7 + <l!f7 (20 '\ti' f8 21 �xc8 .l!l( xc8 22 c3 is very good for White) 2 1 � xh7 + A_g7 22 c3 � h8 (White plays �f4 + after most moves) 23 'i!irf4 + A. f5 24 �xf5 .l!l( xh7 25 �xd6 + \frg8 26 �xc4 A.h6 27 �xh6 � xh6 28 �e3 and with three solid pawns for the exchange White has good winning chances in the ending.

(2) 10 . . . 'jjl(b8 1 1 h4 and now:

(2a) 11 . . . b5 ( 1 1 a5? 1 2 A.h6) 12 A.d5! Hc8 1 3 \fr b 1 ( 1 3 A.xc6 _txc6 1 4 h5 b4 1 5 �ce2 �xh5 1 6 !ll xh5 gxh5 17 �f5 '!J, c7 1 8 A.h6 was also very good for White in Wahls-Ristic, Dort­mund 11 1 989) b4 14 �ce2 � xd5 15 <E)xc6 g xc6 ( 1 5 <E)c3 + loses to 1 6 � xc3 bxc3 1 7 � xb8 )4 axb8 1 8 'lf(d5 � xb2 + 1 9 \fr a 1 �e6 20 �a5) 1 6 exd5 )!! a6 1 7 A.d4 with a n excellent position for White, Mikhalchishin-Sirov, USSR 1 988.

(2b) 11 . . . Jiii c8 12 � b3 a5 ( 1 2 b5 1 3 h5) and White has two

tempting lines: (2bl ) 13 a4 � xd4 14 A.xd4 b5

1 5 <E)d5 e5 16 jtxe5 bxa4 (after 1 6 dxe5 1 7 �xf6 + A.xf6 1 8

�xd7 l;ll a7 1 9 �d6 the only ques­tion is whether the extra pawn is enough to win) 1 7 A.xf6! axb3 ( 1 7

'!ilfxb3 1 8 �e7 + \fr f8 1 9 � xc8 � xc8 20 A.xg7 + <1Jxg7 2 1 �d3 wins) 1 8 c3! g c4 (there is nothing better as White already has a deci­sive attack) 1 9 � xg7 <1Jxg7 20 h5 .l!l( a4 2 1 \frbl �b5 22 �g5! Resigns, D' Adamo-Teipelke, corr. 1 986.

(2b2) 13 h5 a4 ( 1 3 �xd4 1 4 A.xd4 a4 1 5 A. d 5 e 5 1 6 hxg6! is crushing) 14 A.d5 .:2:J xh5 (or 14 . . e6 1 5 hxg6! exd5 1 6 A. h6! fxg6 1 7 A.xg7 f:jxg7 1 8 �h6 + <l! f7 1 9 <E)xd5 with a very strong attack) 1 5 g4 � f6 1 6 � f5 A. xf5 1 7 gxf5 �b4 1 8 fxg6! � fxd5 1 9 A_d4! (a very nice move) A. xd4 (the lines 19 . . . fxg6 20 exd5, 19 . . . <E)f6 20 gxh7 + fJh8 2 1 .l!lt dg 1 and 19 . . .

e5 20 gxh7 + *h8 2 1 Iil dg l f6 22 exd5! exd4 23 � xg7 'l!/ xg7 24 �h6 + 'I!Jf7 25 � g l all win for White) 20 �xd4 e5 2 1 gxf7 + *fll (21 *xf7 22 liii, xh7 + <8ig8 23 �gl + *xh7 24 �g2 wins) 22 �gl -EJf6 23 � h6! -EJxa2 + 24 -EJxa2 �c7 25 -E:lb4 wrxf7 26 � xd6 -EJe8 27 .i;;!. d3 and although material is equal, Black's king is horribly exposed, Agdestein­Karlsson, Gausdal 1987.

1 l j}_b3 Iil fc8 12 *b1 -EJe5

1 2 � ab8 (a bit slow, so White can afford to spend time preparing h4 with g4, which cuts out the reply h5) 1 3 g4 b5 14 h4 b4 1 5 -EJd5 -EJxd5 16 exd5 cZ)xd4 1 7 j}_xd4 j}_xd4 1 8 "tllfxd4 �c5 1 9 �d2 a5 20 h5 a4 2 1 hxg6 axb3 22 gxf7 + * ffi 23 "i!l.f h6 + *xf7 24 �h5 + * f6 25 g5 + Resigns, Schmidt-Bobotsov, Varna 1 964.

13 h4 .f)c4 After l 3 b5 1 4 -E:lcxb5 �xd2

1 5 Axd2 it is hard to believe that Black can have enough for the pawn. Kuzmin-Garcia Martinez, Hastings 1 973/4 continued 1 5 f)c4 1 6 Q)c3 Q)xe4 1 7 .f)xe4 j}_xd4 1 8 Ag5 Ae5 1 9 Axe7 -EJ xb2 20 � xd6! j}_e8 2 1 � d5 with a clear plus for White.

14 Axc4 Iil xc4 15 QJb3 ( 70) "*c7

This appears to be the best square:

( 1 ) 15 . . . �e5 (a recent idea, but there is an obvious danger of the queen being trapped) 16 Ad4

Dragon Variation 83

'i!iite6 1 7 g4 ( 1 7 h5 a5 1 8 h6 Ah8 1 9 Q} d 5 a4 20 j}_xf6 j}_xf6 2 1 cZ)xf6 + 'il!i'xf6 22 QJd4 !;!t ac8 23 � he t a3 was unclear in Hebden-Jonsson, London 1 988) � ac8 1 8 lit he 1 Ac6 1 9 Q)d5 Aa4 20 c3 b5 2 1 'lif'h2 and White has the advan­tage, Koch-Ristic, Dortmund 11 1 989.

(2) 15 . . . �a6 (dubious) 1 6 e5 Q)e8 ( 1 6 dxe5 17 Q)c5 "*'d6 1 8 "*'e2!) 1 7 h 5 ( 1 7 �d5 is also strong) � ac8 ( 1 7 J.. xe5 1 8 hxg6 followed by Ad4 i s good for White) 1 8 hxg6 fxg6 (or 1 8 hxg6 1 9 Ah6 AxeS 20 .!21d5 with a strong attack) 1 9 Ad4 with a distinct plus for White.

(3) 15 . . . "*'d8 (also poor) 1 6 Ah6 J..h8 (perhaps 1 6 � ffi is relatively best, but after 1 7 j}_xg7 "*'xg7 1 8 g4 Black has no real counterplay) 1 7 h5 .i, ac8 1 8 hxg6 fxg6 1 9 e5 .!£1e8 20 exd6 e6 2 1 Ag5 and Black is in a mess, Ciocaltea­Drimer, Romania 1 968.

16 Ad4 One should also take note of a

promising alternative for White here, namely 1 6 g4 and now 1 6 . . .

84 Dragon Variation

� e6 1 7 h5 a5 ( 1 7 � c8 1 8 <21 b5 �d8 19 <2! 5d4 � d7 20 hxg6 fxg6 21 e5 <2! e8 22 *h2 was very good for White in Bangiev-Ovhinikov, USSR 1 976) 1 8 hxg6 fxg6 1 9 �d4 a4 20 �xf6 exf6 (20 . . . �xf6 2 1 .'£)d5 �xd5 22 �xd5 + <l/g7 23 !i, xh7 + wins) 21 <2! d5 �d8 22 <2!d4 J..f7 23 � h2 h6 24 <2!e3 left Black with a miserable position in Gonzales-Schroder, corr. 1 987.

16 J..e6 Or 1 6 . . . .&_c6 1 7 h5 a5 1 8 hxg6

hxg6 1 9 a4 e6 (Black wants to take on a4 without allowing J..xf6 followed by .'£) d5, but this is too slow) 20 M h4 �xa4 2 1 !i dh l Jl xd4 22 <2! xd4 J..e8 23 �db5 with a large advantage for White, W ang Zili-Velimirovic, Thessalo­niki 01. 1 988.

17 h5 a5 18 hxg6 hxg6 19 a4 g b4

After 1 9 � c8 20 g4 )j, xd4 2 1 <2!xd4 'llf b6 Black had some counterplay in Hartston-Hollis, London 1 967, but 20 g h4! looks very unpleasant for Black.

71 B

20 H h4 ( 71) b5

Or 20 !i, c8 (20 �xb3 2 1 cxb3 e5 22 J..e3 H xb3 23 H c l *d7 24 �h6 �h8 25 ll.g5 ll.g7 26 l:![ ch l l:![ xc3 27 �xc3 was ex­cellent for White, Hartston-Kol­bek, Dresden 1969) 2 1 g dh I �xb3 22 cxb3 e5 23 J..e3 .i. xb3 24 � b5 �d8 25 J..h6 �f8 26 J..xf8 <l/xf8 27 � h8 + <2!g8 28 i'i\ xg8 + Resigns, Savic-Simic, corr. 1 979.

21 <2! xb5 �b7 The theoretical main line con­

tinues 21 �c4 22 <2!c3 � ab8 and although this is good for White, it turns out that 22 M dh I ! is even better. Black cannot reply 22 M xa4 because of 23 � xd6 exd6 24 '*'g5 (the main threat is 25 le! h8 + �xh8 26 � xh8 + , but both 25 �xf6 and 25 �xf6 are lesser threats) �h5 25 �xg7 frixg7 (25 )g( c8 26 �c3) 26 ii;\ xh5 with a crushing attack. Therefore Black has nothing better than 22 � c8, but then White has a pleasant choice. He may play 23 <2!c3, effectively gain­ing a tempo over an already pro­mising line, since Black needs to double rooks on the b-file to create any threats. Alternatively he may continue 23 g4 followed by 'l!i\'h2 with an enormous attack.

22 � dh1 The position of the queen on b7

is no better; White's attack on the h-file is just too strong.

22 j( xa4 23 <2!c3

Not 23 <21 xd6 exd6 24 '*g5 because of 24 . . li;'!, xd4.

23 J_xb3 Or 23 .!;1 b4 24 g4 J..xb3 25

'f!fh2 J_xc2 + (25 J_a2 + 26 .£)xa2) 26 *al .£)h5 27 jJ_xg7 'f;;xg7 (27 lll: xb2 28 lll: xh5!) 28 M. xh5! gxh5 29 'f!fxh5 f6 30 'f!fh7 + \t'f8 31 "Wtg6 mating.

24 cxb3 lll: b4 After 24 � xd4 25 'f!fxd4

'f!fxb3 26 .£)d5 e5 27 'f!fc3 White has a clear material and positional advantage.

25 .£)d5 e5 26 lll: h8 + ! Resigns

Because of 26 . . . J_xh8 27 l!!f. xh8 + f'ixh8 28 'f!fh6 + .£J h7 29 !El f6 mating.

Or:

Game l9 Nunn-Mestel

London (GLC) 1986

1 e4 c5 2 .£)f3 d6 3 d4 cxd4 4 .£) xd4 L2)f6 5 .£)c3 g6 6 J..e3 J..g7 7 f3 0--0 8 'f!fd2 L2)c6 9 J..c4 j_d7

10 0-0-0 l!!f, c8 1 1 J.. b3 .£) e5 12 h4 ( 72) h5

( I ) 12 . . . 'f!fa5 1 3 'f;; b 1 �c4 1 4 J..xc4 l!!f. xc4 (the only difference compared to game 1 8 is that Black's rook is on f8 instead of a8, but this difference is crucial) 1 5 h5! ( 1 5 LEJ b3, as i n game 1 8, is still promising, but this may be even better) � xc3 (when the rook is on

72 B

Dragon Variation 85

a8 this move is good) 1 6 h6! (White saves his h-pawn because 1 6 . . . � c5 1 7 hxg7 hits the rook, so Black must play 1 7 lii: fc8 transposing to the game) l!!f, fc8 ( 1 6 . J.. h8 1 7 "ilf xc3 is very good for White) 1 7 hxg7 .!! 3c5 1 8 c3 (it seems to me that 1 8 b4 �b6 19 -E:J b3 i(4 xc2 20 jj_xb6 .!! xd2 2 1 .£)xd2 axb6 22 �c l i s very good for White) 'f!fa4 1 9 J..g5 l! xg5 20 "i!i!fxg5 with advantage to White, Shabalov-Yurtaev, Moscow GMA 1 989.

(2) 12 . . . a5 13 h5 ( 1 3 a4 is also good) a4 ( 1 3 .£) xh5 14 g4 � xc3 1 5 bxc3 .£)f6 1 6 J..h6 a4 1 7 jJ_xg7 f'ixg7 1 8 J..xa4! J..xa4 1 9 � f5 + Resigns, Draskovic-Lazic, corr. 1 974) 1 4 � xa4 J_xa4 1 5 jJ_xa4 .£)c4 1 6 tf"d3 (Black suffers from the lack of a white-squared bishop) 'f!fa5 1 7 J..b3 d5 1 8 hxg6 hxg6 1 9 g4! dxe4 20 fxe4 .£) xg4 2 1 � g 1 'f!fg5 + 2 2 'f;; b 1 �ge5 23 �h3 � fd8 24 J..f2 �d2+ 25 �xd2 'f!fxd2 26 J..e3 'f!fa5 27 !Ele6 fxe6 28 'f!fxe6 + Resigns, Fatali­bekova-Akhsharumova, Moscow 1 975.

86 DraKon Variation

(3) 1 2 . . . Qj c4 (the most im­portant 12th move alternative) 1 3 11.. xc4 !!! xc4 1 4 h5 and now:

(3a) 14 . . . b5? 1 5 hxg6 fxg6 1 6 e5 dxe5 1 7 Qje6 i s bad and 1 4 . . . )il(a5 1 5 * b l ( 1 5 Qjb3 is also good) transposes to line I above.

(3b) 14 . . . 'fi!/c7 1 5 hxg6 fxg6 1 6 Qjde2 !!! c8 1 7 * b 1 11_ e6 1 8 Qj f4 'fi!/d7 1 9 Qjxe6 t¥xe6 20 !!! c l Qjd7 2 1 'fi!/d5 'fi!/xd5 22 Qjxd5 � e8 23 b3 !!! c6 24 c4 and White's superior pawn structure gives him a permanent advantage in �he ending.

(3c) 14 . . . .:Z)xh5 (the critical move) 1 5 g4 Qj f6 1 6 Qj b3!? (73) (rarely played, but promising) and now:

73 B

(3c l ) 16 . . . a5 1 7 11_h6 J..xh6 1 8 � xh6 (with the threat o f � dh I followed by !!l xh7 and '!Wh6) fflg7 (intending .1i h8) 1 9 e5! dxe5 ( 1 9 Qjxg4 20 fxg4 11_xg4 2 1 !!! dh l i s very good for White) 20 g5 .:Z)g8 2 1 'fi!/h2 !!! e8 22 � xh7 + * f8 23 'fi!/xe5 f6 24 gxf6 and White wins.

(3c2) 16 . . . !!! xc3 1 7 bxc3 "Wtc7 1 8 e5 Qje8 19 �h2 h5 20 gxh5

11_xe5 2 1 f4! 'l!ifxc3 (2 1 j_xc3 22 hxg6) 22 11_d4 J_xd4 23 � xd4 with a large advantage for White.

(3c3) 16 . . . � e8 1 7 e5 Qjxg4 1 8 fxg4 J..xg4 1 9 � dg l and now:

(3c3 1 ) 19 . . . h5 20 e6! 11_xe6 (20 f5 21 Qjd5 'l!ltc8 22 Qjd4 and

now 22 * h7? 23 !!! xg4! fxg4 24 'l!ltd3 won for White in Ye Jiangchuan-Velimirovic, Asia-Yugoslavia 1 984) 2 1 11_d4 (Gufeld's suggestion of 2 1 !!l xh5! is a good alternative, the point being that 21 gxh5 loses to 22 11_h6 li! g4 23 li!t xg4 followed by 'l!ltg5) j_g4 22 !!! xg4 hxg4 23 J..xg7 ffjxg7 24 '*' h6 + *f6 25 Qjd5 + ffe6 26 li! e 1 + ffd7 (26

*xd5 27 'i!ll'h l + is attractive) 27 'fi!/g7 !!! f8 (27 . . . *c8 28 'fi!/xf7 *b8, Caturian-Poletaev, corr. 1 983, and now 29 Qj f6 !! h8 30 'fi!/xc4 exf6 3 1 'itxg4 gives White a winning position) 28 .:Z)d2 � c5 29 ll!( xe7 + '*'xe7 30 Qj xe7 *xe7 3 1 Qje4 � f5 32 Qjxd6 with advan­tage to White because of Black's exposed king.

(3c32) 19 dxe5 and now there are two possible moves. My suggestion was 20 '*'xd8 � xd8 21 .:Z:Jd2 (or 2 1 .:Z) a5!?, aiming to keep the knight in a more active posi­tion) � xc3 (21 � b4 22 a3) 22 bxc3, when it seems to me that Black's four pawns are too far back to provide enough compen­sation for the rook. However, in the 1 988 correspondence game Kauranen-Nesis White preferred to play for the attack by 20 '*'g2!?, threatening 2 1 Qjd2 and 2 1 '*'xb7.

This policy was a success after 20 . . . "i!rc8?! (this must be bad; Black should prevent j_h6 and secure the g4 bishop by 20 h5, even if this means sacrificing the queen­side pawns) 2 1 j_h6 jj_ h8 22 �d2 � xc3 23 bxc3 jj_d7 24 �g3 (White has an extra rook and an attack) jj_c6 25 � h2 'ii!ff5 26 �h4 '(il(f6 27 jj_g5 1i/g7 (very thematic for the Dragon; pity about the pawn on e5) 28 j_xe7 h5 (28 . . . e4 29 jj_f6) 29 � h3 a5 30 J..c5 b5 3 1 � e4 b4 32 -2)d6 � b8 3 3 � f5 Resigns.

(3c4) 16 . . . 'ific7 and now there are two promising lines:

(3c4 1 ) 17 e5 dxe5 ( 17 � xg4 1 8 fxg4 ]txg4 19 exd6 exd6 20 ,2)d5 wins) 1 8 g5 ]tf5 1 9 gxf6 exf6 20 j_ h6 g5 2 1 li._xg7 rk;xg7 is recommended by Sapi and Schneider, but 22 Wre3! followed by :t)d2 and 2Jde4 appears good for White. Combined with a later � d2 and :t)d5 Black will be hard pressed to defend his seriously weak pawn at f6. If Black exchanges on e4 he will be mated down the h-file.

(3c42) 17 g5 (normally a bad move, but this position is a special case against White gains a lot of time) ..£)h5 1 8 :t)d5 'ii!fd8 19 J...d4 jj_xd4 ( 1 9 e5 20 jj_xa7) 20 -E}xd4 jj_c6 2 1 b3 j_xd5 (2 1 l;i, c5 22 b4 )ij( c4 23 €le3 is also good for White) 22 bxc4 .;lxc4 23 -2) [5 f6 24 li xh5! gxh5 25 )ij( g ] e6 26 gxf6 + 'ltrh8 27 ..£)e7 1-0, Thes­ing-Ballmann, Zug 1 989.

13 J..g5 (74)

74 B

Dragon Variation 87

13 . . . �c4 1 3 !1 c5 is covered in game

20. Other 1 3th moves: ( I ) 13 . . . a6 (ignoring the

threat) 14 g4 hxg4 1 5 h5 � xh5 1 6 � d5 ,g e8 1 7 ,g xh5 gxh5 1 8 'l!!th2 (this gives White a crushing attack just as in game 20 below) gxf3 1 9 'i!il'xh5 jj_g4 20 '/!!th4 ,g c7 2 1 Ji h l �g6 22 'l!!th7 + '1t1 f8 23 jj_h6 Resigns, Hechler-Masur, corr. 1 980.

(2) 13 . . . �h7 14 J..h6 jj_xh6 1 5 'i'l!Jxh6 � xc3 1 6 bxc3 and now:

(2a) 16 . . . "�ta5 1 7 'fi b 1 ( 1 7 �e2 J.. b5 1 8 � he 1 <2Jf6 1 9 H d4 .i, c8 20 K b4! e6 2 1 �d4 jj_d7 22 'l!!te3 b5 23 '1tib2 � c5 24 �e2 �c6 25 )'tf4 �e8 26 a3 was also good for White in Tal-Mista, Dubna 1 973) )'txc3 ( 1 7 K c8 1 8 f4 <2Jc4 1 9 f5 gives White the advan­tage) 1 8 � e2 'l!!tc5 1 9 g4 <2)xf3 ( 1 9 . . . hxg4 20 f4 'l!!te3 2 1 h5 wins) 20 K d5 ! � f2 2 1 gxh5 g5 (21 '/!!txe2 22 hxg6 <21f6 23 g7 wins) 22 hxg5 'l!!te3 23 l!l hd l ! and White won easily, Geller-Kuzmin, USSR Ch. 1 978.

(2b) 16 . . . 'l!!tc7 1 7 * b 1 ! �c4 ( 1 7 '/!!txc3 leads to line 1) 18 g4

88 Dragon Variation

hxg4 ( 1 8 � f6 1 9 � f5!) 1 9 f4 ( 1 9 h5? g5) I!l c8 20 � d3 'i!ta5 2 1 h5 g5 and now 22 e5!? was good for White in Lobron-Miles, Biel 1 986 after 22 dxe5 23 jl_ xc4 � xc4 24 � b3 'i!tc7 25 fxg5 �f5 26 g6, but according to Lobron it was simpler to play 22 �xc4! li xc4 23 � b3 'i!td8 24 fxg5 "i!ffB 25 e5! "i!fxh6 (25 �f5 26 exd6 �xd3 27 �xf8 + � xf8 28 dxe7 wins) 26 gxh6 �f5 27 exd6 exd6 28 M xd6 when White's material advantage and threats against the Black king give him a winning position.

14 'i!te2 After 1 4 ttd3 Black can reply

1 4 . . . �e5, since 1 5 'i/lj"e2 would be an invitation for Black to sacri­fice at c3.

14 �a5 15 '1i'b1

For the moment Black has no threats because M xc3 isn't dangerous when the knight pre­vents the queen moving to a5, so White takes the opportunity to make a useful consolidating move.

15 a6 This move represents the point

of the manoeuvre �c4-a5. Black threatens to trap the knight on d4 by . . . e5 and if White were now forced to spend a tempo moving the queen Black would play b5 with good counter­play. Everything depends on whether White can successfully ig-nore the threat of e5.

,

16 g4 (75)

16 e5 Or 1 6 �xb3 1 7 axb3 hxg4

(for 1 7 . . . e5, see the note to Black's 1 8th move) 1 8 h5! gxf3 ( 1 8

�xh5 1 9 fxg4 12Jg3 20 *h2 �xhl 21 � xh l leads to mate) 1 9 � xf3 �g4 ( 1 9 . . . 12!xh5 20 � dg l threatening M xh5 i s very strong) 20 ll dg l (20 h6 � h8 2 1 e5 is also good, but not 2 1 h7 + �xh7 when Black can block the h-file by

� h5) gxh5 (20 . . . �xh5 2 1 �xf6 exf6 22 � xh5 gxh5 23 "i!fg2 wins) 21 �xf6 exf6 and now 22 K xg4 hxg4 23 � d4 is one danger­ous attacking possibility.

17 gxh5! Geller-Miles, Linares 1 983,

continued 1 7 � f5 gxf5 1 8 gxf5 �xb3 1 9 axb3 jl_c6 20 � hg l *h7 2 1 1i! g2 � g8 and White went on to win, but Mestel was evidently not convinced about White's compensation, and in this I agree with him.

17 exd4 18 �d5! ( 76)

Geller's notes to his game gave 1 8 h6 � h8 (18 . . . dxc3 1 9 hxg7 flxg7 20 )!;g xd6 wins) 1 9 h7 + fJxh7 20 h5 and wins, but 18 . . .

�xh6! 19 �xh6 � xb3 20 axb3 dxc3 is a big improvement, when the position is totally unclear. Later I discovered that Timman had suggested 18 � d5! in 1 983, but during the game I was un­aware of this!

18 �xb3 The only real defence to the

threat of 19 h6 � h8 20 h7 + *xh7 2 1 h5 is to play gxh5 . Unfortunately 1 8 gxh5 (trying to deflect White's rook by 1 8 d 3 1 9 H xd3 gxh5 fails after 20 � g 1 fi>h7 2 1 �xf6 + �xf6 22 � xd6 � xg5 23 � xg5 winning easily) loses to 19 � hg1 fi>h8 20 e5 dxe5 2 1 �xe5 �xd5 (or 21 � e8 22 � xf6 �xf6 23 �xf6 � xf6 24 � xf6 JJ. e7 25 � xd4 with a won ending) 22 �xg7 + f!Jxg7 23 j_xd8 + f!J h7 24 jlxa5 with a strong attack and a bonus pawn to come. Thus Black tries to exchange on b3 before taking on h5, but in this position White doesn't need to recapture. If Black wanted to exchange on b3 he should have done so at move 1 6, with the continuation 16 . . .

Dragon Variation 89

�xb3 1 7 axb3 e5 1 8 gxh5 exd4 1 9 �d5. I think White can still win even in this case, but it is much more complex: 19 gxh5 20 � dg l lt1h7 (20 f!Jh8 2 1 f4 �g4 22 �xg4 hxg4 23 h5! lt1h7 24 h6 j_ h8 25 �d3 f!Jg8 26 h7 + wins) 2 1 f4 �g4 22 �d3! f!Jg8 (22

f!J h8 23 � xg4 hxg4 24 h5 is worse) 23 � xg4 hxg4 24 h5 (this seems to be good even when White has a tempo less) �xd5 (24

K e8 25 h6 �h8 26 h7 + lt1f8 27 � xf6 wins) 25 �xd8 Q� b4 (25

� xf4 26 '(iffl ) 26 '*'d2 d3 27 cxd3 � c2 28 W!xb4 � xb2 + 29 f!Jc I � xd8 30 d4 and White should win.

19 h6! Stronger than transposing to

the above analysis by 1 9 axb3. 19 j}_b5

Trying to generate counterplay, but there is no defence. The main line runs 1 9 �h8 ( 1 9 . . . � xd5 20 hxg7) 20 h7 + f!J xh7 21 h5 f!Jg8 (2 1 �g7 22 hxg6 + lt1g8 23 � xf6 + �xf6 24 '/j'h2) 22 � h2! (not 22 hxg6 fxg6 23 �h2 f!Jf7) gxh5 (22 � g4 23 fxg4 ytxg5 24 hxg6 mates) 23 LE)xf6 + J..xf6 24 �xh5 li( e8 25 � dg l f!Jf8 2 6 � h6 + j_g7 2 7 �xg7 + f!Jxg7 28 jle7 + and mate next move.

20 �h2 d3 21 cxb3

Not 2 1 c3? d2! with dangerous counterthreats.

21 LE)xd5 After 2 1 � c2 22 hxg7 � xh2

23 gxf8( �) + f!J xf8 24 g xh2

90 Dragon Variation

Black loses the knight on f6. 22 hxg7

Not 22 jj_ xd8? jj_e5. 22 .!! c2

Black cannot avoid a fatal loss of material.

23 jJ_ xd8 K xd8 24 � d2 Resigns

24 .!El e3 is met by 25 �f2.

Game 20 Mestel-Kudrin

Hastings 1986/7

I e4 c5 2 .!Elf3 d6 3 d4 cxd4 4 'El xd4 'El f6 5 .!Elc3 g6 6 jJ_ e3 jJ_ g7 7 f3 0-0 8 �d2 .!Elc6 9 jJ_ c4 jJ_ d7

10 0-0-0 )iil. c8 1 1 jJ_ b3 'El e5 12 h4 h5 13 J..g5 � c5 14 \flbl

This is not only a useful move, it leaves Black with little choice but to play b5, when curiously enough White's kingside break­through becomes stronger because in one variation Black does not have the move �b6 attacking b2.

14 b5 15 g4 (77) 15 aS (78)

Declining the sacrifice, but acceptance is a major alternative for Black: 1 5 hxg4 1 6 h5 and now:

( 1 ) 16 . . . � xc3 with a further branch:

( l a) 17 'l\lltxc3 l2\xh5 1 8 fxg4 jtxg4 19 � dgl 'i'!fd7 is unclear, while 17 jj_ xf6 � xb3 1 8 JJ_xg7 )iil. xb2 + wins for Black.

( 1 b) 17 bxc3 and Black has one bad move and one good move:

(1 b l ) 17 . . . .!Elxh5 and now the simple 18 jJ_h6! (threat 1 9 .!:il, xh5 gxh5 20 '!lit g5) e6 19 � dg 1 gives White a strong attack for no material investment. In Ulybin­Tiviakov, USSR 1 987 White chose a more complex line by 18 .!:i!, xh5 gxh5 1 9 �h2 .!Elc4 (other­wise White plays �xh5 and .!:i!, h l ) 20 'l\lltxh5 f6 2 1 l2\xb5! ll_xb5 (White also has the advantage after 21 fxg5 22 JJ_xc4 + e6 23 l2\xd6) 22 j_ h6 'i'!fd7 (22 �c8 23 � g l f5 24 ll_xg7 \fjxg7 25 exf5 gives White more than enough for the piece according to Ulybin) 23 jj_xg7 \flxg7 24 liil. gl with a crush­ing attack.

( l b2) 17 . . . l2\ xf3! 1 8 l2\ xf1 l2\ xe4 1 9 � h2 l2\xc3 + 20 'fl>cl �a5 2 1 hxg6 (21 � d4 g3 ! IS

unclear) l2\ xa2 + 22 J_xa2 �a3 + 23 'lt>d2 �c3 + 24 *c l (24 'fl>e2

ttxf3 + 25 *e l ? tye4 + and 26 ttxg6 gives Black 5 pawns and

attacking chances for a rook) tya3 + Draw, Lanka-Smirin, USSR 1 989.

( l e) 17 h6! (although never tried in practice I think this is a good move) and now:

( l c l ) 17 . . . .l, xb3 1 8 hxg7 X xb2 + 1 9 '*a l ! (threat 20 j}_ xf6 exf6 2 1 ii h8 + ) •xg7 ( 1 9 . . . <2:l h5 20 J..f6!) 20 j_xf6 + f1xf6 2 1 ttf4 + .g7 22 tyh6 + .f6 23 fxg4! and White has a massive attack.

( l c2) 17 . . . <2:lxf3 I 8 � xf3 X xf3 1 9 hxg7 *xg7 20 jth6 + (20 e5!? is also possible, with the idea that after 20 . . . dxe5 2 1 jth6 + .g8 22 jtg7 il,l h3 23 K xh3 gxh3 24 i!lth6 <2:l h7 25 jt xe5 Black is mated) f1g8 2I jtg7! K h3 (2I <2:1 h5 22 �h6) 22 l\1 xh3 gxh3 23 'Wf h6 <2:1 h 7 24 j_ c3 e5 25 �xg6 + '*h8 26 i!lfxd6 with advantage to White.

(2) 16 . . . <2:lxh5 I 7 <2:ld5 X e8 ( 1 7 l\1 xd5 1 8 jtxd5 is bad since 1 8 . . . gxf3 lets White break through by I 9 <2:lf5!) I 8 ,I xh5 gxh5 I9 tth2 g xd5 (or I 9 K c4 20 j_xc4 bxc4 2I ttxh5 f6 22 f4! and after 22 fxg5 23 fxe5 dxe5 24 .1. hi exd4 25 tth7 + '*f7 26 K fl + '*e6 27 tyxg7 White's attack is too strong, so Karpov­Sznapik, Dubai 01. I 986 finished 22 . . . <2:l f7 23 jth4! ttb8 24 � h i c 3 2 5 b 3 ttb7 26 f5 !Ele5 2 7 <2:1 e6 Resigns) 20 jtxd5 �b6 (now it becomes clear that interpolating fi b i and . . . b5 favours White

Dragon Variation 91

since b2 is not attacked) 2 I ttxh5 e6 22 A b3 and White has a win­ning attack, for example:

(2a) 22 . . . a5 (22 . . . <2:lg6 23 j_e3 tyc7 24 X h i and 22 . . . <2:l xf3 23 <E�f5 <E)xg5 24 <2:l xg7 *xg7 25 tyxg5 + *f8 26 e5! dxe5 27 X hi both win for White) 23 f4 <2:lc4 (23 <E)g6 24 f5) 24 j_h6 tyxd4 (24 . . . *f8 25 j_xg7 + '*xg7 26 ttg5 + '*f8 27 ttf6 or 24 . . . ttd8 25 j_xc4 bxc4 26 e5! f5 27 J..g5 tyc7 28 Af6) 25 X xd4 j_xd4 26 j_xc4 bxc4 27 e5 Resigns, Short-Mandl, Bundes­liga I 986.

(2b) 22 . . . ttc5 23 j_ h6 J..f6 24 f4 <E)g6 25 <E) f5 j_c6 26 ttxg4 d5 27 <2:lg3 tye3 28 e5 jt xe5 29 fxe5 'ijxh6 30 <E) h5 K c8 3I c3 tte3 32 j_c2 '*f8 33 J..xg6 fxg6 34 K fl + *e7 35 i!�tb4 + '*d8 36 tyd6 + Resigns, Chandler-Mestel, Brit­ish Ch. Play-Off 1 986.

16 J..xf6 The alternative I 6 gxh5 has not

been played much but it could also be good, e.g. I6 gxh5 a4 ( 1 6

<2:l xh5 I 7 � d 5 I� e8 I 8 a3 was good for White in Sznapik­Sehner, Slupsk I 987) and now:

92 Dragon Variation

( I ) 17 J..d5 b4 1 8 �ce2 e6? ( 1 8 �xd5 I 9 exd5 � xd5 20 h6 is

better but still good for White according to Sznapik) 1 9 h6! J..h8 20 h7 + ! *xh7 21 h5 exd5 22 hxg6 + *g8 23 � xh8 + *xh8 24 J..xf6 + Resigns, Sznapik­Komeljenovic, Biel Open 1 987.

(2) 17 h6 J..h8 18 J..d5 (after 1 8 h7 + not 18 . . . *xh7? 1 9 h5 �xh5 20 � xh5 + gxh5 21 �h2 and wins but 18 . . . �xh7! with unclear complications) *h7! (White's attack has been blocked) 1 9 a3 �b6 20 J..a2 b4 2 1 axb4 �xb4 22 J..xf6 �xf3! with a very dangerous attack for Black, Uly­bin-Savchenko, Simferopol I 988.

(3) 17 J..xti + !? *xf7! (not I 7 !! xf7? 1 8 hxg6 �xg6 1 9 h5

�f8 20 � dg l *h7 21 e5! with a decisive attack, lvanovic-Feick, Berlin Open I 988) I 8 hxg6 + ( 1 8 f4 b4!) � xg6 I 9 h 5 �e5 20 h6 �c4 (20 J..h8?! 21 f4 �c4 22 �e2! gives White more chances) 21 �g2 J..h8 22 h7 �a5, lvano­vic-Kosanovic, Stara Pazova I 988, and now 23 J..xf6! J..xf6 (23

� b4 24 � b3 axb3 25 cxb3 .txf6 26 �d5! is good for White) is unclear after 24 �d5!? or even 24 h8(�) j_xh8 25 �d5.

16 J..xf6 17 gxh5 a4 18 J..d5

Better than I 8 hxg6 axb3 I 9 �xb3 � xc3 20 gxf7 + *xf7 2 1 �xc3 (with g + 3 Jl, v J.. + � White might appear to have good winning chances, but in fact Black can just hold the balance) �c4 22

�d4 trb6 23 "ttd3 tyxd4 24 tyxd4 J..xd4 25 � xd4 *g6 and Black drew this ending in Chandler-Petursson, Hastings 1 986/7.

18 e6 19 hxg6 exd5 20 h5 A:l( xc3

In Wiech-Nizynski, Poland I 987, Black was successful with 20

J..g5 2 1 •g2 � xc3 22 bxc3 �c4 23 l!;!! dgl "tta5 24 �xg5 �xc3 25 gxf7 + *xf7 26 �xd5 + *e8 27 '!!!!fa8 + *f7 28 �d5 + *e8 and White had nothing better than perpetual check. How­ever 2 1 f4 �c4 (2 1 lli, xc3 22 bxc3 transposes to the main line, but this could be an important finesse of move order for Black if the suggestion of 21 '!!!!fh6 in the next note turns out to be strong) 22 �h2 J..h6 23 �xd5 appears much more dangerous.

21 bxc3 Not 2 1 'j!i!rxc3? since the tempo

saved by keeping the queen active is more important than the damage done to White's pawn structure. However 2 1 �h6 fxg6 22 hxg6 �e7 23 bxc3 is an inter­esting alternative since White transfers his queen to a better position with gain of time.

21 J..g5 22 f4 J..h6 ( 79)

Not 22 �c4 23 �g2 J..xf4 24 gxf7 + *h8 25 h6 � xf7 26 !i dg l J..g5 27 �xg5 � fl + 28 K xfl �xg5 29 :l;! f8 + *h7 30 l!I f7 + * h8 when White plays 3 I h7! �d8 (3 I J..h3 32 � f8 +

fi' xh7 33 �t xh3 + wins the queen) 32 H g l <E:Je5 (relatively the best) 33 H g8 + -.xg8 34 hxg8( '11') + \ti xg8 35 K f6 with excellent win­ning chances.

79 w

23 l!l dgl A very committal move. White

plays for mate, but if he cannot break through the move is proba­bly a mistake. Instead White should prefer 23 exd5, not only improving his material situation but also securing c6 and e6 as potential knight outposts. In this case I believe White has some advantage, although of course the position is still very complicated.

Dragon Variation 93

23 -.e7? Any error will have fatal conse­

quences in such a double-edged position. Here Black misses a good chance with 23 . . . dxe4! 24 gxf7 + (24 �g2 f5! 25 fxe5 dxe5 is unclear) fi'h8 25 �g2 (25 K g6 .:E)xg6 26 hxg6 fi'g7 27 -.h2 ,K h8 28 f5 -.f6 and Black survives) <E:J xf7 26 -.xe4 �e8! (26 . . . ,K e8 27 -.g6 is good for White) 27 -.d5 -.e3! and Black develops counterplay. In this case Black should be at least equal.

24 •g2 j)_xf4 Forced as 24 <E:lg4 (if the

knight goes anywhere else Black is mated by gxf7 + ) 25 f5 <E:Jf6 26 gxf7 + fi' h8 27 ttg6 wins outright.

25 g7 26 h6 27 <E:JfS 28 exfS 29 fxg6

l!:!( b8 <E:Jg6 J..xfS liteS

and Black lost on time while play­ing 29 f5. White wins by 30 � e I j)_ e3 3 1 lii xe3 followed by 'i!\j'xd5 + and mate.

6 Kan Variation

This line, which starts 1 e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 a6, is notable for the flexibility afforded to Black, since by delaying his piece development he keeps the maximum possible range of op­tions open. Because there are many reasonable choices at each move it is pointless to give precise lines against all possible move or­ders, so in this chapter there will be a greater emphasis on general principles. The continuation recommended in this chapter, 5 Ad3, is the most common line in practice. At the moment Black's most popular reply is to set up a 'hedgehog' position by 5 � f6 6 0-0 d6 (or 6 �c7 7 �e2 d6). After 7 c4 Black may choose to develop his bishop on e7 immedi­ately, but he sometimes brings the queenside out first in order to keep open the option of . . g6 and

jlg7. The 'hedgehog' name is derived from the way Black curls up on the first two ranks, moves like � bd7, b6, Ab7,

l:ii, e8, f!Jc7, J_f8, a ac8 and � b8 being typical. Black's slow development invites White to attack, but experience has shown that the unwary attacker can easily impale himself on Black's spines, and such at-

tacks have to be well-organized if they are to stand much chance of success. Moreover White has to watch out for Black's b5 and

. d5 breaks. Game 22 deals with the lines arising after 5 � f6, including the 'hedgehog'. All Black's other 5th moves, such as 5

jlc5, 5 . . �c6, 5 . . . �e7 and 5 g6, are in game 2 1 .

Game 21 Kengis-Nevednici

USSR 1979

1 2 3 4 5

e4 �f3 d4 Qlxd4 Ad3 (80)

c5 e6 cxd4 a6

5 jlc5 Black's alternatives are

arranged in descending order of importance:

( 1 ) 5 • . . � c6 (it is surprising

that this solid line is not p layed more frequently; although. the symmetrical position gives Black few winning chances, it is hard for White to prove any advantage) 6 �xc6 and now:

( l a) 6 . . . dxc6 7 �d2 e5 8 'l!!th5 j_d6 9 �c4 j_c7 (9 .Z:. f6 1 0 <E}xd6 + 'l!!fxd6 1 1 'l!!fe2 j_e6 1 2 0-0 gives White the chance to make his black-squared bishop a potent force, as in Jansa-Cebalo, Sme­derevska Palanka 1 978 after 1 2

�d7 1 3 .l, d l 'l!!fe7 1 4 b3 0-0 1 5 a4 a5 1 6 J..a3 �c5 1 7 j_c4!) 1 0 j_g5 � f6 1 1 'l!!fe2 h6 1 2 j_ h4 'l!!fe7 1 3 0-0-0 ( 1 3 0-0?! allowed Black to stir up trouble by 1 3 g5 14 j_g3 h5 1 5 f3 h4 in Ligterink­Miles, Lone Pine 1 979) il_e6 14 f4 J..xc4 ( 14 j_g4 1 5 il_xf6 il_xe2 1 6 j_xe7 j_xd l 1 7 j_d6 J,.xd6 1 8 � xd6 + <ffd7 1 9 �xf7 wins material for White) 1 5 j_xc4 b5 1 6 J..b3 0-0 1 7 J..xf6 'l!!fxf6 1 8 1, d7 and White's pressure against f7 gives him some advantage, Tseshkovsky-Miles, Bled-Por­toroz 1 979.

( 1 b) 6 . . . bxc6 (out of favour ever since the famous Fischer­Petrosian game mentioned below) 7 0-0 d5 (7 . . . e5 8 f4 j_ c5 + 9 <ffh l � e7 1 0 'l!!fh5 � g6 1 1 f5 � f4 1 2 j_ xf4 exf4 1 3 f6 J.. d4 14 fxg7 j_xg7 1 5 I;! xf4 'l!!fe7 1 6 <2Jc3 is very good for White, Ravinsky­Vorotnikov, USSR 1 963, while 7 . . . g6?! 8 e5 j_g7 9 f4 d6, van der Wiel-Anand, Thessaloniki 01. 1 984, should have been met by 1 0 exd6 'l'!txd6 1 1 �d2 'fjd4 + 1 2

Kan Variation 95

<ffh l � f6 1 3 �c4 �d5 1 4 �d6 + <ffe7 1 5 c3 with a fine position for White) 8 c4 �f6 9 cxd5 cxd5 1 0 exd5 � xd5 (10 • • . exd5 1 1 �c3 J..e7 1 2 'l'!ta4 + l/!l'd7 1 3 K e l ! 'l!!fxa4 1 4 .z:.xa4 j_e6 1 5 j_e3 0-0 1 6 J.. c5 is Fischer-Petrosian, match. 1 97 1 , which was won by Fischer, while 10 . . . 'l'!txd5 1 1 <E}c3 i>!td7 12 J..g5 J..e7 1 3 ll!l'e2 J..b7 14 !l ac1 0-0 1 5 .l. fd l , although keeping Black's pawns intact, gave White a dangerous initiative in Mikhalchishin-Gor­chakov, USSR 1 972) 1 1 j_e4 K a7 ( 1 1 . . . � b8 1 2 'l'!tf3 f5 1 3 j_xd5 'l'!txd5 14 �xd5 exd5 1 5 � d 1 J..e6 1 6 �c3 � d8 1 7 j_g5 � d7 1 8 .Z:.e2 gives White the better ending and 1 1 . . . il_e7 1 2 �c3 J..b7 1 3 'l!l!fa4 + �d7 14 ll!l'xd7 + <ffxd7 1 5 M d 1 is also promising after 15 . . . Ji::( ad8 1 6 �xd5 j_xd5 1 7 j_xd5 exd5 18 � xd5 + <ffe6 19 � xd8 � xd8 20 j_e3, Matanovic-Roos, Le Havre 1966, or 15 . . . j_f6 1 6 �xd5 J..xd5 1 7 j_xd5 exd5 1 8 � xd5 + <ffe6 19 Jii d2 � hd8 20 ii! e2 + , Averbakh-Taimanov, USSR Ch. 1 960) 1 2 'Wt'd4 � d7 1 3 �c3 � xc3 ( 1 3 . . . ,Zl b7 1 4 �xd5 il_xd5 15 j_xd5 l1li, xd5 16 ll!l'a4 + wins the a-pawn) 14 'l'!txc3 and Black's uncastled king gives him plenty of problems, Belyavsky­Kurajica, Sarajevo 1 982.

(2) 5 . . . g6 6 c4 j_g7 (this is an attempt to reach a kind of hedge­hog position, but with the bishop more actively deployed at g7) 7 � b3 and now:

(2a) 7 • . . �e7 8 � c3 d5 (8 . .

96 Kan Variation

0-0 9 0-0 Qlbc6 1 0 jte2 b6 1 1 J..f4 f5 1 2 exf5 ,l:l xf5 1 3 \¥f'd2 with advantage to White, Ljubojevic­Panno, Buenos Aires 1 980) 9 cxd5 exd5 1 0 J..g5! h6 1 1 j_ xe7 jtxc3 + 1 2 bxc3 'i¥f'xe7 1 3 0-0 dxe4 1 4 J..xe4 0-0 1 5 M e1 \¥f'c7 1 6 \¥f' f3 Qld7 1 7 Ql d4 *g7 1 8 � ab 1 and White has a big lead i n de­velopment, Donchev-Prie, Tou­lon 1 988.

(2b) 7 . . . d6 8 Qlc3 Qlf6 9 il_f4 (9 0-0 followed by J..f4 transposes to lines given in game 22, but White can do better here because Black has no time for Qlc6 and

Qle8) 0-0 1 0 jte2 e5 ( 1 0 �e8 1 1 c 5 is particularly un­pleasant when Black's queen is undefended) 1 1 il_e3 j_e6 1 2 0-0 �c6 1 3 f3 � c8 14 J!i! c l , Ljuboje­vic-Rajkovic, Yugoslavia 1 980, with a good position for White. The plan of directly attacking the d-pawn by Af4 and j_e2 is a logical way to exploit Black's jtg7, and in this case it gives White the advantage.

(3) 5 . . . Qle7 (Black aims to play Qlc6, but only when he can recapture with a piece) 6 Qlc3 (6 0-0 <2\ec6 7 c3 is also possible, when 7 jte7 8 j_e3 0-0 9 f4 d6 10 'E}f3 � d7 1 1 Qlbd2 gave White a small but enduring plus in Geor­giev-Peev, Bulgaria Ch. 1 980- 1 ) �ec6 7 � b3 jte7 8 �h5! d 6 9 jte3 Qld7 1 0 f4 b5 1 1 0-0-0 b4?! 1 2 �a4! e5 1 3 f5 0-0 1 4 g4! with an automatic attack for White, Mikhalchishin-Dorfman, Lvov 1983.

(4) 5 . . • *b6 6 c3! d6 (6 � c6 7 0-0 � xd4 8 cxd4 "l!!txd4 9 �c3 is dangerous for Black) 7 0-0 <£�f6 8 a4 jte7 9 Qld2 "l!!tc7 1 0 a5 0-0 1 1 'E}c4 'E}bd7 1 2 \¥f'e2 .l, e8 l 3 j_g5! J..f8?! ( 1 3 0 0 0 h6 1 4 jth4 b5 is just slightly better for White) 1 4 J..h4 with a good game for White, Lazic-Martinovic, Yugoslavia 1 987.

(5) 5 . . . "l!!tc7 6 0-0 <EJc6?! (Black can transpose to game 22 by 6 Qlf6) 7 Ql xc6 "l!!txc6 (or else Black has an inferior version of variation I above) 8 c4 (8 Ql d2 is probably also good) g6 9 <EJc3 J,.g7 10 � e l �e7 1 1 jtg5 d6 1 2 "l!!td2 with a very good position for White.

(6) 5 . . . b5 6 0-0 il..b7 7 'ilfe2 Qle7 8 a4 b4 9 Ql d2 Qlbc6 1 0 Ql4b3 Qlg6 1 1 f4 with advantage to White, Matanovic-Taimanov, Yugoslavia 1 965. One of the main advantages of 5 J..d3 as opposed to 5 � c3 is that an early b5 by Black is hardly ever a worry, since White may undermine Black's queenside pawns by a4 without fearing a loss of time after b4.

6 Qlb3 jta7 It makes little difference

whether Black plays 6 jta7 or 6 . . A b6, since he must exchange on e3 within a few moves in any case.

Or:

7 8 9 9

"l!!te2 Ae3 "l!!txe3

<2\c6 jtxe3

(81) d6

( 1 ) 9 . . . e5 1 0 �c3 �ge7 1 1

81 B

0-0-0 d6 I 2 J.. c4 b5 1 3 J.. d5 � xd5 I4 � xd5 j_e6 I 5 f4 0-0 with a roughly equal position, Wahls­Farago, Altensteig I 987, but I 2 g d2 followed by lt hd I and J.. e2 creates a more dynamic impres­sion.

(2) 9 . . . �f6 (this move order is sometimes played, but it may be inaccurate) IO e5!? (after IO � c3 Black should play IO d6; try-ing his luck with 1 0 0-0 is bad after I I e5 <E)g4 I 2 j_xh7 + !) � g4 I I 'ltg3 h5 (I I �cxe5? I 2 J..e2 h5 I 3 h3 h4 I4 itc3 wins for White) I 2 h3 � gxe5 I 3 'ltxg7 and White has some advantage.

(3) 9 . . . �ge7 IO <E)c3 0-0 I I 0-0-0 'ltc7 (or I I b6 I 2 f4 j_b7 1 3 � he i 'ltc7 I4 *bi d6 I 5 ith3 <E:� b4 I 6 a3 �xd3 + I7 X xd3 � c6 I 8 ith4! with dangerous threats, Vilela-Lebredo, Cuba I 983) I 2 f4 d6 I 3 ith3 � b4 I4 .!ll hfl ! e5 I 5 f5 d5 I6 exd5 �exd5 I7 �xd5 �xd5 I8 ith4 � f6 I 9 g4 e4 20 g5 exd3 2I !;K xd3 'l!lte5 22 gxf6 'ltxf6 23 �xf6 gxf6 24 )!;!!, d6 with a very good ending for White, Korlov­Batakov, corr. 1 984.

Kan Variation 97

10 �c3 �f6 1 1 0-0-0

White has an interesting alter­native in I I g4!?, when 1 1 . . . b5 I 2 0-0-0 0-0 1 3 g5 �e8 I4 f4 left White with a favourable version of the main line in Anand-Ninov, Baguio City 1 987. Anand finished off efficiently: 14 . . . b4 1 5 � e2 a5 I6 <E:�bd4 <E)xd4 I7 �xd4 �b6?! 1 8 e5! J.. b7 1 9 )!;!!, hfl ! dxe5 20 fxe5 J!!!, d8? 21 j_xh7 + ! \tixh7 22 g6 + \tig8 23 ith3 �f6 24 exf6 fxg6 25 fxg7 Resigns. The critical res­ponse to l l g4 is probably 1 1 . . . �xg4 1 2 'ltg3 � f6 1 3 'ltxg7 � g8 14 �h6 j_d7 intending 'lte7 and 0-0-0, and it is not clear how much advantage White can claim.

1 1 0-0 There are quite a few alterna­

tives: ( I ) 1 1 . . . e5 1 2 � d2 ( 1 2 j_e2 0-

0 1 3 f4 exf4 1 4 'ltxf4 �e8 I 5 �d4 J..e6 I 6 � f5 was a little better for White in Psakhis-Vyzmanavin, Moscow Ch. 1 98 1 ) J..e6 1 3 .i_ hd i 'ltc7 I4 J..e2 .i. d8 1 5 g4! 0-0 ( 1 5

h6?! 1 6 f4 exf4 1 7 'ltxf4 �e5 1 8 �d4 0-0 1 9 g5 d5 20 * b l was good for White, Varjomaa-Tor­nefjell, corr. 1 979) 1 6 'ltg3 J..xb3 1 7 cxb3! '2Jd4 1 8 * b l b5 19 j_d3 �c6 20 f3 � fe8 2I �g2 b4 22 g5! and the outpost on c4 for White's bishop aided an already danger­ous kingside attack in Lekander­Schoneberg, corr. I 980.

(2) 11 . . . b5 12 � d2 ( 1 2 g4!? looks good) 0-0 1 3 .i_ hd l -wtc7 1 4 f4 b4 1 5 '2Je2 e 5 1 6 fxe5 dxe5 1 7

98 Kan Variation

l:! fl �d7 1 8 �g3 � b6 1 9 .f) f5 with strong pressure for White, Byrne-Larsen, Biel IZ 1 976.

(3) 1 1 . . . "f!tc7 may transpose into the main line after 1 2 f4 0-0, but it allows White the alternative of 1 2 g4!, which looks unpleasant for Black.

12 f4 "f!tc7 ( 82)

13 g4 There seems to be no general

agreement as to whether White needs to prepare this with K hg I . The alternative attacking ideas are:

( 1 ) 13 "f!th3 � b4 1 4 g4 b5 1 5 g5 � xd3 + ( 1 5 �e8 1 6 "f!th4 f6 1 7 a3 fxg5 1 8 fxg5 � xd3 + 1 9 � xd3 K b8 was unclear in Vogt­Velikov, E. Germany-Bulgaria 1 987) 1 6 K xd3 �e8 1 7 f5 b4 led to a sharp finish in Bronstein­Suetin, Moscow Ch. 1 982 after 1 8 "f!th4 bxc3 1 9 K h3 cxb2 + 20 *b1 f6 2 1 "f!txh7 + *f7 22 K g1 J g8 23 K h6 exf5 24 Ji!t xf6 + � xf6 25 gxf6 *xf6 26 �xg8 j_ b7 27 'l!rh7 j_xe4 28 "f!tg6 + *e5 29 "f!tg3 + Draw.

(2) 13 I, hg1 (this is probably good enough for a slight advan­tage) b5 1 4 g4 b4 1 5 g5 �e8 1 6 �e2 ( 1 6 � b l a5 1 7 .!;l g4!? a4 1 8 � 3d2 j_a6 1 9 j_xa6 � xa6 20 X h4 g6 2 1 "f!l h3 f5 22 gxf6 I, xf6 23 �c4 may be slightly better for White, Wedberg-Spraggett, New York Open 1 987) "f!ta7 ( 1 6 a5 1 7 �bd4 �xd4 1 8 �xd4 j_a6 1 9 *bl was a little better for White in Arnason-Suetin, Sochi 1 980) 1 7 "f!th3! g6 1 8 f5 exf5 1 9 exf5 �e7 20 �g3! "f!te3 + ?! (20 a5 is better, but still favours White) 2 1 *bl j_ xf5 22 j_xf5 �xf5 23 � de l "f!tf4 24 � gfl tth4 25 ttxh4 � xh4 26 g e4 with an ex­cellent ending for White, Arna­son-Kirov, Plovdiv 1 986.

13 b5 Accepting the offer must be a

critical test of White's willingness to play g4 without the prepara­tory !!I hgl . In Short-Velikov, European Club Ch. 1 987 the con­tinuation was 1 3 � xg4 1 4 �g3 <E) f6 1 5 � hg l �e8 1 6 *b1 ( 1 6 f5!? i s natural) � e7?! ( 1 6 b5 and 1 6 f6 have been sug­gested as possible improvements) 1 7 �d4 ttc5 1 8 �f3 f6 1 9 e5! with a very strong attack, Short­Velikov, European Club Ch. 1 987. Velikov must have found an improvement because he repeated this line in a later game Ivanovic­Velikov, Saint John Open 1 988. Unfortunately Ivanovic varied by 16 "f!th4, so we don't know what Velikov's intention was. Despite this hint, I believe White has good

compensation for the pawn and it would require a brave Black player to take this line on.

14 g5 <2:ld7 15 f5!?

1 5 '/th3 <2:lb4 1 6 a3 Qjxd3 + 1 7 '/txd3 d5 gave Black counterplay in Georgadze-Bohlig, Halle 1 978, although White won this game too.

15 b4 16 Qje2 a5

It is almost impossible to assess positions in which the players are attacking on opposite wings. Unless one of the players is well in front it is likely that a single tempo will decide the race and obscure tactical points will often have a crucial influence on the play.

17 '/th3 Attacking e6 directly, and

creating a concealed threat to h7. 17 exf5 18 exf5 Qjde5

The pin along the c8-h3 diag­onal is awkward for White since he cannot move his kingside pawns, nor can he unpin by '/th4 since Qjxd3 + and jtxf5 repulses the attack. Bringing the knight to d5 is the only way to make progress.

19 <2:lf4 a4 20 <2:ld5 '/tdS (83)

Now that b3 and g5 are under attack, White is committed to the sacrificial path.

21 .!;! hgl! Qjxd3 + ? Black decides to eliminate one

of the attacking pieces, but in

Kan Variation 99

doing so brings the d 1 rook into the attack. The best defence was 21 axb3! 22 <2:lf6 + gxf6 23 'tl!f h6! (23 gxf6 + €! g6 leads nowhere as the f5 pawn is pinned while 24 'tl!t'h6 '/txf6 25 fxg6 fxg6 26 jj_xg6 'ltf4 + 27 '/txf4 K xf4 28 ii_e8 + *f8 29 jtxc6 bxa2 30 *d2 lii d4 + wins for Black) <2:lg6! (23 * h8 24 gxf6 K g8 25 14, g7 .f)xd3 + 26 l! xd3 jtxf5 27 !!l h3! leads to mate at h7) 24 fxg6 fxg6 25 ii_xg6 K a7! 26 gxf6 hxg6 (26

'/txf6 27 jtxh7 + *h8 28 � g8 + ! )i xg8 29 jtg6 + mates, or 27 *f7 28 H dfl ) 27 lii xg6 + (27 'tl!fxg6 + fi h8 leads to nothing as the d l rook cannot reach the h-file, Ji d5 being met by

. Qj e5) 'f1f7 reaching a remark­able position in which it appears that White must mate, but it isn't certain that he can do so. In the first edition I commented that I couldn't see a mate after 28 lil g7 + 'f1e6 (28 . 'f1e8 29 '/te3 + Qje5 30 '/txa7 'ltxf6 3 1 axb3 i s very good for White, with material equality but weak Black pawns and an exposed Black king) 29 '/th3 + *xf6 30 '/th6 +

100 Kan Variation

*e5 (30 <IJ'f5 3 1 � g2! does mate), for example 3 1 � g3 jj_f5 32 c4 jj_e4! I still don't see a mate, and nobody wrote to me suggest­ing one, but if we return to the position after 27 <IJ'f7 White can gain a massive advantage by 28 lJ, e l � e5 29 K g7 + *e6 (29

*e8 30 l;l xa7 "ttxf6 3 1 'ith5 + <IJ'd8 32 axb3 is similar) 30 K xa7 bxa2 (30 "ttxf6 3 1 'ith3 + and "ttxb3 + ) 3 1 "tth3 + *d5 32 'itb3 + <IJ'c6 33 "ttxa2, with a slight material plus for White together with a raging attack.

22 K xd3 �e5 Now White wins by force.

23 �f6 + gxf6 24 "tth6!

Once again 24 gxf6 + � g6 re­pulses the attack, but now 24 �g6 loses to 25 M h3 � e8 26 fxg6 fxg6 27 "ttxh7 + *f8 28 "tth8 + and H h7 mate. 24 *h8 25 K h3 J..xf5 26 g6! j_xg6 27 K xg6 is also mate, so Black must take the rook.

24. . . <£J xd3 + (84)

Black is hoping for 25 cxd3 '* h8 when 26 gxf6 X g8 27 X g7 J..xf5 defends, as does 26 g6 fxg6 27 fxg6 "ttc7 + and 28 "ttg7.

25 <IJ'bl! fxg5 25 '* h8 26 g6 fxg6 27 fxg6

threatens both 28 "ttxh7 mate and 28 g7 + , while 25 . . . "it b6 26 M g3 only makes matters worse.

26 f6 "ttxf6 27 "ttxf6 g4

The only way to avoid mate. Although rook, bishop and two pawns amount to enough mater­ial to balance a queen, Black still suffers from his bad king position.

28 "ttg5 + fih8 29 "ttf6 + *g8 30 �d4 �e5 31 h3

Intending 32 hxg4 followed by � f5 .

31 32 "ttg5 + 33 "ttxh5 34 "ttd5

h5 �g6 gxh3

Attacking a8 and g6. White finishes the game with the same energy he has displayed through­out.

34 J..e6 35 � xe6 h2 36 ll xg6 + Resigns

36 fxg6 37 � xf8 + and 'l!!fxa8 wins all the black pieces.

Game 22 Nunn-Gheorghiu Hamburg 1984

1 e4 c5 2 �fJ e6 3 d4 cxd4

85 B

4 �xd4 a6 5 J..d3 � f6 6 0-0 (85)

6 d6 The surprising move 6 . . . e5 (6

d5 7 e5 is very bad for Black since the natural 7 � fd7 loses to 8 �xe6!) was played in Fedor­owicz-Dorfman, New York 1 989, and now 7 J..g5! exd4 (7 h6 8 J..xf6 -rtxf6 9 �e2 d6 1 0 � bc3 J..e6 1 1 f4 is good for White) 8 e5 J..e7 (8 . . . -rta5 9 J..d2! and 8 . . . h6 9 exf6 hxg5 I 0 !:! e I + are very bad for Black) 9 exf6 J..xf6 IO J_xf6 ttxf6 I I !:!ei + lli'f8 12 J..e4 i s a little better for White since Black will still have an iso­lated pawn after White regains the front d-pawn.

Black has a major alternative in 6 . . . -rtc7 7 -rte2 (7 c4 �c6 8 �xc6 dxc6 is now considered sat­isfactory for Black because White has spent a move on c4, which in this position only serves to weaken d4) d6 (7 . . . �c6 is bad because of 8 � xc6 and 9 e5, while 7 . . . J.. c5!? 8 � b3 is a speciality of Eingorn; the two examples 8

Kan Variation 101

. . . J..e7 9 c4 d6 10 �c3 0-0 I I J..d2 b6 12 �d4 j_b7 1 3 lli' h l �c6 14 �xc6 j_xc6 1 5 K ae l ttb7 1 6 a4 �d7 1 7 b4 -rtc8 1 8 f4, Govedarica-Eingorn, Novi Sad I 988 and 8 . . . j_a7 9 lli' h l d6 IO c4 �c6 1 1 J..g5 J..d7 1 2 �c3 �d4 1 3 �xd4 j_xd4 1 4 � ac i -rtc5 I 5 J_d2, Smyslov-Eingorn, USSR Ch. I 988 were both very slightly better for White, but the idea is an interesting one and I expect to see more of it; perhaps 8 J_e3 is the best reply) 8 c4 g6 (this is the move which gives 6 . . . -rtc7 its distinctive flavour; if Black de­velops his bishop at e7 we reach positions similar to those in the main line below, but after . . . g6 White's prospects of a direct kingside attack are reduced; on the other hand d6 is weakened) 9 �c3 (it is also worth considering b3 and J..b2) J..g7 IO l! d 1 (White's idea is to attack d6 by jtc2, �f3 and J..f4) 0-0 (86) and now:

( 1 ) 11 J..c2 � bd7 (1 1 . . . b6 1 2 � f3 � fd7 1 3 J_e3 �c6 1 4 .l, ac1 jtb7 15 J_b1 i s good for White,

102 Kan Variation

Ivanovic-Cvitan, Vrsac 1 987, while 11 . . . '2l c6 12 '2l xc6 bxc6 1 3 JH4 e5 14 j}_e3 J..e6 1 5 J_b3 � fd8, Cabrilo--Gheorghiu, New York Open 1 988, gives White a slight plus after 1 6 h3 or 1 6 � d3) 12 22lf3 <E:�g4 (not 12 b6 1 3 j}_f4! 2£le5 1 4 2£lxe5 dxe5 1 5 J_ e3 J..b7 1 6 2£l a4 b5 1 7 cxb5 axb5 1 8 2£lc5 j}_c6, Ljubojevic-Hulak, Wijk aan Zee 1 987, and now 1 9 b4! followed by j}_ b3 is good for White) 1 3 J.. f4 (not 1 3 J_g5 <E:�de5 14 '2ld2 b5! and Black seized the initiative in Klinger-Jukic, Berne Open 1 988) and now 1 3 2£lde5 1 4 2£ld2 b5?! 15 h3 '2lf6?! 16 cxb5 axb5 1 7 J_b3 b4 1 8 '2lb5 �b8 1 9 '2lxd6 was good for White in Cabrilo--Kovacevic, Vrnjacka Banja 1 9880 However Black's play was poor in this example; had he played 1 3 0 0 0 '2lge5 White would have a slightly inferior version of line 2a below in which his bishop is on c2 instead of the better square b l .

(2) 1 1 22l f3 (this may be more accurate since White often prefers to retreat his bishop to b 1 ) and now:

(2a) 11 . . . '2lbd7 1 2 �f4 <2!g4 1 3 � ac l b6 1 4 A b 1 '2lge5 1 5 b3 '2lxf3 + 1 6 'tttxf3 l£je5 1 7 �e2 and now:

(2al ) 17 . . . jtb7 1 8 �d2 ( 1 8 j}_ e3 � fd8 1 9 �d2 � ab8?! 20 h3 j}_a8 21 f4 2£lc6 22 'tttf2, Armas­Gheorghiu, Romania Cho 1987, was slightly better for White, but Black can try 19 Jt f8 or 1 9 '2lg4) � fd8 1 9 j}_g5! � d7 (Black

would like to play 1 9 f6 20 j}_e3 g5, but this allows 2 1 l£ja4) 20 h3, intending f4, with advan­tage to White, Armas-Ionescu, Romania Cho 1 9880

(2a2) 17 . . . � b8 (this seems to be more accurate; Black first of all secures the weak b6 pawn) 1 8 h3 jtb7 19 �d2 � fd8 20 J..g5 (the immediate 20 j}_e3! transposes to Armas-Gheorghiu above, which was slightly better for White) f6 2 1 jte3 g5 22 j}_d3 (now 22 2£l a4 is met by 22 jL c6) A c6 23 '2lb5 fi/f7 with equality, Wolff­Hulak, Toronto 1 9890

(2b) 11 . . . 2£lc6 (in many lines it makes no difference whether Black plays 0 '2l bd7 or l£jc6 because he will continue with .'£\ f6-g4--e5xf3 + followed by .'£)e5 in any case, but here we examine one independent line) 1 2 .ii.f4 '2ld7 1 3 � ac l <2\de5 1 4 b3 yjfe7 ( 1 4 l£j xf3 + will trans­pose to line 2a) 1 5 j}_ b 1 (White aims to expel the e5 knight by moving his 22lf3 and Jtf4 away) � d8 ( 1 5 il_d7 1 6 .'£) e l � fd8 1 7 'rl!i'd2 jte8 1 8 jtg5 f6 19 jj_h6 .il_xh6 20 �xh6 also gave White an edge in Timoshchenko-­Eingorn, Tallinn 1 989) 1 6 jL e3 jj_d7 1 7 '2l d2! .ii_e8 1 8 f4 '2ld7 1 9 22lf3 );:;!, ab8 and White retains a slight advantage, Kuzmin­Eingorn, M oscow 1 9890

7 c4 (87) J..e7 Other 7th moves: ( 1 ) 7 . . . g6 8 � c3 J..g7 9 '2l b3!

0-0 10 Jt e2 i£jc6 1 1 J..f4 <E:� e8 (Black has little choice as 1 1 0 0 0

2£je5 1 2 c5 is very awkward) and after 1 2 li1d2 b6 1 3 ll fd 1 2E:le5 14 � ac l 'f!!c7 Black equalized in Nunn-Gheorghiu, Vienna 1 986. However 1 2 c5! is unpleasant for Black, for example 12 . . . dxc5 1 3 )\'ltxd8 Qlxd8 1 4 Qla4! or 1 2 . . . e5 1 3 �e3 �e6 1 4 Qld5.

(2) 7 . . . �d7 8 Qlc3 Qlc6 9 Qlxc6 �xc6 1 0 li1e2 �e7 1 1 b3 0-0 1 2 jt b2 l1l[ e8 (12 . . . Q�d7 1 3 f4 is slightly better for White, while after 12 . . . �b8 1 3 a4 � e8 1 4 li\ ae l Q�d7 1 5 f4 � h4! 16 � d l A_f6 1 7 b4!? the position was unclear in Nunn-Bischoff, Dort­mund 1 987; 14 f4! was more ac­curate, not committing the a l rook for the moment) 1 3 f4 d5!? 14 cxd5 exd5 15 e5 Ql e4 1 6 l£l xe4! dxe4 1 7 jtxe4 J..b5 and now 18 ll!rg4?! �d2! 19 jtxh7 + !? was unclear in Arnason-Toshkov, Jurmala 1 987. White can draw by 18 �c2 � xfl 1 9 .�.xh7 + *h8 20 *xfl lii\ c8 2 1 �d3! g6 22 �xg6 hxg6 23 �xg6 �d2, but 18 )\'ltf3!? is the most promising, not only playing for a possible at tack by

Kan Variation 103

�h3, but also lining up against the b7 pawn.

(3) 7 . . . b6. Normally this transposes to lines considered be­low.

8 '2'Jc3 White has two main attacking

plans, which are distinguished by the 4evelopment of his queen's bishop. Firstly he may build up a slow kingside attack by b3, jt b2, 2£jc3, li1e2, f4, ll ae 1 and so on, with the ultimate aim of a breakthrough by f5. The other plan is to prepare for e5 by l£Jc3, 'itfe2, f4, �d2 and l;l ae l . The important point is that with the bishop on b2 the e5 plan is much less effective, because White ends up with a pawn on e5 and this would block the bishop on b2.

8 0-0 9 li1e2

Moves such as "i!!te2, * h l and f4 are logical because they do not commit White to one plan or the other. My view is that the e5 plan is most effective against . . . '2'1 bd7 by Black, because then the queen's knight blocks the retreat of the one on f6. Therefore it is often useful to delay committing the c l bishop until Black has moved his b8 knight. Against 2£jc6 White will take on c6, then play b3 and §l b2, and against '2'J bd7 White will play jtd2 and K ae l .

9 b6 Black also delays for as long as

possible. 10 f4

104 Kan Variation

Despite the above (rather sub­jective) comments it is quite rea­sonable to play b3 straight away, the advantage being that White can sometimes manage without 'itr>h l . After 1 0 b3 jl_ b7 1 1 jt b2 (88) there are two lines:

88 B

( 1 ) 1 1 .:'£\ c6 (as the earlier explanation makes clear, this move plays into White's hands since we reach positions similar to the main line below, but with White having saved about half a tempo by missing out * h 1 ) 1 2 .!£)xc6 jl_ xc6 and Black has been highly unsuccessful from this position:

( l a) 13 ,g adl )'tbS?! 14 a3 (what on earth is this for?) � dS? (it doesn't matter about the tempo spent on a3 in view of the way Black plays) 1 5 f4 .!£)d7 1 6 e,d5! jl_f8 17 B. f3! li eS IS li h3! g6 1 9 )'tg4 )'tdS 20 lil fl jl_g7 2 1 J_xg7 *xg7 22 f5! with a massive at­tack, lvanovic-Ermenkov, Plov­div 1 9S3.

( l b) 13 f4 .!£ld7 14 14, ad l b5? 1 5 cxb5! axb5 1 6 jl_ xb5 11t b6 + 1 7 J!l f2! jl_xb5 I S 'ilfxb5 >�txb5 1 9

.!£)xb5 li;l xa2 20 jl_xg7 !HaS 2 1 i!i xa2 !!! xa2 22 jL d4 e5, Hellers­Adamski, Eeklo l 9S5, and now 23 Jtc3 gives White a won ending.

( l e) 13 � ael !i!, eS 14 f4 g6 1 5 e5! dxe5 1 6 fxe5 jl_c5 + 1 7 * h l .!£)g4 I S jl_e4! .!£)xe5 1 9 jl_ xc6 .!£)xc6 20 )'tf1 again with a tre­mendous attack for White, Ermenkov-Gheorghiu, Prague 1 9S5.

(2) 11 . . . .!E\ bd7 12 !!! ad l ( 1 2 f4 is also playable) � eS ( 1 2 >�tc7 1 3 jl_ b l li;l feS 14 f4 � acS 1 5 2210 jl_ f8 1 6 *hi Jtc6 was less accur­ate and after 1 7 e5! jl_xf3 I S !il, xf1 dxe5 1 9 fxe5 .!£)g4 20 !il, xf7! White had a very dangerous attack in Plachetka-Ravikumar, Copenha­gen 1 9SO) 1 3 Jt b l >�tbS 14 f4 A f8 1 5 * h l !il, a7 1 6 .!£)f3 ;laS with a double-edged position, Akesson­Mestel, Copenhagen 1 9SO, although I still favour White.

Or:

10 jl_b7 1 1 *hl (89)

l l .!£)c6

( 1 ) 1 1 . • . .:'£\ bd7 1 2 Jtd2 >�tc7 1 3 � acl ( 1 3 l!l ae l l!l feS 1 4 � f1 g6

1 5 K g3 ! flh8 1 6 K h3 e5 1 7 �f3 exf4 1 8 J..xf4 J..f8 1 9 'ltf2 �c5 20 jtc2 was also good for White in I vanovic-Peev, Ba1asiha 1 977) g6 1 4 b4 )l ac8 1 5 a3 'ltb8 (the ad­vantage of playing f4 is that the weakening of c4 created by White playing b4 cannot be exploited by . . . �e5) 1 6 �f3 )l fe8 1 7 B, ce 1 J..f8 1 8 �g5 h 6 ( 1 8 e 5 1 9 f5 gave White a strong attack at no material cost in Commons-Naj­dorf, Lone Pine 1976) 19 � xf7! flxf7 20 e5 �g8 21 'ltg4 �e7 22 J..xg6 + ! �xg6 23 f5 �dxe5 24 fxe6 + ! (24 fxg6 + ff g8 is unclear) ffe7 (24 . . . * g7 25 !! xe5 dxe5 26 I, f7 + mates) 25 'ltxg6! (the cli­max of a magnificent combi­nation) ffd8 26 l;i xe5 dxe5 27 J..xh6 11 xc4 (27 J..xh6 28 H d 1 + mates) 28 'ltxe8 + (White gives up his queen after all) ffxe8 29 )l xf8 + fje7 30 )l xb8 jtc6 3 1 � d 1 b5 32 * g l •xe6 33 K b6 Resigns, Commons-Peev, P1ovdiv 1 976.

(2) 1 1 • . . l;l e8 (dubious as it allows White to play for e5 with­out delay) 1 2 �f3 g6 (12 . . . � bd7 1 3 e5 dxe5 1 4 fxe5 <2)g4 1 5 Af4 J..xf3 1 6 )l xf3 and 12 . . . �c6 1 3 e5 dxe5 1 4 fxe5 <EJg4 1 5 J..f4 are good for White) 1 3 e5! <2) h5, Sax­Bellon, Dubai 01. 1 986 and now Sax recommends 14 A e4 � c6 1 5 g4 �g7 1 6 f5 ! , when the lines 16 . . . exf5 17 gxf5 dxe5 1 8 fxg6 hxg6 1 9 �xe5, 16 . . . gxf5 1 7 gxf5 �xf5 ( 1 7 exf5 1 8 jtd5) 1 8 J..xf5 exf5 1 9 <E)d5 and 16 . . . dxe5 1 7 fxg6 f5 ( 1 7 . . hxg6 1 8

Kan Variation 105

� xe5) 1 8 gxf5 exf5 1 9 j)_ d5 + ffh8 20 �xe5 are all good for White.

12 �xc6 jtxc6 13 b3 <E)d7

Or 1 3 'ltc7 1 4 J..b2 .i! ad8 and now:

( I ) 15 � adl g6 1 6 J..b l � h5 1 7 K d3 ( 1 7 g4 �g7 1 8 f5! exf5 1 9 gxf5 was probably better) e5! 1 8 fxe5 dxe5 1 9 M xd8 K xd8 20 �d5 with just an edge for White, Matulovic-Tringov, Vrnjacka Banja 1 986.

(2) 15 K ael (intending �d5) J..b7 1 6 J..b 1 �d7? ( 1 6 g6 is probably better, when 1 7 'ltd3 is only a slight plus for White) and now 1 7 'lth5 H fe8 1 8 li, e3 � f6 1 9 'lth3 g6 20 f5! gave White a deci­sive attack in Nunn-Gheorghiu, Biel 1 983, which is annotated in detail in Secrets of Grandmaster Play by Peter Griffiths and the present author. For some reason Gheorghiu repeated the whole line in the game Mokry-Gheorg­hiu, Prague 1 985. That game con­tinued 1 7 'i'lf'g4 J..f6 1 8 � e3 g6 1 9 K h3 * h8?? 20 � d5! winning, as 20 exd5 is met by 2 1 'lth4. Of course 1 9 )g[ fe8 is better, but 20 'ltg3 intending f5 gives White a dangerous attack in any case.

14 j}_b2 g6 1 4 J.. f6 leaves d6 weak and

after 1 5 K ad l 'i'lf'c7 1 6 J..b l )l fd8?! ( 1 6 � ad8 was a better chance) 1 7 K d3 g6 1 8 .M fd 1 �c5 19 � d5! exd5 20 J..xf6 � xd3 2 1 exd5! �xf4 22 'lt f3 � h5 23 jtxd8 � xd8 24 dxc6 White had a clear

106 Kan Variation

advantage in Marjanovic-Rajko­vic, Yugoslavia Ch. 1 983.

15 K ad1 The purpose of this (rather

than K ae 1 ) is to prevent the de­velopment of Black's e7 bishop to the long diagonal. After 1 5 il..f6, for example, 1 6 _;tb1 �c7 1 7 l!td2 attacks d6.

90 B

15 K e8 16 _;tb1 (90)

16 l!fc7 The alternatives are: ( 1 ) 16 . . . K a7 1 7 a4 (the direct

1 7 l!td3 was also tempting) jj_f8 1 8 jj_c2 l!fa8 1 9 lltf2 �c5? ( 1 9 � c7 was better) 20 Qld5! exd5 2 1 cxd5 jj_xa4 22 bxa4 j_g7 23 !J1xg7 'fixg7 24 e5! and White stands well, Popovic-Kotronias, Pucar­evo 1 987.

(2) 16 . . . *'b8 17 f5 ( 1 7 a4!? is possible, but I like 17 l!fd3 b5 1 8 cxb5 axb5 1 9 �e2! b4 20 Qld4 jtb7 2 1 llth3 with dangerous threats on the kingside) b5 1 8 fxe6 fxe6 1 9 cxb5 axb5 20 � d5!? j_xd5 (20 . . exd5 21 exd5 jj_f6 22 *'d2 j_xb2 23 dxc6 is good for White)

2 1 exd5 e5 22 l!fg4 � f6 23 l!fe6 + '* g7 was unclear in Prasad­Gheorghiu, Biel Open 1985.

(3) 16 . . . jj_f8 17 e5! dxe5 1 8 jj_e4! gives White a dangerous at­tack. The game Vogt-Gheorghiu, E. Germany-Romania 1 984 con­tinued 1 8 . . *'c7 ( 1 8 . . . jJ_xe4 1 9 �xe4 jj_g7 20 jJ_xe5 jj_xe5 2 1 fxe5 'fig7 22 l!tf2 wins after 22 . . . l!fe7 23 *'f6 + ! \t1g8 24 'flixe7 !! xe7 25 � xd7 or 22 . . . � e7 23 �g5 l!fg8 24 �xf7 K f8 25 � xd7!) 1 9 l!ff3 jtxe4 20 � xe4 f5 (20 _;tg7 2 1 fxe5 �xe5 22 � f6 + 'flh8 23 jJ_ xe5 wins material) 21 X xd 7 fxe4 22 'f!fd 1 �c6 23 jj_xe5 X ac8 24 �d4 jte7 (24 . . . '!A, e7 25 � d6 drops the b-pawn, while 24 . . . b5 25 'f!la7! and 24 . . . jj_c5 25 I:I g7 + 'flf8 26 j_d6 + lead to mate) 25 f5! (the immediate 25 _;th8 jj_f8 26 � g7 + 'flxh8 27 � xg6 + is met by e5, but if Black now plays 25 exf5 this line wins) K cd8 (25 gxf5 26 jJ_ h8 J.. f8 2 7 M\ xf5! exf5 28 ll g7 + mates) 26 I:I xe7! (26 fxg6! � xd7 27 gxh7 + 'flxh7 28 I� f7 + \t1h6 29 'W!'e3 + _;tg5 30 l!fh3 + also wins) exf5 (26 K xd4 27 � g7 + is mate next move) 27 !:! xe8 + �xe8 28 �xb6 Resigns.

17 lltd3 A flexible move attacking d6

and preparing l!fd4 or �h3 according to circumstance.

17 jj_f8 This move cost Black forty mi­

nutes, presumably checking that the line 1 8 �d4 e5 ( 1 8 . . . jj_g7 1 9 llfxg7 + wins) 1 9 �d5 jj_xd5 20

'/Wxd5 exf4 2 1 � xf4 �e5 22 Z:�: dfl lli\ a7 presented no dangers .

18 �b5!? A shock for Black. After 1 8

axb5 1 9 'ltc3 eS 20 cxb5 B, ac8 21 bxc6 '/txc6 22 'i!lttf3 Black cannot exchange at f4 since he has no satisfactory way to cover fl, so White gets to play 23 f5, when Black's white squares look very sickly.

18 jixb5 19 cxbS axb5

1 9 . a5 20 !:!: c l �c5 2 1 'ltc3 is much worse as Black has to play the weakening . e5.

20 '/txb5 �b7 The upshot of White's mini­

combination is that he has the two bishops and a queenside pawn majority. Now 2 1 a4! would have been logical, relieving the b l bis­hop of its defensive duty and pin­ning down the b6 pawn. If Black remains passive the b I bishop can eventually move to b5. During the game I didn't like 2 1 a4! �c5 22 �c4 X ac8, but 23 f5 e5 24 ll!. f3 gives Black no way to exploit the position of White's queen (24 � e6 25 fxe6!), and consolidation by 25 'lte2 will be good for White.

21 'lte2?! b5! Black takes the chance to pre­

vent a4. Now . . . b4 would perma­nently cripple White's queenside pawns, so

22 b4 Black cannot transfer his

knight to c4 because b5 is weak, for example 22 '/ta6 (threat . <EJb6) 23 ji d3 '(!lfxa2 24 ji xb5

Kan Variation 107

'/Wa7 (24 ll: ed8 25 jixd7 and .i, a l ) 25 jic6 followed by jid4 and b5, with an excellent position for White.

22 fj!. a6 (91) Another useful defensive move,

covering d6 in preparation for jig?. Here I thought for a long time trying to find a way to keep the advantage.

23 � f3! 24 jixg7 25 iiii fd3 26 � h3

The point of White's ma­noeuvre is that he gains a tempo by attacking b5 to transfer his rook to the h-file, reviving some threats against Black's king.

26 X b6 Not 26 . . . 'ltc4? 27 "*'b2 + and

28 X c3 trapping the queen. 27 '(!lfd2

The threat is 28 e5 d5 29 f5 and the queen gets to h6.

27 .:2)f6 Once again Black finds a good

defence, relieving the pin down the d-file.

28 f5 exfS

108 Kan Variation

28 e5 29 trh6 + \fig8 30 g4! is very unpleasant.

29 exf5 d5! All these difficult moves were

very time-consuming, so that Black had only a couple of mi­nutes left to reach move 40. At first sight White can win by 30 fxg6 fxg6 3 1 trh6 + \tg8 32 Axg6 hxg6 33 tyh8 + (33 trxg6 + tyg7 34 trf5 I. ce6 defends) *f7 34 K h7 + *e6 35 K e1 + (35 trxf6 + *xf6 36 X xc7 X be6 gives Black enough for the pawn in view of his active rooks and king), but then comes 35 *f5 ! 36 D, fl + *g5! 37 h4 + \tg4 and White's queen is trapped.

30 fxg6?! The immediate 30 X fl ! was

much better, when the defence Black plays in the game would have been prevented.

30 fxg6 31 x n

Now truly threatening 32 trh6 + *g8 33 Axg6.

31 <2lg4! Suddenly Black exploits

White's weak back rank. If 32 trxd5? then 32 <2lf2 + wins.

32 trd4 + H f6 (92) White cannot take the rook or

the knight, nor can he play X hf3 . The move played supports the back rank and attacks the knight.

33 Ad3 h5 Forced, but good. The draw

would now be in sight if it were not for Black's time-trouble .

34 M xf6 <2l xf6

35 K f3 K e1 + 36 Afl tye7?!

36 • . . tye5 37 trxe5 K xe5 38 Axb5 gives White good winning chances as the d-pawn cannot pass over the d3 square. However 36 . . . �d6! was more accurate, for example 37 *g1 (37 h3 I! e4 38 'l'!lrc3 K xb4 39 H xf6 trxf6) l!!( e4 38 trc5 (38 tyc3 d4 39 tyd3 <2lg4! 40 g3 X e 1 threatening . . . <2le5 is unpleasant) trxc5 39 bxc5 X e1 and a draw is inevitable (40 H c3 d4).

37 \tg1 H e4 Black can also choose to wait,

since progress isn't easy for White, but in time-trouble it is very natural to break the pin.

38 trc3 d4 Not 38 K xb4 39 X xf6 d4

because of 40 M e6! 39 trd3

With two moves still to make Black suddenly finds himself in trouble over the b5 pawn. 39 . . . trxb4 loses to 40 Ill xf6, so the best chance is 39 <2ld5 40 trxb5 <£! xb4. Then 41 Ac4 H e l + 42 fif2 fih6 isn't danger-

ous, so White's winning prospects are very slight.

39 <£)g4? 40 h3

and Black's flag dropped before he could make his 40th move. After 40 <£)e3 41 iifxb5 <£)xfl 42 � xfl � e2 43 �d3 iife3 + 44 iifxe3 dxe3, suggested by Black after the game, White can win by 45 a4! (but not 45 b5 � xa2 46

Kan Variation 109

!I b l 1! d2! 47 '*fl .1: f2 + 48 <f1el :!:! xg2 49 :!:! b4 K d2 50 b6 H d8 5 1 b7 1! b8 when Black has drawing chances) H b2 46 b5 g a2 47 b6 il xa4 48 H b1 H d4 49 *fl ! (49 b7 e2) H d8 50 b7 .1: b8 5 1 *e2 etc. Obviously Black has other ways to play, but the two con­nected passed pawns give White good winning chances in any case.

7 �aroczy Bind

This most commonly arises if Black plays an early g6, for example 1 e4 c5 2 �f3 �c6 3 d4 cxd4 4 .t:l xd4 g6, aiming to reach a Dragon position without having played d6. This restricts White's options quite severely, since he has to be careful not to allow Black to play d5 in one go, saving a crucial tempo. An effective way out of this dilemma is to continue 5 c4, setting up the formation of pawns on c4 and e4 known as the Maroczy bind. The asset of this formation is the auto­matic restraint of . . . b5 and d5, Black's basic freeing thrusts. Black does sometimes succeed in orgamzmg b5, but this is nor­mally only good when White has made a mistake. White's main asset is his space advantage, lead­ing to the corollary that he should avoid exchanges which would re­lieve the cramp in Black's posi­tion. If Black does succeed in liquidating to an ending, White's c4 pawn and black-squared weak­nesses can become a liability. Play often becomes a matter of slow manoeuvring as White tries to in­crease his space advantage and force weaknesses in the Black position while his opponent re­mains crouched on his back two

ranks waiting for the first sign of over-extension to launch a coun­terattack. The Maroczy Bind can also occur if Black adopts an unusual move order, for example I e4 c5 2 � f3 �c6 3 d4 cxd4 4 .t:l xd4 d6 and now 5 c4 will proba­bly transpose to this chapter after 5 .t:lf6 6 .t:lc3 g6. Although the Maroczy Bind is slightly passive for Black, players such as Larsen, Petursson and Velimirovic have shown that by patiently waiting for a lapse of concentration from White this line can offer winning chances for Black. The theoretical opinion is that White should maintain a slight advantage, but White players should not believe that this is a line in which they cannot lose.

Game 23 Karpov-Kavalek

Nice Olympiad 1974

1 e4 c5 2 �f3 �c6 3 d4 cxd4 4 �xd4 g6 5 c4

Here Black has two possibili­ties. He may play 5 ll.g7 in order to force White's jj_e3, but in doing this he forfeits the chance to take on d4 at a moment when

White must recapture with the queen. The alternative is 5 � f6, which will often transpose to 5 . . . !/,tg7 if Black does not take up the chance to play �xd4. The 5 !/,tg7 systems are exa­mined in game 24, while in this game we look at 5 LE� f6.

93 B

5 �f6 6 �c3 (93)

6 d6 If Black adopts the move order

6 � xd4 7 �xd4 d6 (possibly to avoid 7 �c2 as in the next note) then White has an interest­ing alternative based on playing j_d3 rather than j}_e2, providing the e4 pawn with useful extra pro­tection, for example 6 �xd4 7 'i!)fxd4 d6 8 j}_g5 !/,tg7 9 �d2 0-0 (9 !/,te6 1 0 � cl � c8 1 1 b3 *a5 1 2 f3 h6 1 3 jte3 0-0 1 4 !/,td3 \lih7 1 5 0-0 a6 16 h3 � d7 17 f4 f5 1 8 exf5 J_xf5 1 9 !it e2! was good for White in Polugayevsky­Belyavsky, USSR Ch. 1 975) 1 0 j}_ d 3 ( 1 0 f3 i s also interesting, for example after 1 0 !it e6 1 1 !ii c I �a5 1 2 b3 a6? 1 3 �d5! �xd2 + 1 4 \lixd2 !/,txd5 1 5 cxd5 �fc8 1 6

Maroczy Bind 111

K xc8! K xc8 17 g3 \lif8 18 !/,th3 K c7 19 � c l � xc l 20 *xc l White was winning i n Byrne-Gar­cia Padron, Torremolinos 1 977-1 2 K fc8 was better when 1 3 !/,te2 a6 1 4 �a4 transposes to Karpov-Kavalek) !/,t e6 ( 1 0 a5 1 1 0-0 a4 1 2 K ac l J.. e6 1 3 'lli!'c2 gave White his usual space advan­tage in Portisch-Reshevsky, Petropolis 1973) 1 1 � c l �a5 1 2 0-0 � fc8 1 3 b3 a6 1 4 � fe l *f8! ( 14 b5? 15 -2ld5! 'Wfxd2 1 6 !/,txd2 !/,txd5 1 7 exd5 bxc4 1 8 !/,txc4 \li f8 1 9 b4 with a clear plus for White, Geller-L. Garcia, Bogota 1978) 1 5 Zi c2 ( 1 5 f4 is Jansa's double-edged suggestion) b5 1 6 cxb5 axb5 1 7 � ee l b4 1 8 � b5 � xc2 1 9 � xc2 with just an edge for White, Polugayevsky­Jansa, Sochi 1 974.

7 !it e2 -21 xd4 This is Black's last chance to

force White to recapture on d4 with his queen. If he plays 7 !/,tg7 White should transpose to game 24 by 8 j{_e3. At one time 8 -2lc2 was thought the best reply to 7 . j_ g7, but after 8 �d7 9 j_d2 a5! (not 9 �c5?! 1 0 b4 -2le6 1 1 � c l 0-0 1 2 0-0 f5 1 3 exf5 gxf5 1 4 f4 �ed4 1 5 c2l xd4 �xd4 1 6 J.. e3 with a positional advan­tage for White, Nunn-Rind, Man­chester 1 980) 10 0-0 0-0 1 1 if;\ c I �c5 Black has a much better ver­sion of the Maroczy Bind than in other lines. In Nunn-Petursson, Wijk aan Zee 1 990 I made matters worse by 1 2 b3? -21 b4! and Black had a clear advantage.

112 M aroczy Bind

Or:

8 �xd4 �g7 9 �g5 (94) �e6

( I ) 9 . . . �d7 10 �d2 h6 (after 1 0 ;Lc6 White should continue with 1 1 f3) l l J,. f4 a6 1 2 0-0 J,.c6 1 3 f3 0-0 14 a4 <E) d7 1 5 a5! <E.)c5 1 6

94 B

g a3 with a small advantage for White, Pomar-Cordovil, Malaga 1 972.

(2) 9 . . . h6 1 0 �e3 0-0 1 1 �d2 *h7 ( 1 1 �a5 12 0-0 forces . *h7 in any case) 1 2 O-OJ..e6 1 3 f4 ( 1 3 �d4 � c8 1 4 b3 a6 1 5 ·l!i!fe3 <E.)d7 1 6 �xg7 *xg7 1 7 f4 �b6 1 8 �xb6 �xb6 1 9 f5 �d7 20 � ad l only gave White a slight edge in Timman-Ribli, Amsterdam 1 973) li;l c8 (after 1 3 . . . �a5 14 f5 forces 1 4 . �d7 since 1 4 gxf5 1 5 exf5 �xf5 allows 1 6 � xf5, so White avoids wasting a tempo on b3) 14 b3 �a5 ( 1 4 . . . a6 1 5 � ad l �a5 1 6 � d4 is good for White after both 16 . . . b5 1 7 f5 jtd7 1 8 �xf6! exf6 1 9 <E.)d5 l!i'xd2 20 � xd2, Nunn-van der Sterren, Groningen 1 974-5 and 16 . . . �d7 1 7 J,.xf6 exf6 1 8 �d5 �c5 + 1 9 * h i a5 20 f5, Nunn-1. Ivanov,

London 1 987, so it makes sense for Black to try to force White's rook to the less active square c l ) 1 5 a3 ( 1 5 � ac l i s also possible, although 1 5 . . . a6 I6 f5 �d7 1 7 h3 �c6 I 8 jtd3 �d7 was unclear in Gulko-Petrosian, Biel IZ 1976) a6 I 6 f5 �d7 1 7 b4 �e5 (Larsen­Fischer, match 1 97 1 ) and now, according to various analyses of this famous match, 1 8 � ad l would have been good for White.

10 li;l c1 (95) 1 0 0-0 is possible, and after 1 0 0-0 I I 'l'td2 a6 ( 1 1 � c8 1 2

b3 b5?! is doubtful because of I 3 e5! when 13 . . . dxe5 1 4 'i!ifxd8 � fxd8 1 5 �xb5 gives White the better ending and 13 . . . b4 1 4 exf6 exf6 1 5 _2_ e 3 bxc3 1 6 ')$'xc3 f5 1 7 .k_d4 jtxd4 1 8 'ii!{xd4 gives White the better middlegame, Bukic-Romanishin, Moscow 1 977) 1 2 f3 (not 1 2 � c l allowing 1 2 b5! with equality) �a5 1 3 � fd l � fc8 1 4 � d5 'iti'xd2 1 5 � xd2 ,kxd5 ( 1 5 �xd5? 1 6 exd5 �d7 1 7 J,.xe7! J..h6 and now not 18 � c2 J..f5 and 1 9 � e8, but 1 8 � ad1! 1i_xd2 19 � xd2 � e8 20 J..xd6 with a clear plus for White, Britton-Donald­son, Rhodes 1 980) 1 6 cxd5 w f8 White can claim a very slight plus. In general these endings with an open c-file are very drawish unless Black has weakened his queenside by playing . b5 (which both allows a4 and gives White an entry point at c6), or White can quickly seize the c-file by playing his bishop to the h3--c8 diagonal.

Neither situation exists here, so White's advantage is insignificant and in practice Black would have few problems reaching the draw.

10 �aS If Black omits this move, we

again face move order questions. After 1 0 0-0 1 1 '*"d2 ili c8 1 2

b3 a6 the obvious 1 3 f3 allows 1 3 b5!? 1 4 cxb5 axb5 1 5 2!xb5

g xc 1 + 1 6 �xc 1 i:iia5 + 1 7 'M'd2 g a8 18 �xa5 g xa5 19 a3! .�xb3 20 fTf2 jj_a4 2 1 � b 1 , Tukmakov­Vaganian, USSR 1 984, and now 2 1 jj_xb5 22 � xb5 g xa3 is a simple draw. Therefore White should prefer 13 0-0, when 13 . . . �aS 14 f3 transposes to the next note, while 13 . . . bS 14 cxb.5 axb5 1 5 jj_xb5 '*a5 ( 1 5 g xc3 1 6 -�xc3 �xe4 1 7 1!i1e3) 1 6 jj_d3 also seems good for White, e.g. 16 g xc3 17 �xc3 �xc3 1 8 g xc3 �xe4 1 9 jj_xe4 jj_xc3 20 jj_xe7.

11 'i!i!t'd2 0-0 After 1 1 g c8 12 f3 ! B lack is

in a rather awkward situation since he has the wrong rook on c8 if he wants to castle. The p oint is that with the f-rook on c8 White is

Maroczy Bind ll3

never threatening �d5, because after the sequence �xd2 �xe7 + 1tt f8 White just loses his knight. With the a-rook on c8, however, Black will sooner or later have to waste time meeting this threat. So 12 . . . 0-0 1 3 b3 a6 1 4 0-0, for example, is better for White than the positions in the note to 1 0 .l c l . It is also too dangerous to take the c-pawn, for example 12 . . . jtxc4 1 3 � d5! tf'xa2 1 4 0-0 .:E)xd5 1 5 g xc4! � xc4 1 6 'l!!fxd5 l! a4 1 7 jtb5 + *f8 1 8 � c l ! and in Geller-Stean, Teesside 1 975 Black resigned because of 1 8 jj_d4 + 1 9 '*'xd4! � xd4 20 Ah6 + mating. In fact White had an even more convincing win by 14 � b4, since 14 't)ltb3 1 5 Ad1 traps the queen.

If Black doesn't castle he soon runs out of things to do, e.g. 1 2

a6 1 3 b3 b 5 1 4 � d5! (as men­tioned earlier, b5 makes this a much better proposition) tf'xd2 + 1 5 \fixd2 Axd5 1 6 cxd5 fTd7 1 7 a4 h6 1 8 jte3 � xc 1 1 9 � xc 1 l! b8 20 H c6 with a very good ending for White, Nunn-Reuben, Lon­don 1 978.

12 f3 � fc8 13 b3

By securely defending c4 and e4 White has prevented any tricks based on an immediate b5 so Black has nothing better than to prepare this thrust with a6.

13 a6 14 � a4

White chooses a favourable

1 14 M aroczy Bind

moment to exchange queens. Black's last move weakened b6 and he must waste a tempo pre­venting White's knight fork.

14 �xd2 + After 1 4 iitd8 White may

either play 1 5 c5 (as suggested by Karpov) based on the idea 1 5 . . dxc5 1 6 �xd8 + � xd8 1 7 ..£) b6 when 1 7 � ab8 allows 1 8 J,. f4, or continue more quietly by 1 5 _zte3 � ab8?! ( 1 5 . . ..£)d7) 1 6 J,. a7 � a8 1 7 .ztb6 iltf8 1 8 J..e3 � ab8 1 9 ..£)b6 � c7 20 0-0 ..£)d7 21 ..£) d5 with an excellent position, Nunn­Blum, London 1 979.

96 B

15 fj>xd2 (96)

15 ]l c6 After Black lost with this move

in Karpov-Kavalek attention turned to 1 5 ..£) d7, but this doesn't seem to be any better, for example 1 6 g4 fJ f8 1 7 h4 ll( c6 18 Mt c2 <E)c5 1 9 ..£)c3 a5 20 <E)d5 H e8 2 1 J,.e3 J,.c8 22 h5 e6 23 � c3 f5 24 hxg6 hxg6 25 exf5 exf5 26 gxf5 J..xf5 27 � ee l , Averbakh-Popov, Polanica Zdroj 1 976, or 1 6 h4 (this is perhaps even stronger) fJ f8 1 7 h5 h6 ( 1 7 . . ll( ab8 18

hxg6 hxg6 1 9 '£)c3 is also good for White) 1 8 .zte3 g5 19 g3 � cb8 20 ..£)c3 b5 2 1 �d5!, Psakhis-Pigu­sov, USSR 1 980 when in both cases White had a good ending. It is curious that the exchange of queens is just what White needs to start a kingside attack by h4-h5. The explanation is that Black's counterplay by b5 would be very dangerous with queens on the board, since it would lead to an attack against the centralized white king. With queens off this counterplay is relatively harmless.

16 '£)c3 � ac8 Karpov suggested 1 6 � e8

as a possible improvement, so as to trap White's bishop in case of 1 7 '£) d5 -2\d7 1 8 ..£) xe7 + ?! '1J, xe7 19 .iJ..xe7 f6 20 _id8 b6. 1 6 � cc8 was played i n Sakharov­Pereira, corr. 1976, which finished in a draw after 1 7 ..£)a4 � c6! I don't suppose Karpov would have agreed a draw if Kavalek had 'found' 1 6 � cc8!

17 '£)d5 *f8 18 1l_e3 '£)d7

Defending such an ending is an unpleasant task at the best of times, doubly so against Karpov. White has the choice of expanding on the queenside by a timely b4, or of gaining space on the other flank by g4 and h4, as in the note to Black's 1 5th move. Until White shows his hand Black can only wait.

19 h4 j_xd5 Black resolves to do away with

the dangerous knight. 1 9 . . h5

was well met by 20 <EJ f4 and Kar­pov's suggestion of I9 f5 would require strong nerves in view of Black's king position.

20 exd5 � 6c7 21 h5 wgS?! (97)

This move is probably a mis­take. 2I . li!\ e8 followed by . . e6 would have opened the posi­tion up for White's two bishops, but by activating his rooks on the central files Black would have d e­veloped counterplay again st White's king.

22 f4! Most players would have

rejected this as it allows Black's knight to settle at e4 (supported by f5). Karpov, however, is actually aiming to provoke f5, which gives him the lever g4 by which he can prise open Black's kingside.

22 <EJc5 23 jj_g4 <EJe4 + 24 wd3 rs 25 jj_f3 b5

It looks as though Black's counterplay has got off the

Maroczy Bind 1 15

ground at last but White defuses it adroitly.

26 g4 bxc4 + 27 g xc4 H xc4 28 bxc4 <2Jc5 +

Black had little choice as he could not allow the white rook to occupy the b-file, nor could he play � b8 without losing a pawn at e4.

29 j_xc5! If there are rooks on the board

opposite coloured bishops tend to lose their drawish influence. Here Black runs into trouble because his king is badly placed and he will have two or even three pawns stuck on white squares, where they cannot be defended by his bishop.

29 � xc5 Black plays for a counterattack

by � a5. 29 dxc5 30 h6 jj_ d4 3 I � b I is also unpleasant, the a6 pawn being particularly weak.

30 h6 j_f8 It looks horrible to bury the

bishop but Black lacked a reason­able alternative, for example 30

fxg4 (30 . . . J..f6 3 I lilt b i threatens 32 gxf5 gxf5 3 3 � b8 + wfl 34 j_ h5 mate) 3 1 j_xg4 lt.f8 (3 I J..f6 32 lt.e6 + wf8 33 � bi ) 32 j_e6 + wh8 33 f5 ill( a5 34 g b I � a3 + 3 5 * e2 � xa2 + 36 \\'fl j_xh6 37 f6 and the pawn slips through (37 exf6 38 � b8 + wins a piece).

31 wc3 Karpov also analyses 3 I g5

� a5 22 � b i as good for White,

1 16 Maroczy Bind

but the variations are by no means simple and in practice it is not surprising that he chose to prevent 3 1 . . . lil a5 by simple means (32 \91 b3, and the rook has to go back).

98 B

31 fxg4 32 jj_xg4 (98)

32 Black resolves to extract his

king, even at the cost of the h7 pawn. Despite its dangerous appearance, he would probably have done better to try 32 . . � c7 33 j)_e6 + f1Jh8 34 f5 � b7! (pre­venting 35 M b l j)_xh6 36 f6 exf6 37 SI, b8 + ), when White finds it hard to make progress because of Black's attack on the h6 pawn.

33 Ae6 + f1Jf6 34 jj_ g8 11 c7

34 jtxh6 35 � xh6 \91 g7 leads to a lost rook and pawn ending after 36 � xh7 + and 37 � xe7.

35 j)_xh7 e6 35 'IJfl is refuted by 36 f5 g5

37 f6! exf6 38 Af5 \91 g8 39 h7 + f1Jh8 40 i: b l and 4 1 l! b8.

36 jj_ g8 exd5 37 h7

Not 37 jj_xd5? � h7. 37 jj_g7?

Loses by force. 37 11 xc4 + 38 fld3 jj_g7 39 jj_xd5 (39 h8( "ltf) Axh8 40 11 xh8 )!;. c8 and 4 1 . . . \91 g7) � c8 leads to the same posi­tion as the game but with White having a pawn less. Black would still be worse, but he would have chances of a draw.

38 j)_xd5 jj_h8 39 ftd3 \91f5 40 f1Je3 � e7 + 41 \9113 a5 42 a4 � c7 43 jj_e4 + f1Jf6 44 � h6 � g7

44 ftg7 45 � xg6 + f1Jxh7 46 � g l + *h6 47 � h i + and 48 iii\ h7 + wins the rook.

45 * g4 Resigns Black is totally paralysed.

Game 24 Vaganian-lvkov

Moscow 1985

1 e4 2 Q}f3 3 d4 4 Q}xd4 5 c4 6 j)_e3 (99)

c5 Q}c6 cxd4 g6 jj_g7

6 Q}f6 Black can also play Q}h6

followed by . . . f5, either with or without d6, but this idea does not equalize:

( I ) 6 . . . d6 7 e,c3 � h6 8 jte2 0-0 9 0-0 f5 1 0 exf5 gxf5 ( 10 . Q}xd4 1 1 j_xd4 j_xd4 1 2 '*xd4

99 B

� xf5 1 3 '¥'d2 jj_d7 was good for White after both 1 4 1i_f3, Tal­Kupreichik, Sochi 1 970 and 1 4 j_ g4, Vilela-Estevez, Cienfuegos 1980) 1 1 f4 '®'b6 ( 1 1 _itd7 1 2 �d2 c2l g4 1 3 _itxg4 fxg4 14 c2l d5 is a little better for White, Szabo­Larsen, Vinkovci 1 970) 12 c2lxf5 �xb2 1 3 c2lxh6 + jj_xh6 14 � c l 1i.._g7 1 5 � c2 ( 1 5 �d5 sacrificing the a-pawn was possible) -�a3 16 'i#d2 with an edge for White, Spassov-Nicevski, Sofia 1 976.

(2) 6 . . . �h6 7 �c3 0-0 8 jl e2 f5 9 exf5 jj_xd4 1 0 Ji_xh6 � xf5 1 1 0-0 d6 ( 1 1 ·i!lJ' b6 is met by 1 2 �d5!) 1 2 _itf3 ( 1 2 '¥'d2 'i#a5 1 3 <i1J h 1 � f7 1 4 f4 was also promis­ing in Shamkovich-Vasyukov, USSR 1 965) _itg7 1 3 jj_e3 1i.._d7 14 g e l b6 1 5 1i.._e4 �f7 1 6 1i_ g5! 1Lf6 1 7 .ii_ xf6 � xf6 1 8 �d2 'i!lff8 1 9 g ad l g d8 20 c2ld5 with a clear plus for White, Kudrin-L lvanov, New York 1 983.

In this note we have seen a pawn structure with white pawns on c4 and f4 against black d-, e­and f-pawns. This structure arises frequently in the Maroczy Bind

Maroczy Bind 1 1 7

and i t i s almost always good for White. Black's problem is that any central pawn advance leaves him with either a backward pawn or hanging pawns, while if the pawns stay where they are White can just build up pressure down the d- and e-files. Similar com­ments apply in the case where Black plays f5 and recaptures on f5 with a piece. The hanging pawns are a more important fac­tor than the temporary piece ac­tivity Black obtains.

7 �c3 0-0 Black has a major alternative in

7 c2lg4, which has recently become more popular. This new respectability has been based partly on an original idea for Black involving kingside pawn expansion, and partly on a reali­zation that the older lines are not so bad for Black as had been thought. White has his typical space advantage, but Black's position is solid and Larsen in particular has achieved quite good results for Black.

After 7 c2lg4 8 �xg4 �xd4 (or 8 J.,xd4 9 j_xd4 Q) xd4 1 0 0-0-0 e 5 1 1 '¥' g 3 d 6 1 2 f4 f6 1 3 f5! '11Jf7 1 4 � b5 �xb5 1 5 cxb5 with an excellent position for White, Mestel-Karlsson, Las Palmas 1 982) 9 �d I (100) Black has three possibilities:

( 1 ) 9 . . . e5 1 0 Jl,d3 (this gives White a positional advantage with no risk, but there seems nothing wrong with the older tac­tical line 1 0 Q) b5 0-0 1 1 �d2!

118 Maroczy Bind

/00 B

�h4 1 2 J..d3 d5 1 3 cxd5 <EJxb5 14 J..xb5 �xe4 15 0-0 � d8 1 6 � fd l when 1 6 j_e6 fails to 1 7 f3 �xd5 1 8 'iife2 trapping Black's queen) 0-0 1 1 0-0 d6 1 2 �d2 J..e6 ( 1 2 f5 1 3 exf5 gxf5 14 f4 <EJ c6 1 5 � ad l �e7 1 6 ..§.e2 <EJd4 1 7 ..§.xd4 exd4 1 8 <EJd5 gave White a clear plus, Andersson-Rogers, Malta 1 980) 1 3 l;l ac l a6 14 b3 � c8 1 5 f3 and now both 1S . . . fS 1 6 exf5 gxf5 1 7 f4 �f6 1 8 <EJe2 � cd8 1 9 <EJxd4 exd4 20 J..f2, Ghi­tescu-Radovici, Romania Ch. 1 977, and 1S . . . �aS 16 � fd l f5 1 7 exf5 <2:1 xf5 1 8 A e4, Tal-Partos, Nice 1 974, were very good for White.

(2) 9 . . . <EJc6 l 0 �d2 'iifa5 1 1 � c l 0-0 1 2 J..e2 d6 1 3 0-0 j_e6 1 4

b 3 � ac8 1 5 f4 with a good posi­tion for White, Polugayevsky­Suetin, Kislovodsk 1 972.

(3) 9 . . . <2:Je6 (the main line) 1 0 � c l ( 101) and now:

(3a) 10 . . . �aS and now it is unclear whether the bishop should be developed at e2 or d3:

(3a l ) 11 J..d3 with a further branch :

(3a l l ) 1 1 . . . ..§.xc3 + 1 2 � xc3

/0/ B

�xa2 1 3 '(l1cl �a5 1 4 c5 is ex­tremely dangerous for Black.

(3a 1 2) 1 1 . . . d6 1 2 0-0 (better than 1 2 ·�d2 1l_ d7 1 3 0-0 ..§.c6 14 � fe l 0-0 1 5 1l_h6 �e5! 16 j_ xg7 ·fiixg7 1 7 � cd l <EJc5 1 8 1l_fl a5 with equality, Nogueiras-Korch­noi, Montpellier 1 985) 0-0 1 3 j'j_ b l j_d7 1 4 f4 <EJc5 1 5 <EJ d5 with advantage to White in Mednis-D. Byrne, US Ch. 1 973.

(3a l 3) 1 1 . . . b6 12 0-0 1l_b7 1 3 f4!? (probably better than 1 3'(l1d2 g5 14 � fd l d6 when 1 5 f3 j)_e5 1 6 *hl J..f4 1 7 j)_xf4 <EJ xf4 1 8 j)_fl <EJe6 19 a3 ·;we5 20 <EJ d5 h5 2 1 b4 f!'f8 22 � el � c8 was equal in Popovic-Cebalo, Yugoslavia 1 988, and 1 5 a3 h5 1 6 � c2 j_d4 1 7 b4 �e5 1 8 <EJd5 it_xe3 1 9 fxe3 � c8 20 � fl <2:Jg7 2 1 �f2 f6 was unclear in Ljubojevic-Korchnoi, Tilburg 1987) 0-0 14 .it. b l d6 1 5 � f2 � ac8 1 6 <EJd5 jtxd5 1 7 exd5 <2:Jc5 1 8 a3 and White is better, A. Rodriguez-Hernandez, Cuba Ch. 1 988.

(3a2) 1 1 it_e2 b6 12 0-0 1l_b7 1 3 f3 g5 (Larsen's plan increases the black-squared pressure and re­serves e5 for the queen, but the

obvious danger is that Black's king has to stay in the centre) 1 4 � f2 ! ( a number o f other games had continued with �d2, but the rook transfer to d2 appears the best way of meeting Black's dou­ble-edged plan) h5 1 5 J.. fl '*'e5 1 6 J4 d2 d6 1 7 �d5 ft f8 1 8 b4 jj_ h6 1 9 �b3 g4 20 jtxh6 + � xh6 2 1 'iiJe3 ·f!!g7 22 f4 with a distinct advantage for White, Short-Lar­sen, Hastings 1987/8.

(3b) 10 . . . b6 1 1 _!d3 ( 1 1 b4 is also good, for example 1 1 il_b7 1 2 J:,d3 0-0 1 3 0-0 �d4 1 4 j_ b l �c6 1 5 a 3 d 6 1 6 �d3 � c8 1 7 f4 and White has consolidated his space advantage, Suba-Tai­manov, Bucharest 1 979, or 1 3 � c8 1 4 f4 with attacking chances for White) j_b7 1 2 0-0 ·iSoi'b8 (Black intends a variant of Lar­sen's plan to dominate the black squares on the kingside; normal development would lead to posi­tions similar to Suba-Taimanov above) 1 3 �d2 �d6 14 �d5 g5 1 5 b4 h5 1 6 .!!l[ fd l j1e5 1 7 h3 .ztf4 1 8 jt fi ! and Black's attack has become bogged down while White has all sorts of threats against Black's king and queen, Mocha­lov-Kapengut, USSR l st League 1 976.

(3c) 10 . . . d6 1 1 b4! (when Black's knight is on e6 White should in general aim to play b4 as quickly as possible, preventing Black cementing his knight on c5 by a5) 0-0 1 2 j_ e2 and now:

(3c l ) 12 . . . b6 1 3 �d2 ( 1 3 0-0 may be better, for example 1 3 _ . .

Maroczy Bind 1 19

jtb7 1 4 �d5 �d7 1 5 il..g4! f5 1 6 il_h3 �c7 1 7 �xc7 "jJxc7 1 8 exf5 gxf5 1 9 c5! with a fine game for White, Adorjan-Larsen, Hastings 1 986/7) jib7 14 f3 ( 14 <E:Jd5 <E:Jc7! was played in Rogers-Hernandez, Calcutta 1 988, and now 1 5 0-0 e6 1 6 <E:Jxc7 �xc7 1 7 f3 would have kept an edge for White) f5 1 5 exf5 gxf5 1 6 <E:Jd5 l';;l f7 1 7 0-0 <E:J f8 1 8 ll: fd 1 with advantage to White, Smejkal-Radulov, Skara 1 980.

(3c2) 12 . . . a5 1 3 a3 axb4 1 4 axb4 il_d7 ( 1 4 l';;l a3 1 5 <E:Jd5) 1 5 0-0 jtc6 1 6 "1Jd2 � a3 ( 1 6 jtxc3 1 7 �xc3 jixe4 fails to 1 8 jt h6 g e8 1 9 lil\ ce l followed by jtg4 with a catastrophe at g7) 1 7 <E:Jd5 * h8 1 8 1tb6 �d7 1 9 f4 with a fine position for White, Por­tisch-Pfteger, Manila 1 974.

(3d) 10 0-0 1 1 b4 will quickly transpose into 3b or 3c.

8 jie2 (102) d6

Black may try to do without this move:

( 1 ) 8 . . . a5 9 0-0 a4 1 0 c5!? (an attempt at outright refutation; 1 0 <Z�db5 would b e similar to 9 . . . a5 below) d5 1 1 cxd6 'filxd6 1 2 <E:Jdb5

120 Maroczy Bind

'it b4 ( 1 2 \'txd I 1 3 � axd I gives White some endgame ad­vantage) 1 3 a3 'lta5 14 f4 e5 1 5 fxe5 <2l xe5 1 6 � xf6! A xf6 1 7 <2ld5 j_d8? ( 1 7 g a6! 1 8 � c l ! J..d7 1 9 � c5 �d8 is better, although White has an ominous initiative) 1 8 J,.d4! f6 19 J,.c3 \'ta6 20 <2l bc7 �a7 + 21 J..d4 �b8 22 <2lxa8 with a clear plus for White, Nunn-Haik, Paris 1 983.

(2) 8 • • • b6 (an important alter­native) 9 0-0 J,.b7 1 0 f3 (when Black develops his bishop at b7 the extra protection of the e-pawn afforded by f3 is usually a good idea) and now Black has an exten­sive range of possibilities:

(2a) 10 . . . �h5 1 1 '£) xc6! 11_xc6 ( 1 1 dxc6 1 2 c5 is good for White) 1 2 � c l f5 1 3 exf5 gxf5 14 f4 <2lf6 1 5 J..f3 � c8 1 6 b3 �e8 17 'E)d5 '/iiff7 1 8 '£) xf6 + �xf6 1 9 � c2 J..xf3?! 20 � xf3 d 6 2 1 � d2 lfJ>h8 22 j_d4 gave White his stan­dard favourable positiOn m

Nunn-Ristoja, Malta 1 980. (2b) 10 . . . d6 (this is inconsis­

tent with the choice of b6) 1 1 �d2 '/iild7 1 2 a4!? e6 1 3 � fd l !ill fd8 14 12Jxc6 � xc6 1 5 a5 bxa5 16 <2l b5 with unpleasant threats to d6, a7 and a5, Gheorghiu-Bel­lon, Las Palmas 1 976.

(2c) 10 . . . � c8 1 1 'ii!t'd2 12!h5 ( 1 1 � e8 1 2 i! ac l 'W!c7 1 3 b4! <2lh5 14 <2lxc6 j_xc6 1 5 1£)d5 '/iifb8 1 6 f4 <2:J f6 1 7 J,.f3 d6 18 j_d4 was very good for White in Nunn­Karlsson, Helsinki 1 98 1 ) 1 2 � fd I Qje5 1 3 b3 f5 14 exf5 gxf5 1 5

12!d5! \'te8?!, Kir. Georgiev-Kris­tensen, Saint John Open 1 988, and now 16 f4! J,.xd5 ( 1 6 . . . <2:Jg4 1 7 Qjxf5 Qjxe3 1 8 12Jxg7 <2:Jxg7 1 9 "*'xe3 wins) 1 7 cxd5 d2lg4 1 8 <2lxf5! J,.xa l 1 9 11_xg4 !2l f6 ( 1 9 J,.f6 20 d6! e 6 2 1 IE:le7 + JJ...xe7 22 dxe7 � f6 23 �xd7 wins and 1 9

j_c3 20 �d3 !2lg7 2 1 d6 are no better) 20 <2:J h6 + *g7 2 1 � xa l *xh6 22 f5 + '#; g7 23 j_h6 + '\fi'h8 24 JJ...xf8 IE:lxg4 2 5 �d4 + l£) f6 26 JJ...h6 wins for White.

(2d) 10 . . . IE:lxd4 1 1 J,.xd4 d6 1 2 'ii!t'd2 <2:Jd7 1 3 JJ...xg7 ( 1 3 JJ...e3 is promising) '#;xg7 14 f4 � c8 1 5 � ad I IE:l f6 1 6 e 5 dxe5 1 7 fxe5 IE:lg8 18 'i!ii'e3! with some advan­tage for White, Cvetkovic­Cebalo, Yugoslavia 1 985.

(2e) 10 . . . IE:le8 1 1 "i!Wd2 'E)c7 12 B, ad l IE:l e6 13 IE:ldb5 d6 14 IE:ld5 li!\ b8 1 5 f4 <2:Jc5 1 6 JJ...f3 a6 1 7 IE:ld4 12:! xd4 1 8 J,.xd4 jj_xd4 + 1 9 \'txd4 b5, Agapov-Kimelfeld, USSR 1 985, and now 20 e5 would have kept some advantage for White.

(2f) 10 . . . e6 1 1 "i!Wd2 ( 1 1 IE:ldb5 is probably also good, e.g. 1 1 d5 1 2 cxd5 exd5 1 3 exd5 IE:le7 14 d6 <2:J f5 l 5 JJ...f2 IE:le8 16 d7 IE:l f6 17 g4 IE:le7 1 8 . .\lh4 with advantage, A. Rodriguez-Pinal, Sagua la Grande 1 984) d5 1 2 l£) xc6 J,.xc6 1 3 cxd5 exd5 14 e5 IE:ld7 1 5 f4 �c5 1 6 � ad l ! f6 ( 1 6 IE:le4 1 7 IE:lxe4 fxe4 1 8 �d6! j_ b7 1 9 lt_c4 is also unpleasant) 1 7 12Jxd5 fxe5 18 j_c4 *h8 19 fxe5 � xfl + ( 1 9

j_xe5 20 j_d4 '/iiid6 2 1 j_ xe5 + '/iil xe5 2 2 � fe I .£)e4 23 � xe4! Resigns, Kuporosov-Jak-

ovic, USSR 1984) 20 J;i xfl �h4 2 1 � f4! �h5 22 ttJe7 J.. b7 23 b4 ttJe4 24 �d7 � b8 25 ttJ c8! � xc8 26 �xb7 � d8 27 �xe4 j_ xeS 28 J..e2 Resigns, Kuporosov-Malis­hauskas, USSR 1 985.

9 0-0 (103)

9 J..d7 Or: ( I ) 9 . . . aS (certainly not 9

ttJg4 losing a piece after 1 0 j_xg4 J.. xg4 1 1 tjjxc6) 10 f3 ttJd7 1 1 ttJdb5 tjjc5 1 2 �d2 a4 1 3 � fd l �a5 14 � ac l J,.e6 1 5 ttJd5 �xd2 16 .!;! xd2 j_xd5? ( 1 6 � fd8 is only slightly better for White) 1 7 cxd5 ttJ b4 1 8 .it xc5! dxc5 1 9 J..xc5 tjjxa2 20 j_xe7 and Black is in trouble, Andersson-Larsen, Linares 1 983.

(2) 9 . • . � e8 (another Larsen idea) and now:

(2a) 10 � b 1 a6 1 1 �d2 .ii_d7 1 2 l! fd I � c8 1 3 f3 ttJxd4 14 .ii_xd4 j_e6 1 5 ttJd5 ttJxd5 1 6 cxd5 j_xd4 + 1 7 �xd4 J.. d7 1 8 � be l �a5 with equality, Speelman­Larsen, Hastings 1 988/9.

(2b) 10 f3 ttJd7 1 1 'ii((d2 ttJc5 (taking twice on d4 is also pos-

Maroczy Bind 121

sible) 12 .l fd l '*'a5 13 � ab l ttJxd4 14 J..xd4 J..xd4 + 1 5 �xd4 ttJe6 16 �f2 J..d7 and White has no advantage, Andersson-Lar­sen, Naestved 1985.

(2c) 10 a3 J..d7 1 1 f3 a6 12 b4 � c8 1 3 � c l ttJxd4 14 j_ xd4 J..h6! 1 5 � c2 j_e6 and Black is at least equal, Short-Larsen, Naestved 1 985.

(2d) 10 �d2 <E\g4, 1 0 � c l ttJxd4 1 1 J..xd4 J..h6 and 1 0 f4 e5 are other points of Larsen 's move, so what is the best answer? Proba­bly White should play 1 0 ttJc2, avoiding the knight exchange and preparing solid development by .!;! c l .

10 �d2 It is also possible to move the

knight on d4, thereby frustrating Black's plan of ttJxd4 and J..c6. In Korchnoi-Soos, Rome 1 982, White continued 10 ttJb3 ttJa5 (10 a5 1 1 a4 ttJ b4 12 f3 jj_c6 1 3 �d2 � c8 14 * h l ttJd7 15 ttJd5 was also better for White, Tarjan-Strauss, USA 1 982) 1 1 0 €)xb3 1 2 axb3 a6 1 3 b4 j_e6 14 �d2 .k c8 15 b3 ttJd7 16 � a2, while Schmidt-Kagen, Lucerne 1 982, went 1 0 ttJ c2 j_e6 1 1 '!li!(d2 a5 12 f4 a4 1 3 l! ab l j_g4 14 J..d3 J..c8 1 5 h3 <E\d7 16 .ii_e2 and in both cases White had a good posi­tion. It could well be that these lines are as strong as the traditio­nal continuations 10 J;,i; c l and 1 0 ·�!!i·d2, but they d o not seem to have been tested in recent years.

A separate question is whether White should play 1 0 � c 1 or I 0

122 M aroczy Bind

�d2. I have preferred the latter move for two reasons. Firstly, the rook on c l is sometimes vulner­able to unwelcome attacks by �xd4 and �h6, and secondly White often starts a queenside ad­vance by a3 and b4 later on. If Black plays a5 White needs a

rook on b 1 to force through b4. but playing the rook to c l and then to b 1 loses a tempo. The lines after � c 1 and "ii!id2 are quite simi­lar, so the material from the first edition of the book is also worth reviewing, but here we shall only examine the lines which follow 1 0 �d2.

10 <E:J xd4 10 l!l c8 1 1 f3 a6 1 2 g ac l

<E:Jxd4 1 3 �xd4 �e6 14 b3 is passive and White has a very com­fortable position, Smejkal-Diez del Corral, Skopje 01. 1 972.

1 1 � xd4 �c6 12 f3 a5

The move order is flexible, but White always answers . . . <E:J d7 by �e3 (to avoid freeing exchanges) and a5 by b3 (or else Black plays . . a4 followed by �a5, with an active position).

13 b3 <E:Jd7 Or 1 3 . l!l e8 ( 1 3 . . . <E:Jh5?! 1 4

�e3! f5 1 5 exf5 gxf5 1 6 f4 was good for White in Kavalek-Lar­sen, USA-Nordic match 1 986) 1 4 � fd l <E:J d7 1 5 � e 3 <E:J c 5 1 6 � acl (16 � ab 1 appears more consis­tent) 'lt'b6 1 7 <E:Jb5 li;! ec8 1 8 'lt'el ( 1 8 .:E:ld4 is better) �xb5! 1 9 cxb5 �h6 with equality, Arnason­Karlsson, Helsinki 1 986.

14 �e3 <E)c5 (104) 15 .;, ab1

In order to make progress White must expel the knight from c5, and so he needs to play a3 and b4. 1 5 g ab 1 appears the most logical, and this intuitive assess­ment is supported by the fact that after the alternative 1 5 )!g, ac l (this position can also arise via 10 g e l ) �b6 ( 15 f5 16 exf5 gxf5 1 7 <E)d5 g f7 1 8 � fd l b6! 1 9 �g5 .1. a 7 was unclear in Vaganian­Yudashin, USSR Ch. 1 988) 1 6 <E)b5 � fc8 1 7 � fd 1 �d8 1 8 <E:Jd4 ( 1 8 �fl �f8 1 9 <E)c3 b6 20 <E:Jd5 .l, ab8 21 g bl �e5 22 �h6 �g7 23 �g5 g b7 24 l! e l <E) e6 25 �e3 <E)c5! was level, Sax-Petursson, Reykjavik 1 988), Nunn-Velimir­ovic, Dubai 01. 1 986, Black can play 1 8 �d7 1 9 � b l <E)e6 20 <E:J xe6 �xe6 with equality (instead of 1 8 . . . 'lt'f8?! 1 9 I;K b 1 �f6 20 a3 'W/g7 21 b4 axb4 22 axb4 <E)e6 23 <E) xe6 fxe6 24 b5 �e8 25 f4 with a clear plus for White in the game).

15 'it'b6 Or 1 5 b6 (after 1 5 e6 1 6

� d l ! intending <E) e2-d4, a3 and b4 Black's panic reaction 1 6 . . . f5

1 7 exf5 � xf5 18 -2l e2 was good for White in Tringov-Haik, Vrnjacka Banja 1 986) 1 6 iLd l (White's problem is that the im­mediate a3 may be met by . . a4; the idea of iLd 1 is to be able to take the Black knight when it ar­rives at b3, but White has other ways to nullify . . a4, for example with the slow preparatory plan of � fcl-c2, .:;lfl and Wf2 to line up against the weak pawn on b6) i"<i b8 1 7 a3 � c8 18 2!d5 .k.xd5?! (probably bad, but White is slightly better in any case) 19 exd5 a4 20 b4 2!b3 21 !:ioi e2 with an excellent position for White, Anand-Larsen, Cannes 1989.

16 !i6 fcl After 1 6 -21 b5 (not 1 6 a3?

2!xb3) 8 fc8 17 � fd l � d8 1 8 -dd4 � f8 1 9 a3 ( 1 9 JLfl may be better, but I doubt if White has more than a tiny edge) . .ii.d7! (not 19 _li f6? 20 b4 axb4 2 1 axb4 2!e6 22 2!xe6 fxe6 23 f4 with the same advantage for White as in Nunn-Velimirovic above) 20 b4 axb4 21 axb4 2!e6 22 i±, a l J;t, xa l 23 &i x a 1 ;;,;, a8 24 � xa8 !:ioi·xa8 25 Bxe6 1Lxe6 Black is completely equal, Jansa-Petursson, Naestved 1 988.

16 £ fc8 17 � c2

Again this is useful preparation for a3 and b4. Now Black cannot meet 1 8 a3 by 1 8 � xb3 because 1 9 � d l ! wins material.

17 'i'ii'd8 There is no point to 17 �b4

because after 1 8 "M·c l Black will

M aroczy Bind 123

be driven back by a3 with great loss of time.

18 j_fl li..e5 White intends iioi fl, followed by

a3 and b4, so Black has to orga­nize some counterplay. With 1J. e5 he hopes to become active with . e6 and 'i'i'i h4, but the exposed bishop on e5 is a target which causes White to switch plans away from his queenside pawn advance to a more aggres­sive idea.

19 2!dl !? The knight transfer to h6,

which gains time along the way when Bg4 hits the bishop, is not a very thematic approach, but chess cannot always be played accord­ing to the recipe book.

19 'itie8 ( 105) Preparing b5 is the most

natural way to counter White's slow build-up towards a kingside attack. Moreover the queen defends the f7 pawn which might come under fire after B f2-g4-h6 +

20 �f2 b5?! It would have been better to

124 Maroczy Bind

play 20 J.. g7, so as to meet 21 :£Jg4 by 21 h5. Then White could hardly venture into the lion's den by 22 .£) h6 + \\'h7, so the knight would have to return to f2. After 20 .1J...g7 White should switch plans again by 21 �d3 b6 22 � f2, followed by a3 and b4.

21 .£)g4 .1J...g7 22 cxb5 il_xb5 23 '-2lh6 + *f8

Unfortunately Black has to go to f8 because 23 ffh8 loses a pawn after 24 8xf7 + , while 23

.1J...xh6 24 il_xh6 gives White a permanent advantage.

24 .1J...xb5 �xb5 25 �d5 'i'!(e8 26 e5! � d8?!

In a difficult position Black fails to offer the most resistance. He should try to give up a pawn to reach an ending in which the off-

side knight on h6 gives White problems. Therefore 26 .£)e6 was the best (26 � a6 27 exd6 is bad since 27 � xd6 28 �xf7 + ! wins a clear pawn), with the idea 27 � xc8 � xc8 28 exd6 exd6 29 �xd6 + �e7 30 ¥ifxe7 + ffxe7 with excellent counterplay. White should prefer 29 :£Jg4!, when the weak pawns on d6 and a5 make Black's position unat­tractive, but his chances would certainly be much better than in the game .

27 exd6 exd6 28 � el � ac8?

The final collapse. 28 Qle6 was necessary to meet the threat of 29 � xc5, when 29 .£)g4 gives White a positional advantage but nothing decisive.

29 � xc5! 30 �xc5

� xc5 Resigns

8 Taimanov Variation

The first moves of this system run 1 e4 c5 2 -E:J f3 e6 3 d4 cxd4 4 -E:Jxd4 '-E:lc6 (or 2 -2)c6 and 4 e6). In a way this resembles the Kan Variation, since Black keeps the f8-b4 diagonal open for his bish­op, but here White cannot play JL.d3, so there are fewer options for the first player. The variation I am recommending for White, 5 -E:l b5, is the most obvious way to exploit the substitution of <21c6 for a6. The main line leads to a kind of Maroczy Bind position in which White's knight is on a3 instead of d4. This provides more support for the c4 pawn, so White is not normally forced to play b3, but on the minus side the knight is distinctly offside on a3 and in many lines White tries to bring it back via b 1 or c2. Black's best strategy is watchful waiting as in the Maroczy Bind. One difference between this line and the Maroczy Bind is that Black's bishop is on e7 instead of g7, so that White may be able to play for a direct kingside attack by advancing his g- and h-pawns. As usual, such a strategy carries many risks, but with the knights on a3 and c3 holding up Black's counterplay by

b5 or d5 White may have enough time to break through.

106 B

Game 25 Chandler-Pritchett

British Ch. 1985

1 e4 c5 2 �f3 e6 3 d4 cxd4 4 Ql xd4 .:21 c6 5 .f)b5 (106)

5 d6 Not 5 .R_c5 (5 Q} f6 6

<El l c3 transposes to Chapter 9) 6 j}_ f4 � f6 7 ·l!lfc l 'lfrf8 8 .f) l c3 when both 8 . . . �ge7 9 .R_e3! b6 1 0 �d2 j}_xe3 1 1 �xe3 d5 1 2 0-0-0 a 6 1 3 exd5 exd5 14 �c7 d4 1 5 � xd4 <£) xd4 16 <2Jxa8, Gufeld­Plaskett, Hastings 1 986/7 and 8 . . . a6 9 j}_d6 + �xd6 1 0 Q} xd6 �d4 1 1 �f4 ( 1 1 •d 1 is also good) � f6 1 2 ,g d l •b4 1 3 •d2, Mokry-Plaskett, Trnava 1 984, gave White a clear advantage.

126 Taimanov System

6 c4 �f6 Black can also play 6 . . . a6 7

� 5c3 � f6, when White can transpose to the main line by playing 8 �a3, but in practice White has usually chosen to try exploiting Black's move order by developing his knight to d2 in­stead. After 6 . . . a6 7 � 5c3 � f6 8 J.e2 J.e7 9 0-0 0-0 10 J.e3 Black has tried:

( I ) 10 . . . ijc7?! (usually wrong in this system, since after White plays :i, c l Black will be in danger from tactics based on !E) b5 or �d5) 1 1 �a3 (reasonable now that Black is committed to a bad line) b6 1 2 ijel llb7, Dieks-Mar­janovic, Manila 1 974, and now 1 3 l;l c l followed by f3 and itf2 would be good for White.

(2) 10 . . . lld7 1 1 f4 ijb8 1 2 IE)d2 b6 1 3 a 3 � a7 14 llf3 J.c6 1 5 � b3 �d7 1 6 �d4 J..b7 17 f5 with a slight plus for White, Kar­pov-Taimanov, USSR Ch. 1 976.

(3) 10 . . . � e8 1 1 a3 J..f8 1 2 ijb3 �d7 1 3 � d2 b6 14 �f3 J. b7 1 5 X fd I and again White is slightly better, Bronstein-Lom­bardy, Teesside 1 975.

(4) 10 . . . b6 1 1 ijb3 (the im­mediate 1 1 �d2 is also good, and after 1 1 1, b8 1 2 Ji c l J.b7 1 3 ijb3 � d7 1 4 .i. fd l �ce5 1 5 f4! IE)g6 1 6 � f3 ! ijc7 1 7 f5! �ge5 1 8 IE)d4 � c5 1 9 ijc2 White was doing well in Gufeld-Hort, Dort­mund 1 983) K b8 1 2 l, d l ijc7 1 3 a3!? Jld7 14 �d2 !I fc8 ( 1 4 . . . �a5 is just slightly better for White) 1 5 !;! de l ! ijd8 1 6 ijd l

K c7, Psakhis-Holm, Plovdiv 1 983 and now 1 7 b4! intending !E) b3 is good for White according to Psakhis.

7 � 1c3 Now 7 � 5c3 makes no sense,

for Black gains a tempo by miss-ing out a6.

7 a6 After 7 . . . Jle7 8 llf4 e5 9 J.g5

a6 10 Jlxf6 gxf6 1 1 �a3 � d4 1 2 �c2 �xc2 + 1 3 ijxc2 White has an evident positional advantage, Chiburdanidze-Taborov, USSR 1 979, so Black must push the knight away.

8 �a3 (107)

8 Jle7 The speculative pawn sacrifice 8

. . . d5 9 cxd5 exd5 10 exd5 � b4 was played a few times, but is now considered bad after 1 1 J.e2, for example:

( I ) 11 . . . Jlc5 1 2 Jle3! J.xe3 13 '!fl'a4 + � d7 ( 1 3 b5 14 '!fl'xb4 J.. b6 15 0-0 Jla5 16 jlxb5 + ! axb5 1 7 K fe 1 + \fid7 1 8 iifxb5 + \fic7 1 9 d6 + ! ijxd6 20 K ac 1 wins for White, while 1 3 . . .

j_d7 14 �xb4 �b6 1 5 �xb6 J..xb6 1 6 �c4 j_d4 1 7 !l; d l leaves Black with nothing for the pawn) 14 �xb4 ( 14 fxe3 �h4 + 1 5 g3 �e7 16 !l; d l is also good) lL_c5 1 5 �e4 + <ll'f8 16 0-0 b5 1 7 �c2 and Black has no real compensation for the pawn, Karpov-van der Wiel, Brussels 1986.

(2) 11 . . . J..e7 12 j_O k_ f5 l 3 0-0 0-0 1 4 �c4 j_c5 1 5 J..e3 J..d6 1 6 �xd6 'tifxd6 1 7 'i!]td2 � fd8 1 8 !l; ad l �d3 1 9 jj_e2 � xd5 20 jj_xd3 �xc3 2 1 jj_xf5 � xd l 22 !! xd l �xd2 23 !! xd2 and White won the ending in Lawton-Gay­son, British Ch. 1 987.

Some Black players have exper­imented with delayed castling, but this increases the danger that White will gain time by missing out jj_e3, for example 8 . . • b6 9 J..e2 jj_b7 10 f4 ( 1 0 0-0 is less accurate, because after I 0 � b8! White has no good way to defend the e4 pawn, since the re­ply 1 1 0 cuts out the plan based on f4 recommended in this game) � b8 l l jj_O �bd7 1 2 0-0 j_e7 l 3 'tife2 0-0 1 4 g4!? d 5 1 5 exd5 exd5 1 6 g5 � e8 1 7 �g2 �e4, Geller-1. Sokolov, Panchevo 1 987, and now 1 8 cxd5 is good for White according to Geller.

9 jj_e2 0-0 10 0-0 b6

By far the most popular move. The alternative is to play for b5 directly in order to save time, but the problem is that a prema­ture . b5 just leaves Black with a weak b-pawn, for example:

Taimanov System 127

( I ) 10 • . . £ b8 l l J..e3 �a5 1 2 �c2 j_d7 l 3 f4 b5 1 4 cxb5 axb5 1 5 a3 !! fc8 16 g4 l!l;d8 1 7 g5 � e8 1 8 jj_xb5 �a5 1 9 a4, Mednis­Schmidt, Nice 1977, and Black does not have enough for the pawn.

(2) 10 • • . J..d7 1 1 J..e3 �b8 (1 1 . . . �e5 1 2 f4 �g6 l 3 l!l;el jj_c6 14 jj_O �e8 1 5 �c2 !! c8 1 6 a4 a5 1 7 g3 �d7 1 8 h4 gave White attacking chances in Yudashin­Osnos, USSR 1 986, while 11 . . . � b8 1 2 f4 �a5 1 3 � c2 b5 14 cxb5 axb5 15 a3 is also slightly better for White) 12 0 �a7, Aga­pov-Taimanov, USSR 1987, and now l 3 �d3!? j_c6 14 � fd l �d7 1 5 jj_ fl �e5 1 6 J%'e2 gives White an edge according to Agapov. If Black cannot force through . b5 then his piece deployment looks very strange.

11 j_e3 ( 108)

1 1 �e5 This is a major branch. Black

has a number of possibilities, but since this line is governed largely by general principles rather than

128 Taimanov System

specific variations, we do not waste time considering numerous very similar lines. Instead we only seriously examine two lines, 1 1 J..b7 and 1 1 . . . 2JeS, which represent quite different plans. In the latter case Black intends repositioning his knight to d7, opening the diagonal for the b7 bishop. This line plans dS at some stage. 1 1 iil.b7 is a more restrained line; Black concen­trates on developing his pieces and waits to see what White in­tends. After 1 1 1L b7 (or 1 1 . . . J..d7 1 2 f] -� b8 1 3 '1\oiel 8, a7 14 ·M'f2 );£ b7 l S � fd l � d8 1 6 £ ac l JLe8 1 7 <iW h l �d7 1 8 ·iiii fl JL f6 1 9 �abl �cS 20 f4 j_e7 21 b3 h6 22 � d2 � c8 23 � cd l � d8, Matulo­vic-W. Schmidt, Vrnjacka Banja 1983, and now 24 8 b2! is good for White, while 1 1 . . . J.i e8 1 2 � c l � e S 1 3 f4 2J ed7 14 it_ f] will probably transpose to the main line of the game) 12 'ii! b3 <dd7 ( 1 2

.£) aS?! 1 3 'jojlf xb6 � xe4 14 �xe4 .il..xe4 lS w xd8 Jt. xd8 16 � ad 1 left Black under unpleasant pressure in Karpov-Kasparov, Moscow match 1984/S; nobody has cared to repeat this line) 1 3 � fd l Black may play:

( 1 ) 13 . . . � c8 14 )4 ac l �aS l S 'l!!'a4 fS 1 6 exfS )el xfS 1 7 b4 J..gS, Chandler-Kurajica, Sarajevo 198S, and now 1 8 .zl g4! iL.xe3 1 9 ji xfS J..xc l 20 J..xe6 + 1fr h 8 21 J..xd7 is good for White accord­ing to Chandler, the point being that 21 �gS 22 �dS ;g xc4 23 � xc4 � xc4 is met by 24 J..h3! ,

both defending g2 and threaten­ing i!ffe8 mate.

(2) 13 . . . � e8 (a bit irrelevant) 14 � ac l 1J.f8 l S ji_fl � c8 1 6 ·;..«c2 '-2lceS 1 7 h 3 'liic7 18 f3 with a typical slight plus for White, Kar­pov-Romanishin, USSR 1 98 1 .

(3) 13 . . . J.i a7!? 1 4 £t; d2 ( 1 4 <2\ a4 <2\aS doesn't achieve any­thing since l S �c2 .£)c6 and l S J..xb6 <2\xb6 1 6 �xb6 �xb6 1 7 .£)xb6 jlxe4 are not better for White) jl a8 l S 'l!¥fd l '*b8 1 6 'i!irfl 22! f6, Nunn-Cebalo, Biel 1 986, and now 1 7 f4 intending �f2 would have created awkward pressure against the b6 pawn.

(4) 13 . . . � b8 14 � ac l j_a8 l S � d2 � b7 1 6 �d l �b8 1 7 '*fl <2\ f6?! (Black's plan is similar to that in Nunn-Cebalo above, but this time White finds the right reply; 1 7 J.i\ c7 was more solid) 1 8 f4 � c7, A. Rodriguez-Kirov, Havana 1 986, and now 19 'i!iff2 � fc8 20 � cd 1 followed by J..f3 would, by defending b2 a second time, create a genuine threat to take on b6.

(S) 13 . . . .£)c5 (the most im­portant line) 14 'l!tc2 (109) and now:

(Sa) 14 . . . .ii.f6 l S � ac l and now:

(Sa l ) 15 . . . J..e5 1 6 <2\ab l 'l!¥fh4 1 7 g3 �f6 1 8 f4 J..d4 19 �d2 eS 20 �dS �d8 (20 i!th6 2 1 �El bc3 fS 2 2 exfS � xfS 2 3 J_g4 is good for White) 21 � bc3 J..xe3 + (2 1 0 ltl' h8 22 fS �d7 23 J_f3 jlcS 24 ltl'g2 f6 2S <2\e2 gave White the advantage in Karpov-

Olafsson, Amsterdam 1976) 22 �xe3 exf4 23 gxf4 � e8 24 b4 '2\d7 25 J.. f3 with a slight plus for White, Westerinen-Liberzon, Geneva 1 977.

(5a2) 15 . . . J..xc3 1 6 �xc3 l£] xe4 1 7 �d3 l£]e5 18 �d4 and White will regain the pawn while keeping the two bishops.

(5a3) 15 . . . LtJ b4 16 'l!!l'd2 (if White wants to avoid complica­tions then 1 6 �bl is a safe op­tion) J..xc3 1 7 bxc3 l£]xe4 18 'M b2 '2\c6 1 9 f3 l£]f6 20 J_xb6 ·1!t!fe7 21 '2\c2 � fc8 22 *a3 l£] d7 23 1i_f2 l£] c5 24 � d2 � d8 25 � cd l f6 26 l£]d4 l£]e5 27 l£] b3 l£] xb3 28 axb3 with an edge for White, Mokry­Lobron, Reggio Emilia 1 984/5.

(5b) 14 . . . �c7 and now there is plenty of flexibility about move order, for example:

(5b I ) 15 f3 (the start of a slightly unusual plan based on preparing b4 by � ab I ) ll, fe8 1 6 •d2 X ac8 1 7 � ab ! l£]d7 1 8 f4 '2\cb8 1 9 M be l (having persuaded the knights to retreat the rook may be repositioned) l£]f8 20 �c2 l£] bd7 2 1 l£]d4 .b8 22 b3 ;Ld8 23 J..f3 and Black's passive play gave

Taimanov System 129

White the advantage in Yudas­hin-Dzhandzhava, Simferopol 1988.

(5b2) 15 � ac1 � ac8 ( 1 5 J..f6 16 c2\ab l � ac8 1 7 a3 �b8 1 8 b4 �d7 1 9 'i!-1d2 .iJ...e7 2 0 f4 � fd8 21 'ii-Je l gave White his usual slight plus in Tseshkovsky­Hulak, Dubai 01. 1 986) 1 6 :d ab! ( 1 6 JL.fl )4 fd8 1 7 � b l c2l b4 1 8 �c2 �xc2 was roughly equal in Lobron-Liberzon, Ramat-Hash­aron 1 982) .t.:� e5 17 �d2 :2Jcd7 1 8 a3 1£ fe8 19 b4 �f6 20 h3 �g6 was played in Jadoui-Karpov, Brussels 1986. This position is slightly unusual in that Black's knight has moved away from the queenside to g6. The logical res­ponse would be for White to start action on the relatively bare left flank by 21 � b3 followed by a4-a5.

(5b3) 13 i-iid2 (this appears the most logical because it allows the a3 knight to return to the game via c2) �e5 ( 1 5 � ad8 1 6 �c2 �e5 1 7 f3 �cd7 1 8 � ac t � f6 1 9 �d4 was slightly better for White in Karpov-Small, Lucerne 01. 1982) 1 6 f3 with an edge for White in Akopian-Semkov, Erevan 1 988. It is worth giving the rest of this game in full because it is an excellent example of how Black can be tortured in this variation. White goes round and round, an­noying Black with one little threat after another, always being care­ful not to allow Black to free himself. Finally White adopts the plan of a q ueenside breakthrough,

130 Taimanov System

which eventually proves decisive: 16 ;s_ ac8 1 7 � ac l £ fd8 1 8 i'ii e l -2Jg6 1 9 iii f2 kf6 20 b4 �d7 21 -2Ja4 £ b8 22 Ji.fl ka8 23 'if�> h i h6 24 � d2 JJ.. e7 2 5 � b l i);lb7 26 a3 -2J ge5 27 � f2 � dc8 28 ..t:Jd2 d c6 29 "21 b3 k!, bc8 30 � b2 � 6c7 3 1 ii! d2 -df6 32 i>id4 2\ed7 33 :s. b l � b8 34 .k. f2 -de8 35 i<id2 _k g5 36 . .k.e3 .k.xe3 37 m-xe3 -2)e5 38 a be l -dd7 39 i<id2 :&_b7 40 a4 :s. d8 41 a5 _:&a8 42 -2J a4 � dc8 43 axb6 -2Jxb6 44 -2Jxb6 ·i'tJxb6 45 c5 dxc5 46 -2Jxc5 iLb7 47 E, a l � a8 48 -2Jd7 iila7 49 b5 1J..c8 50 b6 8 Xd7 5 1 bxa7 8 xd2 52 � xd2 'itrf8 53 � d8 'itre7 54 E:; xc8 Resigns.

12 f4 -2Jed7 The knight transfer from c6 to

d7 improves the position of Black's pieces, since he may exert pressure against e4 by J_ b7 and .:dc5, or he may play for

d5. The defect is that it gives White extra time, and this allows him the chance to start a kingside attack.

13 jj_f3 11_b7 14 ·l.l!fe2 li( c8 ( 1 10)

Once again the move order is a matter of personal choice, but this move is bound to be played sooner or later, so Black may gain some flexibility by playing it first.

15 ,g, acl It is unwise not to play this

move, because a poorly defended knight on c3 is an invitation for Black to start a tactical storm by playing d5. This happened after 1 5 g4 h6 1 6 h4 d5! 1 7 exd5

jj_xa3 18 bxa3 Q)xd5 19 cxd5 � xc3 20 j_ d4 ,g, xa3 21 g5 hxg5 22 fxg5 e5!, A. Rodriguez-Polu­gayevsky, Moscow 1 985 and 1 5 I!! ad 1 fi/c7 1 6 g4 d5! 1 7 cxd5 j_xa3 1 8 bxa3 �xc3 19 g5 Q)xd5 20 exd5 j_xd5 2 1 j_xd5 exd5 22 j_d4 �xa3, Hernandez-Renet, Thessaloniki 01. 1 988, with ad­vantage to Black in both cases.

15 �c7 16 ,g, fd1

It is an open question as to whether this further preparatory move is really necessary. The alternative is the immediate 1 6 g4 ( 1 1 1) and now:

( 1 ) 16 . . . Q) c5 1 7 �g2 (after 1 7 jtd2 h6 1 8 h4 Q) h7 1 9 b4 jtxh4

20 frg2! <iJd7 2 1 £il h 1 j_e7 22 g5 hxg5 23 � xh7! frxh7 24 £il h l + ltrg8 25 frgl f5 the complications should have led to a draw in Gufeld-Georgadze, USSR 1 985) d5 (Kasparov suggested 1 7 g5 18 fxg5 <iJ fd7, but not surpris­ingly nobody has cared to try this) 1 8 exd5 ( 1 8 e5 <2l fe4 1 9 cxd5 exd5 20 b4 <iJ xc3 21 � xc3 d4 was unclear in Tseshkovsky-Kas­parov, USSR Ch. 1 979) �d3 19 ;;,( cd l ! .t:lxf4 20 _k xf4 JL.c5 + (20

�xf4 2 1 d6 wins a piece) 2 1 l\' h l �·xf4 22 g5 -2) e8 (22 . .t:ld7 23 dxe6 wins a piece) and now White has a pleasant choice: he can win a pawn for insufficient compensation by 23 dxe6, but 23 d6! appears even stronger to me.

(2) 16 . . . ·i'i.i'b8 (this appears too passive) 1 7 g5 -2Je8 1 8 ii_ g2 g6 19 .il..h3 -2Jg7 20 f5 exf5 21 exf5 gxf5 22 � xf5 <iJ xf5 23 iL xf5 -2l e5 24 ·;>;r h5 -2J g6 25 ..l!Ld4 �;;;; fe8 26 s fl il..f8 (26 g c5 27 -EJd5! is good for White after 27 . . . 1t_xd5 28 1Lxg6 fxg6 29 ·rti h6 or 27 . . . � xd5 28 cxd5 li_f8 29 j_e6! fxe6 30 � f7!) 27 li_xc8 'i'ti' xc8 28 � xf7 'ifi'xf7 (28 .Rh6! 29 .!:;( xh7! wf5 30 .!:;i, h8 + -2Jxh8 3 1 iifxe8 + 1t_f8 is the critical line, but i t i s doubt­ful if Black has enough compensa­tion for the two pawns since White's king can flee to the queen­side) 29 � xh7 + 'i\'e6 30 'iii' xg6 + 'lt,ld7 3 1 ·�f5 + �c7 32 i'i\' xc8 + *xc8, Chandler-Quinteros, Vienna Open 1 986, and now simply 33 h4 would have been good for White.

Taimanov System 131

(3) 16 . . . h6! 17 h4 -2Jc5 1 8 �g2 d5 19 exd5 -2Jd3 ( 1 9 exd5? 20 g5 hxg5 21 hxg5 .t:ld3 22 gxf6 jl xf6 23 . .Rxd5 <iJxcl 24 <2l e4 wins for White, Geller-Franzoni, Berne 1987) with unclear compli­cations. Note that the continua­tion of line 1 is ineffective here, since after 20 &i\ cd 1 (20 g5 lL c5 is unclear) .t:lxf4 21 1t_xf4 li_c5 + 22 'iti' h l &xf4 the h4 pawn is attacked.

16 � fe8 17 g4 h6

If Black plays 1 7 -E)c5, then 1 8 !ifg2 leaves him much worse off than in the last note. There are two reasons for this; firstly the rook on e8 takes away a flight square from the f6 knight, so that there is an immediate threat to win a piece by g5 and b4, and secondly there is no chance that

d5 will work when the rook is on d l .

18 h4 .t:lh7 Black is trying to hold up g5 for

as long as possible. Now 1 9 g5 hxg5 20 hxg5 e5 is fine for Black, so White must take time out to defend h4.

19 �h2 ( 112) 19 'l!!b8

This passive move looks wrong, but Black has not had much luck with the alternatives:

( l ) 19 . . . <iJc5 20 �h3 (defend­ing the bishop on f3, for the im­mediate 20 g5 is met by 20 . . . f5!) JL. f6 2 1 <iJ ab l (meeting the threat to the e4 pawn; White is finally ready for g5!) g6 22 £!l c2 j_g7 23

132 Taimanov System

� cd2 §Lf8 24 g5 h5 25 §Lf2 §Lc6 26 Qla3 Qld7? (26 *b7 was better, although the out-of-play knight at h7 gives White some advantage), Nunn-Cramling, Zurich 1 984, and now 27 e5! is very good for White.

(2) 19 *d8 20 jlf2 g5 (a drastic way to prevent g5 by White; although this gains Black the e5 square he still has problems along the h-file) 2 1 hxg5 hxg5 22 f5 Qle5 23 *g2 §Lf6 24 � h l Qlf8 25 jld4! Aa8 26 i, cfl *e7 27 jle2, Hellers-Wahls, World Junior Ch. 1 986, and the threat of � f3-h3 causes serious problems for Black.

20 g5 White decides to play g5

straight away, even though Black can reply with . . f5. The alterna­tive was 20 * h3 as in line I of the last note.

20 f5 21 *g2

There is a choice of promising lines for White. Chandler decides to keep the h7 knight locked out, but the direct 2 1 gxh6 gxh6 22 exf5 jlxf3 23 *g3 + \trh8 24

*xf3 exf5 25 Qld5 followed by Ad4 was also good.

21 hxg5 22 hxg5 fxe4 23 Qlxe4 d5 24 Qlf2 Ac5!

Black defends well. By ex­changing Black-squared bishops he exposes the weakness of f4.

25 Axc5 Qlxc5 (1 13)

26 l!;! d4 � f8? This move has disastrous

consequences because it gives White the time he needs to revive his kingside attack. Black should have continued his policy of liqui­dation by 26 dxc4; after 27 Ql xc4 Axf3 28 '!!ltxf3 E. cd8 White still has some advantage because of the h7 knight, but with each exchange White's attacking chances are reduced and he may soon have to look to the safety of his own king.

27 g6 .£lf6 28 �h3

Black has no threat, and the free move enables White to create his own unpleasant threats of 29 b4 and 29 Qlg4.

28 £ fe8 Relatively best. 28 . . . d:xc4 29

�g4 j_xf3 (29 �cd7 30 .\;;; xd7) 30 -2lxf6 + � x f6 3 1 'i,t(h7 + <;trf8 3 2 W,fh8 + vee7 33 � xg7 + wins, as does 28 . . • � ce8 29 <;tr g2 followed by � h 1 .

29 B, el Stepping up the pressure on e6

renews the threat of b4. 29 -2lg4 Qjcd7 was less convincing.

29 �d6 30 Rg4

Taimanov System 133

Black cannot bring any further support for e6 and he has no answer to the threat of b4 (30 a5 31 -2l b5).

30 e5 31 fxe5 � xe5 32 � xe5 'iifxe5 33 j}_ xc8 � xd4 34 lL xb7 Resigns.

White has netted a clear piece since the recapture on b7 allows mate in 3 .

9 Sicilian Four Knights

This rather antiquated system is not currently in favour, but fashions can change and it is advisable to be prepared even for less common variations. It exper­ienced a brief surge of popularity when Chandler adopted it for a couple of years, but the generally passive nature of Black's position proved unattractive to other players and it has now virtually disappeared again. Black plays I e4 c5 2 1£) 0 e6 3 d4 cxd4 4 1£) xd4 l£) f6 5 l£)c3 l£) c6 (of course this can arise from other move orders, in particular via 2 �c6). In some ways this is akin to the Kan and Taimanov systems since Black leaves the f8-b4 diagonal open for his bishop, but instead of playing . . . a6 he develops a piece. Naturally this is in Black's favour unless White has some direct method of exploiting the omission of a6, so 6 �db5 is the only move to cause Black problems. Black then very often continues 6

d6 and after 7 ..;L f4 e5 8 J.,g5 we have transposed to the Pelikan, considered in Chapter 4. The point of this move order is that Black avoids the possibility that after 1 e4 c5 2 1£) 0 �c6 3 d4 cxd4 4 €)xd4 €) f6 5 l£)c3 e5 6 �db5 d6 White might play 7 €)d5

or 7 a4. Since we are recommend­ing the main line with 7 j_g5 this transposition is not a worry and therefore 6 d6 just leads to the earlier chapter on the Pelikan. After 6 l£)db5, 6 d5 loses to 7 exd5 exd5 8 ..;Lf4 and 6 ..;Lc5 7 ji_f4 followed by JJ..d6 is un­pleasant for Black, so we need only consider 6 j_ b4 in this chapter. The tactical line recom­mended in the first edition has suffered a serious setback in re­cent years, so this time we only analyse the positional continua­tion 7 a3 JJ..xc3 + 8 �xc3 d5 9 exd5, which either gives Black an isolated pawn after 9 exd5 or gives White a lead in development after 9 . �xd5 1 0 J.,d2. This is a safe line for White in which he is likely to secure a small but perma­nent advantage. In practice it is easy for White to allow the posi­tion to slide towards a draw, and in some ways it is an annoying line to meet because instead of the sharp struggle typical of most Sicilian lines, White is trying to exploit a slight positional edge. Nevertheless it is even more un­pleasant for Black, who can only win if White takes exceptional risks, and so this line is relatively unpopular.

Game 26 Mokry-8. Stein

Gausdal 1988

1 e4 c5 2 22!f3 e6 3 d4 cxd4 4 �xd4 �f6 5 �c3 �c6 6 �db5 .�b4 7 a3 j_xc3 + 8 �xc3 d5 9 exd5 ( 1 14)

9 . . . exd5 This has been almost the only

move played during the past decade, but it is interesting to note that in a very recent game Ulf Andersson preferred the knight recapture. The analysis runs 9 �xd5 1 0 jtd2 and now:

( 1 ) 10 . . . '2!xc3 1 1 j_xc3 'i'ii' xd l + 1 2 � xd l f6 ( 1 2 e5 1 3 .�d3 .ile6 14 0-0 f6 1 5 f4 i s simi­lar, lvkov-Giigoric, Amsterdam 1 964) 1 3 f4 (for some reason Ehl­vest preferred the passive 1 3 f3 and after 1 3 jt.d7 14 . .l\l.d3 0-0-0 1 5 � d2 e5 16 jte4 j_e6 1 7

Sicilian Four Knights 135

;tte2 � xd2 + 18 'lftxd2 ljje7 1 9 jj_b4 tjjd5 20 il_xd5 j_xd5 2 1 'lfte3 the players agreed to a draw in Ehlvest-Andersson, Skelleftea 1 989) l\l.d7 14 . .itc4 0-0-0 1 5 0-0 'Yfl;c7 1 6 � de l � he8 1 7 � f3 j_c8 with a very unpleasant position for Black, Fischer-Addison, USA Ch. 1 962-3.

(2) 10 . . . �h4 1 1 J!lllf3 0-0 (11 . . . �e5 1 2 "iioi'g3 and 11 . . . �d4 1 2 �d3 just make matters worse) 1 2 0-0-0 �xc3 1 3 Jtxc3 e5 14 _il_d3 Jlg4 (14 'il'Jg4 15 Jte4 -rt� xf3 16 j_ xf3 with the typical favourable ending for White, Minic-Gerusel, Halle 1 967) 1 5 �e4 'i'\lfh6 + 1 6 il._d2 lf(g6 1 7 f3 ii_e6 1 8 '¥fxg6 hxg6 1 9 Jte3 and again White has a promising end­ing, Tai-Matulovic, Kislovodsk 1 966.

(3) 10 . . . �b6 1 1 � b5 �d4 12 exd4 �xd4 1 3 j_b5 + jj_d7 14 ]lxd7 + \trxd7 15 0-0 left Black's king badly placed in Kaplan-Sia­peras, Siegen 1 970.

(4) 10 . . . �f6 1 1 �h5 0-0 1 2 0-0-0 ljj xc3 ( 1 2 . . . ti!fxf2 1 3 ljj xd5 exd5 14 Jtd3 is no better) 1 3 jlxc3 "�tf4 + 14 � d2 e 5 1 5 il._b5 with advantage, Matulovic-Kok­koris, Athens 1 969.

(5) 10 . . . 0-0 1 1 �h5 tjj f6 12 )1Hh4 �d4 1 3 j_g5 � d8 14 �xd4 lit xd4 1 5 il. d3 b6 1 6 0-0-0 is the same story as in all the other lines, Gufeld-Hasin, USSR Ch. 1 966.

10 j_d3 0-0 Or 1 0 . . d4 (after 1 0 �e7 +

1 1 �e2 �xe2 + 1 2 � xe2 ljje5 1 3 A,b5 + j_d7 14 j_ xd7 + *xd7,

136 Sicilian Four Knights

Liberzon-Bronstein, USSR 1 972, White could have played 1 5 j_e3 � he8 16 j_d4 with a slight ad­vantage) and now:

( 1 ) 1l �e2 + j_e6 1 2 <2:!e4 <2:!xe4 1 3 �xe4 �d5 14 J..f4 0-0-0 1 5 0-0 g5, Frolov-Maliutin, Jur­mala 1 989, and now 16 it d2 <2:! e5 17 �xd5 lt xd5 18 ltf5 + is a little better for White according to Maliutin and Kimelfeld.

(2) 1 l <2:!e2 lt f5 ( 1 1 0-0 1 2 0-0 transposes into the main line) 1 2 0-0 ltxd3 1 3 �xd3 0-0 1 4 ltg5 h6 1 5 lth4 H e8 16 Jii ad 1 � c8 1 7 c3! (more dynamic than 1 7 a fe1 � e6 1 8 *fl �c7 1 9 j_g3 'ii1'b6 20 b4 with just a microscopic advan­tage for White, Karpov-Kuzmin, Leningrad 1 977) dxc3 18 'i!,!(h3 'fje7 19 <2:!xc3 'fje6 ( 1 9 'iil'e5 20 f4 �e3 + 21 'fjxe3 � xe3 22 j_xf6 gxf6 23 <2:!d5 is very good for White) 20 'fjxe6 fxe6 (20 . . . � xe6 2 1 j_xf6 Ji:ii. xf6 22 )!4 d7 b6 23 H e 1 favours White) 2 1 j_xf6 gxf6 22 <2:!e4 with an endgame advantage for White, Estevez-Chaviano, Santa Clara 1 983.

1l 0-0 (1 15) d4

This is the most logical because it forces White to decide where his knight is going immediately. To avoid liquidation it seems that 1 1

d4 should be met by the rela­tively passive <2:!e2; after other 1 1 th moves White can usually ar­range to meet d4 by the more active <2:!e4, for example:

( 1 ) 1 l . . . a6 1 2 JJ..f4 ( 1 2 j_g5 is also promising) d4 1 3 <2:!e4 <2:!d5 ( 1 3 il_f5 14 ltc7! illustrates why Black should not have delayed) 1 4 j_d6 K e8 1 5 j_g3 f5? (suicide, but even 1 5 . . . il_f5 1 6 <2:!d6 ll_ xd3 1 7 �xd3 '!4 e7 1 8 <2:! f5 is very awkward) 1 6 <2:!d6 � f8 1 7 ltc4 lte6 1 8 .l;ll. e 1 'ltd7 1 9 <2:!xb7 'ltxb7 20 K xe6 <2:!a5 21 lta2 Resigns, Vukcevic-Ervin, USA 1 976.

(2) 1 l . . . h6 12 il_f4 d4 1 3 <2:!b5 ( 1 3 'E}e4 as in line 1 is also pos­sible) <2:!d5 14 'ltf3! j_e6 1 5 K ad 1 'ltd7 1 6 h3 K ad8 1 7 lth2 "fke7 1 8 'tlrg3 and White has a clear advan­tage, Ciric-Rossolimo, Vrsac 1 969.

(3) 11 . . . ltg4 12 f3 j_e6 1 3 ll..g5 h6 ( 1 3 K e8 14 'Wt'd2 d4 1 5 <2:!e2 a6 1 6 <2:! g3, Planinc-Anders­son, Sombor 1 970 and 1 3 'Wt'b6 + 1 4 fl' h 1 <2:!d7 1 5 f4! f5 1 6 "f!lf3, Matulovic-Benko, Vrnjacka Banja 1 973 were also bad for Black) 14 lth4 g5 1 5 il..f2 <2:! h5 1 6 <2:! b 5 and according to Taimanov White has a clear plus.

12 <2:!e2 The available evidence suggests

that this offers the best chances for an advantage. After 1 2 <2:!e4

J..f5 1 3 j_g5 j_xe4 1 4 j_ xe4 h6 1 5

J.. h4 ( 1 5 J..xf6 trxf6 L 6 X e 1 ll, ad8 1 7 tyd3 <2)e5 1 8 tyb3 b6 1 9 I!I e2 g6 20 K ae 1 *g7 2 1 trb5 � fe8 22 Jl d3 <2) xd3 23 ttxd3 X xe2 24 K xe2 K d5 was equal in Kudrin-Chandler, London 1 987) g5 1 6 j_xc6 bxc6 1 7 j_g3 itd5 1 8 f4! <E:�e4 ( 1 8 . . . g4? 1 9 j_ h4 wins) 1 9 fxg5 hxg5 ( 1 9 . . . <2) xg3?! 20 hxg3 hxg5 2 1 tyd3 f5 22 m ae l is dangerous) 20 tyd3 f5 21 K ad 1 � ad8 22 J..f2 c5 23 c3 tyb3! White's advantage was infinitesi­mal in Kir. Georgiev-Chandler, Leningrad 1 987.

12 . • . J..g4 (116) Or 12 h6 ( 1 2 . . . itd5 1 3

<2) g3 gives White an edge) 1 3 h3 (probably best, although 1 3 J.. f4 .£)d5 14 j_g3 tyf6 1 5 � e 1 <2Jde7 1 6 .£) f4 j_ f5 1 7 J..c4 � ac8 1 8 tyd2 � fd8 1 9 <E:�d3 gave White an edge in Geller-Winants, Amster­dam 11 1 987; not, however, 1 3 J..b5 j_g4 1 4 f3 tyb6!) a6 (because now j_b5 is a real threat) and White may try:

( I ) 14 J..f4 !;!! e8 ( 1 4 . . . <2)d5 1 5 j_ h2 tyf6 1 6 <2)g3 gives White an edge) 1 5 � e1 tyd5 1 6 <2)g3 J..d7 17 tyd2 (threat j_xh6) g5 18 J..c7 b5? (Black should have started hacking rooks oft) 1 9 f4 1! ac8 20 fxg5 hxg5 2 1 <2)f5! 1: e3 22 e xe3 dxe3 23 tyxe3 11 xc7 (Black has material equality but too many loose pieces) 24 lt ad 1 'lt'ffi 25 � fl * g7 26 j_ f5 and White wins, Carlier-Winants, Wijk aan Zee 11 1 987.

(2) 14 l!l et <2)d5 15 e, f4 ( 1 5

Sicilian Four Knights 137

J..e4 <2)de7 16 �d3 �b6 17 � d l � d8 1 8 �g3 'flh8 1 9 j_f4 tyb5 20 J..d3 �d5 2 1 j_c7 liK eS 22 c3 dxc3 23 <2)xc3 was good for White in Ernst-Prasad, Subotica 1 987) <2)xf4 16 J..xf4 J.. e6 1 7 � h5 �d7 18 .i, e2 f5 19 � ae l J..f7 20 �f3 g ae8 21 � xe8 � xe8 22 � xe8 + J..xe8 23 j_g3 with a solid positio­nal plus for White, Lobron­Gobet, Biel 1 984.

13 f3 This appears best, because the

exchange on e2 often simplifies Black's defensive task. After the alternative 1 3 j_g5 ( 1 3 J..f4 � e8 14 � e l � b6 1 5 b4 22!e5 1 6 j_xe5 � xe5 1 7 tyd2 j_xe2 18 li( xe2 � xe2 1 9 j_xe2 g6 20 K d l K d8 2 1 J.. f3 gave White a small but lasting advantage in Lobron­Chandler, Biel l 987) tyd6 14 }.! e l ( 1 4 ttd2 J..xe2 1 5 trxe2 K fe8 1 6 tr d l .£)e5 offers White nothing, Short-Wiedenkeller, Esbjerg 1 984) K e8 (after 14 . . . a5 1 5 �d2 <2)d5 1 6 h3 j_xe2 17 j_xe2 h6 1 8 J..h4 Black tried to imprison White's bishop by 1 8 f5 19 c4 .£)[4 20 J.. f3 <2)g6 2 1 j_g3 f4 but

138 Sicilian Four Knights

this rebounded after 22 J..d5 + ltl>h7 23 � e6 )t(d7 24 �d3 �ce7 25 J..h2 �f5 26 � ae 1 � ab8 27 �e4 b6 28 g3 'l!ltd8 29 h4 h5 30 )'!i!lf3 wh6 31 � xf4 + <ff h7 32 'W!e4 � c8 33 'i!il{f3 wh6 34 � l e5 Resigns, R. Mainka-B. Stein, Dortmund 11 1 987) 1 5 �d2 ( 1 5 f3 jth5 1 6 �f4 h6 1 7 �xh5 hxg5 1 8 �xf6 + �xf6 1 9 'ii\'d2 .:£)e5 20 � e4 Draw, Short-Chandler, Hastings 1 987/8) J..xe2 (better than 1 5 � ac8 1 6 �g3 with a clear edge for White) 16 � xe2 E xe2 1 7 �xe2 lii( e8 1 8 �f3 .:2)e5 19 'ii\'f4 �b6 20 J..xf6 .:£)xd3 21 �g3 �xf6 22 �xd3 )'!i!lb6 Black drew easily in Kudrin-Rogers, London 1988.

13 J..h5 14 J..g5 "Wtd6

The position is the same as in Short-Chandler above, except that the moves � e I and E fe8 have been omitted. This dif­ference favours White, as the main line of the game proves.

15 �e1 15 'ii\'d2 � ad8 16 i; ad l !:[ fe8

1 7 J..h4 1l..g6 1 8 J..xg6 hxg6 1 9 J..fl .:£)d5 was level i n Zapata­Chandler, Amsterdam 1 987. The move �e1 has the immediate threat of �h4, but White also intends to step up the pressure on d4 by 'itfl and � ad 1 .

15 J..g6 16 � dl !Ue8

Black must not exchange on d3 as this gives White a free tempo to increase the pressure on d4 by �fl and H fd l . However Black

might have tried to exploit the fact that White's queen is no longer defending the c2 pawn by putting a rook on c8 .

17 'l!i!ffl � adS 18 H d2 � d7

Black decides to meet White's plan passively, even though being forced on the defensive is usually a sign that an isolated pawn posi­tion has gone wrong. However Mokry's suggestion of 1 8 !:[ e5 appears no better after 1 9 J..h4 threatening J..g3.

19 i; fd1 )!;!( ed8 In the line 1 9 .it xd3 20

� xd3 'ite5 2 1 J..xf6 �xe2 22 �xe2 � xe2 23 J..xd4 � xc2 24 J..c3 Black succeeds in exchang­ing his isolated pawn, but only at the cost of giving White a domi­nant bishop and good chances of penetrating to the seventh rank.

20 j_b5! Removing a vital defender in­

creases the pressure on d4 intoler­ably. Black's reply leads to a fatal material loss, but even 20 �e5 21 j_xc6 bxc6 22 J..xf6 gxf6 23 f4 'itb5 24 .:£)g3, threatening f5, is very unpleasant.

20 21

Black's since 2 1 J..xd7.

22 23 24 25 26

White

h6 ( 1 1 7) .:£)xd4! hxg5

moves are all forced, .:£) xd4 loses to 22

�xc6 '!lfxd2 � xd2 � xd2 .:£)xd8 � xfl <ffxfl J..xc2 .:£)xb7

is a pawn up and his

active king makes the task of con­verting his material plus into a point relatively simple.

Sicilian Four Knights 139

26 J..b3 27 '£\aS AdS 28 L£)c6 j)_xc6 29 j)_xc6 '*f8 30 *e3 *e7 31 .d4 *d6 32 j_ bS Resigns

Black did not wish to see the technical phase of the game.

1 0 Lowenthal Variation

This line starts 1 e4 c5 2 e,f3 e,c6 3 d4 cxd4 4 e,xd4 e5 and is slightly akin to the Pelikan in i ts use of an early e5. The 'old' Lowenthal runs 5 e, b5 a6 6 e,d6 + jlxd6 7 '/i!l'xd6 �f6 (7 flte7 8 '/i!l'xe7 + is worse) and Black hopes that his lead in de­velopment will compensate for his black square weaknesses and lack of the two bishops. Current theory suggests that this is a vain hope and White should be able to maintain an advantage. Black has an interesting alternative which has been pioneered by Sveshnikov and other Soviet players. This runs 5 e, b5 d6, and here White has the choice between 6 c4, aim­ing for a firm grip on d5, or 6 e, l c3 as in the Pelikan. If White plays 6 e, l c3 Black plans to reach a superior type of Pelikan varia­tion in which e,ge7 is played instead of e, f6, thus avoiding the doubling of Black's f-pawns.

If, on the other hand, White plays 6 c4 Black will either aim for counterplay by f5 or try to exchange his bad bishop by jle7-g5. This line is still in a state of flux, so we give a more com­plete survey than usual, since at this stage it is impossible to judge which line is best for White.

Game 27 Liberzon--Franzoni Biel (Open) 1980

1 e4 c5 2 e,o :£lc6 3 d4 cxd4 4 e, xd4 e5 5 €)b5 a6

Or 5 d6 (118) (5 . . . �f6 6 e, l c3 transposes to the Pelikan, 5 . . . J.. c5 6 €) 1 c3 !£) f6 transposes to page 63 and 5 . . . h6 6 2Jd6 + J..xd6 7 'i!t!xd6 �e7 8 �d I <E:l f6 9 €Jc3 is good for White) and now there are two lines:

( 1 ) 6 €) 1c3 a6 (6 €l f6 is a Pelikan) 7 €J a3 b5 8 e,d5 €)ge7 (after 8 €) f6 White may trans­pose back to the Pelikan with 9 J..g5, but unfortunately the resulting line is not part of our recommended repertoire; I there­fore suggest 9 e, xf6 + '\irxf6 10 c4

b4 1 1 <E:�c2 Ab7 1 2 A d3 '(Qt d8 1 3 0-0 jle7 14 a3 and White had an edge in Vito1insh-Ambarcumian, USSR 1 988; 8 J...e7 9 c4 b4 10 <E:�c2 a5 1 1 jle3 � b8 12 J...e2 � f6 1 3 iitd3 ®d7 14 tjjxe7 r�Jxe7 1 5 ll( d 1 gave White an edge in Geo. Timoshchenko-Sveshnikov, Moscow GMA 1 989) 9 c4 tt:Jd4 and now:

( l a) 10 cxb5 <E:�xd5 1 1 exdS J...e7 1 2 J...c4 axb5 1 3 tjjxb5 j_a6 14 tjja3 0-0 1 5 0-0 J...f6 16 j_e3 left Black struggling to find compen­sation for the pawn, Klovans­Kiselev, Frunze 1 988.

( l b) 10 jle3 tt:J xd5 1 1 cxdS jle7 1 2 J...d3 0-0 1 3 0-0 f5? ( 1 3 A f6 is only very slightly better for White) 14 jlxd4 exd4 1 5 exf5 J...xf5 1 6 tjjc2 and Black has no compensation for his weak pawns, A. Rodriguez-Estevez, Cuba Ch. 1 988.

(2) 6 c4 j_e7 (6 jle6 generally leads to a transposition) 7 � 1 c3 a6 8 ® a3 and now:

(2a) 8 . . . h6 9 j_e2 jl e6 10 0-0 J...g5 11 tjjc2 transposes to line 2c below, but not 1 1 jlg4 e.f6 1 2 J...xe6 fxe6 1 3 J...xg5 hxg5 14 tt:Jc2 \ti f7! and Black is at least equal, Nikolenko-Sveshnikov, USSR 1 987.

(2b) 8 . . . tt:Jf6 (Black simply develops, abandoning the idea of

. . h6 and J...g5) 9 J...e2 0-0 1 0 j_e3 j_e6 1 1 0-0 l;! c8 1 2 iitd2 tt:J a5 ( 1 2 h6 1 3 l;! acl was slightly better for White in Dol­matov-Guseinov, Klaipeda 1988) 1 3 tjjd5 J...xd5 14 exd5 b6 1 5

Lowentha/ Variation 141

l;! ac l ®b7 16 f4 tt:Jd7 1 7 tt:Jc2 a5 1 8 tjja3 with advantage to White, Dolmatov-Minasian, USSR 1 988.

(2c) 8 . . . jle6 9 .£)c2 h6 10 j_e2 jlg5 1 1 0-0 �f6 ( 1 1 tt:Jge7 1 2 �d3! ®g6 1 3 g3 jlxc l 1 4 !;l axcl iitg5 1 5 �d5 jlxd5 1 6 cxd5 with an edge for White, Geller-Lpu­tyan, Moscow 1 987) 1 2 *d3 J...xcl ( 1 2 iitc7 1 3 K d l lll[ d8 14 J...xg5 hxg5 1 5 K ac l �b6 1 6 b 3 i s a little better for White, Geller-Lputyan, USSR 1 987) 1 3 i4, axc l 0-0 14 K fd l �b6! 1 5 '!i'xd6!? (or 1 5 a3 �a5! 1 6 �xd6 .£:! xc4 1 7 J...xc4 �xd6 18 l;l xd6 jlxc4 19 K b6 l!l ab8 20 tt:J e3 jl e6 2 1 !! d 1 li fe8 22 b3 \fi f8 with equality, Dvoiris-Tiviakov, USSR 1 988) 'ii!i xb2 16 ·�d3 � b6 1 7 � b l �a7 1 8 �d5 .£)d7 19 C?Jc7 <E:�c5 20 *'e3 ii;[ ab8 is unclear according to Tiviakov.

In practice 6 c4 has been the most popular response, but 6 .£J i c3 may be better, even though this allows Black the chance to transpose into the Pelikan.

6 tt:Jd6 + jlxd6 7 'f!l'!'xd6 'i'itf6 ( 1 19)

7 �e7 8 '!i'dl <E:� f6 9 tt:Jc3 threatening JJ..g5 is good for White.

8 �d1 White has a wide variety of

queen moves and most of them are good! There seems little doubt that 8 'i'itc7, which has always been highly regarded theoreti­cally, gives White a good game but I have not recommended it

142 Lowenthal Variation

here because White must always be careful that his queen is not trapped, so the simpler �d I seems preferable. One should note that 8 �xf6 is also quite good, e.g. 8 �xf6 <£)xf6 9 <£)c3 and now:

( I ) 9 . . . d5 10 Jl..g5 d4 (10 . . . ldb4 1 1 J.. xf6 gxf6 12 <£)xd5 <£)xc2 + 13 lt'd2 8 xa l 14 �c7 + 'I!Je7 1 5 <£)xa8 J..e6 16 �b6 JLxa2 1 7 frc3 and 10 . . . -d xe4 1 1 �xd5 0-0 12 ii_e3 are also good for White) 1 1 ll_xf6 dxc3 1 2 .1J..xg7 � g8 1 3 J.. h6 c.2) b4 14 0-0-0 <£)xa2 + 1 5 fr b l .§_e6 1 6 1i d6 h g6 1 7 jj_e3 <£) b4 1 8 jlc5 with advantage for White according to Gligoric.

(2) 9 . . . <£) b4 and now: (2a) 10 ll_d3 l2\ xd3 + ( 1 0 h6

1 1 b3 d6 1 2 j}_a3 <£)xd3 + 1 3 cxd3 f,;e7 14 f4 *e6 1 5 f5 + lt;e7 1 6 � d l � e8 1 7 d4 exd4 1 8 � xd4 'l!f f8 1 9 jj_xd6 + *g8 20 0-0 b5 2 1 e 5 Resigns was a drastic finish, Byrne-Evans, USA Ch. 1 98 1 ) 1 1 cxd3 h6 1 2 b3 with an edge for White.

(2b) 10 lt'd2 d5 1 1 a3 d4 1 2 axb4 dxc3 + 1 3 fre3 <£)g4 + 14 'I!Je2 f5 15 bxc3 l2!f6 1 6 .!ill a5

<£)xe4 17 f3! <£)d6 1 8 � xe5 + and Black has very little for the lost pawn, Velimirovic-Ristic, Yugos­lavia 1 979.

8 . . . �g6 8 <£)ge7 9 �c3 0-0 (9

�g6 transposes to the next note) 1 0 j}_e3 b5 1 1 �d2 �g6 1 2 f3 d6 1 3 0-0-0 li!( d8 14 * b l J..b7 1 5 g4 f6 16 <£)d5 QJxd5 1 7 �xd5 + is also good for White, Gligoric­Benko, Dublin 1 957.

9 �c3 ( 120)

9 d5!? For a time this move caused a

revival of the Lowenthal, but now White has found a way to defuse the complications and liquidate to a favourable ending. The older line runs 9 . . . � ge7 10 h4! h5 (10 . . . d5 1 1 h5 �d6 12 h6! g6 1 3 exd5 and now 13 . . . �d4 and 13

<£)b4 allow 14 <£Je4, while 10 . . . h6 1 1 h5 �f6 12 j}_e3 0-0 1 3 �d2 b 5 14 0-0-0 b4 1 5 � a4 a 5 1 6 � b6 � b8 1 7 �d6 gave White an excellent ending in Boleslavsky­Sakharov, USSR 1 957) 1 1 J_g5 d5 (the only move that makes sense, for example 1 1 . . . b5 1 2

�d3 .i:..b7 1 3 0-0-0 l,;;l d8 14 'i!i!(d6 1 5 l,;;l xd6 f6 16 ..:d_e3 �c8 1 7 l,;;l d2 <2:! 6e7 1 8 j_d3 d6 1 9 l! hd 1 was very good for White in Hazai­Csom, Warsaw 1 987) 1 2 exd5 (the tempting 1 2 J..xe7 is met by 1 2 . d4!) <2:! b4 ( 1 2 <2:! d4 1 3 J..d3 J.. f5 14 J..xf5 -EJ exf5 1 5 �d3 f6 1 6 j_e3 i s very good for White) 1 3 JJ...xe7 f'ixe7 1 4 JJ...d3! (much better than the often recom­mended 14 d6 + since White reaches the same type of ending, but with his d-pawn securely defended) <2:!xd3 + 1 5 �xd3 �xd3 1 6 cxd3 and now:

( I ) 16 . . . b5 17 a3 and Black cannot recover his pawn, for ex­ample 17 . . . J..f5 1 8 f'ld2! llil, h6 1 9 iiii he 1 'l;d6 20 � ac 1 , Svesh­nikov-Panchenko, USSR 1 977 or 17 . . . J..b7 1 8 0-0-0 'l;d6 (sug­gested by Sveshnikov) 1 9 d4, or finally 1 7 lliL b8 (suggested by Baumbach) 1 8 0-0-0 b4 1 9 axb4 � xb4 20 .i. he! Jii! xh4 (or else H e4) 2 1 d4!, and in all cases White has a good ending.

(2) 16 . . . H h6 1 7 0-0-0 .i g6 1 8 Jiii he1! li;1 xg2 (18 . . . f6 1 9 d4 and 18 . . . 'l;d6 19 d4 are also good for White) 1 9 .i, xe5 + ( 1 9 d4!? is interesting) 'l;d6 ( 1 9 wd8 20 lll[ del j_d7 2 1 d6 threatens lll( xh5) 20 d4 followed by -EJe4 + and again White has the advantage.

10 �xd5 �xe4 + l l j_e3 �d4

This move, which is the only reasonable reply to the threat of -EJc7 + , is the idea behind 9 d5!?

Lowenthal Variation /43

12 -E:lc7 + f'le7 12 . . . *d8? allows White to

take the a8 rook, while after 12 . . . *f8? White can either play 13 � cl or take the exchange by 13 'ii(d3 <E�xc2 + 14 '*'d2 �xd3 + 1 5 jl.xd3 -EJxe3 1 6 �xa8 �d5 1 7 liii ac I or take the rook -a pleasant choice!

13 iiii cl! Until this move was discovered

Black had been doing rather well against 1 3 21 xa8?! and 1 3 'i'ii'd3.

13 _itg4 If Black moves the rook on a8

then 14 c3 is very strong. 14 'i'ioid3 'i'ii' xd3 15 ]Lxd3 8 d8 I6 h3 (121 )

16 j_c8 The alternative is 1 6 j_h5 1 7

f4 f6 1 8 'f1 f2 't1d6 1 9 c3 and now: ( 1 ) 19 . . . *xc7 is bad after 20

cxd4 + <If b8 21 fxe5 fxe5 22 lii!, c5. (2) 19 . . . <2:! b3 20 axb3 'f1xc7 2 1

J..e4 �e7 2 2 Jiil. hel <2:! d 5 2 3 fxe5 -EJ xe3 24 <If xe3 fxe5 25 llit fl 1il deS 26 l,;;l f5 J.. g6 27 liii g5 j_xe4 28 ki\ xg7 + f'lc6 29 'llrxe4 with a winning position for White, Mar-

144 Lowenthal Variation

janovic-Simic, Yugoslavia 1 983 . (3) 19 . . . Qjc6 20 J,.b6 exf4 2 1

c4 <2:lge7 22 J..e4 Qjc8 2 3 c5 + ltrd7 24 <2:ld5 <2:lxb6 25 cxb6 frd6 26 Qjxf4 J..f7 27 !;!\ hd I + lfie5 28 lfle3 with a clear plus for White, Winsnes-Hillarp, Rilton Cup 1 988.

17 f4 exf4 18 jtxf4 Qje6

Otherwise White castles and Black is unable to develop his king's rook while e8 is covered.

19 Qjxe6 jtxe6 20 0-0

The outcome of the opening is very favourable for White. He has two active bishops supporting a queenside pawn majority and while so many pieces remain on the board Black's king is not well placed on e7.

20 21 a3 22 .td2 23 � eel 24 � f3 25 b3 26 c4 27 .txc4?!

<2:lf6 Qjd5 Ji;l d7 !;!l c8 b5 h6 bxc4

It was more important to drive away Black's centralized knight than to keep the queenside pawns intact. After 27 be <2:l f6 28 .t b4 + White has a passed pawn and an attack against Black's king.

27. . . !4 aS?! (122) A passive and nervous move. 27

11. c6 is better. 28 a4

Preparing both J.. c l -a3 + and b4-b5.

28 frd6 An attempt to bring the king

over to help in the fight against White's queenside majority, but two rooks and two bishops are a formidable attacking force and the king soon runs into trouble.

29 b4 frc7 30 b5 axb5 31 J,.xb5 � d6 32 M dl

With the sneaky threat 33 J,. f4 <2:l xf4 34 Jil c3 + winning the exchange.

32 \fib7 33 J_b4 � dd8 34 Jil fd3 '\1rc7

Trying to unpin the knight. 35 � cl + '\1rb6

35 fr b8 36 )!l c5 J! a7 37 J..c6 wins material.

36 J,.c5 + 37 l!il cdl 38 J,.e7 39 J..xe8 40 l!I xd5 +

'\1ra5 Jil ac8 � e8 liil xe8 Resigns

1 1 Pin Variation

There is no generally accepted name for this variation, which runs I e4 c5 2 �f3 e6 3 d4 cxd4 4 �xd4 � f6 5 �c3 j_ b4. There is certainly a pin involved, so 'Pin Variation' is a reasonable name. Until ten years ago this was thought to be a very poor line for Black, but round about 1 979 it suddenly reappeared with Black's play being based on a new idea involving an exchange sacrifice. After a few years during which it was used in occasional Grand­master games it entered a decline and is now very rarely seen. How­ever it is worth studying because there are a lot of tricky tactics in the Pin Variation, and White players who do not know the cor­rect antidote may well find them­selves in trouble.

Game 28 Wagman--Barle Biel (open) 1981

1 e4 c5 2 �f3 e6 3 d4 cxd4 4 � xd4 �f6 5 � c3 _itb4 6 e5 ( 123)

The only move to cause Black any difficulties.

6 f)d5

Black's two alternatives are close to losing by force:

( I ) 6 • • . "iifa5 7 exf6 j_xc3 + 8 bxc3 "iifxc3 + 9 'i!i!\'d2 'i!i!\'xal 1 0 c3 (threat 1 1 �b3 '*bl 1 2 �d3) 'i!i!\'b1 1 1 ltd3 'i!i!\'b6 12 fxg7 li g8 1 3 �h6 and wins.

(2) 6 . . . � e4 7 'i!i!\'g4 'i!i!\'a5 (7 �xc3 8 �xg7 liK f8 9 a3 and now the lines 9 . . • lta5 1 0 lth6 "i'#fe7 1 1 � b3 and 9 "i'#fa5 10 � b3 'i!i!\'d5 1 1 J..d3 are winning for White so Black must try 9 . • . �b5 + 1 0 axb4 �xd4 1 1 j_g5 �b6 1 2 j_h6 'l!l!'xb4 + 13 c3 � f5 14 cxb4 �xg7 1 5 j_xg7 .!J( g8 1 6 ltf6 but White's black square pressure gives him a very favour­able ending) 8 'i!i!Jxe4 j_xc3 + 9 bxc3 'i!i!Jxc3 + 10 f8id1 >etxa l 1 1 � b5 d5 1 2 'ltb4 .Q)a6 ( 1 2 'i!i!\'xe5 1 3 f4 �c6 1 4 fxe5 � xb4 1 5 � c7 + f;>d8 1 6 � xa8 b6 1 7 j_a3

146 Pin Variation

is winning for White) 1 3 o£)d6 + wd7 14 .2_xa6 bxa6 1 5 o£) xf7 � g8 16 lffd2 d4 (or else j_a3) 1 7 j_b2 �Wxa2 1 8 � a l �d5 1 9 !;;;[ a5 and White has a decisive attack (analysis by Euwe).

7 il..d2

van Gent 1 986, and now 1 3 g6 fxg6 14 j_h6 wf7 1 5 1! xf8 + *xf8 1 6 j_xf8 wxf8 1 7 j_e2 is good for White) 10 J.,h6 g6, but the main line still favours White after 1 1 h4! ( 124) and now:

Originally theory gave 7 ·�1·g4 124 as best, but after 7 . 0-0 (the new B idea mentioned above) 8 kh6 g6 9 j_xf8 'i'fi·xf8 Black has reason-able compensation for the exchange with play against c3 and e5.

7 �xc3 Or 7 .2_xc3 8 bxc3 0-0 9 lt_d3

d6 10 'i'S' h5 (10 exd6 'M xd6 1 1 0-0 was at least slightly better for White, Geller-Tseitlin, Moscow 1982, while 10 f4 dxe5 1 1 fxe5 o£)d7 12 'i'ti h5 g6 13 rt, e2 i't:r C7 14 c4 'ii& b6 15 �f3 J?Je7 16 iil_ c3 Bc5 17 �d2 proved good for White in Epishin-Uiybin, Tbilisi 1989) g6 1 1 � e2 dxe5 1 2 i�'i'xe5 -::Jd7 1 3 �d6 w f6 14 0-0 b 6 1 5 J?Jc6 J?Jc5 1 6 c4 e5 1 7 ·r.;; xe5 with a clear plus for White, Vogt-Ermenkov, Ber­lin 1982.

8 bxc3 iL a5?! A major decision point for

Black. 8 :ill._a5 keeps the pres­sure against c3 but leaves the kingside dangerously bare. Black should adopt the alternative var­iation 8 !iie7 9 ·i£ig4 0-0 (9 . . . g6 1 0 h4 h5 1 1 i'l\ g3 J?Jc6 12 J?J b5 was good for White in Rabar­Fuster, Munich 1942, while the remarkable 9 . . . g5 I 0 h4 h5 1 1 hxg5 !? hxg4 1 2 � xh8 + .k f8 was played in Grosar-De Waal, Sas

( I ) 1 1 . . . d6 1 2 h5 �a5 ( 1 2 dxe5 1 3 Ad3! exd4 14 hxg6 fxg6 1 5 jlxg6 wins) 1 3 0-0-0!? dxe5 (13 . . . 'ii!i'xc3 14 .!it h3 ..W a l + 1 5 wd2 'iii1xa2 1 6 J.,d3 �a5 + 1 7 c3 'tlifxe5 18 hxg6 fxg6 1 9 j_xg6 � xf2 + 20 *cl and 13 . . . �xe5 14 hxg6 fxg6 1 5 J..d3 � f6 1 6 � de l 't!l!fd5 17 j_g5 Z! f7 18 j_xe7 � xe7 1 9 J.,xg6 are winning for White) 1 4 <£) b 5 a6 1 5 hxg6 fxg6 1 6 _;txf8 J.,xf8 1 7 � xh7! 'l.rxh7 1 8 j_d3 wins for White.

(2) 1 1 . . . �a5 1 2 'l\ti'g3 M. d8 (after 1 2 . d6 1 3 exd6 Black may play 13 . . . J..xd6 14 �xd6 Ji d8 1 5 *' b4 'l!lol'xb4 1 6 cxb4 )i xd4 1 7 c3 with the better ending for White or 13 • . • J..f6 14 0-0-0 �xa2 1 5 h5 <£)c6 1 6 <£)xc6 'l!lol'a3 + 1 7 1ti'd2 bxc6 1 8 j_xf8 wxf8 1 9 hxg6 hxg6 20 J..c4 and White has the advantage) 1 3 h5 d6 14 hxg6 fxg6 1 5 .ZI.f4 dxe5 1 6 J\l.xe5 )i d5

1 7 f4 <£) d7 1 8 1l..c4 � xe5 (after 1 8 .!! xe5 + 1 9 fxe5 'l'#xe5 + 20

�xe5 �xe5 21 J,Lb3 the e6 pawn is too weak) 1 9 1l..xd5 �xd5 20 fxe5 �e4 + 2 1 'tffd2 1l..d7 22 � ae 1 '¥'d5, Wedberg-Pokojowc­zyk, Copenhagen Open 1 984, and now 23 � xh7! 1frxh7 24 £ hl + 'tffg7 25 �h2 1trf7 26 'i!!Jh7 + 1tre8 27 'i!lixg6 + 1trd8 28 g h8 + should win for White.

9 �g4 White can also use the move

order 9 ll d3 d6 10 'i!\f' g4 after which Black has nothing better than 1 0 0-0 transposing into the game.

9 0-0 10 j_d3 d6 11 �f3

Byrne and Mednis suggest 1 1 J,Lg5 with the lines 1 1 . . . J,Lxc3 + 1 2 lfr fl f5 1 3 exf6 � xf6 14 � d 1 and 1 1 . . . 'W:fc7 1 2 J.. f6 �xc3 + 1 3 *e2 'ii\'{d2 + 14 1tr fl g6 1 5 £ d I . This has never be.en tried in practice but looks good to me.

1 1 g6 Or else ll_xh7 + is crushing, for

example 1 1 dxe5 1 2 j_xh7 + \\>xh7 1 3 �h5 + 1trg8 14 ®g5 £ e8 1 5 )\i1xf7 + *h8 1 6 �h5 + followed by 'l)ith7 + , �h8 + , � xg7 + and ® f7 + picking up the queen.

12 h4! The latest twist-White just

plays for mate. 12 �g5?! h5! (not 1 2 dxe5? 1 3 �xh7! lfr xh7 14 1l..g5 J..xc3 + 15 *e2 �d4 1 6 'i!\f'h3 + *g8 1 7 ll_f6) 1 3 �g3 dxe5 14 ®e4 �d7 1 5 0-0 1trg7

Pin Variation 147

enabled Black to defend in Peters-Arnason, New York 1 980, while 1 2 0-0 dxe5 (Sigurjonsson gives the attractive line 12 . . . � c6 1 3 J.,g5 �c7 14 il, f6 j_xc3 1 5 'i!\f'h4 <£)xe5 1 6 �g5 h5 1 7 �xh5) 13 �xe5 followed by f4 only gave White an edge in Sigurjonsson­Arnason, Iceland 1980.

12 . . . dxe5 1 2 . . . h5 is met by 1 3 'i!\f'g3 or 1 3

'i!\f'f4. 13 h5 f5 (125)

14 1L xf5! exf5 14 a xf5 1 5 hxg6 h5 16

;g xh5 ·ii-J f6 ( 1 6 £ xh5 1 7 �xh5 '!i1c7 1 8 0-0-0 wins, while other­wise Black can hardly meet the threat of � h8 + ) 1 7 0-0-0 fol­lowed by � dh l with a decisive attack.

15 "ii'J'C4 + £ f7 1 5 W;g7 1 6 hxg6 '1!rxg6 1 7

�xe5 + wins. 16 hxg6 hxg6 17 �g5 wc7

If the queen defends the rook from any other square White wins by 1 8 � xf7 '@' xf7 19 � h8 +

18 ·i'i)' h4 'lt;f8 19 �xt7?

148 Pin Variation

Having conducted the attack so well up to here it is surprising that White should miss 1 9 �h8 + 'f!Je7 20 <E:!xf7 when he is material up with a mating attack. Fortunately White is still winning even after 1 9 <E:!xf7?

19 '1Jxf7 1 9 'l'fxf7 20 �h8 + �g8 2 1

tt-f6 + wins the queen. 20 �h7 + '1ie6

20 'f!Jf6 2 1 j_g5 + 'ffje6 22 �xg6 + 'ffjd5 23 0-0-0 + 'ffjc5 24 j_ e3 + 'ffi b5 25 � d5 + is even worse.

21 22 23 24

'jifxg6 + R h6 �g8 + � bl!

White correctly adheres to the rule applying to king-hunts that it is more important to cut off the king's escape route than to give check.

24 25 26 27

j_e3 + j_xd4 + � f8 +

b5 <Eld4 exd4

White misses it the first time round

27 28 �g8 + 29 l!I xb5 + ! but spots it the second! 29 'f!Jxb5 30 �d5 + �c5 31 a4 + Resigns

It is mate next move.

1 2 2 � f6

Although this move surfaces from time to time, its appearances at the Grandmaster level are very rare. White has a range of pos­sible lines against 2 .:2) f6, which have all achieved good practical results. The variation 3 ..£Jc3, which was analysed in the first edition of this book, has per­formed well in the intervening years, while the more recent idea 3 e5 .:2)d5 4 .:2)c3 e6 5 .:2)e4 has also acquired a good reputation. How­ever, in this second edition, we will return to the 3 e5 .:2)d5 4 .:2)c3 e6 5 .:2)xd5 exd5 6 d4 variation which was recommended in the first edition. Concentrating on this one line I will give a more detailed coverage and show that White can be sure of at least a slight advantage against 2 Bf6.

Game 29 Rhine-Sprenkle

USA 1981

1 e4 c5 2 .:2)f3 .:2)f6 3 e5 .:2)d5

3 .:2)g4 4 h3 .:2)h6 may be met by 5 d4 or 5 c3, with advan­tage to White.

4 .:2)c3 e6 Or 4 . . .:2)xc3 (4 .:2)b4 5

J..c4 and 4 . . . .:2)c7 5 d4 cxd4 6 �xd4 Qlc6 7 '¥'e4 g6 8 ll.c4 ll.g7 9 0-0 0-0 I 0 lii e I , Kindermann­Ostl, Bundesliga 1 987, are good for White) 5 dxc3 ( 126) and now:

126 B

( I ) 5 . . . d5 6 exd6 �xd6 (6 exd6 7 J..c4 ll.e7 8 ll.f4 0-0 9 �d2 followed by 0-0-0 puts severe pressure on the backward pawn) 7 �xd6 exd6 8 §tf4 ll.g4 (8 . . d5 9 0-0-0 §te6 1 0 .:2)g5 followed by g3 and either ll.g2 or <E:lxe6 and ith3, when Black's central pawns will be fortunate to survive) 9 0-0-0 <El d 7 I 0 11. c4 (even better than j}_xd6, since . . . 0-0-0 is pre­vented) followed by j}_xd6 win­ning a pawn.

(2) 5 . . . <E:lc6 6 ll.f4 h6 (6 . . . e6 7 �e2 �a5?! 8 <E:l d2! 'f!lc7 9 <E:lc4 f6 1 0 � d6 + j}_xd6 1 1 exd6 �a5 1 2 h4 was good for White in van der Wiei-Bjelajac, Novi Sad 1 982)

150 2 !iJ/6

7 'i!ife2 flfc7 8 0-0-0 b6 9 �e3 e6 1 0 lE:!d2 lE:Je7 1 1 lE:Jc4 lE:!f5 ( 1 1 lE:!d5? 1 2 lii( xd5 exd5 1 3 lE:!d6 + \frd8 14 e6 wins) 1 2 '!Wh3 Ab7 1 3 lii( g l intending g4 and White has some advantage, van der Wiel­Murei, Moscow 1 982.

(3) 5 . . . g6 6 Ac4 Ag7 7 Af4 0-0 8 'i!itd2 followed by 0-0-0 and h4 gives White a strong attack.

(4) 5 . . . b6? 6 e6! dxe6 (6 fxe6 and 6 . . . f6 are both met by 7 lE:Je5!) 7 �xd8 + \frxd8 lE:Je5 *e8 9 Ab5 + Ad7 10 lE:Jxd7 lE:Jxd7 1 1 Af4 and White is close to winning already.

5 � xd5 exd5 6 d4 lE:Jc6

If Black doesn't like to sacrifice a pawn he can try 6 d6, but after 7 Ab5 + (127) Black can­not equalize, as the following analysis shows:

127 B

( 1 ) 7 . . . Ad7 8 Axd7 + 'l!lrxd7 (not 8 . . . lE:Jxd7 9 dxc5) 9 0-0 Q�c6 10 exd6 Axd6 ( 1 0 . . 'l!\rxd6 1 1 dxc5 'i!ifxc5 1 2 Ae3 is also good for White) 1 1 l;il e l + lE:Je7 1 2 dxc5 Axc5 1 3 Ag5 0-0 14 �d3 f6 ( 14 . . . h6 1 5 Axe7 Axe7 1 6 l! ad l

)l! ad8 1 7 c4 Af6 1 8 cxd5 Axb2 1 9 d 6 and White's passed pawn is very dangerous) 15 Ae3 and Black has a poor isolated pawn position in which he has no active pieces to compensate for the static weakness.

(2) 7 . . . lE:Jc6 8 0-0 Ae7 (8 A e6 is also met by 9 c4 when Black has nothing better than to transpose by 9 Ae7) 9 c4 Ae6 (9 . . . dxc4 10 exd6 tltxd6 1 1 d5 a6 1 2 Axc4 and 9 . . . a6 10 Axc6 + bxc6 1 1 cxd5 cxd5 1 2 exd6 �xd6 1 3 dxc5 'i!ifxc5 14 A e3 are very good for White) 10 Ae3 �b6 (White threatened exd6) 1 1 a4 a6 1 2 a5 �c7 1 3 exd6 'l!\rxd6 14 dxc5 �d8 1 5 Axc6 + bxc6 1 6 lE:Je5 'f�Jc7 1 7 'il!Ya4 with a horrid posi­tion for Black, Unzicker-Pomar, Bad Aibling 1 968.

7 dxc5 Axc5 8 '{ltxd5 'i!itb6

Here there is an important alternative: 8 d6 9 exd6 'i!if b6 ( 128) (giving up another pawn to allow Black's pieces to come into play more rapidly) and now:

( 1 ) 10 Ae3!? (for brave players

only) lWxb2 ( 10 A:xe3 1 1 fxe3 �xe3 + 12 �e2 � e6 13 �g5 is good for White) 1 1 2b5!? and now 11 . . . 'ii!l'xa1 + 1 2 lfi'e2 ·iiofc3 1 3 Axc6 + bxc6 (not 13 . . . \fi'f8? 14 '*xc5 nor 13 • • • lfi'd8? 1 4 �xf7) 14 �xc6 + <�rf8 1 5 'ii!l'xa8 �xc2 + 1 6 �d2 j_xd6 1 7 � b l leads to an advantage for White. 11 . . . �xb5 1 2 iWxc5 is depress­ing for Black, while 1 1 . . . 0-0 1 2 0-0 �xe3 1 3 fxe3 Ae6 14 'ii4'c5 was good for White in Boll-Lanz, corr. 1 982, so the best line is 11 . . . JLb4 + 1 2 'itre2 ·l:l!fxc2 + 1 3 '-2:!d2 JLe6 with a total mess.

(2) 10 ·i*e4 + �e6 ( 1 0 'itrd8 1 1 j_g5 + f6 1 2 0-0-0! 8 e8 1 3 �h4 is good for White) 1 1 !#h4! ( 1 1 ll_c4!? is an interesting recent idea, when van der Wiei-Short, Wijk aan Zee 1 990 continued 1 1

'i!irb4 + 1 2 '-2:!d2 0-0 1 3 0-0 � ae8 14 c3 'i'11b6 1 5 � f3 h6 1 6 b4 ll_xd6 1 7 11..e3 'i/iic7 1 8 �xe6 � xe6 1 9 �c4 and Black has in­sufficient compensation for the pawn) with the variations:

(2a) 1 1 . . . f6 1 2 JLd3 0-0-0 ( 1 2 � b4 1 3 d7 + ! ll_xd7 14

�h5 + followed by 0-0 i s good for White) 1 3 0-0 and there is a further branch:

(2a l ) 13 • • . JLxd6 14 j_e3 �xb2 1 5 � ab l ili'fa3 16 J..c4 gives White good attacking chances for no sacrifice.

(2a2) 13 • • • � xd6 1 4 j_e3 (I am not sure that it was necessary to return the pawn since Black has no immediate threats; 14 a3 intending b4 appears promising)

2 CZJ.f6 151

11..xe3 15 fxe3 �xe3 + 16 fih1 '/f(c5? ( 16 �b6 i s just slightly better for White) 1 7 � ae l JLd7 1 8 '/f(g3 g6 19 �d2! f5 20 '-2:!b3 �b4 21 a3 winning material, Chandler-Arnold, Bundesliga 1 987.

(2a3) 13 . . . h5 14 j_e3 �xe3 1 5 fxe3 �xe3 + 1 6 '\fi' h 1 j_g4 ( 1 6 � xd6 1 7 !:( ae l �b6 1 8 ;ii xe6 � xe6 19 1.L f5) I 7 itf g3 ( 1 7 Jii ae I �c5 1 8 j,Le4 is good for White according to Gutman) �c5 1 8 � ad I ltrb8 1 9 �e4 �e5 20 �xe5 J.xd I , Hansson-Fernandes, Lon­don 1 984, and although this game has appeared a number of times in print, nobody seems to have noticed that after 2 1 ili'fxg7 Black can quite reasonably resign.

(2b) 11 . . . .f.Lxd6 1 2 j,Le2 (even 1 2 �d3 '-2:!b4 1 3 0-0 '-2:J xd3 14 cxd3 0-0 1 5 'l:l!fe4 h6 1 6 .f.L e3 'ii\'xb2 1 7 M fb I was slightly better for White in Hellers-Ivarsson, Swe­den 1 985) _11.__ e7 ( 1 2 . '-2:!b4 1 3 0-0 is good for White after 13 . . . 0-0 14 c3 or 13 . • • . il.e7 14 '&{e4 f5 1 5 �e5 8xc2 1 6 jj_ g5) 1 3 we4 0-0-0 14 0-0 .k d5 (after 14 �d4 1 5 f)xd4 a xd4 1 6 � e3 H_c5 1 7 'i'l\' C3 White stands very well) 1 5 w g4 + 'iti>b8 ( 1 5 jJ_ e6 1 6 '£!a4), Chandler-Bartsch, Bundesliga 1 985, and now 16 c4! is good for White.

9 k c4 .k. xf2 + 10 '\ti>e2 0-0 1 1 � fl H_c5

Black has regained the sacri­ficed pawn, but now f7 is exposed to attack.

152 2 !iJ/6

129 B

12 �gS ( 129) �d4 + Transferring the knight to e6 in

order to shield fl. 12 -dxe5? ( 1 2 d6? 1 3 � xfl! �d4 + 14 <ttd l j_g4 + 15 g f3 + 1Ji>h8 1 6 -�g8 + and mate) 1 3 -�xe5 d 5 1 4 �xd5 Jg, e8 + ( 1 4 lll._g4 + 1 5 )4 f3 _r& g l 1 6 \fr fl ! g ad8 1 7 �e4 � d l + 18 1Ji>e2 1_xf3 + 1 9 gxf3 � fd8 20 R.xfl + 'itff8 2 1 fj xh7 + lJi>xfl 22 �g5 + 'l\>f8 23 � f5 + 'l;e7 24 'i>'\' fl + wd6 25 §1.[4 + <�rc6 26 'i<i'c4 + ·;;.tc5 27 ffl'xc5 + §J..xc5 28 � xd l � e8 + 29 �e4 Resigns was Prokopchuk-Kuz­netsov, USSR 1972) 1 5 <tt f3 1£r f6 + 1 6 'l\>g3 lll._d6 + 1 7 g f4! .§J..e6 18 fj xe6 � xe6 19 'ii'i xd6 ·�g6 + 20 8 g4 � e3 + 21 .Jl.xe3 ·if(xd6 + 22 w f2 l;t\ e8 23 � f4 � e7 24 J[.b3 ·i>�ie5 25 8 e l g5 26 � f3 'it;g7 27 � d l f6 28 <�rgl g4 29 .,Rd4 Resigns, Spassky-Ciric, Marianske Lazne 1 962.

13 'it;d1 fje6 14 2Je4

Here White has various possibi­lities, but this move attacking c5 and restraining d6 looks best.

14 . . . d6

14 . j e7 is too passive and in Savkin-Tseitlin, corr. 1 972 White obtained a strong attack after 1 5 c3 ( 1 5 2Jd6 is also good) d6 1 6 exd6 i;; d 8 1 7 lfrc2 jlxd6 1 8 � xfl! ft>xfl 19 2Jg5 + lt>e8 ( 1 9 'it;g8 20 ·ffl'e4 h6 21 :ZJ..e3 i:iif a5 22 J1. xe6 + 1L xe6 23 ·;i,f h 7 + and � fl + wins) 20 fjxe6 � f2 + 2 1 *b3 )i(b6 + 22 il..b5 + lll._d7 23 fjc7 + ! and Black resigned with­out waiting to see one of the lines 23 . . . \ifXC7 24 ·i'fi·g8 + _Rf8 25 ,A f4! or 23 . . . 1L xc7 24 -m g8 + 'lrJe7 25 Jl..g5 + '1id6 26 � d l +

15 exd6 £ d8 1 5 Jl..xd6? is a miscombina-

tion which rebounds after 1 6 �xd6 � d8 1 7 J..f4 d Xf4 (Black saw the danger in Zaretdinov­Pugachevsky, USSR 1 977 but still lost after 17 h6 1 8 J.e5 2Jg5 1 9 l;ll xfl etc) 1 8 �xfl + 'lrJh8 1 9 �g8 + Resigns, Unzicker-Sar­apu, Siegen 1 970.

16 j}_d3 j_xd6 17 ·�hS fS 18 �xd6 'l!!Jxd6

After 1 8 . £ xd6 19 'l!!Jxf5 the threats to f7 and h7 force 1 9 Jii: xd3 + , but Black does not have enough compensatioTh

19 'iifxfS ( 130) 19 . . • 'ill!xh2

Or 1 9 � f8 ( 1 9 g6 20 � fl + 'itr'h8 2 1 � f6 + '\t>g8 22 ,kf4 liquidates to an ending in which White has a clear extra pawn) 20 '!!!1f7 + '\fih8 2 l ll!ff4 and now:

( 1 ) 21 . • . �cS 22 j_e3 'i!l!Jh5 + (22 j}_g4 + 23 '\fid2 �a5 + 24

130 B

b4 �h5 25 �g5 exchanges queens) 23 *d2 ;le6 24 �g5 (24 ;1 d4! looks very good to me since 24 . . . �g6 allows 25 jlxg7 + fl/xg7 26 �f6 + and 24 . . . �dS 25 �e5! �xg2 + 26 � f2 �g4 27 h3! is a disaster) �e8, A. Rodri­guez-Diaz, Cuba Ch. 1 983, and now 25 � f4 intending J.. d4 gives White a clear advantage.

(2) 21 . . . �e7 22 �g5 �e8, Short-Minic, Banja Luka 1 985, and now Minic gives the line 23 J..d2 J.. e6 (23 �a4 24 ;lc3! lt;l xd3 + 25 f#j c l �g6 26 ll;! f8 + mates) 24 ;lc3 �g6 2 5 fl/d2 l! d5 26 �g3 �d7 27 g ae l , assessing the final position as slightly better for White. I suspect that White's advantage is considerably greater than this; he is a pawn up with the two bishops, and if he consoli­dates with f#j c l he must be win­ning. Therefore Black should play 27 J.. f5, but after 28 .i, e3 followed by * c l White is a pawn up for nothing.

20 �ti + fl/hS 21 J..gS � gS 22 J..e3! �dS

2 W6 153

After this move White can gain a clear endgame advantage with no risk. The critical move is 22 . �xg2! 23 'lth5! ( 131) (I gave 23 ll f2 in the first edition, but 23 �g4 + 24 'iti'd2 'i!tb4 + 25 rfi d l �g5! i s good for Black) and now:

( I ) 23 . . . �f8?! 24 � f4! (intending !;:;\ h4 and mate on h7) �h3 (seemingly forced) 25 i\!i xh3 ..@_xh3, Odeev-Varlamov, corr. 1 987, and now 26 *d2 il. g2 (26

� d8 27 i:i h l JL.g2 28 i;i. h2 J..c6 29 Ji;\ fh4 h6 30 .i.d4 wins the exchange) 27 g g 1 . .ll c6 28 � h4 (intending i;i. g3-h3) i;i. d8 29 i;i. g3 J.. d7 30 J..d4 and White has a large advantage.

(2) 23 . . . g6 24 'i'li h4! (24 ,@.d4 + <2)g7 25 J..xg7 + 'l/xg7 26 �e5 + <�ih6 is a draw) � g7 (24 • • • Ji:i: d8 25 'il!f f6 + * g8 26 'ii1 f7 + lfJh8 27 J.. d2! and 24 . . . ll..d7 25 .li, f7 .i;i. g7 26 J.. e4! are very good for White) 25 i;i. g l ·�f3 + (25 'l!l1d5 26 .i;i. xg6 threatening 27 {i!(xh7 + �·xh7 28 � h6 mate is crushing) 26 *d2 .Rf5 reaching a position in which White has a

154 2 t}jf6

very strong attack for the pawn. I have not been able to find a forced win for White, but Black has a difficult defence in prospect, e.g. 27 1i g3 �d5 28 g hi h5 29 _:&d4 (29 fii g5? Jii gd8!) _axd3 (the threat was 30 'iii xh5 + ) 30 cxd3 'lt'h7 (to defend against both J;i xg6 and J;i g5 followed by � xh5 + ) 3 1 � g5 ·� e6 (3 1 � f7? 32 ·�e4 � f5 33 ;g; hxh5 + gxh5 34 g xf5 wins) 32 ;,;; e5 'M c6 33 li;\ c l �g2 + 34 fi, e2 'M'd5 (34

�g4 35 ·M-xg4 hxg4 36 '£ e7 wins) 35 � c5 iird6 36 g e7 ;.:;. ad8 37 �c3 and Black is in big trouble since 38 fi, cc7 is threatened and 37 � c8 fails to 38 8 xg7 + while 37 li;\ d7 loses after 38 J;i xh5 +

23 'ii?f2?! As Rhine correctly points out,

23 'i'tl'f4! i'li' xf4 (23 'i'f'i xg2 24 i;;;\ g l '!!i¥d5 25 �d2 and &:; h i ) 24 8 xf4 would have given White a very favourable ending at no risk.

23 o.?Jc6 24 �d2 %'d6?

The bishop on d3 is the main enemy and Black should have tried to eliminate it by 24 . o.?Je5, when 25 _axh7 (25 Jii hi -2) xd3) 8 d8 + ! (25 o.?Jc4 + 26 l\>c3 we5 + 27 _:&d4 �a5 + 28 l\>xc4 _&e6 + 29 l\>d3 and White evades the checks) 26 iL.d3 -2)xd3 27 cxd3 �d6 28 ·i"f h4 + l\>g8 29 #e4 leaves White with some advan­tage, but in view of the opposite coloured bishops it isn't clear how many winning chances he has.

25 J1! hl h6 26 _lieS! ( 132)

White's ambition is to gain f4 for his queen, when a sacrifice on h6 will be inevitable.

26 "¥d5 26 'I!J!ic7 is spectacularly

refuted by 27 ·;i1 f6! 26 . . . )iie5 27 � ae l ·�g5 + 28 JL.e3 followed by .;ixh6 also loses quickly.

27 '1%'f4 � d8 27 ¥·xg2 + 28 'ltfc3 doesn't

help Black in his efforts to combat Jiit xh6 +

28 Jiit xh6 + '!fig8 29 � h8 + ! wxh8 30 i$'h4 + '\ffg8 31 �h7 + f;;f7 32 �g6 + f;;g8 33 �h7 + frf7 34 � n + .ars

34 'l;e6/e8 35 � g6 + ltrd7 36 li;\ f7 + mates.

35 .!;;:. xf5 + !ii xf5 36 �xf5 + frg8 37 ft'cl Resigns

Quite apart from his material disadvantage there is no defence to the threat of � c4 + .

1 3 2 a6

This is often called the O'Kelly Variation after the late Belgian Grandmaster who played it with some regularity. Since a6 is almost universal in the Sicilian Black gets it over with as soon a� possible, reserving his options as to which Sicilian system to adopt. White should not play 3 d4? cxd4 4 �xd4 � f6 5 � c3 e5 when Black has a favourable version of the Najdorf in which his king's bishop can emerge actively at c5 or b4. 3 c3 is a sensible reply, which tends to lead to 2 c3 Sicilian positions in which Black has played the unusual move a6 which is perhaps not the best wa; to spend a tempo. However the strongest reply of all is 3 c4, which either leads to Maroczy Bind positions or to a sort of hedgehog.

In view of the rare occurrence of 2 a6 in practice, it perhaps does not rate a chapter of its own, but this did give me the excuse to include another of my own games in the book!

Game 30 N unn--Surtees

Basingstoke Open 1977

1 e4 c5 2 �f3 a6 3 c4 ( 133)

3 .dc6 Or: ( ! ) 3 . . . d6 (this may lead to a

type of hedgehog) 4 d4 cxd4 ( 4 1tg4 is an interesting move, since 5 dxc5 1l._xf3 6 'ii-1Xf3 dxc5 leaves Black with a grip on d4 to com­pensate for the two bishops; 6 gxf3! dxc5 7 �xd8 + lf!>xd8 8 2Jc3 is possible, but the simplest line is 5 d5 with a positional edge for White) 5 ,f)xd4 2J f6. 6 -?Jc3 b6 (6

e6 7 j_d3 leads to positions from Chapter 6, so we concen­trate here on Black's attempt to develop early pressure against e4 which is unique to 2 . a6) 7 k d3 lt b7 8 0-0 (it is more accurate to play 8 *"e2 :£J bd7 9 b3-see next

?�te) -?J bd7 9 i't\'e2 e6 (9 -?Je5 ts Interesting since I 0 lL c2 ;:; c8 1 1 k a4 + -2l fd7 is unclear, so White would have to allow Black to take on d3) 1 0 b3 ( 1 0 f4 l\'t c7 1 1 lf!> h l

156 2 a6

§J_e7 1 2 §J_d2 is also good, when Nunn-Franklin, London 1 985, continued 12 . . h5 13 � ae 1 h4 14 f5 '-i:lf8 1 5 fxe6 fxe6 16 '-i:ld5! �d8 17 e5 dxe5 18 "tl!l'xe5 �xd5 1 9 �xe6 �d6 20 �xd6 .ZLxd6 2 1 cxd5 Resigns) l¥fc7 1 1 §J_b2 "*c5?! (Black plays too ambitiously with his king stuck in the centre- I !

. ZLe7 followed by . 0-0 is better) 12 � ael ! (exploiting the tactical point 1 2 . l¥fxd4 13 '-2l a4 White prepares a breakthrough by '-i:ld5) b5 ( 1 2 �h5 1 3 'i'l!fd2 g6 14 f4 §J_h6 1 5 '!Wf2 g5 16 j}_e2 g4 1 7 �g3 M g8 1 8 §J_ d l ! was also good for White in Nunn-Frank­lin, London 1977, since e5 is imminent, while Franklin's later suggestion of 1 3 g5 allows 14 �d5! with added effect as the 15 square is now available) 1 3 cxb5! �xd4 14 bxa6 j}_c6 1 5 '-21 b5 1¥1 b6 16 §J_d4 'l'l!fb8 1 7 � c l '-i:lc5 18 a4! (this nullifies the threat of � xd3 and prepares to break open the c-file by b4) e5 1 9 §J_e3 Ae7 20 b4 '-i:lxd3 2 1 li xc6 � xb4 22 a7 'll!(b7 23 � b6 (heading for b8) � xe4 24 '-i:lc7 + '1t>d7 25 l¥tb5 + 'f:;xc7 26 li e ! + '-i:lc2 27 M xc2 + Resigns, Nunn-Franklin, Not­tingham 1 979, as 27 *'xc2 28 li b7 + 'it'd8 29 li b8 + 'f:;c7 30 §L b6 is mate.

(2) 3 . . . e6 4 '-i:lc3 �c6 5 d4 cxd4 6 �xd4 A b9 (6 '-21 f6 7 �c2! is good for White) reaching an unusual position which does not seem to be considered by theory. 7 '-i:l c2 j}_xc3 + 8 bxc3 is one possibility, but I like 7 '-i:lxc6.

Then 7 . . • bxc6 8 "iii"d4 looks very awkward since 8 . . . � f6 and 8 . . . "iii"f6 are both met by e5, so 7 . . . dxc6 8 "iii"xd8 + '1Jxd8 is best. Then White plays 9 jl_f4, intend­ing 0-0-0 + and � a4 with good play against the weak black squares at c5 and b6. If Black exchanges at c3 White had the dream square d6 for his bishop .

4 d4 cxd4 5 �xd4 '-21f6

Or 5 . . e5 (5 . . . e6 6 '-i:lc2 is still good for White, and 5 . . . g6 allows 6 '-i:l xc6 and 7 l¥td4) 6 '-i:l f5 d5 (6 �f6 7 '-i:lc3 transposes to Nunn-Surtees) 7 cxd5 Axf5 8 exf5 '-i:ld4 9 '-i:lc3 �e7 (again 9 .:£) f6 transposes) 10 .ztd3 ( 10 f6 is also promising) �exf5 1 1 0-0 jl_d6 1 2 f4 with a dangerous initiative for White.

6 '-i:lc3 e5 7 '-i:lf5 d5

After 7 . . . d6 8 §J_g5 (8 '-i:l e3 controlling d5 is also good) ll_xf5 9 exf5 �d4 1 0 ll_d3 White's con­trol of e4 and d5 gives him an excellent position.

8 cxd5 jl_xf5 9 exf5 �d4

10 jl_d3 �xd5 1 1 0-0 (134) jl_b4

Black has tried a variety of moves in this position, but with­out coming close to equality, for example 1 1 . . . � xc3 1 2 bxc3 '-i:lc6 (once the knight has to move from d4 the only asset of Black's position vanishes) 13 � b l � b8 14 "it f3 fl!c7 15 jl_ e4, Ravinsky­Kliascicki, USSR 1 966, 1 1 . . .

134 B

'-E.)f6 1 2 � e 1 '-E.)c6 1 3 �b3 Ji.b4 14 l! d 1 �e7 15 Ji.g5, Rogacovski­Konova1ov, corr. 1 972 or 1 1 . . . !tJ...e7 1 2 !tJ...e4 �xc3 1 3 bxc3 '-2\c6 14 � b 1 �c8 1 5 'I'Mg4, Matano­vic-Perez, Belgrade 1 96 1 with a clear plus for White in every case.

12 !tJ...e4! '-2\ xc3 13 bxc3 jj_ xc3 14 � b1 0-0

In Altshuler-Fink, corr. 1 960 Black tried to hold on to the pawn but after 14 . . � b8 1 5 �g4 g6 1 6 !tJ...g5 gxf5 1 7 Axf5 f6 1 8 �h5 + White had a winning position.

15 � xb7 This simple move was sug­

gested by Gligoric and Sokolov as an improvement over the unclear continuation 1 5 �g4 �d6 1 6 � d l � ac8 (but not 1 6 )g. ad8? 1 7 � d3 A b4 1 8 f6!) 1 7 � d3 � c4 in which White lacks a knock-out blow.

15 � d6 (135) White's main threat was 1 6

jj_a3, driving the rook away from the defence of f7, followed by � h5 and if Black manages to defend f7 White still has the crushing blow f6 in reserve.

2 a6 157

Black's �d6 is of course designed to prevent jj_a3, but un­fortunately the move loses by force. He had to try 1 5 � b8 although 16 � xb8 �xb8 1 7 f6 gives White a strong attack with no material investment.

16 ;,:, b3! White utilizes the undefended

queen to threaten li.xa8, � xc3 and i_a3. Black's reply is forced.

16 . . . i,f ac8 17 ,&a3 �d8

If the queen moves anywhere else 1 8 j}_xf8 .f)xb3 1 9 j}_xg7 wins a pawn and demolishes Black's kingside.

18 � b7 ,g e8 ( 136)

158 2 a6

19 Ji!\ xti A piece of rather unnecessary

flashiness since 1 9 �h5 ·£ f6 (or 19 '8. c7 20 f6) 20 R_d5 wins quite easily.

19 -;ng5 19 l!txf7 20 ·i'ti"h5 + <ttg8 (20

lfi> f6 2 1 iYi' Xh7 � f7 22 i\-fh5 + forces the king to g8 in any case) 2 1 f6 g6 22 1/. xg6 'i'ii' d7 23 f7 + wins, but Black can hardly hope to survive long after losing the vital f7 pawn.

20 jtd5 'it;h8

21 f6 g6 21 gxf6 22 !1J..e7 costs mater-

ial. 22 �a4 e4

Losing at once, but Black's king would have succumbed soon m any case.

23 if:\ xh7 + Resigns After 23 lfrxh7 24 � d7 +

lt>h6 25 "!Wg7 + *h5 26 'i'l!/h7 + lfrg4 (26 'ii\'h6 27 g4 + lti>g5 28 f4 + exf3 29 . .&c l + ) 27 ·ll'fh3 + ltr f4 28 . .&.d6 + Ji!\ e5 29 "i'rl xc8 the position speaks for itself.

1 4 Unusual Lines

The material in the first 1 3 chapters will be sufficient to pre­pare the reader for the vast major- E ity of the games he will have as White against the Sicilian, but there remain a substantial number

(3 d5) 4 �xd4 iJ..c5 or 4 j_b4 +

Unusual Black second moves: I e4 c5 2 �f3 g6 (2 ·liifc7, 2

b6).

A of unusual variations which Black players might adopt. Only a few have any pretentions to respecta­bility and we concentrate on these few in this chapter. Wilder eccen­tricities are usually best dealt with by an application of common sense and straightforward de­velopment. The following break­down of lines considered in this chapter will aid the reader in locating the variation he is look-ing for.

A The pseudo-Dragon 1 e4 c5 2 .S f3 �c6 3 d4 cxd4 4 �xd4 cEJ f6 5 �c3 g6.

B Unusual lines involving 2 �c6 apart from the pseudo­Dragon: I e4 c5 2 � f3 �c6 3 d4 cxd4 (3 d5) 4 �xd4 � f6 (4 a6, 4 d5, 4 fifc7, 4 . 'i!ii' b6) 5 � c3 l'l!b6.

C Unusual lines involving 2 d6: 1 e4 c5 2 �f3 d6 3 d4 cxd4 3 �f6) 4 �xd4 � f6 5 �c3 e5 (5 �bd7) 6 lt.. b5 +

D Unusual lines involving 2 e6: I e4 c5 2 � f3 e6 3 d4 cxd4

137 w

1 2 3 4 5

e4 �f3 d4 �xd4 �c3

c5 �c6 cxd4 �f6 g6 (137)

This is an attempt by Black to reach accelerated Dragon posi­tions without allow White the op­tion of playing the Maroczy bind. It has been played a few times in Grandmaster chess, but White can obtain a clear advantage with accurate play.

6 �xc6 bxc6 Or 6 dxc6 7 �xd8 + *xd8

8 .f,tc4 *e8 (not 8 . . . b5? 9 JJ..xf7 e6 1 0 JJ..g5 J....e7 1 1 0-0-0 + win-

160 Unusual Lines

ning, while 8 . . . f;J_g7 9 JL.f4 *e8 1 0 0-0-0 l£!d7 1 1 Ac7! is good for White since Black is not allowed to castle) 9 e5 .!E)g4 1 0 f4 h5 (after 10 . . . Af5 1 1 h3 .'£!h6 1 2 g4 Black has to go back since 1 2 jL xc2 13 H h2 j_a4 14 �xa4 b5 1 5 jlb3 bxa4 1 6 f;J_xa4 K c8 17 )ii c2 *d7 1 8 Ae3 is winning for White) 1 1 Ad2 h4 (or 1 1 JL.f5 1 2 h3 €:) h6 1 3 0-0-0 Ag7 1 4 K he l with ad­vantage to White) 1 2 �e4 �h6 1 3 j_c3 h3 14 e 6 K g8 1 5 ex£7 + l£lxf7 1 6 0-0-0! with a clear plus for White, Maus-Schlick, B un­desliga 1 987.

7 eS l£!g8 After 7 . . . � d5 8 l£! xd5 cxd5 9

�xd5 K b8 1 0 e6! (with 'il!J'e5 if the pawn is taken) f6 1 1 JL.f4 li b4 1 2 Ad2 � b6 1 3 11.b5 K d6 14 "�tc4 White has a winning position.

138 B

8 jtc4 ( 138)

8 jtg7 Other moves are no better: ( I ) 8 . . . dS 9 exd6 �xd6 (9

exd6 10 i!ff3 d5 1 1 �xd5 cxd5 1 2 JL.xd5 �e7 + 1 3 j_e3 � b8 1 4 0-0 jtg7 1 5 Af4 with a decisive at-

tack) 10 0-0 'l'!fxd 1 1 1 l{ xd 1 Ah6 1 2 Axh6 �xh6 1 3 K d2 �f5 1 4 �e4 i s very pleasant for White, Geller-Stein, USSR Ch. 1 966--7.

(2) 8 . . . 'l'!fa5 9 0-0 and now: (2a) 9 . . . 'i'!fxeS 10 K e 1 'l'!fa5

(or 10 . . . 'l'!fb8 1 1 "ijd4 f6 1 2 �e4 11.g7 1 3 Af4 'l'!tb6 14 l£!d6 + * f8 1 5 'l'!td3 Ab7 1 6 f;J_xg8 K xg8 1 7 'l'!fc4 Resigns, Tiviakov-Muger­man, Pinsk 1 989) 1 1 b4 'l'!td8 leaves Black in a dreadful mess. In Karaklajic-Ivanovic, Yugoslavia 1 974, White won Black's queen by 1 2 l£le4 e6 1 3 ll_ b2 f6 1 4 ll_xe6 dxe6 1 5 �xf6 + i!fxf6 1 6 ll_xf6 l£!xf6, which proved sufficient in the end, but I would not be sur­prised if White had an even stronger continuation.

(2b) 9 . . . jj_g7 10 !#f3 f5 ( 10 . . . e6 1 1 1L.f4 and 10 . . . f6 1 1 g e l are also good for White) 1 1 it.f4 leads to the note to Black's I Oth move below.

9 'i\'rf3 fS Relatively best, for example 9

e6 10 ]Lf4 %'a5 1 1 0-0! J..xe5 1 2 b4 "ijc7 1 3 � b5 i!fb8 14 JJ..xe5 �xe5 1 5 H ad 1 d5 1 6 H fe 1 'l'!tb8 1 7 ll..xd5 cxd5 1 8 'il!J'xd5 *f8 1 9 iird8 + *g7 20 �c7 � f6 2 1 'W!e7 and White wins.

10 J1.f4 e6 Or 1 0 H b8 ( 1 0 'W!a5 1 1

0-0 Jl.xe5 1 2 b4 i!fc7 1 3 l£! b5 �b8 14 J..xe5 'l'!fxe5 1 5 � fe 1 'il!J'b8 1 6 'l!lfc3 is now instantly decisive) 1 1 0-0 e6 1 2 � ad 1 followed by K fe 1 and just a s i n the main line White has an unpleasant bind, Anders­son-Bilek, Teeside 1 972. Black

has no way to solve the problem of his backward d-pawn and the g7 bishop is inactive . White players must be on the lookout for the exchange sacrifice . � b4xf4, which can be good for B lack if he can get the e5 pawn, but provided White keeps his bind on Black's position he can be op­timistic about the future.

11 0-0 The correct choice; in other

games White played 0-0-0, but this gives Black counterplay down the b-file.

1 1 -2) h6 We give the rest of the game

Short-Sosonko, Wijk aan Zee 1 986, which is a model example of how to play such positions. Black is never allowed to free himself and is finally overcome by the problems resulting from the back­ward d-pawn: 1 2 � ad l 1/ic7 1 3 � fe l � f7 1 4 �g3 0-0 1 5 h4 l:frh8 1 6 �a4 a5 1 7 b3 � e8 1 8 '!We3 h6 1 9 g4! � g8 20 J,.g3 J.. f8 2 1 Ytib6 )!! a7 22 f3 1/ixb6 + 23 � xb6 j_c5 + 24 J..f2 J,.xf2 + 25 frxf2 fxg4 26 fxg4 *g7 27 -2) a4 g5 28 h5 l!! f8 29 *g3 �d8 30 � c5 � f4 3 1 a4 * f8 3 2 J..d3 *e7 3 3 j_g6 l!! a8 34 � e3 .1ii( b8 35 � ed3 � bb4 36 � xd7 + j_xd7 37 .lii( xd7 + Resigns. B l e4 c5

2 � f3 �c6 3 d4 cxd4

3 d5 4 exd5 \ltxd5 5 � c3 �e6 + (or 5 1/ih5 6 d5) 6 J.. e3 cxd4 7 � xd4 �d7 8 � d b5 .1ii( b8 9 �e2 and White is probably win-

Unusual Lines 161

ning already, Boleslavsky-Gurge­nidze, USSR 1960.

4 �xd4 1/ib6 4 . . . a6 c4 transposes to

Chapter 13 , 4 . . . d5 5 -2)c3 dxe4 6 � xc6 �xd l + 7 *xd l bxc6 8 �xe4 j}_f5 9 J.d3 0-0-0 10 *e2 e6 1 1 J..f4 *b7 12 .lii( ad l isjust good for White and 4 . . . � f6 5 ,.£)c3 )li!J b6 transposes to the main line. That leaves 4 tfHc7, which can be met by 5 ,.£) b5 �b8 6 c4 �f6 7 2\ 5c3 e6 8 f4 (8 j}_ e3 allows either 8 . . . j_d6!? or 8 . . . b6 followed by .A_c5) d6 (8 1J..c5 9 e5 -2:Jg8 I 0 2le4) 9 J\i_ e2 1J..e7 10 . .!e3 0-0 1 1 0-0 b6 (White has a favour­able version of Chapter 8 in which Black has lost time with his queen) 1 2 -E:�d2 j_b7 1 3 j_O � d8 14 a3 (better than 14 � c l i f8 1 5 ·¥' e2 d5!? with unclear play, Chandler-Barlov, Haninge 1 988) 1J..f8 15 j}_f2 -E:�d7 16 b4 with advantage to White, Karpov­Kurajica, Hastings 1 97 1 /2.

5 -E:�b3 ,.£)f6 6 -E:�c3 e6 7 J..e3 'it!Jc7 8 j}_d3 ( 139)

162 Unusual Lines

8 'ke7 Or 8 a6 (8 ,k b4 9 0-0 0-0

10 2:!b5 io� b8 1 1 f4 was good for White in Gheorghiu-Forintos, Monte Carlo 1 968) 9 f4 d6 I 0 � f3 (10 �e2 is also possible, as in some examples below, while White may start his kingside pawn advance immediately, e.g. 10 g4 b5 1 1 g5 2:!d7 12 M- d2 ii_b7 13 0-0-0 -2:)c5 14 '!r,) f2! with some advantage to White, Belyavsky­Gufeld, Suhumi 1 972) b5 (if Black plays k_e7 we transpose to the main line below) 1 1 0-0-0 ii_b7 1 2 'i11i> b l 2\a5 1 3 2\ xa5 ii'fxa5 14 g4 0-0-0 1 5 g5 2\ d7 1 6 a3 lfi b8 1 7 !l'i f2 1i.e7 1 8 k d4 e 5 19 fxe5 dxe5 20 .Jo. a7 + 1Wa8 21 2:!d5 with a clear plus for White, King-Wirth­ensohn, Berne 1 988.

9 f4 d6 10 f'>'if3

It is useful for White to delay committing his king, since he can reasonably castle on either side. However, there is an argument for developing the queen to e2, for example 10 ·ri'ie2 a6 and now l 1 0-0-0 0-0 1 2 g4 ii;; e8 1 3 g 5 � d 7 14 ld f3 g6 1 5 )"( afl b5 1 6 8 h3 b4 17 -2:\d I , A. Rodriguez-Carlier, Amsterdam 1 987 or l 1 0-0-0 b5 1 2 1Wbl 2:! b4 1 3 g4 2:!d7 14 g5 .Ja b7 15 a3 2\ xd3 16 cxd3 g6 1 7 ;,;; c l .2Jc5 1 8 2l d4 ·M d8 1 9 h4, Todorovic-Bosic, Novi Sad Open 1988, with good attacking chances for White in both cases. In Wedberg-Benko, New York Open 1989, the continuation I 0 0-0 0-0?! (castling into the storm is

wrong; 10 a6 was better) 1 1 g4 (White is even better off than in the main line, as his queen may go directly to h5) �d7 12 g5 ii;;� e8 1 3 ii;;l f3 a6 14 ii;;� h3 gave White a massive attack; the finish was 1 4

� f8 1 5 'i'i1h5 �b4 1 6 ii;;l fl jj_d8 17 a3 � xd3 18 cxd3 b5 1 9 f5 exf5 20 Qjd5 �d7 2 1 ji_d4 fxe4 22 ii;;� h4 e3 23 �xe3 :g e6 24 � hf4 � g6 25 � xf7 'tl!fxf7 26 :g xf7 � xg5 + 27 !¥'xg5 jLxg5 28 lll: xg7 + Resigns.

10 a6 ( 140) Black also delays castling since

1 0 0-0 1 1 g4 l;:;! e8 1 2 g5 .z:,d7 13 h4 � b4 14 h5 jLf8 1 5 0-0-0 a6 1 6 g6 gave White an immense attack in Jansa-Martinovic, Vrnjacka Banja 1 982.

l1 g4 White has a choice of good

lines. After 1 1 0-0 0-0 (not 11 . . . b5 1 2 e5, but castling invites the kingside pawn storm, so 1 1 .�d7 may be best) 1 2 g4 (in some games White played ii;;\ ae l , but this preparation is not necessary) Black has fared very badly in practice, for example 12 . . . lli\ e8

1 3 g5 �d7 14 �h5 � f8 1 5 f5 �e5 1 6 f6 �d8 1 7 � d4 b5 1 8 � £2 l.:. b7 1 9 � afl b4 20 �ce2 � fg6 2 1 .£lg3, Hawelko-Sznapik, Slupsk 1988 or 12 . . . �b4 1 3 g5 -2)d7 14 'i>jh5 g6 1 5 �-h6 � e8 16 � ad l b5 1 7 a3 �xd3 1 8 � xd3 1.:.f8 19 �h4 lt_b7 20 j;_d4, G. Mainka-Martinovic, Dortmund 11 1 988, and White's attack I S

overwhelming in both cases. 11 b6

This rather odd move is designed to support c5 in anti­cipation of the manoeuvre �d7--c5. 1 1 . . . h6 1 2 0-0-0 b5 1 3 � hg l '-2\ d 7 14 wf2! (this possi­bility explains Black's preference for b6 in the main line, since once he has played b5 the knight on d7 is hard to redeploy) l.:. b7 1 5 <ttr b l l.:. f6? (Black's 6 renders the kingside too danger-ous for 0-0-he should have played 1 5 2l b4) 1 6 e5! dxe5 1 7 j_ xb5 0-0 1 8 g5 hxg5 1 9 fxg5 �e7, Estrin-Kopylov, USSR 1 973, and now 20 -�-h4! axb5 2 1 � g3 would have given White a decisive attack according to Estrin. Perhaps Black should try 11 . . . b5 but White still has the advantage.

12 g5 2!d7 1 3 0-0-0 .£lc5 14 w b l (to

answer .£lxd3 by cxd3) iJ_d7 15 h4 i'<1b7 1 6 JJ_e2! (now the c5 knight isn't doing much) 2\a7 1 7 f5 2\ b5 1 8 li_ d4 kc6 19 fxe6 -2l xc3 + 20 ji_ xc3 2\ xe6 2 1 � hfl 0-0 22 lt.d3 b5 23 a3 2\c5 24 2\xc5! dxc5 25 i-i f5 h ae8 (or 25

Unusual Lines 163

c4 26 ioii e5 f6 27 gxf6 _&xf6 28 i»e6 + ;..: f7 29 ii xf6!) 26 l.:.xg7! kd7 (26 '/ixg7 27 e5) 27 ·ffl"e5 il_d8 28 i<i d6 rk;xg7 29 � h6 + lfth8 30 e5 f5 3 1 exf6 .i_e6 32 £ de I c4 33 li_g6 � f7 34 JJ_xf7 � xf7 35 g6 i>i'xg6 37 f7 Resigns, Kavalek-Hiibner, Buenos Aires 1978. C l e4 c5

2 2\f3 d6 3 d4 cxd4

Black quite often plays 3 -2l f6 in order to avoid the line 3

cxd4 4 � xd4. White should reply 4 2\c3 when Black is obliged to play 4 cxd4 5 -2)xd4 trans­posing to normal lines.

4 -2l xd4 2\ f6 5 2\c3 e5 ( 141)

Or 5 -2l bd7 6 JLc4 21 b6 (Black's development is not easy because e6 and .1J. .e7 will allow a il_xe6 sacrifice, while 6 g6 7 f3 kg7 8 ji_e3 0-0 9 'iti'd2 is a Dragon in which Black has de­veloped his knight to the inferior square d7) 7 k b3 e5 8 �de2 it_e6 9 .1J. g5 .i_e7 10 . .&.xf6 ji_xf6 1 1 2\d5 2\xd5 12 it_xd5 ·;,;;- b6 1 3

141 w

164 Unusual Lines

jtb3 0-0 14 c2Jc3 and White's con­trol of d5 gives him a clear advan­tage, R. Byrne-Cuellar, Siegen 1 970.

6 ll_b5 + One of the points of 5 a6 is

to prepare e5 by preventing this move, so it is the only logical reply to 5 e5.

6 !Elbd7 After 6 ll_d7 7 jtxd7 +

�xd7 8 !Elf3 (8 !El f5 allows Black to complicate the issue by 8 �xe4) the exchange of white­squared bishops enhances the weakness of d5.

7 �f5 a6 8 j'j_ xd7 + 'Wfxd7 (142)

The critical moment. White has a number of possible plans, but it is not clear which is the best:

( 1 ) 9 jigS � xe4 10 � xg7 + jtxg7 1 1 � xe4 and now:

( l a) 1 1 . . . d5?! 1 2 � f6 + jtxf6 1 3 jtxf6 0-0 ( 1 3 � g8 14 0-0 is also very bad) 1 4 '/iYd3 e4 1 5 llt'd4 � e8 16 g4! �d6 1 7 0-0-0 'V!ff4 + 1 8 lfl b l � e6 1 9 g5 � e8 20 j_h8 f6 2 1 gxf6 'V!ff5 22 � hg l + fif7 23 � g7 + lfje6 24 itb6 + Resigns,

Camacho-Cruz Lima, Cuba 1986.

( l b) 11 . . . 0-0 (this pawn sacri­fice is the point of the variation, but it may not be correct) 1 2 �xd6 f6 ( 1 2 �f5 1 3 <El f6 + Axf6 1 4 �xf6 'lj'e4 + 1 5 lfl fl jth3 1 6 f3 �c4 + 1 7 *e l �b4 + 1 8 jtd2 is good for White) and I doubt if Black has enough for the pawn. After 13 jte3 �g4 14 �c3! jtf5 15 �d5 + � f7 16 h3 �g6 1 7 0-0-0! H c8 ( 1 7 jtxc2 1 8 H d2 jtf5 1 9 g4 is good for White) 1 8 � d2, Klovan-Mocalov, USSR 1981 or 13 �xd7 jtxd7 14 j_d2 Af5 15 f3 Jlg6 16 0-0-0 � ac8 1 7 jtc3 b5 1 8 a3 H c6 1 9 H d7 jtf5 20 1i;!t a7 jth6 + 2 1 jtd2 jtxd2 + 22 lflxd2, Perenyi-Bielczyk, Ber­lin Open 1 988, White had the ad­vantage and although Black managed to draw the first of these games the ending cannot be pleasant for him.

(2) 9 <Ele3 �c6 and now: (2a) 10 �d3 (this gives White a

small but safe advantage) ll_e6 1 1 0-0 1i;!t c8 1 2 Ad2 ( 1 2 a4 jte7 1 3 �cd5 jtxd5 14 exd5 'i!lfc7 1 5 a5 g6 1 6 b3 0-0 1 7 �c4 H fe8 18 .il_e3 A f8 was equal in Popovic-Rajko­vic, Vrsac 1 987) jte7 ( 1 2 . . . g6 is possible, but White is still slightly better) 1 3 � cd5 jtd8 (now 1 3 . . j}_xd5 14 exd5 f!fc7 1 5 <El f5 is good for White; jtd2 is much more useful than a4) 14 c4 0-0 1 5 !;[ ac l � e8 1 6 b3 �d7 1 7 �b4 with some advantage for White, Ehlvest-Kupreichik, Moscow TV 1 987.

(2b) 10 �ed5 �x:d5 1 1 �xd5 il_e6 0-0 �c8 13 c4!? �·xc4 14 �b6 �xe4 15 �xc8 il_xc8 1 6 £ e l �g4 1 7 f3 with a small plus for White, Krnic-Jovanovic, Yugoslavia 1 982.

(2c) 10 0-0 .£\xe4 ( 10 1Le6 is

Unusual Lines 165

some point, but attempts to exploit this directly don't work. White has to be a bit more careful to gain the advantage against 4

.k.c5. 5 2lb3 kb6 ( 143)

probably better, when White may 143 have nothing better than 1 1 �d3 w transposing to line 2a) 1 1 .£\xe4 �xe4 1 2 �d5 � b8 1 3 b3 J. f5 14 c4 f6 1 5 il..a3 ltrf7 1 6 �d2 � d8 1 7 � fe l 'i!l(g4 1 8 �e3 �·g5 1 9 �d5 + with advantage to White, L. Schneider-Bator, Sweden Ch. 1 986.

(2d) 10 )li!ff3!? b5 1 1 0-0 j_b7 1 2 �cd5 -2\xd5 1 3 -2\xd5 iifxc2 14 jigS il_xd5 1 5 exd5 f6 1 6 � ac l 'i!{(g6 1 7 kd2 .zte7 18 � c7 with good compensation for the pawn, Kudrin-Conquest, London 1 986. D 1 e4 c5

2 2!f3 e6 3 d4 cxd4

3 d5 4 exd5 exd5 5 1t. b5 + -2\c6 gives Black an uncomfor­table isolated pawn position after 6 -2\c3 or 6 0-0.

4 -2\xd4 _k c5 Or 4 k b4 + , when White

can transpose to chapter 1 1 by 5 -2\c3 -2:1 f6, but it is also possible to play 5 c3 .k.e7 6 c4, with a Mar­oczy bind position.

The idea of 4 .k.c5 is to reach a position similar to that after 1 e4 c5 2 -2:1£3 e6 3 d4 cxd4 4 2J xd4 a6 5 ]Ld3 .�.cS (see game 2 1 ), but without wasting a tempo on a6. Naturally this exposes Black to the possibility of -2\ b5 at

6 .EJc3 Simple development guarantees

at least a slight advantage. The ambitious 6 c4 is also promising, for example 6 c4 -2\ e7 7 -2\c3 (White must prevent d5) 0-0 (or 7 -2\bc6 8 ]Lf4 e5 9 _kg5 f6 10 _.ik d2 d6 1 1 ·@ h5 + 'i\>f8 12 .ik.d3 . .k..e6 1 3 0-0 and White is better, Howell-S. Arkell, London 1 986) 8 il.f4 (it is essential to reach d6 with the bishop before Black prevents it with . f5, e.g. 8 jie2?! f5 and _k. f4 is impossible) f5 9 .1L d6 -2\ bc6 10 1/._ e2 and now if Black plays 10 . . . fxe4 1 1 c5 1Lc7 1 2 -2\ xe4 White's hold on d6 cannot be broken, while after 10 . . . f4 aiming to play .dg6-e5 (after a rook move, of course) White might even consider 1 1 il.h5!?

6 J"ije7 7 1L.d3

166 Unusual Lines

Or 7 ji_g5!? f6 8 jj_ h4 0-0 9 ·M' h5 o.dbc6 10 0-0-0 o.de5 l l jj_g3 o8 7g6 1 2 lti> b l f5 1 3 f4 o.dc6 14 ..2.c4 'f:& f6 1 5 e5 ii-ie7 1 6 � b5 a6 1 7 o.dd6 ltih8 18 h4 -?) a5 19 -2) xa5 Resigns, Sibarevic-G. Welling, Lugano Open 1 989. Weak play by Black, but this direct plan could be dangerous.

7 0-0 8 0-0 o.dbc6 9 jj_f4 f5 (9 d5

10 exd5 <dxd5 l l .£) xd5 '!f(xd5 1 2 c4 'l!il(d8 1 3 j_d6!) 1 0 Jt.d6 f4 l l �h5 f3 1 2 g3 JrLc7 1 3 e5 g6 14 �g5 §txd6 15 exd6 .£)f5 1 6 'iilf xd8 Q)xd8, Wedberg-Nunn, Helsinki 1 983, and now 1 7 11_xf5 iiil xf5 1 8 Z!d4 l\i f8 1 9 a4! a 5 20 � a 3 would have been good for White. E l e4 c5

2 -tl f3 g6 One of the most important lines

in this chapter, since it has oc­curred many times in Grandmas­ter chess and White can probably only secure an edge against it. Other second moves are very unusual and can be met by nor­mal development, e.g. 2 . . . b6 3 d4 cxd4 4 o.d xd4 j_b7 5 Z!c3 Qjc6 (or 5 . . . a6 6 j_d3 g6 7 f4 !JJ...g7 8 <d O d6 9 0-0 followed by 'il'.\'e l-h4 with attacking chances) 6 _1 f4 � c8 7 o.dxc6 dxc6 8 "i¥f0 'l!iifd4 9 � d l 'l!l1c5 1 0 e5 i!j\ d8 l l � xd8 + 'ltrxd8 1 2 j_e2 *e8 1 3 0-0 f5 14 e6 Z!f6 1 5 � d 1 �d5 16 jj_e5 Resigns, Belyavsky-Quinteros, Vienna Open 1 986, or 2 . . . 'f!tlc7 3 c3 (it is probably not a good idea to play 3 d4 since 3 cxd4 4 .£)xd4 Z!f6 5 o.dc3 a6 followed by

e5 gives Black a type of Naj­dorf position in which his king's bishop can still be developed ac­tively at c5 or b4) .:2) f6 4 e5 .:2)d5 5 d4 cxd4 6 cxd4 d6 7 4Q a3 a6 8 !JJ...d3 e6 9 4Qc4 dxe5 10 dxe5 b5 1 1 4Qe3 �b7 1 2 0-0 followed by a4 when Black's queenside pawn structure will be weakened.

3 d4 The attempt to reach a Mar­

oczy bind position by 3 c4 !1.tg7 4 d4 can be met by 4 �a5 + , when it is not at all easy for White to maintain any advantage.

3 !JJ...g7 3 . . cxd4 4 .:2)xd4 transposes to

lines examined earlier, for ex-ample 4 .:2)c6 5 c4 and 4 !JJ...g7 5 c4 end up in Chapter 7, 4

4Qf6 5 .:2)c3 d6 is chapter 5, and 4 . . . .:2) f6 5 .:2)c3 .:2)c6 leads to line A in this chapter.

4 dxc5 *a5 + (144) 4 .:2) a6 5 J..xa6 �a5 + 6 c3

'i'llfxa6 7 �e2 -tWc6 8 J..e3 'il'l'xe4 9 .:2)bd2 '/itc6 1 0 0-0 4Q f6 1 1 .:2)d4 'l!il(c7 1 2 .£) b5 'i'ir'd8 1 3 !JJ... f4 was good for White in Rajna-Nagy, Hungary 1 960.

5 -2Jc3 Natural, but 5 c3 may be

stronger, for example 5 "" xc5 6 _ke3 )5dc7 7 .ik.d4 e5? (7 -2l f6 8 e5 -2\g4 9 -2l a3 ! is better, but White can still claim a modest advantage) 8 li_e3 -2l f6 9 -2J a3 ! 0-0 10 -2l b5 i<dc6 I I -2)xe5 -;.Ii xe4 1 2 -2Jxf7! with a large plus for White, Maric-Tringov, Bar I 977.

5 -2)f6 Or 5 JLxc3 + (5 � xc5 6

Bd5 e6 7 b4 )i-; f8 8 -2)c7 + ltrd8 9 -2J xa8 _kxa l 1 0 .:;ig5 + it.J6 1 1 JLxf6 + -2Jxf6 12 � d4 i'tle7?! 1 3 1L. b5 b6 14 -2l xb6! axb6 1 5 iii xb6 + t\'e8 1 6 0-0 was very good for White in Mohrlok­Breazu, corr. I 987) 6 bxc3 "i'<> XC3 + 7 li_d2 i>i"XC5 8 _kd3 (8 1L e2 -21 f6 9 e5 -21 g4 10 0-0 -2l xe5 1 1 11_e3 -2) xf3 + 1 2 j_xf3 fi, c7 1 3 ·i>i" d4 was also promising in Petrov-Limonikov, corr. I 974) -2) f6 (8 d6 9 0-011_g4 1 0 £ b l ·if[c7 1 I g b3 -2:! d 7 1 2 ;,; c 3 -2lc5 1 3 h3, Ambrosz-Petran, Czechoslo-

Unusual Lines 167

vakia I 979 gave White more than enough for the pawn) 9 0-0 0-0 1 0 e5 2}g4 I I g b1 2}xe5 1 2 g b5 -2:! xf3 + 1 3 �xf3 ·iifc7 14 JL.h6 � e8 15 g f5 g f8 (New in Chess gave . f6 as unclear, but 1 6 g e 1 ! appears crushing after 1 6 tiic3 I 7 .k c4 + ! ·ii'i' xc4 1 8 g xf6) I 6 JL.xf8 gxf5 1 7 JLxe7 d5 1 8 � e1 .k.e6 1 9 J.:.f6 -2)d7 20 .it..d4 ii'i'd6 21 k xf5 k xf5 22 -� xf5 -2:! f8 23 i!i' g5 + -21 g6 24 h4 g f8 25 h5 Resigns, Frivaldszky-Monostori, corr. 1986.

6 it..d3 "ffl' XC5 7 _ke3 ·;.<; a5

7 'if! h5 is possible, but I cannot find any practical ex­amples of it .

8 iii d2 -2\ c6 9 0-0 0-0 1 0 h3 d6 1 I a3 . .k .e6 I 2

-2)g5 d5 ( 1 2 1J. d7 1 3 f4) 1 3 exd5 kxd5 1 4 b4 i>ii d8 1 5 � ad ! and White has a slight advantage, Sveshnikov-Romanishin, USSR eh. 1 977.

Index of Variations

1 e4 c5 2 �f3

Now: A B c D

A

2 2 2 2

ci)c6 d6 e6 others

2 . . • �c6 3 d4 cxd4

3 d5 161 4 �xd4 �f6

4 *'b6 161 4 d5 161 4 ·r!fc7 161 4 a6 5 c4 156 4 0 0 0 e5 5 � b5 a6 ( 1 5 � f6 6

� l c3 63; 5 d6 140-1; 5 j}_c5 6 � I c3 .t) f6 63; 5 0 h6 140) 6 �d6 + 1t.xd6 7 tifxd6 141

4 g6 5 c4 � f6 (5 j}_g7 6 j}_e3 1 1 7) 6 � c3 d6 (6 � xd4 Ill) 7 j}_e2 j}_g7 (7 ci)xd4 Ill) 8 j}_e3 (8 �c2 I l l) 1 19

4 e6 5 � b5 d6 (5 �f6 6 � l c3 135; 5 1t.c5 125) 6 c4 � f6 (6 a6 126) 7 � l c3 a6 8 �a3 j}_e7 (8 0 0 0 d5 126) 9 jj_e2 0-0 1 0 0-0 b6 ( 1 0 others 127) 1 1 ;Le3 �e5 ( 1 1 others 127-30) 1 2 f4 �ed7 130

5 Q}c3 d6 5 0 0 0 e6 6 .£�db5 J.. b4 (6 0 d6 7

J.. f4 e5 8 j}_g5 -see 5 e5) 7 a3 ;Lxc3 + 8 �xc3 135

5 g6 159 5 0 0 0 e5 6 �db5 d6 (6 0 0 0 others

63) 7 ;Lg5 a6 (7 j}_e6 63) 8 � a3 b5 (8 0 !1Le7 63; 8 0 0 0 d5 64; 8 Jte6 64) 9 §J..xf6 gxf6 ( 1 9 �xf6 68) 10 � d 5 f5 ( 1 0 jj_g7 68-70) 1 1 j_d3 !1Le6 12 0-0 !1Lg7 ( 1 2 Jtxd5 70) 1 3 �h5 71

6 ;Lg5 e6 6 Jtd7 7 Jl_e2 59 6 g6 57 6 �a5 58 6 '/!!fb6 58 6 a6 58

7 �d2 a6 7 h6 40 7 � xd4 41 7 1J..e7 8 0-0-0 0-0 (8

�xd4 41; 8 a6 49) 9 <2J b3 a5 (9 d5 43; 9 �a5 43; 9 h6

43; 9 a6 43-4; 9 t'rlb6 44-5) I 0 a4 d5 1 1 iL b5 46

B

8 0-0-0 h6 8 $t e7 49 8 . .1d7 9 f4 b5 (9 h6 50; 9

k e7 51) 10 _k xf6 gxf6 52 9 :;ie3 1J..d7

9 others 54 10 f4 55

2 3 d4 4 :£:)xd4 5 .£�c3

d6 cxd4 .£�f6

Index of Variations 169

Now: 12 h4 h5 B 1 5 e6 1 2 j.i·a5 85 B2 5 g6 1 2 a5 85 B3 5 a6 1 2 -2Jc4 86 For 5 Bc6 see A and for 5 13 k.g5 ;g_ c5

e5 and 5 -2\ bd7 see p. 1 63. 13 a6 87 1 3 -2Jh7 87

81 1 3 -2Jc4 87 5 ... e6 14 '!irb1 90 6 g4 h6

6 e5 32 83 6 -2Jc6 7 g5 Bd7 8 h4 33 5 a6 6 a6 7 g5 -2\ fd7 8 h4 36 6 f4 e5 6 1J..e7 7 g5 -2\ fd7 8 h4 37 6 <2Jc6 2

7 h4 -2Jc6 6 g6 2 7 k.e7 20 6 e6 7 'i!i;f3 2 7 a6 8 k g2 21 6 �bd7 7 �e2 3

8 )4 g1 h5 6 . 'i!i;c7 7 <-2lf3 <2J bd7 (7 e6 8 j_d7 26 8 �d3 6) 8 jj_d3 7 8 d5 26 7 �f3 �bd7

9 gxh5 i!1 Xh5 28 7 fijc7 8 a4 9 8 a4 JJ..e7

82 8 f;tc7 9 5 g6 8 d5 12 6 .1J..e3 1J..g7 9 $td3 0-0

6 a6 78 10 0-0 12 7 f3 0-0

7 a6 78 c 7 -2\c6 78 2 e6

8 •i>i-!d2 -2\c6 3 d4 cxd4 8 d5 79 3 d5 165

9 j_c4 ii.d7 4 � xd4 a6 9 a6 80 4 Ac5 165 9 a5 80 4 $t b4 + 165 9 -2\xd4 80 4 � c6-see A 9 �¥·a5 80 4 � f6 5 �c3 jtb4 (5 9 �d7 80-1 d6-see B 1 ; 5 � c6 134) 6 e5

10 0-0-0 � c8 145 1 0 ftiC7 82 5 Ad3 �f6 1 0 ·;;.;-b8 82 5 � c6 94 1 0 �a5 1 1 .k.b3 83 5 g6 95

1 1 JJ..b3 -2\e5 5 . . . :t:J e7 96

1 70 Index of Variations

5 �b6 96 5 �c7 96 5 b5 96 5 A_c5 6 � b3 96 5 d6 6 0-0 � f6 10/

6 0-0 d6 6 'ft(c7 101-2 6 e5 101

7 c4 102

D 2 �f6 149 2 a6 155 2 g6 166 2 b6 166 2 'fife? 166

BATSFORD

publish the world's best

CHESS BOOKS

written by the world's best

CHESS PLAYERS

KASPAROV - KARPOV - SHORT

NUNN - SPEELMAN - CHANDLER

FLEAR - SMYSLOV - KORCHNOI

ADOR]AN - KOTOV - KEENE

PLASKETT - HODGSON - RETI

NIMZOWITSCH - TAIMANOV

LASKER - SVESHNIKOV

POLUGAYEVSKY - SUETIN

STAl JNTON