American Batsford Chess Library

Beating the Anti-King's Indians

Joe Gallagher

An ICE Book
International Chess Enterprises, Seattle
International Chess Enterprises, Inc.
2005 Fifth Avenue, Suite 402
Seattle, Washington 98121-2850

P.O. Box 19457

Seattle, Washington 98109-1457

First published 1996

Copyright© 1996 by Joe Gallagher

All rights reserved. No part of this book may be

reproduced, by any means, without prior permission
of the publisher.

Typeset by John Nunn

and printed in Great Britain by
Redwood Books, Trowbridge, Wilts
for the publishers,
B. T. Batsford Ltd, 4 Fitzhardinge Street, London WlH OAH

British Library Cataloging-in-Publication Data.

A catalog record for this book is
available from the British Library.

First published in the United States in 1996 by

International Chess Enterprises, Inc.
Originally published in Great Britain in 1996 by
B. T. Batsford.
Library of Congress Catalog Card Number: 96-077758
ISBN 1-879479-36-2 (An ICE Book: pbk.)

First American edition - 1996

Printed in the United Kingdom

All first editions are printed on acid-free paper

A BATSFORD CHESS BOOK

Editorial Panel: Mark Dvoretsky, Jon Speelman
General Adviser: Raymond Keene OBE
Specialist Adviser: Dr John Nunn
Commissioning Editor: Graham Burgess
Contents

Symbols 4
Introduction 5

Section 1: King's Indian

1 Four Pawns Attack 9

2 h3 Systems 31
3 Averbakh 53
4 White plays Bg5 72
5 Exchange Variation 85
6 5 i.d3 91
7 5 llJge2 97
8 Unusual Lines 104

Section 2: White plays without c4

9 Trompowsky 111
10 Torre Attack 137
11 London System 148
12 Kingside Fianchetto 157
13 Veresov 170
14 Barry Attack 180
15 Blackmar-Diemer Gambit 185

Index of Variations 190

Symbols

+ Check
++ Double Check
# Mate
Good move
" Excellent move
? Bad move
?? Serious blunder
!? Interesting move
?! Dubious move
;!; Small advantage to White
+ Small advantage to Black
+ Large advantage to White
+ Large advantage to Black
+- Decisive advantage to White
-+ Decisive advantage to Black
aO Unclear position
- Equal position
1-0 White wins
0-1 Black wins
112-l/2 Draw
Ch Championship
Echt European team championship
Wch Wor ld championship
Wcht World team championship
OL Olym piad
z Zonal
IZ Interzonal
Ct Candidates event
corr Correspondence g ame
(n) nth match g ame
(D) Diagram follows
Introduction
'What on earth are the Anti-King's
Indians?' you must be asking your­
selves. Well, I have taken the liberty
of defining them as all variations of
the King's Indian except for the
Classical, the Samisch and the Fian­
chetto; plus all the lines where White
doesn't play an early c4 (Trompow­
sky, Torre Attack, etc.).
Much literature has been devoted
to the King's Indian in recent years,
but a large percentage of it has con­
centrated on the 'main lines' . For
example, Nunn and Burgess have
produced a mammoth 640-page
work (in two volumes) uniquely on
the Classical Variation, while I my­
self chipped in with a 240-page ef­
fort on the Sarnisch and I believe that
Batsford have a project on the Fi­
anchetto variations in the pipeline.
Even books dealing with the whole
King's Indian tend to treat our vari­
ations as an afterthought. For ex­
ample, The Complete King's Indian
by Keene and Jacobs (Batsford
1992), devotes a mere 13 pages (out
of 272) to the variations covered in
this book.
The neglect of these 'Anti-King's
Indians' seemed a little unfair to me.
Taking my own games as an example
I found that about 50o/o of my King's
Indian's over the last five years have
been Classicals, Samisches or Fi­
anchettoes while the other 50% have
been made up from the King's Indian
lines in this book. And, what's more,
these figures do not take into account
the 25% or so of my games in which
White played 1 d4 and didn't follow
up with c4. The idea for this book
was beginning to take shape.
Of course the material was much
too vast to consider an extensive ref­
erence book, so the by now familiar
concept of a repertoire book was the
answer. Against each of the vari­
ations in this book I have selected
one main defence for Black but you
will also find plenty of alternatives in
the notes in case the main line ever
runs into trouble.
As a quick overview, here are the
principal recommendations against
each system:
1 Four Pawns Attack: 6 ... lDa6
2 h3 systems: Main line with
6...e5, although Black can also play
. . . ltJa6 first.
3 Averbakh: 6... ltJa6
4 Early i.g5: Benoni style ... c5
5 Exchange variation (strictly
speaking this is a Classical but it may
also be considered as the ultimate
Anti King's Indian): Old line with
9 . .. Ite8 based on 13...ltJd7.
6 5 i.d3: 6...ltJc6 and 7...ltJh5
7 5 ltJge2: a quick ...a6 and ... c6
8 Unusual Lines (King's Indian):
see chapter 8
9 Trompowsky: 2.. . ltJe4
6 Introduction
10 Torre Attack: 4 ...0-0, delaying
the central strike until White has re­
vealed his set-up.
11 London System: Playing for
...e5
12 Fianchetto Variations (with-
out c4): Pirc style set-up.
13 Veresov: 3 ...lbbd7
14 Barry Attack: Quick ...c5
15 Blackmar-Diemer: Take the
money and run.
I have tried to vary the type of de­
fence that Black adopts as recom­
mending ...e5 against everything
would have been a little dull. Per­
haps this may seem like extra work
for the reader, but in the long term,
varying your approach will have
beneficial effects on your game and
increase your understanding of chess
in general.
Although this book has 'King's
Indian' in the title many of the lines
in the second section (Chapters 9-
15) will be of interest to anyone who
plays 1. ..lbf6 against 1 d4 (and even
to those who play 1...d5 in the case
of the Blackmar-Diemer Gambit). It
has to be said, though, that the de­
fences I have selected are geared to­
wards the aggressive King's Indian
player rather than the solid Queen's
Indian exponent.
A lot of the lines in Section 2 have
been neglected by the very top play­
ers but amongst everyone else (and
that includes your average grand­
master) they are extremely popular. I
rarely play a tournament without
having to face at least one of them.
The Trompowsky is, in fact, the larg-
est chapter in this book (even though
I have only examined the one line
2... lt:Je4) and the single variation you
are the most likely to face. As the
world is still awaiting Hodgson's
version of events this line is covered
in considerable detail.
The material is examined through
the context of twenty nine complete
and annotated games. This is impor­
tant as .I believe that it is impossible
to get to grips with an opening if you
only ever study the first fifteen
moves or so. Where variations are
new to you I think a good approach
would be to concentrate on the an­
notated games and the text, only
turning to the fine print when you
have grasped the basic ideas and
have perhaps played a game or two
in the line. I'm sure you will find that
learning theory (if that is your de­
sire) will be a much simpler and less
unpleasant business once you have a
few practical outings behind you.
A quick word about 'Beating'
which appears in the title of this
book is in order. Don't expect to get
a winning position out of the open­
ing every time as this is impossible,
especially with the black pieces.
What I have aimed for is double­
edged middlegame positions in
which Black can confidently play for
the win. Even this has been ex­
tremely difficult in some cases
(Chapters 5 and 12 spring to mind)
but if White is hell bent on a draw
then there is very little you can do
except for outplaying him in a drawn
position, which will, of course, give
you a great deal of satisfaction.
 Introduction 7

Over the years I have had a great their soundness. I hope they bring
deal of experience with many of the you as many points and as much en­
lines in this book and can vouch for joyment as they have brought me.
1 The Four Pawns Attack
The Four Pawns Attack, in which
by the fifth move White has already
constructed an enormous centre
stretching from c4 to f4, is undoubt­
edly White's most ambitious set-up
against the King's Indian. In the
early part of this century such an edi­
fice would have been regarded as a
decisive advantage and the player of
the black pieces ridiculed for such
weak opening play. Then along came
the hypermoderns who taught us that
there are ways of battling against
such centres. They pointed out that
while Black has been concentrating
on development and getting his king
into safety, White has invested valu­
able tempi on the construction of his
centre. It follows, therefore, that
Black must strike quickly and try
and open the position before White
can consolidate the space advantage
that his centre has gained him. Until
recently it was assumed that the only
acceptable way for Black to do this
was by playing a quick ... c5, but
times have changed and now I am
able to recommend a system which
is based mainly on Black playing for
... e5. As the immediate 6 ...e5 is pre­
mature (7 dxe5 dxe5 8 l'bxe5 is good
for White) this advance has to be pre­
pared and the best way of doing this
is with 6...l'ba6. The move ...l'ba6, in
general, has breathed new life into
many variations of the King's Indian
and the Four Pawns Attack did not
escape this phenomenon.
One of the main reasons for play­
ing this system is that the theory is
still undeveloped. For example, one
of the most important sources of
opening theory is the Encyclopaedia
of Chess Openings. Volume E, pub­
lished in 1991, considered 6...l'ba6 to
be worth just one line {plus foot­
notes) out of a 12 page coverage on
the Four Pawns Attack (and this
doesn't even include the main line,
6...c5 7 d5 e6 8 �e2 exd5 9 cxd5,
which is classified as a Benoni and
dealt with in Volume A). Many other
books on the King's Indian hardly
mention, or don't mention at all,
6.. . l'ba6 including some published
well into the 1990s. Only Burgess's
The King's Indian for the Attacking
Player (Bats ford 1993) deals with
this variation in a thorough manner
and this is a book that your average
Four Pawns practitioner is unlikely
to possess as it is principally aimed
at the black player.
This does not mean, though, that
6... l'ba6 is just some tricky little side
line; in fact at international level it
is now the most common choice
against the Four Pawns and has re­
cently received no less than Kas­
parov's seal of approval. The Four
Pawns Attack, though, has never at­
tained the popularity of the Classical
10 The Four Pawns Attack
or Samisch variations, and conse­
quently, theory moves at a snail's
pace in comparison. Much of the
credit for developing this system
belongs to the Russian master Igor
Belov although several other players
jumped on the bandwagon pretty
quickly (I can number myself amongst
them).
The first two games below deal
with 7 i.e2 (and unusual 7th moves
for White) against which Black
should play 7... e5 ! ; Game 1 exam­
ines White accepting the pawn sacri­
fice, 8 dxe5 dxe5 9 1Vxd8 (9 lDxe5)
9 ... :xd8 10 lDxe5 , while Game 2
deals with the positional alternative
8 fxe5 dxe5 9 d5. The two other prin­
cipal 7th moves for White, 7 i.d3
and 7 e5 are dealt with in Game 3.
Game 1
Naumann - Gallagher
Hastings 199011
1 d4 llJf6
2 c4 g6
3 l2Jc3 i.g7
4 e4 d6
5 f4 0-0
6 lLJC3 lDa6!? (D)
7 i.e2
This is the most popular move al­
though 7 i.d3 and 7 e 5 are also quite
common; these are examined in
Game 3, whilst below we talce a brief
look at a couple of rarer alternatives:
1) 7 i.e3 l2Jg4 8 i.gl c5 !. This
energetic reaction ensures Black of a
good game. White can now try:
la) 9 d5 f5 ! 10 exf5 (10 e5 dxe5
11 h3 e4) 10 ... i.xf5 11 h3 lDf6 12 g4
(12 i.f2 is suggested by Kiseleva -
probably she's hoping for that in her
next game) 12... i.d7 13 i.g2 \ra5 14
1Vd2 b5 15 l2Jg5 (15 cxb5 i.xb5 16
lDxb5 \rxb5 -+) 15. ..bxc4 16 i.e3
lDb4 17 0-0 l2Jd3 with an excellent
game for Black, Dekusa-Kiseleva,
Ukrainian Ch 1993.
lb) 9 dxc5 lDxc5 10 i.xc5 (10 h3
i.xc3+ I 1 bxc3 lDf6) 10 dxc5 11
.••
\rxd8 i.xc3+ 12 bxc3 :xd8 is
clearly better for Black according to
Serikov and Kiseleva, but White could
have kept her pawn structure intact
by 12 \rd2! . Therefore Black should
probably play 10 i.xc3+ with a
...
good game.
2) 7 c5 dxc5 8 d5 e6 9 i.xa6
bxa6 10 0-0 exd5 1 1 e5 and now:
2a) 1 1 i.b7!? was a typical sac­
...
rifice from GM Kovalev who is cer­
tainly of the opinion that a couple of
pawns and a big centre are worth a
piece. The game Riedel-Kovalev,
Munich 1992/3 continued 12 exf6
i.xf6 13 f5 :es 14 fxg6 hxg6 15
i.f4 c4 16 \ra4 c6 17 �hl \rb6 18
1Vc2 i.c8 19 i.g5 i.f5 201Vd2 i.h8
with chances for both sides.
2b) 1 1 l2Je4 is the sane person's
.••
choice. The game Riedel-Held, also

from Munich 1992/3, continued 12
'ifxd5 'ifxd5 13 4Jxd5 i.b7 ! 14 l2Je3
l%.ad8 with advantage to Black as
White will struggle to get his queen­
side out.
7 •.• e5!
This pawn sacrifice did not work
on move 6 as Black was unable to in­
crease the pressure on e4 in time;
now with .. . 4Ja6-c5 available things
are completely different.
8 dxe5
The alternative capture, 8 fxe5,
has been more popular recently and
that is the subject of the next game.
White has one other possibility, 8
0-0 which has never really caught on
as after the sequence 8 . exd4 9
..
4Jxd4 4Jc5 10 i.f3 l%.e8 1 1 :%.el
Black has the strong possibility of
1 1 i.g4! (D)
•.•
The only time Black failed to play
this move was in Vincent-Gallagher,
Lyon 1993 where 1 l ... a5 was my
choice� this was not because I didn't
see 1 l. . . i.g4 but because I mistak­
enly t hought it would lead to exces­
sive simplifications. Now, however,
White has to fight for equality. For
example:
The Four Pawns Attack 11
1) 12 i.xg4? ! ltJxg4 13 4Jb3?
(White must play 13 'ifxg4 although
Black is better after l 3... i.xd4+)
13 ... 4Jxb3 14 axb3 'ifh4 15 h3 i.d4+
and Black wins.
2) 12 i.e3 and now there are two
lines:
2a) Black can even consider play­
ing 12 4Jcxe4?! 13 i.xg4 ltJxg4 14
.•.
'ifxg4 4Jxc3 15 bxc3 c5 16 4Jc2 ! (16
4Jb5 a6) 16...i.xc3 although I'm not
advocating this line of play as if
White ever manages to get organised
there may be a heavy price to pay for
the weak dark squares and the hole
on d5.
2b) 12... i.xf3 13 gxf3 4Jh5 14
�h l, Schon-Fleck, Porz 1988, and
now 14... 'ifh4 with an edge for Black
according to Knaak.
3) 12 e5 i.xf3 (the immediate
12.. .dxe5 is certainly worth consid­
ering) 13 4Jxf3 dxe5 14 4Jxe5 c6 15
'ifxd8 l%.axd8 16 i.e3 4Jfe4 17 4Jxe4
4Jxe4 18 i.xa7 f5! 19 i.e3 i.xe5 20
fxe5 l%.xe5 and Black was at least
equal in McNally-Bennett, Scottish
Ch 1994.
4) 1 2 4Jb3 i.xf3 13 'ifxf3 4Jxb3
14 axb3 c6 is+ according to Knaak.
This may seem a surprising assess­
ment but Black's position has been
eased by the exchange of a pair of
minor pieces each and White also
possesses the most serious weakness
in the position - the e4-pawn. A
good plan would be to double rooks
on the e-file although I must admit
that I don't consider Black to be bet­
ter after 15 i.e3 ( )
= .
5) 1 2 4Jc2 i.xf3 13 'ifxf3 c6 is
also pleasant for Black.
 12 The Four Pawns Attack
8 ... dxe5
9 1fxd8
White can also capture on e5 at
once: 9 lbxeS (9 fxe5 1fxdl 10
.txd 1 lbg4 11 .tf4 lbb4 + Belov)
9 lbc5 10 .tf3 (101fc2? lbfxe4 11
•..
lbxe4 .tf5 12 .tf3 .txe5 13 fxe5
.txe4 14.txe41fh4+ 15 g31fxe4+
161f xe4 lbxe4 with a clearly better
ending for Black) 10 ...1fxdl+ 11
'it>xd1 lld8+ 12 'it>c2 (12 'it>e2 .te6
13 lbd5 lbfd7! 14 .te3 lbxe5 15
fxe5 lbd7 16 .tg5 .txd5! 17 cxd5
lle8 was slightly better for Black in
Gorelov-Belov, USSR 1987) and
now (D):
1) 12...lbfxe4 13 lbxe4 .tf5 14
llel .txe5 15 fxe5 lld4 when there
is:
la) 16 'it>c3 lld3+ 17 'it>b4 (17
'it>c2 lld4 ) 17... lba6+ 18 'it>a5 b6+
= lbb4+ 17 'it>b2 lbd3+ 18 'it>c2 lbb4+
19 'it>xa6 .tc8+ 20 'it>b5 .td7+ 21
'it>a6 is a pretty draw given by Belov
which later occurred in Ca.Hansen­
Berg, Arhus 1991.
1b) 16 b3 is a risky winning at­
tempt. After 16 lbxe4 (16....txe4+
.•.
17 .txe4 llxe4 18 .ta3 llxel 19
llxel lbe6 20 lldl looks like an edge
for White) 17 'it>b2 lbc5 18 'it>a3
lbd3 19 lldl Black has a choice be­
tween two lines:
l bl ) The game Namgilov-Sepp,
Rostov 1993 continued 19 c6 20 g4
•.•
.te6 21 .tg5 when 21 lbxe5 22•••
llxd4 lbxf3 23 lld8+ llxd8 24.txd8
.txg4 25 .tf6 is an inadvisable ex­
change sacrifice and 21 ...b5 22 cxb5
cxb5 23.txa8? lla4+! 24 bxa4 b4#
is pure fantasy on account of 23
llxd3!. Black should probably play
21...lld7, maintaining the equilib­
rium as Sepp did in the game.
1b2) The fact that there are no
queens shouldn't prevent us from
playing for mate: 19 lle8, intend­
.•.
ing ... lle6-a6 would be more ambi­
tious, e.g. 20 .tb2 (20 .te3 llxe5!)
20... lle6 (20... llxe5? 21 llxd3) 21
.txb7 lld8 (threatening ... llb6) 22
c5 .ie4! (22. ..llb8 23 c6 llxb7 24
llxd3 llxc6 25 lld8+ 'it>g7 26 .td4 !
= ) 23 c6 lld5 with a strong attack for
Black.
2) 12 .te6. If Black wants to
••.
keep the game going he must try this.
Possible continuations are:
2a) 13 llel lbfd7! (Black relies
heavily on ... lbfd7 in this variation)
14 lbxd7 llxd7 and now 15 i.e2
.txc3 should be a little better for
Black but after 15 b3 lbd3 16 lle2
I can't see a convincing way for
Black to continue the struggle.
2b) 13 lbdS lbcxe4 14 lbxc7
i.f5 15 g4 lbxg4 16.txg4.txg4 17
lbxa8 ( 17 lbxg4 llac8 18 lbb5
llxc4+ 19 'it>b3 llc5 looks very nasty
for White) 17 ... i.f5 is a suggestion
by Belov. After 18 llel, 18 llxa8 is
..•
possible, but perhaps 18 ...f6 is the

most promising as both 19 '£k7 fxe5
20 tlJd5 ltJc5+ and 19 tlJxg6 hxg6 20
tlJc7 tDc5+ 21 �c3 l!d3+ 22 �b4
i.f8 are very good for Black.
2c) 13 i.e3 tDcxe4 14 tDxe4 i.f5
15 l!hdl i.xe4+ (15... tDxe4? 16 g4)
16 i.xe4 tDxe4 17 llxd8+ llxd8 18
l!dl must be level, although Black
can unbalance the game by continu­
ing 18 ... llxdl 19 �xdl tlJd6 20 b3
i.xe5!? 21 fxe5 tDc8.
2cl ) If Black can then get his
king to e6 he will be able to claim an
edge. For example, after 22 �c2 �f8
23 �d3 �e7 24 �e4 �e6 the knight
will be able to get out via a7 (if
White prevents tDe7).
2c2) Unfortunately, White can
prevent the king manoeuvre with 22
i.cS! after which 22. . .b6 23 i.a3 c5
24 b4! cxb4 25 i.xb4 f5!? (25... �g7
26 'iii> e2 f5 27 exf6+ �xf6 is also
equal) 26 exf6 (26 e6 �g7 27 i.c3+
�f8 28 i.b4+ �e8 followed by
... tlJe7-c6 is a slightly risky winning
attempt, but 28... �g7 is drawing)
26...�f7 27 �e2 �xf6 28 �d3 �e5
29 i.c3+ �e6 also leads to a draw.
Perhaps you think that I have got a
bit far in my attempts to keep the
game going, but with the word
'Beating' in the title of this book I
don't want to allow our opponents a
forced draw straight out of the open­
ing. Of course it is no easy matter
playing for a win with Black, espe­
cially if White is content to draw, but
if you opt for variation '2' you will at
least give your opponent a chance to
go wrong.
9••. llxd8
10 tlJxe5
The Four Pawns A ttack 13
10 fxe5 tlJg4 11 i.f4 lle8 12 l!d1
has occurred a couple of times when
Black has responded rather weakly
with 12 tDxeS and 12 �f8. Best is
••• •••
12 tlJcS intending 13... tlJe6 which
••. ,
is also the reply to 13 h3. The point is
that if the bishop drops back to g3
then the g4-knight will hop into e3
and if it moves along the c l -h6 di­
agonal then Black will simply recap­
ture on e5.
10 •.• tlJc5
1 1 i.f3
1 1 tlJd5 was tried in Chibur­
danidze-Xie Jun, Manila worn Wch
(12) 1991. After l l ... c6 12 tDe7+
�f8 13 tlJxc8 l!axc8 14 i.e3 tDfxe4
15 0-0 f6 16 tlJf 3 f5 the game was
level.
11 ••• i.e6 (D)
The immediate 11 ttJfd7 has also
•••
been seen. After 12 tlJxd7 i.xc3+
(perhaps l 2 ...i.xd7 is better, as 13
e5 f6 gives Black good play) 13 bxc3
i.xd7 14 i.e3 tlJd3+ 15 'iii>fl i.e6 16
c5 i.c4 1 7 �g1 V.I vanov considers
that Black has compensation for his
pawn.
12 ttJd5
Alternatively:
 14 The Four Pawns A ttack
1) 1 2 i.e3 (awarded an l!' by
V.Ivanov in Jnformator but he only
took variation' la' into account) and
now:
la) 12 ...ltJd3+ 13 ltJxd3 l:.xd3 14
'it'f2 i.xc4 is given as by Belov,
=
who points out that 15 i.e2? l:.xc3!
16 bxc3 ltJxe4+ 17 'it'fl i.xe2+ 18
'it'xe2 ltJxc3+ is good for Black, but
fails to spot that 15 e5! wins mate­
rial. V.lvanov gives 15... ltJd5 16
ltJxd5 i.xd5 17 i.e2l:.xe3 18 'it'xe3
i.xg2 19 l:.hgl +.
lb) 12... ltJfd7! is a much better
move. Sutter-Gallagher, Suhr 1992
continued 13 0-0 (13 ltJxd7 4Jd3+!
is good for Black) 13 ... ltJxe5 14 fxe5
ltJd3 (14 . ..ltJd7!? deserves attention
as after 15 ltJd5 i.xe5 16 i.g5 Black
now has 16... f6) 15 ltJd5 i.xe5 16
i.g5l:.f8 (16 ...i.xd5 17 i.xd8 i.xc4
18 i.h4 ltJxb2 is winning for Black,
but White can play 17 cxc15 f6 18
i.e2 !) 17 i.f6 (although I had al­
ways intended to capture on f6 I re­
member getting a nasty shock when I
realized that my opponent was
threatening to deliver mate in one)
17... i.xf6 18 ltJxf6+ 'it'g7 19 ltJd5
l:.ac8 20 b3 c6 with a favourable
endgame for Black.
2) 1 2 0-0! (perhaps this is the
only way for White to obtain an
equal position) 12 ... ltJfd7 13 ltJxd7
i.d4+! (13. ..l:.xd7 14 i.e3 is prob­
ably favourable for White) 14 'it'h1
l:.xd7 15 lL\d5 c6 and now 16 l:.dl ?
i.g7! 17 l:.bl (or 17 i.e3 ltJa4!)
17...l:.e8 18 b3 ltJxe4! 19 i.xe4 cxd5
20 cxd5 i.g4 21 i.f3 i.f5 won mate­
rial for Black in A.Geller-Belov,
USSR 1988 but 16 i.e3! cxd5 17
i.xd4 dxe4 18 i.xc5 exf3 19 l:.xf3
i.xc4 is completely equal.
12 ..• ltJfd7! ( D)
A player coming across this posi­
tion for the first time could be for­
gi ven for thinking that Black is in
serious trouble. A quick pawn-count
will reveal a slight deficit for Black,
whilst the most eye-catching fea­
tures of the position are the seem­
ingly dominant white knights on e5
and d5. But, to borrow a cliche from
football, chess can be a funny old
game. Black's last move guarantees
the removal of the knight on e5,
whilst the other one on d5, unless it
helps itself to a c-pawn laced with
poison, will soon be expelled by
... c6. From then on, with their diago­
nals cleared, Black's bishops will be­
gin to show what they are capable of
and White may also begin to regret
having moved all his central pawns
leaving so many weak squares be­
hind them.
13 ltJxd7
As we have already commented,
13 ltJ xc7? is not really an option. Af­
ter 13 ...ltJxe5 14 fxe5 Black has sev­
eral tempting continuations, e.g.
 The Four Pawns A ttack 15

14... tt:Jd3+ (the other possibilities 2) 16...fS is more prudent. Cher­

are 14....txc4 and 14....txe5) 15 niakov-Belov, Podolsk 1990 con­
�fl (15 �e2 .txc4 16.tg5 lDxe5+ tinued 17 exf5 gxf5 18 lldl llad8 19
17 �f2 f6 18 tt:Jxa8 fxg5 is similar) lDc2 .tf6 20 .te3 llJd3 21 .txa7
15....txc4 16 .tg5 tt:Jxe5+ 17 .te2 lDxb2 22 llxd7 llxd7 23 lle1 .txc4
f6 18 tt:Jxa8 fxg5 19 llJc7 lld2 20 24 lDe3 llJd3! with some advantage
.txc4+ tt:Jxc4 21 lDd5 lDxb2 22 llcl for Black, but perhaps 18 g4 would
.td4 with an enormous attack for be more critical, after which Burgess
Black. recommends 18 ... lDd3 19 gxf5.tf7
13 ... llxd7 with plenty of activity for Black.
14 0-0 16 b3
Both 14 �e2?! lle8 15 e5 c6 16 Of course 16 eS is answered by
.te3 llJa4 17 b3 cxd5 18 bxa4 d4 19 16... llxf4 and 16 lldl llxdl+ 17
.td2 f6, Vaiser-Weindl, Chiasso 1989 lDxd 1 .txc4 is also no solution to
and 14 eS?! c6 15 llJb4 a5 16 .te3 White's problems as e5 can always
.tf8 17 lDxc6 llJd3+ 18 �f l tt:Jxb2 be met by .. .f6.
19 lDd4.txc4+, Stokstat-Berg, Cop­ 16 ••.tt:Jxe4
enhagen 1991 led to excellent posi­ Watch this piece carefully as it
tions for Black. performs impressive pirouettes in
14 •.• c6! the heart of White's position. Each
14 .td4+ transposes to '2' in the
.•. time White thinks he has consoli­
note to White's 12th move. dated, the knight does another little
15 lDe3 lld4! number just to emphasise who's in
I recall feeling q uite pleased with control.
myself after finding this move, which 17 .tb2 tt:Jd2! (D)
wins back the pawn without relin­
quishing any of the positional pres­ i.-
,r@f:, � -· �. w�
sure. Afterwards I discovered that
15....td4 was the recommended move
w�
-�� � ��-�
d ,, ,
and although this is quite good I still -�-.t-�-
- �� � ;ff$@
prefer my choice. So that you ca n •• ••
make up your own mind, here are a -�·
� " ·�
�h � / /

couple of examples after 1S .td4..•

-�- ��·
r� � p
16 �hl:
I) 16 tt:Jd3 17 f5 f led to great lSg • ,"- Y � �
a f{tf t�!ff ���
••.

complications in Inkiov-J.Ivanov, ;/,

/
pa � . ,,
�/;f, , �
:;;;

Bulgaria 1992, which finally settled

down into an equal endgame after
17... tt:Jxcl 18 fxe6 fxe6 19 .tg4!
lle8 20 lldl ! .txb2 21 llxd7 .txal
22 c5 .tf6f 23 llJc4 lle7 24 llxe7
.txe7 25 .txe6+ �f8 =.
This key move is the tactical justi­
fication behind 15... lld4.
18 :n
After 18 .txd4 .txd4 19 llfe1
lDxf3+ 20 gxf3.txal 21 llxal White
16 The Four Pawns A ttack

would be in for a long and difficult Black is a pawn up with a much

defence. 21. a5 ! looks like a good
.. better position. The remaining moves
start to the technical exercise. were: 30 h3 llg8 31 cJ;h2 llb2 32
18 .•. lld3 lbg4 lDf5 33 llf3 lbh4 (the final
19 i.xg7 cJ;xg7 knight dance) 34 .f:g3 llg6 35 lld 1
20 .f:e l llad8 h5 36 lld8+ cJ;g7 37 lld7+ cJ;f8 38
With his base camp established lbe5 llxg2+ 39 llxg2 llxg2+ 40cJ;h1
far into enemy territory, Black has a :e2 41 lld4 lbf5 42 lbd7+ cJ;e7 0-1
clear advantage.
21 i.dl lbe4! Game 2
The knight heads for c3 from Lautier - Kasparov
where it can attack the base of the A msterdam 1995
white pawn chain.
22 :m lbc3 1 d4 lb f6
23 i.c2 lld2 2 c4 g6
24 rs 3 lbc3 i.g7
White understands that he won't 4 e4 d6
be able to hold his queenside to­ s f4 0-0
gether (24 lla1 lle2 looks winning 6 lbf3 lba6
for Black and 24 a4 would fatally 7 i.e2 es
weaken b3) so he seeks counterplay 8 fxeS dxeS (D)
on the kingside, only to see it snuffed
out by another neat knight manoeu­
vre.
24 ... gxfS
2S i.xfS llxa2
26 :a lb e2+!
27 cJ;bt lb d4
28 llg3+ cJ;bS
29 i.xe6 fxe6 (D)

9 dS
9 dxeS?! transposes to the line 8
dxe5 dxe5 9 fxe5 ! ? , discussed on
page 12.
9 lbxeS?! is also doubtful but
worth looking at in a little m ore de­
tail. After 9 cS! 10 i.e3 (10 d5
..•

lbxe4 11 lbxe4 i.xe5 is good for

Black) there is:

1) 10 ... cxd4 1 1 i.xd4 and now:
la) Several commentators give
1 1 llJg4 an '! ', quoting the game
.••
Bystriakova-Umanskaya, Stavropol
1989, where Black had good posi­
tional compensation for the pawn af­
ter 12 lDf3 i.xd4 13 \i'xd4 lDb4 14
0-0-0 \i'xd4 15 l!xd4 lDc6 16 l!d2
�g7 17 h3 llJf6 18 l!e1 i.e6.
1 b) Whilst this seems OK for
Black, I don't see any real necessity
to head for an ending a pawn down.
One idea is l l . . lDb4?!, but after 12
.
i.c5 \i'a5 13 i.xf8 i.xf8 we have
probably given up a bit much mate­
rial for our beautiful position.
1c) Perhaps the best line of all is
1 1 ... \i'e7 ! as 12 0-0 l!d8 looks ex­
tremely awkward for White.
2) 10...llJg4!? occurred to me af­
ter seeing variation '1a' above. Al­
though it is playable, it is no more
effective than 10 ...cxd4. I have ex­
amined:
2a) 1 1 i.xg4 i.xg4 (you may wish
to amuse yourself with 11. ..cxd4) 12
lDxg4 (12 \i'xg4 cxd4 13 i.g5 \i'a5)
12...cxd4 13 lDh6+ �h8 14 0-0 f5!
with good play for Black.
2b) 1 l lDxg4 cxd4 ( 11. .. i.xg4 12
d5 !) 12 lDh6+ �h8 ( l 2...i.xh6 13
i.xh6 \i'h4+ would be fine for Black
but White should play 13 i.xd4 !
with the advantage) 13 i.f2 ! i.e6 14
liJd5 \i'a5+ 15 �fl i.xh6 16 i.xd4+
i.g7 17 i.xg7+ �xg7 18 1Wd4+ f6
19 �f2 lDb4 ! with adequate com­
pensation for the pawn.
9 .
.. lLJcS
This was thought suspect until re­
cently, as White will soon be able to
kick the knight with b4. To do so,
The Four Pawns A ttack 17
though, he will have to part with the
important dark-squared bishop. Of
the alternatives, 9 c6 (D) is the most
•••
promising.
White then has :
l ) 10 i.e3?! llJg4 11 11d2 ( 11
i.g l i.h6 leaves Black very active)
1 l. ..cxd5 12 cxd5 f5 13 i.xa6 bxa6
14 0-0 fxe4 15 lDg5 i.f5 16 l!adl
l!b8 17 b3 i.h6 with good play for
Black, Siegmund-Schafer, Nettetal
1991.
2) 10 0-0. Black now usually
transfers his king's knight to the
blockading square d6, but it's not
clear whether he should first give a
check with his queen:
2a) 10 \i'b6+ 11 �h1 cxd5 12
•..
cxd5 tiJe8 13 b4! ? \i'c7? ! (13 ... llJd6
looks more to the point) 14 lDb5
\i'd8 15 d6 i.d7? (15...i.e6) 16 i.g5
f6 17 lDxe5 ! i.xb5 18 .ixb5 \i'xd6
19 \i'b3+ �h8 20 lDxg6+ ! hxg6 21
\i'h3+ �g8 22 i.c4+ llf7 23 l!adl
\i'c6 24 \i'b3 \i'c7 25 e5 ! �f8 26
i.xf7 \i'xf7 27 e6 \i'c7 28 l!d7 \i'e5
29 e7+ and White soon won, Mich­
aelsen-Schafer, B undesliga 1993.
2b) 10 llJeS 11 dxc6 (11 i.e3
•••
cxd5 12 cxd5 llJd6 will perhaps be
18 The Four Pawns A ttack
tested in the future) 11. .. Wb6+ 12
cst>h1 bxc6 13 a3 (after 13 ltJa4 Wb4
14 a3 We7 15 i.g5 i.f6 16 i.e3 ltJg7
17 Wel ltJe6 Black had every reason
to feel satisfied with the outcome of
the opening in Vaiser-Lane, Cappelle
la Grande 1994) l 3...ltJac7 14 ltJa4
1ib8 15 b4 llk6 16 c5 and Knaak con­
cludes that White has an edge. He
is probably right; after 16... ltJ8c7,
White should not play 17 Wd6 ? lld8
18 Wxc6 as 18 ... i.b7 costs him his
queen, but simply 17 i.b2, with the
intention of meeting 17 ltJb5 by 18
•••
i.xb5 cxb5 19 lDc3 ltJd4 20 ltJd5
and 17 lldS by 18 Wb3!.
••.
3) 10 i.g5 h6 1 1 i.b4 (11 .lxf6
Wxf6 12 0-0 is equal according to
Lukin and l l . .. i.xf6 12 Wd2 i.g7
13 0-0 cxd5 14 cxd5 Wb6+ 15 �hl
i.d7 was comfortable for Black in
D.Bischoff-Schafer, Mengen 1990)
1 1 ... Wb6 12 Wd2 ltJc5 13 i.xf6 (13
i.f2? ltJfxe4 ! and 13 Wc2? ltJfxe4
14 ltJxe4 ltJxe4 15 Wxe4 i.f5 are a
couple of lines for White to avoid)
13...i.xf6 and now:
3a) The game Bagaturov-Lukin,
USSR 1989 went on 14 llbl ?! cxd5!
15 cxd5 (15 ltJxd5 ltJxe4! is good for
Black) 15... i.g7 16 b4 ltJa6 17 ltJa4
Wd6 18 i.xa6 Wxa6 19 ltJc5 Wd6
and, with . .. b6 and .. .f5 to come,
Black stood very well.
3b) 14 Wxh6 leads to a complex
struggle after 14 Wxb2 15 llc 1
.•.
i.g7 16 We3 f5 !? (16...tlJd7 is solid)
17 0-0 and now:
3bl ) Knaak gives 17 ltlle4 18
•.•
ltJxe4 fxe4 19 llbl Wc2 (19...Wxa2?
20 llal Wc2 21 llfcl Wb2 22 llcbl
wins the queen as 22...Wc2 loses on
the spot to 23 lbel) 20 llfcl Wa4 21
ltJh4 llf6 22 Wxe4 as ;t. He points
out that 22 %114 23 Wxg6 llxh4 is
•••
refuted by 24 Wes+ i.f8 25 llb3 ! ,
but doesn't consider 2 2 cxd5 23
•••
Wxd5+ i.e6 24 Wxb7 lld8, which
looks like fair compensation for a
pawn to me.
3b2) Withdrawing the queen at
once by 17 Wb6 !?deserves serious
•..
consideration.
10 i.g5
White would like to play 10 Wc2
but this runs into 1o.. ltJfxe4 ! 11
.
ltJxe4 i.f5 12 i.d3 i.xe4 13 i.xe4
f5, when Black regains his piece
with the better game.
10 •.• h6
1 1 i.xf6
1 1 i.h4 g5 12 i.g3 ltJfxe4 13
ltJxe4 ltJxe4 14 i.xe5 g4 ! 15 i.xg7
gxf3 wins for Black since 16 .lxf3
�xg7 17 i.xe4 drops a piece after
17... Wh4+ and 16 i.xf8 Wh4+ 17 g3
f2+ 18 cst>n Wh3 is mate.
11 ••• Wxf6
12 b4 ll)a6 !
By attacking b4 Black gains the
time he requires to blockade the po­
sition with ... c5. The alternatives are
less promising:
1) 12 ltJxe4?! 13 ltJxe4 Wf4 14
.••
ltJfd2 f5 15 g3 ! We3 16 Wb3 Wd4 17
lDc3 e4 18 llcl f4 19 liJcxe4 i.f5 20
Wd3 Wes 21 gxf4 Wxf4 22 :n ! and
Black didn't have enough for his
piece in D.Ilic-Certic, Belgrade 1989.
Later Ilic felt that 21 0-0 would have
been even better, giving 21... fxg3 22
hxg3 llae8 23 i.f3 as +-. rm not so
sure about this assessment as after
23...Wxg3+ 24 ltJxg3 i.xd3 25 llfel

.:txe1 + 26 .:txe1 .ic3 27 .:te3 .id4
Black is not without hope.
2) 12 ttJd7 13 c5 ! a5 14 a3 axb4
•••
15 axb4 .:txal 16 1Wxal 1Wf4, Haus­
ner-Khalifman, Bundesliga 1990/1,
and now Khalifman gives 17 1Wa2!
f5 18 g3 ! 1We3 191Wd2 as clearly bet­
ter for White.
13 a3
After 13 .:tbl c5, 14 a3 transposes
to the game while 14 b5 Ci:Jc7 15 d6
tf:Je6 I 6 Ci:Jd5 may have gained a
tempo on Lautier-Kasparov, but is in
fact much worse for White because
Black's bishop is actually better off
on c8 in this position.
13 •.• c5! (D)
14 .:tbl .id7 ! ?
It would be interesting to know if
this novelty was prepared before­
hand or was simply a piece of over­
the-board improvisation after being
surprised by his opponent's choice
of opening. An earlier game, O.Rod­
riguez-Dorfman, Costa Catalana
1994, went instead 14 1Wb6 !? 15
.••
1Wd2 .:td8 16 h4 h5 17 b5 tf:Jc7 18
1We31Wd6 19 0-0 .:tf8 20 �h1 b6 21
.:tf2 Ci:Je8 with good play for Black.
15 b5 !?
The Four Pawns A ttack 19
The main idea behind Black's last
move was to protect b5 so that he can
play ...1Wd6 in one go. For example,
Gurevich gives 15 0-0 1Wd6 (the line
15... cxb4 16 axb41Wb6+ is also in­
teresting since I 7 c5 is met by
I 7... Ci:Jxc5 ! ) 16 Ci:Jb5 (161Wb3 is met
by 16... fS and 16 b5 !De? leaves
Black free to concentrate on a king­
side attack) 16...1Wb6! 17 �hl cxb4
18 axb4 .ixb5 19 cxb5 tf:Jc7 with ad­
vantage to Black. Lautier, not satis­
fied by such variations, selects a
more aggressive continuation.
15 .•• tf:Jc7
16 d6
A double-edged choice, but other­
wise Black will achieve his block­
ade.
16 •.. l2Je6
Black avoids 16••. ttJes on account
of 171Wd5!.
17 ttJd5 1Wd8
18 1Wcl2 Ci:Jd4
18 f5 would be very risky as
.••
White could then whip up a strong
attack by 19 tf:Je7+ �h7 20 h4!, in­
tending to meet 20 .. .fxe4 with 21
Ci:Jg5+ .
19 0-0 .ie6
20 tf:Jxd4 .ixd5 (D)
Black avoided 20 exd4 because
•••
of 21 tf:Je7+ �h7 22 1Wt4 followed
by e5. However Ftacnik proposes
22 ... .:te8 followed by 23... .:txe? as a
reasonable exchange sacrifice.
21 llJc6!
Pretty and forced; 21 cxdS exd4
22 1Wf41We8!, followed by ...1We5, is
good for Black according to Gure­
vich.
21 .•• bxc6
20 Th e Four Pawns Attack
21 ...ii'xd6 drops a piece to 22
l2Ja5 ! and 21 i.xc6 22 bxc6 bxc6
..•
23 l:b7 is horrible for Black.
22 cxd5 cxb5
Obviously not 22 cxd5 23 ii'xd5
.••
when White is dominant on the light
squares as well as retaining his mon­
ster on d6.
23 l:xb5 ii'xd6
24 'Ci'a5 f5 !
Black could have kept his c-pawn
with 24 i.f6 as 25 l:xc5 ? loses to
•.•
25 ...i.dS. However after 25 l:b7!
i.d8 26 ii'b5 i.b6 27 i.c4 he would
have had no winning chances.
25 l:xc5 fxe4
26 l:c6 l:xfl +
27 i.xfl 'ii'f8
28 d6
28 l:xg6 e3 29 ii'e l ii'f 4 ! would
give Black a strong initiative. Gure­
vich continues with 30 ii'g3 �h8 !
31 l:xg7? ii'f2+ ! 32 ii'xt2 exf2+ 33
�xf2 �xg7 winning for Black.
After 28 d6 the players agreed to a
draw. Several commentators justi­
fied this with the variation 28 l:d8
.•.
29 ii'd5+ �h7 30 ii'xe4 l:xd6 31
l:xd6 ii'xd6 32 i.d3 but, as Gure­
=,
vich pointed out, 28 e3! would have
•••
left White with some work to do.
Game 3
Vaiser - Berkovich
Tel Aviv 1992
1 d4 lDf6
2 c4 g6
3 l2Jc3 i.g7
4 e4 d6
5 f4 0-0
6 l2Jf3 l2Ja6
7 i.d3
There is one other seventh move
that has to be examined, namely 7 e5
(D).
This is certainly White's most ag­
gressive choice against 6...l2Ja6 and
in some ways the most logical as
well. After 7 l2Jd7 (7 ... l2Je8 does
••.
not fit in weJJ with the strategy of
undermining the white centre but
7...l2Jh5 !? worked well after 8 i.e2
i.h6 9 f5?! i.xcl 10 ii'xcl i.xf5 11
ii'h6 f6 12 g4 i.xg4 13 l:gl i.xf3 14
i.xf3 li:Jg7 15 i.xb7 ii'b8 ! , Vokac­
Babula, Lazne Bohdanec 1996; 9 g3
must be the right move) White has
quite a few possibilities:
1) 8 h4 dxe5! (8...c5 9 d5 dxe5 10
h5 is playable but dangerous for
Black) 9 dxe5 (9 fxe5 c5 !) 9 ... li:Jdc5

IO i.e3 (10 Wxd8 l:.xd8 11 ti:Jd5
i.g4 gives Black good play accord­
ing to Sokolin) 10...i.f5 11 llxi41Wd7
12 tt:Jxf5 1Wxf5 13 g4 We6 (White al­
ready looks over-extended) 14 1Wf3
:ad8 15 :di f6 16 b4 fxe5 ! 17l:.xd8
:xd8 18 ti:Jd5 (18 f5 is well met by
I8 . .. e4 ! ) 18 ...exf4 19 tt:Jxf41Wf7 ! 20
bxc5 e5 21 i.g2 exf4 22 Wxf4 Wxf4
23 i.xf4 c6 with a clear advantage
for Black, Vaiser-Avrukh, Moscow
Tai mem rpd 1992.
2) 8 We2 c5 9 d5 ti:Jb6 (D) and
now:
2a) 1 0 lbe4 i.g4 11 tt:Jeg5 f6 !
(destruction of the centre is what it's
all about) 12 h3 (Kuzmin considers
both 12 lt:Jxh7? ! �xh7 13 lt:Jg5+
fxg5 14 1Wxg4 ti:Jb4 ! and 12 exf6
exf6 13 tt:Je6 i.xe6 14 dxe6l:.e8 15
g4 f5 to be very good for Black)
12 ... i.xf3 13 ti:Jxf3 e6 ! (continuing
the plan) 14 dxe6 fxe5 15 fxe5 We7 !
16 i.e3 (Kuzmin considers 16 i.g5
to be White's best but I'm not sure
what he has in mind after l 6...Wxe6
since 17 exd6 Wxd6 18 i.e7 allows
the reply 18 . ..1Wg3+) 16... dxe5 17
0-0-0, Glek-Kuzmin, Podolsk 1990,
and now 17 ... e4 18 lt:Jg5 liJa4 is the
The Four Pawns Attack 21
recommendation of Kuzmin, who
doesn't fear 19 l:.d7 as after 19 ...1Wf6
White's queenside is likely to cave
in (20 l:.xb7 l:.ab8 21 tt:Jxe4 Wxe6
doesn't solve White's problems while
20l:.xg7+ looks insufficient as well).
2b) 10 h4 e6 11 h5 exd5 12 hxg6
hxg6 13 cxd5 ti:Jb4 14 a3 lt:J4xd5 15
lt:Je4 dxe5 16 f xe5 i.g4 17 i.h6
i.xh6 18 l:.xh61We7 19 tt:Jeg51Wxg5
20 ti:Jxg5 i.xe2 21 i.xe2l:.fe8 22 e6
f6 (22...�g7 23 l:.h7+ �f6 24 exf7
l:.e7 is good for Black, but 23 exf7 !
leads to a draw after 23 ...l:.xe2+ 24
�xe2 �xh6 25 lt:Je6 ti:Jf4+ ! 26 ti:Jxf4
�g7 = ) 23 l:.xg6+ �f8 24 i.b5 (24
ti:Jh7+ �e7 25 i.g4 l:.g8 ! is good for
Black) 24 . . . fxg5 25 i.xe8 l:.xe8 26
0-0-0 and White eventually managed
to draw this inferior ending, Vaiser­
Hebden, London 1991.
3) 8 c5 (a pawn sacrifice to take
the pressure off e5) 8 dxc5 9 d5 and
..•
now:
3a) 9 tt:Jdb8, aiming for . ..c6,
•••
was the choice in the prototype game
with 8 c5, Semkov-Hebden, Tou­
louse 1989. But this undevelopment
is a little slow and after 10 h4 ! c6 11
h5 ti:Jb4 12 hxg6 hxg6 13 e6! fxe6 14
tt:Je5 exd5 15 tt:Jxg6 i.f5 16 1Wh5 !
White had built up a powerful attack,
though Black did manage to escape
into an extremely messy ending after
16...:f6 171Wh7+ �fl 18 lbe5+ �f8
191Wh8+ ! i.xh8 20 l:.xh8+ �g7 21
l:.xd8 lt:Jc2+ 22 �f2 tt:Jxal 23 g4l:.f8
24l:.xf8 �xf8 25 gxf5 ti:Jc2 which he
managed to hold with difficulty.
3b) 9...ti:Jb6 10 a3 (otherwise
...ti:Jb4 will pick off the d-pawn) with
a choice for Black:
22 The Four Pawns Attack
3bl) 10 e6 (10 .. .f6 has also been
•••
suggested) 11 .ixa6 (11 .ie3 exd5
12 .ixa6 d4) l l. .. bxa6 12 .ie3 exd5
(12 ... lLJxd5 13 l2Jxd51Wxd5 141Wxd5
exd5 15 .ixc5 is unclear) 13 .ixc5
lle8 14 .ixb6 axb6 151Wxd5 .ig4 16
ltJgS? 1Wxd5 17 l2Jxd5 .ixe5 ! and
Black won material in Birens- Van
Laatum, Ostend 1992. Better was 16
O·O, although Black's position is still
preferable.
3b2) 1 0 lLJbS! ? was my selec­
.•.
tion in Videki-Gallagher, Kecskemet
1990, as I didn't want my pawn
structure to be ruined and felt that
playing for ... c6 was the right idea.
The position is similar to that of
Semkov-Hebden, with the important
difference that here the black knight
is more actively placed on b6, as op­
posed to a6 in that game. Play con­
tinued 11 .ie3 c6 12 dxc6 l2Jxc6 13
.ixc5 .ig4 14 .ie2 llc8 (I think
Black already has an edge) 151Wxd8
llfxd8 16 lldl llxdl+ 17 �xdl
.ixf3 18 .ixf3 l2Jxe5 ! 19 .ixb6
l2Jxf3 20 .ixa7 .ixc3 21 bxc3 llxc3 !
22 gxf3 llxa3 and now White gave
back the piece with 23 llel, but the
rook ending proved to be untenable.
I had expected the self-bottling 23
.igl, after which Black should prob­
ably take the two f-pawns and hope
to convert his material advantage.
4) 8 .ie2!? cS 9 exd6 (9 d5 dxe5
10 0-01Wc7 is not worth a pawn) and
now:
4a) White's devious idea is re­
vealed after 9... exd6 10 d5 (D).
Then a position akin to one of the
main lines of the Four Pawns Attack
has been reached. If you compare the
position in the diagram to the one af­
ter 1 d4 l2Jf6 2 c4 g6 3 l2Jc3 .ig7 4 e4
d6 5 f4 0-0 6 l2Jf3 c5 7 d5 e6 8 .ie2
exd5 9 exd5 you will notice that the
sole difference is in the positioning
of the black knights: a6 and d7 in the
diagram as opposed to b8 and f6 in
the 9 exd5 variation. One may argue
that the knight on a6 is better placed
than the one on b8, but the knight on
d7 is certainly worse off than the one
on f6. I doubt whether Black has a
better move than 1O.. .l2Jf6 (1O .. .f5
possibly, but I don't like the look of
it) transposing directly into the
9 ... l2Ja6 line of the exd5 variation, a
line which is supposed to favour
White.
4b) 9 cxd4 10 l2Jxd4 (IO dxe7?
•••
1Wxe7 11 l2Jxd4 l2Jb6 with ...lld8 to
follow is out of the question for
White) 1 0... l2Jb6!? 1 1 0-0 1Wxd6
(l l. .. exd6 12 lLJdb5 .ie6 13 l2Jxd6
1Wc7 is a pawn sacrifice which de­
serves consideration) 12 .ie3 1Wc5!?
(not 12 ...lldS? 13 lLJdb5 ! li'xd l 14
llfxdl llxdl+ 15 llxdl .ie6 16 b3
with a very good ending for White,
while 12 ...1Wb4?! intending to grab a
pawn is likely to end in heartbreak)
13 �41Wc7 141Wb3 (14 c5 l2Jd5 15

lLJb5 Wc6 16 i.d4 lLJe3 ! 17 i.xe3
1Wxe4 + is a line given by Knaak)
14 Jld8 and now:
.•
4bl) Vokac-Kovalev, Osttava 1993
continued 15 l:.adl i.d7 16 c5 lLJd5?
17 WxdS i.c6 18 Wc4 i.xe4 19 llJe6!
with advantage to White. However
Black should have played 1 6 fuc5
•••
after which Knaak gives 17 lLJxc5
1Wxc5 18 lLJb:5 Wf5 19 ltJxa7 lLJc8 20
tt:Jxc8 l:.axc8 21 i.b6 i.e6 22l:.xd8+
:xd8 23 Wa3 as = but this seems to
underestimate Black's chances. Af­
ter 23 ...l:.d2 ! 24 i.f3 l:.xb2 Black is
much better as 25 1Wxe7 loses to
25 ...i.f8 followed by 26...l:.xb6 and
27 . ..i.c5+.
4b2) Knaak proposes instead 15
lLJb5 1Wc6 16 ltJxa7 1Wxe4 17 i.f3
(17 Wxb6 l:.d6 when 18 1Wb3 l:.xa7
19 i.f3 Wxe3+ wins for Black and
18 i.f3 l:.xb6 {18 ...Wxc4 transposes
to '4b223'} 19 i.xe4 l:.e6! 20 tt:Jxc8
l:.xe4 21 lLJb6 l:.d8 looks at least
slightly better for Black) and now:
4b21) He continues 17 Wxe3+
.••
18 Wxe3 i.d4 .19 Wxd4 l:.xd4 20
ltJxc8 l:.xc8 21 i.xb7 ltJb4 (White
also retains winning chances after
21...l:.cxc4 22 i.xa6 l:.c2 23 l:.f2
l:.xf2 24 q;xf2 l:.d2+ 25 �f3 l:.xb2
26 a4) 22 i.xc8 lLJxc8 23 b3 lLJd6 ±.
4b22) However, Knaak doesn't
consider 1 7...Wxc4 (D), which ap­
pears to equalize the game.
For example:
4b221) 18 i.xb6? i.d4+.
4b222) 18 lLJxc8 l:.axc8 191Wxb6
l2Jb4! ? 20l:.ac1(after 20 i.xb7 l:.b8
21l:.acl1Wd3 22l:.fdl {22 i.f2 lLJd5
appears to be favourable for Black}
22.. .Wxd1+ 23l:.xd1l:.xd1+ 24 �f2,
Th e Four Pawns Attack 23
the solid move 24 ...l:.d7 gives Black
good winning chances) 20 ...Wxcl+
21 l:.xcl l:.xcl+ 22 �f2 lLJd3+ 23
q;g3 (23 �e2 l:.d6 ! followed by
...l:.el+) 23. . l:.cc8
. with unclear play.
Black should aim to play ...e5 to ex­
pose the white king.
4b223) 18 Wxb6 l:.d6 with a fur­
ther branch:
4b2231) 19 l:.fcl (19 l:.acl gives
Black the option of 19. ..1Wxa2, which
is probably quite good as 20 1Wb5
i.d7 211Wxb7 l:.b8 looks very sus­
pect for White) 19 ...l:.xb6 (19 ...1Wd3
is met by 20 l:.d I ! and 19 ...We6 20
1Wb3 i.xb2 21 l:.xc8+ l:.xc8 22
tt:Jxc8 i.xal 23 lLJxd6 exd6 24 1Wxe6
fxe6 25 i.xb7 lLJc5 26 i.c6 leaves
Black with some work to do) 20
l:.xc4 i.e6 21 i.xb6 (21 l:.c2 l:.b4!)
21...i.xc4 22 l:.cl (after 22 l:.el,
22...e6 23 b3 i.d5 24 i.xd5 exd5 25
:e7 looks like an edge for White, so
Black should probably play 22 ...i.e6
and if 23 b3 then 23 ...lLJb4 gives ac­
tive play; note that i.xb7 is always
met by ...l:.b8) 22 ...i.xa2 23 i.xb7
l:.b8 24 l:.c8+ l:.xc8 25 lLJxc8 lLJb4
26 tt:Jxe7+ �h8 and as White can't
save his pawn on b2 it's time to agree
a draw.
24 The Four Pawns Attack
la) Vaiser-Gallagher, Suhr 1990
continued 10 0-0 exd5 11 exd5 (11
cxd5 l:.e8 places the White centre
under a lot of pressure, e.g. 12 h3
i.f5! 13 e5 i.xd3 14 ifxd3 dxe5 15
fxe5 llJb4) l 1. . .l:.e8 12 i.d2 ifd7!
(it's important to be able to meet h3
with ... i.f5) 13 a3 llJc7 14 b4?! cxb4
15 ax b4 b5 ! (in order to protect his
strong point on d5 White would like
to play b3 but it's too late for that
now) 16 l:.a5 a6 17 ilb3 i.xf3! 18
l:.xf3 llJg4 19 llJd1 (Vaiser avoided
19 llJe2 on account of 19 ... llJxd5 20
cxd5 ifa7+ 21 'i!thl llJf2+ 22 l:.xf2
if xf2 but White may be able to de­
fend after 23 ifdl) 19 ... bxc4 20
i.xc4 llJb5 21 i.xb5 axb5 22 h3 llJf6
23 llJf2 il b7 +.
1 b) During the post-mortem of
the above game Vaiser thought that
10 dxe6 would have been better,
meeting 10 . .. i.xe6 with 11 f5 and
also preferring White's structure af­
ter 10.. .fxe6. Indeed in ECO he gave
10 dxe6 fxe6 11 0-0 as it, which
rather excited me as I had found a
massive improvement for Black. I
didn't have long to wait until the
pleasant San Bernardino tournament
where I once again found myself fac­
ing Vaiser with the black pieces. The
morning was spent fine tuning my
novelty and lunch was then taken in
a confident state. Afterwards, I de­
cided to take in some alpine air and
as I strolled around admiring the
mountain peaks who should sud­
denly join me but. .. Vaiser. To my
horror he started complaining about
his health and then made me what he
thought to be a very generous draw
offer. When I hesitated he took of­
fence and exclaimed "You don't
think rm going to play the Four
Pawns, do you? It's going to be llJf3
and g3 and you'll never beat me in a
million years". So, I accepted his of­
fer and five years later I still haven't
got to play my novelty and, unfortu­
nately for me, rm so bad at keeping
secrets that I can't stop myself from
sharing it with you. The point is that
after 10 dxe6 i.xe6 Black doesn't
have to worry about 11 f5 (the threat
of .. . llJg4 means that Black is also
well-placed after other moves) as he
has a powerful piece sacrifice in
11 i.xf5!! 12 exf5 l:.e8. Now with
• .•
13 ifd2 or 13 ile2 losing to a combi­
nation of . .. llJg4 and ... i.h6, White
must defend his bishop with 13 �d2
when Black can blow open the centre

with 1 3 d5 ! 14 cxd5 (no choice as
••.
14... d4 and 14 ... dxc4 were threat­
ened) 14 lLJxd5 (D) and now:
.••
lbl) 15 lLJxd5 Wxd5 16 Wb3
l%.ad8 ! 17 Wxd5 (17 .ig5 is met by
17 ... c4) 17 ...l:.xd5 catches White in a
deadly pin (18 �e2 just exchanges
one deadly pin for another deadly
pin and loses to 18....th6). A combi-
nation of ...l:.ed8, ..tt:Jb4 and ...c4 is
. 2) 8 0-0 and now Black has a
going to win Black his piece back
with a winning position.
lb2) 15 .ig51Wa5 (15...tt:Jxc3 16
bxc3 Vd7 is tempting but I can't find
anything against 17 �c2 !) when
White has several defensive tries:
lb21) 16 1Wb3 c4 ! 17 .txc4 (on
17 Vxc4, 17 l:.ac8 is not so clear on
•.•
account of 18 l:.ael !, but 17 tt:Jxc3
•••
18 bxc3 l:.ac8 wins easily) 17 ... lLJc5
18 Wc2 lLJe4+ 19 �cl lLJdxc3 20
bxc3 and Black has a monstrous at­
tack. One way to win is 20... l:.ac8 21
li'b3 Wxc3+ 22 Vxc3 .ixc3, but
there may be even better.
lb22) 16 l:. c l lLJab4 looks very
good for Black, for example 17 Wb3
lLlxd3 18 �xd3 lLJb4+ 19 �d2
:ad8+ ! 20 .txd8 Wxd8+ 21 ttJd5
.ih6+ 22 �d l lLJxd5 wins while 17
The Four Pawns Attack 25
.tc4 is strongly met by 17 ... lLJb6
when Black has similar ideas based
on ... l:.ad8+, e.g. 181Wb3 l:.ad8+ 19
.ixd8 lLJxc4+ 20 Wxc4 Wxd8+ and
wms.
I b23) 1 6 f6 may delay the attack
for a move or two but it doesn't alter
the fact that the white king is stuck
on open files in the middle of the
board. After 16 ... lLJxf6 I can't see
anything resembling a defence for
White: 17 Va4 is met by 17 ...lLJe4+ !
and after 17 � cl J:.ad8 there seems
to be no defence against ...c4 , while
1 7 'ifh3 c4 ! ? (17...l:.ad8) 18 .txc4 (18
Wxc4 l:.ac8 19 Wb3 lLJc5) 18 ... lDe4+
19 �c2 lLJb4+ 20 �bl lLJxg5 21
lLJxg5 1Wf5+ also loses. Perhaps
these variations are not the whole
story but I have faith in the black at­
tack.
choice:
2a) 8 e5 (the most common
•.•
continuation but it's safer to employ
the move-order of variation '2b') 9
fxe5 tt:Jd7 (D) (9...dxe5 10 d5 i) and
now, apart from 10 i.e3 c5 11 d5
transposing to '2b', White has a cou­
ple of interesting possibilities:
26 The Four Pawns Attack
2al) 10 i.e2 i.xf3 (10 ...dxe5 11
d5 looks a little better for White) 11
gxf3 c5 (l l. ..dxe5 12 d5 i) 12 exd6
cxd4 (12 ...i.xd4+ 13 @hl Wf6 14 f4
Wxd6 15 e5 is good for White) 13
lbd5 tt:Jac5 14 �hl b6 15 i.f4 with a
clear advantage to White, Blokh­
V.lvanov, Moscow Ch 1992. He did­
n't take long to finish Black off
either: 15 ... d3 16 i.xd3 i.xb2 17
i.h6 l:e8 18 tlJe7+ @h8 19 f4 tlJf6
20 e5 tlJxd3 21 Wxd3 i.xal 22 i.g5
tlJg4 23 tlJxg6+ fxg6 24 i.xd8 1-0.
2a2) 10 Wet is the latest word,
after which Black has:
2a21) 10 c5 11 i.g5 ! (to make
•..
this move possible is one of the main
points behind Wel) l1. ..Wa5 12 Wh4
i.xf3 13 l:xf3 with very good at­
tacking chances for White. Belov
gives 13 ...cxd4 14 tlJd5 dxe5 15
tlJe7+ �h8 1 6 l:h3 h5 17 Wg4, but
fails to spot 1 6 W xh 7+ ! �xh7 17
l:h3+ i.h6 18 i.xh6 tlJf6 19 i.d2+
+-.
2a22) 10 i.xf3 11 gxf3! dxe5
..•
(l l ...c5 12 d5 tlJxe5 13 i.e2 f5 14 f4
tlJd7 15 exf5 is clearly better for
White according to Belov) 12 d5
lDac5 13 i.c2 a5 14 i.e3 We7 15@hl
with an edge for White, Hubert-Be­
lov, Porz 1995.
2a23) 10 dxe5 11 d5 llld c5 12
..•
i.bl (12 i.e2) 12 ...c6 13 i.e3 cxd5
14 tlJxd5 (14 cxd5 looks more natu­
ral) 14 ... llle6 15 Wg3 i.xf3 16 Wxf3
lll d 4 with a double-edged game,
Sutter-Gallagher, Bern 1995, but as
this was in a quickplay event it is
perhaps best not to take much notice
of it.
2b) 8...llld7 when White can try:
2bl ) 9 d5 c6 10 i.e3 tlJac5 11
i.c2 Wb6 12 l:bl 1i'b4 13 i.b3
tlJxb3 14 Wxb3 a5 with a decent game
for Black, Khan-Panzer, Budapest
1993.
2b2) 9 i.e2 is interesting since
9 e5 10 fxe5 transposes to '2a' and
.. .
9 cS 10 d5 looks quite good for
..•
White. Perhaps Black should play
9 i.xf3!? before committing him­
.. .
self in the centre. 1 0 gxf3 would be
quite strange now while 10 i.xf3 e5
11 dxe5 (11 fxe5 c5! is now quite
good for Black as he will obtain con­
trol over ...e5) l l... dxe5 12 f5 is
probably the critical line. An inter­
esting idea for Black is 12 ... tlJb6, in­
tending to meet 13 b3 or 13 We2
with 13...Wd4+! whilst after 13 Wxd8
l:fxd8 14 b3 c6 the position looks
very comfortable.
2b3) 9 i.e3 e5 1 0 fxe5 cS! 1 1 d5
(11 exd6 obviously loses to 1 l... cxd4
and on 11 dxc5 both 11... dxe5 and
11. .. dxc5 look entirely playable)
l l tt:Jxe5 1 2 i.e2 lllxf3 + (less good
...
are 12 ...i.xf3 13 gxf3! f5 14 f4 tlJf7
15 exf5 gxf5 16 @h1, Petronic-Be­
lov, Pravets 1989 and l2 ...i.d7 13
Wd2 tt:Jc7 14 i.g5 Wes 15 @hl a6
16 a4 b6 17 Wel f6 18 i.d2 tlJf7 19
i.d3, as in Arkhipov-Belov, Mos­
cow 1987, with the better chances
for White in both cases) 13 i.xf3
(now 13 gxf3 i.h3 14 l:f2 f5 is quite
good for Black) 13 i.xf3 1 4 Wxf3
.•.
(D) with a branch:
2b31) 1 4...lll b8? 15 i.f4! and
the knight will never make it to its ·
dream home on e5.
2b32) 1 4 We7 15 i.f4 tfJc7 (the
•••
lines 15...i.e5 16 i. xe5 W xe5 17

1ii'f6 and 15 . ..f5 16 1ii'g3 :adS 17
ltJb5 are both good for White) 16
1ii'g 3! (ensuring that ....te5 doesn't
equalize at once) 16...:adS 17 �h1!
(17 .tg5 .id4+ 1 S �h1 f6 is level)
17 . . . .td4 ! 1 S :ae1 f6 19 l'lJe2 .te5
20 l'lJgl ! a6 (20... b5 could have been
played without preparation) 21 l'lJf3
.txf4 22 1ii'xf4 b5 23 b3 : bS ? (ac­
cording to Fta�nik Black could still
have held the balance by blocking
the queenside with 23 ... b4 and then
carrying out the torturous manoeu­
vre . . . l'lJc7-aS -b6-d7-e5; now White
is able to crash through on the queen­
side) 24 b4! cxb4 25 c5! :bdS 26
llkl4 (perhaps 26 c6 is even stronger)
26 ... dxc5 27 l'lJc6 1ii'd 7 2S l'lJxdS
1ii'x dS 29 :ct ! (forcing the c-pawn
to advance deprives Black of an out­
post on d4 for her knight) 29 .. . c4 30
d6 l'lJe6 31 1ii'd2 1ii'd7? (the last
chance was 3 1 .. . a5) 32 1ii'x b4 l'lJg5
3 3 1ii'c5 :es 34 :eel ! :e6 35 e5 f5
36 a4 ! (White plans to invade on the
a-file) 3 6 ... l'lJf7 37 axb5 axb5 3S
1fic7! 1ii'xc7 39 dxc7 :c6 40 e6 :xc7
(40 ... l'lJd6 loses to 41 :dl ) 41 e7
lLld6 42 eSW+ l'lJxeS 43 :xeS+ �f7
44 :bs c3 45 �gl 1-0 Zsu.Polgar­
Chiburdanidze, St Petersburg 1995.
The Four Pawns Attack 27
2b33) 14 l'lJb4 has been sug­
.•.
gested by Arkhipov and deserves an
outing. Black is now threatening to
get his knight to one of the key cen­
tral squares via d3 or c2 so White must
play 15 1ii'e2 whereafter 15 ... 1ii'e7 16
a3 (perhaps White can improve on
this move but the knight does have
annoyance value on b4) 16 ... l'lJa6 17
.t f4 .te5 (Black should still avoid
17 ... l'lJb8 , this time on account of 18
l'lJb5 :<ls 19 1ii'f 3 ! when White is
threatening to capture on d6) looks
roughly equal. Basically, Black's
knight foray has prevented White
from getting his queen to g3 .
8 fxe5
Just as with the bishop on e2
(Grune 1) accepting the pawn prom­
ises White nothing: 8 dxe5 dxe5 and
now:
1) 9 fxe5 l'lJc5 ! 10 .tc2 (10 exf6
l'lJxd3+ 11 �e2 l'lJxc1+ 12 :xc1
1ii'xf6) 10... Wxdl+ 11 �xdl l'lJg4 12
�e2 l'lJxe5 with a long-term posi­
tional advantage for Black.
2) 9 l'lJxe5 l'lJc5 (9 ... l'lJg4 10
l'lJxg4 .txg4 11 .te2 .txc3+ 12 bxc3
Wxdl+ 13 .txdl .txdl 14 �xdl
:adS+ 15 �c2 :fes is supposed to
give compensation according to
Sokolin but I find this line slightly
less convincing) 10 .te3 (10 .tc2
Wxdl+ 11 �xdl l'lJg4! 12 l'lJxg4
.txg4+ 13 �el .txc3+ 14 bxc3
l'lJxe4 15 .txe4 :feS is a little better
for Black) 10... l'lJxd3+ (10 . ...tg4 11
.te2 .txe2 12 WxdS :fxdS 13
.txc5 ! is good for White) 11 1ii'xd3
1ii'xd3 12 ltJxd3 :es (not l 2 . .. l'lJxe4
13 l'lJxe4 :es 14 l'lJe5 f6 15 l'lJxf6+
.txf6 16 .td4 + ) 13 e5 (much safer
28 The Four Pawns Attack
is 13 tt:Je5 with an equal game after
13 ... tlJg4 1 4 tlJxg4 i.xc3+ 1 5 bxc3
i.xg4 16 e5 f6 as 17 i.d4 can be met
by 17 ...c5) 13 ...tlJg4 14 i.gl i.f5 15
tlJc5 b6 16 tt:Ja6 i.xe5! ? 17 fxe5
I:.xe5+ 18 �d2 l:.d8+ 19 tt:Jd5 c6 is
an extremely sharp variation given
by Knaak, who considers this posi­
tion to be unclear. We can take this
line just a little further: 20 h3 cxd5
(20 ... tt:Jf6 or 20...tlJh6 are met by 21
i.d4 ) 21 hxg4 dxc4+ 22 �c3 i.xg4
and here Black has three pawns, a
more co-ordinated position and a con­
tinuing attack for his piece (of course
23 �xc4 loses to 23 ...i.e2+).
8..• dxe5
9 d5' c6 (D)
9... tt:Jc5 (quite effective with the
bishop on e2) 10 i.c2 a5 is suspect
as Black won't be able to prevent a3
and b4 in the long term .
w iYb3 lLxl6 17 tt:Ja4 i.g7 18 4Jb6 l:.b8
10 0-0
Alternatively 1 0 i.g5 with two
lines:
1) 10 ... ifb6 and now:
la) 1 1 ifd2? ! tlJc5 12 %:.bl cxd5
13 cxd5 tlJg4 ! (Sokolin) and White
can't prevent .. .f5 as 14 h3 is refuted
by 14 .. .4Jf2 L
1b) 1 1 tlJa4! renders the above
line redundant. After 1 1 iYaS + 1 2
.•.
i.d2 ifd8 13 0-0 cxd5 14 cxd5 tt:Je8
1 5 i.xa6 bxa6 16 i.b4 4Jd6 17 %:.cl
White had the better chances in
Garcia Palermo-Danailov, Alicante
1992, although l l .ifb4+ 12 i.d2
..
ife7 must be a slight improvement
for Black.
2) 1 o h6! 11 i.xf6 c11 i.h4 ifb6
...
is now fine for Black as after 12 tlJa4
ifb4+ there is no 13 i.d2) 11. .. i.xf6
12 %:.bl tt:Jc7 (perhaps there are more
dynamic moves available in the posi­
tion, 12 . ..ifb6 for example, but one
shouldn' t grumble about the text as
Black will feel very comfortable
once the knight has arrived on d6) 13
0-0 cxd5 14 cxd5 tt:Je8 15 �hl a6
(the knight doesn' t take up immedi­
ate residence on d6 as this would al­
low White to exchange it off with
tt:Jb5, which would be a rather sad
end to such a lengthy manoeuvre) 16
19 %:.be1 i.g4 with a roughly level
game, Garcia Palermo-Comas, lber­
caja 1992.
10 cxd5
1 1 cxd5 tlJe8
12 ife2
On 12 i.x a 6 Black has the reply
,
12 ... ifb6+.
12 ••. tlJc5
13 i.g 5
It is somewhat surpns1ng that
Vaiser allowed the exchange of his
bishop as Black seems to have com­
fortable equality afterwards. Knaak
suggests 13 i.c2 b6 14 b4 i.a6 1 5 b5
i.b7 which he assesses as ;!;. Perhaps
this is true as White does have a
 The Four Pawns Attack 29

passed pawn and a potentially deci­ cha nce to play the u ndermini ng
sive outpost on c6, but these assets move .. .f5.
will be very difficult to exploit; the 19 i.f2 f5
d-pawn is firmly blockaded and the 20 a4 a6
white knights are currently in no po­ 21 b4 :.ac8!
sition to occupy c6 . Black should Of course Black doesn't play
complete his development with a 21 ...li'xb4 which would allow White
combination of ... :.cs, . ..lLJd6, and to infil trate to the seventh rank after
. . .1!i'd7 (or . ..li'e7) and then, depend­ 22 :.abl , but instead prepares a
ing on the circumstances play . . .f5 or bishop-activati ng exchange sacri­
perhaps double his rooks on the c­ fice. Perhaps White should not ac­
file. I am sure that the majority of cept the offer although this woul d
King's Indian players would happily be inconsistent with the move b2-b4
settle for this position. which has, incidentally, already com­
13 •.• f6 promised White's position o n the
14 i.h4 queenside.
White has developed his bishop in 22 i.c5 :.xc5
this fashion in order to hold up .. .f5. 23 bxc5 li'xc5+
14 lLJxd3 24 �hl 1!i'd4!
15 li'xd3 lLJd6 This is probably what White had
16 lLJd2 underestimated or overlooked.
Planning to exchange the strong 25 :.o
knight on d6 but this is all very time- 25 :.adl :.cs is very uncomfort­
consuming. a ble for White as 26 :.f3 would lose
16 ••. i.d7 to 26 . ..:.xc3 ! .
17 lLJc4 li'c7 25 ••. :.cs
18 lLJxd6 1!i'xd6 (D) 26 :.e 1 i.h6!
With the awkward threat ... i.d2.
27 exf5 (D)
and a draw was agreed, some­
what prematurely on Black's part.

White's passed pawn is compen­

sated for by the bishop pair, the prob­
able first use of the c-file and the
30 The Four Pawns A ttack
After 27 �xfS 28 'fixd4 exd4 29
...
lDd l ( 29 lDe2 l:.e8 ! ) Black's power­
ful pair of bishops and dangerous
d-pawn provide ample compensa­
tion for the exchange. He can also
consider 27 �d2! ?, as I can' t see
••.
anything better for White than 28
'fixd4 exd4 29 lDe4 �xe l 30 lDf6+
�f7 31 llJxd7 (31 fxg6+ hxg6 32
llJxd7+ �e7 is similar) 31. gxf5 32
..
l:.xf5+ �e7 33 lDe5 �d6 with what
should be a won endgame for Black.
2 h3 Systems

A n early h 3 b y White i n the King 's The material is split up as follows:

Indian has two main ideas behind it. Grune 4: White delays, or omits
The first is to pave the way for .le3 altogether, �f3 .
as Black will now be unable to harass Grune 5 : White plays h3, �f3 and
the bishop with . . . �g4. Of c0urse .le3.
White quite often plays .lg5 but he Game 6: White plays h3, �f3 and
stil l needs to be ready to meet . . . h6 .lg5 .
with .le3. The second idea is to sup­ Game 7 : A fashionable line from
port the advance g4, which can be the Petrosian System.
played to gain space on the kingside
or to dissuade Black from playing Game 4
. . . f5 . Although these intentions are Paunovic - Kupreichik
similar to those of White in the Yugoslavia 1992
Samisch the two variations lead to
quite different types of game. In h3 's 1 d4 �f6
favour is that it keeps open the op­ 2 c4 g6
tion of �f3 and doesn' t weaken the 3 �c3 .lg7
dark squares as much as f3 , whilst 4 e4 d6
on the other hand White does noth­ s h3 0-0 (D)
ing to support the all-important e4-
square.
My main recommendations are
centred around Black playing the tra-
ditional . . . e5 in conjunction with the
modern . . . �a6 which has, in fact,
been the main line for many years.
The struggle is often very sharp and
bo th players can attack on either
wing ( t he centre is almost always
blocked). White tends to win games
by taking control over the crucial e4-
squ are, whilst Black's victories usu­ 6 .lgS
ally occur when he achieves the After 6 .le3 �a6 there are three
advance . .. e5-e4 or when White has lines :
neglected the safety of his king (or a 1) 7 �f 3 e5 8 d5 transposing to
combination of both). Game 5 .
32 h3 Systems
2) 7 g4 e5 8 d5 l2Jc5 (8 . . . c6 is an
alternative) 9 f3 a5 10 'ii'd 2 c6 1 1
dxc6 ! ? bxc6 1 2 0-0-0 l2Jb7 1 3 c5 d5
14 exd.5 ( 14 g5 d4 ) 14 ... liJxd.5 15 .lc4
( 1 5 l2Jxd5 cxd5 1 6 'ii' xd5 'ii' xd5 1 7
l:txd5 .le6) 1 5 . . . l2Jxe3 1 6 'ii' xe3 'ii'e7
with a roughly level game, Gomez­
Topalov, Seville 1992.
3) 7 .l d3 (more common) 7 eS ...
8 dS and now we shall examine a
couple of possibilities for Black:
3a) 8 liJhS 9 g3 'ii'e8 (or 9 . . . l2Jc5
...
1 0 .le2 l2Jf6 1 1 'ii'c 2 a5 1 2 0-0-0 a4
1 3 g4 , Knaak-Piket, Hamburg 1 99 1 ,
and now instead of the immediate
1 3 . . . l2Je8, Piket suggests 1 3 . . . .ld7,
delaying . . . l2Je8 until l2Jf3 has been
played as it will then be more diffi­
cult for White to advance on the
kingside) 10 .le2 l2Jf6 1 1 l2Jf3 ltJd7 !
(more logical than 1 1 . . . .ld7 which
has been played a few times ; . . . l2Jh5-
f6-d7 may look like a waste of time
but White has only gained .ld3-e2
and g2-g3, which don ' t really help
him) 1 2 g4 f5 1 3 a3 l2Jac5 1 4 gxf5
gxf5 1 5 l2Jd2 l2Jf6 1 6 .lxc5 dxc5 1 7
'ii'c 2 l2Jxe4 1 8 l2Jdxe4 fxe4 19 l2Jxe4
.lf5 20 .ld3 'ii' h 5 2 1 'ii'e 2 'ii' h4 with
an active game for Black, I.Sokolov­
Van Wely, Groningen 1994. It will be
difficult for White to maintain both
his blockade on e4 and material
equality.
3b) 8 l2Jd7 and now ( D) :
...
3bl ) 9 l2Jge2 l2Jdc5 1 0 .lc2 f5 1 1
exf5 ( 1 1 f3 .lh6) 1 1 . . . gxf5 1 2 0-0 f4
1 3 .l xc5 l2Jxc5 1 4 f3 is unclear ac­
cording to Kuzmin.
3b2) 9 g4 l2Jdc5 10 .lbl ? ! ( 1 0
.lc2) 1 0 . . . f5 1 1 exf5 gxf5 1 2 l2Jge2
'ii' h4 1 3 a3 e4 1 4 gxf5 .l xf5 1 5 l2Jd4
.lg6 1 6 .lc2 l2Jd3+ 1 7 .lxd3 .lxd4 !
was excellent for Black in Avshalu­
mov-Kupreichik, Blagoveshchensk
1 988.
3b3) 9 a3 l2Jac5 10 .lc2 f5 1 1 b4
( 1 1 f3 .lh6) 1 1 . . . l2Jxe4 1 2 .lxe4
fxe4 1 3 l2Jxe4 'iih4 1 4 g4 !, Barlov­
Kir. Georgie v, is given as better for
White by several commentators, but
the simple 13 l2Jf6! promises Black
••.
a good game. White is far too under­
developed to maintain his grip on e4.
14 .lg5 should be met by 14 . . . .lf5 .
6
••• l2Ja6
..•6 cS is quite popular, but after 7
d5 e6 8 .ld3 exd5 9 cxd5 ! we are in
the Modern Benoni, which is outside
the scope of this book.
7 .ld3
There are a couple of alternatives,
line 'l' being the most important:
1 ) 7 l2Jf3 'ii'e8! (Black plans
8 . . .e5 but with this tricky move order
he can meet 9 d5 with an immediate
. . . l2Jh5 ; it is also possible to play
7 . . . e5 as 8 d5 transposes to game 6
and 8 dxe5 dxe5 9 'ii xd8 l:txd8 10
l2Jd5 l:td6 is fine for Black) and now
(D):
l a) 8 .l d3 e5 9 d5 (this falls in
with Black's plans, but 9 0-0 didn' t

look very impressive in Shabalov­
Edelman, New York 1 993: 9 . . . li)h5
1 0 :e l exd4 l l lbd5 f6 1 2 .lc l c6
1 3 li)f4 li)xf4 1 4 .lxf4 'ii'd S 1 5
li)xd4 f5 with advantage to Black)
9 . . . li)h5 1 0 g3 f5 1 1 .ld2 li)c5 1 2
.lc2 fxe4 1 3 li)xe4 li)xe4 1 4 .lxe4
li)f6 1 5 .lc2 e4 with a decisive ad­
vantage for Black, Bronstein-Nij­
boer, Wijk aan Zee 1 992. Bronstein
can't have been concentrating that
day.
l b) 8 li)d2 is also met by S . . . e5 9
d5 li)h5 .
le) 8 eS is perhaps more critical
but so far experience is limited to
one quick draw: S . . . dxe5 9 dxe5 lbd7
10 'ii'e 2 f6 1 1 exf6 exf6 1 2 .le3 112_ 11z
Bykhovsky-Istratescu, Tel Aviv 1994.
1 d) 8 g4 can be met by 8 eS 9 d5
. ..
li)d 7 transposing to the main game,
or perhaps more logically by Bur­
gess's suggestion 8 cS 9 dS e6 10
•••
dxe6 when Flear-Wood, London
1 993 continued 10 'ii'xe6 1 1 'ii'e 2
.••
:es 1 2 0-0-0 li)xe4 1 3 li)xe4 'ii'xe4
1 4 'ii' xe4 :xe4 1 5 :xd6 :es with a
c omfortable game for Black, while
B urge ss recommends 10 .lxe6 1 1
...
'ii'xd6 .lx g4 1 2 hxg4 li)xe4 l 3 li)xe4
'ii'xe4 + 1 4 .le2 :res but does n ' t
h3 Systems 33
take 1 5 .le3 into account [yes in­
deed, this line was somewhat over­
op timistic! - editor 's note]. I would
prefer 1 1 h6 1 2 .le3 :ds 1 3 'ii'g 3
. ..
li)b4 1 4 :c 1 b6 with plenty of com­
pensation for the pawn.
l e) 8 'ii'c 2 c6 ! ? 9 :d l e5 1 0 d5
cxd5 1 1 cxd5 .ld7 1 2 li)d2 :cs 1 3
.lxa6 bxa6 1 4 'ii' b 3 .lb5 1 5 'ii'a3
li)h5 ! (Miles clearly enjoys himself
when he plays the King 's Indian) 1 6
g 3 f6 1 7 .le3 f5 l S .lxa7 :n 19
.lb6 .ld3 20 .la5 ? ! (maybe White
could have tried 20 'ii'x d6) 20 . . . 'ii'e7
2 1 h4 'ii' a7 (threatening . . . li)xg3) 22
:h2 li)f6 23 'ii' x d6 li)g4 24 'ii'e6
:xc 3 0- 1 Comas-Miles, Benasque
1 995 .
2) 7 g4 seems to be played only
against me. 7 cS is obviously an op­
...
tion but both my games have gone
7 eS 8 dS 'ii'e8 (8 ... c6 might be bet­
.. .
ter) 9 li)ge2 and now:
2a) The game Suba-Gallagher,
Kuala Lumpur 1 992 went 9 'iti>h8 . ..
(the idea is to be ready to play
. . . li)gS and . . .f5 at a moment's no­
tice) 1 0 'ii'd 2 c6 1 1 li)g3 cxd5 1 2
cxd5 .ld7 1 3 .le2 li)c5 1 4 f3 li)gS !
1 5 h4 f5 1 6 gxf5 gxf5 1 7 exf5 .lxf5
l S li)ce4 ( I S li)xf5 ) 1 S . . . .lxe4 19
li)xe4 li)xe4 20 fxe4 'ii' a4 ! 21 .ld3
li)f6 22 'ii'e2 li)d7 with an unc lear
position. White has the bishop pair
but his king is exposed.
2b) In Cramling-Gallagher, Biel
1 994 I opted for 9 c6, not because
.••
it was a prepared improvement but
because I'd completely forgotten the
Suba game (a large percentage of
novelti es are born this way). Play
continued 1 0 li)g3 cxd5 1 1 cxd5
34 h3 Systems
.ld7 1 2 .ld3 lbc5 ( 1 2 . . . b5 1 3 a4
lbc7 1 4 axb5 lbxb5 15 .lxb5 .lxb5
1 6 'iif3 !, winning a piece, is a nasty
trap and typical for this line) 1 3 .lc 2
a5 1 4 a4 'ti'd8 ! (it was around about
here that I felt that 8 . . . c6 may have
been better) 1 5 l:lbl l:lc8 1 6 <ifi>fl
lba6! ? l 7 .ld3 l2Jc5 1 8 .lb5 (my op­
ponent was of the opinion that I
had made a serious strategic error in
permitting the exchange of light­
squared bishops, whilst I believed
the exchange to be in my favour as
it brought me some much-needed
space; I must confess that the trans­
parent trap, 1 8 <ili>g2 lbxd3 1 9 1Wxd3
lbxd5 !, helped to tip the scales in fa­
vour of 1 6 . . . lba6) 1 8 . . ..lxb5+ 19
axb5 'iid 7 with a roughly equal posi­
tion.
2c) I should also point out that I
refrained from the natural 9 lbd7 ...
on account of 10 l2Jg3 lbdc5 1 1 a3 !
f5 1 2 b4 l2Jxe4 1 3 lbcxe4 fxe4 1 4
l2Jxe4 when White is much better.
7 •.• e5
8 d5 'iie8
Black unpins, the first step to­
wards playing . . .f5. The alternative is
8 c6 ( D) with the foil owing possi­
... , White an edge) 1 4 . ..l:lfc8 1 5 <ifi>h l
bilities:
1) 9 'ti'd2 (probably not very ac­
curate as White neither wants to cas­
tle long nor play .lh6) 9 . . . cxd5 1 0
cxd5 lbc5 l l .lc2 a5 l 2 l2Jge2 .ld7
1 3 g4? ! ( 1 3 a4 would have main­
tained equality) 1 3 ... b5 14 l2Jg3 b4
1 5 llJd 1 'ii b6 1 6 i.h6 .lxh6 1 7
'iix h6 <ifi>h8 ! (intending to drive the
queen away with . . . l2Jg8 ; a common
ploy in the King's Indian) 1 8 l2Jh5 ?
(this just opens more lines in Black's
favour) 1 8 . . . l2Jxh5 1 9 gxh5 'iid 8 ! 20
hxg6 fxg6 2 1 l2Je3 'iif6 22 0-0-0 a4
0- 1 S teinbacher-Brunner. Of course
White is losing but it was a little
unsporting to resign before any
blood was drawn.
2) 9 lbge2 lbc5 1 0 .lc2 'ii b 6
( 1 0 .. . a5 is also possible) 1 1 0-0 cxd5
1 2 cxd5 ( 1 2 .lxf6 .lxf6 1 3 l2Jxd5
'iid 8 1 4 b4 lbe6 1 5 l2Jxf6+ 'ii xf6 1 6
'ti'd2 b6 ! is nothing for Black to
worry about according to Kasparov)
and now:
2a) 12 .ld7 1 3 l:lbl a5 1 4 'iid 2
...
( 14 .le3 transposes to Petursson­
Zliger, Horgen 1 994 when 14 . .. l:lfc8
1 5 f3 'ti'd 8 1 6 a4 'iie 8 1 7 g4 b5 was
about equal, but 1 6 a3 may give
'iid 8 1 6 a3 a4 17 l:lbe l l:la6 1 8 lbc l
'iia5 1 9 Ji.bl l2Jh5 20 l2Jd3 l2Jb3 2 1
'iie 3 f6 2 2 .lh6 g5 2 3 .lxg7 <iftxg7
with equality, Ivanchuk-Kasparov,
Horgen 1 994.
2b) Kasparov believes that White
has no more than a draw after
12 'ti'xb2 He gives 1 3 l:lbl 1Wa3 1 4
••• .
lbb5 1Wxa2 1 5 l:lal ( 1 5 lbxd6 1Wa5
and 1 5 .lxf6 .lxf6 1 6 Jla l 'ii'c 4
{ 1 6... 'ti'b2 l 7 it)ec3 ! } 17 lbxd6 'ii'b4
are both OK for Black) 1 5 . . .'ii b2 1 6
 h3 Systems 35

llbl ( 1 6 tDec3 tiJfxe4 ! ) 1 6 . . . 1Wa2 1 7 here against Cvitan. After 1 0 ...tiJdc5

llal =. 1 1 .lc2 f5 1 2 b4 tDxe4 1 3 tlJxe4 fxe4
9 g4 14 .lxe4 Black should have brought
A ltern atively : his knight back into play at once
1 ) 9 tt)ge2 tiJd7 10 0-0 f5 1 1 exf5 with 1 4 . . . tiJb8.
gxf5 1 2 f4 e4 1 3 .lc2 1i'h5 ! (the f4 After 10 tlJge2, Kupreichik pro­
blockade is very common in the posed 10 fS 1 1 exf5 e4 1 2 tDxe4
••.

King's Indian, but here the bishop on gxf5 1 3 gxf5 tlJe5 but, as Howell
g5 is misplaced) 14 tiJd4 'ti'xd 1 1 5 pointed out, this is a load of garbage;
llaxd l tiJb4 1 6 .l b l tlJc5 with a 14 f6! is very good for White so long
good game for Black, Cramling­ as after 1 4 . . . 1i'h5 he plays 1 5 tiJd4 !
Gallagher, Bern 1 992 (although 1-0 and not 1 5 fxg7??, which allows
after a mammoth 1 24 moves, 47 of mate in three. Instead of 1 0 . . . f5 I
which were spent mating with suggest 10 tDdcS 1 1 .lc2 f5.
•••

bishop and knight against king). 10 fS

2) 9 a3 tlJc5 10 .lc2 a5 l 1 'ti'f3? ! 1 1 gxfS gxfS
tijfd7 1 2 tiJb5 tlJa6 1 3 g4 tiJdc5 1 4 12 llgl
tlJe2 ( 1 4 b4 axb4 1 5 axb4 tiJxb4 ! is 12 exfS is very strongly met by
good for Black) 14 . . . f5 15 1i'g2?! 1 2... e4 ! .
fxe4 1 6 .lxe4 tlJxe4 1 7 1i'xe4 .ld7 ! 12 'lii>h 8
1 8 a4 tlJc5 1 9 1i'e3 .lxb5 20 axb5 13 lbh4 liJdcS
e4 ! and the rest was carnage, Cher­ 14 .lc2 fxe4
nin-Gallagher, Basie rpd 1995 . What 15 tlJxe4 l2Jxe4
I like about these h3 sy stems is that 16 .lxe4 tlJcS
when things go wrong for White 171i'c2
they go really wrong. The bishop must be protected as
9 ••• tiJd7 (D) retreating it would allow 17 . . . e4 ! .
17 ••• tlJxe4
w J m.tB'W••B As Kupreichik points out 17 ... :.J4
is met by 1 8 .lxh7 and 17...'ti'hS 1 8
. , ._. , . , .lxh7 ! 1Wxh7 19 tiJg6+ 'lii>g 8 20
�··� �. � -
·•
� .& •
­ tlJe7+ leads to a draw.
• - � �,] • 18 1i'xe4 (D)
-�-�-�- 18 1i'hS! ?
� �
.••

�� M A
y

18 'ti'f7 , intending 1 9 ... .lf.S, also

� JiL. � � 0
A� � ·@,:�
� ��y� ;ff�
• ••

�
�
'{Pbf.;;-;,
looks quite good.
0,, � ;ff ;@:,
19 llcl
:/.
� ,, ,, @, i?,., �, ?'.//:10

D
"',: .' :� ··�/:Wf,f:/,.;:-....;�?0;::. 'if·=
� '&;
0 ,/
� f; ' ,,
� � . �, ·� :,
«
,
�
With the intention of swin g ing his
q ueen's rook to the kingside.
1 0 tiJf3 19 ... .ld7
In th e same Basie quickplay men­ An attacking player like Kuprei­
.
tt on ed above, Chernin played 10 a3 chik would be considering . . . llf4 at
36 h3 Systems

:xe7 :xf2 and 24 fxe3 :xe7 25

:xg7 1i'xh4+ 26 �d l 1i'h5 + 27 �c l
1i'e8 ! 28 1W g5 :ef7 ! Black would
.

have to work a bit harder after 24

each turn - here it was rejected be­
cause of 20 .lxf4 1Wxh4 2 1 1i'g2 -
but it is clear that it, and other attack­
ing ideas, will be more effective once
the queenside is developed.
20 :c3 :aeS
21 :.cg3 b5 !
21 ... lH4 22 1i'g2 ! is less good as
Black can 't play . . . e4 .
22 1i'g2 e4!
23 .le7 (D)
After something like 23 .le3
.lxb2 (or maybe 23 . . . .lf6) White has
no killer blow.
:xe3 but Kupreichik points out that
24 . . . .lxb2 ! should be decisive. It's
not so much the pawn that is impor­
tant but the possibility of playing
. . . .lc3+, e.g. 25 .lxfS .lc3+ !, 25
1We4 :n ! and 25 1i'g3 :gs ! 26 1i'h2
.lc3+ L
24 .•. :xe7
And White resigned as after 25
:xg7 1Wxh4+ there is no good
square for his king.
If a game is lost so quickly with­
out any clear error then it must mean
that the whole strategy is wrong. I
have serious doubts about White's
combination of g4 and lbf3 in this
game.
Game 5
Chemin - J.Polgar
New Delhi 1990

1 d4 lbf6
2 c4 g6
3 lbc3 .lg7
4 e4 d6
5 h3 0-0
6 lbf3
Of course 5 lbf3 0-0 6 h3 is the
same.
6 ••. e5 (D)
It is equally possible for Black to
play 6 lba6 7 .le3 e5, which has
•••

23 e3! the advantage of avoiding the boring

24 f3 ending considered in the next note
Comp letely hopeless, but so are and gives him the op tion of meeting
24 .lxfS exf2+ 25 �f l 1i'e2 mate, 7 .lg5 with 7 . . . 1We8 .
24 :xg7 exf2+ 25 1i'xf2 :xe7+ ! 26 7 d5

7 dxe5 dxe5 8 'fixd8 D.xd8 is no
more dangerous for Black here than
in the Classic al Variation (see Chap­
ter 5). In fact the extra h 3 White has
gained seems to make no difference
at all to the line recommended there.
For exam ple, Cvetkovic-Zontakh,
Arandjelovac 1 993 continued 9 .i.g5
D.e8 1 0 �d5 �xd5 1 1 cxd5 c6 1 2
.i.c4 cxd5 1 3 .i.xd5 �d7 1 4 liJd2
�5 1 5 �c4 .i.f8 1 6 0-0 .i.e6 1 7
.i.xe6 �xe6? 1 8 .i.f6 with an edge
for White, but if the reader turns to
page 89 he wil l find that 17 . . .D.xe6 !
gives Black a good game in the posi­
tion with the pawn on h2, so it surely
does here as well.
7 ••• �a6
7 �h5 has received a bad press
.• .
over the years, the general opinion
being that Black should wait for .i.e3
before p laying . . . �h5 as then the f­
pawn wil l have something to latch
onto. The supposed antidote runs 8
�h2! 'fies 9 .i.e2 lt.Jr4 10 .i.f3 rs 1 1
g3! �xh3 1 2 i.g2 (D) and now:
1 ) 12 fxe4 1 3 .i.e3 .i.f5 (1 3 ... 'fie7
..•
14 'fid2 h6 1 5 �xe4 is also good for
W hite) 1 4 �g 4 ! with advanta ge to
Wh ite, Bag irov-Vukic, Banja Luka
1 976.
h3 Systems 37
2) 12 ...f4 1 3 �f3 g5 ( 1 3 . . . �xf2! ?
1 4 �xf2 .i.g4 certainly deserves fur­
ther investigation and could well put
this line back in business; Vilela­
Frolov, Havana 1 99 1 continued 15
llxh7 .i.xf3 1 6 lhg7+ �xg7 1 7 .i.xf3
�d7 but perhaps more serious tests
of Black's sacrifice c an be found) 1 4
D.xh 3 ( 14 .i.xh3 g4 1 5 .i.g2 comes to
the same) 1 4 . . . g4 1 5 D.h l gxf3 1 6
'fixf3 'fig6 1 7 .i.h3, Vaganian-Chi­
burdanidze, USS R 1 98 1 , l 7 . . . .i.xh3
1 8 :xh 3 �d7 1 9 .i.d2 D.f7 20 0-0-0
with an edge for White.
8 .i.e3
8 .i.g5 is the next game.
8 ••• ltJh.5
8 �c5 9 tlJd2 aS is an alternative
.•.
plan, with the following possibili­
ties:
1 ) 10 a3 (recommended by ECO)
1 0 . . . �fd7 ! ? (but they don' t consider
this, nor 1 O . . . a4, which is an interest­
ing pawn sacrifice) 1 1 b4 f5 ! 1 2 exf5
gxf5 1 3 bxc5 f4 1 4 liJde4 fxe3 1 5
fxe3 ( 1 5 f3 �xc5 is also very good
for Black) 15 . . . �xc5 1 6 'fic2 .i.h6
with an advantage bordering on the
decisive, Verat-Hebert, Paris 1 995 .
2) 10 .i.e2 �8 ( I 0 . . . c6) 1 1 g4
( 1 1 0-0 f5 1 2 exf5 .i.xf5 1 3 �f3 'fie7
38 h3 Sy stems
1 4 .J:le l b6 = Gligoric-Byrne, Lenin­
grad IZ 1 97 3; presumably Black re­
frained from 1 2 ... gxf5 on account of
1 3 f4) 1 1 . . .f5 1 2 gxf5 ( 1 2 exf5 gxf5 1 3
.J:lg l 4Je6 ! ? is suggested by Knaak
and is a familiar motif in these lines)
1 2 . . . gxf5 1 3 exf5 ? ! ( 1 3 'ii'c 2 is bet­
ter) l 3 .. . iLxf5 1 4 4Jf3 ii.g6 with an
active game for Black, Cramling­
Djurhuus, Reykjavik 1995 .
3) 10 g4 c6 (it seems logical to
dissuade White from castling long
once he has played g4 ; on the other
hand 1 0 . . . 4Je8 1 1 'ii'e 2 ! f5 1 2 gxf5
gxf5 1 3 0-0-0 is promising for White)
1 1 iLe 2 (intending h4) 1 1 a4!? (this
•••
is Dolmatov's patent) and now (DJ :
3a) 12 i.xc5 dxc5 l 3 4Jxa4 'ii'a5
( l 3 . . . iLh6 was Dolmatov 's original
recommendation) 1 4 4Jc3 i.h6 1 5
'ii'c 2 ii.d7 1 6 .J:lgl 1/2- 1/2 Bagirov-Dol­
matov, Lucerne Wcht 1 993. Black's
dark-squ ared grip gives him enough
compensation for the pawn. I imag­
ine Dolmatov has analysed this in
some depth so keep a look-out for his
next game as it will no doubt prove
more informative than this one.
3b) 12 b4 axb3 1 3 axb3 l:.xal 1 4
'ii' xa l 4Ja6 1 5 ltJfl ? ! ( 1 5 'ii' b l c5 =)
1 5 ... lbe8 I 6 4Jg3 iLf6 ! 1 7 'ii'c l iLh4
1 8 iLd 3 4Jg7 with good play for
B lack, Shabalov-Dolmatov, Phila­
delphia 1 993 . As in line ' 3a', the
white king will struggle to find a safe
home.
3c) An important point behind
Dolmatov 's idea is that 12 h4? !
'ii'a5 ! leaves White facing the un­
pleasant threats of 1 3 . . . 4Jxe4 and
l 3 ... a3.
9 4Jh2
White usually takes some defen­
sive measures against the threat of
. . . f5 and this strange-looking knight
move has become the main line. The
alternatives are:
l ) 9 4Jd2 'ii'e8 10 4Jb3 (9 lDd2
looks more natural than 9 4Jh2 but
the problem is that after 1 0 iLe2 lDf4
1 1 iLf3 Black has 1 l . . . lDd3+) and
now (D):
l a) 10 15 1 1 c5 with a further
•••
branch:
l al ) 11 14 1 2 iLd2 4Jxc5 1 3
•••
4Jxc5 dxc5 1 4 b4 ! ? ( 14 iLe2 .J:lf6 ! 1 5
4Jb5 'ii'e 7 1 6 iLa5 f3 ! 1 7 iLxf3 4Jf4
1 8 ii.xc7 c4 + Markov-Sirota, corres
1 987) 1 4 . . . cxb4 1 5 4Jb5 f3 1 6 g 4
4Jg3 1 7 lDxc7 'ii' t7 1 8 lDxa8 ltJxh 1

is unc lear according to Koopman . I
suspect that White is better, though,
as his knight is much more likely to
escape than Black's. Black should
probably look for an improvement
earlier; one possibility, instead of the
flashy 1 6 . . . 4Jg3 , is 16 4Jf4 as 17
.•.
4Jxc7 fails to 1 7 . . . 4Jg2+ 1 8 i.xg2
fxg 2 19 D.g l Wf7 ; therefore White
should play 17 Wxr3 and the best I
can see for Black is 17 . . . 4Je6 ! fol­
lowed by 1 8 . . . 4Jd4 with a mess that
could easily work out in his favour.
l a2) l l 4Jr4 ! ? 1 2 cxd6 ( 1 2 g3
.•.
4Jh5 with ... f4 to follow) 12 . . . cxd6
1 3 exf5 gxf5 14 i.xf4 exf4+ 15 i.e2
(B urgess points out 1 5 We2 4Jb4
1 6 0-0-0 4Jxa2+ ! 1 7 4Jxa2 Wa4)
15 . . . f3 ! 1 6 gxf3 i.d7 1 7 D.g l �h8 1 8
Wd2 D.c8 1 9 0-0-0 4Jb4 20 �bl f4
2 1 4Jd4 We5 ! + Koopman-Burgess,
Biel 1987 .
1 b) 10 b6!? may be a tempo
••.
well spent. Klimenok-Lybin, corres
1 993 continued 1 1 i.e2 4Jf4 1 2 i.f3
f5 1 3 0-0 ( 1 3 h4 ! ?) 1 3 ... llJc5 14 4Jxc5
bxc5 1 5 D.e l Wt7 1 6 D.b l a5 with a
comfortable game for Black.
2) 9 4Jgl Wes 10 i.e2 and now:
2a) 10 4Jr4 l l i.r3 rs 1 2 g3 ( 1 2
.•.
4Jge2 We7 1 3 exf5 was Vilela-Pe­
corelli, Havana 1 99 1 , when Vilela
gives 1 3 . . . gxf5 1 4 i.xf4 exf4 1 5 0-0
i.e5 +) and now B urgess's sugges­
tion 12 4Jb4 looks like fun. After
•.•
13 gxr4 rxe4 14 4Jxe4 exr4 White
has a cou pie of bishop moves :
2al ) lS i.d2 i.f5 ! ? ( 15 . . . 4Jd3+
16 �f l 4Jxb2 17 Wb3 4Jxc4 1 8
Wxc4 i.xa l regains some material
but after 1 9 Wxc7 Black has lost the
initiative) 1 6 i.xb4 i.xe4 1 7 We2
h3 Systems 39
( 1 7 �fl i.xb2) 1 7 . . . i.xf3 1 8 4Jxf3
Wa4 19 Wd2 a5 20 i.c3 D.ae8+ 2 1
�fl Wxc4+ 22 �g2 D.e2 23 Wd 1
i.xc3 24 bxc3 D.f5 ! and Black will
pick up a third pawn leaving the situ­
ation unclear.
2a2) lS i.cl i. f5 1 6 We2 Wa4 !
with . . . D.ae8 to follow looks awk­
ward for White.
2b) It would be interesting to see
an example of 10 rS !?. I'm not sure
•..
if White can do better than 1 1 exf5
4Jf4 1 2 i.xf4 (with the knight on h2
White can castle here with a roughly
level game) 1 2 . . . exf4 1 3 fxg6 Wxg6
14 �fl 4Jc5 15 4Jf3, transposing to
the note to move 1 5 in the main
game.
3) 9 a3 is a different approach
whereby White hopes to lock the
knight on a6 out of the game. How­
ever, it seems risky to ignore Black's
kingside play. Flear-Cvitan, Bern
1 993 continued 9 . . .f5 (Piket once
played 9 . . . We8 but it is unnecessary
to support the knight on h5 since 10
exf5 gxf5 1 1 4Jxe5 We8 ! wins mate­
rial for Black as . . . f4 is coming) 1 0
b4 �h8 ! ? (Cvitan had previously
played 1 0 ... 4Jb8) 1 1 D.c 1 c5 1 2 dxc6
bxc6 1 3 i.e2 fxe4 14 4Jxe4 d5 1 5
cxd5 cxd5 1 6 i.g5 Wd7 1 7 4Jc5
4Jxc5 1 8 D.xc5 i.b7 and Black's
centre was very impressive.
9•.• Wes
10 i.e2 rs
10 4Jr4 1 1 i.r3 rs (D) is an im­
•..
portant alternative:
1 ) 12 h4? ! , intending to hit the
knight with g3 ( 1 2 g3? 4Jxh3 1 3
i.g2 f4 doesn't work here) disap­
peared from practice after the game
40 h3 Systems
Kavalek-Kasparov, B ugojno 1 982:
12 . . . 'iVe? 1 3 g3 lDb4 ! (a typical piece
sacrifice for this variation) and now:
l a) 14 'iVb3? was the move Kav­
alek actually played, but is 'boldness
bordering on su icide' (Kasparov) :
1 4 . . . lDfd3+ 1 5 'i!fe2 f4 1 6 i.d2 and
now 16 . . . lDxf2 ! was best. Kaspa­
rov 's analysis runs 1 7 'i!fxf2 lDd3+
1 8 'i!f g2 fxg3 1 9 'i!f xg3 :r4 ! 20 i.xf4
(20 lDg4 h5 ! 2 1 lDe3 i.f6 ! 22 lDg2
i.xh4+ 23 :xh4 'iVg5 + is winning)
20 . . . exf4+ 2 1 'i!f g2 'iVxh4 22 :hn
i.h3+ 23 'i!fh 1 i.xfl 24 :xn ll)f2+
25 :xf2 'iVxf2 26 'iVxb7 :f8 with ad­
vantage to Black.
1 b) Alternatively 14 gxf4 loses to
1 4 . . .fxe4 ! 1 5 lDxe4 ( 1 5 fxe5 lDd3+
1 6 'i!fe2 :xf3 ! 1 7 lDxf3 i.g4 1 8
lDxe4 lDxe5 -+) 1 5 . . . exf4 1 6 i.d2
lDd3+ 1 7 'i!fe2 lDc5 .
l e) White had to try 14 0..0
against which Kasparov proposes
1 4 . . . g5 .
2) 1 2 O·O lDc5 ( 1 2 . . . g5 1 3 :e t
'iVg6 1 4 ll)fl 'iVh6 15 exf5 i. xf5 1 6
i.g4 i.g6 1 7 lDg3 Ieft White with a
firm grip on the central light squares
in A nastasian-A. Kuzmin, Blagove­
shchensk 1 988; 1 2 . .. 'iVe? and 12 . . . b6
have also been played but these are
less active than 1 2 . . . lLJc5) 13 exf5 ( 1 3
li'c2 a5 1 4 :ad l b6 1 5 :re l 'iVt7 ,
Guseinov-A.Kuzm.in, USSR 1 99 1 ,
1 6 exf5 i.xf5 1 7 tDe4 with equality;
1 3 i.xc5 dxc5 1 4 'iVd2 i.d? is un­
clear) 1 3 . . . gxf5 ( 1 3 . . . i.xf5 was rec­
ommended by Kuzmin) 14 i.xf4
exf4 1 5 :e t 'iVd8 1 6 'iVd2 'iVg5 17
lDb5 lDe4 ! 1 8 'iVd3 :n 1 9 :e2? ( 1 9
lDd4 lDc5 2 0 'iVd 1 i.e5 2 1 b4 lDe4
22 :c 1 would have given White some
queenside counterplay) 1 9 . . . i.e5 20
lDd4 lDc5 2 1 'iVd 2 :lg? 22 'i!fh 1 'i!f h8
2 3 :g1 'iVf6 24 :d t i.d? with an ex­
cellent position for Black, Nikcevic­
Miles, Toulouse 1994.
1 1 exfS
Other moves don ' t seem to pose
any problems for Black, e.g. :
1 ) 1 1 i.xhS gxh5 1 2 lDf3 fxe4 1 3
lDd2 'iVg6 14 'iVe2 i.f5 15 0-0-0? ( 15
a3 is better) 15 ... lDb4 with a clear ad­
vantage to Black, Anastasian-Neve­
rov, Minsk 1 990.
2) 1 1 i.f3 lDf6 ( 1 1 . .. lDf4 trans­
poses to the note to Black's 1 0th
move) 1 2 g4 fxe4 1 3 i.xe4 b5 ! with
good play for Black, Berkovich­
Reeh, Budapest 1 99 1 .
3) 1 1 0-0 lDf6 1 2 exf5 gxf5 1 3 f4
exf4 1 4 i.xf4 , Gheorghiu-Cooper,
Novi Sad OL 1 990 and now Gheor­
ghiu gives 14 . . .lDc5 ! 15 lDb5 'iVf7 1 6
lDd4 lDce4 as unclear.
1 1 ... lDf4 (D)
12 i.xf4?
More prudent is 12 0-0, when
C.Hansen-Kasparov, Tasinge 1 990
continued 1 2 ... i.xfS 1 3 :et ! (in­
tending ll)fl �g3 and eyeing up the
black queen) 13 'iVf7 14 lDfi ( 1 4 a3
..•
lDc5 1 5 i.xc5 dxc5 1 6 i.f3 e4 ! gave

Black active play in Ibragimov­
Kruppa, Kherson 1 99 1 ) and now:
1) Kasparov sacrificed unsoundly
with 14 i.xh3? and after 15 gxh3
•••
ltJxh3 + 1 6 �g2 ltJxf2 1 7 'ii b l ! (not
allowing . . . 'iff5) 1 7 . . . e4 1 8 ltJg3 !
(threatening :n ) 1 8 . .. 'ifd7 1 9 i.xf2
%txf2+ 20 �xf2 'ifh3 2 1 'ifxe4 ! tl)c5
(2 l . . . %tf8 + 22 i.f3 'ifh2+ 23 �f l
'iix g3 24 �e2 'ifh2 + 25 i. g2 tl)c5
26 'ifg4 and Black has at most a few
swindl ing chances) 22 'fle7 i.e5
(22 . . . %tf8 + 23 'ifxf8 +) 2 3 11g l %tf8 +
24 �e l i.xg3+, the improvement 25
%txg3 ! 'iix g3 26 �d2 would have de­
fused the attack.
2) 14 ltJb4 is better and after 15
.••
ltJg3 :
2a) The game Flear-McDonald,
Dublin 1 99 1 continued 15 i.d3 1 6
•••
'ifd2 e4? ! ( 1 6 . . . i.xe2 �) 1 7 ltJcxe4
ltJxe2 + 1 8 ltJxe2 i.xe4 1 9 'ifxb4
i.xg2 20 �xg2 'iff3 + 2 1 �h2 %tae8
22 'ifd2 i.e5 + 23 ltJg3 h5 24 'ife2
'if xe2 (24 . . . h4 25 'ifxf3 %txf3 26
�g2) 25 %txe2 h4 26 %tae l hxg3+ 27
fxg3 %tf3 2 8 i. f4 and Black's enter­
taining play had only succeeded in
bringing him a lost ending.
2b) Kasparov gives the variation
1 5 ltJc2 1 6 i.xf4 ltJxe 1 1 7 ltJxf5
•..
h3 Systems 41
gxf5 1 8 i.d2 ltJxg2 1 9 �xg2 �h8 as
unclear.
12 .. • exf4
13 fxg6 'ifxg6
14 �
14 i. g4 i. xg4 1 5 hxg4 %tae8+ is
equally unpleasant for White.
14 .•• ltJc5 (D)
For a pawn White has forfeited his
right to castle, parted company with
his important dark-squared bishop
and opened files on the kingside for
the black rooks ; a heavy price by
anyone's standards. In addition, with
the extra pawn defending his king it
will only be in the ending that he can
hope to derive any benefit from it.
The odds are heavily stacked against
him arriving there with the rest of his
army intact.
15 %tel
White gives himsel f the option
of playing b4 . Wegner-Cramling,
Hamburg 1 99 1 went instead 15 ltJf3
i.d7 1 6 �g l �h8 1 7 �h2 %tg8 1 8
%tg l 'ifh6 1 9 'iid 2 i.f6 (threatening
. . . %txg2+) 20 �h l a5 2 1 %tae l %tg6
22 ltJh2 %tag8 (Black has calmly and
purposefully built up her position)
23 i.f3 i.f5 24 ltJe4 ltJxe4 25 i. xe4
42 h3 Systems

i. xe4 26 l%.xe4 i.e5 ! (26 ... 'ifxh3 27 move, 22 liJd4 loses to 22 . . .i. xd4 !
,

'ifxf4 is not so clear, but now it is 23 'ifxd4 lbe5 24 i.e2 f3 ! .

definitely a threat) 27 :ct.ee l (27 liJf3
l%.xg2 !) 27 . . .'fig7 28 f3 i.d4 and
Black won the exchange and eventu­
ally the game. A model performance
from Black.
15 •.• i. f5
16 li)f3 i.f6!
Preventing liJh4 and clearing the
g-file for the rooks.
17 'it'g l?!
As in Wegner-Cramling White
plays 'it'g l -h2 in order to connect his
rooks, but this is very slow. Polgar 22 ••• l%.xg2+!
suggests 17 liJd4 as an al ternative Not the sort of opportunity that
and after 1 7 . . . i.xd4 ! ? 1 8 'ifxd4 f3 ! Judit Polgar is likely to let pass by.
1 9 gxf3 ( 1 9 i.xf3 l%.ae8 20 'iii' g l liJd3 23 l%.xg2 i.xh3
2 1 1%.d l liJe 1 ! ) 1 9 . . . l%.t7 she concludes The rook on g2 has no square ( 24
that Black has a strong attack. 1%.g l i.fl + ) and apart from the mat­
17 ... 'it'h8 ing threats there is the simple threat
18 'it'h2 l%.g8 to play 24 . . . i.xg2 25 'it'xg2 i.xc3 26
19 1%.gl l%.xc3 'ifg7 + picking up the rook.
19 i.fi allows 1 9 . . . liJd3 ! . Hence White's reaction.
19 ..• 'iib6 ! 24 llle4 llle5!
20 i.n 25 lLJxe5
White digs in, defending his most If 25 liJ xf6 , then 25 ... i.g4 + wins
sensitive spots on g2 and h3. A more a whole wad of material , while 25
carefree move, such as 20 b4, would 'it'gl i.xg2 26 <it'xg2 l%.g8 + 27 'iii' f l
have led to a similar fate to the one 'ifh l + is also immediately decisive .
that White soon met: 20. . . l%.xg2+ ! 2 1 25 •.• i.xe5
l%.xg2 i.xh3. 26 llJg5
20 ••. 26 f3 loses prettily: 26 . . . i.xg2+
21 b4 27 'it'xg2 l%.g8 + 28 'it'fl 'ifh 1 + 29
22 i.d3? (D) 'iii' e2 l%.g2+ 30 liJf2 l%.xf2+ ! 3 1 'iii' xf2
S uddenly, Chemin can take the i.d4+ 32 'it'e2 'ifg2+ 33 'iii'e l 'iff2#.
suffering no longer and makes a 26 •.. i.xg2+
doomed bid for freedom. An under­ 27 'it'xg2 'ifx g5+
standable reaction as there was not 28 'it'f3
much future in waiting for Black to Or 28 'it'n f3 ! .
complete her rather plodding build­ 28 ••. l%.g8
up and the onl y other active-looking 29 'it'e2
 h3 Systems 43

29 'ite4 'ii'g 6+ 30 'itf3 1i'g4+ 3 1 by 1 2 . . . c6) 1 2 . . . c6 1 3 i.h6 cxd5 14

'ite4 f3+ is the end. i.xg7 'itxg7 15 1i'xd5 ( 15 cxd5 b5 is
29 .. . f3+ good for Black) 1 5 . . . i.c6 1 6 1i'xd6
0-1 lbxe4 1 7 lbxe4 i.xe4 1 8 !the 1 ltfd8 !
30 'iif el i.f4 and 30 'itxf3 1i'g4+ 1 9 1i'xe5+ 1i'xe5 20 lbxe5 lbb4 ! 2 1
3 1 'ite3 i. f4+ are the reasons for b3 lbxa2+ 22 'itb2 lbb4 23 f3 i.c2 !
White 's resignation. 24 ltxd8 l:txd8 25 'iifc 3 a5 with an
edge for Black, Anka-Gallagher,
Game 6 Biel 1 992.
Raetsky - Gallagher 2) 10 ltgl 'ith8 11 1i'd2 lbdcS
Hastings 1 99213 (D) and now:

1 d4 lbf6
2 c4 g6
3 lbc3 i.gT
4 e4 d6
s ll)f3 0-0
6 h3 eS
7 dS lba6
8 i.gS (D)

2a) 12 0-0-0 i.d7 1 3 h4 f5 (a

good rule of thumb is that when
White is threatening the asphyxiat­
ing h5 , it's time to lash out with . . .f5 )
1 4 gxf5 gxf5 1 5 i.h6 i. xh6 1 6
1i'xh6 1i'e7 1 7 lbg5 fxe4 ! ? (White's
position looks more menacing than
it actually is; Dolmatov also gives
1 7 . . . l:tf6 1 8 1i'h5 ltaf8 as ;) 1 8
8 . .. h6 lbcxe4 lbxe4 1 9 lbxe4 ltf4 20 i.d3
The other way to unpin, 8 1i'e8••• , lbb4 2 1 i.bl lbxa2+ 22 'itd2 i. f5
is more fashionable. After 9 g4 (oth­ 23 'ite3 San Segundo-Cvitan, Mos­
erwise . . .lbh5 ) 9 lbd7 (9 . . . c6 !?­
.•. cow OL 1 994, and now instead of
Dolmatov) there is: 23 ltxh4 ? 24 1i'g5 !, with the point
..•

1 ) 10 1i'd2 lbdc5 1 1 0-0-0 i.d7 that after 24 . . . 1i'xg5 25 lbxg5 White

(it looks risky to play . . . f5 before is threatening lbf7#, 23 :rs would
..•

connecting the rooks) 1 2 i.e2 ( 1 2 be very good for Black.

i.e3, as played in S an Segundo­ 2b) 1 2 lbh4 is presumably moti­
Illescas, Madrid 1 994, should be met vated by a desire to prevent . . . f5 and
44 h3 Systems
a reluctance to commit the king to
the queenside where it may come un­
der attack. It is awarded an ' ! ' by
Burgess and Hjartarson who both
give 12 .i d7 1 3 %1bl lLJa4 1 4 lLJxa4
•..
.ixa4 1 5 f3 .i d7 1 6 %te l ! 0sten­
stad-Maki, Haifa Echt 1 989. This as­
sessment is perhaps debatable, but
anyway Black's play wasn't very
convincing . I would prefer 12 c6 .. .
and after 1 3 %1bl ( 1 3 dxc6 1ixc6)
1 3 . . . cxd5 14 cxd5 .id7 Black doesn't
have to worry about 15 b4 lLJa4 1 6
lLJxa4 .ixa4 1 7 b5 lLJc5 as he can
free his bishop with . . . a6. Although
White can undoubtedly do better
than this variation, his underl ying
problem, king security, is not going
to disappear.
3) 1 0 lLld2 rs 1 1 gxf5 gxfS 12
%1g l ( 1 2 exf5 lLJdc5 1 3 f6 .ixf6 1 4
.ih6 1ig6 1 5 .ixf8 �xf8 1 6 1if3
.if5 gave Black plenty for the ex­
change in Shirov-Badea, Moscow
1 99 1 ) 12 �hS (D) and now:
•..
3a) 13 .id3?! lLJdc5 1 4 .ibl ,
Gulko-Hj artarson, Reykjavik 1 99 1 ,
and now 1 4 . . . fxe4 1 5 lLJdxe4 .ixh3
1 6 .ie3 .if5 1 7 lLJxc5 lLJxc5 1 8
i.xf5 %1xf5 1 9 .i xc5 dxc5 20 lLJe4
1ie7 followed by . . . .ih6 gives Black
a clear advantage according to Hjar­
tarson.
3b) 1 3 a3 lLJf6 1 4 'ii'c 2 lLJc5 1 5
.ie3 lLJcxe4 l 6 lbdxe4 fxe4 1 7 0-0-0
.i f5 was unclear in Bagirov-Ling­
nau, Cuxhaven 1 994 .
3c) 13 exf5 lLJdc5 ( 1 3 . . . %1xf5 1 4
lt.Jde4 i) 1 4 f6 .ixf6 1 5 .ih6 %1g8 1 6
%1xg8+ 1ixg8 1 7 1ih5 .id7 (San
Segundo gives the nice variation
1 7 . . . lLJb4 1 8 lLJde4 lLJxe4 1 9 lLJxe4
lLJc2+ 20 �d2 lLJxa l 2 1 lLJxf6 11g6
22 1ixg6 hxg6 23 .id3 .if5 24 .ixf5
gxf5 25 �c 1 when the black pieces
are strangely paralysed) 1 8 0-0-0
.ie8 1 9 1if5 .ig7 20 .ie3 .ig6 2 1
11g4 1it7 with about equal chances,
San Segundo-Shirov, Madrid 1 994 .
9 .ie3 lLJcS
There are a number of differences
created by the black pawn being on
h6. One of them is that 9 lLJhS is not
..•
so good as in Game 5 , as after 1 0
lLJh2 1ie8 1 1 1id2 White gains a
tempo hitting the h-pawn. Comas­
Komljenovic, S an Sebastian 1 99 1
continued 1 1 . . . �h7 1 2 0-0-0 .id7 1 3
lLJg4 f5 1 4 exf5 gxf5 1 5 .ixh6 ( 1 5
.id3 ! ? e4 1 6 .ic2 is worth consider­
ing as 1 6 . . . fxg4 1 7 hxg4 is danger­
ous for B l ack) 15 . . . 1ig 6 ( 1 5 . . .fxg4
1 6 .i d3+ gives White a strong at­
tack) 1 6 .i x g7 �xg7 1 7 lLJe3 f4 1 8
lLJc2 .if5 1 9 .id3 2;, Black has some
positional compensation in return
for the pawn.
10 lLJd2 a5
1 1 g4
1 1 .ie2 is a more sol id approach
and this actually transposes to Grune
7.

I haven' t seen any examples of 1 1
a3 here, but I suggest that Black con­
tinues as in Game 6 ( 1 l . . . lZJfd7 1 2 b4
f5) although it won' t be quite as dev­
astating as he doesn't have . . . i.h6.
11 c6
12 i.e2 lZJe8!? (D)
Playing on both wings may seem
to be over-ambitious and although it
worked well in this game many play­
ers may prefer line '2' below. If
Black does wish to play with . . . lZJe8
he should wait for i.e2 so that the
dangerous pl an of 'iVe2 and 0-0-0 is
no longer available to White. Alter­
natives for Black are:
1) 12 a4. Recommended with
..•
the pawn on h7, but as one of the
main points behi nd the pawn sacri­
fice is to activate the bishop via h6, it
is obviously less effective here.
2) 12 ... i.d7. Conversely, this is
one of the situations where it seems
favourable for Black to have a pawn
on h6 as White doesn't have the an­
noying reply g5 . Instead, there is:
2a) 13 f3 l:tb8 1 4 lZJfl ( 14 lZJb3
b6 15 'iVd2 cxd5 16 cxd5 a4 1 7 lZJxc5
bxc5 was pleasant for Black in Gri­
vas� Ehl vest, Komotini 1 992) 1 4 . . . h5
h3 Systems 45
( 1 4 . . . cxd5 1 5 cxd5 b5 looks worth a
try) 1 5 i.g5 cxd5 16 cxd5 'iVb6 1 7
'iVd2 lZJh7 1 8 i.h4 i.f6 1 9 i.xf6 1h- 1h
Grivas-Smirin, Komotini 1 992.
2b) 13 b4?! a4 14 gS ( 1 4 i.xc5
dxc5 15 lZJxa4 cxd5 16 lZJxc5 dxe4 is
promisi ng for Black according to
Nunn; White may pick up a pawn but
his king has no home) 14 ...bxgS and
now:
2bl ) 15 hxgS lZJh7 1 6 l:tg l ( 1 6
lZJf3 'iVaS) 1 6 . . .'iVaS 1 7 l:.c l cxd5 1 8
cxd5 b5 with a healthy initiative for
Black, Zeller-Poldauf, Berlin 1993.
2b2) 15 i.xgS 'iVa5 16 'iVbl cxd5
1 7 cxd5 b5 1 8 a3 l:.fb8 with an excel­
lent game for Black, Chiburdanidze­
Nunn, Linares 1988.
13 lZJb3
13 b4 is clearly an alternative to
which Black should reply 1 3 . . . f5 .
13 ... cxdS!
Perhaps White missed that Black
would be able to support his knight
on c5 with . . . b6. Of course 1 4 lZJxc5
is now met by 1 4 ... d4 ! .
1 4 cxdS b6
1 s h4 rs
16 gxfS gxfS
17 exrs i.xrs
1 8 lZJxcS bxcS
19 i.d3 (D)
19 ... e4!
Naturally White is not going to be
allowed to cal mly blockade the e4-
square.
20 lZJxe4
A fter 20 i.xe4 i.xc3 + 2 1 bxc3
Black can choose between 21 i.xe4 .•.
22 'iVg4+ lZJg7 23 1Vxe4 'iVf6 and
21 ... 'iVf6, both of which offer plenty
of compensation for the sacrificed
46 h3 Systems

pawn. During the g ame I intended 32 ltb7 ltf7 33 ttJc4 ttJxd5 34 ttJxd6
the latter. ltxb7 35 ttJxb7 the ending looks like
20 ... ltb8! a draw.
21 ttJg3 .ixd3 29 .•• 11417
22 Wxd3 Wf6! Now everything is under control
Combining defence of the king­ again.
side with the attack on the queen­ 30 &4 lt:JxdS
side. 31 ltxf7 ltxf7
23 ttJe4 Wxb2 32 ll\xd6 Wat+
24 ltdl Wes 33 'itie2 Wxa2+
25 ltgl ltb4! 34 'itiel Wal+
26 ttJd2 ltxh4 35 'itie2 1ii'h2+
Black's miraculous major pieces 36 'itiel Wc3+
have netted a pawn through some 37 'itie2 lte7
heavy industry. White lost on time in a hopeless
27 Wg6 lt:Jc7! situation. Incidental ly, 37...ltxf.2+
28 ltbl ! would have been stronger as after 38
With his whole position collaps­ 'itixt2 Wxe3+ 39 'itifl 'ifc 1 + 40 'itif2
ing White finds the only chance. Wd2+ 4 1 'itig3 We3+ White can re-
28 ... lthf4? ( D) sign.
Under severe time�pressure I was 0 .. 1
quite pleased with this move but, in
fact, it throws the win away. Much Game 7
better was 28 ltb4! (not 28 ... ttJxd5 ?
••• Alexandrov - Zakharevich
29 ltb7) 29 'ifxh6 (29 ltxb4 axb4 30 St Petersburg 1 994
'ifxh6 Wal + fol lowed by . . . Wxg l )
29 . . .ltxbl + 30 lt:Jxbl lt:Jxd5 -+. 1 d4 ttJf6
29 ltb7? 2 c4 g6
And fortunately White fails to 3 ttJc3 .ig7
take his chance. After 29 Wxg7+! 4 e4 d6
Wxg7 30 ltxg7 + 'itixg7 3 1 .ixf4 ltxf4 S .ie2 0-0

6 lbf3 e5
7 d5 a5
8 h3 (D)
This line used to be abou t as far
off the beaten track as you could get
(in fact the first time I came across
the move was when Novik played it
against me three years ago), but a
glance at Info rmator 64 will reveal
no fewer than four games with 8 h3 .
Its main champion is the young Rus­
sian Zviagintsev, but other strong
grandmasters, such as Beliavsky and
Gulko, have also thought it worth a
whirl . Although strictly speaking the
vari�.tion belongs to the Petrosian
S ystem its spiritual home is to be
found amongst the h3 systems of this
chapter and, indeed, there are many
direct transpositions to the other
games. However, it seems to me (and
others) that an earl y .ie2 reduces
White 's options and denies him
some important resources. For ex­
ample, .ih3 is no longer available
and the queen's path to the kingside
has been blocked. True, Black has
committed himself to . . . a5 but this is
a much more useful move which will
very rarely harm him. My conclusion,
h3 Systems 47
therefore, is that whilst 8 h3 is fash­
ionable at the moment it is destined
for a return to the wilderness.
8 ... lba6
8 lbbd7 is less flexible so the
..•
only alternative worth considering is
8 lbh5 after which there is :
.•.
I ) 9 g3 f5 (9 . . . lba6 ! ?) I 0 exf5
gxf5 1 1 lbg5 ( 1 1 .ig5 'ife8 is un­
clear) 1 1 . . . lbf6 1 2 g4 (White wishes
to control e4) 1 2 . . . 'ife7 1 3 gxf5 .ixf5
1 4 .ig4 lbxg4 1 5 hxg4 .ig6 1 6 l2Je6,
Pogorelov-Becerra, Cordoba 1 995 ,
and now as Pogorelov points out,
1 6 . . . lba6 1 7 lbxf8 ( 1 7 lbxg7 'ifxg7
1 8 .ih6 'iff6 and 1 7 .i g5 'iff7 are
also playable for Black) 1 7 . . Jtxf8
wou ld give Black a lot of play for
the exchange (which makes it hard
to understand why he sprinkled
White 's preceding moves so liber­
ally with exclamation marks).
2) 9 lLld.2 lbf4 1 0 .ifl lba6 1 1 g3
lbh5 (both sides have wasted time on
the kingside and Black hopes to
profit from the weakness on h3) 1 2
lbb3 (on 1 2 .ie2, Beliavsky recom­
mends 1 2 . . . lbf6 followed by . . . c6,
while Knaak suggests l 2 . . . lbc5 , not
fearing 1 3 .ixh5 gxh5 1 4 'if xh5 f5 ;
but here Beliavsky proposes 1 3
lbb3) 1 2 . . . c6 1 3 .ie3 .id7 (intend­
ing . . . a4) 1 4 a4 lbb4 15 lZ.c l 'ife7 1 6
c5 ! ? c xd5 1 7 c xd6 'ifd8 ! ( 1 7 . . . 'ifxd6
1 8 .ic5 'iff6 1 9 lbxa5 !) 1 8 lbxd5 ( 1 8
exd5 b6 !) 1 8 . . . lbxd5 1 9 'ifxd5 lbf6
20 'ifd3 .ia4 with a double-edged
game, Zviagintsev-Beliavsky, Yu­
goslavia 1 995 .
9 .ig5
9 .ie3 has not been played much
so much here, as White prefers to
48 h3 Systems
encourage Black to play . . . h6. In fact
the position after 9 . . . lt:Jc5 1 0 llJd2 is
considered in Game 5 .
9
•.• ii'e8
9 h6 is Black's main alternative .
•.•
After 10 .ie3 llJcS 1 1 lt:Jd2 we are
going to concentrate on 1 1 liJh7, .••
which is an attempt to exploit the
free . . . h6 that Black has been given.
Before doing so though, it should be
pointed out that 1 1 llJeS 1 2 g4 c6 ! ?
...
( l 2 . . .f5) transposes to Raetsky-Gal­
lagher, Game 6.
After 1 1 llJh7 (D), there is:
..•
1 ) 1 2 g4 f5 1 3 gxf5 gxf5 1 4 ltg l
'it>h8 1 5 exf5 .ixf5 and Nunn, who
originally su ggested 1 1 . .. llJh?, com­
ments that White's position is simply
not strong enough for him to gain
control over e4 . For example, after
1 6 .ixc5 dxc5 17 .ig4 ii'h4 ! White's
king is far from safety.
2) 1 2 h4 h5 ! ? ( 1 2 ... f5 1 3 h5 ii'f6
14 ii'c2 llJa6 1 5 f3 f4 1 6 .if2 g5 was
also satisfactory for Black, Yermo­
linsky-Hellers, New York 1 993) 1 3
llJb3 b6 1 4 f 3 f5 1 5 lt:Jxc5 bxc5 1 6
ii'c2 llJf6 17 .i d 3 f4 1 8 .i n .id? ,
Zviagintsev-Nevednichy, Yu gosla­
via 1 995, with about equal chances.
3) 1 2 llJb3 llJxb3 ( 1 2 . . . b6 ! ?) 1 3
axb3 i.d7 ( 1 3 . . .f5 14 b4) 1 4 ii'd2 h5
1 5 0-0 .if6 1 6 c5 i. g5 1 7 cxd6 cxd6
1 8 llJb5 .ixb5 1 9 .ix b5 .ixe3 20
ii'xe3 llJf6 2 1 ltfc 1 and I can' t de­
cide whether White has an edge or
not. In Hjartarson-Hellers, Oster­
sund Z 1 992 Black managed to hold
the balance with the aid of some re­
sourceful play: 2 1 . . . h4 22 lta4 ltc8 !
2 3 ltxc8 ii'xc8 24 ii'g5 (24 ltxa5
ii'c2) 24 ...'it>g? 25 ltc4 'ti'd8 26 ii'e3
(26 ii'xh4? ii'b6 27 i. a4 ii'a6 and
. . . b5 is going to hurt) 26 . . . llJh5 27
ltc3 ( so that the bishop can return
home in the event of . . . llJf4) 27 . . . llJf6
28 l:.c4 4Jh5 29 ltc3 1'2- 1'2 .
4) 12 0-0 .id? ( 1 2 . . . f5 1 3 exf5
.ixf5 ! ?) 1 3 ii'c2 llJg5 ! ? 14 h4 lbb7
1 5 h5 llJf6 (of Black's last five
moves, four have been with this
knight and it has ended up where it
started; in return for this scandalous
waste of time he has lured the white
h-pawn forward and White is about
to take back his move ii'c 2) 1 6 'ti'd l
gxh5 17 .ixh5 tt:Jd3 1 8 .if3 ii'c8 1 9
ii'c2 llJf4 20 c5 dxc5 2 1 .ixc5 lte8
22 tiJe2 .ib5 23 lDc4 llJ4xd5 ! with a
good game for Black, Beliavsky­
Sher, Bern 1 995 .
1 0 llJd2
The immediate 10 g4 may simply
transpose after 1 o llJd7 1 1 l:.gl
•..
ltJcicS (D) and then 1 2 lZJd2, but there
are a couple of important 1 2th move
alternatives for White:
1 ) 1 2 ii'd2 'it>h8 with a couple of
examples:
l a) 13 h4 f6 14 .ie3, Zviagintsev­
Novik, S t Petersburg 1 994, and now
Glatman gives 14 fS 1 5 gxf5 gxf5
.•.

1 6 1'.h6 l:.g8 17 1'.xg7+ l:.xg7 1 8
l:.xg7 'fi;xg7 1 9 1fg5+ 1Vg6 20 1ile7+
1iff7 2 1 1ilg5 + as equal. There's ob­
viously plenty to investigate here but
it's clear that Black should prefer
14 . . .f5 to 14 1'. d7 15 h5, which was
••.
Novik's choice in the game.
1 b) 1 3 0-0-0 1'. d 7 14 b4 and
now:
1 bl ) 14 it)b4!? (threatening to
.. .
play 15 . . . it)xa2+) 15 'fi;bl f5 1 6 gxf5
it)xe4? l 1 7 it)xe4 1'.xf5 1 8 1'.d3 1fa4
19 b3 1fa3 20 1fe2 a4 2 1 1'.c 1 (Black
must have missed this when he em­
barked on the combination) 2 1 ... axb3
22 1'.xa3 l:.xa3 , Piskov-Damlj ano­
vic, Belgrade 1 995 , looks l ike pure
fantasy to me, although judging by
Piskov's notes in Informator he was
a worried man at the time. He man­
aged to beat off the attack by return­
ing a large chunk of material : 23
ft)fg5 ! l:.fa8 24 axb3 l:.xb3+ 25
1*'b2 ! l:.xb2+ 26 'fi;xb2 it)xd3+ 27
l:.xd3 h6 28 f3 hxg5 29 hxg5 with a
clear advantage for White.
l b2) 14 fS looks better to me.
•••
After, for example, 1 5 gxf5 gxf5 1 6
1'.h6 1'.xh6 1 7 1fxh6 1ile7 Black is
re ady to expel the queen with ... l:.f6
and, if necessary, . . . 1'.e8.
h3 Systems 49
2) 12 h4 and now:
2a) 12 c6 1 3 h5 cxd5?! (better is
.••
1 3 ... h6 14 1'.e3 g5 , although after 1 5
it)d2, intending ft)f1 -g3, White still
has an edge) 14 h6 ! 1'.h8 1 5 cxd5 (15
it)xd5 is also good) l 5 ... 1'.d7 16 a4
with a clear advantage for White,
Piket-J.Polgar, Aruba 1 995 .
2b) Piket prefers to drive the
bishop away immediately and pro­
poses 12 ••• b6 or 12 f6 as possible
•••
improvements, but I believe the criti­
cal line to be 12 'fi;h8 1 3 h5 (13 1fd2
•..
transposes to line ' l ' ) 1 3 . . . gxh5 ! 1 4
gxh5 f5 with a double-edged game.
10 ••• llxl7
Black intends . . . it)d7-c5 followed
by .. .f5 . I believe this plan was intro­
duced into practice by your not so
humble author who knew the idea in
similar positions. Another method is
1 0 b6 1 1 1'.e3 lDh7 12 g4 f5 , but the
•• •
only comment I have seen on this is
' unclear, Piskov-Arsovic, Belgrade
1 995 ' ; not very illuminating.
1 1 g4
1 1 a3!? (D) forces Black to switch
plan as 1 1 . .. ttJdc5 would now be met
by 1 2 b4. Alternatively, Black can
try:
50 h3 Systems

1 ) 1 1 f5 1 2 exf5 and now since

..•
12 gxf5?? 1 3 .th5 drops a queen,
..•
Black can play 1 2 Axf5 (which I
.•.
don ' t like) or try 12 e4! ? (as sug­
.•.
gested by Poluliakhov). After 13 fxg6
'ifxg6 1 4 lDdxe4 tDac5 1 5 ttJxc5
tDxc5 1 6 .te3 'ifxg2 1 7 .tf3 'ifg6,
Black certainly has good play, but
I' m not so sure about 13 lDdxe4 gxf5
1 4 ltJd2 ! , e.g. 1 4 . . . 'ifg6 1 5 f4 .t xc3
1 6 bxc3 h6 1 7 .th5 ifg7 1 8 .th4
13 a3 presents Black with no par­
ticular problems. M .lvanov-Cvitan,
Cappelle la Grande 1 995 continued
1 3 . . . .td7 1 4 h4 f5 1 5 g xf5 gxf5 1 6
h5 .tf6 1 7 .t xf6+ Axf6 1 8 'ifc2
'iff8 ! 1 9 0-0-0 fxe4 20 ttJcxe4 llf4
2 1 f3 .tf5 +.
13 . .. f5!
In Novik- Gallagher, Oberwart
1 993 Black didn' t have the courage
of his convictions and played 13 c6 ..•

'ifxg2 1 9 'iff3 ! looks very difficult
for Black.
2) 1 1 f6 1 2 .t h4 and now:
• .•
2a) 12- .th6 l 3 b4 .tg5 1 4 .txg5
. .td7 1 6 h4 b5 1 7 'ifd2 b4 1 8 lDd l f6
fxg5 1 5 c5 dxc5 1 6 bxa5 lDf6 1 7
ttJc4 was good for White in Kra­
senkov-Zakharevich, S t Petersburg
1 994.
2b) Perhaps Black can try 12 g5 •..
1 3 .tg3 f5 1 4 exf5 ( 1 4 f3 lDf6)
14 . . .ltJf6 after which I can't see any­
thing devastating for White, e.g. :
2bl ) 15 'ifc2 ltJh5 ! when 1 6 .tg4
should be met by 1 6 ... ttJf4! rather
than 16 ttJx g3 1 7 fxg3 h5 1 8 f6 !
.••
after which he was a little fortunate
not to get into difficulties. The game
continued 1 4 ltJg 3 cxd5 1 5 cxd5
1 9 .te3 'ife7 (intending . . . f5) 20 h5
gxh5 2 1 f3 ! ? hxg4 22 fxg4 and now
22 . . . f5 ! gained some desperately
needed space on the kingside. The
position is unclear (0- 1 , 50 after
many adventures).
14 gxf5 gxf5
15 ltJg3 (D)
Novik points out that 15 .th5
'ifd7 1 6 lDg 3 f4 1 7 lDf5 can be met
by 17 ttJxe4; even if this tactic were
.•.

when White gains control of e4 . not available, the exchange sacrifice

2b2) 1 5 h4 gxh4 1 6 .t xh4 .txf5 with 17 Axf5 seems to be pretty
•.•

looks fine for Black. crushing.

2b3) 15 b4 .txf5 and White has
no good way to defend his pawn on
b4 .
2b4) 15 ltJde4 .txf5 ! ? 1 6 ttJxg5
lDc5 1 7 0-0 a4 with what looks like
a good pawn's worth of compensa­
tion.
11 ••. ttJdc5
12 llgl 'liih8
13 lDO!?
If Black is going to play . . . f5 then
he must do so now as once the knight
arrives on g3 it will be very risky. 15 . .• fxe4! ?

If I remember correctl y my post­
mortem with Novik concluded that
15 �f6 was quite good for Black; I
•••
don' t think we examined the text,
which commits Black to a queen sac­
rifice, albeit an extremely powerful
one.
16 ltJhS
This looks very strong but Black
has a nasty surprise in store for his
opponent. Alternatively, 1 6 ltJcxe4
ltJxe4 1 7 ltJxe4 �f5 18 �h5 'ii'b8 (to
avoid exchanging the l ight-squared
bishops; I suppose 1 8 . . . �xe4? ! is a
bit fanciful and unnecessary) 1 9 'ii'e2
is u nclear according to Knaak. If
Black is feeling ambitious he may
try l 9 . . . b5 intending to ' sac' the ex­
change after 20 �e7 (20 cxb5 �xe4
followed by . . . ltJc5) with 20 . . . bxc4
(or 20 . . . �xe4) 2 1 �xf8 'ii'xf8 22
'ii'xc4 � xe4 23 "ii'xe4 ltJc5 24 "ii'f3
'ii'e 7. White has a rough ride ahead of
him
16 •.• 'ii't7
Forced, as 16 llgS 1 7 ltJxg7
•••
<Jitxg7 1 8 �e7+ ! wins for White.
1 7 �h4
1 7 ll g 2 is met by 1 7 . . .� xh3 and
1 7 ltJxg7 'ii'x g7 looks favourable for
Black, who can of course also inves­
tigate 17 ... 'ii'xf2+.
17 ••. �h6!!
Perhaps White had been expect­
ing 1 7 ll gS when 1 8 llxg7 ! llxg7
..•
1 9 �f6 'ii'x f6 20 ltJxf6 llgl + 2 1 �fl
� xh3 22 'ii'h5 ! gives him a winning
attack (Knaak).
1 8 �f6+ "ii'xf'6
19 ltJxf'6 llxf'6 (D)
S o what does Black have for his
queen? Well , materially speaki ng
h3 Systems 51
he has bishop, knight and pawn,
which works out as a two-point defi­
cit (please forgive me for being so
basic) or the equivalent of an ex­
change less. Quite a meagre amount,
in fact, when you take Black's mas­
sive positional advantage into ac­
count. White's king is stuck in the
middle and the black rooks will soon
be doubled against one of White's
most sensitive points, namely f2.
The black minor pieces are also a
group worth envying; the bishops
simply radiate power and the knights
are poised to infiltrate the enemy
camp. I suspect that the white posi­
tion is already beyond redemption.
20 a3 �rs
21 llg3
White's last two moves have been
geared towards preventing . .. ltJd3+.
A position such as the one that arises
after 21 b4 ltJd3 + 22 �xd3 exd3 23
'ii'h 5 llaf8 would be completely
hopeless for him.
21 ••• llaf8
22 'ii>f1
The king attempts to hide in the
corner - probably the only chance.
22 ••• � g6
23 <Jitgl
52 h3 Systems

23 l:lg2 doesn't help on account of 26 ..• ltJc2!

23 . . . i.e3 . 27 i.x g6
23
... llxf2 (D) 27 l:.bl loses to 27 . . . ltJe3 and 27
l:la2 i. xh5 28 'ifxh5 llfl + 29 �h2
l:l8f2+ 30 l:lg2 i.f4 i s mate .
27 ... ltJxa l
28 .txe4
28 \i'xa l hxg6 29 'ifd 1 (29 l:lxg6
i.e3) 29. . . e3 is given by Belov.
28 . •. ltJab3
29 ith5 :2r6
30 l:lf3?
This loses material . White had to
play 30 i. g2 although Black has a
winning advantage after 30 . . . ltJ<l4.
24 b4?! 30 .. . ltJxe4
Against Knaak's suggestion 24 3 1 ltJxe4
l:lbl Black can maintain the bind with White must have overlooked that
24 . . . a4 or search for something more 3 1 l:lxf6 is met by 3 1 . . . i.e3 + ! 32
devastating. It may be that 24 i.h5 is �g2 ltJxf6.
White's last chance. 31 ... l:[xf3
24
•.• axb4 32 'ifxh6 ltJd4
25 axb4 ltJxb4 33 h4 l:l3f4
Black has now achieved material 34 ltJg5 :n+
parity. 3 5 �g2 l:l8f2+
26 i.h5 36 'itJ>h3 llhl+
White can exchange a rook with 0-1
26 l:la8 l:lxa8 27 �xf2 but this does 36 �g3 ltJf5+ and 36 �g4 l:lf4+
not lessen Black's pressure . are the end.
 3 The Averba kh
The Averbakh system is charac­
terized by the moves 1 d4 liJf6 2 c4
g6 3 liJc3 i. g7 4 e4 d6 5 i.e2 0-0 6
i.g5 (D).
The original idea was to try to sti­
fle Black's traditional kingside coun­
terplay by making it difficult for
Black to play . . . e5 and . . . f5 . The first
point is that 6 . . . e5?? loses at once : 7
dxe5 dxe5 8 1Wxd8 llxd8 9 liJd5. So
if Black is not going to abandon the
plan of playing for . . . e5, he must first
. play a preparatory move. Until re­
cently there were thought to be only
a couple of options. Firstly, he could
drive the bishop away with 6 . . . h6
and then play . . . e5. Bu t . . . h6 is not al­
ways a desirable move for Black to
play in the King's Indian. Not only
does it weaken his kingside but very
often White will be able to gain a
tempo with 1Wd2. The second idea
was to support the advance . . . e5 by
means o f 6 . . . liJbd7, but this too has
its downside. As the knight on d7
cramps the whole of the black
queenside it cannot remain there too
long. This means that B lack has to
play an early ... llJc5 and . . a5, which,
.
although a solid idea, often leaves
Black passively placed.
Following the modern trend in the
King' s Indian a third idea has come
to Black' s rescue, namely 6 . . . llJa6.
The advance . . .e5 is now playable as,
for example, 7 1ffd2 e5 8 dxe5 dxe5 9
1Wxd8 llxd8 1 0 lDd5 leads nowhere
as Black can simply reply 1 O . lld6.. .
Although the basic idea behind
6 . . . lDa6 is to prepare . . . e5 , the move
is very flexible and in certain cases
Black can fol low up with . . . c5 or
even . . . c6 followed by . . . llJc7.
Grune 8 concentrates on 7 f4 and a
few unusual White moves, Game 9
game deal s with 7 lDf3 and 7 h4,
whil st Game 1 0 tackles the main
l ine, 7 1Wd2. I am proposing that
Black continues 7 . . . e5 8 d5 1ffe 8 ! ? .
Grune 8
Yakovich - Smirin
Munich 1 993
1 d4 lDf6
2 c4 g6
3 lZJc3 i.g7
4 e4 d6
5 i.e2 0-0
6 i.gS llJa6
54 The A verbakh
7 f4 (D)
The other main lines, 7 h4, 7 ttJf3
and 7 1Wd2 will be considered in sub­
sequent games, whils t below we
examine a few rarely played vari­
ations:
1 ) 7 .i f3 c6 8 1Wd2 e5 9 ttJge2
1We8 10 0-0 h5 1 1 d5 and now:
1 a) Hort-Cramling, Prague 1 995
continued 1 1 cxdS 1 2 .ixf6 ! ? .ixf6
•••
1 3 tlJxd5 .ig7?! ( 1 3 . . . .id8 may leave
the black king a little bare but would
at least all ow th e knight on a6 to get
into the game) 14 tlJec3 .id7 15
l:tfd 1 l:td8 16 b4 .ie6 17 1We3 b6 1 8
a4 with a big plus for White.
l b) 11 tlJh 7 1 2 .ie3 c5 would
...
h ave been my choice and the logical
follow-up to 1 0 . . . h5 .
2) 7 .id3. White often reposi­
tions this bishop on c2 in the Aver­
bakh, but he usually waits until a
decent interval has passed before
moving it again� if I had played 5
.ie2 followed by 7 .id3 for my old
school team I would have been
benched for the rest of the season
and I'm sure that a similar fate would
have befallen B areev, notwithstand­
ing any elaborate explanations he
may have offered. 7 . . . e5 (I suppose
7 . . . c5 also comes into consideration)
8 d5 c6 (8 . . . 1We8 ) 9 tlJge2 ltJc5 1 0
.ic2 a5 1 1 1Wd2 (after 1 1 a3 cxd5 1 2
cxd5 Black c an sacrifice a pawn with
12 a4!?, or allow White to advance
•.•
his b-pawn as 12 i.d7 1 3 b4 axb4
••.
1 4 axb4 l:txal 1 5 1Wxal ttJa6 is un­
clear) 1 1 . . . cxd5 1 2 exd5 ! ? ( 1 2 cxd5
.id7 intending . . . b5 looks OK for
Black whil st 1 2 ttJxd5? ltJcxe4 ! 1 3
.ixe4 ttJxe4 1 4 .ixd8 tlJxd2 15 .ie7
ttJxc4 1 6 .ixf8 �xf8 leads to an un­
pleasant ending for White) 1 2 . . . .id7
1 3 0-0 1Wb6 14 tlJg3 ? ! (Gelfand pre­
fers the immediate 14 .ie3) 14 ... l:tfc8
1 5 .ie3 1Wa6 1 6 1We2 tlJe8 1 7 f4 f5
1 8 tiJb5, Bareev-Gelfand, Linares
1 994, and now, according to Gelfand
1 8 . . . e4 ! would be better for Black.
After 19 ttJd4 tiJd3 ! 20 b3 (20 .ixd3
exd3 2 1 1Wxd3 1Wxc4 :f) 20 . . . tlJb4 !?,
Black threatens 2 1 . .. tlJxd5 22 cxd5
.ixd4 as well as . . . a4.
3) 7 1Wcl (perhaps this avoids
some tricks based on . . . ttJfxe4, but
it still leaves me pretty confused)
7 . . . e5 8 d5 c6 9 tiJf3 1We8 (9 . . . cxd5
10 cxd5 .id7 followed by . . . 1%.c8 is
suggested by Glek) 1 0 0-0 ttJh5 (or
I 0 . . . c5 ! ?) 1 1 dxc6 ! bxc6 12 l:td 1 f6
1 3 .ie3 f5 ! ? 14 l:txd6 f4 15 .id2 g5 !
and Black had good attacking pros­
pects in return for his pawn, Palat­
nik-Glek, Philadelphia 1990.
4 ) 7 f3 has little independent
significance. After 7 . . . e5 8 d5 1We8
(8 . . . c6) 9 1Wd2 we have transposed
elsewhere in this chapter.
7 ••• c6! ?
In the Pour Pawns Attack proper
such a plan would be considered too

slow but here, with the white bishop
on the double-edged g5-square, other
factors come into play. Black intends
to follow up with . . . liJc7 after which
he will be ready to challenge the
white centre with . . . d5 or embarrass
the bishop on g5 by . . . liJe6.
7 1We8 preparing . . . e5, is an im­
•.• , (Black's position may look passive,
portant alternative, with the foil ow­
ing possibilities:
t ) 8 liJf3 and now:
l a) 8 eS 9 fxe5 (9 dxe5 dxe5 10
•••
liJxe5 liJc5 and the threats of . .. llJe 6
and . . .llkxe4 provide adequate com­
pensation for the pawn) 9 . . . dxe5 10
d5 ( t 0 liJxe5 c5 is good for Black
and t 0 dxe5 liJg4 t t liJd5 liJxe5 t 2
Jl.e7 c6 is a perfectly playable ex­
change sacrifice) I O . . h6 1 1 Jl.xf6
. lent game for Black. The remaining
( t t Jl.h4 liJg4 allows Black to be­
come very active) t t . . . Jl.xf6 t 2 a3
1i'e7 t 3 0-0 :ds t 4 :bt c5 t5 1i'd2
Jl.g7 t 6 :rd t Jl.d7 t 7 b4 1i'd6 with a
rou ghly level position, Mohr-Miles,
Bad Worishofen t 990.
t b) 8 h6 {perhaps an improve­
.•.
ment on the immediate . . . e5) 9 Jl. h 4
es and now:
t bt ) 10 dxeS dxe5 t t liJxe5 liJc5
( t l . .. g5 ! ?) t2 Jl.f3 ( t 2 1i'c2? liJfxe4 !
t 3 liJxe4 Jl.f5 t 4 Jl.d3 liJxd3 + t 5
1Wxd3 Jl.xe5 t 6 fxe5 1Wxe5) t 2 . . . g5
(now t 2 . . . liJcxe4 J 3 liJxe4 liJxe4 t 4
Jl.xe4 f6 i s not playable because of
the weakness of g6) t 3 il.f2 ( 1 3 Jl.g3
g4) t3 ... liJcxe4 ! t4 liJxe4 (or t4 Jl.xe4
gxf4) t 4 . . . liJxe4 t 5 Jl.xe4 gxf4 t 6
Jl.d4 c5 ! t 7 Jl.c3 .i.xe5 with a clear
advantage for Black.
t b2) 1 0 fxeS dxe5 t t Jl.xf6. The
problem for White is that 1 1 dS al­
lows t t .. .liJg4, but after the text he is
The Averbakh 55
playing variation ' l b' with a tempo
less. In Sorin-Ubilava, lbercaj a
1 992, Bl ack decided to use this
tempo to prevent White from playing
c5 without having to play ... c5 him­
sel f. The game continued 1 1 . . . Jl.xf6
1 2 d5 Jl.e7 ! ? 1 3 a3 liJb8 14 b4 b6
but as long as he avoids being steam­
rollered on the q ueenside his bishop
pair and potential kingside attack
will give him a good game) 15 liJb5
il.d8 t6 0-0 a6 17 liJc3 Jl.e7 1 8 �h t
il.d6 t 9 liJe t 1i'e7 20 liJd3 liJd7 2 t
:c t a5 22 1i'b3 Jl.a6 23 Jl.g4 :tbs
24 :r3 ? ! liJf6 25 il.h3 liJh7 ! 26 :f2
h5 27 g3 h4 28 :en hxg3 29 hxg3
liJg5 30 Jl. g4 �g7 with an excel­
moves were 3 t :r.h2 :hs 32 :rf2
liJh7 33 bxa5 bxa5 34 a4 1i'g5 35
il.f3 liJf6 36 :xh8 :xh8+ 37 �g2
Jl.c8 38 liJe2 1i'e3 39 c5 liJxe4 40
cxd6 liJxf2 4 t liJdc t .i.h3+ 42 �g t
liJd3+ 0- t .
2) 8 1i'd2 (White believes that his
first priority is to establish some con­
trol over the dark squares) 8 . . . e5 9
fxe5 dxe5 t 0 d5 liJc5 1 1 1i'e3 liJa4
t 2 liJb5 1i'e7 t 3 0-0-0 a6 t4 d6 cxd6
t 5 liJxd6 1i'c7 t 6 �bt liJc5 t 7 Jl.xf6
Jl. xf6 t 8 liJxc8 :axc8 t9 Jl.g4 llJe6 !
20 Jl.xe6 fxe6 2 t :c t b5, Tuk­
makov-Mortensen, Reykj avik t 990,
and now 22 c5 ! would have been
equal.
8 ltJf3
8 1i'd2 4Jc.7 is likely to trans pose
back in to the main line after 9 liJf3,
but in the game Moskalenko-Nadyr­
khanov, Alushta t 994, Black tried
8 bS! ?:
.. .
56 The Averbakh
1 ) The game continued 9 eS b4
1 0 exf6 bxc3 1 1 bxc3 ( 1 1 ifxc3 exf6
1 2 �h4 lle8 is less good as White
may experience some trouble on the
e-file and his d4-pawn is less secure)
l l . . . exf6 12 �h4 ifa5 1 3 tiJf3 �f5
1 4 0-0 llfe8 with an equal position
according to Belov.
2) Belov al so examines White
accepting the pawn sacrifice, giving
9 cxbS cxb5 1 0 �xb5 llb8 1 1 �e2
ifb6 1 2 llbl ttJc7 1 3 tiJf3 �b7 as
unclear. A possible continuation is
1 4 �xf6 �xf6 15 0-0 ( 1 5 d5 �xc3 !
1 6 ifxc3 f5 looks quite good for
Black) 1 5 . . . ttJe6 16 :tfd 1 ( 1 6 e5
�xf3 ) 16 . . . tiJxd4 ! 1 7 ttJxd4 e5 with
at least equal chances for Black.
8
... ttJc7 (D)
8 bS has also been tried here. Af­
...
ter 9 cxb5 cxb5 1 0 �xb5 llb8 1 1
ife2 tfJc7 1 2 �c4 d5 1 3 exd5 ttJcxd5
1 4 �xf6 ttJxf6 1 5 0-0 tiJh5 Black
had decent positional compensation
for his p awn, Gulko-B arlov, Mont­
real 1 992.
9 ifd2
9 dS prevents Blac k's two main
ideas ( . . . d5 and . . . ttJe6) but allows a
third, 9 . . . tiJh5 ! . Seirawan-Gelfand,
Tilburg 1 990 continued 1 0 f5 ( 1 0
ifd2 f6 l 1 �h4 �h6 12 g3 e5 ! is good
for Black as after 1 3 dxe6 ttJxe6 the
bishop on h4 is in trouble) 10 gxfS?
..•
1 1 exf5 tiJf6 1 2 dxc6 bxc6 1 3 g4 !
llb8 14 ifd2 ltJa6 1 5 tiJd4 with good
prospects for White. Seirawan, how­
ever, considers that 10 cxdS! 1 1
•••
cxd5 tiJf6 ! 1 2 fxg6 hxg6 1 3 ifd2
ttJa6 is very good for Black.
9 � h4 has also been seen. Sorin-
0.Foisor, Olot 1992 continued 9 . . d5 .
1 0 cxd5 cxd5 1 1 e5 ltJe4 1 2 ifb3
�h6 1 3 g3 ( 1 3 �g3 may be better)
1 3 . . . b6 1 4 0-0 �b7 1 5 llad l lbe6 16
tiJe l ttJxc3 17 ifxc3 f6 ! with good
play for Black.
9
..• dS
9 ttJe6 10 �h4 �h6 1 1 g3 li.Jh5
...
is also interesting .
10 �xf6 exf6!
A big improvement on 10 �xf6 ...
1 1 cxd5 cxd5 1 2 e5 �g7 1 3 h4, with
ad vantage to White, Moskalenko­
Nijboer, Wijk aan Zee 1 992.
1 1 0-0 dxe4
12 ltJxe4 � g4! (D)
Black's whole strategy is based on
pressurizing the d-pawn. By elimi­
nating the knight on f3, playing . . . f5
 The A verbakh 57

and a rook to the d-file Black should 17 cxd5

be able to force the pawn to advance
to d5 . He will then hope to blockade
this pawn with a knight on d6, which
in conjunction with his better bishop
and safer king should, at the very
least, compensate for White's central
passed pawn. I am, of course, writing
this with the benefit of hindsight, but
I' m sure that Smirin envisaged all
this when he played 1 o exf6, and
. . .
most probably at home before the
game.
17 i.xd5 llfd8 1 8 llfe l 1Wf6 (but
not 1 8 . . . 1Wd6? 1 9 i.xt7+) 1 9 lle5 ( 1 9
b 3 lLlxd5 20 cxd5 1Wd6) l 9 . . . lLlxd5
( 1 9 . . . 1Wb6) 20 llxd5 llxd5 2 1 cxd5
lld8 is a little better for B lack as hi s
pieces are more active than their
counterparts.
17
•.. 1Wd6
18 g3 tLlb5
19 i.g2 llac8? !
Black starts t o lose the thread
around here. B etter was 1 9 1Wb6, •••

13 llad l 1We7 intending . . . lDd6, as 20 d6 would

14 tLln simply be pushing the pawn to its
This is not an especially good doom.
square for the knight and it might 20 �bl llc4
have been better to play 14 tLlc3. It still wasn't too late for . 1Wb6.
. .

Smirin then gives 1 4 . . . llad8 15 h3 21 lLld3 llfc8

i. xf3 1 6 i. xf3 f5 17 d5 1Wc5+ and 22 llcl h5?
1 8 . . . 1Wxc4, but White should be able This is a serious error which could
to improve on his 1 5th move; 15 llfel have proved costly. Black should
looks logical, upon which Black can have settled for the equal ending that
try 1 5 . . . lLle6. he could have forced by chopping all
It may, in fact, not be essential for the rooks off.
White to move his knight as after 14 23 1We2?
llfel 1Wxe4 1 5 i.d3 the black queen White misses his chance: 23 llxc4
is trapped. Instead Black should play llxc4 24 lLle5 ! llc8 (24 . . . i. xe5 25
14 lladS when White should avoid
.•• 1We2 ! ) 25 Ile 1 would have pushed
1 5 h3 on account of 1 5 . .. i.xh3 ! . Black onto the defensive.
14
.•• i.xf3 23
•. • 1Wc7 !
15 i.xf3 f5 24 llxc4 1Wxc4
Black must, of course, play this 25 llcl? (DJ
before White can play f5 himself. White should have kept his hands
16 d5 off the black queen and played 25 a3
Perhaps White pushed his d-pawn instead, which Smirin assesses as +.
at once to pretend that he was ad­ 25
..• \i'xcl+!
vancing out of choice, rather than ne­ 26 lLlxcl llxcl+
cessity. 16 llfel 1Wd6 followed by a 21 i.n lLld6
r ook to d8 would have been similar Theoretically speaking Black has
to the game. insufficient material for the queen,
16
••• cxd5 but the truth is that his position is
58 The Averbakh

close to winning. The pieces that he about . . . tDc3. Now Black has to
does have co-ordinate beautifu lly back-pedal a little.
and his ki ng is completely safe. 39 libs+ 'it>h7
White, on the other hand, has an ex­ 40 •es 'it>g7!
tremel y exposed ki ng and a bishop Perhaps Black had originally in­
that is virtually irrelevant. The d­ tended 40 i.12+ 4 1 'it>fl tlJxg3+ 42
•.•

pawn also remains firmly blockaded 'it> g2 but this would lose control of
and it is ironic that White would have the position.
more chances without it as then his 41 •es+ 'it>gS
bishop would become a piece again. 42 libs+ i.f8
28 b3 l::t b l 43 i.f3?
29 ..d3 l::t b 2 The only chance was to play 43
30 i.g2 d6!. After 43 . . . ttJxd6 44 i.f3 l::tc2 !
White lets his a-pawn go because 45 i.d5 l::tc 5 Black should stil l win
of the variation 30 a4 tDe4 3 1 _.b5 but his task would be more compli­
i.d4 when he is liable to get mated. cated.
30 ..• l::txa2 (DJ 43 .•. tiJd6
But now Black can combine his 44 1n>6 l::ta3
threats against the king with pushing 4S i.dl lbe4
a passed pawn on the queenside. 46 i.f3 l::t a l+
White's position is hopeless. 47 'it>e2 tiJd6!
31 h4 as Now Black is threatening to pick
32 •e3 bS up the b-pawn.
33 lib6 i.f8 48 'it>d3
34 i.f3 l::taJ 4S 'it>d2 a4 ! decides the issue.
3S 'it>g2 b4 48 ..• l::tc l !
36 i.dl l::ta2+ Black just needs one passed pawn.
37 'it>fi tLJe4 49 •xaS l::tc3+
3S 'it>el i.cS? SO 'it>e2 l::txb3
38 l::ta l would have won on the
.•. S l •a4 l::t b2+
spot as there is nothing to be done S2 'it>d3 b3
 The Ave rbakh 59

53 i. d l lLJe4 h5 b5 ! ? 14 cxb5 i.b7 15 i.h6 \i'xd6 !

and White Resigned as 54 �c4 1 6 \i'g5 (and not 1 6 \i'xd6 i. xh6+)
J:.bl 55 i.xb3 loses to 55 .. .l::t xb3 ! . 16 . . . i.xh6 1 7 \i'xh6 \i'e7 (Nunn pre­
fers l 7 . . . \i'c7) 1 8 \i'g5 lDg7 1 9 i.f3
Gaine 9 e4 20 ltJxe4 i. xe4 2 1 i.xe4 h6 22
V. Milov Gallagher
- \i'e 3 ltJxe4 23 \i'xe4 g5 24 ltJe2 (24
Bad Ragaz 1 994 lDf3 f6 is unclear) 24 . . . a6 ! he prob­
ably wished his king was el sewhere
1 d4 ltJf6 (0- 1 , 49).
2 c4 g6 l a2) Consequentl y, in his next
3 lDc3 i. g7 outing a few rounds later against
4 e4 d6 Nunn, he preferred simpl y 12 Ild l ,

5 i. e2 o.o intending to castle short some time

6 i.gS ltJa6
7 ltjf3
7 h4 is a speciality of Bareev's al­
though it has usually provided him
with a nice round zero in the tourna­
ment chart. Black can respond in
Benoni style or in traditional King 's
Indian fashion:
1 ) 7 cS 8 d5 lDc7 9 \i'd2 e6 (D)
•..
and now:
l a) 10 eS dxeS 1 1 d6 liJce8 was
tried a couple of times by Bareev
during Hastings 1 992/3.
lal) In the first gaine, against Ju­
dit Polgar, he played 12 0-0-0 but af­
ter 1 2 . . . \i'd7 ! (blocking the d-pawn
and unpinning the knight on f6) 1 3
in the future. After 12...\i'd7 ! :
l a2 1 ) That gaine continued 13
\i'e3 (1 3 lDf3 e4 14 lDe5 \i'xd6! 15
\i'xd6 liJxd6 16 Ilxd6 liJe8 1 7 ltJxf7 !
�xf7 1 8 Ild2 i.xc3 19 bxc3 lDf6, in­
tending . . . e5 followed by . . . i.e6 is
about equal according to Nunn)
1 3 . . . b6 14 lDf3 lDh5 ! (Black is more
than happy to part with an exchange
if he can retain his central pawn
mass) 1 5 i.e7 (1 5 ltJxe5 i.xe5 1 6
\i'xe5 f6) 1 5 . . . f6 ! (originally Nunn
intended 1 5 . . . lDf4 but he suddenly
noticed 1 6 \i'xe5 ! ! i. xe5 17 lDxe5
when despite having only one piece
for the queen White has the better
gaine) 1 6 i.xf8 i.xf8 1 7 lDe4 lDf4
1 8 0-0 i. g7 ! (ruling out any tricks
based on ltJxf6+ and lDxe5 ; the d­
pawn can be rounded up later after
Bl ack has consolidated his posi­
tion) 1 9 Ilfe l i.b7 20 i.fl i.h6 2 1
\i'c3 i. xe4 2 2 Ilxe4 ltJxd6 with a
clear advantage to Black (0- 1 , 44).
1 a22) White's l atest try is 1 3 hS.
Zak:harevich-Dolmatov, Kazan 1 995
continued 1 3 . . . b5 ! 1 4 cxb5. i. b7 1 5
i.h6? ! ( 1 5 hxg6 fxg6 1 6 lDf3 is un­
clear) 1 5 . . . i.xg2 1 6 Ilh2 i.d5 1 7
60 The A verbakh
i.xg7 and now best was 1 7 . . . lll xg7 .
Zakharevich gives the following
variation: l S 1i'g5 lllfh5 ! 1 9 i.xh5
(or 19 lllxd5 exd5 20 1i'xe5 :aeS 2 1
1i'xd5 lll f4 + ) 1 9 . . . f6 20 1i'g3 gxh5
2 1 lll x d5 exd5 22 :xd5 1i'b5 with
advantage to Black.
lb) 10 hS is perhaps the most
logical move and in Onishchuk­
Wegner, Berlin 1 993 White obtained
the advantage after 1 0 . . . exd5 1 1
exd5 bS !? 12 cxb5 i.b7 1 3 i.f3 1i'd7
1 4 lll ge2 lll x bS 1 5 1i'f4 lll x c3 16
lllxc3 :res+ 1 7 'iii>f l ii'fS?! lS 1i'xf5
gxf5 1 9 h6 i.hS 20 :h4. Black
should not despair, though, as there
are plenty of possible improvements;
1 1 :es, 14 lll g4!? and 17 :eS ! ?
.•. . .• .••
spring to mind.
2) 7 eS 8 dS h6 9 i. e3 lllcS 10
•..
1i'c 2 ( 10 f3 doesn ' t fit with an early
h4 on account of 1 0 . . . lllh 5, but 1 0
i.f3 was tried out i n Bagonyai-Col­
linson, Balatonbereny 1 992; after
1 0 . . . a5 1 1 g4 c6 1 2 g5 hxg5 1 3 hxg5
lll h7 1 4 1i'd2 cxd5 1 5 cxd5 a4 1 6
i.d l i.d7 1 7 f3 :es 1 S lllh 3 b5 1 9
lll f2 1i'a5 20 llle 2 a draw was agreed
in an unclear position) 10 c6 1 1 hS
.••
( 1 1 b4? lll cxe4 ! ) and now (D) :
2 a) In Bareev- Kasparov, Lina­
res 1 992 Black played 1 1 gS. This
•..
advance is usually a sign that things
have gone wrong for Black, but this
position seems to be an exception.
Firstly, White is not ready to exploit
the weakening of f5 and secondly,
Black does not mind the kingside be­
ing blocked up as his chances on the
queenside are in no way inferior to
White's. It should be mentioned that
the idea of playing . . . g5 in this posi­
tion is not new, but previou sly it was
only played after an exchange on d5 .
I think Kasparov wanted to retain
control over b5 for as long as possi­
ble, as back in 1 9SO he suggested,
after the moves 1 1 . cxdS 1 2 cxd5
•.
g5 1 3 g4 a5 1 4 f3 i.d7, that White
should play 1 5 i.b5 . Let 's return to
Bareev-Kasparov which continued
1 2 f3 ( 1 2 b4 lll c xe4 1 3 lll xe4 lll xe4
1 4 1i'xe4 f5 1 5 1i'c2 cxd5 1 6 cxd5
f4 gives Black a lot of play for the
piece according to Bareev) 1 2 . . . a5
1 3 g4 i.d7 1 4 lll h 3 a4 1 5 1i'd2 cxd5
1 6 cxd5 1i'a5 1 7 lll b l lll fxe4? !
( 1 7 . .. 1i'xd2+ 1 S lll x d2 b5 is a lot
safer) l S fxe4 lll xe4 1 9 1i'xa5 :xa5
20 lllc 3 ! lllg 3 2 1 :g l ltJxe2 22
'iii>xe2 e4 23 :ac 1 ! f5 24 gxf5 :xf5 ?
(24 . . . i.eS 25 lll x g5 hxg5 26 :xg5
'iii> h S 27 lll xe4 i.xb2 is given as
equ al by B areev, but Black doesn' t
look out of the woods yet to me) 25
lll f2 i.eS 26 :h 1 i.b5 + 27 ltJxb5
:xb5 2S :cs+ 'iii>h7 and now instead
of the inaccurate continuation 29
:dl ? :xb2+ 30 :d2 a3 ! , which al­
lowed Black to escape with a draw,
29 :bl :rxd5 30 lll xe4 should be
winning for White.

2b) 1 1 cxdS 1 2 cxd5 .id? ! ? (this
.••
suggestion from S hereshevsky has
become q uite topical recently) 1 3
hxg6 fxg6 1 4 b4 ( 14 .ixh6 .ixh6 1 5
l:xh6 <titg? 1 6 l:h l l:h8 17 l:xh8
'ii xh8, with good play for the pawn,
is the j ustification of Black's 1 2 th
move) 1 4 . . . ltJa6 1 5 a3 h5 1 6 f3 ltJc7
( 1 6 ... ltJh? ? ! 17 ltJb5 ! was favourable
for White in B areev -Gelfand, Biel
IZ 1 993) 17 ltJh 3 ltJh7 1 8 'iid 2 ( 1 8
l:c l ? ! l:t7 1 9 'ii'd 2 .if6 + Raetsky­
Glek, Bad Ragaz 1 994) 1 8 . . ..if6 1 9
0-0-0 was Zakharevich-Poluliakhov,
Azov 1 995 and now, because the
immediate 1 9 . . . a5 is met by 20 b5 ,
Poluliakhov suggests preparing this
advance with 19 'ili'e8.
•••
We now return to the position af­
ter 7 ltJf3 (D).
7 ••• h6!
8 .ih4
Not the only bishop retreat:
1 ) 8 .if4? ! eS ! 9 dxeS ltJhS 10
.ie3 dxeS and now:
l a) 1 1 o ..o c6 1 2 'ili'xd8 l:xd8 1 3
l:fd l l:e8 1 4 g3 ltJf6 1 5 ltJd2 ltJg4
1 6 .ix g4 .i xg4 1 7 f3 .ie6 1 8 b3
ltJ b4 1 9 l:ac l f5 with advantage to
Black, Milov-Smirin, Haifa 1 995 .
The Averbakh 61
l b) 1 1 'ili'cl <tith? 1 2 0-0 ( 1 2 c5 f5
1 3 exf5 gxf5 1 4 'ili'c2 ltJb4 1 5 'ili'b3
ltJd3+ 1 6 .ixd3 'ili'xd3 was good for
Black in Korsunsky-V.lvanov, Mos­
cow 1 992) 1 2 . . . c6 1 3 l:d l 'ili'e7 14
ltJe l ltJf6 15 f3 ltJc5 I 6 ltJc2 ltJe6 1 7
'ili'd2 l:d8 1 8 'ii'e 1 ltJd4 ! 1 9 ltJxd4
exd4 20 l:xd4 (20 .ixd4 l:xd4 2 1
l:xd4 ltJg4) 20 ... :Xd4 2 1 .ixd4 ltJg4!
22 e5 (22 .ixg7 ? 'iic 5+ leads to mate
and 22 'ii'd 2 'ili'd6 is also good for
Black) 22 . . . .ixe5 23 fxg4 .ixd4+
with an excellent game for Black,
Uhlmann-M . Schafer, German Ch
1 99 1 .
2) 8 .ie3 e5 9 0-0 ltJg4 1 0 .ic l
c6 1 1 d5 ! ? f5 1 2 ltJe l ltJf6 1 3 exf5
gxf5 1 4 f4 cxd5 1 5 cxd5 ltJg4 1 6
ltJd3 'ii'b6+ 1 7 <tith l , Farago-Howell,
B ad Wildbad 1 990, and now Farago
gives 1 7 . . . e4 1 8 .ixg4 exd3 1 9 .if3
'ili'd4 as unclear. Although White can
win the d-pawn with 20 ltJb5, Black
will have a strong initiative after
20 . . . 'ili'c5 2 1 'ili'xd3 ltJb4 22 'iie2
i.d? 23 ltJc3 l:fe8.
8
••• es
8 cS!? is tempting as 9 d5 g5 1 0
...
.i g 3 ltJh5 looks like a good Benoni
for Black.
9 0-0
9 dS? ! allows 9 . .. g5 1 0 .ig3
ltJxe4 ! 1 1 ltJxe4 f5 12 ltJfd2 fxe4 1 3
ltJ xe4 .if5 when 14 .id3 g4 1 5 0-0
h5 1 6 f3 .ixe4 17 .ixe4 ltJc5 1 8 .ic2
e4 ! was promising for Black in Ro­
gers-Mortensen, Vejstrup 1 9 89 and
14 f3 g4 15 0-0 'iie 8 1 6 .if2 'ii'g6 17
ltJg3 .ic2 18 'ii'd 2 gxf3 1 9 .ixf3
l:xf3 ! ? 20 gxf3 l:f8 gave Black good
play for the exchange in Pliester­
Reinderman, Wij k aan Zee 1 994,
62 The Averbakh
10 i.xf6! ?
I was quite pleased with the out­
come of the opening and had been
expecting an easy game after some­
thing like 10 dxeS. The text move
completely puzzled me, but I soon
discovered what it was all about as
after...
10 ... i.xf6
My opponent didn't even hesitate
before playing . ..
1 1 cS! ?
My first reaction was one of deep
scepticism that White's idea could be
any good but I was still pretty wary
as I had obviou sly tumbled into a
prepared variation . White intends to
ruin Black's queenside pawn struc­
ture with i.xa6 and then try to prove
that the resulting position i s better
suited to his knights than Black's
bishops.
1 1 ... exd4?!
I arrived at the text, after consider­
able thought, as I felt that the posi­
tion should be opened up for the
bishops and because I underesti­
mated White's 1 4th move. I did well
to avoid 11. .. llJbS? though as after
1 2 lt:Jd5 i.d8 1 3 cxd6 cxd6 1 4 dxe5
dxe5 1 5 Wc l ! White wins a pawn.
Maybe the best move is 1 1 i.g7, •••
e.g. 12 i.xa6 bxa6 13 dxeS dxeS 14
llJdS 1Vd8 15 lt:Jb4 Wxdl 16 Afxdl
fS! (D) and now:
I ) Ryskin-lskusnykh, Azov 1995
continued 17 llJdS ? fxe4 1 8 llkl2
i.g4 ! and White must have realised
only now that he can't move the rook
on accoun t of l 9 ... llad 8 winning
material . Therefore he gave up the
exchange with 19 lt:Jxe4 i.xd 1 20
Axd 1 Afd8 2 1 c6 and punted a draw
offer, which was accepted despite
the fact that Black i s close to win­
ning. The white knights may look at­
tractive but he is caught in a nasty
pin on the d-file. Black should play
2 1 . . �f7, to avoid any checks, fol­
.
lowed by . . .l::tb 8 and ... llb5 . If White
ever supports his knight on d5 with
lt:Jec3 then Black will be able to lib­
erate his bishop with e4. . . .
2) 17 lbd2 is perhaps the critical
line. Although Black can win a pawn
with 17 ...l::tb 8 1 8 a3 a5 l9 llk6 l::txb2
 The Averbakh 63

the position is a real mess after 20

lbc4.
12 ltJdS J.d8
13 J.xa6 c6!?
This was the move I had been
banking on. 13 bxa6 leads to a bad
•••

position after 1 4 cxd6 (or 14 1ixd4

dxc5 15 1Wxc5 as 15 ... 1i'xe4? loses to
1 6 1Wc6 !) 14 . . . c6 15 lbc7 J.xc7 1 6
dxc7 1Wxe4 1 7 1Wxd4 1Wxd4 1 8 lDxd4
when 18 J.d7 is met by 19 l::t ac 1
•••

and 18 J. b7 by 1 9 life 1 .
... able to generate counterplay along
More tempting was 13 dxcS and ••• the b-file or by pushing an a-pawn ;
although White's position is a little and no matter how sick they may be,
awkward after 14 J.d3 c6 1 5 l£lf4 for the moment he still possesses an
J.c7 1 6 1id2 J.g4 I can't believe that extra pawn.
Black has enough for a piece. 18 l::tfcl 1We7
14 1ixd4! 19 ltJeS!? J.e6
This is much stronger than 14 Pinning the knight by 19 1Wd6 •••

l£lf4 1Wxe4 ! 15 1id2 ( 15 l£lxg6 1Wxg6 was also possible although after 20
1 6 J.d3 1Wf6 is good for Black) l::td l 1i'b8 (not 20 . . . 1Wc7 2 1 lbxg6)
15 . . . J.g5 ! with advantage to Black. 2 1 b3 (2 1 1Wf6 1Wxb2 allows another
14 •.• dxcS pin) 2 1 . . . J.e6 22 l::t ac l White retains
I had originally assumed that his edge.
14 cxdS would give me a good
••• 20 1i'e3!
game, but after 1 5 1Wxd5 ! dxc5 1 6 Of course White is not interested
J.xb7 J.e6 (or 1 6 . . . J.xb7 1 7 1i'xb7 in 20 lbxc6 1i'b7 when Black takes
J.f6 1 8 l::t ae l with advantage to over the initiative.
White) 1 7 1Wxc5 l::tb 8 both 18 J.dS 20
••• 1i'd6
and 18 J.c6 J. b6 1 9 1Wc3 are in 2 1 1Wxc5 lixcS
White's favour. The only other real 22 l::txcS l::tab8
alternative is 14 bxa6 but this just
••• 23 b3 l::tbS
transposes to the game after 15 lDf6+ 24 l::ta cl?
J.xf6 16 1Wxf6 dxc5 . A serious mistake which changes
15 l£lf6+! J.xf6 the whole complexion of the game.
16 1Wxf6 bxa6 White should have played 24 l£ld3 as
17 1if4 <j;b7 (D) 24 . . . l::td 8 can be met by 25 l::tc3 ! .
The tactical phase of the game 24
..• l::td8! (D)
has ended and White has emerged Suddenly the black rooks have
with the advantage. Black's queen­ sprung to life. I felt quite relieved
side is a wreck but his game is not round about here as if I was going to
completely hopeless as he may be lose it would at least be without the
64 The A verbak.h

32 ..• .ic4
33 h5 a3
34 hxg6?
34 �h2 would have avoided the
next note.
34 •• . fxg6?
I didn' t have the time to work out
the consequences of 34 .:al+ 35
•..

�h2 a2 but later analysis showed

suffering I had envisaged a few
moves ago. In fact Black is no longer
worse as , due to the weakness o f
his back rank, White can ' t prevent a
rook from penetrating into his posi­
tion.
25 .:xb5
This must have hurt White but
there is no real choice as after 25 f4
.:xc5 26 .:xc5 .:d 1 + White's queen­
side will drop off.
that White is unable to defend, for
example 36 11a8+ �g7 37 gxt7 �xt7 !
(37 . . . .ixf7 3 8 ltJf5 + �h7 39 .:a7
.:d 1 40 .:xn+ �h8 4 1 .:a7 draws)
38 lbc6 (38 ltJf5 .:d l !) 38 . . . .ib3 !
(3 8 . . ..:d l 39 lbe5+ �e7 40 .:a7+
followed by lbxc4 draws) 3 9 ltJd 4
.ia4 ! and Black queens the a-pawn.
Perpetual check can always be
averted by marching the king to the
queenside.
35 �h2! .tn
Now on 35 .:al White has 36
•.•

25 ... cxb5 lbc2 .

26 f4 .:d2 36 �g3 .ixg2
27 .:c7 �g7 ? ! 36 ... .:xg2+ 37 �f3 a2 was much
27 �g8 would have been more
..• too risky. White can choose between
accurate, but time trouble was now 38 lllxb5, 38 e5 and 38 f5.
upon us. 37 lbxb5
28 ltJfJ ! .:xa2 White sensibly aims for the draw.
29 ltJd4 �f8 The alternative 37 e5 .id5 3 8 e6
30 .:xa7 a5 .:g2+ 39 �h3 .:d2 40 e7+ �f7 4 1
31 h4 lbxb5 .ie6+ 42 �g3 a2 is favour­
Although Black will be able to able for Black.
create a strong o utside passed pawn 37 ... .ixe4
White is not without his chances on 38 llk3 .:g2+
the kingside. 39 �h3 .if3
31 •.• a4 40 .:xa3 .:b2
32 b4?! 41 b5 g5
I had expected 32 bxa4 .:xa4 33 42 fxg5 hxg5
lbxb5 .:xe4 34 g3 with an inevitable 43 �g3 g4
draw. The text is an extremely risky 44 �f4 �e7
winning attempt.
 The Averbakh 65

Game 1 0 would also be fatal) 22 . . . lLJh3 23

Petursson - Grivas 'ii' h4 lLJxf2 ! 24 @xf2 i.b6 25 @e l
Katerini 1 993 i. xe3 with an overwhelming posi­
tion for Black, Uhlmann-J.Polgar,
1 d4 lLJf6 Aruba 1 992.
2 c4 g6 2) 10 0-0 lLJc5 1 1 i. xf6 i. xf6 1 2
3 lLJc3 i. g7 lLld5 ( 1 2 b4 lLJe6 1 3 lLld5 i.g7 with
4 e4 d6 . . . c6 to foil ow is also quite good for
5 i. e2 0-0 Black) 1 2 . . . i.d8 1 3 'ii'e 3 llkl7 (Pol­
6 i.g5 lLJa6 gar would no doubt have played
7 ti'd2 e5 (D) 1 3 . . . lDe6 here as 14 lLJxe5 c6 15 lLJc3
i.g5 looks very awkward for White)
1 4 c5 a5 1 5 lLJd2 c6 1 6 lLJc3 b5 ! 17
cxb6 i.xb6 1 8 ti'h6 lLJf6 and Black's
bishop pair eventually made them­
selves felt in Uhlmann-Podzielny,
Bundesliga 1 992.
8.•• 'iVe8
It doesn't take a genius to work
out the point behind this move - the
knight on f6 is unpinned so that it
can get out of the way of the f-pawn .
The queen is, i n fact, not s o badly
8 d5 placed on e8; in some variations it
8 lLJf3 is well met by 8 ...'ii'e8 ! . can assist the advance . ..b5 whilst in
Now 9 d5 transposes to line ' 1 ' i n the others it may spring out on the king­
note to White's 9th move and the side. One thing that Black has to
only other sensible way of meeting watch out for is an annoying lLJb5 .
the threat 9 . . . exd4, viz. 9 dxeS, al­ 8 c6 is a major alternative and al­
..•

lows Black a very comfortable game.
After 9 dxeS there are a couple of
.••
examples:
1 ) 10 l:ldl lLJc5 1 1 i.xf6 i.xf6 12
lLld5 i.d8 13 'iVe3 lLJe6 ! ? ( l 3 . . . lLJd7
is also quite playable) 14 lLlxe5 c6 1 5
lLJc3 i.b6 1 6 'ii'h 6 i.c7 ! 1 7 lLJg4 f5
1 8 exf5 lLJf4 1 9 fxg6 hxg6 20 lLJe3
l:lf7 2 1 'ii'g5 i.f5 ! ( White has two
extra pawns bu t his two most im­
portant pieces are very unhappily
placed) 22 g3 (this loses, but perhaps
it is already too late; 22 0-0 'ii'e 5 !
though it's not my main recommen­
dation (it was a very close call but
finally I felt that 8 . . . 'ii'e 8 was more
energetic) here is a summary of the
current state of affairs :
1 ) 9 i. d3 (on 9 i.d l , 9 . . . lLJc5 1 0
i.c2 transposes but 9 . . . cxd5 1 0 cxd5
b5 ! ? gives active play) 9 . . . lLJc5 1 0
i.c2 a5 1 1 lLlge2 transposes to Game
8, line '2' in the note to White's 7th
move .
2) 9 i.f3 cxd5 1 0 ltlxd5 (a speci­
ality of Farago although his results
66 The Averbakh
have not been very encouraging; 1 0
cxd5 .id? 1 1 lLlge2 b5 would hand
Black the initiative) 1 0 . . . &5 1 1
lLlxf6+ .ixf6 1 2 .ixf6 1i'xf6 1 3 lLle2
b6 1 4 b3 .ib7 1 5 lLlc3 lLle6 1 6 0-0
lLld4 1 7 .ie2 'ii h4 1h- 1h Farago­
Groszpeter, B udapest Elekes mem
1 99 3 .
3) 9 lLlf3 lLlc5 1 0 .ixf6 1i'xf6 1 1
b4 lLla6 1 2 a3 c5 ! (a typical idea in
many lines with the knight on a6) 1 3
llbl 1i'e7 1 4 0-0 f5 1 5 lLle l, Bareev­
G .Kuzmin, US SR Ch (Leningrad)
1 990 and now Bareev gives 1 5 . . . fxe4
1 6 lLlxe4 .if5 1 7 .if3 b6 1 8 lLld3
llac8 1 9 bxc5 .ixe4 20 .ixe4 lLlxc5
as equal.
4) 9 f3 cxd5 10 cxd5 .id7 (D)
and now there are several possibili­
ties for White:
4a) 1 1 .ixa6 bxa6 1 2 lLlge2 1i'b6
( 1 2 ... llb8 1 3 .ie3 lib? 14 0-0 lLle8 is
an alternative) 1 3 .ie3 1i'b7 1 4 0-0
lLle8 ( 14 . . .lLlh5 ! ?) 1 5 llac l ( 1 5 b3 f5
16 exf5 gxf5 1 7 .ih6 .ixf6 1 8 1i'xh6
llf6 is equal according to Dolmatov)
1 5 . . . f5 1 6 exf5 gxf5 17 f4 lLlf6 1 8 h3
lLlh5 ! was fine for Black in the game
Yermolinsky -Dolmatov, Groningen
1993.
4b) 11 .ib5 (a positionally justi­
fiable exchange but it wastes time)
1 1 . .. .ixb5 ( 1 1 . . . 'iia5 ! ?) l 2 lLlxb5 1i'b6
1 3 lLlc3 lLlc5 (threatening . . .1i'xb2)
1 4 b3? ! (best is 14 lld l when Dol­
matov suggests 1 4 . . . a5 1 5 lLlge2 a4)
14 . . . lLlh5 1 5 .ie3 lLlf4 16 g3, Peturs­
son-Dolmatov, Lucerne W cht 1993 ,
and now 16 . . . llac8 � 1 7 lld l ! ( 17
gxf4? exf4 1 8 .id4 lLld3 + 1 9 'iixd 3
"ti'xd4 is good for Black) 17 ... lLlfd3+
1 8 'iiif 1 'iia6 1 9 'iit g 2 b5 is unclear
according to Dolmatov.
4c) 1 1 g4 h6 ( l 1 . .. 1i'a5 1 2 h4
transposes into the next note, but
1 2 lLlh3 ! is quite good for White,
for example 12 . . . llfc8 13 lLlf2 lLlc5
1 4 llbl 1*'b4 1 5 lLlb5 ! ? 1i'xd2 + 1 6
'iit xd2, Petursson-Kotronias, Reyk­
javik 1 992, with a typical Averbakh
ending slightly in White 's favour) 1 2
.ie3 ( 1 2 .ixh6 lLlxe4 is a trick that
should be familiar to all King's In­
dian players) 1 2 . . . h5 1 3 h3 lLlc5 1 4
0-0-0 1i'b8 1 5 'iii b l, Alterman-Xie
Jun, Cap d ' Agde 1 994, and now Al­
terman gives 1 5 . . . b5 ! 1 6 b4 lLla4 1 7
lLlxa4 bxa4 1 8 'iit a l llc8 as = .
4d) 1 1 h4 1i'a5 12 g4 ( 1 2 lLlh3
can be met by 12 . . . lLlh5) 12...h5! 13
.ixf6 .ixf6 14 gxh5 'iit g7 15 hxg6
fxg6 (D) is another typical motif that
should be familiar to all King's In­
dian fans as this sort of sacrifice can
occur in several variations . For his
pawn Black has obtained a relatively
secure king position and chances for
active play on both the dark squares
and the queenside .
loseliani-Gallagher, Biel 1 990
now continued 16 h5 llh8 17 h6+
'iit h 7 1 8 lLlh3 .ixh3 1 9 llxh3 llac8

20 �fl ll:lc5 2 1 :bl 1i'dS 22 �g2
J.g5 23 1i'el ltJd7 ! (perhaps the white
king has a magnet attached to it) 24
1i'g3 J.f4 25 1i'g4 :es ! 26 �h l lLlf6
27 1i'g2 :gs and Black, with several
tasty outposts on the kingside, had
more than enough compensation for
the pawn.
Ioseliani was clearly impressed as
a year later she was to be found on
the black side; Gaprindashvili-Iosel­
iani, Tbilisi 1 99 1 went instead 16
J.xa6 bxa6 1 7 1i'g2 :hs I S h5 :h6
19 ll:lge2 :bs 20 0-0-0 1i'b4 2 1 :d2
:gs 22 :c2 �f7 23 hxg6+ :gxg6
24 1i'f 1 1i'b6 with roughly equal
chances, although 1 -0, 5 1 .
5 ) 9 h4 cxd5 1 0 cxd5 J.d7 (or
1 0. . .1i'a5 !? 1 1 f3 lLlh5 1 2 g4 ll:lg3 1 3
:h3 ll:lxe2 1 4 ll:lgxe2 b5 1 5 a3 f6 1 6
J.e3 h5 17 :g3 hxg4 I S fxg4 J.d7
1 9 h5 g5 20 ll:lc l 112- 1/2 Serper-Ye
Jiangchu an, Jakarta 1 994) 1 1 f3 ( 1 1
J. xa6 bxa6 1 2 h5 1i'a5 1 3 ll:lge2
:abS 14 f3 1i'b6 ! was good for Black
in Glek-Moskalenko, Odessa 1 9S9)
1 1 . . . 1i'a5 an d we have transposed to
variation '4d' .
Let us now return to the position
after S . . . 1i'eS ( D) .
9 J.dl
The Averbakh 67
a /
/
This funny-looking move has, to
date, been White's most popular
choice . The bishop vacates e2 for
the king 's knight and heads for c2
from where it will have far more in­
fluence on events (remember that
Black nearly always plays ... f5 ). The
main drawback to this pl an is that it
is very time-consuming but White
hopes that with the centre blocked
this will not prove too serious. There
are a whole host of alternatives,
some of them with similar ideas to
the text whilst others are less subtle:
1 ) 9 ll:lf3 lLlhS ! 10 g3 fS with a
couple of examples:
l a) 11 exfS gxf5 12 J.h6 f4 1 3
J.xg7 , Gulko-Djurhuus, Manila OL
1 992, and now instead of 13 �xgT .•.
14 :g t ! which allowed White to de­
velop some initiative, Black should
play 13 ll:l xgT with an equal game
•••
(Gulko).
l b) l l lLlh4 f4 1 2 g4 lLlf6 1 3 f3
h5 ! 1 4 g xh5 lLlh7 15 ll:lxg6 ( 1 5 :g t
ll:lxg5 1 6 :xg5 1i'e7) 15 . . . ll:lxg5 1 6
ll:lxfS �xf8 1 7 0-0-0 1i'xh5 and
Black eventually converted his ma­
terial advantage in Tisdall-W.Wat­
son, Oslo 1 99 1 . These variations
illustrate why White normally tries
68 The A verbakh
to develop his king's knight to e2
rather than f3 .
2) 9 0-0-0 (D) when Black has:
2a) 9 ltJcS 10 f3 ltJh5 1 1 b4 ( 1 1
•..
.ih6 a5 1 2 .ixg7 <j;xg7 1 3 .id3 f5
1 4 .ic2 fxe4 1 5 lbxe4 lbxe4 1 6
.ixe4 ltJf6 1 7 .ic2 b5 ! was good for
Black in Jasnikowski-Piket, Novi
S ad OL 1 9 90) 1 1 . . . ltJa6 1 2 a3 ltJf4
1 3 .ifl f6 1 4 .ih4 .ih6 1 5 <j;b2 f5
1 6 1i'c2 fxe4 1 7 ltJxe4 .if5 1 8 .if2
lbb8 112- 1'2 Raetsky-Golubev, Biel
1 994. A bit more of this game would
have been helpful but it looks OK for
Black.
2b) 9 ltJhS ! ? (this seems play­
•..
able even when White hasn't wasted
a tempo on 9 .id 1 or 9 .if3) 10 .ixh5
gxh5 1 1 .ih6 f6 !? ( l 1 . .. f5 looks more
natural) 1 2 .ixg7 <j;xg7 1 3 f4 <j;b8
1 4 ltJf3 ltJc5 1 5 f5 ( 1 5 llhfl !? has
been suggested) 1 5 . . . llg8 16 1i'h6
1i'f7 1 7 b4 ltJd7 1 8 lbe2 a5 1 9 b5
lbc5 20 ltJg3 llxg3 ! ? 2 1 hxg3 ltJxe4
22 1i'xh5 1i'xh5 23 llxh5 lbxg3 24
llh3 lbe2+ ! 25 <j;d2 li)f4 26 llh6
<j;g7 27 lldh l .ixf5 28 g3 ltJd3 29
g4 .ie4 30 <j;e3 lbc5 and Black had
good compensation for the exchange,
S .Ivanov-Kovalev, Minsk 1 995.
3 ) 9 h4 ltJcS (9 . . . ltJh5 is sug ­
gested by Burgess but there seems to
be no harm in waiting to collect the
w bite p awn that is destined to arri ve
on h5) 10 1i'c2 ( 1 0 f3 ltJh5 is good
for Black while 1 0 .i xf6 .ixf6 1 1
ltJf3 a5 1 2 0-0-0 .ig7 1 3 h5 f5 1 4
hxg6 1i'xg6 was about level in Piket­
J . Polgar, Aruba 1 995) 1 0 aS 1 1 hS
••.
(consistent; nobody has tried 1 1
0-0-0) 1 1 ltJxhS 12 .ixhS gxhS and
..•
now:
3 a) 13 ltJbS ? ! f5 ! 1 4 ltJxc7 1i'g6
1 5 ltJh3 fxe4 1 6 .te3 1i'xg2 1 7 0-0-0
.ixh3 1 8 ltJxa8 ltJd3 + 1 9 <j;bl .ig4
20 lldgl 1i'f3 21 ltJc7 ltJxf2 22 .ixf2
e3 23 .ixe3 .if5 24 .ih6 .i xc2+
25 <j;xc2 1i'e4+ 0- 1 was a real ham­
mering for White in K watschewsky­
Gross, Balatonbereny 1995.
3 b) 13 .ie3 ltJa6 ( 1 3 . . . b6 !?) 14
ltJge2 rs 15 f3 1i'g6 16 llh2 occurred
in Kakhiani-Kovalev, Helsinki 1 992
and although White eventually won
the game I can ' t believe that Black
stands badly here.
3bl ) Knaak gives the variation
16 .ih6 1 7 .ixh6 ! 1i'xh6 1 8 1i'd2
.••
1i'g6 ( 1 8 . . . 1i'xd2+ 1 9 <j;xd2 fxe4 20
lbxe4 b5 ! ? is also suggested by
Knaak who continues 2 1 cxb5 ltJb4
22 ltJ2c3 .i b7 23 llxh5 lbxd5 24
llah 1 ;t) 1 9 0-0-0 with attacking
chances for White, but 1 9 . . . ltJc5 20
lldh l llt7 (20 . . . fxe4 2 1 llxh5 might
be good for White) 2 1 llxh5 llg7
looks fine for Black to me.
3 b2) 16 h4! ?, trying to use the
•••
extra pawn is another idea, e.g. 1 7
0-0-0 f4 1 8 .i f2 h3 1 9 llg l (or 1 9
gxh3 1i'h5 ) 1 9 . . . 1i'h5 with an ex­
tremely playable position for Black.

4) 9 f3 lllh S 10 .t d l ( I O g4 is re­
served for greedy masochists; after
1 0 . . . lll f4 1 1 .txf4 exf4 1 2 'ii'x f4 f5
White was already struggling in the
game B uckley-Fishbein, Philadel­
phia 1 99 1 ) 10 fS 1 1 lll ge2 (D) and
.•.
now Black has several possible con­
tinuations:
4a) 1 1 ..id7 1 2 .tc2 lllb4 ! ? (seiz­
••
ing the chance to gain some space on
the queenside) 1 3 .tbl a5 14 a3 llla6
1 5 b3 f4 ( 1 5 . . . lll c5 , maintaining the
tension, also comes into considera­
tion) 1 6 .tc2 .tf6 1 7 .txf6 llxf6
( 1 7 . . . lll x f6 is more natural and if
White continues as in the game
Black saves several tempi; perhaps
he was concerned about 1 8 g3) 1 8
· 0-0-0 lllc5 1 9 �b2 b5 ! (otherwise
Black will be pushed back by b4) 20
cxb5 .txb5 2 1 lll xb5 'ii' xb5 22 lllc 3
. 'ii' b7 2 3 �a2 llff8, followed by trans­
ferring the knight on h5 round to the
queenside, gave roughly equal play
in S . lvanov-Ryskin, St Petersburg
1994.
4b) 11 lilcS 1 2 b4 fxe4 !? 13 bxc5
•.•
( 1 3 fxe4, perhaps) 1 3 . . . exf3 14 gxf3
llxf3 with unfathomable complica­
tions. The game Kriszany-Czebe,
The Averbakh 69
Budapest 1 993 continued 15 .tc2 e4
1 6 0-0-0 e3 1 7 'ii'e l 'ii'e5 1 8 h4 lllg 3
19 llld4 llf2 20 llg l .t f5 2 1 .txe3
lll e2+ ! 22 lllc xe2 'ii' xe3 + 23 lld2
lle8 24 lll x f5 gxf5 25 .td3 f4 26 h5
�h8 27 h6 .tf6 and White was com­
pletely tied down (0- 1 , 37).
4c) ll fxe4 12 lll xe4 ( 1 2 fxe4
•.•
'ii'f7 1 3 .te3 lllf4 1 4 lll g 3 h5 )
1 2 . . . lll f4 1 3 0-0 'ii' f7 1 4 .tc2 h6 1 5
.th4 g5 1 6 .t f2 is given as unclear
by Sokol in.
4d) 11 'ii'f7 12 .tc2 f4 intend­
••.
ing . . . .tf6 was suggested by Knaak.
5) 9 .tf3 h5 ! ? (9 . . . lllh5 1 0 .txh5
transposes to the main game) 1 0 h3
(standard continuations such as 1 0
lll ge2 lllh7 1 1 .th6 f5 were not very
appealing to White, but it's hard to
believe that this is an improvement)
1 0 . . . lllh 7 1 1 g4 lllxg5 1 2 'ii'xg5 f6 1 3
'ii'd2 ( 1 3 'ii'h4 hxg4 1 4 hxg4 �f7 ! is
mentioned by Glek) 13 . . .h4 ! 14 .tg2
f5 1 5 lll f3 ( 1 5 g xf5 gxf5 1 6 exf5
e4 ! and 1 6 lll f3 fxe4 followed by
l 7 . . . llf4 are not better) 1 5 . . . fxg4 1 6
lllxh4 llf4 ! with very active play for
Black, F.Portisch-Glek, St lngbert
1 99 1 .
6) 9 .td3 has hardly been seen
here, presumably because of 9 . . . lllh5
(although 9 . . . lllg 4 and 9 . . . llld 7 may
also be worth investigating). The one
example I 've seen, Petursson-Nunn,
London Lloyds Bank 1 994, actually
continued 9 . . . .td7 1 0 lll g e2 lllc5 1 1
.tc2 a5 1 2 f3 transposing to line '2'
below.
9 ••• lllhS !? (D)
I have a vivid memory o f the first
time I saw such a knight move. Play­
ing through the games of the 1 972
70 The Averbakh
World Championship match I was
astounded by Fischer's l l . . . lbh5 ! in
Game 3 . Even after all the fancy
ex planations I had great difficulty
in coming to terms with the move.
Twenty years on I have finally
grasped what it's all about and I am
even recommending such a course.
Of course after 10 .ixh5 gxh5 Black's
kingside pawns are shattered, but as
compensation for this he will have
gained the bishop pair and an open
g-file for his maj or pieces to oper­
ate on. Even so, the text may not be
to everyone' s taste so let's take a
look at the alternative, 9 lbcS. Af­
.•.
ter 10 .ic2 as 1 1 lb ge2 ( 1 1 lbb5?
lbfxe4 1 2 .ixe4 lbxe4 1 3 'ii'e 3 lbxg5
1 4 lbxc7 'ii'd 8 1 5 lbxa8 f5 was tre­
mendous for Black in Amura-C.Foi­
sor, S ubotica worn IZ 1 99 1 ) there is
(D):
1 ) 1 1 lbhS 1 2 lbb5 ! 'ii'd7 1 3
..•
0-0-0 b6 1 4 f3 a4 and now 1 5 <iii> b l
followed by lbec3 would give White
an edge . Instead Seirawan-Piket,
Wijk aan Zee 1 99 1 continued 15
g4? ! lbf4 1 6 lbxf4 exf4 1 7 .ixf4
.ia6 1 8 'ii' b4 ? ( 1 8 lba3 was better)
18 . . . .ixb5 1 9 'ii'xb5 'ii'e7 20 .id2 a3
2 1 b4 'ii'e5 ! with a decisive attack for
Black. Yet another example of how
dangerous it is for White to open the
long diagonal in the King's Indian.
2) 1 1 . . .id7 12 fJ w ith a further
.
branch:
2a) 12 hS, intending . . . lbh7 and
•..
. . . f5 , is well met by the prophylactic
1 3 .ie3 ! . Petursson-Djurhuus, Gaus­
dal 1 995 now continued 13 . . . lbh7 1 4
0-0-0 b6 ( 1 4 . . . f5 ? 15 .ixc5 dxc5 1 6
d6 c6 17 lba4) 1 5 h3 h4 1 6 g3 ! 'ii'e7
1 7 .:tdg 1 a4 1 8 f4 with a clear plus
for White.
2b) 12 <iii>hS is an alternative way
•.•
of preparing . . . f5 . A fter 1 3 0-0 lbg8
1 4 .:tae l f6 1 5 .ie3 f5 1 6 exf5 gxf5
17 <iii> h l b6 1 8 f4 e4 1 9 g4 lbh6 20
gxf5 lbxf5 2 1 .ixc5 e3 ! 22 .ixe3
lbxe3 23 .:tg 1 lbxc2 24 'ii'x c2 .if5
Black possessed a powerful bishop
pair as compensation for the pawn,
Tisdall-Manninen, Gausdal 1 99 1 .
2c) 12 bS is also quite playable.
.••
After 1 3 cxb5 .ixb5 1 4 lbxb5 'ii'x b5
White can claim, at most, a small
edge.
10 .ixhS
10 f3 would be pretty bizarre in
this particular position but would in
fact transpose to line '4' in the note
to White's 9th move.
 The Averbakh 71

10 •.• gxhS lS f3 b6!

1 1 lLJge2 f6! With the simple idea of capturing
An important improvement over on e4 and playing . . . lLJc5 .
the game Petursson-Glek, Belgrade 16 exfS i.xrs
1 988, which went 1 1 ... rs 1 2 exf5 17 lLJg3 i.d7 ! ?
i.xf5 1 3 lLJg3 e4 ( 1 3 .. .'ti'g6 1 4 lLJxf5 17 lLJcS l 8 lLJxf5 'ti'xf5 1 9 0-0 a5
•..

l:txf5 15 i.e3 'ti'xg2 1 6 0-0-0 is good was good enough for equality.
for White) and now White should 1 8 0-0?
have played 14 i.h6! lDc5 15 0-0 a5 The white king would have been
1 6 i.xg7 �xg7 1 7 f3 ! with some ad­ better off on the queenside according
vantage as Black's king is very ex­ to Grivas.
posed . 18 .. • h4?
12 i.h6 Black misses a golden opportu­
One of the main points behind nity. After 1 8 :r4! Grivas gives 1 9
...

l l . . . f6 is that after 1 2 i.e3 Black lLJge4 l:taf8 20 �h i lLJc5 2 1 l:tfe l h4

doesn ' t play . . . f5 at once, but first 22 h3 i.f5 with advantage to Black.
l 2 . . . h4 ! in order to prevent lLJg3 . 12 19 lLJge4 ltJcs
i.h4 is also not recommended on ac- 20 'ti'gS ! 'ti'xgS
count of 1 2 . . . 'ti'g6. 21 lLJx gS as
12 ... i. xh6 22 l:tadl l:tf4
13 'ti'xh6 'ti'g6 (D) The game is level. The remaining
Black 's opening problems are al­ moves were 23 lLJge4 l:taf8 24 h3
ready history. lLJxe4 25 fxe4 �g7 26 l:txf4 l:txf4 27
:n l:txfl + 28 �xfl �f6 29 �f2
�g5 30 �e3 i.e8 3 1 b3 i.d7 32
lLJd 1 i.e8 33 lLJf2 i.d7 34 lLJd3 i.e8
35 b4 axb4 36 lLJxb4 �f6 37 lLJa6 c5
38 dxc6 i.xc6 39 lLJb4 i.b7 40 a4
�g5 4 1 �f3 �f6 42 lLJc2 i.c8 43
�e3 i.e6 44 lLJa3 �e7 45 �d3 �d7
46 lLJbl �c6 47 lLJd2 �c5 48 �c3
Ji.. f7 49 lLJb3+ �c6 5 0 a5 bxa5 5 1
lLJxa5+ �b6 52 lDb3 i.g6 53 �d3
�c6 54 lLJd2 �c5 55 lLJf3 i.f7 56
lDxh4 i. xc4+ 57 �e3 d5 58 exd5
14 'ti'd.2 rs lf2- 1h
4 Wh ite plays i.g5

The first two games of this chapter is much less popular and much
are concerned with the Smyslov Sys­ sharper than the Smyslov System.
tem in which White plays i.g5 and
follows up with the solid e3. One of Game 1 1
his principal ideas is to limit the ac­ Dely - Haik
tivity of the King's Indian bishop but France 1970
if Black does manage to prise open
the long diagonal then the absence of 1 d4 lt:Jf6
White's dark-squared bishop from 2 c4 g6
the queenside may be keenly felt. 3 ltJc3 i.g7
Therefore my main suggestion is for 4 liJf3
Black to attack the centre with . . . c5, 4 i. gS is almost certain to trans­
and this is the su bject of Game 1 2 . pose to lines considered later.
Playing for ... e5 is , as ever, an impor­ 4 •.. d6
tant option but I feel that here it plays If Black had played 4 . . 0-0 instead
.

into White's hands by increasing the then there is another interesting plan
relevance of the bishop on g5 . In against 5 i.g5, based on playing . . . c5
fact, there doesn' t seem to be any and then . . . d5 . This runs 4 0-0 5 •••

clear path to equality for Black after i.gS cS (D) and now White can sup­
playing . . . e5 , which, along with a port or block the centre (6 i.xf6
couple of promising sidelines, is the i.xf6 7 lt:Je4 ti'b6 =):
subject of Game 1 1 . A word about
chasing the bishop from g5 : it almost
always makes sense for Black to play
. . . h6 but he should be wary about fol­
lowing up with . . . g5 which involves
a much more serious weakening of
the kingside. A good rule is that . . . g5
should only be played when there is
a concrete follow-up in mind, such
as gaining the bishop pair with
. . . lt:Jh5 xg3, or removing the threat
against e7 so that . . . ti'b6 can be
played. I ) 6 e3 cxd4 7 exd4 (7 lt:Jxd4
The final game of the chapter ti'a5 !) 7 dS! ?. Normally one would
•..

deals with i.g5 followed by e4, which be hard-pressed to find similarities


between the King's Indian and the
Caro-Kann but, amazingly enough,
this position is actually classified in
the ECO code under B 14. The Panov
Attack move-order goes 1 e4 c6 2 d4
d5 3 exd5 cxd5 4 c4 llJf6 5 llk3 g6 6
.ig5 .i g7 7 llJf3 0-0, arriving at the
same position as after 7 . . . d5 . White
now has:
l a) 8 cxdS (rare, but perhaps
only because White feels he should
be trying to refute the pawn sacrifice)
8 . . . llJxd5 9 'ilb3 llJxc3 (9 ... lbb6 ! ?)
1 0 bxc3 llJc6? ! (I would prefer
I 0 . . . 'ifc7 to this unwieldy move) 1 1
.ie2 b6 1 2 0-0 'ild6 1 3 l:l.ad l e6 1 4
llJd2 lLJa5 1 5 'ilb4 ! 'ifc7 ( 1 5 . . . 'ii'xb4
1 6 cxb4 llJc6 1 7 .if3 .ib7 1 8 llk4 is
promising for White) I 6 lbe4 h6 1 7
'ife7 'ifxe7 1 8 .ixe7 l:l.e8 1 9 llJf6+
.ixf6 20 .ixf6 with advantage to
White, Smyslov-Martinovic, Gron­
ingen 1 989/90.
l b) 8 .ixf6 .ixf6 with a further
branch:
l bl ) 9 cxdS .i g4 (9 . . . e6 should
probably be met by 1 0 .ic4 accord­
ing to S .Pedersen and Ca.Hansen) 1 0
.ic4 ( l 0 .ie2 llJ<l7 1 1 0-0 .ixf3 1 2
.ixf3 'ilb6 1 3 'ifa4 tlxb2 1 4 tlxd7
tlxc3 1 5 tlxb7 11z. 11z Pekarek-Zu­
ger, Prague 1 989) I 0 . . . 'ilb6 (ECO
gives 1 0 . . . llJd7 1 1 0-0 .ixf3 1 2
'ifxf3 .ixd4 1 3 tle4 .if6 1 4 l:l.ad 1 ;1;
S myslov-Taimanov, USSR 197 1 ) 1 1
.i b3 llJd7 1 2 0-0 .i xf3 1 3 'ifxf3
.ixd4 14 l:l.ad l Uac8 15 :t'el .ixc3 ! ?
1 6 bxc3 'ild6 1 7 tle3 l:l.fe8 1 8 tlxa7
l:l.xc3 1 9 'if xb7 l:l.c7 20 'il b5 l:l.ec8 2 1
'ife2 lLJc.s 22 'ife3 (22 .ic2) 22 ... lbxb3
23 axb3 l:l.b8 24 'ilg3 'ii'x g3 112_11z
Sadler-Nunn, Oviedo rpd 1 992.
White plays .ig5 73
Black's positional compensation
meant that he was never really in
danger.
I b2) 9 llJxdS (to the uninitiated
this may seem like a freebie but the
weakness of d4 means that it will be
impossible for White to retain his
extra pawn without making serious
positional concessions) 9 ..ig7 (D)
••
and now White has tried:
I b2 1 ) 10 llJc3 .ig4 1 1 .ie2 ( 1 1
d5 'ifa5 1 2 'ifc2 .ixf3 1 3 gxf3 llJd7
14 .ie2 l:l.ac8 1 5 0-0 llJb6 1 6 'ifb3
.ixc3 1 7 bxc3 l:l.c7 18 l:l.fd 1 l:l.fc8 :t=
is similar, Moore-Burgess, Frome
1 99 1 ) 1 1 . . . llJc6 1 2 d5 .ixf3 1 3 .ixf3
llJa5 ( 1 3 . .. l2Je5 ! ?) 14 .ie2, as in Pach­
man-Andersson, Geneva 1 977, and
now 1 4 . . . .ixc3+ ! 1 5 bxc3 'ifc7 1 6
tla4 l:l.fc 8 enables Black to regain
his pawn with some advantage.
l b22) 10 .ie2 llJc6 1 1 0-0 1id6
1 2 llJc3 llJxd4 1 3 llJxd4 'ifxd4 1 4
'ii' b 3 with rough equality, Kristins­
son-H.Olafsson, Reykjavik 1 984.
l b23) 10 llJe3 lLJc6 (ECO gives
10 . . .'ifa5+ 1 1 'ild2 'ifxd2+ 12 �xd2
l:l.d8 1 3 .id3 llJc6 1 4 llJc2 .i g4 1 5
d5 .ixf3 1 6 gxf3 l2Je 5 1 7 .ie2 e6 1 8
lbe 3 = Lputian-Gufeld, USSR 1 98 1 ;
74 White plays 1'.g5
1 8 . . . 1'.h6 wins the pawn back) 1 1
d5? (better is 1 1 llJc2, which Griin­
berg assesses as = after l l . . . 1'.g4)
l 1 . . . 1fa5 + ! 1 2 ttJ<l2 (White probably
intended 1 2 'ifd2, spotting too late
1 2 . . .1'.xb2 ! ) 1 2 .. . 1'.xb2 ! 1 3 l:tbl ( 1 3
dxc6 1'.xa l 1 4 'ifxa l :d8 1 5 'ifb2
bxc6 1 6 1'.e2 :b8 ! 1 7 'ifc2 :b7 is a
variation provided by Grunberg;
White , s d a ys are clearly numbered)
1 3 . . . 1'.c3 1 4 dxc6 :d8 1 5 llJd5
1'.xd2+ 1 6 1f xd2 1fxd2+ 1 7 �xd2
bxc6 and Black eventually converted
his extra pawn in Knaak-H.Grtin­
berg , E.German Ch 1 989.
2) 6 d5 (in general White seems
reluctant to play d5 but Ehl vest is ob­
viously an exception) 6 ... d6 7 lLJd2
h6 8 1'.h4 and now:
2a) 8 g5 9 1'.g3 llJh5 1 0 e3
. ..
llJxg3 1 1 hxg3 f5 ( 1 1 . . .e6 is an alter­
native) 1 2 1'.d3 llJd7 1 3 llJf3 e6 1 4
dxe6 llJb6 1 5 g4 ! 1'.xe6 1 6 gxf5
1'.xf5 1 7 1'.xf5 l:txf5 1 8 'ifc2 'iff8 1 9
llJe4 ! , Ehlvest-Pugachev, S t Peters­
burg 1 994. Here we have a conflict
of opinions as Ehlvest believes that
White has a clear advantage and
Glatman assesses the position as =
after 1 9 . . . d5 20 llJg3 :n; somehow,
I don' t think he took 2 1 llJxg5 ! into
account.
2b) Kasparov preferred to take
the Benko option and his game with
Ehl vest continued 8 a6 9 e4 b5 10
...
1'.e2 (White declines the offer as h4
is not where the dark-squared bishop
belongs in a Benko) 1 0 . . . b4 l l llJa4
llJh7 1 2 0-0 ttJ<l7 1 3 'ifc2 g5 1 4 1'.g3
lLJe5 1 5 :ae I a5 1 6 llJf3 llJg6 1 7
e5 ! ? g4 1 8 lLJh4 llJxe5 ( 1 8 . . . llJxh4 1 9
1'.xh4 1'.xe5 20 1'.d3 with the double
threat of llJxc5 and 1'.xh7 +) 1 9 llJf5
1'.xf5 20 1fxf5 1fc8 2 1 1fxc8 :axc8
22 1'. xe5 dxe5 23 1'.xg4 f5 24 1'.d 1
llJg5 ! (on d6 the knight will feel like
an octopus) 25 f3 lLJf.7 26 1'.c2 lLJd6
27 b3 e4 ! , keeping the knight out of
the game, gave Black slightly the
better of a draw i n Ehlvest-Kaspa­
rov, Horgen 1 995 .
5 1'.g5 0-0
It is surprising that 5 c6! ? is not
.•.
tried more often as after the virtually
automatic 6 e3 Black has 6 . . . 1fa5 !
(intending 7 . . . llJe4 ), which seems to
equalize at once . Schmidt-Hug, De­
brecen Echt 1 992 continued 7 'ifd2
(nobody has played 7 1'.d3 , when
7 ... 1'.f5 looks sensible and 7 . . . 1'.g4 8
1'.h4 { 8 1'.f4 ? e5 ! } 8 . . . 1fh5 interest­
ing) 7 . . . 1'.g4 8 1'.e2 (8 b4 ! ? 1fxb4 9
:bl 1fa5 1 0 :xb7 llJbd7 ! is unclear
according to S . Pedersen) 8 . . . 1'.xf3 9
1'.xf6 1'.xf6 1 0 1'.xf3 llJd7 1 1 0-0 0-0
= 1 2 :fd l :fd8 1 3 :abl 1fc7 1 4 b4
a5 1 5 b5 llJb6 1 6 'ifd3 c5 1 7 ttJ<l5
llJxd5 1 8 1'.xd5 'ifb6 1 9 dxc5 dxc5
20 1fe4 :ab8 2 1 'iff3 �g7 22 a4
1/z- 1/z . Perhaps this is not the sort of
game that the majority of King's In­
dian players are after but ifs worth
bearing in mind if you only need a
draw.
If this line catches on then White
players will probably start looking
for an alternative 6th move; 6 h3 and
6 e4 spring to mind. Here are a cou­
ple of suggestions, off the top of my
head, which you shouldn ' t take too
seriously. Against 6 h3 Black could
try 6. . . 1fa5 7 'ifd2 b5 ! ? whilst on 6
e4 chasing the bishop doesn't look a
bad idea, e.g. 6. . . h6 7 1'.h4 g5 (Black

could also throw in 7 . . . 'ii'a5) S i.g3
lLlh5 and White usually prefers his
pawn back on e3 in this type of posi­
tion.
6 e3
The position after 6 e4 (more
commonly reached via the move or­
der 1 d4 lLlf6 2 c4 g6 3 lLlc3 i.g7 4 e4
d6 5 lLlf3 0-0 6 i.g5) was once fa­
voured by East German grandmas­
ters Uhlmann and Malich, but when
they dropped it from their repertoires
the line virtually ceased to exist. It
is, I suppose, an Averbakh without
the flexibility. The generally recom­
mended course for Black is to follow
the game Uhlmann-Fischer, Havana
OL 1 966: 6 . . . h6 7 i.h4 g5 (7 . . . lLla6 ! ?
i s worth thinking about as S i.e2 e5
9 d5 g5 1 0 i.g3 lLlxe4 transposes to a
line considered in Game 9) S i.g3
lLlh5 9 i.e2 e6 (preparing . . . f5) 10 d5
( 1 0 0-0 llJc6 1 1 d5 CiJe7 1 2 l:.c l lLlr4
was unclear in Malich-A.Rodriguez,
Halle 1 976) 1 0 . . . f5 l l lLld4 lLlxg3 1 2
hxg3 fxe4 1 3 lLlxe6 i.xe6 1 4 dxe6
i.xc3 + ! ? ( 1 4 . . . 'ii' f6 is more solid ;
ECO gives 1 5 lLlxe4 'ii'xe6 1 6 'ii'd5
:es 1 7 f3 llJc6 l S 0-0-0 =) 1 5 bxc3
'ii'f6 1 6 e7 ! :es ( 1 6 . . . 'ii'x f2+ 1 7 �d2
:es l S :xh6 leaves the black king
too exposed) 1 7 :bl lLla6 l S 'ii'd4
�g7 1 9 :xb7 :xe7 20 'ii'xf6+ �xf6
2 1 :xh6+ �g7 22 :h5 �g6 with
equality according to ECO.
6 • .• lLlbd7 ( D)
My main recommendation, 6 cS, •••
is the subject of the next game.
7 'ii'c 2!
Accurate and unpleasant to meet
(at least in my experience). With
Black 's sixth move signalling his
White plays i.g5 75
intention to play . . . e5 White 's first
priority is to put a rook on the d-file,
after which the pressure in the centre
will force Black to make a conces­
sion such as . . . :es .
7 i. e2 would allow Black a much
easier ride. Ruban-J .Polgar, Gronin­
gen 1 993 continued 7 . . . e5 S 0-0 h6 9
i.h4 g5 1 0 i.g3 lLlh5 1 1 dxe5 lLlxg3
1 2 hxg3 dxe5 1 3 'ii'c2 f5 14 :ad l c6
1 5 lLld2 h5 ! ? 1 6 e4 (of course 1 6
i.xh5 is met by 1 6 . . . g4) 1 6 . . . f4 1 7
gxf4 gxf4 l S i.xh5 ( a hot pawn but
Black would have a strong attack
anyway) 1 S . . . 'ii' h4 1 9 i.f3 lLlf6 20
'ii' b 3? ! �hS 2 1 :re 1 lLlg4 22 i. xg4
i. xg4 23 f3 :adS ! 24 lLlfl (24 'ii'c2
would have put up more resistance)
24 . . . i.xf3 ! 25 gxf3 :gs 26 lLlh2
i. fS + 27 �h l i.c5 2S :e2 :d7 ! 29
lLla4 i.f2 0- 1 (the threat of . . . 'ii'xh2+
is decisive).
7 ... es
7 c6 is also possible when play is
••.
quite likely to transpose back into
the main line . However Black does
have the option of playing a q uick
. . . a6 and . . . b5 , e.g. S :d l h6 (or
S . . . a6 9 i.e2 b5 1 0 a3 i.b7 1 1 0-0
'ii'c7 1 2 lLld2 llacS 1 3 i.h4 � Inkiov­
Gallagher, Toulouse 1 993) 9 i. h4
76 White plays i.g 5

'ifa5 (9 . . . 'ife8, intending . . . e5 , is 12 h3!?

worth consideration) 1 0 i.e2 a6 1 1 W hite takes precautions against
0-0 b5 1 2 a3 bxc4 1 3 i.xc4 llJb6 14 ... g5 and . . . llJh5 .
i.e2 llJbd5 1 5 llJxd5 cxd5 16 b4 12 ... exd4
'if b6 1 7 :c 1 i.d7 , Inkiov-Soltis, 12 llJfS could be met by 1 3 i.g3
..•

Moscow 1 989, is given as = by ECO, with ideas of c5 .

although I would have thought that 13 llJxd4 llJc5
Black may still have some work to 14 i.f3 (D)
do after 1 8 'ifc7 .
8 :dt h6
Black usually flicks in . . . h6 at
some point as it is very useful to have
the option of . . . g5 .
9 i.h4 :es
As previously mentioned this can
be considered at least a minor con­
cession as Black would prefer to
keep his rook on f8 to support the
thematic advance . . . f5 . But the threat
to the e-pawn has to be dealt with
and 9 'ife7 10 llJd5 and 9 'ife8 1 0
.•• ..• This sort of position often gives
llJb5 are both out o f the question, Black dynamic possibilities in the
whilst 9 ... exd4 1 0 llJxd4 is at least an King's Indian but here his d-pawn is
edge for White . That leaves 9 ... g5, especially vulnerable.
but Black should be wary about 14 .•. llJe6
playing such a move when he has 1 5 i.g3 llJd7
castled and White hasn't. After 1 0 1 5 llJg5 fails to 16 c5 ! .
. ••

i.g3 llJh5 1 1 dxe5 llJxg3 1 2 hxg3 16 llJxe6

dxe5 1 3 i.d3 Black has problems on 16 liJdb5! ? is unnecessary.
the light squares. 16 ... :xe6
1 0 i.e2 c6 17 i.g4
1 1 0-0 'ifc7 Forcing Black to weaken his king­
1 1 ...'ifa5 transposes to Smyslov­ side.
Westerinen, Hastings 1 972 wher e 17 f5
White obtained the better game after 18 i.e2 llJe5
1 2 dxe5 dxe5 1 3 llJd2 llJfS 14 a3 19 e4!
i.f5 1 5 'ifc l g5 1 6 i.g3 llJg6 1 7 b4 And now White opens the posi-
'ifc7 1 8 f3 :ad8 19 c5 llJd5 20 llkie4 tion for his better developed pieces.
llJxc3 2 1 'ifxc3 ! whilst l 1 'ife7 1 2
.•. 19 .•. :es
b4 'iff8 1 3 b5 llJh7 14 bxc6 bxc6 15 20 f4 llJf7
d5 was good for White in Bosboom­ 2 1 i.d3 'ifa5
Nijboer, Dutch Ch 1 991 . 22 Ji.el !? fxe4
 White plays i.g5 77

23 lbxe4 11b6+
24 1l.f2! i.f5 (D)
White was not concerned about
24 i.d4 as even 25 i.c3 (25 �h 1 is
•••

also very good) 25 . . . i. xf2+ 26 'ti'xf2

'ti' xf2+ 27 � x f2 g i v es tremendous
compensation for the exchange.

1 ) 7 dxc5. Not mentioned in any

sources I've seen and obviously not a
critical test but it may be of concern
i f you 're facing a muc h weaker op­
ponent. After 7 . . . dxc5 8 'ti'xd8 1l.xd8
9 i.xf6 i.xf6 1 0 lbd5 , lO llJc6 is
...

very comfortable for Black whilst

10 i. xb2 ! ? 1 1 1l.bl i.f6 1 2 llJc7
•.•

25 g4 ! i.xe4 i.c3+ 1 3 �e2 i.f5 1 4 1l.d l llJc6 1 5

26 i.xe4 llJxa8 1l.xa8 looks like good value
Now Black's kingside is going to · for an exchange, and you may find
fall apart. something even better in this line .
26 ••• i.d4 2) 7 d5 is another plausible move
27 i.xg6 i.xf2+ never mentioned . Black could try
28 i.xf2 'ti'c7 7 h6 8 i.h4 1i'b6, hoping to reach
...

29 i.b7+ <iii>f8 similar positions to the main line,

30 'ti'g6 1-0 whilst the Benko option (7 ... b5 ! ? or ,

perhaps 7 . . . 1i'a5 followed by . .. a6

Game 1 2 and ... b5) is also worth considering .
Pachman Smyslov
- 3 ) 7 h3 provides a haven for the
Amsterdam 1994 bishop on h2, thereby preventing the
. . . h6, . . . g5, and . . . llJh5 idea. Black
1 d4 llJf6 has:
2 c4 g6 3a) 7 i.f5!? 8 g4 i.e4 9 i.g2
...

3 llJcJ i. g7 cxd4 1 0 exd4 llJc6 1 1 i.e3 i.xf3 1 2

4 llJfJ 0-0 i.xf3 lD<l7 1 3 0-0 e5 1 4 dxe5 l2Jdxe5
5 i.g5 d6 1 5 i.d5 llJe7 1 6 i.g5 112-112 Hort­
6 e3 c5 (D) Kindermann, Munich 1 99 1 .
7 i.e2 3b) 7 llJc6 8 d 5 llJa5 9 lD<l2 a6
...

Other moves are rare. They in­ 1 0 a3 b5 1 1 cxb5 axb5 1 2 i.xb5 i.d7
clude: 1 3 i.e2 h6 1 4 i. f4 1l.b8 1 5 1l.bl 'ti'c7
78 White plays Ji..g5

1 6 0- 0 c4 with compensation for We shall just examine one of the

B lack, Ku ligowski-Hawelko, Po­ alternatives, 8 g5, which leads to a
•••

lanic a Zdroj 1 984, is an example sharp struggle in which Black weak­

from ECO. ens his kingside in order to obtain
7 ... h6 the bishop pair (or, perhaps more
8 Ji..h4 (D) precisely, to force White to open the
long diagonal). Even if you intend to
play 8 . . . Ji.. f5 it could be worth taking
a quick look at the variations below
as the resulting type of position may
arise in many openings.
After 8... g5 9 Ji.. g 3 %5 10 dxc5
(White was not successful with 1 0
ec2 g4 1 1 lbh4 cxd4 1 2 exd4 l2Jc6
1 3 d5 llJd4 14 ti'd2 e5 1 5 dxe6 Ji.. xe6
+ S myslov-Tal, US S R 1 973, nor
with 1 0 0-0 lbc6 1 1 d5 lba5 1 2 :c 1
a6 1 3 lLxi2 lbxg3 1 4 fxg3 ? ! e6 +

8 ••• Ji.. f5! Wexler-Fischer, Mar del Plata 1 959)

Black has tried numerous moves 10 ... lbxg3 1 1 hxg3 dxc5 1 2 ti'c2
in this position, but my attention was ( swapping queens would diminish
drawn to the modest-looking text the importance of Black 's weakened
when I spotted that it was the Smys­ ki ngside while, on the other hand,
lov's choice when somebody had the the bi shop pair would remain an
cheek to play his own system against extremely relevant factor) 12 ...e6
him. There are at least a cou ple of (White was threatening 1 3 lbxg5 )
go od reasons for putting the bishop there is (D) :
on f 5. The first is to introduce the
possibility of . . . ll)e4. Exchanging his
knight on f6 for White 's on c3 is al­
most always a good deal for Black as
with all the excess baggage removed
from the long diagonal the full force
of the King's I ndian bishop is likely
to be felt. The second reason for put­
ting the bishop on f5 is that it covers
bl and in a surprising number of
variations this allows Black to mount
a decisive assault against the b2-
pawn, which in this variation is more 1 ) 13 :tdl ti'e7 14 l1Jdl and now
likely to have a coating of sugar than Black can choose between one sharp
the usual arsenic. and one solid move:

l a) 14 rs 1 5 g4 (the drawback
..•
of playing . . . f5 too early is revealed
as White is able to launch an attack
on the light squares; nevertheless,
Black's position remains viable)
15 . . . f4 1 6 'ife4 (Bagirov later gave
1 6 exf4 l:txf4 1 7 g3 as � but this is
debatable � after 1 7 . . . l:.f7 both 1 8
1We4 lbd7 19 i.d3 lbf6 and 1 8 lbde4
lbc6 1 9 lbd6 llkt4 20 l:.xd4 cxd4 2 1
lbxf7 'if xf7 22 lbe4 i.d7, intending
. . . i.c6, look fine for Black) 16 .. .fxe3
1 7 fxe3 lbd7 1 8 i.d3 :n 1 9 lbf3
lbf6 20 'i6g6 lbxg4 2 1 lbe4 l:txf3 !
(avoiding 2 1 . . . lbxe3 ? 22 lbfxg5 !
lbxg2+ 23 'iii>d2 hxg5 24 lbf6+ ! 'ifxf6
25 l:th8+ ! and mate) 22 gxf3 llle5 23
'ifh5 i.d7 with fully adequ ate com­
pensation for the exchange, Bagirov­
Kelecevic, Sarajevo 1980.
l b) 14...lbc6 15 0-0 l:tb8 16 a3 a6
1 7 g4 ( 1 7 lbde4 f5 1 8 lbd6 l:td8
poses no problems) l 7 . . .i.d7 18 llkie4
lbe5 l 9 lbg3 i.c6 20 ltJce4 (20 ttJh5
lbg6 2 1 lbxg7 'iii> x g7 is also fine for
Black; note that he is willing to part
with either bishop if White has to
waste several tempi in collecting it)
20 . . .lbg6 2 l lbh5 f5 22 gxf5 exf5 23
lLJc3 i.xc3 ! 24 'ifxc3 l:tfd8 25 l:txd8+
l:.xd8 26 l:td 1 :xd 1 + 27 i.xd 1 'li'e5
28 'ifxe5 lbxe5 29 i.e2 'iii>f7 with a
fractionally ad vantageou s ending
for Black, Alekseev -S chekachev,
Moscow 199 1 .
2) 13 g4 ltJc6 14 a3 l:tb8 15 lbd2
'ife7 1 6 0-0-0 ? ! (playing with fire;
16 0-0 would be similar to line ' l b' )
16 . . . a6 1 7 l:.h5 b5 ! (the race is on) 1 8
cxb5 ( 1 8 l:.dh l b4 1 9 l:.xh6 f5 ! )
1 8 . . . axb5 1 9 lbce4 c4 20 lbxg5 f5 !
2 1 gxf5 (there 's no time to retreat)
White plays i.g5 79
2 1 . . . hxg5 22 fxe6 i.xe6 23 g4 (23
'ifh7 + 'iii> f7 24 lbe4 l:th8 ! 25 lbd6+
{ 25 lbxg5 + 'ifxg5 ! } 25 . . .'iii> f8 and
Black wins) 23 . . . c3 24 'ifh7+ 'iii>f7 25
lbe4 cxb2+ 26 'iii> b 1 l:.h8 27 lbd6+
'iii> f8 28 'ife4 lbe5 29 f4 l:txh5 30
gxh5 lbf7 0- 1 S .Pedersen-Gadjily,
Duisburg jr Ech 1 992.
9 0-0
9 h3 has been pl ayed but after
9 . . . lbe4 White is simply a tempo
down on the next note.
I haven' t seen any games with 9
i.d3 but perhaps White should al­
ready be thinking of how to maintain
the balance with a move such as this .
9
..• lbbd7! ?
Preparing . . . 'li'b6, which is not
playable at once in view of 1 0 i.xf6
followed by lbd5 . Black could of
course play . . . g5 , but as the game
progresses you will discover his rea­
son for not doing so. 9 lbc6? is a
•••
mistake as after 1 0 d5 lbb4 1 1 a3
lba6 1 2 lbd2 i.d7 1 3 e4 e5 1 4 l:tbl
b6 15 b4 Black found himself very
passively placed in Haik-Sax, Bag­
neux 198 1 .
9 lbe4 is an important alterna­
•.•
tive, though. After 10 lbxe4 i.xe4
( D) there are a couple of examples:
1 ) 1 1 'iid2 g5 12 i. g3 'ifb6 1 3
l:tfd 1 lbc6 (White is already lost) 14
l:.ac 1 (after 14 d5 llkt8 White can kiss
goodbye to his b-pawn) 14 . . .l:tad 8 ? !
(perhaps a touch too sadistic; Black
could have cashed in at once with
the same mini-combination that he
played on his next move) 15 b3 ( 1 5
d 5 would have saved the pawn al­
though Black would still be much
better) 15 ... i. xf3 1 6 i.xf3 cxd4 1 7
80 White plays i&.g5
i&.xc6 dxe3 ! and Black soon won,
Skare-Westerinen, Gausdal 1 992.
2) 1 1 ltJd2 i&.f5 1 2 i&.f3 ltJc6 1 3
i&.xc6 ( 1 3 liJb3 cxd4 1 4 exd4 g5 1 5
i&.g3 'ii b6 is very good for Black)
1 3 . . . bxc6 1 4 e4 i&.e6 ( 14 . . . i&.c8 ! ? 1 5
dxc5 i&.xb2 1 6 :bl i&.g7 is at least an
edge for Black) 1 5 d5 i&.d7 1 6 'fic2
llb8 1 7 :abl cxd5 1 8 cxd5 f5 + Le­
bel-Sharif, French League 1 992 .
I think that we can conclude that
White is badly in need of an im­
provement against 9 . . . ltJe4.
10 dS?!
Alternatively :
1 ) 10 'fid2 ltJe4 gives comfort­
able equality, whilst 10 'fib6!? 1 1
.•.
ltJd5 ? ! ltJxd5 1 2 cxd5 :te8, with
ideas of .. i&.e4, also deserves consid­
eration .
2) 10 ltJd2 (threatening 1 1 g4)
10 . . . g5 1 1 i&.g3 i&.g6 12 ltJb3 :cs 1 3
:c 1 ltJe4 1 4 ltJxe4 i&.xe4 1 5 i&.d3
liJf6 1 6 f4 i&.xd3 17 'fixd3 ltJh5 1 8
i&.e l e6 1 9 'fie2 ltJf6 20 fxg5 hxg5 2 1
dxc5 dxc5 ( 2 1 . . . ltJe4 ! ? ) 22 e4 ltJd7
23 h4 gxh4 24 'fih5 ltJe5 (24 . . .'fic7 ! ,
in tending to centralise the black
queen, is preferable) 25 :d l 'fie8 26
i&.c3 with some advantage to White,
Shrentzel-Enoshi, Tel-Aviv 1 98 8 .
However, Black 's play can certainly
be improved upon . 1 1 . .. i&.g6 and
1 2 . . . llc8, for example, were pretty
listless moves just when Black
should have been looking to create
concrete threats or to increase the
activity of his pieces. 1 1 cxd4 1 2 •••
exd4 'fib6 ( 1 2 . . . e5 ! ?) 1 3 ltJb3 a5 ! ,
emphasising the fact that that the
knight on b3 has simply journeyed
from one insecure home to another,
would have been a more dynamic re­
action. If White now lashes out with
1 4 f4 then Black should probably
play 1 4 gxf4, rather than 14...g4 1 5
..•
i&.f2 ! .
3 ) 10 ll cl (probably best) and
now :
3a) Zangiev-Nadyrkhanov, Kras­
nodar 1 995 continued 10 gS 1 1 i&.g3
•..
ltJh5? ! ( l l . . .ltJe4 should be fine for
Black) 12 ltJxg5 ltJxg3 1 3 fxg3 hxg5
1 4 :xf5 e6 1 5 .:n ( 15 :f3 ! ?)
15 ...cxd4 1 6 exd4 'iib6 1 7 �h l 'fixd4
1 8 'fic2 ! with an edge for White.
3b) 10 'fib6 would now be met
...
by l l b3 .
3c) Nadyrkhanov's suggestion
10 lLJe4 1 l ltJxe4 i&.xe4 12 ltJd2 g5 !
.•.
( l 2 . . . i&.f5 1 3 e4) deserves closer ex­
amination.
10 . .. '11b 6!
No time is wasted in attacking the
most sensitive spot in the enemy
camp, whilst the trap 10 ltJe4? 1 1
..•
ltJxe4 i&.xe4 1 2 ttJd2 i&.f5 1 3 e4, is
avoided.
1 1 tt:Ja4
Ugly, but 1 1 'fid2 g5 12 i&.g3 lbe4
1 3 ltJxe4 i&.xe4 and 1 1 'fi b3 g5 1 2
i&.g3 ltJe4 1 3 ltJxe4 i&.xe4 1 4 ltJ<l2
i&.g6 both seem to lose a pawn.
 White plays .tg5 81

11 fi'a5 17 fi'xb4
12 tiJd2 tlJb6 There is no choice for White as 1 7
13 tlJc3 (D) fi'xe7 fi'xb2 loses material.
13 tlJxb6 fi'xb6 again leaves the 17 ... cxb4
b-pawn in difficulties. Now White 18 tlJb5 ttJxd5 !
hopes that his problems can be 19 .tf3 .td3 !
solved by advancing e4, but Smys­ 20 .txd5 .txb5
Jov was ready for that one. 21 .txa8 .txn
22 .te4 .ta6
0- 1
Not surprisingly, Pachman denied
S myslov the opportunity to exercise
his legendary technique. A very ele­
gant game full of neat tactical ideas.
The moral of the story : don' t play the
S myslov S ystem against Smyslov,
even a 73-year old Smyslov.

Game 1 3
Spassky Fischer
-

13 1Wb4! Sveti Stefan/Belgrade ( 1 6) 1 992

14 fi'b3
14 e4 fi'xb2 ! 1 5 l%c l .td7 looks 1 d4 ltJf6
like a relatively safe pawn. 1 6 f4 2 c4 g6
would probably be met by 1 6 . . . ttJh7. 3 ttJc3 .tg7
Perhaps this line was still the lesser 4 e4 d6
evil for White as Black's next move 5 .tg5 (D)
was quite devastating.
14 .•. ttJbxd5 !
And now we know why SmysJov
dido ' t succumb to any urge he may
have felt to push his g-pawn. The
exposed position of the bishop on h4
is the key point in this simple, but
pleasing combination.
15 cxd5
Or 15 l"Dxd5 tlJxd5 16 fi'xb4 tlJxb4
1 7 i. xe7 l:[fe8 1 8 i. xd6 l%ad8 - + .
15 ... fi'xh4
16 fi'xb7 1Wb4! A slightly Jess respectable sys­
Black's last few moves illustrate tem than the one considered in the
weH the power of the queen. previous two games.
82 White plays i.g5
s .
. . h6
It makes sense to put the question
to the bishop before committing one­
self in the centre.
6 i.h4
Alternative retreats for the bishop
are:
1 ) 6 i.f4 ltJc6 7 d5 e5 S i.e3 and
here Black has the choice between
8 ltJe7 and 8.. ltJd4!? .
... . move like 2 1 i.bl Black can simply
2) 6 i. e3 ltJg4 7 i. c 1 e5 (perhaps
7 . . . c5 ! ?) S d5 (S dxe5 should of
course be met by S . . . ltJxe5) S . . . f5 9
i.e2 ltJf6 10 exf5 gxf5 1 1 i.h5+ ( 1 1
f4 ? 0-0 1 2 ltJf3 ltJe4 ! 1 3 fxe5 ltJxc3
1 4 bxc3 dxe5 1 5 0-0, Petrosian­
Torre, Tilburg 1 9S2, and now ac­
cording to Petrosian 1 5 . . . c5 ! would
have given Black a clear advantage)
1 1 . . . ltJxh5 1 2 flxh5+ �fS 1 3 ltJge2
fies 14 ltJg3 ltJa6 1 5 0-0 i.d7 1 6 b4
�g s 1 7 ltbl �h7 I S ltJb5 l%f8 1 9
fixes ltaxeS 20 ltJxa7 i.a4 ! 2 1 i.d2
i.c2 22 libe l i.d3 23 ltfe l e4 112- 112
I.Sokolov-Ivanchuk, Linares 1 995.
I wonder what Senor Rentero had to
say.
6•.• cs
Delaying this advance would give
White the chance to play f4.
7 dS
7 dxcS flaS 8 i.d3 fixes (better
than S . . . dxc5? ! 9 f4) 9 ltJge2 ? ! (9 f3
would allow White to keep his im­
portant bishop; even so, after 9 . . . ltJc6
1 0 ltJge2 ltJe5 1 1 i.f2 fla5 Black
would have a comfortable Maroczy
Bind) 9 .. gS 10 i.g3 ltJhs 1 1 :ct
. 10 i.xe4
ltJc6 12 a3 lLJxg3 (a strong case
could be made out for continuing
1 2 . . . a5) 13 ltJxg3 i.eS! ? 14 b4 flb6
and now:
1 ) Bakic-Mozetic, Yugoslavia
1 992 continued l S ltJfS? ! i.xc3 + !
1 6 :xc3 i.xf5 17 exf5 fld4 I S fld2
a5 with advantage to B lack who has
the far superior minor piece.
2) lS ltJdS is better and Balde as­
sesses the position after 1 5 . . . fldS 1 6
0-0 e6 1 7 ltJe3 a5 l S b5 ltJe7 l 9 ltJg4
i.d4 20 ltJh5 ltJg6 as unclear. After a
support his bishop with 2 l . . . flb6,
not fearing 22 liJh(g)f6+ �e7 .
7 ..• gS ! ?
7 e6 is another idea, whereas
•..
7 flaS S i.d3 g5 9 i.g3 transposes
••.
to the main line and this was in fact
the move order employed in the
Stein-Geller game given below.
8 i.g3 flaS
9 i.d3 (D)
9 fld2 ltJh5 would not disturb
Black too much.
9 ... ltJxe4!
This combination is the justifica-
tion of Black's play.
i.xc3+
1 1 bxc3 flxc3+
12 �n fS!
Black now wins back one of the
white bishops and should remain a

pawn up. In such a sharp position,
though, material is only of secon­
dary importance.
13 .tlcl
This was suggested as an im­
provement over 13 ll:Je2, which led
to a crushing victory for Black in
Stein-Geller, Moscow 1 966. Play
continued l 3 . . . 1i'f6 14 i.c2 f4 1 5 h4
:rs ! 1 6 hxg5 hx.g5 17 ll:Jxf4? (based
on an oversight; 1 7 i.h2 f3 ! is also
very bad while Geller offers l 7 1i'e l
as the only chance, w hen he gives
Black the choice bet w een 1 7 .. .fxg3
and 17 . . . ll:Jd7 !? 1 8 i.h2 ll:Je5 1 9 f3
.tlh8) 1 7 ... g xf4 1 8 i.h2 (S tein had
intended 1 8 i.h4 but at the last mo­
ment noticed 1 8 . . . .tlh8 !) 1 8 . . . ll:Jd7 1 9
g3 ll:Je5 20 1i'h5+ �d8 2 1 gxf4 ll:Jg4
22 .tlel .:r.h8 23 i.h7 1i'g7 and White
resigned as after 24 i.g3 ll:Jf6 Black
wins a piece.
13 ..• 1i'f6 (D)
14 h4
The point behind 1 3 .tlc 1 was to
give White the option of 1i'h5+ with­
out then having to worry about a
hanging rook on al . Ftacnik suggests
14 1i'h5+ here, considering the posi­
tion after 14 ... �d8 1 5 h4 g4 ! 1 6 i.d3
White plays i.g 5 83
f4 1 7 i.xf4 1i'xf4 to be unclear. I pre­
fer Black, not so much because he is
a pawn up but because I think his
king has better long-term prospects
on the queenside than White's on the
kingside. A possible continuation is
1 8 ll:Je2 1i'f6 1 9 ll:Jg3 :rs 20 ll:Je4
1i'f4 with advantage to Black.
14 •.. g4!
Geller only considered 14 ....fxe4,
when 1 5 1i'h5+ favours White. Many
people criticised Fischer for his anti­
quated openings in this match, but if
you had a whole stack of novelties
gathering dust on the shelf after a
twenty year lay-off, I'm sure you
would also be trying to get them in
when you made your comeback.
15 i.d3
Ftacnik suggests 15 ll:Je2 fxe4 1 6
�gl i.f5 17 �h2 lLJd7 1 8 .tlel . This
may be an improvement over the text
but White certainly hasn' t enough
compensation for two pawns. An­
other idea, suggested by Polugaev­
sky, is 15 i.c2. After 1 5 . . . f4 1 6 i.a4+
�d8 1 7 ll:Je2 fxg3 1 8 ll:Jxg 3 he was
of the opinion that White had good
play for the pawn, but I'm also scep­
tical about this. Black should start
with 1 8 .. ..tlf8 and follow up, a la Fis­
cher, with a quick . . .ll:Jd7-e5.
15 ... f4
16 ll:Je2
16 i.h2 g3 1 7 1i'h5+ �d8 i s good
for Black.
16 ••. fxg3
17 ll:Jxg 3 :rs
18 :c2 ll:Jd7 ?
Black is more than happy to give
back his extra pawn if it involves the
rapid development of his queenside
84 White plays �g5

and increased attacking chances on mention his rook on h 1 and leave it

the kingside via the open g-file. at that.
19 •xg4 tt:Jes 23 f4 l2Jxc4!
20 •e4 24 tllh S •n
An important point is that 20 Spassky obviously saw this but
•hs+ �d8 doesn' t help White as he probably fell that without his c-pawn
then has no good way to deal with (which blocked in his bishop) he
the threat to his bishop. For example, might be able to generate some swin­
2 1 •e2 �g4 ! 22 •e4 tll xd3 when dling chances.
23 •xd3 loses to 23 . . ...a l + and 23 25 •xc4 •xhS
•xg4 to 23 ... tll x f2. 26 ltb2 ltg3!
20 •.• �d7 Black prepares the fatal doubling
21 �g l 0-0-0 of his rooks on the g -file, having cal­
22 �n ltg8 (D) culated 27 •a6 •xd5 ! 2 8 •xa7
�c6.
27 �e2 •n
Matanovic has pointed out that
27.....-g6! was possible as after 28
•a6 ltxg2+ 29 �fl ltg8 White has
no mate.
28 �f3 ltdg8
29 ..b3 b6
30 •e3 •r6
31 lte2 �bS!
Not falling for 31. eS? 32 dxe6
••

� c6 3 3 �xc6 ! ltxe3 34 .:.xe3 when

From the mess that was on the White would be very much back in
board nine or ten moves ago Black the game.
has clearly emerged victorious. He 32 ltd2 eS!
has completed his development and 33 dxe6 �c6
his forces co-ordinate beautifully, 34 �n �xf3
whereas for White . . . well , let's just 0-1
5 The Excha nge Variation
Practitioners of the Exchange Vari­
ation can be divided into three cate­
gories. Firstly, there are the endgame
lovers who play this system with the
intention of grinding you down in a
long boring endgame. These people
deserve some respect, although our
main feeling towards them should be
one of sympathy for having such a
feeble system against the King's In­
dian . The second category are the
psychologists. These are the tricky
characters who select this variation
because they feel that it is the most
unpleasant for you to play against,
especially if you are noted as a tacti­
cal player. The third, and in my expe­
rience by far the most numerous
category, are the wim ps. They select
this variation with the idea of killing
the game and achieving an easy
draw. On no account should they be
given one until every las t possibility
has been exhausted . Perhaps they
will achieve their objective in the
end, if they play extremely well, but
they should at least be made to suffer
for it.
Game 1 4
Acebal - Gallagher
Candas 1 992
1 d4 liJf6
2 c4 g6
3 liJc3 �g7
4 e4 d6
5 lijf3 0-0
6 �e2 e5
7 dxe5 dxe5
8 'ii'xd8 llxd8 (D)
9 �g5
By far the most common choice.
White now threatens to win material
with 10 l0d5. 9 liJxe5? would be j ust
a mistake. Sanchez-Geller, Stock­
holm IZ 1 952 continued 9 ... liJxe4 1 0
liJxe4 �xe5 1 1 0-0 ( 1 1 �g5 lld4 ! i s
g ood for Black) 1 l . .. liJc6 1 2 l:e l
rJ;g7 1 3 a3 �f5 1 4 liJg3 �e6 1 5 �fl
a5 16 .tibl a4 and Black had the more
active pieces as well as a positional
advantage on the queenside.
9 liJd5 occurs from time to time.
9 .tid7 is an interesting reply, but
.••
the simplest is to play 9 ltJxd5 1 0
...
cxd5 c6 1 1 �c4 cxd5 1 2 �xd5 liJd7
when White has nothing better than
1 3 �g5 which transposes to the
main line after 1 3 . . . .:es.
86 The Exchange Variation
9 •.• lte8
Black has a l arge choice in this
position: 9 liJbd7, 9 llf8, 9 liJa6
•• • ••• •••
and the modern pawn sacrifice 9 c6 •••
are all quite playable alternatives but
I am of the opinion that the reliable
old line, based on 1 3 . . . ltJ<l7, poses
White the most difficulties. As far as
I can see, White has absolutely no
chance of obtaining the better game
and must even pJay welJ to avoid
ending up in an inferior position.
10 liJd5
The main alternative is 10 0-0-0
which usually leads to a lengthy ma­
noeuvring game with few piece ex­
changes. White has slightly more
space but also a hole on d4 (Black
will cover his d5-square with . . . c6)
which has to be protected at all
times. Black has numerous possibili­
ties but we are going to concentrate
on the one I believe to be the most
logical, 10 ltJa6 ( D)
.•. . whether he wants a knight or a
1 ) White can now take a pawn
with l l liJxe5 as 1 1 ltxe5 12 lld8+
.•.
liJe8 1 3 f4 lte6 1 4 .tg4 is supposed
to. be good for h im , although I' m
not completely sure of this after
1 4 . . . llb8 ! . Any way 1 1 liJc5 is a
•.•
much safer way to play as White
won't be able to hang on to the pawn.
A couple of examples:
la) 12 qjf3 tlJtxe4 13 liJxe4 ltJxe4
1 4 .te3 liJxf2 ! 15 .txf2 .th6+ ! fol­
lowed by taking on e2 left Black a
pawn up in Malich-Peterson, Riga
1 96 1 .
l b) 12 .txf6 .txf6 1 3 f4 c6 14 b4
.t xe5 1 5 fxe5 liJd7 16 e6 llxe6 1 7
.tg4 lte7 1 8 b5 rJ;g7 with a roughly
level game, Haik-Spassky, French
Ch 1 99 1 .
2) 1 1 ltJe l is the standard choice.
The knight heads for c2 from where
it will still have d4 under control and
help prepare an eventual queenside
advance. At the same time White
will now be able to secure his e­
pawn with f2-f3 . 1 1 c6 ( l 1 . . . ltJc5
.•.
1 2 f3 lDe6 1 3 .te3 c5 followed by
. . . liJd4 also looks adequate) 12 ll:k2
( D) and now Black has to decide
bishop on e6:
2a) 12 .te6 1 3 b3 ltJc5 1 4 f3
...
a5 1 5 lld2 liJfd7 1 6 lthd l f5 1 7 liJel
fxe4 ! ? 1 8 ltJxe4? ! (according to the
theory of the superfluous piece
White should recapture with the

pawn) 1 8 . . . llJxe4 1 9 fxe4 llJc5 20
i.f3 a4 with a good game for Black.
Serrer-Uhlmann, B undesliga 1 99 1
continued 2 1 liJd3 i.f8 22 i.e3 lLJd7
23 � b2 b5 ! 24 c5 axb3 25 axb3 b4 !
when the threat of ... .:ta3 is difficult
to meet. If 26 llJxb4, then both
26 .. . llJxc5 and 26 . . . .:teb8 look quite
promising, while the move played in
the game, 26 .:tat, simply lost a
pawn after 26 . . .:xal 27 �xa l i.xb3
as 2 8 liJxb4 is met by 28 . . . i.xc5 29
bxc5 llJxc5 30 llJxc6 i.a4 ! 3 1 .:td6
(3 1 .:td5 liJb3+) 3 1 . . . llJb7 32 .:tf6
�g7.
2b) t2 llJc5 1 3 f3 4Je6 1 4 i.e3
•••
liJf4 (the immediate . . . i.f8 is also
possible but Black w ishes to soften
up the white kingside) 1 5 i.fl h5 16
h4 i.f8 (the bi shop has more future
on this diagonal) 1 7 b4. This is pri­
marily played to prevent . . . llJe6 fol­
lowed by . . . i.c5 , but one of the
reasons I prefer Black in thi s sort o f
position (I'm not claiming an advan­
tage, but I would take the black
pieces if offered the choice) is his
greater king security. Although the
queens have been exchanged all the
other pieces remain and with the
q ueenside quite likely to open at
some point the white king may yet
find itself uncomfortably pl aced. A
good example is the game Lesiege­
Smirin, Biel IZ 1 993 which contin­
ued 1 7 . . . llJe6 1 8 a3 b6 1 9 �b2 i.b7
20 g3 �g7 2 1 i.h3 .:tab8 22 �b3
i.c8 23 i.c 1 a5 24 i.b2 i.a6 25 lbe2
�g8 26 llJe l (26 i.xe5 l2Jc5+ 27
bxc5 .:txe5 28 cxb6 .:txb6+ 29 �c3
i.g7 ! gives Black a very strong at­
tack) 26 ...a4+ 27 �c3 lLJd4 ! 28 liJxd4
The Exchange Variation 87
exd4+ 29 .:txd4 c5 ! 30 .:td 1 cxb4+
3 1 axb4 i.g7 32 �c2 i.xc4 with a
clear plus for Black.
to ... llJxd5
t t cxd5 c6
t2 i.c4 cxd5
t3 i.xd5 liJd7! (D)
Much stronger than the more fre­
quently played t3 ...llJa6 or t3 llJc6
...
which do allow White some chances
of a nagging edge. Black now threat­
e ns to gain the bishop pair with
. . .liJf6 (perhaps preceded by . . . h6) so
White's choice is quite limited.
t4 liJd2!
The only move. By defending his
e-pawn White renders the ...llJf6 idea
harmless and at the same time moves
his knight nearer to the more active
squares on the queenside. S urpris­
ingly often, though, White has ig­
nored Black's threat. For example:
1 ) t4 :ct h6 1 5 i.e3 ( 1 5 i.h4 g5
1 6 i.g3 4Jf6 is good for Black)
1 5 . . . liJf6 16 i.b3 (White can't allow
l 6 . . . llJxd5 so he must seek compli-
cations) 1 6 . . . llJxe4 1 7 .:tc7 i.e6 1 8
i.xe6 .:txe6 1 9 .:txb7 .:ta6 20 a3
liJd6 ! 2 1 .:tc7? ! (it was better to re­
tain control of the b-file, even though
88 The Exchange Variation
2 1 .f.tb4 .f.tc6 22 0-0 f5 was still very
pleasant for Black in the game
Teschner-Fischer, Stockholm IZ
1 962) 2 1 . . . l:tb8 22 .f.tc2 e4 23 llJd4
.f.ta4 ! 24 lbc6 l:txb2 25 .f.txb2 �xb2
26 llJxa7 llJf5 with a clear advantage
for Black, Capusciotti-Gallagher,
Forti 1 992 .
2) 14 0-0-0 h6 1 5 �e3 ( 1 5 �h4
l2Jf6 1 6 �xf6 �xf6 is slightly better
for Black but perhaps the lesser evil)
1 5 . . . l2Jf6 1 6 tlJe l ? ! l2Jxd5 1 7 exd5
�f5 1 8 llJc2 .f.tac8 1 9 .f.td2 .f.tc4 ! and
Black was already close to winning
in Tillmann-Gallagher, Bern 1995.
14 ... tlJcS
14 l2Jb6 and even 14 �f8 are
••• ••.
not so bad for Black but the text is
the most active.
15 l2Jc4
Perhaps not the best as it allows
Black some tricks based on l2Jxe4 ,
while the knight may also get booted
by . . . b5 at some point. The alterna­
tives are:
1 ) 15 0-0 �e6 and now:
l a) 16 �e3? �xd5 17 �xc5 i.c6
l 8 l2Jc4 .f.ted8 ( 1 8 . . . �xe4 is an inter­
esting exchange sacrifice but there is
no need for it) 19 f3 b6 20 �e3 f5
with a clear advantage for Black,
Mtilbach-Gallagher, Bern 1 993 .
l b) 16 �xe6 l2Jxe6 17 �e3 lDd4
l 8 llJb3 ! lbc2 19 .f.tac l llJxe3 20 fxe3
.f.tac8 and despite W hite's ugly cen­
tral pawns he should have no trouble
holding the draw.
2) 15 We2 l2Je6 16 �e3 l2Jf4+
(this . . . l2Je6-f4 manoeuvre can be
considered the key to the position in
a number of lines) 17 �xf4 exf4 and
the openi ng of the long diagonal
assures Black of an active game. A
couple of examples:
2a) 18 l:tacl �xb2 1 9 .f.tc7 �e6
20 �xe6 .f.txe6 2 1 .f.txb7 �c3 22
llJbl llxe4+ 23 'iPf3 .f.tb4! with advan­
tage to Black, Vanheste-Gallagher,
Metz 199 1 .
2b) 18 llJc4 �e6 1 9 Wf3 �xd5
(it looks risky to give White a passed
d-pawn but he doesn' t seem to be
able to keep it) 20 exd5 .f.ted8 2 1
.f.tad 1 .f.tac8 22 b3 .f.tc5 23 d6 b5 24
l2Ja5, H .Olafsson-Ehlvest, Erevan
1 9 88, and now I think 24 . . . �f8 !
would be good for Black. 25 l2Jb7 is
nothing to worry about on account
of 25 . . . .f.tc3+ followed by 26 . . . .f.td7 ,
whilst 25 d7 .f.tc7 26 l:td5 .f.tdxd7 27
.f.txb5 .f.tc3+ ! looks extremely good
for Black.
3 ) 15 0-0-0 l2Je6! (not 1 5 ... lDd3+
16 Wbl l2Jxf2 1 7 .f.tdfl ! and White
wins) 16 i.e3 l2Jf4 17 �xf4 exf4 ( D)
and now White has:
3a) 18 Wbl �e6! 19 �xe6 .f.txe6
20 f3 (20 .f.the l .f.td8) 20 f5 ! with a
••.
couple of possibilities:
3al ) 21 exf5 gxf5, with an edge
for Black in Andersson-Zsu.Polgar,
Bilbao 1987, but 21 ... .f.te2!? looks

even more promising, e.g. 22 fxg6
h xg6 23 g3 b5 ! and White is com­
pletely tied down; 24 :bet is met by
24 . . . l:xh2 and 24 gxf4 l:d8 25 �c2
i.h6 ! wins a piece as 26 �c3 can be
met by 26 . . . b4+ and 26 l:hel by
l:dxd2+.
3a2) 21 l:cl ! (this seems the best
defence) 2 1 . . . .t:d8 22 l:c2 l:ed6 (I
went for the promising bishop v
knight ending but 22 ... i.d4 was a se­
rious alternative) 23 lZJc4 l:d 1 + 24
l:xd 1 l:xd 1 + 25 l:c 1 l:xc 1 + 26
�xc 1 fxe4 27 fxe4 (27 lZJd6 e3 28
b3 ! may also enable White to draw
by setting up a fortress position)
27 . . . b5 28 lZJa3 ! a6 29 lZJc2 (the
knight is heading for its ideal square:
d3) 29 . . . �f7 30 lZJb4 a5 3 1 lZJd3 g5
and in Ekstrom-Gallagher, Villars
1 995 , I made what I thought was a
generous draw offer in order to se­
cure first place in the tournament.
My opponent accepted it seventy­
five minutes later! A possible con­
tinuation is 32 �d2 �e6 33 �e2
i.e5 ! 34 �f3 h5 35 h3 a4 with an
edge for Black, although our post­
mortem concluded that White should
be able to hold.
3 b) After the above Andersson­
Polgar game 18 f3 was suggested as
an improvement for White, but here
too Black has an easy life: 1 8 . . . i.e6 !
19 lZJb3 i.xd5 20 l:xd5 (20 exd5
l:e2 wins a pawn) 20 . . . f5 ! (the by
now familiar way of activating the
black rooks) 2 1 exf5 l:ac8+ 22 �b l
l:e2 23 f6 ! ? (White reserves the e4-
square for later use) 23 . . . i.xf6 24
l:d2 and now P.Cramling-Gallagher,
B iel 1 99 1 was agreed drawn after
The Exchange Variation 89
24 .:ceS and P.Cram ling-Grivas,
•..
Debrecen Echt 1 992 was agreed
drawn after 24 .t:e7.
..•
15
••• i.f8
This keeps the knight out of d6,
but the bold could investigate giving
up an exchange with 15 i.e6 1 6••.
lDd6 i.xd5 .
1 6 0-0 (D)
The more aggressive 16 0-0-0 led
to a quick defeat for White in the
game S algado-Gallagher, L'Hospi­
talet 1 992 after 1 6 . . . i.e6 1 7 �bl
l:ac8 ( 1 7 . . . lZJxe4 1 8 i.xe4 i.xc4 1 9
i.xb7 l:ab8 is fine for Black but I
was playing for tricks) 1 8 i.e3 ? (bet­
ter is 1 8 l:he 1 when 1 8 . . . i.xd5 1 9
exd5 should be slightly better for
Black as White's d-paw n is more
likely to turn out weak than strong)
l 8 . . . lZJxe4 ! 1 9 i.xe4 l:xc4 20 i.xb7
l:b8 2 1 i.d5 i.f5+ 22 �a l l:c2 and
unfortunately for White 23 i.b3
loses to 23 . . . l:xb3 ! 24 axb3 l:c6 ! .
Therefore he tried 23 i.xa7 but after
23 . . . l:bxb2 24 i.e3 i.b4 25 g4 i.c3
he was probably wondering why he
had even bothered to get out of bed.
16 •.. i.e6
17 i.xe6 l:xe6!
90 The Exchange Variation

At first glance one could dismiss

this position as dead drawn, but a
closer inspection will reveal a size­
able initiative for Black. His rooks
are more active, White's bishop is
offside on g5 and the knight on c4
will soon be hit by . . . b5 .
t8 f3 bS
t9 llle3 h6
20 .i h4 llld 3
2t liJdS llc8
22 b3 llc2
23 llfdt lllb4!
Stronger than 23 ... .icS+ 24 �fl
.id4 which allows White to solve his
problems after 25 llabl lllc 5 26
llbc l ! . People often talk about the power
24 lllx b4 of doubled rooks on the seventh
Perhaps White should have tried r ank, but a rook and bishop? In fact,
24 llact as 24 llxa2 25 llc8 is awk­
••. W hite is totally paralysed and can
ward to meet. Instead Black should only watch while Black calmly im­
play 24 llxcl ! 25 llxc l lll xd5 26
•.. proves h is posi tion by bringing the
exd5 lld6 reaching a favourable end­ king to the centre and playing . . .f5 .
game. 27 .icS aS .

24 ... .ix b4 28 �n llc6

2S .if2 a6! 29 b4 a4
I was tempted by 2S lla6, but
..• 30 llabt �g7
this would have been a false trail. Af­ 3t :at rs
ter 26 a4 .ic3 27 llac 1 llxc 1 28 32 �gt �f6
llxc l bxa4 29 llxc3 axb3 30 llxb3 ! 33 :n �e6
llal + 3 1 .iel llxel + 32 �1"2 llal 33 34 llf2?
llb5 ! f6 34 llb7 White will have no This loses material but passive de­
trouble holding the draw. fence would also have lost. One plan
26 a3 for B lack would be to play . . . f4 fol­
Now 26 a4 could simply be met lowed by ... .ie3+.
by 26 . . . bxa4 as after 27 llxa4 a5 34 •.• ll6xcS !
White remains bottled up. On 26 .ie3, 3S exfS+ gxfS
with the idea of preventing . . . .id2, 36 llxd2 llxd2
B lack cou ld play 26 . . . .ia3 so that 37 bxcS llc2
27 .ixh6 can be met by 27 . . . .ib2 0-t
6 5 i.d3

Whilst one wouldn't call this a popu­

lar way of meeting the King's Indian
it is , nevertheless , a system which
has developed considerabl y over the
last few years. White has a simple
plan: complete his kingside develop­
ment (i.d3, l2Jge2, 0-0) as quickly as
poss ible and be ready to meet any
subsequent . . . f5 by Black with exf5
and f4. This is a solid line in which
Black's chances of a successful king­
side attack are slim and it can number exd5 (this can be delayed) 9 cxd5 is
amongst its regular users grandmas­ quite a tricky one for h im.
ters such as Seirawan , Christiansen The traditional King 's Indian re­
and Marin. The main drawback with sponse, 6 e5, has been s lightly ne­
.••

an early i.d3 is that it slackens glected here, even though after 7 d5

White's already rather shaky grip on ltJhS Black seems to have a perfectly
d4 . Therefore, it is no great surprise playable position, e.g. :
that Black's most popular defence is 1 ) 8 h3?! (8 f3 1i'h4+ 9 g3 1i'h3
based on a qu ick assault against this was also quite good for Black in
point. Hodgson-Pein, British Ch 1 987) 8 .. .f5
9 exf5 gxf5 1 0 g4 l2Jf4 ! ( 1 0 . . . fxg4?
Game 1 5 1 1 hxg4 i.xg4 1 2 1i'c2 followed by
Christiansen - Gallagher 1 3 i.e3 and 14 0-0-0 allows White a
Bern 1996 strong kingside attack) 1 1 l2Jxf4 ( 1 1
i.xf4 exf4 1 2 1i'c2 l2Ja6 1 3 a3 l2Jc5
1 d4 l2J r6 14 0-0-0 f3 1 5 l2Jg3 1i'g5+ is given as
2 c4 g6 unclear by Nadyrkhanov but I sus­
3 l2Jc3 i. g7 pect he has underestimated Black's
4 e4 d6 chances) 1 1 . .. exf4 1 2 i.xf5 i.xf5 1 3
5 i.d3 0-0 exf5 IZ.xf5 14 1i'g4 1i'f6, with advan­
6 l2Jge2 (D) tage to Black, Tunik-Nadyrkhanov,
6 ••• l2Jc6 Voskresensk 1 993.
Black can also react in Benoni 2) 8 0-0 C5 9 exfS gxfS 10 C4 (this
fashion with 6 c5, although I think
••• move represents White's big idea)
the resulting variation, 7 d5 e6 8 0-0 and now :
92 5 j_dJ
2a) 10 e4 1 1 j_c2 c5 1 2 j_e3 !?
•••
(12 dxc6 ltJxc6 1 3 j_e3 j_e6 14 j_b3
{ if 14 b3 then 14 . . .ltJf6 followed by
... d5 } 1 4 ... ltJf6 1 5 h3 f/e7 1 6 ltJb5
ltJe8 1 7 ltJbd4 ltJxd4 1 8 ltJxd4 j_d7
was about level in Basagic-Djeno,
Zagreb 1993) 12 ...l2Jcl7 13 f/d2 :f6?!
( 1 3 . . . ltJdf6 is ;!; according to Marin)
14 g4 ! :g6 1 5 g5 ltJb6 ( 1 5 . . . h6 1 6
ltJxe4 ! ) 1 6 b 3 h 6 1 7 :f2 ! hxg5 1 8
fxg5 f4 1 9 ltJxf4 f/xg5+ 20 ltJg2
fle5 2 1 :an j_h3 22 ltJxe4 :es 23
ltJg5 ! j_xg2 24 :xg2 :f6 25 :xf6
flxe3+ 26 flxe3 :xe3 27 :xd6 1 -0
Marin-Llanos, Berga 1 993 .
2b) ECO suggests 10 ltJd7 giv·
•.• , <sth2 j_xc5 1 9 :f3 gave White good
ing 1 1 :b t exf4 1 2 ltJxf4 ltJxf4 1 3
j_xf4 llle5 as =.
7 0-0 ltJhS ! ?
It would usually be pretty pro­
vocative for Black to move h is king's
knight before play ing . . . e5, but here
White's hands are tied by the threat
to his d-pawn. The text has been gain­
ing in popularity recently, mainly
because Black is in bad shape in the
main line. This runs 7 eS 8 d5 ltJd4
•.•
9 ltJxd4 exd4 1 0 ltJb5 lte8 1 1 :e 1
ltJg4 ( 1 1 ... a6 1 2 ltJxd4 ltJxd5 1 3 j_fl
is an edge for White) 1 2 h3 a6 1 3
hxg4 axb5 14 cxb5 fih4 and now 15
g3 fih3 1 6 j_ft flxg4 17 f/xg4
j_xg4 1 8 <stg2 gives White a favour­
able ending on account of his poten­
tial passed a-pawn and pressure on
the c-file, while Milov 's new move,
15 j_ f4 may be even stronger.
, fih3 ltJg4+ 1 9 <stg2 when Black's in­
7... !JJd7 is another idea. After 8
j_e3 eS 9 dS !JJd4 (9 . . . llle7 is less ef­
fective here) B lack's position is not
as good as it may seem at fust sight.
A couple of examples (D):
1 ) 10 j_c2!? ltJxc2 1 1 flxc2 f5
1 2 exf5 gxf5 1 3 f4 ltJf6 14 h3 � 5
( 14 ... j_d7 is more solid) 15 c5 ! ? exf4
1 6 ltJxf4 ltJxf4 1 7 j_xf4 j_d4+ 1 8
play for a pawn in Seirawan .. Van
Wely, Wijk aan Zee 1 995.
2) 10 :ct c5 1 1 dxc6 bxc6 1 2 b4
f5 1 3 exf5 gxf5 14 j_xd4 exd4 1 5
ltJa4 f/g5 1 6 f4 f/g6 1 7 c5 w ith ad­
vantage to Wh ite, Agdeste in-Dol­
matov, Tilburg 1 993.
8 j_e3
As 8 dS ltJe5, with . . . c6 to follow,
cannot be good for White, the only
other move is 8 j_c2. Aleksandrov­
Golubev, Nikolaev 1 993 continued
8 ...e5 9 dxe5 (9 d5 llle7 is surely not
an improvement on the main l ine)
9 ltJxeS 1 0 b3 f/h4 1 1 j_d2 ltJg4
•.•
(odd; 1 1 . . . f5 is more natural) 1 2 h3
ltJe5 13 f4 j_xh3 ( 1 3 . . . ltJc6 would
leave Black worse but is s ounder
than the text) 1 4 fie l ! f/g4 15 f/f2
f5 1 6 <sth2 j_xg2 17 f/xg2 Wih4+ 1 8
vestment had clearly not paid off.
9 dxeS is playable, but slightly pas­
..•
sive: 10 ifxd8 :xd8 ( 10 . . . ltJxd8 ;!;)
1 1 ltJd5 j_e6 ( 1 1 . . . :d7 12 j_a4 ! +)
1 2 :d t ;!;.

8 .•• e5
9 d5 !i:Je7
9 !i:Jd4 ( D) is obviously an im­
...
portant alternative, and then:
1 ) 10 .ixd4?! exd4 1 1 !i:Jb5 c5 1 2
dxc6 bxc6 1 3 !i:Jbxd4 c5 and Black
regain s the sacrificed pawn with ad­
vantage.
2) 10 !i:Jb5 !i:Jxe2+ ( 1 0 . . . !i:Jxb5 1 1
cxb5 f5 is an alternative) 1 1 .ixe2
!i:Jf4 12 .if3 f5 1 3 !i:Jc3 .id7 14 exf5
gxf5 1 5 .ixf4 exf4 1 6 'A'd2 'A'h4 1 7
llfe l .ie5, Kiselev-Zaitsev, Moscow
1 992 . Here we have a conflict of
opinions as B elov considers White to
be better whilst Knaak prefers Black.
This suggests that the chances are
about equal.
3) 1 0 ltcl a6 1 1 b4 'A'e8 1 2 .ibl
!i:Jxe2+ 13 !i:Jxe2 f5 14 exf5 gxf5 15
f3, Tunik-Fedorov, Minsk 1995 is
assessed as ;t by Belov but Black's
1 0th move looks like the play of an
indecisive man. 10 rs or 10 cS look
••• .••
more to the point.
4) 1 0 'A'd2! ? deserves an outing
as this introduces the possibility of a
later .ig5.
10 'A'd2 rs
1 1 exr5 gxf5 (D)
5 .id3 93
Black can also play 1 1 !i:Jxrs ••.
when 1 2 .ig5 .if6 1 3 .i xf6 'A'xf6
was about equal in Piket-J .Polgar,
Amsterdam 1 995. White might do
better, though, to avoid exchangin g
dark-squared bishops: 12 !i:Je4 !i:Jf6
1 3 !i:J2c3 has been suggested as ;!;,
Further tests required.
1 2 .ig5! ?
1 2 C4 is more natural, after which
Black should play 1 2 !i:Jg6!. White
.••
then has to decide what to do abou t
the pressure on f4 :
1 ) 13 rxe5 (I would certainly be
extremely reluctant to play such a
move) 1 3 . . . dxe5 14 .ig5 'A'd6 15
!i:Jb5 'A'd7 16 !i:Jg3? (White inflicts
serious kingside weaknesses upon
himself; better would have been 1 6
llael a6 1 7 !i:Jbc3 'A'd6 with an un­
clear game according to Belov)
1 6 . .. !i:Jxg3 17 hxg3 a6 18 !i:Jc3 'A'd6
with advantage to Black, Tunik­
Shulman, Minsk 1 995.
2) 13 g3 is Tunik' s latest try. Af­
ter 1 3 . . . .id7 ( 1 3 . . . exf4 ! ?) 14 ltae l
a6 1 5 b3 exf4 1 6 !i:Jxf4 !i:Jg xf4 1 7
gxf4 'A'h4 1 8 .ie2 llf6 1 9 .ixh5
'A'xh5 20 !i:Je2 l:g6+ 2 1 !i:Jg 3 'A'h3
22 'A'g2 'A'xg2+ 23 <iit x g2 lte8 a
94 5 i.d3
roughly level ending had arisen,
Tunik-Mochalov, Orel 1 995 .
12 ••• f4!
I wasted an awful lot of time on
this, mainly because I didn't w ant to
be laughed at for playing such an
anti-positional move. Black cedes
control of e4 but in return gains f5
for his pieces and a potential attack
along the g-file. It seems to me that
control of e4 is not nearly as impor­
tant here as, for example, in a posi­
tion without f-pawns .
13 f3
Otherwise B lack may advance
... f3
13 ••• 'ird7!?
I'm sure many people would have
played 13 i.f6 but I don' t like giv­
•••
ing up my King's Indian bishop even
when it appears to be a miserable
lump of wood (or was it plastic?). I
was already dreaming of its trium­
phant emergence on the a7-gl diago­
nal and was also slightly concerned
about 1 4 i.xf6 ltJxf6 1 5 g3 !? , but
thi s was probably just an excuse to
keep the bishop. I feel vindicated by
the fact that it captured a rook just
seven moves later.
14 i.xe7
An understandable decision be­
cause 1 4 i.h4 ltJf5 1 5 i.f2 i.h6
would be a nerve-racking experience
for White.
14 ... 'irxe7
15 'irc2 i.h6
After 15 i.f6, the bishop would
•••
be impeding its own queen.
16 JZ.ael :t7 (D)
After puzzling over my oppo­
nent' s last move for a few moments
I calmly prepared the transfer of my
rook to the g-file.
17 ltJd4!
This came as a complete shock. I
had actually considered 17 ltJxf4 but
felt that there wasn't much chance of
that one working as f4 is defended by
three pieces in addition to the pinned
e-pawn. In fact, not seeing ltJd4
probably worked in my favour as
there was no really good way to pre­
vent it ( 1 6 . 'irg7 but I wanted the
.. ,
queen to go to h4 ) and I didn ' t have
to waste a lot of time worrying about
it.
17 ••• �h8
A useful move as there were some
variations where the h-pawn could
be taken with check and others
where White could profit from my
king being on the a2-g8 diagonal.
18 c5 i.g5
I didn't even consider 18... dxcS as
this w ould weaken the crucial e­
pawn, preferring instead to activate
my bishop.
19 li)e4!?
Just three moves ago I was slowly
building up my kingside attack, con­
fident that my opponent was devoid

of counterplay. Suddenly, the board
was ablaze with his pieces.
19 •.. � h4!
I did well to avoid 19 exd4 20
•..
ltJxd6 cxd6 2 1 l:.xe7 �xe7 (2 l . ..lhe7
22 cxd6 l:.g7 23 l:.e 1 is also ex­
tremely dangerous) 22 �g6 ! when
Black is in serious trouble.
20 liJb5
The main point behind my last
move was that after 20 l:.e2 exd4 ! 2 1
ltJxd6 cxd6 22 l:.xe7 llxe7 23 cxd6
l:.g7 the e 1-square is covered so
White's rook won ' t be able to get in
the game. Christiansen prefers to
give up the exchange and collect the
d6-pawn, but he overlooked a nasty
tactical point. White could have first
captured on d6 before playing liJb5
when play would almost certainly
transpose to the note to the 2 1 st
move.
20 .•. � xe l
2 1 l:.xel a6! ? (D)
Objectively speaking it may have
been better to play 21 l:.g7 when 22
••.
cxd6 cxd6 23 liJbxd6 �h3 gives
Black good attacking chances. How­
ever, I felt sure that my opponent
hadn ' t seen what was coming and,
anyway, who can resist a whole rook?
5 �d3 95
22 cxd6
Obvious ly retreating the knight is
hopeless , but White had an alterna­
tive sacrifice in 22 ltJxc7 ! ?. After
22 ... it'xc7 23 ltJxd6 l:.e7 24 ltJxc 8
(or 24 it'c 3 liJf6 ! 25 ltJxc 8 ltJxd5)
24 ... it'xc8 (24 . . . l:.xc8? 25 d6 it'a5 26
b4 ! it'xb4 27 lle4 it'xc5+ 28 it'xc5
l:.xc5 29 dxe7 liJf6 30 g3 ! leaves
Black fighting for a draw) 25 d6 l:.g7
26 l:.xe5 liJf6 the situation is very
unclear. White has three pawns and a
positional advantage for his rook. In
the post mortem Christiansen felt
that this would give him a decisive
advantage whilst I was of the opin­
ion that Black 's problems were not
insurmountable.
22 .•. Vi'h4!
23 it'c3 axb5
24 it'xe5+
On 24 dxc7 Black has the reply
24 . .. l:.xc7 ! (25 it'xe5+ l:.g7) .
24•.• l:.g7 !
Not 24 . . �gS? on account of 25
.
ltJg5 ! cxd6 26 it'e8+ l:.f8 27 �xh7+
�g7 28 l:.e7+ and White wins.
25 l:.cl ?
This just leaves White a rook
down for not much. He should have
played 25 it'e8+ llg8 26 it'e5+ when
Black must try 26 . . . ltJg7 if he's go­
ing for the win. After 27 ltJg5 ! l:.e8
28 it'xe8+ ltJxe8 29 llxe8+ �g7 30
l:.e7+ �f8 3 1 liJxh7+ �g8 32 l:.e8+
�g7 33 l:.e7+ it'xe7 (33 .. . �h6?
loses to 34 liJf8 ! and 33 . . . �g8 is a
draw) 34 dxe7 �f7 (34 . . . �d7 35
liJf8 !) 35 �xb5 �xe7 an unclear
ending has arisen in which White
has a slight material advantage but
B lack's pieces are more active (even
96 5 J..d3

if the rook and bishop are still at

home).
25 .... cxd6
26 llJxd6 i.d7
27 llJf7+ � g8
28 liJb6+ �f8!
29 llet
After 29 liJfS J..xf5 30 1fxf5+ both
30 l2Jf6 and 30 1ff6 win for Black.
••• •••

29 ••• .:e s
Avoiding the trap 29 . 1ff6? 30
. .

1fxh5 1id4+ 3 1 �h l 1fxd3 32 1'e5 ! . 32 1ixf6+ l2Jxf6

30 1id6+ llge7 33 llgS llg7
31 :es 1if6! (D) 34 b4 llxgS
White could have res igned here 35 bxgS llJxdS
but played a few more moves be­ 36 J..x h7
cause of my slight time pressure. 0-1
7 5 tbge2

I'm not sure who first thought of this In the notes below you will also
system but it was quite prominent find a d iscussion of the main line
amongst Hungarian players in the against ltJge2 (5 . . . 0-0 6 liJg3 e5) as
1 960s before disappearing and re­ this is also quite reasonable for
turning to enjoy a slight renaissance Black.
period in the 1 980s . To spend a cou­
ple of tempi manoeuvring a knight to Game 1 6
g 3 (it obviously can't remain on e2 Flear - Gallagher
where it clogs up the whole kingside, San Bernardino 1 991
the one exception being when White
fianchettoes his king's bishop) may 1 d4 liJf6
seem like strange behaviour, but 2 c4 g6
from there the influence it exerts on 3 ltJc3 Jl.g7
the e4- and f5-squares makes it more 4 e4 d6
difficult for Black to achieve his tra­ S liJge2 (D)
ditional kingside counterplay. Con­
sequently, my favoured approach for
B lack's is to initiate queenside pro­
ceedings at once, even delaying cas­
tling as the tempo saved may be put
to good use on the queenside and an
early h4 by White will be less men­
acing. The basic position is arrived at
after the moves 1 d4 liJf6 2 c4 g6 3
4X3 Jl.g7 4 e4 d6 5 ltJge2 a6 6 liJg3
c6 when White has an important de­
c ision to make; whether to stop . . . b5
with a4 (which has certain positional S a6
disadvantages after the reply . . . a5), B lack can equally play S c6, the
•••

or to ignore B lack' s queens ide dem­
onstration in favour of quick devel­
opment or attackin g in the centre.
The h-pawn plays a very prominent
role in this variation and a lot of
g ames will see either a quick h4 by
White or an early . .. h5 by Black.
choice being dependent on what you
play against the Samisch (a variation
beyond the scope of this book, which
we have transposed into after 6 f3). If
your variation contains neither an
early . . . c6 nor an early . . . a6, then you
can study the material j ust below on
 98 5 !iJge2
5 . . . 0-0, or play the percentage game
- th is means assuming that people
who play the !iJge2 system are not
going to transpose into the Samisch.
However, take care if you find your­
self Black against Novikov.
As mentioned above, 5 0-0 6 •..
!iJg3 e5 7 d5 is considered to be the
main line . Here is a summary of the
current state of affairs:
1 ) 7 !iJg4 (speculative) 8 �e2
•..
1i'h4 9 !iJb5 (9 f3 ? !iJxh2 ! ) 9 . . .!iJa6
1 0 �d2 �h6 ( 1 0 . . . c6 1 1 1i'c l ! 1i'd8
1 2 !iJc3 ;I; Forintos-Sinko v ics , Hun­
gary 1 986) 1 1 �xh6 !iJxh6 1 2 0-0 f5
1 3 exf5 gxf5 1 4 f4 exf4 1 5 !iJh5 f3
1 6 �xf3 1fxc4 1 7 a4, Serper-Dan­
nevig, Gausdal 1 99 1 , and now ac­
cording to Knaak Black should have
tried l 7 . . . 1i'h4.
2) 7 a5 8 �e2 !iJa6 9 h4 (9 0-0
••.
is not very promising, e.g. G.Geor­
gadze-Akopian, Tbilisi 1 989 contin­
ued 9 . . . !iJc5 1 0 b3 �d7 1 l llb l h5 1 2
�g5 1i'e8 1 3 1i'd2 !iJh7 14 �h6 h4
1 5 �xg7 �xg7 1 6 !iJh l 1i'e7 1 7
l:tbe l 1i'g5 with a favourable game
for Black) 9 c6 (9 . . . h5 is also play­
.. .
able, e.g. Gurevich-Nijboer contin­
ued 1 0 �g5 1i'e8 1 1 1i'd2 !iJh7 1 2
· �e3 �d7 1 3 a4 1i'e7 14 !iJb5 f5 !
with unclear play) 10 h5 and now:
2a) 10 cxd5 (virtually everyone
•. .
adopts this move order but it seems
inferior to '2b ' ) 1 1 cxd5 !iJc5 1 2 a4 !
�d7 1 3 l:ta3 l:tc 8 1 4 �e3 1i'b6 1 5
hxg6 fxg6 1 6 f3 with an edge for
White, Novikov-Hernandez, Pam­
plona 1 99 1/2.
2b) 10 lt:Jc5 ! 1 1 �e3 (playing
•• .
1 1 a4 makes little sense before Black
h as exchanged on d5 ) 1 1 . . . cxd5 1 2
cxd5 1i'b6 ! 1 3 l:tb l ( 1 3 b3 �d7 1 4
o-o 1i'b4 ! ? 15 1i'd2 ! :fcs 1 6 a3 1i'b6
1 7 :abl 1i'd8 is roughly leve l, Kor­
chnoi-Nunn, Wijk aan Zee 1 992)
l 3 . . .�d7 14 !iJfl a4 1 5 !iJd2 1fa5 1 6
hxg6 hxg6 1 7 f3? ! ( 1 7 �fl was bet­
ter) 1 7 . . . b5 1 8 a3 !iJh5 with good
play for B lack, Lutz-Gelfand, Hor­
gen 1 994.
3) 7 c6 8 �e2 cxd5 (8 . . . a6 !?) 9
••.
cxd5 a6 (9 . . . !iJbd7 has little inde­
pendent significance; if Black fol­
lows up with . . . a5 we are likely to
transpose to '2', whilst if he follows
up with . . . a6 we are likely to trans­
pose to one of the lines considered
below) and White has tried several
moves (D) :
3a) 10 0-0 !iJbd7 1 1 �e3 b5
( 1 l . .. h5 is also possible) 1 2 b4 !iJb6
1 3 a4 !iJxa4 14 !iJxa4 bxa4 1 5 D.xa4
h5 ! is unclear according to ECO.
3b) 1 0 �e3 h5 ! (often a good
response to �e3 as the bishop now
belongs on g5 and White is usually
wil ling to invest a tempo to put it
there) 1 1 �g5 ( 1 1 h3 !iJbd7 1 2 !iJf l
b5 1 3 !iJd2 !iJh7 1 4 !iJb3 f5 1 5 exf5
gxf5 1 6 �xh5 b4 17 !iJa4 f4 1 8 �d2
e4 1 9 �xb4 !iJe5 was excellent for

B lack in the g ame Szabo-B asagic,
Sarajevo 1 972) l t . .. li:Jbd7 1 2 'i'ld2
'i'lb6 1 3 0-0-0? ! ( 1 3 0-0 is safer)
1 3 ... li:Jh7 14 i.e3 'i'la5 1 5 'iti>bl ?! ( 1 5
i.xh5 ! ?) 1 5 . . . h 4 1 6 tt:Jn l2Jc5 with a
good game for Black, Serper-Api­
cella, Asiago 1994 .
3b) 10 h4 h5 1 1 i.g5 li:Jbd7 1 2
tt:J n 'i'le8 1 3 li:Jd2 b 5 14 a3 li:Jh7 1 5
i.e3 f5 1 6 f3 li:Jdf6 1 7 i.d3 i.d7 1 8
'i'le2 'iti>h8 1 9 0-0-0 'i'lb8 20 b4 gave
rise to a complex strategic struggle
in Nov ikov-Xie Jun, Helsinki 1992,
eventually drawn after 5 1 moves.
6 li:Jg3 c6
7 a4
This has been White's most popu­
lar choice, the main alternative being
7 i. e2 . After 7 ... b5 (7 . . . h5 ! ? or 7 . . . e5
can also be played but it seems a lit­
tle inconsistent not to play . . . b5)
there is (D) :
1 ) 8 e5! ? (a little premature per­
haps) 8 dxe5 (8 . . . li:Jfd7? 9 exd6
•• .
exd6 1 0 i.f4 li:Jf6 1 1 li:Jge4) 9 dxe5
'i'lxdl+ 10 li:Jxdl ( 1 0 i.xd l li:Jg4 1 1
f4 bxc4 1 2 i.e2 i.e6) 10 ...li:Jfd7 1 1
f4 with a couple of examples:
l a) In Tyrtan ia-Gallagher, Bad
Worishofen 1 993 Black attacked the
5 li:Jge2 99
centre at once with 1 1 f6 . After 1 2
••.
exf6 exf6 1 3 li:Je4 f5 (on 1 3 . . . 'iti>e 7 I
w as afraid of 1 4 i.d2 ! ) 14 li:Jd6+
'iti>e7 1 5 li:Jxc8+ llxc8 1 6 i.d2 ( 1 6
i.e3 ! ? li:Jf6) 1 6. . . lDc5 ! 1 7 i.b4 li:Jbd7
1 8 liJc3 ( 1 8 llcl is well met by
1 8 . . . a5 ! as although 1 9 i.xc5 li:Jxc5
20 cxb5 cxb5 2 1 i. xb5 may appear
to win a pawn it loses the game after
2 1 ... �3+!) 1 8 ...'iti>t7 19 lld l i.f8 the
game was about level.
l b) Perhaps 11 li:Jb6 offers more
..•
chances of a complex middlegame.
The game Goormachtigh-W. Watson,
Brussels 1 986 continued 12 li:Je3
i.e6 1 3 i.d2 li:J8d7 1 4 llc l i.h6 1 5
llfl ( 1 5 0-0 looks more natural, al­
though White may have been wor­
ried about some combination of
. . . bxc4 and . . . li:Jxe5, exploiting the
undefended state of the bishop on
d2) 1 5 . . . 0-0 1 6 b3 llad8 1 7 li:Je4
bxc4 1 8 li:Jxc4 li:Jxc4 1 9 i.xc4 i.f5 !
20 li:Jg3 li:Jxe5 2 1 li:Jxf5 gxf5 22
i.xa6 lld4 ! (White is allowed no
peace) 23 llc2 li:Jg4 24 h3 li:Jf6 25
i.c3 :e4+ 26 lle2 lla8 2 7 i.xf6
llxa6 28 i.xe7 llxa2 ! 29 llxe4 fxe4
leading to a favourable endgame for
Black.
2) 8 cxb5 axb5 9 b4. This is often
an effective way to counter an early
. . . b5 by Black in the King's Indian ­
first blockade and later seize the in­
itiative with a4 - but here White has
a slightly inferior version as he has
already committed his kn ight to g3.
In the s imilar pos itions arising
from the Samisch, for example, the
e2-knight usually settles on the more
active b3-square. 9 . . .li:Jbd7 10 i.b2
0-0 1 1 0-0 i.b7? ! (this is what John
100 5 �ge2
Nunn would call a lazy move; Black
doesn't want to have to calculate the
consequences of d5 at each tum so
he prevents it but in doing so he con­
demns his bishop to passivity; better
was l l . . . �b6, keeping open the op­
tion of . . . �e6) 1 2 Wb3 �b6 1 3 l:tfd l
h5 1 4 �fl e6 1 5 �c l �fd7 1 6 �e3
�a4 17 lbd2 e5 1 8 �f3 We7 1 9
l:tac I with an edge for White� Rem­
linger-Djurhuus, Gausdal 199 1 .
3) 8 0-0 0-0 (8 . .. bxc4 ! ? 9 � xc4
d5 I O �b3 ! dxe4 1 1 �gxe4 �5
{ 1 1 . . . �xe4 1 2 �xe4 Wxd4 would
be too risky } 1 2 Wf3 0-0 1 3 �g5
l:ta7 14 l:tad l h6 15 �c l e6 16 l:tfel
is a little better for White, Novikov­
Kruppa, USSR 199 1 ) 9 f4 ! ? e5 I O
fxe5 dxe5 1 1 d5 cxd5 1 2 �d5
(Black does not fear 1 2 cxd5 as he
will be easily able to blockade the
passed pawn and perhaps under­
mine it later with . . . f5) 1 2 . . . �xd5 1 3
Wxd5 Wb6+ 1 4 c5 Wc6 I 5 a4 �e6 !
1 6 axb5 axb5 1 7 Wxc6 �xc6 1 8
l:txa8 l:txa8 1 9 � xb5 �d4 20 �d3
l:tb8 ! 21 �e2 �f8 and Black re­
gained the sacrificed pawn without
relinquishing his positional advan­
t age, Remlinger-Hebden, Gausdal
1992.
Before moving on, 7 h4 deserves
a quick mention. Liardet-Gallagher,
Geneva 1 993 continued 7 . . . h5 8 �e2
b5 9 cxb5 axb5 I O b4 0-0 (often
when Black has played . . . h5 in re­
sponse to h4 he has to worry about
piece sacrifices on h5, but delay ing
castling until White has played a
move such as b4 renders it extremely
unlikely that White can conduct a
kings ide attack without al lowing
serious counterplay on the queen­
side) 1 1 �g5 �bd7 1 2 iid2 �b6 1 3
0-0 ( 1 3 l:tc l) 1 3 . . . �h7 ! 1 4 �e3 e6 (it
turns out that it's White who has
problems on the kingside) 1 5 d5 cxd5
1 6 �xb5 ( 1 6 exd5 bxc3 17 Wxc3
�xd5 is good for Black) 1 6 . . . d xe4
1 7 �gxe4? (1 7 a4 d5 is less clear)
1 7 . dS 1 8 �g5 ? d4 1 9 l:tfdl e5 and
. .
Black won material.
7 ... aS! (D)
To the uninitiated, playing . . . a6
and then . aS may seem like a crimi­
..
nal waste of time, but the point is that
Black has now secured outposts for
himself on the queenside - b4 at
once and the more important c5 -
square after an eventual . . . e5 .
8 �e2 0-0
A case can be made out for play­
ing 8 eS although the most likely
.••
result is transposition to lines con­
sidered below. A couple of inde­
pendent examples:
I) 9 dxeS d xe5 I O Wxd8+ 'iii> x d8
should be fine for Black. A possible
continuation: 1 1 f4 �bd7 1 2 0-0
<i/e7 1 3 �e 3 exf4 1 4 � xf4 �e8
when 15 �g5+ should not be met by
15 �f6? 1 6 l:txf6 ! �exf6 1 7 e5
...

when White wins but by 15 f6. It is
••.
quite noticeable how badly pl aced
the knight on g3 is.
2) 9 dS l2Ja6 10 h4 h5 :tl :a3 ! ?
(now it would be dangerous for
Black to castle in view of i.g5 and a
sacrifice on h5, so he has to come up
with something else) 1 1 . . . lDd7 ! 1 2
i.g5 ( 1 2 dxc6 bxc6 1 3 Wxd6 i. f8 !)
12 . . . i.f6 13 dxc6 bxc6 14 Wxd6
l2Jb4 ( 1 4 ... i.xg5 15 hxg5 lDb4 looks
more accurate; perhaps Black was
worried about 15 Wxc6 but 15 . . . :bs
leaves him with the devastating
threat of ... lDb4) 15 Wd2 ( 15 i.xf6
was better even though Black would
still have adequate compensation)
15 . . . i.x g5 1 6 hxg5 l2Jc5 ! 1 7 l2Ja2
Wxd2+ 1 8 �xd2 �e7 19 �c3 l:.b8
20 llJc 1 :ds with tremendous posi­
tional compensation for the pawn,
Rasmussen-Berg Hansen, Danish
Ch 1 992.
9 f4?!
The most aggressive choice but
it doesn' t fit in well with White's
queenside pawn structure. The alter­
natives are:
1 ) 9 0-0 es and now:
l a) 1 0 dS l2Ja6 1 1 i.e3 lDc5 1 2
: e 1 (perhaps White could have dis­
pensed with this move; if he was in­
tending to free fl for the knight then
h e later changed his mind and re­
treated it to h 1 ) 1 2 . . . h5 (other ideas,
s uch as 1 2 . . . cxd5 1 3 cxd5 l2Je8 de­
serve consideration) 1 3 f3 h4 1 4
lDh l lDh5 15 lDf2 i.f6 ! ( 15 . . . l2Jf4 1 6
..t n cxd5 1 7 cxd5 b6 1 8 �h l was
good for White in Verdikhanov­
Kruppa, Nikolaev Z 1 993) 1 6 lDd3
.
lDxd3 17 Wxd3 i.g5 1 8 i.fl i.xe3+
5 lDge2 101
1 9 Wxe3 with an unclear position ac­
cording to Verdikhanov.
l b) 10 i.e3 l2Ja6 1 l Wd2 l2Jg4 1 2
i.xg4 i.xg4 1 3 f3 exd4 ! ? 1 4 i.xd4
i.e6 1 5 :ad 1 lDc5 1 6 i.xg7 �xg7
1 7 �h 1 f6 1 8 Wf4 i.xc4 1 9 :xd6
We7 20 :fd l :ad8 2 1 :xd8 :xd8
22 :xd8 Wxd8 23 e5 i.f7 24 exf6+
Wxf6 25 Wc7 Wd4 26 h3 lDxa4 27
lDxa4 Wxa4 28 We5+ �g8 29 lDe4
Wb5 30 lDf6+ �h8 3 1 We7 i.g8 ! 32
We8 Wc4 33 lDxg8 We t + 34 �h2
Wf4+ 35 �h 1 We 1 + 36 �h2 Wf4+
37 �h 1 112- 112 Gulko-Benj amin, Los
Angeles 1 99 1 . A simple but elegant
game.
2) 9 h4 e5 10 d5 (after playing h4
White must block the centre) 10 . . . h5
(although l O . . . lDa6 1 1 h5 Wb6 1 2
:a3 l2Jc5 also looks O K for Black it
is worth taking a time-out to fix the
kingside ; the only drawback is that
Black will have to concern himself
with i.xh5 sacrifices, but usually
these will not work) 1 1 i.g5 Wb6
12 :a3 ( 12 Wc2 l2Ja6) 12 . . . l2Jbd7 !
( 12 . . . lDa6? ! 1 3 i.xf6 i.xf6 1 4 i.xh5 !
gxh5 15 lDxh5 is one example of a
dangerous i.xh5 - note that the rook
on a3 is in position to swing) 1 3 Wc2
lDc5 14 i.e3 ( 1 4 i.xf6 and 15 i.xh5
is less effective now as White's queen
is worse on c2 and Black's knight is
already on c5) 14 . . . i.d7 ( 14 . . . l2Jg4 !
1 5 i.xg4 i.xg4 is better according to
Bologan; White must deal with the
threats of 1 6 . . . Wxb2 and 1 6 . . . Wb4)
15 lDfl :ac8 1 6 lDd2 Wb4 (now
Black is threatening to play . . .cxd5 ,
an exchange which he has carefully
avoided making until now in order
to deny White use of the c4- and
102 5 tl:Jge2

b5 -squares) 17 !l:Ja2 li'b6 1 8 !l:Jc3 is certainly the sort of move that has
li'd8 19 li'bl 'it>h7 ( 1 9 . . . !l:Ja6) 20 b4 a strong unsettling effect.
axb4 2 1 li'xb4 .ih6 22 .ixh6 'it>xh6 16 !i:Jb5! .if8
with an edge for Black, Ionov-Bol­ 17 .ig5
ogan, USSR Ch 199 1 . 17 lDd.6 .ieS looks excellent for
9 ••• e5! Black so White continues with his
10 dxeS unbalancing campaign.
10 fxe5 dxe5 1 1 d5 li'b6 ! looks a 17 ••. cxb5
little awkward for White 18 cxb5!
10 dxe5 Freeing c4 for his bishop was one
1 1 li'xd8 l:.xd8 of White's main ideas.
12 f5 18 !i:Jb4
White tries to make it difficult for 19 .ixd8 l:.xd8 (D)
B lack to develop his queenside and
doesn ' t allow him u se of the e5-
square but he is, nevertheless, still
balancing on the edge of a positional
prec1p1ce.
. .

12 ••• !l:Ja 6
13 0-0 !i:Jd7
14 .i e3 !i:Jdc5
15 l:.adl .id7 (D)

20 fxg6 fxg6?!
I was bluffed into this anti-posi­
tional continuation. After 20 bxg6 •••

2 1 .ic4 !l:Je6 22 .ixe6 fxe6 23 l:.f6

I was feeling quite content at this
stage of the game; the game plan was
one more sem i -developing move
( . . . .ieS) followed by gradual ly con­
verti ng my positional advantage.
Suddenly my opponent threw a giant
spanner in the works. Perhaps 16
!i:Jb5 is insufficient for equality but it
'it>g7 ! 24 l:.xe6 'it>f7 25 l:.xe5 'it>f6 the
white rook is trapped.
2 1 .ic4+ 'it>b8
22 l:.t7 l:.c8
I had been relying o n this move
but it seems insufficient. Probably
White has enough activity to com..
pensate for his slight material defi­
cit.
23 h3
A sensible decision as it's unclear
if White can equalise after 23 .:.Cxd7
!i:Jxd7 24 .ie6, e.g. 24 . . . !i:Jd3 ! ? 25
 5 llJge2 103

checks than from the ending after 27

l:txc l ? i.xc l 28 i.xd7 i.f4 29 h4
llJd3 ! w hen he is caught in an ex­
tremely nasty pin.
27 i.gl+
28 �hl i.d4+
29 �h2 i.g l +
30 �h l l:tel
Unfortunately for Black his knight
is completely cut off from the king­
side so he can' t increase the pres­
sure. The text is a rather half-hearted
attempt to continue the game.
31 i.c4!
Threatening to defend the king­
side. Black cannot afford to waste
24 l:tfxd7 ! i.e3+ any more time.
25 �h2 llJxd7 31 •.. i.cS+
26 i.e6 l:tcl?! 32 �h2
The last chance was to prevent 32 i.fl i.e3 ! (or 32 . . . i.f2) or 32
White's rook from becoming active llJO l:txe4 would be rather silly.
with 26 i.d4.
••• 32 ••. i.g l +
27 l:txd7! 33 �b l i.cS+
White correctly judges that there 34 �h2
is far less to fear from the discovered
8 U n usual Li nes

In this Chapter we examine a few
very rare lines. The material is split
up as follows:
A : 1 d4 lt:Jf6 2 c4 g6 3 l'Dc3 .ig7 4
There are many other moves that
Black can play but they are likely to
transpose to the Classical, not cov­
ered in this book. 6 lt:Jg4 leads to in­
•.•

e4 d6 5 lt:Jf3 0-0 6 .ie3 dependent play but I don 't really

B : 1 d4 lt:Jf6 2 c4 g6 3 lt:Jc3 .ig7 4 trust it.
e4 d6 5 .ie2 0-0 6 .ie3 (6 g4, 6 h4) 7 dxeS
C: 1 d4 lt:Jf6 2 c4 g6 3 lt:Jc3 .ig7 4 7 .i e2 transposes to the Gligoric
lt:Jf3 d6 5 e3 Variation.
7 ..• dxeS
A) 8 'ii'xd8 .:.Xd8
1 d4 lt:Jf6 9 liJdS :d7
2 c4 g6 9 llJa6 is an important alterna­
•••

3 lt:Jc3 .i g7 tive leading to the followi ng possi­

4 e4 d6 bilities:
s lt:Jf3 0-0 1 ) 10 :dl and now:
6 .i e3 (D) la) 10.. ..ig4?! runs into 1 1 .igS!
when 1 1 :d6 loses to 1 2 lt:Jxf6+
.•.

and 1 1 :xdS 1 2 cxd5 lt:Jxe4 1 3

•..

.ie7 lt:Jd6 14 .ixa6 bxa6 1 5 :c t was

very good for White in Toth-Morten­
sen, Thessaloniki OL 1 984.
1 b) 10 :rs!? 1 1 lt:Jxf6+ .ixf6
••.

1 2 a3 b6 1 3 .id3 .id7 14 �e2 :fd8

A favourite variation of Larsen
and Rivas. White's main idea is to
achieve a favourable version of the
Exchange Variation, but as we shall
see Black has nothing to fear in the
resulting ending.
6 •.• eS
1 5 h3 :ab8 1 6 b4 c5 was fine for
Black in Cifuentes-Nijboer, Wijk
aan Zee 1 99 1 .
l e) 10....te6 1 1 .ig5 .ixd5 1 2
cxd5 lt:Jc5 l 3 llJd2 h6 1 4 .ixf6 .ixf6
1 5 :c t .ie7 1 6 b4 lt:Jd7 1 7 :xc7
.ixb4 1 8 .ib5 (18 :Xb7 a5 19 f3 b8
is dangerous for White) 1 8 . . . lt:Jf6 1 9
.id 3 :ac8 20 :xc8 :xc8 2 1 �e2
lt:Je8 is = according to Hubner in
ECO.
2) 10 0-0-0 and now (D):

2a) 10 i.g4 leading to a further
.•.
branch :
2a l ) 1 1 h3 i.xf3 1 2 gxf3 lLld7 ! ?
( 1 2 . . .c6 is also perfectly playable) 1 3
h4 c6 14 ltJc3 i.f8 15 h5 ltJdc5 with
good play for Black, Bjarnason-Van
Wely, Lyngby 1 990.
2a2) 1 1 cS 4Jxd5 12 exd5 lLlb4
13 i.c4 b5 ! 14 i.xb5 ltJxa2+ 15 <iii b l
llab8 16 <iiix a2 llxb5 gave Black a
good game in Rivas-Blees, Amster­
dam 1 986.
2a3) 1 1 i. g5 (this looks the most
testing) 1 l . . . lld6 12 h3 i.xf3 13 gxf3
lle6 ( 1 3 . . . <iii f8 ! ?, intending . . . ltJg8
and . . . c6, has been suggested by
C .Hansen and 13 . . . h6 ! ? 14 4Jxf6+
i.xf6 1 5 i. xh6 llxd l+ 16 <iiix d l
ltJc5 17 h4 lld8+ 1 8 <iilc2 ltJe6 1 9
i.e3 ltJd4+ 20 i.xd4 llxd4 was quite
comfortable for Black, despite his
pawn minus, in Cifuentes-Geenen,
Belgium 1 993, though I have to say I
would be extremely reluctant to give
up a pawn like this) 14 h4 c6 15 lbc 3
( 1 5 4Jxf6+ i.xf6 16 i.h3 i.xg5+ 17
h xg5 lie? and 1 6 lid? ltJc5 1 7 llc7
i.xg5+ 1 8 hxg5 a5 are both comfort­
able for Black) 15 . . . ltJe8 ( 1 5 . . . llee8
deserves consideration, intending to
meet 16 lld6 with 16 ... lDh5) 16 i.h3
Unusual Lines 105
lld6 17 llxd6 ltJxd6 1 8 b3 ltJe8 1 9
lid 1 i.f6 20 ltJa4 ! with an edge for
White, Cifuentes-1.Sokolov, Wijk
aan Zee 1 993.
2b) 10.... i.e6 1 1 4Jxf6+ i.xf6 1 2
c5 ( 1 2 llxd8+ llxd8 1 3 a3 c5 14 i.e2
1h-•h J .Horvath-Groszpeter, Hungar­
ian Ch 199 1 ) 1 2 . . . llxd l+ 1 3 @xd l
lld8+ 14 <iilc l 4Jb4 15 i.c4 ltJd3+ 1 6
i.xd3 ( 1 6 <iii c2 lLlf4 1 7 h 3 i.d7 was
quite good for Black in Ubilava-Ku­
preichik, Kuibyshev 1 986) 1 6 . . . llxd3
1 7 <iiic 2 lld8 1 8 h3 i.d7 1 9 lLld2
i.b5 20 b3 i.e7 with an equal posi­
tion, L.Hansen-Kupreichik, Copen­
hagen 1 988.
10 4Jxf6+
10 ltJxeS? ltJxd5 1 l liJxd7 ltJxe3
wins for Black.
10 •.. i. x f6
1 1 c5 4Jc6
12 i.b5 (D)
12 ..• @f8 ! ?
12 lld8 i s more common. After
••.
13 i.xc6 bxc6 there is:
1 ) 14 0-0 llb8 (it is usually a
good idea for Black to force b3 so
that he doesn't have to worry about a
white rook sw inging to a3 later) 15
b3 i.a6 16 llfel i.g7 17 llac l h6 1 8
106 Unusual Lines
ltJd2 l:.b4 1 9 g4 i.d3 with equality,
Renet-Zsu .Polgar, Brest 1987 .
2) 14 l:tdl .i.a6 ! 15 l:txd8+ l:.xd8
1 6 liJd2 i.e7 17 g4 l:.b8 1 8 b3 f6
with an edge for Black, Barbero­
Khalifman, Plovdiv 1986.
3) 14 llxl2 i.a6 ( 14 ... l:.b8 15 0-0-0
i.e6 1 6 b3 i.e7 17 liJbl f5 1 8 f3
l:.xd l + 1 9 @xd l fxe4 1/2- 1'2 Rivas­
Lukin, Leni ngrad 1984 ) 1 5 0-0-0
i.e2 1 6 l:.de I i.d3 17 f3 a5 18 ltJbl
i.h4 19 g3 i.e7 was about level in
Rivas-J.Polgar, Madrid 1 993.
13 @e2 @e7
14 a3
White doesn ' t really know what
to do� he is reluctant to play i.xc6 as
Black will have saved the tempo
which he normally spends on unpin­
ning and, especially after his 1 3th
move, . . . i.a6 will prove annoying.
14 i.g7
15 i.c4 h6
16 h3 ltJd8
17 b4 ltJe6
18 g3 l:.d8
19 l:.adl i.d7
20 h4 i.c6 (D)
Black has a slightly better ending
as 2 1 i.dS can be met by 2 1 . . . i.b5+
and 22 . . . c6. Kosten-King, Hastings
1990/1 continued 2 1 l:.xd8 l:.xd8 22
ltJd2 ltJd4+ 23 .i.xd4 l:.xd4 24 i.d3
a5 25 l:.bl axb4 26 axb4 i.d7 27 f3
i.e6 28 b5 l:.a4 29 l:.b2 i.f8 30 l:.c2
@f6 3 1 c6 b6 32 i.c4 l:.b4 33 i.d3
i.c5 34 l:.c I i.d4 35 l:.c2 @e7 36
l:.c I l:.b2 37 l:tc2 l:.xc2 38 i.xc2
@d6 39 i.d3 @c5 40 @d I i.f2 4 1
ltJfl i.c4 42 i. xc4 @xc4 43 @e2
i.c5 44 g4 @xb5 and Black soon
won.
B)
1 d4 ltJf6
2 c4 g6
3 ltJc3 i.g7
4 e4 d6
5 i.e2 0-0
6 i.e3
A move which, for no particular
reason, has been completely ne­
glected by chess theory. One cou ld
argue that it is not inferior to 6 ltJf3
which has just had a 640-page work
written on it, while 6 i.e3 merits
25o/o of a footnote in ECO. There are
obvio u s similarities to the 6 . . . h6 7
i.e3 line in the Averbakh and al­
though there are some transpositions
it sh ould be in Black's favour to have
his pawn on h7 . White has a couple
of other unusual and aggressive al­
ternatives:
I ) 6 g4 a6 ! ? (Black signals his
preference for the Benko option,
very logical after a move like g4 as
White must be at least thinking about
castling long) 7 g5 (the reason for
Black delaying . . . c5 is to enable him
to switch plans after 7 a4, which can
be met by 7 . . e5 ! 8 d5 a5 ! ) 7 . . . liJfd7
.

(7 . . . ltJh5 led to very wild play after S
i.e3 b5 9 i.xh5 gxh5 1 0 1i'xh5 llJc6
1 l lLlge2 lLlb4 12 <ii'd 2 c5 in Bareev­
Dj uric, Bled 1 9 9 1 ) S i.e3 (S a4? !
llJc6 9 i.e3 e5 1 0 d5 lLJd4 is good for
Black) S . . . c5 9 d5 b5 ! 1 0 lLlf3 ! ( 1 0
cxb5 axb5 1 1 ltJxb5 ? i. xb2 1 2 l:.bl
1i'a5 + ! and 1 1 i.xb5 i.a6 12 i.xa6
lLlxa6 1 3 lLlf3 1i'b6 14 1i'e2 l:.fbS
with a very good Benko are vari­
ations given by Lanka) 1 0 . . . ltJb6 ! 1 1
cxb5 axb5 1 2 i.xb5 i.a6 1 3 1i'e2?
( 1 3 1i'd3?? loses at once to 1 3 ... c4 ! ,
1 3 a4 i.xb5 1 4 axb5 l:.xal 1 5 Wxal
e6 is very good for Black and 1 3
1i' b3 is well met by 1 3 . . . WcS ! , al­
though this is probably the best that
White can do) (D)
1 3 . . . ltJa4 ! ! (a brilliant and unex­
pected tactical shot) 1 4 i.d2 (1 4
lLlxa4 1i'a5+ 1 5 i.d2 1i'xb5 16 1i'xb5
i.xb5 1 7 ltJc3 i.d3 ! is disastrous
for White) 1 4 . . . lLlxc3 1 5 bxc3 ( 1 5
i.xc3 ? loses to 1 5 . . . i.xc3+ 1 6 bxc3
1i'a5 ) 1 5 . . . i.xb5 ! 16 Wxb5 e6 (the
white position is riddled with weak­
nesses and his king has no home) 17
0-0 ( 1 7 dxe6 fxe6 l S 1i'e2 d5 19 0-0
l:.a4 ! demonstrates the energy in
Black's position) 1 7 . . . exd5 lS exd5
Unusual Lines 107
WcS ! 19 l0h4 1i'h 3 ! 20 lLlg2 lLJd7 21
f4 l:.abS 22 1i'e2 l:.b2 23 l:.fbl i.xc3 !
24 l:.xb2 i.xb2 25 l:.bl i.d4+ 26
i.e3 ? ! (26 <ii' h l h6 !) 26 . .. :es 27
l:.b3 1i'f5 ! 2S 1i'c4 l:.xe3 ! 29 ltJxe3
(29 l:.xe3 lLlb6 ! ) 29 . . . 1i'xf4 30 1i'e2
c4 ! 3 1 l:.a3 c3 ! 32 <ii'g 2 i.xe3 33
l:.xc3 1i'xg5+ 34 <ii'h 3 (34 <iti'f3 lLle5 +
35 <ii'e 4 f5#) 34 . . . i.c5 ! 35 Wes+
<ii' g7 36 1i'xd7 1i'h5+ 37 <ii'g 2 1i'e2 +
3S <ii'h 3 Wn + 39 <ii'h 4 1i'f6+ 0-1
Hort-Lanka, Manila OL 1 992.
2) 6 h4 ltJc6 (I don' t suppose
there is much wrong with 6 . . . c5 or
6 . . . e5 either) 7 i.e3 (7 d5 lLle5 , un­
clear, is better) 7 . . . e5 S d5 lLld4 ! 9
lLlh3 (9 i.xd4 exd4 10 1i'xd4 l:.eS 1 1
1i'd3 1i'e7 1 2 f3 ltJh5 is good for
Black) 9 . . . c5 1 0 dxc6 bxc6 1 1 ltJg5
d5 ! 1 2 cxd5 cxd5 1 3 exd5 lLlxd5 14
lLlxd5 1i'xd5 with advantage to Black,
Alvarez-Palacios, Seville 1 992.
6
••• es
6 c5 is not a bad move, but as the
••.
6 ... h6 7 i.e3 c5 line in the Averbakh
is not part of our repertoire I can't re­
ally recommend it.
7 dS lLla6
7 c6 S g4 cxd5 9 cxd5 1i'a5 1 0
.•.
<ii' f l ( 1 0 i.d2 ltJa6 1 1 h4 lLlc5 12 f3
1i'b6 1 3 l:.b l a5 14 i.e3 h5 1 5 ltJh3
a4 1 6 ltJf2 1i'a5 was unclear i n
Sadler-McDonald, British Ch (East­
bourne) 1990) l O . . . lLla6 1 1 a3 i.d7
1 2 g5 ltJeS (Black ends up in a pas­
sive position after this but it would
take a brave man to play 1 2 . . . ltJh5 )
1 3 b4 1i'dS 14 h4 f6 1 5 ltjf3 i.g4 1 6
l:.c l ltJac7 1 7 ltJd2 i.xe2 l S Wxe2
w ith an edge for Wh ite as Black's
minor pieces have very little scope,
Sadler-Krasenkov, Pamplona 1 990.
108 Unusual Lines
8 g4 ltJcS
9 f3 a5 (D)
10 h4
10 1i°d2 h5 ! ? ( 1 O . lDe8 1 1 h4 f5
. .
12 g5 .id7 13 ltJh3 c6 1 4 exf5 ! .ixf5
1 5 ltJf2 was a little better for White
in Conquest-Gallagher, Douai 1 993
and 10 ... .id7 1 1 h4 h5 12 g5 ltJh7 1 3
ltJh3 is an inferior version of the
main line) 1 1 h3 ltJh7 1 2 0-0-0 h4 1 3
l:.el .if6, with ... .ig5 to follow, was
about equal in Marjanovic-Martin­
ovic, Belgrade 1 977 .
10 ... h5
1 1 g5 ltJh7
This is better than 1 1 ltJfd7 and
...
1 1 ltJeS (both of which have been
•.•
played against Sadler, the main cham­
pion of this system) as Black must
seek immediate counterplay with
. . .f6 to avoid being squashed. We
have, incidentally, now transposed
into an old line of the Averbakh.
12 ltJh3 f6
12 c6 13 1i°d2 cxd5 1 4 ltJxd5 l:.a6
•••
1 5 ltJ f2 ltJe6 1 6 0-0-0 l:.c6 1 7 q;b 1
was favourable for White in F.Olafs­
son-Keene, Reykjavik 1 976 but the
player of the black pieces obviously
wasn't idle in Reykjavik as the main
l ine is taken from a game played a­
few rounds later in the same tourna­
ment.
13 1i'd2 .ixh3
14 l:.xh3 fxg5
15 hxg5 l:.f4!
An excellent, if typical, exchange
sacrifice which has to be accepted as
otherwise Black w ill simply pick up
the g5-pawn.
1 6 .ixf4 exf4 (D)
17 0-0-0?
White had to try 17 1i°xf4, even if
after 1 7 . . . .ie5 1 8 1i'e3 ltJxg5 1 9 l:.h 1
1i'f6 Black still has a complete bind
on the dark squares.
After 17 0-0-0, the game Gunnars­
son-Keene, Reykjavik 1 976 contin­
ued l 7 . . . ltJxg5 1 8 l:.h2 1i°f6 1 9 l:.g 1
ltJf7 20 l:.hg2 ltJe5 with a position­
ally won game for Black. In retur n
for a minimal material investment
he has excellent outposts for both
knights, a powerful battery on the
long diagonal, two potential passed
pawns on the kingside and an a­
pawn which could cause White some
grief. Considering that Keene's hero
is Nimzowitsch I should not omit to
mention his blockade of the centre.

C)
1 d4 �f6
2 c4
2 �f3 g6 3 b4 is a weird move-or­
der often employed by the Swiss Th1
Kanel. One example is Kanel-Gal­
lagher, Villars 1 995 which continued
3 . . . i.g7 4 i.b2 d6 5 e3 e5 6 c4 (6
dxe5 �fd7) 6 . . . 0-0 7 i.e2 exd4 (on
several occasions I have reached
the position after 7 . . . �bd7 against
Kanel) 8 �xd4 c5 ! 9 �b5 �c6 1 0
bxc5 dxc5 1 1 0-0 i.e6 1 2 �d2 fie7
1 3 1i'c 1 l:tfd8 with an edge for Black.
Note the similarities with the Dreev­
Shirov game given below.
2
••• g6
3 �c3 i.g7
4 �f3 d6
5 e3 (D)
At first glance this looks like a very
passive system against the King's In­
dian but White wants a solid centre
and kingside as he has aggressive in­
tentions on the queenside.
5 ... 0-0
6 i.e2 �bd7
7 0-0
In Dreev-Shirov, Lvov 1990 White
delayed castling in favour of 7 b4.
Unusual Lines 109
There followed 7 . .. e5 8 i.b2 exd4 9
�xd4 c5 ! ? 1 0 bxc5 dxc5 1 1 llJdb5
�b6 12 1i'h3 �g4 1 3 h3 �e5 14 0-0
i.e6 1 5 l:tad l 1i'h4 1 6 �c7 �xc4
1 7 �xe6 �xb2 1 8 �xf8 and the
players agreed to a draw in view of
1 8 . . . �xd l 1 9 �xg6 hxg6 20 l:txd l
with a roughly level endgame.
7
- · e5
8 b4
8 b3 is more solid. ECO gives
8 . . . l:te8 9 i.a3 e4 (9 . . . exd4 10 �xd4
�c5 1 1 1i'c2 �fe4 1 2 �xe4 �xe4
1 3 i.b2 a.5 14 �b5 i.xb2 1 5 1Wxb2
b6 = Barcza-Bolbochan, Helsinki
OL 1 952 ) 10 llJd2 �f8 1 1 1i'c2 i.f5
12 i.d l c5 13 d5 �6d7 14 i.b2 1i'g5
1 5 i.c2 �f6 with an unclear game,
Barcza-Planinc, Ljubljana 1969.
8
•• • l:te8
9 a4 exd4
9 e4 is doubtful si nce White's
.• .
queenside attack is already well un­
der way.
10 exd4
10 t0xd4 is met by 10 . . . c5 .
10
.. . c5
Black can react in several differ­
ent ways to White's queenside pawn
armada. 10 dS 1 1 1i'b3 is perhaps a
...
l ittle better for White while 1 0 aS ...
1 1 b5 �b6 was my choice in the
game Wirthensohn-Gallagher, Wa­
hlen 1993, the idea being to dissuade
White from playing c5 as then Black
will be able to occupy the d5-square.
Play continued 12 i.f4 i.f5 1 3 !:te l
h 6 14 h3 g5 1 5 i.h2 �e4 1 6 �xe4
i.xe4 17 c5 dxc5 ! ? 1 8 l:txc5 c6 1 9
bxc6 bxc6 20 �e5 �d7 ! with com­
plications that were not unfavour­
able for Black.
110 Unusual Lines

1 1 :bl cxb4
12 :xb4 liJb8!? (D)
A grandmasterly move. Black
repositions his knight in order to cre­
ate some play against the hanging
pawns. Spiridonov-Hort, Brno 1 975
continued 1 3 h3? ! ltJc6 1 4 :b5 ltJa5
1 5 i.e3 b6 1 6 :b4 i.a6 1 7 d5 ltJg4
1 8 i.g5 fllc7 1 9 liJb5 i.xb5 20 axb5
ltJf6 2 1 i.e3 :xe3 ! 22 fxe3 f//c5 23
:b 1 fllxe3+ 24 �h2 ltJh5 with a
clearly better game for Black.
9 The Trom pows ky
The last ten years or so have seen the
Trompowsky develop into a fairly
respectable system. Much of the
credit for this belongs to English
Grandmaster Julian Hodgson who
for years never played anything else.
Even in top tournaments where his
opponents had days to prepare for
him he still managed to prove that
the Trompowsky can be a dangerous
weapon . I can recall very well his
initiation into the Trompowsky. In
1 985 Julian, Mark Hebden and I
took part in a small round-robin tour­
nament in Alicante, Spain. Amongst
the opposition were the then un­
known Spaniards Illescas and de la
Villa, who turned out to be Trom­
powsky fanatics. As five of our six
games against this pair were with
Black many of our nights were spent
digging deep in search of the Trom­
powsky 's secrets, although it has to
be admitted that our duty-free stock
also took a terrible battering durin g
these sessions. Some incredible
ideas were found, even if they didn't
all look quite so promising by the
following afternoon. At any rate, be­
fore the end of the tournament Julian
was convinced that the Trompowsky
was the opening for him and the rest
is public knowledge . Even I took it
up for a year afterwards with excel­
lent results before finally getting a
bit bored playing the same positions
all the time. In fact the popularity of
the Trompowsky is largely due to the
fact that it eliminates the need to
learn masses of theory on the King's
Indian, Nimzo-lndian, Queen's In­
dian and many other lines. For some
time Hodgson couldn't even be both­
ered to learn how to play the Queen's
G ambit and started to play 2 i.g5
in response to 1 . . d5 . . . which hap­
.
pens to remind me of one of my
weirdest ever games. I could say that
what follows is a good example of
the strategically complex and tacti­
cally uncompromising struggles that
arise from Trompowsky-style posi­
tions, but that would be a load of old
Chapter 1 4. In reality I have always
slightly regretted that this game has
never been published so I' m going to
seize my chance despite its irrele­
vance: G allagher-Crouch, Notting­
ham 1 987 1 d4 d5 2 i.g5 f6 3 i.h4
lZJh6 4 e3 lZJf5 5 i.g3 h5 6 i.e2 h4 7
i.h5 + �d7 8 i.g4 e6 9 i.f4 g5 1 0 e4
dxe4 1 1 i.c l ! (D).
This is the position I really
wanted to show you. I bet your club­
mates won' t be able to guess how
this came about ! White has good
compensation for a pawn and the
remaining moves provided even
more fun (for me) : l l . . . �e7 1 2 c3
'ii'd5 1 3 lZJh3 tlJci6 1 4 0-0 i.d7 1 5 b3
lZJc6 1 6 i.e3 b5 1 7 a4 llb8 1 8 axb5
'ii' x b5 1 9 lZJd2 lZJf5 20 lZJxe4 'ii'x b3
1 12 The Trompowsky
2 1 ti'f3 �f7 22 i. h5 + �g7 23
ltJhxg5 fxg5 24 ti'g4 i.e7 25 i. xg5
�f8 26 i.xe7+ ltJcxe7 27 ti'g5 ! :h6
28 g4 ! hxg3 29 fxg3 llxh5 30 ti'xh5
i.c6 3 1 lLlg5 �g7 32 ti'h7 + �f6 33
h4 i.e8? 34 ltJe4 mate. OK, let's
quickly get back to the repertoire be­
fore the editor notices [Eh ? What
was that? - editor's note].
After l d4 ltJf6 2 i.g5 Black has a
large number of moves but I think
we can rule out 2 . .. e6 and 2 . . . d5 as
being inappropriate for King's In­
dian players . 2 . . . c5 is interesting but
my preference is for the most critical
line, 2 . . . llJe4, mainly because I don' t
think you should let Trompowsky­
ites play i. xf6. White normally re­
plies 3 i.f4 and this, along with 3
i. h4 and the eccentric 3 h4 are ex­
amined in the games below.
G ame 1 7
Bellon W. Watson
- Chiasso 1 99 1 after the moves 4 . . . d5
Hastings 198516
1 d4 lLlf6
2 i.gS ltJe4
3 i. h4 (D)
This retreat was once quite popu­
lar but is now rarely seen. This is no
doubt due to a number of heavy de­
feats which White suffered in the
latter half of the 1 980s. The draw­
back of i. h4, as opposed to i.f4, is
that Black is able to force the ex­
change of knight for bishop and con­
sequently obtains counterplay on the
dark squares. This may sound con­
tradictory to those of you who have
just read the i ntroduction to this
chapter where I stated that 3 i. xf6
should be avoided, but the difference
there is that the pawn structure
makes it hard for Black to achieve
active play.
3 .•. gS
3 cS 4 f3 g5 5 fxe4 gxh4, trans­
•.•
posing back into the game, is an
equally valid move-order.
4 f3
4 i.g3 would be an admission of
defeat whilst 4 ti'd3 led to an active
game for Black in JokSic-Gallagher,
5 f3 gxh4 6 fxe4 dxe4 7 ti'xe4 c5 !
(the initiative is what counts in such
positions) 8 dxc5 ltJc6 9 c3 i.h6 1 0
ltJf3 (not 10 ti'xh4? i.c l ! winning
material; this is an extremely impor­
tant trick which occurs time and time
again in this variation) 9 . . . h3 1 0 g4

.i.c I ! I I 'iic 2 .i.e3 I 2 b4 ( 1 2 .i.xh3
h5 !) I 2. . . 'tid5 ! ? ( 1 2. . ..i.xg4 I3 'iie4)
I 3 'iid 3 'tixd3 I 4 exd3 .i.xg4 I 5
ltJbd2 0-0-0 1 6 b5 ltJa5 17 i.e2
.i. xf3 I S ltJxf3 llhgS +.
4 ... gxh4
S fxe4 cS
S ...eS is the move that we spent
most of our time on in Alicante, but it
is less trustworthy than 5 . . .c5 . After
6 ltJf3 (6 e3 'tig5 is quite good for
Black) Black has (D):
I ) 6...exd4 7 'iix d4 :gs S 'tie5 +
'tie7 (as S . . . i.e7 9 'ii h5 looks good
for W hite it is preferable to sacrifice
the c-pawn) 9 'tixc7 and now:
I a) de la Villa-Gallagher, Alicante
I 9S5 continued 9 ltJc6 I 0 ltJc3
...
i.g7 I I liJd5 'iixe4 I 2 0-0-0 ltJb4 I 3
ltJxb4 'tixb4 I 4 c3 'tia4 I 5 e3 'tixa2
I 6 .i. b5 .i.f6 ! I 7 .i.xd7+ .i.xd7 l S
'tixd7+ �fS I 9 'iid6+ i.e7 20 'tid2
:cs 2 I ltJd4 :g6 with quite a good
game for Black, but I didn't feel con­
fident enough to repeat this in a later
round against Illescas. Perhaps I 5
:d3 would have posed more serious
problems.
I b) I n a l ater game, Keitling­
haus-Knaak, Bundesliga I 99 1 Black
The Trompowsky 1 13
opted for 9 ltJa6 and after I 0 'tic4
.•.
b5 ! ? I I 'tid5 ltJc7 I 2 'iid 3 .i.b7 I 3
ltJc3 .i.h6 I 4 ltJd4 :gs ! (Black pre­
pares to meet I 5 ltJf5 with I 5 . . . :xf5
leaving his fate in the hands of his
powerful bishops) I 5 e3 b4 I 6 ltJcb5
ltJxb5 I 7 ltJxb5 d5 ! I S 0-0-0 a6 I 9
ltJd4 dxe4 20 'iib3 the game was un­
clear. Black's pieces are more active,
but his pawns are ragged.
2) 6 .i.h6! ? (Black parts com­
.••
pany with a central pawn in order to
activate his bishops as quickly as
possible) 7 ltJ xe5 d6 S ltJf3 0-0 9
ltJc3 f5 I O exf5 .i.xf5 I I e4 .i.g4 1 2
.i.e2 h3 I 3 g3 l2Jc6 (I think we had
this position on the board the night
before the game and had concluded
that Black stood very well, but we
had underestimated White's next
move) I 4 liJd5 ! .i. xf3 I 5 .i.x f.3 ltJx d4
I 6 'iix d4 llxf3 I 7 �e2 :f8 I S llafI
with an edge for White, Illescas-Gal­
lagher, Alicante I 9S5 .
3) I seem to recall 6 ltJc6! ? be­
.••
ing tried by S peelman, and this may
be a more effective sacrifice since
White's d-pawn is forced to advance
if he wants to collect his booty.
6 e3 .i.h6
The assault on the dark squares
commences.
7 �f2
The only way to protect the pawn
as 7 'iif3 'iib6 is out of the question.
White has also tried 7 .i.c4, with the
obvious point 7 i.xe3?? S 'ii f3 ,
..•
but 7 e6 looks quite promising, e.g.
..•
S 'iih5 'tig5 (S . . . .i.xe3 9 'tixc5) 9
'tixg5 .i. xg5 I O ltJc3 ( 1 0 �f2 cxd4
I I exd4 .i.c I ! ) I O . . . .i. xe3 I I ltJb5
�dS 12 liJf3 a6 I 3 ltJd6 �e7 I4 e5
114 The Trompowsky
( 1 4 lt:Jxc8+ llxc8 1 5 d5 is better ac-
cording to S chmidt, but 1 5 . . . b5 still
looks good for B lack) 1 4 . . . cxd4 1 5
c3, Bellon-Schmidt, Biel 1 990, and
now 1 5 . . . dxc3 1 6 bxc3 lt:Jc6 would
have been the simplest, with an ex­
cellent game for Black.
7 ••• cxd4
8 exd4 iib6
.
8...eS t ? 1s
. . an mteresting new 1'dea.
• .
Voloshin-Golubev, Alusta 1 993 con­
tinued 9 lt:Jf3 lt:Jc6 1 0 c3 ( 1 0 d5
'1Vb6+) 1 O d6 1 1 lt:Jbd2 ( 1 1 i.b5 0-0
. . .
1 2 '1Ve2 f5 ! 1 3 lld l ? fxe4 1 4 '1Vxe4
d5 and White resigned in Espin
Martinez-S orin, Benidorm 1 992 as
he loses a piece) 1 1 . . . i.g4 ( 1 1 . . . 0-0
followed by . . . f5 looks tempting here
as well) 1 2 h3 i. h5 1 3 i.e2 '1Vf6 1 4
d5 lt:Je7 1 5 llfl lt:Jg6 and Black had a
firm grip on the dark squares as well
as attacking chances against the
white king.
9 lt:Jc3 e6!
Taking the b-pawn is a risky busi­
ness as the foJJ owing example dem­
onstrates: 9 '1Vxb2?! 1 0 lt:Jd5 �d8
.••
1 1 llbl '1Va3 1 2 '1Vh5 '1Vd6 1 3 lt:Jf3
'1Vg6? 1 4 '1Ve5 1 -0 Ker-Pomeroy,
New Zealand Ch 1 994. Anyway the
main purpose behind . . . '1Vb6 was to
create pressure along the important
g l -a7 diagonal which happens to
number the white king amongst its
occupants. The text, of course, main­
tains the position of the queen by
preventing tt:Jd5 .
10 lt:Jf3 lt:Jc6 (D)
1 1 lt:JbS?!
The start of an incredibl y opti­
mistic plan. White should probably
play i.b5xc6 in order to relieve the
pressure on his centre. On 1 1 e5,
l 1 . .. f6 ! 12 exf6 0-0 is recommended
by Vujacic in Jnformator but his as­
sessment of ' unclear' perhaps under­
estimates Black's chances. If Black
can play 1 3 . . . llxf6 he will have a
strong attacking position and the
only move to prevent this, 1 3 lLJe4, is
met simpl y by l 3 . . d5 .
.
11
... 0-0
12 c4 d6!
This makes it difficu lt for White
to drive back the black queen.
13 b4
13 lt:Jxd6 is strongly answered by
1 3 . . . e5 ! , for example 1 4 '1Vb 3 exd4
1 5 '1Vxb6 i.e3+ 1 6 �e l ax b6 1 7
lt:Jxh4 lt:Jb4 1 8 �d 1 i.g4+ ! 1 9 i.e2
i.d7 20 lt:Jb5 d3 2 1 i.fl i.x b5 22
cxb5 llfc8 with a winning position
for Black, Aleksandrov-Loginov,
Kstovo 1 994.
13
.•• '1Vd8
13 lt:Jxb4 1 4 llbl a6 has been
•..
recommended, but I prefer Watson's
choice as he is now ready to play
. . . e5 without having to worry about
the reply c5 .
14 llbl
14 a3 would have avoided the
game cont inuation but Black would
 The Trompowsky 115

still be able to open the position by 21 exf5 �xf5

playing . . . e5 followed by . . . f5 . 22 :n 'ii'xa2
14 ..• e5! 23 'ii'xa2 :xa2
15 d5 (D) The queen exchange offers White
1 5 dxe5 'ii' b6+ is also very good no relief whatsoever.
for Black. 24 c5
White has no t ime to take for 24
:xb7 as 24 . . .:al + 25 �f2 :xn + 26
�xfl e4 is the end.
24 ..• �e3
This tightens the noose around the
white king, but 24 .Lt+ would have
..

been good enough as well.

25 ltJxh4
25 cxd6 fails to 25 . . . :al + 26 �d l
�d3.
25 ..• :at+
26 �di �g4!
15 •.• ltJxb4! 27 ltJf3 e4
Black u tilises a little tactical trick 28 :b4 :r4!
to open the queenside which will 29 cxd6 exf3
make it much easier for him to get at 30 :x14 12+
the white king. 0·1
16 :xb4 1Wb6+
17 �e l a6 Game 1 8
18 1Wb3 Hodgson Nunn -

18 :a4 �d7 ! 1 9 ltJc3 'ii'e 3+ 20 English Ch 1991

ltJe2 �xa4 2 1 'ii'x a4 'ii'xe4 is not an
improvement. 1 d4
18 .... axbS 2 �gS
19 :xb5 'ii'a7 3 h4 (D)
I think Black has at least two
pawns' worth of compensation here
but he isn't even any material down.
20 �e2
20 c5 doesn ' t work: 20 . . . �d7 ! 2 1
:xb7 and now both 2 1 . 'ii'a5 + and the
. .

piece sacrifice 2 1 . . . 'ii' xc5 are very

good for Black.
20 •.• f5 !
Now a dangerous passed e-pawn
can be added to Black's list of b11mps.
116 The Trompowsky
The ' h4 Tromp' , introduced as a
stopgap between the fading 3 .i. h4
and the rising 3 .i.f4, has been the
subject of much ridicule and laugh­
ter over the y ears. From a purely
chess point of view it is hard to be­
lieve in this line but in practice White
has actually scored quite well. Per­
haps this is due to the psychological
impact that a move such as 3 h4 can
have on the opponent. On seeing 3
h4 Black probably begins to feel
very confident (how can he play such
rubbish?) but he may also experience
some difficulty in talcing the game
seriously; not a good combination. I
can recall one of the first games in
this line, Hodgson-Gufeld, Hastings
1 986/7 . Black played the opening
quite well and by move 1 5 was a
pawn up for nothing. But soon after­
wards things started to go wrong;
he castled queenside rather riskily
and White was able to sacrifice a
piece for unclear complications after
which ' Big Eddie' proved to be no
match for a Hodgson in his element,
eventually overstepping the time
limit in a lost position. What made
this game so memorable though is
the reaction of Gufeld to his defeat.
For the next hour he remained alone
on the stage, a tragic figure with his
head clasped in his hands and for the
rest of the tournament he could be
heard explaining to anyone who
would listen that this game was not
chess and how disgraceful it was that
someone could conduct a game in
the way Hodgson had just done.
My main recommendation against
the ' h4 Tromp' , apart from staying
calm, is 3 . ..c5 but I have also given a
summary of the theory of 3 . . . d5 ,
w hich may be less dynamic than
3 . . .c5 but is probably the simplest
way to equalise.
3 cS...
3 dS 4 ltJd2 with several possi­
.•.
bilities:
1 ) 4 ltJxgS S hxgS .i. fS (5 . . . c5 is
.••
interesting, leading to an unclear
game after 6 dxc5 e5 7 e4 ! ) 6 e3 and
now:
1 a) The aforementioned Hodg­
son.. Gufeld, Hastings 1 986/7 contin­
ued 6 e6 7 g4 .i.g6 8 f4 c5 (8 . . . lLld7
. ..
9 fi'f3 .i.xc2 10 llc 1 .i.a4 1 1 .i.d3 g6
1 2 f5 gxf5 1 3 gxf5 fi'xg5 14 ltJh3
fi'h4+ 15 ftjf2 ..g5 1 6 ltJh3 fi'h4+
1 7 ltJf2 fi'g5 I 8 ltJh3 i12- 112 Hodgson­
Zagrelbelny, Manila OL 1 992) 9
ltJgf3 (9 fi'f3 ! ?) 9 . . . l2Jc6 10 c3 fi'b6
1 1 ltJh4 ? ! ( I I 1i'b3 is more to the
point) I I . . fi'x b2 I 2 ltJxg6 fxg6 1 3
.
llbl fi'xc3 1 4 llb3 fi'a5 1 5 llxb7 c4
1 6 @f2 0-0-0 ( 1 6 . . .fi'xa2 ! is best) 1 7
llbl .i.b4 1 8 ltJxc4 ! dxc4 1 9 .i.xc4+
@c7 20 .i.xe6 llhe8 2 1 fi'b3 lld6 22
d5 + llexe6 ! 23 dxe6 lld2+ 24 @g3
fi'b5 25 llhc l a5? 26 fi'c4 ! fi'xc4 27
llxc4 .i.e7 28 llh I @d6 29 llxh7
@d5 30 llc l .i.a3 3 1 llb l .i.b2 32
llh2 ! llxh2 33 @xh2 .i.a3 34 f5 gxf5
35 gxf5 .i.e7 1 -0.
I b) 6 cS 7 g4 (7 dxc5 e6 8 ltJb3
•.•
.i.xc5 9 .i.d3 .i.xd3 I 0 fi'xd3 is con­
sidered as clearly better for White
by Kosic but I wonder if he took
I O . . . fi'xg5 ! into account, e. g. 1 1
fi'b5+ ltJc6 I 2 ltJxc5 fi'xg2 1 3 fi'xb7
0-0 14 fi'xc6 fi'xhl 15 O-O-O fi'h4 and
Black may even be better) 7 . . . .i.d7 8
g6! (a thematic pawn sacrifice in this

line) 8 . . . fxg6 9 i.d3 'ii' b 6 10 dxc5
'ii'f6 1 1 'ii'f3 e5 1 2 'ii'x f6 gxf6 1 3
i.xg6+ @e7 1 (an improvement on the
original h4 Tromp game, Depas­
quale-Kudrin, London 1986, which
went 1 3 . . . @d8 14 g5 ! with a clear
advantage for White) 1 4 f3 ltJa6 15
D.xh7+ D.xh7 16 i. xh7 i.h6 17 ltJfl
ltJxc5 and Black had some pressure
in return for his pawn, M .Hansen­
Fedorov, Tastrup 1992.
2) 4 ltJxd2 5 i. xd2 occurred in
.••
the game Hodgson-Hebden , Candas
1 992. One may look at the position
in the following way: from the start­
ing position give White the move d4
and Black . . . d5 , remove a knight
from each player and then offer
White a couple of free moves� not
many players would choose i.d2
and h4. However, Hebden now opted
for S eS?! which is exactly the sort
. ••
of reaction that the ' h4 Tromp' lures
people into. After 6 dxe5 ltJc6 7 ltJf3
i.g4 8 i.g5 i.e7 9 'ii'd2 'ii'd7 10 0-0-0
0-0-0, the simple 1 1 e3 (instead of 1 1
'ii'f4? f6 1 ) would have left White on
top. A more natural course for Black
to follow would have been to play
. . . e6 and . . . c5 , perhaps even without
. . . i.f5 w hich does rather invite
White to expand on the kingside. In
my opinion this type of 'French'
bishop is often just as well placed in­
sid e the pawn chain.
3) 4 i.rs (D) is the most solid,
...
no-nonsense approach to the 'h4
Tromp' and is responsible for Hodg­
s on's solitary defeat in this line.
White has:
3a) 5 ltJxe4 i. xe4 6 f3 h6 7 fxe4
(7 i.f4 is less risky, but after 7 . . . i.h7
The Trompowsky 1 1 7
8 e3 e6 9 i.d3 i.xd3 10 'ii'x d3
ltJd7 ! ? l l ltJe2 i.e7 12 g3 c5 Black
had an active game in Kosic-Droz­
dov, Bukovice 1 993) 7 . . . hxg5 8 'ii'd3
(8 ltJf3 ! ?) 8 . . . e6 9 'ii' b5 +? ltJc6 10
'ii'xb7 ltJb4! (perhaps White had only
considered 1 O . . ltJxd4, after which
.
1 1 0-0-0 would give him good play
for the pawn) 1 1 'ii'b5+ 'ii'd7 1 2 'ii' b7
'ii'c8 13 'ii'b 5+ 'ii'd7 14 'ii'b 7 'ii'c 8 15
'ii' b5+ c6 16 'ii'a4 dxe4 with advan­
tage to Black, Hodgson-S alov, Wijk
aan Zee 1 993.
3 b) 5 e3 h6 6 i.f4 e6 7 g4 i.h7 8
ltJxe4 i.xe4 9 f3 i.h7 10 i.d3 i.xd3
1 1 'ii'xd3 c5 ! 12 i.xb8 ! ? %txb8 1 3 f4
cxd4 1 4 exd4 i.d6 15 ltJe2 h5 with a
double-ed ged game, Hodgson-Be­
liavsky, Groningen 1 994 .
4 dxcS
In his most recent outings with
the 'h4 tromp' Hodgson has pre­
ferred 4 dS, after which there is:
1 ) 4 'ii'b6 (risky) 5 lZJd2 ltJxg5 6
...
hxg5 'ii'x b2 7 e4 with good compen­
sation for the pawn.
2) 4 ltJxgS 5 hxg5 g6 (not just
...
with the intention of placing the
bishop on its best diagonal, but also
with the id ea of preventing any awk­
ward g6 pawn sacrifices from White)
118 The Trompowsky
6 lbc3 d6 7 a4 i.g7 8 1i'd2 'iib6 9 l:.a2
liJa6 1 0 e4, Hodgson-Adams, Wijk
aan Zee I 993 , and now I like the
look of I O . . . liJb4, e.g. 1 1 l:.a3 e6 12
i.b5+ �f8 13 dxe6 i. xe6 with an
active game for Black. However, it
wouldn't surprise me if White could
improve on his 7th or 8th moves.
3) 4 g6 S 1i'd3 and now :
•••
3a) S ...lDxgS 6 1i'c3 ! ? f6 (6 . . . l:g8
7 hxg5 i.g7 8 1i'b3 c4 9 1i'a3 b5 1 0
l:.xh7 was very messy in Hodgson­
P.Cramling, Bern 1 992 but White
does have an extra pawn) 7 hxg5
i.g7 8 liJd2 (perhaps White should
play 8 gxf6 in order to force Black to
recapture with the bishop) 8 . . . d6 9
gxf6 exf6 ! (9 . . . i.xf6 may seem more
natural, but after the text Black has a
much healthier pawn structure) 10
1i'g3 (threaten ing l:xh7) 1 0 ... 0-0 ! 1 I
1i'h4 h6 1 2 1i'g3 g5 with an excellent
game for Black, Hodgson-Gufeld,
London rpd I 995.
3b) S 1i'a5+ (this avoids Hodg­
...
son's 1i'c3 idea, although as we saw
above it's not clear if it's worth pre­
venting) 6 liJd2 liJxg5 7 hxg5 i.g7 8
c3 (an important move, limiting the
scope of Black's strong bishop)
8 . . . d6 9 e4 liJd7 I 0 a4 l:.b8 I I liJc4
1i'c7 I 2 f4 a6 ! ( Black must open the
queenside as quickly as possible in
order to try to punish White for mov­
ing so many pawns) I 3 a5 b5 I 4 axb6
liJxb6 I 5 1i'c2 e6 I 6 dxe6 i. xe6 17
ltJe3 and now Black played I 7 ... i.d7
in Hodgson-Emms, British Ch I 992,
drawing the sting from any f5 by
White and preparing to pressure the
e-pawn with . . . i.c6. This, though, is
a rather slow idea an d it could well
be worth investigating the razor­
sharp I 7 . . . d5 ! ?.
4 ..• 1i'a5+
As one would expect, the theory
of this line is still largely undevel­
oped and there is no clear consensus
as to what Black's best continuation
is; therefore, I have examined all the
main alternatives below:
I ) 4 liJa6 (D) has been Black's
.••
most popular choice but according
to Adams, in lnformator, Hodgson
considers it inferior to 4 . . .1i'a5+.
White has now tried a couple of
moves:
l a) S 1i'd4 liJaxc5 (5 . . . 1i'a5+ 6
liJd2 liJxg5 7 hxg5 1i'xc5 , as played
in Miladinovic-Adams, Moscow OL
I 994, also looks comfortable for
Black) 6 ll:\c3 liJxc3 7 1i'xc5 liJe4 8
1i'd5 liJxg5 (8 . . .liJf6 9 i.xf6 gxf6 1 0
0-0-0 d6 1 1 e 4 a 6 led to a fairly typi­
cal S icilian position w ith chances
for both sides in Kosic-S hipov, Bel­
grade I 994) 9 hxg5 1i'b6 I O 0-0- 0
1i'xf2 1 1 liJf3 ( I I 1i'e4 is suggested
by Miladinovic but it's hard to be­
lieve that White has enough for the
pawn after something like I I . . . 1i'b6
I 2 liJf3 g6 when 13 1i'e5 can simply

be met by 1 3 . . .f6 1 4 gxf6 'ii'x f6)
l 1 . .. 'ii'e 3+ 1 2 <it'bl e6 13 'ii'c4 d5 1 4
'ii' b 5+ .i.d7 1 5 'ii'x b7 llc8 1 6 llJd4
.i.c5 1 7 :h3 'ii'e5 1 8 llc3 <it'e7 1 9 e4
:c7 with a clear advantage for Black,
Kosic-Kiselev, Yugoslavia 1 993 .
1 b) S ltJd2 (this seems the better
option) s ltJaxc5 with a further
.. .
branch:
1 b l ) 6 ltJ xe4 ltJxe4 7 'ii'd4 'iia 5+
8 c3 ltJxg5 (8 . . . d5 ! ? has been sug­
gested here) 9 hxg5 'ii'xg5 10 e4 'ii'a5
1 1 ltJf3 and somewhat surprisingly
White seems to have good play for
the pawn. For example: l l d 6 1 2 ...
e5 ! dxe5 1 3 ltJxe5 .i.e6 1 4 b4 ! 'ii'a4
1 5 :d 1 with a very strong attack
for White, Depasquale-Lanka, Mel­
bourne 1 99 1 ; perhaps 1 1 'ii'b 6 can
••.
be tried, but after 1 2 'ii' x b6 axb6 1 3
ltJe5 e6 1 4 .i. b5 ! .i.d6 1 5 l2Jc4 .i.c 7
1 6 e5 Black will find it difficult to es­
cape the bind, at least with his mate­
rial advantage intact.
1 b2) 6 ltJgf3 'ii' b 6 7 ltJxe4 ltJxe4
8 'ii'd4 'ii'xd4 9 ltJxd4 a6 1 0 g3 e5 1 1
ltJb3 d5 1 2 .i.g2 has occurred several
times and it appears that White
might have an edge in this ending,
for example 12 ltJxgS 1 3 hxg5 .i.e6
..•
1 4 0-0-0 (better than 1 4 f4 0-0-0 1 5
<it'f2 .i.e7 1 6 llad 1 d4 1 7 .i.h3 .i.xh3
1 8 llxh3 f6 with at least equality
for Black, Adams-Lesiege, Oakham
1 992) 1 4 0-0-0 1 5 .i.h3 ! or 12 fS
. . . .•.
1 3 .i.xe4 fxe4 1 4 c3 b6 15 0-0-0 .i.e6
1 6 f3 exf3 1 7 exf3 .i.d6 1 8 f4, Hodg­
son-Hebden, Cappelle la Grande
1 992, and in both cases White had
slightly the better chances.
2) 4 ltJxgS 5 hxg5 e6 6 ltJf3
•.•
.i. xc5 7 e3 (7 l2Jc3 1!Wb6 8 llJe4 'iixb2 !
The Trompowsky 119
9 ltJxc5 'ii'c 3+ 1 0 ltJd2 'ii'xc5 is good
for Black) 7 . . . ltJc6 (7 . . . 'ii' b6 8 ltJbd2
d5 9 ltJb3 .i.e7 1 0 'ii'd2 l2Jc6 1 1 0-0-0
.i.d7 1 2 'ii'c3 f6 is unclear according
to Timoshenko) 8 'iid3 (8 g6 looks a
little premature as after 8 . . . fxg6 9
.i.d3 0-0 White has nothing immedi­
ate and moves such as 1 0 ltJc3 or 1 0
<it'e2 are well met by 1 o . ltJb4; 8
. .
.i.d3 , threatening to play g6, looks
quite natural though) 8 . . . 'iib6 9
ltJbd2 (9 'ii'b 5 or 9 'ii' b 3 were more
solid) 9 . . . 'iix b2 1 0 llbl 'ii'x a2 1 1
'ii'c 3 ( 1 1 llxh7 is better), To�i<:-Vara­
vin, Alushta 1 994, and now Black
failed to play the decisive 1 1 ...'ii'xbl + !
1 2 ltJxbl .i.b4.
3) 4 lLJc6 5 llJd2 ltJxc5 !? 6 e4 d6
..•
7 ltJgf3 .i.g4 8 c3 g6 9 .i.e3 .i.g7 1 0
.i.xc5 dxc5 1 1 'ii'b 3 'ii'c7 was fine for
Black in Miladinovic- Sulskis, Mos­
cow OL 1 994.
4) 4 h6!? 5 .i.e3 e6 6 ltJd2 ltJxc5
..•
7 ltJgf3 d5 8 c3 b6 9 g3 .i.b7 1 0 .i.g2
.i.e7 1 1 0-0 0-0 with a roughly level
game, Miladinovic-Svidler, Yugosla­
via 1 995 .
S ltJd2 ltJxgS
6 hxgS g6
The immediate 6 'ii'xcS is doubt­
•.•
ful on account of the familiar pawn
sacrifice 7 g6 ! . After 7 . . . fxg6 8 e3
l2Jc6 9 ltJgf3 (9 .i.d3 allows 9 . . .ltJe5,
but even here 1 0 ltJgf3 ltJxd3+ 1 1
cxd3 d6 1 2 ltJh4 is quite awkward
for Black) and White will continue
with .i.d3 creating threats of .i. xg6
and llxh7. On 9 . . . ltJe5 White can
play 1 0 ltJxe5 'ii'xe5 1 1 .i.d3 'ii'f6 1 2
c3 followed by 'iic2 with a strong in­
itiative .
7 c3
120 The Trompowsky

I recall that Hodgson once derived l2Jf8 ! (Black now has the better pros­
great pleasure from the rook ma­ pects) 1 9 c4 a5 20 a3 axb4 2 1 axb4
noeuvre 7 lth4!? l2Jc6 (7 . . . 'flxc5?? 8 'fld7 22 l2Jc3 l:a3 23 l2Jdb5 ?? (the
l:c4) 8 llc4 . N ogueiras and Estevez wrong one) 23 . . . l:xb3 0- 1 .
now give 8 . . . .ig7 9 c3 l2Je5 1 0 l2Jb3 8... .ig7
'fie? 1 1 llh4 a5 1 2 a4 as good for 9 e3 l2Jc6
White (Black won ' t get his pawn 10 l2Jb3 'ft6
back), but why not 8 . . . liJe5 when the 1 1 a4 d6
consistent 9 llc3 .ig7 Jooks good for 12 a5 'ilic7
Black? 13 lla4
7 ••.
'fixes (D) We've already seen the manoeu ­
vre llh4-c4, and now Hodgson at­
tempts to treat us to the manoeuvre
l:a4-h4 . People usually talk about
open files for rooks, but sometimes
open ranks can be just as effective.
Take a look at the splendid game
Karpov-Hort, Moscow 197 1 for con­
firmation (beyond the sc ope of this
book, I'm afraid).
13 ..• .id7
14 'flat! ?
White's imagination is working
8 l2Jgf3 overtime. With the a-pawn rein­
8 l2Je4 'fies (perhaps there was no forced llah4 is now a serious threat,
need to encourage White's next move) hence Nunn's reaction, liquidating
9 l:h4 .ig7 10 l2Jf3 'flaS 11 b4 'flc7 the h-file pressure.
12 llcl 0-0!? 1 3 e3 d6 14 .ic4 lLJd7 14 .0 hS
lS .ib3 b6 16 �d2 .ib7 17 'flhl l S gxh6 l:xh6
(one of the few white moves that I 16 llxh6 .ixh6 (D)
managed to predict in this game)
17 llfc 8 and now:
.••

1 ) 18 llxh 7 .ixe4 19 llxg7+

�xg7 20 'flh6+ �g8 21 llhl fails to
2 1 . . . 'flxc3+, whilst other 2 1 st moves
for White are also insufficient, e.g.
21 c4 .ixf3 22 gxf3 (22 l:h 1 .ih5)
22 ... d5 ! 23 f4 (23 llhl 'fle5) 23 ... dxc4
24 lib 1 'fld6+ 25 �e2 'fid3+ 26 �f3
'flc3 ! and there is no mate.
2) The game T.Wall-Gallagher,
Grangemouth 1 990 continued 18 l0d4

The position is roughly level. The
remaining moves were 1 7 a6 l:.b8 J 8
lDbd4 �g7 1 9 axb7 lDxd4 20 lDxd4
l:.xb7 2 1 l:.a2 e5 22 lDc2 �e6 23 l:.a6
'iWb8 24 1Wa5 �f6 25 lDb4 �d8 26
1Wa4 + �d7 27 'iWd l �e7 28 �e2
l:.b6 1'2-1/2
Game 1 9
Gerstner - Gallagher
Biel 1993
1 d4 lDf6
2 �gS lDe4
3 �f4 cS (D)
3 . . . d5 is equally playable but the
text is probably more suitable for
King's Indian players.
4 dS
1 ) 4 f3 is the next game.
2) 4 dxcS is almost never played;
4 . . . lDc6 looks like a decent repJy.
3) 4 c3 i s u nambitious and in
Teren tiev-Gallagher, Liechtenstein
l 990, it was only my opponent's
u nsporting behaviour that deprived
me of my quickest ever victory.
4 ... 1Wb6 seemed to me the most logi­
cal continuation but I was loathe to
exchange queens so early. Suddenly
The Trompowsky 121
a beautiful trap flashed through my
mind and . . . 1Wb6 was played with a
trembJing hand. The next few moves
were banged out as we followed
my opponent's home preparation : 5
'iWb3 cxd4 ! 6 1Wxb6 axb6 7 �xb8 (7
cxd4 lDc6 is promising for Black)
7 . . . dxc3 ! 8 �e5 l:.xa2 ! ! (D).
Only here did he stop to think; a
bit late as skilful play is no w re­
quired to restrict his losses to a rook
and three pawns !
4 ..• 'ffb6
Only in this way can White be pre­
vented from developing his pieces
smoothly.
s lDd2
As the b-pawn can only be de­
fended with awkward moves White
often leaves it to its fate. It should be
noted, though, that Hodgson has no
confidence in this particular offer.
The alternatives are :
I) S b3?? 'iWf6 ! .
2) S lLJc3 1Wb4 (5 . . .1Wxb2 6 lDxe4
transposes to the game) 6 a3 lDxc3 7
axb4 lDxd l 8 �xd l cxb4 9 �xb8
l:.xb8 1 0 l:.xa7 e6 1 1 dxe6 dxe6 1 2
e 3 �c5 with advantage to Black,
Serrano-Kolev, St Cugat 1 992.
122 The Trompowsky
3) S li'cl c4 ! ? (5 . . . g5 also de­
serves consideration because Black
quickly achieved the better game in
Van der Sterren-Yusupov, Amster­
dam 1978 after 6 .ie5 f6 7 .ig3 i.. g7
8 c3 f5 9 e3 0-0, but White should
probably have tried 7 .ixb8) 6 e3
1i'a5+ 7 lDc3 ( 7 lDd2 c3 !) 7 . . . lDxc3 8
1i'd2 e6 ! 9 bxc3 (9 d6 g5 10 .ig3
.ig7 1 1 lDe2 b5 is very good for
Black) 9 . . . exd5 with some advantage
for Black. My source is the Informa­
tor editorial team, but Kasparov has
also suggested 5 . . .c4.
4) S .icl is the main alternative
to sacrificing the b-pawn. White's
bishop has taken three moves to get
nowhere, but Black's knight is sus­
pended in mid-air and his queen is
al so exposed on b6. Black must act
vigorously to prevent White from
taking over the centre which, for
some reason, he often fails to do in
practice. We are going to examine a
couple of possibilities:
4a) S e6 6 f3 (D) and now:
...
4al ) 6... lDf6 7 c4 exd5 8 cxd5 c4
(Black could have played a Benoni
with an extra tempo but . . . 1i'b6 is not
of much use ; instead he talces his
chances on the dark squares) and
now:
4al 1 ) 9 e4 .ic5 1 0 lDh3 d6 1 1
.i xc4 .ixh3 1 2 gxh3 0-0 1 3 l2Jc3
and now Gurevich an d Chernin pro-
pose 1 3 . . . a6 ! (presumably the imme-
diate 1 3 . . . 1i'd8 is met by 1 4 lDa4,
not fearing 14 . . . lDxe4 15 fxe4 1i'h4+
1 6 �d2) 1 4 �fl 1i'd8 followed by
. . . lDh5 and .. .f5 . It looks like enough
play for a lousy pawn to me.
4al 2) 9 e3! is Hodg son 's new
move, which seems to place Black
in some difficulties. His game with
Stohl, Isle of Man 1 995 continued
9 li'aS+ 1 0 lDc 3 b5 1 1 1i'd4 ! .i b4
•••
1 2 1i'e5+ �f8 ( 1 2 . . . �d8 1 3 .id2
l:e8 14 1i'g5 ! +) 1 3 a3 .ib7 14 axb4 !
li'xa l 1 5 lDge2 1i'a6 1 6 ltJ<l4 with
excellent compensation for the ex­
change. 9 i.. cS would have been
...
another try when both 1 0 .i xc 4
1i'b4+ 1 1 lDd2 .ixe3 1 2 1i'e2? ! 0-0
and 10 lDc3 0-0 1 1 �f2 l:e8 look
fine for Black but I ' m not sure what
to do after 10 �f2 ! as 1 0 . . . 0-0 1 1
.i xc4 l:e8 1 2 1i'b3 looks insuffi­
cient.
4a2) 6 li'aS+ 7 c3 lDf6 8 e4 d6
.••
(8 . . . exd5 9 e5) and now White has:
4a2 1 ) 9 .id2 exd5 1 0 c4 1i'c7 1 1
cxd5 g6 with a reasonable Benoni
according to Kasparov, but he also
suggests 1 0 exd5 as an improvement
without considering 1 0 . .. lDxd5 . Af­
ter 1 1 c4 lDb4 1 2 a3 .if5 I can't see a
good continuation for White.
4a22) 9 lDa3 exd5 1 0 exd5 .ie7
1 1 lDc4 1i'd8 was roughly level in
Van der Wiel-Kasparov, Moscow
IZ 1 982. According to Kasparov,
Van der Wiel now began to play

s trangely : 1 2 tlle 3 0-0 1 3 tlle 2 l:.e8
1 4 g4? (White calmly gains space on
the kingside , ignoring the fact that
h is king is still in the middle on an
open file; after 40 minutes thought
Kasparov found the refutation of
White's plan) 14 ...lilid7! 15 tllg 3 ..lg5
1 6 �f2 tlle5 1 7 ..lb5 ( 1 7 h4 ..lxe3+
1 8 ..lxe3 1if6 ! and the threat of
1 9 . . . tll x g4 is decisive) 1 7 . . . ..ld7 1 8
..lxd7 tD bxd7 1 9 tllef5 c4 ! 20 tllh5 ?
tll d 3+ 2 1 �g3 ..lxc l 22 l:.xc l g6 !
and White resigned, realising that
after 23 1id2 Black is not forced to
capture towards the cen tre with
23 gxfS, which loses to 24 1ih6,
....
but can play 23 gxhS winning at
••.
once.
4b) s g6 6 f3 tlld6 ! ? 1 e4 ..tg7 8
...
tlld2 0-0 (8 . . .f5 and 8 . . . e6 also come
into consideration) 9 f4 ! (9 ..ld3 f5 !
1 0 tlle 2 c4 ! and 9 tllh 3 e6 1 0 ..le2
exd5 1 1 exd5 c4 are both good for
Black according to Rotshtein) 9 . . . e6
1 0 e5 tll f5 ( 1 0 . . . exd5 1 1 exd6 l:.e8+
1 2 ..le2 c4 may bring you success at
blitz) 1 1 tllc4 'iid8 and now instead
of 12 tlle 3? tll x e3 1 3 ..lxe3 d6 ! 1 4
tllf3 dxe5 15 dxe6 ..lxe6 1 6 tll xe5
tlld 7 1 7 tll x d7 ..lxd7, which gave
Black an overwhelming position in
Liogky-Rotshtein, Cannes 1992,
White should play 12 dxe6? dxe6 1 3
'iix d8 l:.xd8 1 4 c3 b6 1 5 ..le2 i.b7
1 6 ..lf3 ..ld5 ! with an equal game.
Let us return to the position after 5
tll d2 (D).
S •.• 1ixb2!
Black makes use of a little tactic
to prevent White developing his
forces in the most harmonious man­
ner. After S tll x d2 6 ..lxd2 1ixb2
•••
The Trompowsky 123
7 e4 I believe that White does have a
dangerous initiative for the pawn.
6 lhxe4
There is little choice as 6 l:.bl
1if6 7 tll xe4 'iix f4 8 tll x c5 e6 is fa-
vourable for Black.
6.. . Wib4+
7 'iid.2
7 c3? ?, which rules out a later
. . . 'ii b4 , was a new try in Djurhuus­
Tisdall, Norwegian Ch 1 995. After
7 . . . 1ixe4 8 e3 e6 9 dxe6 •xe6 1 0
tll f3 ..le7 1 1 ..ld3 b6 (Black is not
tempted by . . . d5 as this would allow
White to open the centre with c4 or
e4) 1 2 'iic 2 g6 13 h4 ..lb7 14 h5 l:.g8
15 hxg6 hxg6 Tisdall states that
White has no concrete compen sa­
tion for his pawn apart from having
achieved a complicated position, a
fair enough comment. After the fur­
ther moves 1 6 l:.d 1 liJc6 1 7 tll g 5
..lxg5 18 ..lxg5 tlle5 19 ..le4 d5 20
..ld3 (20 ..lxd5 ..lxd5 2 1 'iid2 is bril­
liantly refuted by 2 1 . . . l:.d8, when 22
..lxd8 fails to 22 ... ..lxg2 ! and 22 e4
to 22 . . . l:.d7 !) 20 . .. ..tc6 2 1 ..le2 f6 22
..lf4 0-0-0 Black's advantage was
obvious.
7.•• 1ixe4
8 C3?!
124 The Trompowsky
More logical is 8 e3 (D) after
which Black has the choice between
countering in the centre with 8 . . . e6
or evacuating his queen by 8 . . . 'ii'b4 .
1 ) 8 e6, with the following pos­
..•
sibilities:
la) 9 c4 e5 (it's interesting that
Black waits for c4 before blocking
the centre; White no longer has ac­
cess to the square c4 and the scope of
his light-squared bishop is also re­
duced) 1 0 f3 'ii'f5 l l .id3 ?? 'ii'f6 1 2
.ig3 e4 ! and White lost a whole
piece in Hodgson-Chandler, Hast­
ings 1 99 1 .
1 b) 9 dxe6 'ii' xe6 (9 . . . dxe6 1 0
.ixb8 ! l:.xb8 l l .ib5+ �e7 1 2 0-0-0
'ii'd 5 1 3 'ii'c 3 gives White a very
strong attack while 9 . . . fxe6 weakens
the kingside but could still be worth
investigating) 1 0 ll:\e2 J b6 (Black
feels that his main priority is to cover
the d5-square; 1 O .ie7 1 1 ll:\c3 .if6
. . .
1 2 lDb5 ! is quite threatening, but
Black may be able to bail out with
1 2 . . . 0-0 1 3 ll:\c7 'ii'c 6 1 4 l:.d l b6 1 5
ll:\xa8 'ii'x a8 with some compensa­
tion for the exchange) 1 1 ll:\c3 .ib7
1 2 l:.d 1 .ie7 1 3 ll:\b5 0-0 ( 1 3 . . . ll:\a6
14 .ic4 d5 15 .ixd5 .ixd5 1 6 'ii'xd5
'ii' xd5 17 l:.xd5 ll:\b4 1 8 Ci:Jc7+ �f8
1 9 l:.d2 l:.d8 20 �e2 l:.xd2+ 2 1
�xd2 f5 ! 22 a3 Ci:Jc6 23 �c3 �f7 led
to approximate equality in Morta­
zavi-Howell, British Ch 1 992) 1 4
ll:\c7 'ii'x a2 15 c4 'ii'b 3 1 6 f3 .ih4+
17 .ig3 .if6 1 8 �f2 Ci:Jc6 l 9 ll:\xa8
.ixa8 20 .ie2 ll:\a5 2 1 'ii' xd7 ll:\xc4
22 l:.d3 'ii'a2 23 'ii'c7 ll:\b2 24 l:.d7 c4
25 �fl , Klinger-Dimitrov, Velden
1 993, and now 25 . . .'ii' b l + 26 .ie l c3
27 .ic4 c2 28 l:.xf7 'ii'xe l+ 29 �xel
cl 'ii' + and 30 ... 'ii'x c4 would have
won for Black.
l e) 9 ll:\e2 ! ? 'ii' x d5 1 0 'ii' x d5
exd5 l l ll:\c3 appears to give White
enough play for the pawns. Klinger­
Akopian, Palma 1 989 continued
l 1 . . . d6 1 2 ll:\xd5 �d8 1 3 0-0-0 .ie6
l 4 ll:\c3 ll:\d7 (Black cannot save his
pawn as 1 4 . . . �d7 is strongly met by
15 ll:\e4 d5 1 6 c4 ! when 1 6. . . �c6 1 7
ll:\c3 dxc4 loses to 1 8 l:.d8) 15 .ie2
f6 1 6 .if3 l:.b8 17 ll:\e4 ( 1 7 .ixd6
.ixd6 1 8 l:.xd6 �e7 is fine for Black)
l 7 . . . ll:\e5 1 8 ll:\xd6 .ixd6 1 9 .ixe5
fxe5 20 l:.xd6+ �e7 2 1 l:.hd 1 l:.hd8
22 l:.xd8 :xd8 23 l:.xd8 �xd8 24
.i xb7 .ixa2 25 �b2 .ie6 26 .ia6
1f2- 1h.
2) 8 'ii'b4 9 c3 'ii'a5 (D) and
...
now:
2a) 10 ll:\f3 d6 with a further
branch:
2al ) 1 1 l:.bl ? ! g6! (immediately
targeting the weak point of White's
position - c3) with a couple of exam­
ples:
2al l ) 12 e4 .ig7 1 3 e5 ( 1 3 l:.c l
would be embarrassing and 1 3 l:.b3
ugly but both are probably better
than 1 3 e5) 1 3 . . . 0-0 14 .ie2 ll:\d7 1 5

l::tb5 (the only way to save the e­
pawn) 1 5 . . .li'c7 1 6 exd6 exd6 17
.lh6 llJf6 18 .lxg7 �xg7 19 c4 li'e7
and White had nothing to show for
his pawn in Adams-Gelfand, Tilburg
rpd 1 993 (although White did man­
age to draw). It should be noted that
the actual move order of this game
was 8 llJf3 d6 9 e3 li'b4 1 0 c3 li'a5 .
2a 12) 12 i. bS + llJd7 (possibly
I 2 . . . .ld7 is better since 1 3 .lxd7+
llJxd7 14 l::txb7 llJb6 is not an option
for White; I actually avoided this line
on account of 1 3 c4 as I thought the
ending might be difficult to win) 1 3
.lxd7 + .lxd7 1 4 l::t x b7 i.g7 15 0-0
( 1 5 l::tb3 +) 15 . . . .lxc3 1 6 li'c2 (on 1 6
li'd3 I had prepared a beautiful vari­
ation : 1 6 ... .lc8 17 l::tb 5 .la6 ! 1 8
l::ttb l .lb4 ! 1 9 l::t l xb4 cxb4 2 0 li'd4
li'xb5 ! 2 1 li'xh8 + �d7 22 li'xa8
•fl #) 1 6 . . . .lc8 1 7 l::t xe7+? (totally
u nsound; better is 17 l!b3) 17 ... �xe7
1 8 li'e4 + �d7 1 9 .lxd6 l::te 8 (Black
could also win with 1 9 . . . �xd6) 20
i.e5 li' b4 ! 0- 1 Gilles-Gallagher,
Bern 1995.
2a2) l l .ld3 g6 1 2 l::tc l i. g7 1 3
h4 ! ? .lg4 1 4 h5 .lxf3 ( 14 . . . i.xh5 1 5
lDh2 f5 1 6 llJ f l ! followed by llJg3
didn' t appeal to me) 15 gxf3 llJd7 1 6
The Trompowsky 125
c4 ! (sensible) 1 6 . . .li'xd2+ 17 �xd2
0-0-0 and w h ilst White's space ad­
vantage gave him some compensa­
tion for the pawn I would still assess
this position as �, Weindl-Gallagher,
San Bernardino 1994.
2b) 1 0 d6! ? llJc6 1 1 llJf3 g6 (the
line 1 l . . .exd6 1 2 .lc4 .le7 1 3 0-0 b5
looks good for Black but perhaps
White could try to keep Black bot­
tled up with 1 3 l::tb l ) 1 2 l::tc l i.g7 1 3
dxe7 and now:
2b l ) Bellon-Dorfman, Logrono
1 99 1 continued 13 dS ?! 14 e4 !
...
llJxe7 ( 1 4 . . . dxe4 15 llJg5 is danger­
ous) 1 5 exd5 0-0 1 6 d6 llJc6 17 i.c4
( 17 d7 i.xd7 1 8 li'xd7 l::t fe8 + allows
Black a strong attack) 1 7 . . . li'a4 1 8
.ld5 .le6 1 9 i.e3 .lxd5 20 li'xd5
li'a6 ! 2 1 d7 llJe7 22 li'xc5 llJf5 and
in this highly unclear position the
players chickened out and agreed to
a draw.
2b2) The simple 13 llJxe7 looks
•••
good for Black. 14 .ld6 doesn't lead
anywhere after 14 . . . llJf5 1 5 .lc4
li'b6 whilst the immediate 14 .lc4
can be met by 14 . . . d5. On 14 .leS
0-0 1 5 i.xg7 ( 1 5 i.c4 d5) 15 . . . �xg7
W h ite's development is too poor to
exploit the weakened dark squares
around the black king.
I think it would be fair to conclude
that White is struggling to demon­
strate full compensation for the
pawn after 8 . . fib4.
.
8 ... li'd4!?
8 li'b4 is an equally valid ap­
. ..
proach here but I felt that Black's su­
perior pawn structure would give
him the better chances in any ending,
even if White won his pawn back.
126 The Trompowsky

9 11'xd4 cxd4 (D) 14 0-0-0 d6 (D)

10 .ie5 ?! The opening phase of the game

I was expecting 10 l;tdl, against
which I intended 1 0 . . . f6 ! ? ( 1 0 . . . g6
and 10 . .. d6 are also reasonable) with
the idea of exploiting the suspect po­
sition of the white rook after 1 1
:xd4 e5 12 :c4 ( 1 2 dxe6 dxe6 fa­
vours Black) 1 2 . .. tlJa6. For example
1 3 .id2 b5 14 :c3 tlJc5 15 e4 b4 and
now 16 :c4? loses an exchange after
1 6 . . . a5 1 7 .ie3 d6, so White has to
play 16 :e3 to which 1 6 a5 looks . . .
the most natural reply.
10• .• e6
1 1 .ixd4?
White should have settled for the
slightly worse position that arises af­
ter 1 1 dxe6 dxe6 ( 1 1 . . . tlJc6 ! ?) 12
.lxd4 tlJc6. The text is very careless
as Black's extra pawn is actually
worth quite a lot.
11.•. exd5
12 .ib2?
12 0-0-0 is better.
12.•. .lb 4+
1 3 c3
has concluded and it is time for
White to take stock. He has lost a
pawn, reduced his once proud dark­
squared bishop to a tragic state and
failed to even contemplate the devel­
opment of his kingside. As if this
were not enough, he has also posi­
tioned his king in the firing line of
the adversary 's two powerful bish­
ops. I still can ' t believe it took me
another 25 moves to win this game.
15 c4 0-0!
16 cxd5 .le3+
17 �bl
Or 17 �c2 .if5 + 1 8 �b3 tlJa6
with a murderous attack.
17 .•• .if5 +
18 �al .ic2?
Although it was physically pain­
ful to part with one of my glorious
bishops , I nevertheles s cashed in,
assuming that my opponent was
about to resign. A little more thought
would have produced the devastat­
ing 18 :cS?, with the point 19 .id4
••.

1 3 �f2 0-0 is also very unattrac­ :c t + ! 20 :xc l .ixd4+.

tive. 19 .icl .ixdl
1 3 ... .ic5 20 .ixe3 llxl7

Of course Black is completely
winning but, in comparison with
I 8 . . . l:tc8, White is now able to get his
kingside out and present the oppo­
nent with some technical difficulties.
The remaining moves were 2 1 .i.f4
lLJe5 22 lLJh3 l:tac8 23 e4 .i.c2 24
.i.e2 f5 ! ? (I was quite willing to re­
turn some of my material advantage
to regain the initiative and simplify
the position) 25 lLJg5 ! fxe4 26 lLJe6
exf3 27 gxf3 l:txf4 28 l'LJxf4 i.f5 29
l:td l ? (the last chance was 29 lLJe6,
although 29 . . . <j;(f7 30 f4 lLJg6 should
still be winning for Black) 29 . . . l:tc2
30 h4 <j;(f7 3 1 l:te l a6 32 lLJe6 <j;; f6 33
f4 lLJg6 34 h5 lLJe7 35 .i.d l ? l:tc l +
36 <j;; b2 l:tbl + 37 <j;;a 3 lLJxd5 38 tLJd4
lLJb6 39 lLJb3 lLJc4+ 40 <j;; b4 lLJb2 0- 1
Game 20
I. Sokolov - Smirin
Wijk aan Zee 1 993
1 d4 lLJf6
2 .i.gS lLJe4
3 .i.f4 cs
4 f3 1WaS+
At first glance this check may
seem difficult to comprehend, but
the point is that by forcing White to
play c3 (5 lLJd2 lLJxd2 6 .i.xd2 1i'b6
is equal, while after 5 . . . lLJf6 there is
probably nothing better than 6 c3,
transposing to lines considered later)
his options are reduced. For exam­
ple, the position after 4 . . . l'LJf6 (in­
stead of 4 . . . 1Wa5 +) 5 d5 1i'b6 6 e4 (6
lLJc3 is another extra possibility
available to White) 6 . . . 1Wxb2 7 tLJd2
1i'c3 has been reached on several oc­
casions and the general consensus is
The Trompowsky 127
that White has fair play for his pawn;
but if Black had flicked in 4 . .. 1Wa5+,
then the move . . . 1Wc3 would have
captured a second pawn.
Another variation which Black's
queen check avoids is 4 . .. lLJf6 5 dxc5,
which usually leads to a very sharp
Sicilian type position - perhaps not
to everyone 's taste. A drastic exam­
ple is Landenbergue-Roder, Bern
1993: 5 . . .�6 (5 ...1Wa5+ 6 lLJc3 1Wxc5
7 e4 d6 8 1i'd2 a6 is an alternative
way for Black to play) 6 lLJc3 lLJxc5
7 e4 d6 8 1i'd2 i.d7 9 0-0-0 1i'a5 1 0
<j;; b 1 ltd8 ?? 1 1 lLJd5 ! 1-0.
5 c3 lDr6
6 dS
6 lLJd2 is a major alternative, em­
ployed recently by S alov, Adams
and Hodgson amongst others. After
6 cxd4 7 lLJb3 1i' b6 (7 . . . 1Wf5 ! ? is a
•••
recent try but I ' m sticking with the
older and more trustworthy 7 . . . 1Wb6)
White can choose to conduct the
game with or without queens:
1 ) 8 1Wxd.4 (D) and now:
l a) 8 1Wxd4 9 cxd4 d5 is prob­
•••
ably a l ittle better for White ; prac­
tice suggests that he can transform
his slight lead in development into a
128 The Trompowsky
space advantage or something more
concrete . For example, 1 0 e3 e6 1 1
g4 ( 1 1 l:.c l lDc6 1 2 g4 .ld7 1 3 lZJc5
.l xc5 1 4 l:.xc5 1;e7 1 5 1;d2 l:.hc8
1 6 .ld3 was also pleasant for White
in Hodgson-Tiviakov, Calcutta 1 993)
l 1 . .. lZJc6? ! (better is 1 1 . . . .lb4+ fol­
lowed by 1 2 . . . 1;e7 withjust an edge
for White) 1 2 .t b5 .ld7 1 3 a3 ! 1;d8
1 4 l:.c l llJe8 1 5 lDc5 .lxc5 1 6 l:.xc5
1;e7 1 7 lZJe2 lZJd6 1 8 .ld3 l:.ac8 1 9
b 4 ( 1 9 e4? is careless: 1 9 . . . dxe4 20
fxe4 lZJxd4 ! 2 1 l:.xc8 l:.xc8 22 lZJxd4
e5 !) l 9 ... b6 20 l:.c3 a5 2 1 b5 lZJa7 22
a4 l:.xc3 23 lZJxc3 l:.c8 24 1;d2
lZJc4+ 25 1;c2 f6 26 e4 dxe4 27 fxe4
lZJd6 28 1;b3 lZJf7 2 9 e5 ! l:.h8 30
lZJe4 h5 3 1 g5 fxg5 32 lZJxg5 lZJh6 33
.te l ! lZJc8 34 l:.fl h4 35 h3 l:.g8 36
.lh7 l:.h8 37 .lg6 .le8 38 .la3+
1;d7 39 .le4 lZJe7 40 .lxe7 1;xe7 4 1
l:.c l i.d7 42 l:.c7 1;d8 43 l:.b7 1 -0
Salov-Akopian, Wijk aan Zee 1 993.
Fortunately for the game of chess
there are people with worse tech­
nique than Salov.
l b) 8 lZJc6! 9 li'xb6 axb6 (D)
. .. .
leads to a more dynamic position
where Black's central superiority
should compensate for his weakened
queenside. White has now played:
1 b1 ) 1 0 e4? ! d5 ! (an important
point) 1 1 exd5 ( 1 1 .ld3 e5 12 .lg5
.le6 1 3 lZJd2 lZJd7 14 exd5 .t xd5 1 5
.lc4 .lxc4 1 6 lZJxc4 b5 1 7 lZJe3 f6
1 8 .lh4 lZJc5 was better for Black in
Ochoa-Dorfman, New York 1 989)
l 1 . . . lZJxd5 1 2 .ld2 e5 and Black
has a very active game. One example
is Rausis-Mukhutdinov, Moscow
1 992 which continued 1 3 a3 .le6
( 1 3 . . . i.f5 is also quite good) 1 4 c4
lZJf6 1 5 .le3 lZJd7 1 6 l:.c 1 lDc5 1 7
lZJxc5 .lxc5 1 8 .lxc5 bxc5 l 9 llJe2
1;e7 20 lZJc3 l:.hd8 2 1 .le2 lLJd4
with a clear advantage for Black.
1 b2) 10 a3 ?!. White plans an as­
sault on the b-pawn but ftrst wants to
rule out . . . b5-b4. Greedy and time
consuming is the verdict. 1 0 . . . d5 ? ! (I
think 1 0 . . . e5 ! 1 1 .le3 d5 is more ac­
curate as White has hardly anything
better than 12 .lxb6 transposing to
the game, while avoiding 1 1 lZJd4)
1 1 .lc7? (Smirin believes White
should have played 1 1 lZJd4, al­
though he still considers Black to
have an edge after l l . . .e5 !? 12 lZJxc6
exf4 1 3 lZJd4 .ld6 1 4 1;f2 0-0 15 g3
lZJh5 ! 1 6 gxf4 lZJxf4 1 7 e3 lZJe6)
l l . . . e5 ! 1 2 .lxb6 d4 ! 1 3 cxd4 .le6
14 llJc5 ( 1 4 dxe5 lZJd7 costs White a
piece and 1 4 d5 lZJxd5 1 5 i.c5 lZJa5 !
1 6 lZJxa5 .lxc5 1 7 lZJxb7 .ld4 ! is
tremendous for Black) 1 4 . . . lLJd5 1 5
lZJxe6 fxe6 1 6 .lc5 lZJxd4 ! 1 7 .lxd4
( 1 7 .lxf8 l:.xf8 is also excellent for
Black as Smirin demonstrates with
the following variation: 1 8 l:.c l lZJe3
1 9 1;f2 lZJdc2 20 lZJh3 h6 ! 2 1 'tt g 3
l:.a4 ! -+ 22 lZJf2? lZJf5+ 23 1;h3
l:.h4#) 17 . . . exd4 1 8 l:.c l .ld6 1 9 e4?
(an understandable bid for freedom

which hastens the end; 19 g3 9i;e7 20
f4 l:thc8 2 1 l:txc8 l:txc8 22 9i;d2 l'tJe3
2 3 lDf3 was a better chance although
B lack is still much better) 1 9 . . . dxe3
20 .lc4 .le5 2 1 .lxd5 exd5 22 l:tc2
9i;d7 23 lDe2 l:thc8 24 9i; d l l:tc6 ! 25
f4 .lf6 26 l:txc6 bxc6 27 9i;c2 9i;d6
28 l:td l c5 29 l:td3 d4 0- 1 V.Kova­
cevic-Smirin, Zagreb Z 1993.
1 b3) 10 .le3 (greedy and sensi­
ble) and now:
1 b3 1 ) 10 bS 1 1 lLJd4 ( 1 1 .ld2
•••
e5 1 2 a3 and now 12 dS? ! 1 3 e3
•••
lDa7 1 4 f4 ! ? exf4 1 5 exf4 .ld6 al­
lows White some positional advan-
tage, but 12 d6 followed by . . . .le6
••.
should be fine for Black) 1 l . . . lDxd4
( 1 1 . . . b4 1 2 lDb5 ! is quite good for
White) 1 2 .lxd4 e6 1 3 e4 b4 1 4
.lb5 ! is ! according to Milov.
1 b32) 10 dS ! ? 1 1 .lxb6 e5 1 2
•..
e4 ! ( 1 2 e3 lLJd7 1 3 .lc7 lLJc5 would
be very bad for White) 1 2 . . . .le6
(better than 1 2 . . . dxe4 1 3 .lc4 exf3
14 lLJxf3 when White's pieces are
very active) 1 3 .lb5 lDd7 14 .lf2
dxe4 ( 1 4 . . . d4? ! 15 lLJc l dxc3 1 6
bxc3 .lc5 1 7 lLJge2 9i;e7 1 8 lLJd3
.lxf2+ 1 9 9i;xf2 l:ta3 20 l:thdl l:tha8
2 1 l:td2 f6 22 l:tad 1 was a little better
for White in Tregubov-Nadyrkha­
nov, S ochi 1 994) 1 5 lLJd2 exf3 1 6
lDgxf3 .le7 1 7 0-0 0-0 1 8 l:tfe l l:tfd8
with chances for both sides. We
probably need further tests in this
line before any judgement can be
made.
1 b4) 1 0 lLJd4 (D) with the further
branch:
l b4 l ) 10 llJdS? 1 1 lDb5 ! l:ta4
•••
( 1 1 . . . e5 is welJ met by 1 2 e4 and
1 1 . .. l:ta5 1 2 llJc7 + lLJxc7 1 3 .lxc7 e5
The Trompowsky 129
1 4 e3 .lc5 15 b4 lLJxb4 1 6 cxb4
.l xb4+ 17 <j/;f2 .lc3 1 8 .lxb6 is
winning for White) 12 .ld2 lDa5 1 3
0-0-0 ! d6 ( 1 3 . . . l:txa2 14 9i;bl l:ta4 15
e4 d6 1 6 b4) 14 e4 lDf6 15 9i;bl lLJc4
1 6 .le 1 with a clear advantage for
White, Rausis-VJvanov, Riga 1 993 .
1 b42) 10 eS!? 1 1 lLJxc6 exf4 1 2
•.•
lLJd4 d5 ( 1 2 . . . lDd5 ! ?) 1 3 <j/; f2 .ld6
1 4 e 3 fxe3+ 1 5 9i;xe3 0-0 is as­
s essed as � by V.I vanov, but he, no
doubt, grew up in the Soviet school
of chess where they value things
such as pawn structure more highly
than us (remember the Short-Kas­
parov match) . I have to confess that I
wrote the previous sentence before
having actually examined the posi­
tion, confidently assuming that I
would find some way to trouble the
white king on e3. I can't, so I'll have
to agree with Ivanov after all.
1 b43) 10 lLJxd4 1 1 cxd4 d5 1 2
•.•
.lc7 i s assessed a s � by variou s
sources, whilst Nadyrkhanov gives
12 . . . e6 1 3 .lxb6 lDd7! as =F, which I
find quite puzzling as I can't see any­
thing obvious after 1 4 .lc7 . Instead
of 1 3 . . . lLJd7 though, I would J ike to
suggest the paradoxical 13 9i;d7!? ...
(D).
130 The Trompowsky
The idea is simply to win back the
material with . . . �c6 and I can't see
anything convincing for White, e.g. :
1 b43 1 ) 14 l:tcl �b4+ 15 �f2
l:txa2 1 6 l:tc7 + �d6 with good play
for Black.
l b432) 14 � cs �xc5 1 5 dxc5
�c6 16 b4 and now both 16 lhJ and
..•
16 d4 look very good for Black.
.••
1 b433) 14 a4 �b4+ 15 cat>f2 �c6
1 6 a5 lbd7 1 7 l:tc l + cat>d6 1 8 �c7+
'3ie7 and Black wins back his pawn
( 1 9 l:tal b6) with a good game.
2) 8 cxd4 (D) and now:
2a) 8 d5 is by far Black's most
...
common, but perhaps not the best,
choice. White now has:
2al ) 9 l:tc l lt:Jc6 1 0 e3 a5 ! 1 1 a4
e5 ! ? 1 2 dxe5 ( 1 2 �xe5 lt:Jxe5 1 3
dxe5 1i'xe3+ i s excellent for Black)
1 2 . . . lt:Jh5 1 3 �b5 ! lt:Jxf4 14 exf4
�b4+ 15 �fl 0-0 16 lt:Je2 �e6 with
an extremely unclear game, Hodg­
son-Nunn, Pardubice 1 993.
2a2) 9 e3! . After the above game
Hodgson came to the conclusion that
l:tc 1 was a bit of a luxury and that the
most important thing for White to do
was to rush his king's knight to c3.
Black has now tried:
2a2 1 ) 9 �f5 1 0 lt:Je2 lt:Jc6 1 1
.••
lt:Jc3 e6 1 2 l:tc 1 l:tc8 ( 1 2. .. �e7 1 3
lt:Jb5 i s good for White and 1 2 . . . a6 1 3
lt:Ja4 ! ? ii'b4+ 1 4 lt:Jd2 foil owed by
a3 and b4 should also give White an
edge) 1 3 g4 �g6 1 4 h4 h6 15 �f2
1i'd8 1 6 �b5 �d6 1 7 lt:Je2 0-0 1 8
lt:Jc5 �xc5 1 9 l:txc5 ii'b6 20 �xc6
l:txc6 2 1 l:txc6 1i'xc6 22 g5 hxg5 23
hxg5 lt:Jh7 24 1i'b3 ;t Hodgson-Wojt­
kiewicz, Rakvere 1 993 .
2a22) 9 lt:Jc6 1 0 lt:J e2 e5?! (the
..•
alternative 1 0 . . . a5 1 1 lt:Jc3 a4 1 2
lt:Jd2 1i'a5 1 3 lt:Jb5, Hodgson-Anka,
Metz 1 994 was also a little dubious
so Black should probably settle for
1 0 . . . e6 although after 1 1 lt:Jc3 White
has a small edge) 1 1 �xe5 lt:Jxe5 1 2
dxe5 1i'xe3 1 3 1i'd4 ! ( 1 3 exf6 �b4+
1 4 lt:Jd2 0-0 is extremely dangerous
for White) 1 3 ... ii'xd4 14 llJexd4 gave
White a very good endin g in Gal­
lagher-Forster, Metz 1 994. Spending
a tournament with Hodgson can eas­
ily lead one into picking up bad hab­
its. In Metz, not only did he persuade
me to play the Benko Gambit, an
opening I haven' t touched since I
was in short trousers, but he also
talked me into wheeling out my first
Tromp for many years. He is o f

course blameless for the fact that I
fai led to win my game against For­
ster.
2b) 8 e6! 9 .i.d2 (one of the main
•••
points behind the flexible 8 . . . e6 is
that 9 e3 is now met by 9 . . . llJ<l5 ;
therefore White has to waste time
with his bishop before he can get his
kingside out) 9 ltJc6 1 0 e3 a5! ?
...
(slightly more aggressive than the
10 . . . .i.b4 which Hodgson had faced
in the previous round although there,
too, Black achieved a comfortable
game after 1 1 ltJe2 .i.xd2+ 1 2 1ixd2
0-0 1 3 ltJc3 d6 1 4 g4 e5 1 5 d5 ltJe7
16 0-0-0 a5 1 7 �b l a4 1 8 ltJc l a3 1 9
b 3 l:r.a5 2 0 .i.c4 .i.d7 Kengis­
=,
Hodgson , Bern 1 995) 1 1 a4 .i.b4 12
.i.b5 liJd5 1 3 1ie2 (the e-pawn needs
protection but now White has to de­
velop h is knight to the edge of the
board) 13 0·0 14 ltJh 3 d6 (as in
• ••
Hodgson-Kengis Black aims for
. . . e5) 15 ltJg5 ( 1 5 liJf2 is safer)
1 5 e5 16 .i.xb4 liJdxb4 17 dxeS
•.•
(D) and now:
2bl ) Hodgson-Sue tin, Bern 1 995
continued 17 dxe5 1 8 l:r.d 1 h6 1 9
•••
ltJe4 .i.f5 20 0-0 l:r.ad8 2 1 .i.c4 11'c7
2 2 l:r.c l 1ie7 23 liJbc5 b6 24 ltJg3
The Trompowsky 131
.i.c8 25 ltJce4 .i.e6 with a level posi­
tion.
2b2) I believe that Hodg son was
more concerned about 17 . liJc2+!?.
..
After 1 8 1ixc2 1ixe3+ 1 9 �fl ( 1 9
1ie2 1ixb3 20 exd6 liJd4 2 1 1ie4
11'b4+ followed by . . . .i.f5 is awful
for White) 1 9 ... ti'xg5 20 exd6 White
has an extra pawn but an unhappily
placed king.
Let us now return to the position
after 6 d5 (D).
6 ... 'iib6
Very often Black simply plays
. . . d6 and . . . g6, or 6 ... e6 straight
away, but I believe that the attack on
b2 poses White the most problems.
7 b3
White accepts a slight weakening
on the a l-h8 diagonal (which can
prove relevant) in order to maintain
the material balance. The alternatives
involve giving up a large amount of
material for uncertain compensation
and are therefore rarely seen. They
are, however, not without danger for
Black.
1 ) 7 e4. Trompowsky players
don' t usually lose too much sleep
over the fate of their b-pawn, but
132 The Trompows/cy
here this sacrifice, thanks to . . . 'ii'a5 +,
involves throwing in the c-pawn as
well. After 7 . . . 'ii'x b2 8 ltJd2 'ii' xc3 9
�c4 d6 10 ltJe2 'ii'a5 White has a se­
rious lead in development but Black
is solid and two pawns ahead. I, my­
self, have played this position with
White (a long time ago) in several
quickplays but without success.
White should probably now play 1 1
a4 to prevent . . . b5 and l l ... g6 would
be a sensible reply.
2) 7 'ii'd2 is the move White
would like to play, but it does allow
7 ltJxdS 8 'ii'xdS 'ii'x b2 9 'ii'b3
.••
'ii'xa l after which Black has won the
exchange and two pawns but got his
queen into a tight spot. If we con­
tinue with the natural moves 10 e4
ltJc6 (D) White then requires another
four moves to win the queen; � on
f l somewhere, ltJge2, 0-0 and ltJa3.
Let's have a look at a few lines and
Black's attempts to counter this plan .
2a) 1 1 � c4 e6 1 2 'ii'c 2 (Black
was threatening . . . ltJa5) l 2 . . . b5 ! 1 3
�xb5 ( 1 3 �e2 b4 ! ? 1 4 �d2 bxc3 1 5
�xc3 ltJd4 ! ) 1 3 . . . e5 ! 1 4 �c l ( 1 4
�xc6 l:tb8) 1 4 . . . l:tb8 1 5 a4 a6 and
Black wins.
2b) 11 �bS d6 and now there are
two possibilities:
2bl ) 12 � c4 e6 1 3 'ii'c 2 b5 (the
line 1 3 . . . �d? 1 4 �c l ltJe5 1 5 �b2
1Vxb2 l 6 'ii'x b2 ltJxc4 1 7 'iVxb? ltJb6
followed by . . . �e7 and . . . 0- 0 also
looks good for Black) 14 �xb5 �d7
1 5 �c l l:tb8 1 6 ltJa3 l:txb5 ! ? (other
moves such as 1 6 . . . ltJe5 or l 6 . . . �e7
might be even better, but the text is a
good example of the sort of tactics
that are available to Black in his at­
tempt to extricate his queen) 17 ltJxb5
ltJb4 ! with advantage to Black.
2b2) 12 ltJe2 �e6 1 3 �xc6+ bxc6
1 4 'ii' b7 l:tc8 1 5 ltJc l g6 (Black sim­
ply plans . . . �g7, . . 0-0 and . . . l:tb8)
.
1 6 �g5 f6 17 � d2 �g7 1 8 ltJa3 <i;f]
1 9 ltJc2 l:tb8 20 'ii' x b8 'ii'xc l + 2 1
�xc l l:txb8 and Black is winning.
2c) 1 1 �d3 d6 1 2 ltJe2 ( 1 2 �c l
ltJe5 ! 1 3 �b5+ <i;d 8 1 4 �e2 c4 1 5
'ii'c 2 �d7, setting up ideas of . .. ltJd3
and . . . �a4 , is winning for Black)
1 2 . . . �e6 1 3 'ii'xb7 l:tc8 14 �c4 l:tb8 !
1 5 'ii'xc6+ �d7 1 6 �xfl + <i;d 8 and
Black wins. Basically, once the b-file
is open White has very little chance
of achievin g his aim. These vari­
ations are not meant to be conclusive
but are intended to demonstrate the
sort of resources available to Black
and to help you pluck up your cour­
age before sending your queen into
the unknown. Perhaps the most im­
portant clue as to the status of 7 'ii'd 2
is that none of the major Trom­
powskyites is willing to try it, which
is especially revealing in the case of
Hodgson who has a soft spot for al­
lowing . . .'ii'xal .
7 .•. e6

It is al so quite reasonable for
B lack to continue in Benoni fashion.
A recent example is Adams-Tka­
chev, Wijk aan Zee 1 995 : 7 d6 S e4
..•
g6 9 �d3 �g7 1 0 l2Je2 0-0 1 1 l2Jd2
l2Jbd7 1 2 l2Jc4 1//c 7 1 3 a4 b6 1 4 0-0
a6 15 1//d 2 :es 1 6 l2Jg3 �b7 with
about equal chances.
8 e4
8 dxe6?! looks very anti-posi­
tional. After S . . . 1//xe6 9 c4 d5 1 0 l2Jc3
d4 1 1 l2Jb5 l2Ja6 1 2 e4 dxe3 1 3 l:tc l ?
( 1 3 1//e 2) 1 3 . . . l2Jh5 ! 1 4 l2Jd6+ �xd6
1 5 �xd6 �d7 1 6 g4 �c6 ! White
was already completely lost in Alek­
sandrov-Akopian, Oakham 1 992.
8 .•. exdS
9 exdS �d6 (D)
10 l2Jh3
Smirin comments, in his excellent
notes in New in Chess, that he was
more afraid of 10 �gS upon which
he intended 10 �e7. I believe that
..•
Black stands quite well here since
White's development is rather poor
and his unprotected bishop on g5 ex­
poses him to some tactical tricks. For
example :
1 ) 1 1 c4 0-0 ( 1 1 . . . 1//d6 !? also de­
serves serious consideration as it is
The Trompowsky 133
unlikely that White can get away
with 1 2 l2Je2 1//e5 1 3 �f4) when
White needs to just play �d3 and
l2Jge2 to consolidate his position, but
this is difficult to arrange, e.g . 12
l2Jc3 l:teS creates awkward pressure
on the e-file ( 1 3 �d3 l2Jxd5 !) whilst
1 2 �d3 l2Jxd5 ! 1 3 �xh7+ <it(x h7 1 4
1//xd5 �xg5 ( 1 4 . .. �f6? 1 5 1//f5+ ! )
1 5 1//x g5 l:teS+ should be very good
for Black.
2) After 1 1 l2Ja3, Black should
probably avoid 11 11/aS 1 2 l2Jc4 ! ?
..•
1//xc3+ 1 3 �d2 1//d4 1 4 d6 �xd6 15
l2Je2 1//d 5 16 l2Jf4 �xf4 17 �xf4,
which gives White strong pressure
in return for his pawns, and instead
play 1 1 d6 1 2 l2Jc4 1//dS or possibly
•..
even l l 1//d6!?.
...
10 ... 0-0
10 �xf4 1 1 l2Jxf4 1//d6 1 2 1//e2+
•..
�dS 1 3 1//d 2 (;!; Schussler) 1 3 . . . g5
14 lllli3 is assessed as ;!; by Sokolov,
but I'm not so sure. After 14 . . . h6 15
�c4 l:teS+ 16 �f'2 violent reactions
such as 16... bS or 16...g4 1 7 l2Jf4
gxf3 l S gxf3 l:te4 are perhaps inap­
propriate owing to Black's lack of
queenside development. However,
there are more measured approaches,
e.g. 16 a6!? 1 7 a4 b6 1 S l2Ja3 l2Jc6
..•
or 16... b6 1 7 l2Ja3 a6 ! l S l:tae l l:txe l
1 9 l:txe l b5 when 20 �xb5 will at
most lead to unclear complications,
and probably less.
1 1 1//d2 :es+
12 �e2 (D)
S miri n comments that the white
pieces are placed somewhat awk­
wardly, but that if he manages c3-
c4 he will be able to complete his
development without hindrance and
134 The Trompowsky

18 'iicS+ followed by ... d6 is an ­

•..

other idea.
17 •.• 'iieS
18 ltJc3 ltJxdS!
19 'iixdS
On 1 9 l:tcl, 19 tt:Je3 looks goo d
••.

but 19 ltJxc3 20 l:txc3 l:txa2! is even

...

more convincing (2 1 1fxa2 1fxc3+

22 'iid2 'iic 5).
19 ..• 1ixc3+
20 � (D)
place Black under positional pres­
sure. Hence Black's next move.
12
••• c4!
13 �xd6
1 3 bxc4 �c5 looks awful for
White. On 1 4 �g5 Black can still
play 14 . . . d6 as 1 5 �xf6 g xf6 1 6 1ih6
fails to 1 6 ... 'iib2 ! .
13 ... 1ixd6
14 bxc4 bS!
White is allowed no respite. The
text both breaks up the white centre
and starts to tackle the only draw­ 20 .. . �b7!!
back to Black's position - his lack of It took Smirin 25 minutes to find
queenside development. this star move, after first being de­
I S cxbS a6 pressed by variations along the lines
l S ti:JxdS is also playable but
..• of 20 1ie3+? 2 1 �g3 l:te6 22 �d3
••.

. . . a6 ensures that even more lines with 23 l:the 1 to follow and 20 l:ta4
••.

will be opened. 2 1 l:the l .

16 c4 axbS
17 cxbS?
Too optimistic according to Smi­
rin. White should have settled for 17
ti:Jc3 bxc4 1 8 0-0 with a complex
game ahead . Sokolov now gives
the variation 18 l:ta3 1 9 liJe4 tt:Jxe4
.•.
20 fxe4 l:txe4 2 1 tlJg5 with an attack
for White, but I can' t see anything
devastating after 2 1 . . . l:th4 22 h3 (22
g3 l:txg3+) 22 .. .f6. Smirin, on the
other hand, suggests 18 ti:Jc6 while
•••
21 1fc4
The far from obvious point be­
hind Black's last move is that 21
'iix b7 costs White his queen after
2 1 . . . 'iic 5+ 22 �g3 l%a7 ! . 21 1id3
also fails to 2 1 ... 'iic5+ 22 �fl l:ta3 .
21 ... 1ie3+
22 �g3 hS?
Although this move forces trans­
position into an ending a piece up it
seems to let slip Black's advantage .
The alternatives are :
 The Trompowsky 135

1 ) 22...1i'xe2? 23 llhe 1 1i'xe 1 +

24 llxel llxe l 25 liJg5 lie? 26 1i'c7
lle8 27 1i'xb7 with a winning posi­
tion for White.
2) 22 lle4 23 1i'c 1 ! 1i'xe2 24 fxe4
. ..

1i'xe4 25 liJf4 g5 26 a4 (perhaps you

believe this to be a misprint, as I did
at first) 26...gxf4+ 27 1i'xf4 and White
is w inning according to Smirin. The
p oint is that after 27 . . . 1i'xg2+ 28
�h4 Black has no choice but to ex­
change queens (28 . . . 1i'e4) where­
upon White's passed pawns will
triumph over Black's pitiful pieces .
3) 22 dS! is the solution accord­
•.•
ing to Smirin, who provides the fol­
lowing variations:
3a) 23 1i'c7 h5 ! 24 1i'xb7 lla4 !
with a decisive attack.
3b) 23 1i'f4 1i'xf4+ 24 liJxf4 g5
For his piece White has two
passed pawns, including the incred­
ibly powerful one on b5 which domi­
nates the whole queenside. After
serious thought Smirin came to the
conclusion that his only chances lay
in a counterattack on the kingside.
26
•.• fS!
26. llaxa2 ? 27 llxa2 llxa2 28 lie?
•.

25 liJh5 llxe2 and Black has a much lla7 29 b6 and 26 �b7 27 lie? �g6
•••

better version of the game as he has a 28 llxb7 �xh5 29 a4 are very good
square on d7 for his knight and he for White.
has retained his h-pawn. 27 llc7 f4+
3c) 23 1i'd3 1i'xe2 (23 . . . 1i'e5+ 24 28 �b3 g4+!
�f2 liJd7 is also good) 24 llhe l 29 �b4!
1i'xe l + 25 llxe l llxe l and Black's 29 fxg4 �xg2+ gives Black a dan­
material advantage should be suffi­ gerous f-pawn and 29 �xg4 llxg2+
cient to win the game. 30 �xf4 lla4+ allows him to activate
23 1i'f4! his pieces.
Forced because 23 libel loses to 29
•••
gxf3
23 .. Jle4 and 23 .ld3 to 23 . . . d5 24 30 llxb7
1i' b3 (24 1i'c3 lla4) 24 . . . 1i'e5+ 25 Better than 30 liJxf4, which w ould
�t2 1i'd4+ . leave White fighting for a draw after
23 ••• 1i'xf4+ 30 . . . lle4 3 1 llxb7 llxf4+ 32 �g3
Black too has no choice since llf8 33 gxf3 lla3.
23...1i'xe2 24 llhe l 1i'xe l + 25 llxe l 30 ... fxg2
llxe l 26 1i'c7 .ld5 27 liJf4 is out of 31 liJxf4 llf2
the question. 32 tl)xg2!
24 liJxf4 gS Otherwise ... llf 1 wins.
25 liJxhS llxe2 32
••. llxg2
26 libel (D) 33 llcl !
136 The Trompowsky

There is no time to push the a­ The players now agreed to a draw

pawn on account of 33 .. . �n, threat­ (in a position where Black's d-paw n
ening to whip up a mating attack by and queen's knight remain unmoved)
34 . . . 4Jc6. in view of 34 �g3 :hxa2 3 5 :cs+
33
••• :xh2+ �n 36 :bxb8.
1 0 The Torre Attack
The Torre Attack is one of the most
solid systems in this book and a
popular choice for those who wish to
avoid anything resembling a theo­
retical battle (for a variety of reasons,
but a lack of time for the amateur and
a lack of the work ethic for the
grandmaster are the most common) .
After the moves 1 d4 l2Jf6 2 l2Jf3 g6 3
.i.g5 .i.g7 White usually plays 4
l2Jbd2 intending to play e4. Then af­
ter 4 . . . 0-0, 5 e4 is not very highly
thought of because of 5 . . . d5 ! (see
Game 2 1 ) so it is better for White to
wait with 5 c3. Then Black can
choose between the solid 5 . . . d5 and
the more dynamic 5 . . . d6, which after
6 e4 actually transposes to the Pirc
Defence. Black then has the choice of
playing for . . . e5 (6 . . .'iie 8, 6 . . . l2Jbd7 )
or playing 6 . . . c5 , which is the course
I am recommending and which re­
cently received the PCA World
Champion's seal of approval. The
details are to be found in Game 22.
Game 2 1
Bogdanovich - Cvitan
Liechtenstein 1 994
1 d4 l2Jf6
2 l2Jf3 g6
3 .i.gS .i. ,;T
4 l2Jbd2
Sometimes White plays 4 c3 (or
even 3 c3) to dissuade Black from
playing a very quick . . . c5 , but as this
is not our intention it will merely
lead to a transposition of moves.
4 •.• 0-0
It's too early for Black to declare
his hand in the centre as, depending
on the white set-up, he can strike
with either his d-, c-, or e-pawn.
5 e4
5 c3, actually White's most popu­
lar choice, is the subject of the next
game.
5 e3 is extremely passive but has
its supporters (usually pretty rock­
solid characters). After 5 d6 (D) we
•.•
consider various moves by the king's
bishop, but not 6 c3 which will just
transpose to one of the other lines. In
each case Black plays for . . . e5 , sup­
ported by ... l2Jbd7 and . . . 'iie 8, which
seems to be the best reaction when
White has played e3 . This line can be
annoying to meet in a must-win situ­
ation, but then that's life.
 138 The Torre Attack
1 ) 6 �d3 liJ bd7 7 0-0 eS 8 c3 h6
9 � h4 11'e8! (threatening . . e4) and
. 11'c8 20 11'e7 g5 2 1 �g3 g4 22 lbh4
now:
l a) 1 0 ltJel ( 1 0 11'c2 could also
be met by 1 0 . .. d5) 1 0 . . . d5 !? 1 1 liJb3
a5 1 2 a4 liJb6 ! ? 1 3 ltJc5 liJfd7 14
11'b3 ltJxc5 15 dxc5 ltJc4 16 �xc4
dxc4 1 7 11'xc4 �e6 1 8 11'e2 11'c6 1 9
e4 11'xc5 with an edge for Black,
Moiseev-Bronstein, Moscow 1 968 .
l b) 1 0 e4 (a bit embarrassing as
White is simply a tempo down on a
respectable line where Black plays
. . . d6 and . . . e5 against e4) 1 0 ... liJh5
1 1 lle l f5 ! ? (Black puts his extra
tempo to violent use; l l . . . liJf4 1 2
�fl liJb6 i s a solid alternative) 1 2
exf5 gxf5 1 3 dxe5 dxe5 l 4 liJd4 liJb6
1 5 liJb5 liJf4 1 6 �fl 11'f7 with a dou­
ble-edged position in which I prefer
Black, de Guzman-Gutierrez, Ma­
nila 1 99 1 .
2) 6 �e2 11'e8 7 0-0 liJbd7 (7 ... e5
8 c3 liJc6) 8 c3 (8 c4 is more aggres­
sive but doesn't really fit with liJbd2)
8 . . . e5 9 dxe5 dxe5 1 0 e4 h6 1 1 �h4
ltJc5 1 2 11'c2 �d7 1 3 llfe l a5 14
�g3 liJh5 1 5 ltJc4 liJxg3 1 6 hxg3 b5
1 7 ltJe3 c6 +, Kaber-Lutz, Bundes­
. liga 1 990.
3) 6 �c4 liJbd7 7 0-0 es (7 . . .11'e8
would avoid ' 3a') with a couple of
examples:
3a) 8 dxeS dxe5 9 ltJe4 11'e8 1 0
liJxf6+ �xf6 1 1 e4 11'e7 1 2 �xf6
liJxf6 was soon agreed drawn in
Dreev-Khalifman , Vilnius 1988.
3b) 8 c3 11'e8 9 b4 a5 10 �b5 c6
1 1 �e2 lD<l5 1 2 �c4 liJ7b6 ( the al­
ternative 1 2 . . . ltJxc3 1 3 11'b3 axb4 1 4
11'xb4 is not so clear) 1 3 �xd5 liJxd5
14 1l'b3 h6 15 .i.h4 �e6 1 6 c4 liJxb4
1 7 a3 liJa6 1 8 11'xb7 exd4 1 9 exd4
1l'd8 23 1l'xd6 1l'xd6 24 �xd6 llfd8
25 �g3 llxd4 with an excellent end­
ing for Black, Kraut-Mohr, B undes­
liga 1 990.
S .•• dS! (D)
This strong central thrust seems to
equalise effortlessly (the sort of
words which can return to haunt one)
and has cast a cloud over the natural
5 e4.
6 �d3
The alternatives have not brought
White much joy either:
1 ) 6 � xf6 (this relieves the cen­
tral pressure but at quite a high price;
it has, though, twice been the choice
of Salov) 6 . . . �xf6 (6 . . . exf6 is als o
interesting : 7 �e2 dxe4 8 liJxe4 liJd7
9 0-0 f5 10 liJed2 c5 1 1 c3 cxd4 1 2
liJxd4 liJc5 was about equal in Guse­
inov-Petrushin, Tallinn 1 983) 7 exd5
(after 7 e5 �g7 Black is ready to
strike back with both . . . f6 and . . . c5,
whilst 7 c3 �g7 8 exd5 11'xd5 9 �c4
11'f5 { 9 . . . 1l'd8 } 1 0 0-0 c5 1 1 lle l
cxd4 1 2 liJxd4 11'c5 1 3 11'e2 e6 was
level in Salov-Ye Jiangchuan, Til­
burg rpd 1 994) 7 . . . 11'xd5 8 �c4 1l'd8

9 c3 lt:Jc6 ( .. . e5 is coming) 10 li.Je4
i.g7 1 1 0-0 i.g4 (l l . .. e5) 12 h3 i.xf3
1 3 1ifxf3 e5 14 lt:Jc5 ! exd4 1 5 lt:Jxb7
lt:Je5 ! 1 6 1ife4 1iff6 17 i.b3 dxc3 1 8
bxc3 l:tab8 1 9 l2Jc5 with equality, Sa­
lov-Hebden, Moscow 1 986.
2) 6 es llJe4 with several possible
moves:
2a) 7 i.d3 i.f5 (7 i.d3 was rec­
ommended by Hodgson but he just
g ave 7 . . . lt:Jxg5 8 lt:Jxg5 c5 9 h4 ! , al­
though even here 9 . . . c4 followed by
... f6 looks OK for Black) 8 i.f4 c5 9
dxc5 lt:Jc6 1 0 lt:Jb3 (I assume that in
reply to 1 0 0-0 Black would have
simply played 1 0 . . . lt:Jxc5 as after 1 1
i.xf5 gxf5 he has a strong central
grip in return for his slightly exposed
king position) 1 0 . . . f6 ! ? (Black pre­
fers to face the future with a big cen­
tre as opposed to spending some
time recuperating the pawn� I ' m
sure, though, that a case could also
be made out for 1 o . . 1ifc7 ) 1 1 exf6
. 1 0 ... 1ifc6 1 1 0-0 lt:Jxc3 12 1ife l i.e6 !?
i.xf6 1 2 c3 e5 1 3 i.h6 :n 14 0-0
1ife7 1 5 1ifc2 l:td8 1 6 l:tae l i.g7 1 7
i.xg7 �xg7 1 8 lt:Jfd2 lt:Jxc5 1 9
i.xf5 gxf5 20 f4 e4 2 1 l:te3 lt.Je6 with
advantage to Black, Vaisman-Mag­
erramov, Nimes 1 99 1 .
2b) 7 i.e3 cS and now:
2bl ) 8 c3 cxd4 9 cxd4 l2Jc6 1 0
i.e2 1ifb6 1 1 1ifb3 lt:Jxd2 1 2 lt:Jxd2
i.e6 (not 1 2 . . . lt:Jxd4 ? 1 3 1ifxb6 l2Jc2+
1 4 �d l lt:Jxe3+ 1 5 1ifxe3 ! - it may
seem trivial to include a note like this
but it's the sort of thing that can be
e asily overlooked in practice) 1 3
1ifxb6 axb6 14 0-0 f6 15 exf6 i.xf6
16 lt:Jt1 i.g4 with a quite satisfactory
p osition for Black, Van Beers-Ro­
gers, Ostend 1 992.
The Torre Attack 139
2b2) 8 dxcS l2Jc6 (8 . . . lt:Jd7 !?) 9
lt:Jxe4 dxe4 1 0 1if xd8 l:txd8 1 1 lt:Jd2
lt:Jxe5 1 2 h3 i.e6 ! ? ( 1 2 . . . f5) 1 3 a3
( 1 3 lt:Jxe4 l2Jc4 14 i.xc4 i.xc4 15 c3
and now 15 . .. i.d5 !? or 15 . . . i.d3 1 6
lt:Jd2 e5 1 7 0-0-0 f5 give Black good
play for the pawn although it is still
slightly surprising that he preferred
this to 1 2 ...f5) 13 .. .f5 14 0-0-0 h6 15
i.e2 g5 16 g3 lt:J g 6 17 i.h5 i.f7 1 8
lt:Jc4 lt:Je5 1 9 i.xf7 + �xf7 20 lt:Ja5
lt:Jc6 ! and Black was better in C.Hor­
vath-Khalifman, Leningrad 1 9 89 as
2 1 lt:Jxb7 is met by 2 1 . . . i.xb2+.
2c) 7 i.f4 c5 8 c3 lt:Jc6 9 dxc5
lt:Jxc5 1 0 lt:Jb3 lt:Ja4 1 1 1ifd2 i.g4
with a strong initiative for Black,
S�rensen-Hebden, London Lloyds
Bank 1 99 1 .
3 ) 6 exdS ltJxdS with a couple of
examples:
3a) 7 c3 c5 8 dxc5 1ifc7 9 i.c4
1ifxc5 1 0 b4 ? (an incredible move)
1 3 i.xe6 1ifxe6 14 1ifxe6 fxe6 with a
clear advantage to Black, Manor­
Smirin, Tel Aviv 1 99 1 .
3b) 7 lt:J b 3 h6 (7 . . . a5 ! ?) 8 i.d2
lt:Jd7 9 i.e2 e5 1 0 dxe5 lt:Jxe5 1 1 0-0
( 1 1 lt:Jxe5 i.xe5 1 2 i.xh6 l:te8 1 3 c4
1ifh4 is dangerous for White) 1 1 . . . c6
1 2 c3 1ifc7 1 3 l:te l lt:Jf4 1 4 lt:Jxe5
i.xe5 15 i.fl 1/z- 1/z Bareev-Khalif­
man, Wijk aan Zee 199 5.
6... dxe4
7 ltJxe4 ltJxe4
8 i.xe4 cS
9 dxcS
9 c3 led to a quick defeat for
Whi te in Dunne-Wolff, Philadel­
phia 1 99 1 : 9 . . . cxd4 1 0 lt:Jxd4 1ifa5 !
1 1 1ifd2? ! ( 1 1 i.d2 e5 1 2 lt:Jb3 1ifc7
140 The Torre Attack

holds no hardship for Black but is grandmaster who would take the
better than the text which leads to big white pieces, given the choice.
problems on the d- file) 1 1 . . . :dS ! 1 2 12 .•• b6
0-0-0? (suicide; 1 2 0-0 is met by The bishop clearly belongs on the
1 2 . . . .txd4 1 3 b4 .txc3 !, so White long diagonal and it is worth talcing a
should probably try the ugly 1 2 b4 little bit of time to develop it effec­
although 1 2 . . . 'it'b6 ! looks like a good tively.
reply) 1 2 . . ..txd4 1 3 cxd4 'it'xa2 1 4 13 :tel eS!
d 5 .tg4 (threatening . . . :tcS+) 1 5 More to the point than 13 :teS, ...

'it'b4 lba6 1 6 'it'xe7 :tacS+ 1 7 '.td2 whereafter 1 4 'iie 2 .tb7 1 5 :tad l a6

'it'xb2+ l S '.te3 :teS 1 9 'it'f6 'it'e2+ 1 6 'it'e3 :tacS 1 7 :te2 b5 1 S :ted2
0- 1 . .t aS was roughly level in Cifuentes­
9 'it'c7 Wolff, Wijk aan Zee 1992.
10 c3 lba6 ! 14 'it'cl?!
1 1 0-0 lbxcS White intends to exchange dark­
12 .tel (D) squared bishops but in doing so he
seriously compromises his position.
14 'it'e2 looks better.
14 ... .tb7
15 lbh4?
When I first played over this game
I was astounded by this move. White
w as obviously not very impressed
with 15 'it'e3 f6 1 6 .th4, but this was
the course he had to follow.
15 :tfe8
16 .th6 .tf6
17 .tgS .tg7
"White is better due to his queen­ 18 .th6 'it'e7
side pawn majority" is a familiar 19 .txg7 '.txg7
phrase from my chess youth and also 20 g3
a great fallacy. The benefits of a A sad necessity.
queenside pawn majority can nor­ 20 ..• 'it'f6
mally only be appreciated deep into 21 'it'e3 :tad8
the endgame and before he gets there 22 b4? !
White (in this case) will have to ne­ By fatally weakening his queen­
gotiate a tricky middlegame where side White invites the coming com­
Black possesses an extra central bination. 22 :tadl was natural when
pawn. I'm not going to go as far as to Black would have to find 22 . . . e4 ! to
suggest that Black is better in the maintain a serious advantage. Then
diagram position, but I think you after 23 :txdS :txdS 24 lbg2 lbd3 ,
would be hard-pressed to find a 2 5 :tbl loses to 25 . . . lbxf2 !, 25 :tdl
 The Torre A ttack 141

to 25 . . . llJxb2 26 :xd8 \i'xd8 27

i.x e4 \i'd 1 + 28 llJe l llJc4 and 25 w

i.xd3 exd3 re-opens the long diago­

nal with predictable consequences.
22 •.• \i'c6
23 f3 (D)

There is no rush for Black to con­

vert his positional advantage into a
material one. As long as White is de­
nied counterplay the pawns will soon
start dropping off. The text, apart
from centralising the king, prevents
llJg5.
23 ... e4! ! 32 lbd4
24 fxe4 32 lbd2 i.d3 and 32 :et i.xf3
The main line runs 24 bxcS exf3 both lose at once.
25 \i't"2 (25 'ti'xf3 \i'xc5+ 26 \i'f2 32 ... :ct !
\i'xf2+ 27 �xt"2 :d2+) 25 . . . :e2! 26 33 llJbS?
:xe2 fxe2 27 llJg2 :d2 and Black Obviously a blunder but after 33
wins. :e3 :h1 34 llJf3 :a 1 35 :e2 :c 1 36
24 . .. llJxe4 :e3 :c2+ Black wins a pawn with­
25 i.xe4 :xe4 out relinquishing any of his posi­
26 \i'xe4 \i'xe4 tional trumps.
27 :xe4 i.xe4 33 i.d3
28 :et rs 34 llJxa7 i.xe2
W hite has made it to the ending, 35 �xe2 :xc3
but one in which he will require a 36 bS :c2+
miracle to survive. His queenside is 37 �f3 0-t
ripe for plucking whilst his knight is
a pitiful creature in comparison with Game 22
Black's majestic bishop. Yusupov - Kasparov
29 �f2 :d2+ Riga 1995
30 :e2 :d t
3 t llJf3 t d4 llJf6
Of course 3t :et loses to 3 t . .:xel
. 2 llJf3 g6
32 �xel g5. 3 i.gS i.g7
3t ..• �f6! (D) 4 c3 0-0
142 The Torre Attack

s lbbd2 d6
S dS is an important alternative
•..

even if such a move is against the

nature of many King 's Indian play­
ers. Althou gh I' m not really recom­
mending it, I feel that a summary of
the current theoretical situation is
worthwhile. 6 e3 gives Black three
possibilities, of which line 3 is by far
the most important (at the moment):
1 ) 6 b6 7 a4 ! (the most annoy­
••.

ing move for Black to face) 7 . . . c5 S 3b) 9 b4 (with his rock-solid cen­
i.d3 i.a6 9 i.xa6 lbxa6 IO 0-0 'fie? tre it's logical for White to expand on
I I 'fie2 'fib7 1 2 h3 :res I 3 lLJe5 ;!;; the wing) 9 c6 with several tries for
•.•

Malaniuk-Gufeld, Calcutta 1 993. White:

2) 6 c6! ? 7 i.e2 i.g4 S 0-0 lLJbd7
•.•
9 b4 (9 h3 ! ? i.xf3 I O lLJxf3 lLJe4 I I
i.f4) 9 ... i.xf3 ! ? (9 . . . a5 IO b5 a4 I I
ltc I ltcS I 2 c4 'fia5 I 3 h3 i.xf3 I 4
i. xf3 e6 I 5 bxc6 bxc6 I 6 'fic2 ;;t;
Kamsky-Anand , Las Palmas 1 995)
1 0 lbxf3 ( 1 0 i.xf3 e5 I I e4 h6 I 2
i.xf6 lLJxf6 1 3 dxe5 lbxe4 = Be­
liakov) 1 0 . . . lLJe4 I 1 'fib3? ( I I ltc 1
= ) Beliakov-Gudzovaty, Yalta 1 995,
, Emilia I 993 continued I 6 bxa5 ltxa5
and now Black could have picked up
an exchange by l 1 . . . h6 ! 1 2 i.f4 g5
1 3 i.g3 g4 I4 lLJh4 lLJd2.
3) 6 lbbd7 (Black plans . . . :es
•..
and ... e5) 7 i.e2 (7 i.d3 lteS S c4 { S
0-0 e5 = } S . . . e5 9 cxd5 exd4 1 0
lbxd4 lbb6 I I lbe4 lbbxd5 was at
least for Black in B arbero-Gal­
= i.h4 dxc4 I 3 lbxc4 lbb6 I 4 ·'fib3 g5
lagher, San Bernardino I 99I while 7
b4 c6 S i.e2 lteS 9 0-0 e5 is consid­
ered later) 7 lte8 8 0-0 es (D) and
• •.
now:
3a) 9 dxeS (timid) 9 . . . lbxe5 I O
lbxe5 ltxe5 I I lLJf3 lteS 1 2 a4 h6 1 3
i.xf6 .ixf6 I 4 'fib3 c6 1 5 a5 a6 1 6
ltfd l 'fic7 1 7 c4 dxc4 I S i.xc4 i.e6
= Hug-Adams, Biel IZ I 993 .
3bl ) 10 i.h4 a5 I l a3 e4 I2 lLJel
h6 1 3 lbc2 lLJfS I 4 c4 g5 I 5 i.g3
lLJg6 i s a fairly typical p osition for
this line. White hopes that his queen­
side advance will create weaknesses
to attack or some entry squares into
the heart of the opponent's position,
while Black is dreaming of glory on
the kingside. Salov-Gelfand7 Reggio
I 7 lbb4 ltaS I S cxd5 lLJxd5 I 9 lLJxd5
cxd5 20 'fic2 lte6 2 I ltfc I ltc6 22
'fib3 f5 23 i.h5 lLJfS 24 h3 lLJe6 25
'fid I f4 26 i.h2 'fid6 27 a4 Jlxc I and
a draw was agreed in a position
where I prefer Black.
3b2) 1 0 c4 exd4 I I lbxd4 h6 I 2
I5 i.g3 lLJe4 I 6 ltfd I 'fie7 1 7 ltac I
ltdS was at least equal for Black in
Kallai-Gyorkos, Hungarian Club Ch
1 993 .
3b3) 10 ltcl a5 (10...e4; IO...'fie7)
1 1 b5 c5 1 2 dxe5 lbxe5 I 3 c4 ! lbxc4
I 4 lbxc4 dxc4 I5 'fixdS ltxdS I 6
i.xc4 h 6 (Malaniuk-Stohl, Brno
I 993) and now Stohl gives I7 i.xf6!

i.xf6 I 8 ltfd I ltxd I+ I 9 ltxd I i.g4
20 ltd6 r.itg7 2 I ltJd2 ltd8 22 ltxd8
i.xd8 23 f3 i.d7 24 a4, wi th i.d5
and lL\c4 to follow as ;;!;.
3c) 9 ltJb3 ! ? . This modest-look­
ing move of Miles's has posed Black
the most problems recently. The idea
is to vacate d2 for the king's knight
from where it will support c4 and
prevent an annoying . . . ltJe4 (after
. . . h6 and . . . g5 for example). Miles­
Nunn, London Lloyds Bank I 993
now continued 9 . . . c6 (9 . . . a5 IO a4 c6
I I c4 exd4 I2 ltJbxd4 ltJb6 I 3 ltJd2
h6 I4 i.h4 g5 I 5 i.g3 fie7 I 6 f/c2
dxc4 I7 i.xc4 ltJe4 I 8 ltJxe4 fixe4
I 9 i.d3 was quite good for White in
Stangl - Har-Zvi, Altensteig I994)
IO ltc I a5 ( 1 0 ...fib6? ! I I ltJfd2 ltJf8
I 2 dxe5 ltxe5 1 3 i. f4 lte8 I 4 c4 ;t
Miles-Gdanski, Iraklion I 993) I I c4
a4 (in his notes Miles points out a
couple of other ideas for Black -
I l . . . exd4 or I l . . . dxc4 I2 i.xc4 a4
I 3 ltJbd2 exd4 I 4 ltJxd4 ltJb6 - but
these remain untested) I2 ltJbd2
exd4 I 3 ltJxd4 fia5 I 4 cxd5 fixd5
I 5 i.f4 ltJe5 I 6 fic2 ( 1 6 h3 would
have been smoother) I 6 . . . i.g4 I 7
i.c4 ! f/a5 I 8 h3 i.d7 I 9 i.e2 ;!;
ltac8 ? (the start of a dubious plan)
20 ltfd I b5 2 I ltJ2f3 ! ltJxf3+ 22
i.xf3 ltJd5 23 i.d6 fib6 24 fic5
fixc5 25 i.xc5 ltJf6 26 i.a3 and the
black c-pawn dropped off ( 1-0, 45)
6 e4 cS (D)
7 dxcS
The standard reaction. 7 dS h6 8
i. h4 e6 does n ' t look very promis­
ing for White whilst other moves
are sli ghtly frowned upon because
White ' s d-pawn may become weak,
The Torre Attack 143
for example: 7 i.d3 cxd4 8 cxd4 h6
9 i.h4 ltJh5 1 0 0-0 gS 1 1 i.g3 ( I I
ltJxg5 hxg5 I2 fixh5 gxh4 I3 e5 lte8
I 4 ltJf3 looks dangerous for Black,
but I I . . . ltJf4 ! is a much safer way to
win a piece) 1 1 ...g4! ? ( I I . . . ltJc6 I 2
d 5 ltJb4 I3 i.c4 a5 i s a safer, but
probably not better, way to play) 12
ltJh4 ltJxg3 13 hx g3 i.xd4 14 ltJfS
i.xfS 15 exfS hS 16 f6 (otherwise
Black will blockade with . . . ltJd7-f6)
and now:
I ) Timman-Topalov, Belgrade
I 995 continued 16 ltJd7 I 7 fxe7
•.•
fixe7 I 8 i.f5 ltJc5 I 9 ltJc4 f/f6 20
fid2 d5 2 I ltJe3 tlJe4 22 i. xe4 dxe4
23 ltJd5 f/e5 24 ltad I fixd5 25
fixd4 fixd4 26 ltxd4 f5 and White
was able to hold the ending.
2) I hesitate to suggest 16 ...i.xf6
to you as Timman must have planned
something and Topalov must have
had his reasons for rejecting it, but
let's just say that I can't see anything
very convincing for White, e.g. I 7
fia4 ( I7 ltJe4 i.g7 looks good for
Black) I 7 . . . d5 I 8 f/f4 and now Black
should avoid 18 eS on account of
•.•
I 9 fih6 e4 20 ltJxe4 ! when he gets
mated and play instead 18 r.itg7 as
•.•
I 9 f/f5 is simply met by I 9 .. . lth8.
144 The Torre Attack
7 ••• dxcS
In such a position Black's first pri­
ority will be to establish some con­
trol over e5 so that White can't
advance his e-pawn under favour­
able circumstances. The radical way
to do this is to play . . . e5 himself but
this leaves a gaping hole on d5 ;
therefore it is preferable for Black to
achieve his aim through piece play.
In the longer term Black will be
looking to expand on the queenside
or to gain the bishop pair, whilst oc­
casionally he may just have to react
to whatever action White has taken.
8 i.e2
Sometimes White plays 'iVc2 first
but this just transposes. A slightly
different idea is 8 i.c4. After 8 fbc6
•..
(D) we have:
1) 9 0-0 fba5 ! ? 1 0 i.e2 i.e6 1 1
.:tel ( 1 1 i.e3 'iVc7 1 2 ibg5?! i.d7 1 3
f4 h6 14 ibgf3 ibg4) 1 1 . . . a6 1 2 'iVc2
( 1 2 a4 is not on due to the weakness
on b3) 1 2. . . b5 13 ibb3 ibxb3 14 axb3
'iVc7 (the immediate 14 . . . 'iVb6 may
be better) 15 i.h4 h6 16 lbd2 'iVb6
1 7 ibfl .:tfd8 1 8 ibe3 .:ta7 1 9 .:tad l
.:tad7 with a roughly level game,
Malaniuk-Marin, Calimanesti 1 992.
2) 9 'iVe2 'iVc7 1 0 0-0 h6 1 1 i.h4
ibh5 12 'iVe3 lba5 1 3 i.e2 ( 1 3 i.d3
would transpose into Zilberman­
Yurtaev, Frunze 1989, where after
1 3 . . . .:td8 14 i.c2 g5 15 i.g3 lbxg3
1 6 fxg3 ! ? i.e6 1 7 e5 'iVc6 18 lbe4
lbc4 ! 1 9 'iVxc5 'iVxc5+ 20 lbxc5
i.d5 Black had achieved good play
in return for what is almost certain! y
a temporary pawn sacrifice) 13 . . . ibf4
14 ibb3 fbxe2+ 1 5 'iVxe2 ibxb3 16
axb3 i.e6 = Espig-Reeh, Potsdam
1988.
8
•.. lbc6
9 0-0
In Sharif-A. Kuzmin, Doha 1993,
White tried 9 h3 'iVc7 10 ibh2 hop­
ing to develop a quick kingside at­
tack. However, after 10 . . . ibd8 ! 1 1
i.e3 lbe6 1 2 g3 b6 1 3 'iVc2 i. b7 14
lbg4 .:tac8 ! ? 15 ibh6+ r.ith8 16 i.c4
(the point of Black's mysterious 14th
move is revealed after 16 0-0-0
ibd4 ! ) 16 . . .'iVc6 17 f3 b5 he must
have wished that he had treated the
opening more conventionally.
9
••• 'iVc7
More accurate than 9 h6 which
•••
gives White the extra option of i.f4.
10 'iVc2 (D)
10 .:tel .:td8 1 1 'iVc2 transposes.

10 ... ltd8! ?
A clever new move retaining the
option of developing the queen ' s
bishop at e6 or on the long diagonal,
according to Stohl. Whilst this is true,
I suspect that Kasparov had some­
thing much more devious in mind
(see the note to White' s 1 2th move).
The main alternative is 10 h6, after
..•
which White has the usual choice:
I ) 1 1 i.h4 lLJh5 1 2 ltfe l i.e6 1 3
i.fl ltad8 1 4 lDc4 ltd7 1 5 lted l
:fd8 1 6 ltxd7 ltxd7 I 7 lLJe3 lbe5 1 8
lbxe5 i.xe5 1 9 i.g3 lbxg3 20 hxg3
c4 with an edge for Black, Kiselev­
Yurtaev, B arnaul 1 98 8 . This is simi­
lar to the main game.
2) 1 1 i.e3 b6 1 2 h3 i. b7 (after
1 2. . .lLJh5 1 3 ltfel lLJf4 14 i.fl g5 1 5
a4 ltd8 1 6 lbc4 lbg6 1 7 a5 White
had a lot of pressure, Ye Rongguan g­
Wang Zili, Chinese Ch 1994) 1 3 lLJh2
(White prepares f4 and e5, a typical
plan for this variation; there is no
need for Black to panic, though, as
once White plays e5 he will obtain
squares of his own) 1 3 . . . ltad8 14 f4
e6 1 5 r.iih I \i'b8 (giving the option of
. . . \i'a8 ) 1 6 ltad l ltfe8 17 e5? ! lbd5
18 i.gl lbce7 ! I 9 lbe4 lLJf5 (White's
next move loses by force but he is al­
ready in grave difficulties having
boxed his own king in) 20 ltf3
lbxf4 ! ! 21 ltxd8 ltxd8 22 lLJg4 (22
ltxf4 i.xe4 with . . . lbg3# to follow)
22 . . . i. xe4 23 \i'xe4 lbxe2 24 \i'xe2
h5 and Black soon won, Malaniuk­
Tkachev, London 1 994. A fine per­
formance from Black against a
leading specialist in the Torre Attack.
1 1 ltfel h6
12 i.h4
The Torre Attack 145
The alternative is 12 i.e3, when
the normal move is 1 2 . . . b6, but I'm
sure that Kasparov had prepared
12 lbg4! . After 13 i.xc5 ( D):
•.•
I ) Stohl gives 13 b6 14 i.a3
.•.
lLJd4 1 5 \i'c I lbxe2+ 1 6 ltxe2 as � '
but I doubt the validity of this assess­
ment as after 1 6 . . . i.a6 1 7 ltel ( 1 7
c4 ltac8) 1 7 . . . lbe5 ! 1 8 lbxe5 i.xe5
1 9 lLJf3 ( 1 9 g3 is very weakening)
19 . . .i.f4 20 \i'c2 i.d3 Black is ex­
tremely active and the white queen
short of squares.
2) Whilst variation ' I ' is quite at­
tractive ifs also academic because
13 lLJceS! is extremely strong, e.g . :
•••
2a) 14 lbxe5 (best) 1 4 . . . i.xe5 1 5
i. xg4 i.xh2+ 1 6 r.iih I i.xg4 gives
Black some advantage as 17 i.xe7?
fails to 1 7 ...ltxd2! 1 8 \i'xd2 i.f4 .
2b) 14 i.a3 lLJxf2 ! 1 5 lLJxe5 ( 1 5
r.ii x f2 fails against 1 5 ... lbg4+ 1 6
�g l \i'b6+) 1 5 . . .\i'b6 ! 1 6 r.iif l (what
else?) 1 6 . . . lbh3 ! and White will be
mated.
2c) 14 i.d4 loses a piece and 14
i.e3 is too sick a move to analyse.
12 •.• llJhS!
A key move whenever the white
bishop drops back to h4 . i. g3 is
146 The Torre Attack

prevented and Black has the option outcome is equality (and in addition,
of . . . ltJf4 or . . . g5 . Yusupov had to contend with the in­
13 ltJc4 i.e6 timidating flourish that Kasparov
14 ltJe3 must have played 1 8 . . . b5 with).
According to Stohl White should 19 c4
have played 14 a4 to prevent Black's 20 .:edl ltJh5
queenside expansion. He then gives 21 i.g3 ll\xg3
1 4 . . . ltJf4 1 5 i.fl g5 1 6 i.g3 i.xc4 22 hxg3 'iib6
1 7 i.xc4 ltJe5 1 8 ltJxe5 i.xe5 =. Black has some advantage as his
14
••. llJf4 bishop has more mobility and more
15 i.n llJe5 targets than White's. These sort of
16 ltJxe5 i.xe5 opposite-coloured bishop positions
17 ltJc4 i.xc4 are extremely unpleasant to defend,
18 i.xc4 b5! (DJ although Yusupov did a very good
job until he cracked up just before
the end.
23 a4! a6
23 i.xg3 24 axb5 allows White
•..

counterchances.
24 axb5 axb5
25 .:xa8 .:xa8
26 g4 e6
27 i.e2 .:a2
2s �n i.b8
Intending . . . i.a7.
29 'ifd2
19 i.n Avoiding 'ifd8+.
1 9 i.xb5 c4 (threatening . . . ltJd3) 30 g5 h5
20 i.g3 .:ab8 (20 . . . ltJd3 2 1 i.xe5 31 g3 i.e5
'ifxe5 22 i.xc4 ) 2 1 a4 ltJd3 22
= 32 'ifd7 .:as
i.xe5 'ifxe5 23 .:e3 ltJxb2 ! 24 'ifxb2 33 'ife7
a6 is about equal according to Stohl, Fta�nik points out that it would be
bu t Yusupov must have felt that this difficult for Black to increase his
line was too risky. There are many slight edge after 33 'ifd2. One idea
other dangerous ideas lurking just would be to try . . . .:hs and ...h4.
beneath the surface, 21. .:d6 to
•• 33 .•• .: a7
mention just one. I think the vast ma­ 34 'ife8?! (DJ
jority of players would have reacted White p laces his queen in a pre­
like Yusupov, accepting a slight dis­ carious position when, instead, 34
advantage rather than heading into .:d7 .:xd7 35 'ifxd7 would have
complications where it is easy to given him a tenable ending.
drop a piece and the best p ossible 34 ••• i. d6!
 The Torre Attack 147

represent his best chance. After

35 ... %le7 (not 35 . .. i.xe5? 36 %ld8, but
35 . . . i.e7 is safer) 36 'ilc8 (36 l:txd6
'ilxd6 37 'ilxb5 'ild5 !) 36 . . .i.xe5 37
l:td8 %lc7 38 %lg8+ �h7 39 ilf8 %la7
40 i.f3 (Stohl) 40 b4! White's pieces
...

are, thankfully, less menacing than

they appear.
35••• i.e7
36 l:thl? 'iih7 !
Netting the queen.
35 �g2 37 i.f3 :as
Perhaps Yusupov had originally 0- 1
intended 35 eS, and this would s till 38 e5 ilxf3+ !.
1 1 The London System
This chapter deals with all the lines
where White plays an early .if4 (ex­
cept for the B arry Attack, Chapter
1 4). Game 23 examines an early c4
by White whereas Game 24 concen­
trates on the more cautious c3 . The
latter is a favourite amongst those
who disregard opening theory or
those who wish to bore you out of
your mind. Against both c3 and c4 I
am recommending that Black plays
for . . . e5, which is more attractive
here than in the previous chapter as
when . . . e7 -e5 is achieved it will gain
time hitting the bishop on f4 (admit­
tedly Black sometimes has to play
. . .'ife8 and ... fie7 in order to achieve
. . . e5 ) . The white bisho p usually
drops back to h2 from where it can
either play a pivotal role in a white
queenside attack or find itself com­
pletely out of play. Obviously we
shall be trying to bring about the lat­
ter.
Game 23
Yusupov - Thkmakov
USSR 1978
1 d4 llJf6
2 llJf3 g6
3 .if4 .i g7
4 e3
4 c3 is the next game whilst 4
llJbd2 also occurs from time to time.
It seems dubious for White to play an
early e4, though; Dominguez-Cv i­
tan, Novi Sad OL 1 990 continued
4 . . . 0-0 (after 4 llJbd2) 5 e4 d6 6 .id3
(6 .ie2 llJbd7 7 c3 llJh5 8 .ie3 e5 9
fic2 fie8 1 0 llJgl f5 1 1 g4 fx g4 1 2
.ixg4 llJf4 1 3 .if3 exd4 1 4 cxd4 c5
gave Black good play in Quinteros­
Uhlmann, Leningrad 1 973, although
it has to be admitted that White lost
his head a little) 6 . . . c5 ! (Black finds a
nice way of increasing the effect of
. . . e5) 7 c3 cxd4 8 cxd4 e5 ! 9 .ie3
exd4 1 0 .ixd4 ( 10 llJxd4 llJg4)
1 0 . . . llJc6 1 1 .ie3 d5 ( I l . . . llJg4 1 2
.i g5 f/b6 looks like an alternative
idea) 1 2 0-0 .ig4 1 3 .ig5 .ixf3 14
llJxf3 dxe4 15 .ixf6 fixf6 1 6 .ixe4
fixb2 17 .:t.bl fixa2 1 8 .:t.xb7 .:t.fd8
1 9 .:t.d7 llJd4 ! (otherwise the game
would peter out to a draw) 20 .:t.xd4
.ixd4 2 1 llJxd4 .:t.ac8 22 fib3 fixb3
23 llJxb3 %lc3 24 llJa5 .:t.dc8 and
B lack eventually won this favour­
able endgame.
4 ... d6 (D)
It's too early to chase the bishop
as after 4 llJhS 5 .ie5 f6 6 g4 !
.•.
Black is in trouble.
5 h3
Now, however, 5 .ie2 should be
met by 5 llJhS, e.g. 6 .ig S h6 7
...
.ih4 g S 8 llJfd2 llJf4 ! 9 exf4 gxh4
10 c3 cS when White has the alterna­
tives:
1 ) 1 1 dxcS dxc5 1 2 llJf3 fic7 1 3
llJxh4 .if6 1 4 llJf3 fix f4 with an

active game for Black, Lopushnoi­
Bologan, Kazan 1 995 .
2) 1 1 d S (this looks more testing)
l I . . . ibd7 1 2 0-0 ibf6 1 3 tbc4 h5 1 4
tbe3 �f8 (Black didn ' t play . . . i.h6
at once on account of 15 i. xh5+
i.xf4 1 6 'ifa4+) 1 5 ibd2 i.h6 1 6 g3
i.h3 1 7 .:t.e l hxg 3 ( 1 7 . . . .:t.gS ! ?) 1 8
fxg3 h4 with unclear p l ay in Rivas­
Romero, Leon 1 995.
5 ... 0-0
6 i. e2
6 c3 and 6 i.c4 are considered in
the next game whilst the immediate
6 c4 is very rarely played. It does,
however, pose us a slight problem as
Black will be unable to transpose
into the main line (White will play
tbc3 next move to prevent . . . tbe4).
One possibility is to play 6 tbe4! ? ..•
whilst another is to try 6...cS which
has more effect now that White can­
not play the deadly dull c3 in re­
sponse. A couple of examples of the
latter:
I ) 7 tbc3 cxd4 8 exd4 d5 9 i.e5
i.e6 (9 . . . tbc6 !?) 1 0 'ifb3 ibbd7 1 1
cxd5 tbxe5 1 2 dxe5 tbxd5 1 3 .:t.d I
tbxc3 1 4 .:t.xd8 i.xb3 1 5 .:t.xa8 .:t.xa8
1 6 axb3 4Jd5 1 7 i.c4 .:t.d8 1 8 0-0 e6
and Black has emerged from the
The London System 149
complications with a slightly better
ending in view of White's weak
pawns, Anastasian-Xu Jun, Beijing
1 99 1 .
2) 7 dS tbe4! looks very annoy­
ing for White (8 'ifc2 'ifa5+) but I
would also like to have a quick look
at 7 :es since after 8 llJc3 we have
.••
transposed to Sitanggang-Tkachev,
Djakarta 1 994, whose actual move
order (I d4 ibf6 2 c4 g6 3 lDc3 i.g7
4 ibf3 d6 5 h3 0-0 6 i.f4 c5 7 d5 :es
8 e3) is not examined elsewhere in
this book. The game continued 8 . . . e6
9 dxe6? ! (as White has no intention
of grabbing the hot pawn on d6 he
might have settled for 9 i.e2, when
9 . . . exd5 1 0 cxd5 a6 1 1 a4 tbe4 !?
would be an interesting way for
Black to continue) 9 . . . i. xe6 I 0 i.e2
d5 (of course) 1 1 0-0 h6 1 2 lbb5
tba6 1 3 ibd2 .:t.e7 1 4 i.g3 .:t.d7 1 5
'ifa4 lbb4 1 6 a3 tbc6 1 7 i.f3 d4 1 8
exd4 ibxd4 1 9 ibxd4 .:t.xd4 with ad-
vantage to Black.
6•.• ibbd7 (D)
,
7 0-0
7 c4 allows Black to play . . . e5
without any further preparation, e.g.
7 e5 ! 8 dxeS dxeS 9 tbxeS (9 i.xe5
••.
150 The London System
lDxe5 1 0 lDxe5 lDe4 1 1 "flxd8 l:xd8
1 2 lDd3 gives Black good play for
the pawn; 1 2 . .. l:xd3 is one possibility
while 1 2 .. . i.e6 is another) 9 ttlli5 !
•••
(D) and now:
1 ) 10 i.x h 5 ltJxe5 1 1 i.e2 ltJxc4
( 1 1 . .."i'lxd l+ and 1 2 . . . lDd3+ is also
good) 1 2 "flxd8 l:xd8 1 3 i. xc4 ( 1 3
lDc3 lDxb2 1 4 l:c 1 c6 is a safe extra
pawn) 1 3 . . . i.xb2 1 4 i.xc7 l:d7 is
winning for Black.
2) 1 0 ltJxd7 ltJxf4 1 1 lDxf8
lDxg2+ 1 2 �fl "i'lxd l + 1 3 i.xdl
i.xb2! and Black will pick up the
two trapped pieces and emerge with
the advantage.
7
•.• "fies
B lack aims to play . . . e5, more
logical than . . . c5 once . . . 4Jbd7 has
been played. 7 l:e8 doesn't help but
•.•
7 llJe4!? is certainly worthy of at­
•••
tention as Black may be able to save
a tempo on the main line, e.g. after 8
lDbd2 lDxd2 9 "i'lxd2 e5 1 0 i.h2
"fie? Black has achieved . . . "fie? in
one go. Other tries don't look too
convincing for White either. 8 i.d3
should be met by 8 . . . f5 whilst the
one game rve seen with 7 . . . ltJe4,
Ehrke-Volke, Munich 1 992, quickly
turned sour for White: 8 "fld3 f5 9
l:d l e6 1 0 c4 b6 1 1 llk3 i.b7 1 2
b4 ? ! e5 1 3 i.h2 lDxf2 ! 1 4 �xf2 e4
and Black was much better.
8 c4
8 c3 transposes to the next game.
8 •.. e5
9 i.h2 ltJe4
This seems to be the best move.
9 "i'le7 is quite often played but, in
••.
my view, this doesn' t really solve
Black's opening problems. After 10
lllc3 (D), there is:
1 ) 1 o c6 (this weakens the h2 -
•. •
b8 diagonal, but it's the only sensible
way to maintai n the tension in the
centre) 1 1 b4 ( 1 1 c5 ! ?) and now :
l a) 1 1 b6 1 2 c5 ! bxc5 1 3 bxc5
•.•
dxc5 14 lDxe5 i.b7 15 lDc4 l:fd8 1 6
"i'lb3 with a pleasant game for White,
Bellon-Tai, European Club Ch 1984.
lb) 1 1 exd4 12 exd4 ( 1 2 lDxd4)
...
1 2 ... d5 1 3 c5 llle4 ! 14 l:c l ( 14 lDxe4
dxe4 1 5 i.d6 "flf6 is fine for Black)
1 4 . . . f5 1 5 l:el l:f6 ! ? 1 6 b5 l:e6 17
bxc6 b xc6 18 i.fl h6 ( 1 8 . .. 4Jdxc5
1 9 4Jxd5) 19 "i'lb3 ! liJdf6 (Black
wasn't keen on allowing 20 llJe5 but
the d5-square needed bolstering;
both l 9 . . . 4Jg5 and l 9 . . . �h7 would

have been strongly answered by 20
llJxd5 ! ) 20 llJe5 'iti>h7 2 1 i.d3 llJd7?!,
Anastasian-Wang Zili, Beijing 199 1 ,
and now instead of 22 liJxdS, I be­
lieve that 22 llJxc6 would have been
a more promising piece sacrifice:
22 . . . .J:r.xc6 23 'ii'x d5 i.b7 24 llJxe4
fxe4 25 i.xe4 looks good for White.
2) 10 e4 l l lZXl2 and now:
•••
2a) I ' m not certain if I played
1 l ... b6? in this particular position or
in a very similar one - thankfully, I've
lost the scoresheet - but 1 2 llJdxe4 !
is very embarrassing for Black.
2b) 1 1. c6 1 2 b4 d5 1 3 cxd5
•.
cxd5 1 4 llJb5 llJe8 1 5 'ii'b3 is clearly
better for White according to Bellon.
2c) 1 1 .J:r.e8 (probably the best
•••
although White has some dangerous
piece sacrifices) 1 2 llJb5 'ii'd8 1 3 c5
a6 1 4 cxd6 ( 1 4 llJxc7 ! ? 'ii'x c7 1 5
lDc4 i s also mentioned by ECO)
1 4 ... axb5 1 5 dxc7 'ii'e7 1 6 i.xb5 'ii'b4
( 1 6 . .. i.f8 1 7 llJc4 'ii'e6, Spassky-Bu-
kic, Bugoj no 1 97 8 , and now 1 8
'ii'b 3 ! would have been very good for
White) 17 a4 .J:r.e6 1 8 'ii'c2 liJb6 19 b3
i.f8 with unclear p lay according to
ECO, Eslon-Gallego, San Sebastian
1984.
Before moving on it is also worth
mentioning Torben S0rensen's idea,
9 ttJhs. After 10 dxe5 ( 1 0 0-0 and
••.
1 0 g4 are clearly options) 10 . . . dxe5
1 1 liJb5 \i'd8 1 2 g4 llJhf6 1 3 llJxe5
llJe4 1 4 llJd3 'ii'h 4 1 5 i.f3 llJdc5 1 6
i.xe4 llJxe4 1 7 'ii'f3 i.d7 ! 1 8 llJxc7
( 1 8 'ii'xe4 i.c6) 1 8 . . . f5 ! 1 9 llJxa8
i.c6 Black had excellent attacking
chances in Fedder-T. S0rensen, Co­
penhagen Ch 1 992.
10 liJbd2
The London System 151
10 llJc3 is perhaps a more critical
move. After 10 llJxc3 ill bxc3 'ii'e7
.•.
( 1 l . .. b6 ! ? 1 2 a4 a5 ) 12 'ii'b 3 'iti>h8 13
.J:r.adl rs 14 'ii'a3 (D) :
1 ) Andrianov-Burger, New York
1 990 continued 14... e4 1 5 liJd2 g5
1 6 c5 f4 17 cxd6 cxd6 18 exf4 gxf4
1 9 f3 ! exf3 ( 1 9 . . . e3 20 llJe4 wins a
pawn and it is unlikely that Black
will be able to generate sufficient
counterplay) 20 i.xf3 'ii'e 3+ 2 1 'iti>h 1
i.xd4 22 i.g 1 (22 llJe4 looks very
strong) 22 . . . i.c5 23 i.xe3 i.xa3 24
i.d4+ llJe5 25 llJc4 i.c5 26 i.xc5
llJxc4 27 i.d4+ llJe5 28 c4 ! and the
powerful bishops gave White a clear
advantage.
2) In his notes to the above game
Andrianov suggests 14 cS as a pos­
. ••
sible improvement and this does in­
deed seem to be the case although
Black must still take care, e.g . 15
dxeS ( 1 5 dxc5 llJxc5) and now:
2a) 15 dxeS?! 1 6 .J:r.d5 llJb6 (per­
•••
haps 1 6 . . . b6 but Black is very tied
down after 1 7 l:ifd l ) 1 7 .J:r.xe5 ! (bu t
not 1 7 .J:r.xc5 ? llJa4 ! ) 1 7 . . . i.xe5 1 8
i.xe5+ 'iti>g8 1 9 .J:r.d 1 with tremen­
dous compensation in return for the
exchange.
152 The London System

2b) 15 ... ltlxe5 16 l:.d2 when: 16 ••• f4!

2bl) The natural move 16 i.e6 ••• 16 dS looks natural but Tukma­
• •.

allows White some advantage after kov points out that White could have
17 l:.fd l b6 1 8 1fa4 ! ( 1 8 l:.xd6 tlJxc4 then played 1 7 i.e5 ! followed by f4.
1 9 i.xc4 i.xc4 20 l:.d7 l:.fd8 ! is OK 17 exf4 g4
for Black), e.g. 1 8 . . . i.d7 19 'ifb3 18 hxg4
i.c6 20 l:.xd6 tDxf3+ 2 1 i.xf3 i.xf3 White could have reopened the
22 gxf3 l:.ad8 23 l:.d5 and Black diagonal for his bishop at once with
doesn' t have a great deal in return 18 f5! ?, when Tukmakov considers
for his pawn. 1 8 . . . gxh3 1 9 g4 llJe8 ! , intending to
2b2) But the solid 16 llJf7 ! 17
•• . play 20. . . h5, as unclear.
l:.fd l l:.d8 will make it very difficult 18 ..• llJxg4
for White to improve his position. 19 i.xg4
Black, on the other hand, can de­ There was little choice, e.g. 19
velop his queen's bishop (probabl y i.g3 e3 ! and 19 lbc2 llJxh2 20 'iii> xh2
on b7) and attack c3 with . . . 'ii'f6. i.h6 ! 2 1 g3 l:.xf4 ! .
Somebody good (it might have been 19 •.• i.xg4 (DJ
Bronstein) once commented that f7
is the perfect square for a knight in
the King's Indian.
10 llJxd2
11 'ii'xd2 e4!?
12 ltJel 1fe7
13 ltJc2 f5
14 b4 g5
The position demands that White
attacks on the queenside and Black
on the kingside, as is so often the
case in the King's Indian.
15 c5 ltJf6 20 .J:.ael dxc5
16 ltJa3 (DJ 20 l:.ad8! would have been bet­
• .•

ter according to Tukmakov. Now

White gets rid of his weak d-pawn.
21 dxcS l:.ad8
22 'ifcl l:.d3
23 l:.e3 l:.fd8
24 'ii'c4 + •n
25 'ii'xe4 i.fS
26 'ii'c4
White heads for the ending as 26
'ii'xb7 l:.xe3 27 fxe3 1fxa2 would be
very risky for him.
 The London System 153

26 ••• llxe3 s h3 o ..o

27 fxe3 lld2 6 e3
28 'ii'xf7+ �xt"I Or 6 tiJbd2 tiJbd7 7 e4 (on 7 e3 I
29 tlJc4 llxa2 quite like the idea of playing . . . b6
30 tlJeS+ .txeS and . . . .tb7 before playing for . . . e5)
112-1/2 and now:
l ) 7 es 8 dxe5 dxe5 9 tlJxe5 (9
••.

Game 24 .te3 'iie7 10 'ii'c 2 b6 1 1 0-0-0 .tb7

Zach - Bangiev was at least equal for Black in
Binz 1 994 Braga-B ass, Leon 1 990) 9 . .. tlJxe5
(or 9 . . . lle8 1 0 tlJxd7 tlJxe4 ) 10
=

1 d4 tiJf6 .t xe5 tlJxe4 1 1 .t xg7 lle8 1 2 tlJxe4

2 llJf3 g6 llxe4+ 1 3 .te2 'ii'xd 1 + 1 4 llxd 1
3 .tf4 .tg7 �xg7 with a level endgame. Note
4 c3 (D) that 1 5 lld8 leads nowhere in view of
1 5 . . . b6 threatening 16 ... llxe2+.
2) If you are not happy with a
dull endgame you can try 7 tlJhS .••

before playing . . . e5.

An extremely solid and unambi­
tious move which makes no attempt
to take the initiative. White' s princi­
pal aim is to block the King's Indian
bishop out of the g ame, but ironi­
cally, it's often his own bishop on h2
w hich ends up as a mere spectator.
There are various move orders that
White can employ but for practical
purposes it is easier if we get c3 out
of the way at once.
4••• d6
Better than castling as White now
has to deal with the positional threat
of 5 ... llJh5.
6 ...... tiJbd7
7 .te2
Occasionally White develops his
bishop more actively to c4 but this
need not dissuade us from our plan,
e.g. 7 .tc4 'ii'e8 8 0-0 e5 9 .th2 (9
dxe5 dxe5 1 0 .th2 b6 is clearly fine
for Black, while 9 . . . tlJxe5 has been
played several times in similar posi­
tions) 9 . . . b6 1 0 0-0 .tb7 with a satis­
factory position for Black. At some
point in the near future White is quite
likely to have a poke on the queen­
side with a4 against which Black
should probably respond with . . . a6.
7 ... 'ii'e8
8 0-0 es
9 .th2
We have the same position as the
previous game except that there is a
white pawn on c3 instead of c4; con­
sequently there is even less pressure
on the black position.
154 The London System

9 ..•
Wle7 23 dxc5 l:tb8 24 lDa6 i.xa6 25 i.xa6
10 a4 (D) l:td8 with advantage to Black, De­
noth-Gallagher, Chiasso 1 992.
1 1 lDrd2 l:te8
12 c4
12 a5 a6 1 3 b4 lDf8 1 4 lDa3 h5 1 5
b5 lD8h7 1 6 c4 lDg5 1 7 'iib 3 h 4 1 8
bxa6 bxa6 1 9 l:tac 1 i.e6 gave Black
quite a good attacking position in
J.Garcia-Vaganian, Dubai OL 1 986.
12 ..• lDf8
13 lDc3 hS
A standard move in such posi­
tions. Black can now defend his e­
10 .•. e4! ? pawn with ... i.f5 without having to
Black stakes his future on a king­ worry about g4 and he can also re­
side attack but in doing so liberates deploy his knight on f8 to g5 via h7.
the bishop on h2 and allows White a 14 l:tcl
free hand on the queenside. Less 14 iLJdS? just loses a pawn after
boat-burning alternatives are: 1 4 . . . lDxd5 1 5 cxd5 'iig 5 with a dou­
1 ) 10 &4 1 1 lDfd2 lDxd2 1 2
..• ble threat to h3 and d5.
lDxd2 f5 1 3 :C l �h8 1 4 a5 a6 with a 14 •.• c6
more or less level game, A.Hoffman­ 15 as tlJSh7
Epishin, St Barbara 1 992. 16 l:te l i.rs
2) 10 &8 is another method of
..• 17 Wla4 a6
preparing kingside play. A.Hoffman­ 18 i.fi (D)
C Foisor, Zaragoza 1 992 continued
1 1 a5 ( 1 1 c4 f5 l 2 lDc3 c6 1 3 'iic2 g5
is given by Bangiev) 1 1 ... �h8 1 2 a6
b6 1 3 i.b5 c5 1 4 lDa3 llJc7 1 5 i.xd7
i.xd7 1 6 dxe5 d5 ! 1 7 'iie 2 i.c8 1 8
lDc2 i.d7 1 9 lDa3 i.c8 and White
should have agreed to a repetition.
3) 10 �h8 1 1 lDa3 (perhaps 1 1
•.•

c4) l 1 . . . lDe8 1 2 b4 f5 1 3 b5 g5 1 4 c4
f4 (Black is not so much playing for
mate as to lock the bishop on h2 out
of the game; once this is achieved he
will be quite happy to counter on the White has now achieved his opti­
queenside) 1 5 lDc2 h6 1 6 l:ta3 llkif6 mum defensive position and is ready
1 7 exf4 exf4 1 8 l:tel 'iff7 1 9 a5 c6 20 to turn his attention to the queenside
bxc6 bxc6 2 1 lDb4 'iic 7 22 c5 dxc5 where he will hope to breakthrough
 The London System 155

before Black can arrange a success­ possible king moves, all of which
ful sacrifice on h3. The position is suffer the same fate:
difficult to assess.
t8 ... � h6
In his notes Bangiev preferred
t8 ..if8, which might have saved him
••

a tempo on the game continuation.

t8 ...lllg 5, or perhaps t8 .h4 first,
.•

also suggest themselves.

19 b4 'ii'd8
Holding up b5, which would now
lose the a�pawn. 19 g5 would have
••.

been the consistent follow-up.

20 lllb 3 �e6
2 1 lLJd2 d5 !? I ) 30 �g3, not surprisingly, is
Black declines the tacit draw of- swiftly dealt with by 30 . . . h4+.
fer. 2) 30 �gt ltlf3+ ! 3 1 gxf3 exf3
22 lllb 3 �f8! and despite his enormous material
23 liJc5 �c8 advantage White gets mated or loses
24 'ii'a2 �d6 all his pieces (I ' ve checked with a
Despite being theoretically the computer), e. g. 32 �xh3 'ii'h4 or 32
'bad' piece, White's dark-squared 1ia4 l:tf8 ! (not 32 . . . �xfl 33 'ii'd 7 ! ,
bishop was clearly outperforming its nor 32 . . . 'ii'h 4 3 3 'ii'xe8+ �g7 34
opposite number. Therefore it's a llle6+ ! ) and White has only delayed
good idea to exchange it off. the inevitable.
25 cxd5 ?! 3) 30 �h t �xg2+ ! 3 1 �xg2 (3 1
White should have taken on d6 �xg2 lllf3 32 �h3 'ii'd6 is all over)
first as after 25 �xd6 'ii'xd6 26 cxd.5 3 1 . . . 1id6 32 �c4 'ii'h2+ 33 �el l2Jf3
cxd5 27 lll3 a4 we have transposed and Black wins.
back into the game. 28 �gl lllg5
25 ... �xh2+ 29 �el
26 �xh2 cxd5 Perhaps White should have tried
27 lll3a4 'ii'd6+? ! 29 lllb6 . Obviously Black doesn' t
27 lt:Jg5 ! was more accurate leav­
••. w ant to allow lllxc8 because without
ing Black more options with his his bishop he has little chance of a
queen. After 28 lllb6 (the logical fol­ successful kingside attack, whilst
low-up whilst on 28 �e2, 28 . .. �xh3 29... �xh3 30 ltJxb7! ? 'ii'e6! 3 1 1ie2 ! ?
29 gxh3 'ires could be worth a try; i s open for debate.
the white knights are miles away from 29
•.• l:tb8
their king) 28 �xh3! 29 lllxa8 (2 9
••. 29... �xh3!? 30 gxh3 lllx h3+ 3 1
gxh3 lllf 3+ 30 �g2 'ii'd6 is good for �g2 lllxf2 ! looks quite promising.
Bl ack) 29 lllg4+ (D) White has three
.•• 30 lLJb6 �f5
156 The London System

The same sacrifice as in the pre­

vious note could have been played
here as well.
31 h4
Although Black never actually
plucked up the courage to sacrifice
on h3 things have not worked out too
badly for him. He has managed to
preserve his bishop and will have
new attacking opportunities based
on advancing his g-pawn and the
weakness of White's g4-square. 37 Ahl �g4
31 ••• llJgh7 38 �n �f3? !
32 bS 38 l:tg8 was more accurate.
•..

32 'iic2 might be a better try, in­ 39 l:tgl l:td8

tending to move the knight from c5 40 'jffb3?!
and try to exchange queens. 40 l:tg3 offered some defensive
32 axbS chances.
33 �xbS l:te7 40 ..• 'ii'hl !
34 g3 gS! (D) 41 �d7 l:texd7
The white king has every reason 4 1 l:tg8 also looks pretty termi­
•..

to feel displeased with his subjects; nal.

just when they are most needed the 42 llJcxd7 lLlg4
minor pieces have disappeared en 43 l:txg4 0-1
masse and the white queen finds it­ In view of 43 ... �xg4 44 �el 'iig l +
self cut off from the action. followed b y munching through the
35 �g2 gxh4 pawn chain with check and then talc­
36 gxh4 �h8 ing the knight on d7.
1 2 The Ki ngside Fia nchetto
This chapter deals with all the lines
where White fianchettoes his king's
bishop but doesn' t play c4. Instead
he usually aims to steer the g ame
into Pirc territory by playing e4, but
Black has no reason to fear this; even
if the Pirc is not part of his normal
repertoire the variations are quite
limited and there is an easy plan for
Black to follow. The main problem is
th at White can, and often does, ex­
change on e5, leading to deadly dull
positions. But one shouldn' t be too
disheartened as, with the right atti­
tude, virtually any position can be
won against weaker opposition and
against strong opponents . . . well, a
draw is not too bad with the black
pieces.
Game 25
Espig - Gallagher
Bad Worishofen 1 994
1 d4 ltjf6
2 llJf3 g6
3 g3 i. g7
4 i.g2 0-0
s 0-0 d6
Those of you who are happy to
p lay a Griinfeld or the symmetrical
line of the Fianchetto variation can
play S ... dS .
6 llJc3
White hopes to take the game into
a g3 Pirc, but Black should not be
afraid of this as the fact that the white
knight is on f3 makes it a slightly in­
ferior version for White. The most
popular alternative is 6 c4, transpos­
ing into one of the main lines of the
King's Indian which is outside the
scope of this book. All the other al­
ternatives are examined in g ame 26 ,
except for 6 b3 which we shall look
at here.
6 b3 (D) is certainly one of the
most tedious variations against the
KID but Black should still take great
care in the opening so as to avoid
slipping into a prospectless position
where White has a nagging edge. I
am going to examine a couple of
possibilities but, it has to be said,
there are no miracle solutions for
Black to liven the game up :
l ) 6 cS and now:
•.•
l a) 7 c4 d5 ! ? (7 . . . llJc6 8 i.b2
cxd4 9 lLlxd4 i.d7 is a solid alterna­
tive) 8 cxd5 lLlxd5 9 i.b2 lLlc6 10
158 The Kingside Fianchetto
"i'i'd2 ! (by defending his queen's
bishop White cuts out a lot of tricks)
1 0 . . . ltJc7 ! ( 10 . . . cxd4 l l ltlxd4 ltlxd4
1 2 �xd4 ltJc3 1 3 �xg7 "fi'xd2 1 4
ltlxd2 ltlxe2+ 1 5 �hl �xg7 1 6 llfel
is good for White) 1 1 lld 1 cxd4 1 2
ltlxd4 ltlxd4 1 3 �xd4 �xd4 1 4
"fi'xd4 "i'i'xd4 1 5 llxd4 lbb5 ! 1 6 lld5
ltld6 1 7 ltlc3 �e6 1 8 lldd 1 llfc8 1 9
llac 1 llab8 ! resulting i n a level
ending, A.N .Panchenko-Gallag her,
Bad WOrishofen 1 994.
l b) 7 �b2 cxd4 8 ltlxd4 dS 9 c4
(9 ltla3 e5 1 0 ltlf3 e4 1 1 ltld4 lLlc6
1 2 c4 transposes to variation ' 1 b2' )
and now:
1 bl ) 9 dxc4 1 0 bxc4 ( 1 0 ltla3 ! ?
•.•
cxb3 1 1 "i'i'xb3 "i'i'b6 1 2 ltlc4 "fi'xb3
1 3 axb3 ltlbd7 1 4 llfc l a6 1 5 llc2
llb8 1 6 llac l lle8 gave White some
pressure for his pawn in Bistric-Vogt,
Bugoj no 1983) 1 0 . . . "i'i'b6 1 1 ltlb3
( 1 1 "i'i'b3 is well met by l 1 . . . ltlfd7 ! )
l l . . . lld8 1 2 "fi'c l , Azmaiparashvili­
Kochiev, USSR 1 98 1 , and now Ko­
chiev gives 1 2 . . . ltJc6 1 3 lLlc3 �e6 as
the most comfortable for Black.
l b2) 9 ... eS 10 lLlf3 ( 1 0 lLlc2 dxc4
1 1 �xe5 cxb3 1 2 axb3 ltJc6 1 3 �b2
�e6) 1 0. . . e4 l l lLld4 lLlc6 ( 1 l . .. dxc4
1 2 bxc4 ltJc6 13 ltlxc6 bxc6 1 4 lLlc3
could be an edge for White) 1 2 ltla3
ltlxd4 1 3 "fi'xd4 �g4 1 4 "fi'e3 ( 1 4
cxd5 �xe2 1 5 llfel �d3 looks good
for Black, but 1 4 llfe l is possible)
1 4 ... "fi'e7 15 llabl �f5 16 h3 ? ! d4 !
1 7 �xd4? ! ( 1 7 11xd4 ltlh5 1 8 'ii'd2
e3 1 9 "fi'xe3 "fi'xe3 20 fxe3 �xb 1 2 1
llxbl offers a much better chance)
17 ..."fi'xa3 1 8 �c5 'ii'a5 19 �xf8 llxf8
with advantage to Black, Danielsen­
Deep Blue, Copenhagen 1 993.
2) 6 eS 7 dxeS and now (D):
. ..
2a) 7 lLlg4?! 8 �b2 lbd7 9 "fi'c l !
•.•
ltlgxe5 (or 9 . . . dxe5 1 0 h3 ltlh6 1 1
ltla3 lLlf5 1 2 lld 1 lle8 1 3 e3 lbd6 1 4
c4 e4 1 5 ltle l �xb2 1 6 "i'i'xb2 "i'i'e7
1 7 lld2 ;!; Stangl-Gallagher, Kecske­
met 1 990) 10 ltlxe5 ltlxe5 1 1 f4 ltlg4
1 2 � xg7 �xg7 1 3 h3 ltlf6 14 e4
with an edge for White, Smyslov­
Xie Jun, Prague 1995.
2 b) 7 lLlfd7 8 �b2 dxe5 9 e4 a5
...
1 0 a4 ltla6 1 1 ltlbd2 ltlac5 1 2 "fi'e 2
b6 1 3 llfd l �a6 1 4 "fi'e3 "fi'e7 with
an equal g ame, Dizdarevic-Kozul,
Zagreb Z 1 993.
2c) 7 dxeS with the branch:
.. .
2c l ) 8 �b2 (not 8 ltlxe5? ltlg4)
8 ... e4 9 "fi'xd8 llxd8 10 ltlg5 �f5 1 1
g4 ( 1 1 �h3 �xh3 1 2 lLlxh3 lLlc6 1 3
lLla3 lLld4 1 4 �xd4 llxd4 1 5 llad l
llad8 was very pleasant for Black in
Fuster-Gligoric, Portoroz IZ 1 958)
1 1 . . . � xg4 1 2 ltlxe4 ltlxe4 13 �xg7
�xg7 14 �xe4 lLlc6 15 ltlc3 ltld4 1 6
f3 �f5 17 �xf5 ltlxf5 and Black's
superior pawn structure gave him a
small advantage in Filip-Geller, Am­
sterdam 1956.
2c2) 8 �a3 lle8 9 ltlc3 ltlc6 1 0
ltlg5 �f5 1 1 ltlge4 ltlxe4 1 2 ltlxe4

· 1i'xd 1 1 3 l:fxd 1 lLJd4 1 4 l:d2 l:ad8
. 1 5 )f;>fl ! lDb5? ! ( 1 5 . .. i. xe4 !? 1 6
i.xe4 lLJb5 1 7 l:xd8 l:xd8 1 8 i.b2
c6 1 9 c4 !fJc7 and even though White
has the bishop pair he cannot claim
any advantage) 1 6 l:xd8 l:xd8 1 7
i.b2? ! ( 17 i.e7 ! gives White a clear
advantage after 1 7 . . . l:d4 1 8 llJf6+
)f;>h8 1 9 .lxb7 i.xc2 20 l:c 1 and an
edge after 1 7 . . .l:e8 1 8 lDf6+ i.xf6
1 9 i.xf6 lLJd6 20 c4 lLJe4 2 1 i. xe4
i.xe4 22 l:dl i.f5) 17 .. b6 18 e3 l:.d7
. e4 l:a6 !? 1 1 f3? ! c6 1 2 dxc6 bxc6 1 3
1 9 )f;>e 1 1h- 1h Ekstrom-Gall agher,
Swiss League 1 994.
6••• lLJbd7
Once White has blocked his c­
pawn a strong case may be made for
changing plans and playing . . . d5. I
am, however, sticking with the Pirc
set-up as our main line since this can
arise from a variety of 6th moves. It
is worth, though, examining a couple
of examples of 6 dS:•••
1 ) 7 l:el (White's most logical
plan is to play for e4) 7 . .. tlJe4 ! ?
(7 . . . c6 8 e4 dxe4 9 lDxe4 lLJxe4 10
l:xe4 lLJd7 i s an alternative) 8 i.f4
lLJc6 9 lLJb5 a6 ! 1 0 lLJa3 ( 1 0 lLJxc7
l:a7 and the knight on c7 will not es­
cape) 1 0 ... lLJxd4 1 1 lDxd4 e5 1 2 i.e3
exd4 1 3 i.xd4 i.xd4 1 4 1i'xd4 c5 1 5
1i'e3 l:e8 1 6 l:adl d4 1 7 1i'f4 tllg 5
1 8 lLJc4 i.h3 19 1i'd6 b5 20 flxd8
l:axd8 and a draw was agreed i n
Romanishin.. Magerramov, Helsinki
1 992 although after 2 1 lLJa5 i.xg2
22 )f;>xg2 lbe4 Black must have some
ad vantage.
2) 7 tlleS c6 8 e4 i.e6 9 exd5
cxd5 1 0 tlJe2 llk6 1 1 tllf4 i.f5 1 2 c3
i.e4 1 3 i.h3 1i'c7 1 4 lDfd3 i.xd3 1 5
lLJxd3 e6 1 6 i.f4 1i'd8 1 7 l:el l:e8
The Kingside Fianchetto 159
1 8 i.g2 ttJd7 1 9 h4 h5 20 i.f3 b5 !
and Black's minority attack eventu­
ally triumphed in lvkov-Fischer,
Santa Monica 1966. The bishop pair
is not an advantage in such positions
as White's light-squared bishop is
biting on granite.
7 e4
7 dS was the strange choice in
Rottstadt-McNab, Hastings 1 99 1 .
After 7 . . . lLJc5 8 lLJd4 a5 9 a4 i.d7 10
i.e3 1i'b8 1 4 b3 e5 1 5 ttJde2 llle6 16
fld3 i.c8 1 7 l:ad 1 .:d8 1 8 1i'd2 d5
Black's powerful centre gave him a
decisive advantage.
7 ... eS (D )
8 h3
White's actual move order is not
easy to predict but over the next few
moves he almost invariably plays a
combination of h3, l:e l and a4. It
doesn' t make much difference for
Black, who can j ust get on with his
plan. One alternative which does
change the character of the game is
the insipid 8 dxeS, which has, some­
what surprisingly been favoured by
Smyslov. One examp l e is Smyslov­
S ax, Tilburg 1 979, which continued
160 The Kingside Fianchetto
S ... dxe5 9 b3 (9 We2 c6 1 0 lldl Wc7
1 1 b3 lies 1 2 �a3 Wa5 1 3 �b2 tlJc5
14 llJd2 �e6 15 ltJc4 Wc7 + Barcza­
S ax, Hungary 1 9S3) 9 . . . axb6 1 0 a4
�b7 1 1 llJd2 lies 1 2 �a3 �f8 1 3
�xfS ltJxf8 1 4 ltJc4 ©e6! 1 5 lle l
( 1 5 ltJxe5 ? Wd4 ! is very good for
Black) 15 . . . Wd4 ! ? ( 1 5 . . . ltJd4 looks
more natural) 1 6 ltJd5 <j}g7 and now
in stead of t7 Wf3?! �xd5 l S exd5
e4 1 9 Wd 1 ltJxd5 20 �xe4 lladS 2 1
�xd5 Wxd5 which gave Black a
nagging edge in the endgame, S ax
gives t7 Wxd4 ltJxd4 1 S ltJxc7 ltJxc2
1 9 ltJxeS+ llxeS 20 ltJd6 lle7 2 1
llac l ltJxe l 22 llxe l lld7 23 ltJxb7
llxb7 24 llc l a5 as =.
8 ... c6
The plan I am suggesting for
Black is based on holding his e5
strong-point. With his secure posi­
tion in the centre he will be able to
calmly complete his development:
. . . Wc7, ... b6 and . . .�b7 (or . ..llbS and
. . . b5 if White plays a5 ), . . . .:es and
. . . lladS (optional). Once done, he
will then be able to contemplate ac­
tive operations on the queenside.
9 a4 (D)
In the game Timman-J .Polgar,
Madrid 1 995, White ignored the
'threat' of . . . b5 : 9 llel b5 ! ? (there is
nothing wrong with 9 . . . Wc7 and if
White still refrains from a4 he will
not have the same queen invasion as
in the game) 1 0 dxe5 dxe5 1 1 Wd6
� b7 1 2 �g5 lleS 1 3 :ad l (White's
position looks attractive but it is not
easy to strengthen) 1 3 .. . Wb6 1 4 �e3
Wa5 15 a3 lladS ( 1 5 ... �fS is prob-
ably more accurate) 1 6 ltJd2 �f8 17
ltJb3 Wa6 1 s Wd3 �cs 19 lld2 Wb7
20 lled l ltJb6 ! (fine judgement as
after 2 1 WxdS llxdS 22 llxdS ltJbd7
White will have to make some con­
cessions to extricate his rook, e.g. 23
�g5 <j}g7 24 �xf6 ltJxf6 is quite
good for Black and 23 tlJc5 ! Wc7 24
ltJxd7 ltJxd7 { 24 . . . WxdS 25 ltJxe5
Wes 26 ltJxc6 ! gives a lot of play for
the queen } 25 lies ltJf6 26 lledS
ltJd7 27 lleS ltJf6 is a draw by repeti­
tion unless White risks 2S llxf8+
<j;xf8 29 �c5 + <j}g7 30 �d6 Wa5 !
when 3 1 �xe5 is met by 3 1 . . . b4 ! and
3 1 b4 Wxa3 32 �xe5 Wxb4 33 lld6
�e6 is unclear) 2 1 We2 llxd2 22
ltJxd2 �e6 23 f4 �g7 24 Wf2 Wes
25 <j;h2 exf4 26 gxf4 Wc7 27 �xb6?
(this appears to be the result of a tac­
tical oversight) 27 . . . axb6 2S e5 lDh5
29 ltJde4 ltJxf4 30 ltJd6 �xe5 ! and
White resigned on account of 3 1
ltJxeS ltJxh3+!.
9 ••• Wc7
In practice B lack very often
chooses to block the further advance
of the white a-pawn with 9 a5. •••
However, I don't believe this to be a
good idea; firstly because a4-a5 from
White is not a serious positional
threat and secondly, once Black has

played . . . a7-a5 his dynamic potential
on the queenside is considerably re­
duced.
10 :et
The moves :et and ... :es are oc­
casionally omitted. White can play
1 0 aS, when Malaniuk-Gallagher,
Hamburg 1 995 continued 10 . . . :bS
( 1 0 . .. :es 1 1 :e t is considered later,
whilst on 1 1 �e3 ? ! Black should not
play 1 1 . . .:bs, on account of 1 2 dxe5
dxe5 1 3 �xa7 , but 1 1 . . . exd4 1 12
l!Jxd4 l!Jc5 when 1 3 f3 is answered
by 1 3 . . . d5 ! and t 3 l!Jdb5 looks un­
sound) 1 1 �e3 b5 1 2 axb6 axb6
when White thought for an hour and
played 1 3 dxe5 accompanied by a
draw offer (accepted) . Similar posi­
tions are considered in more detail in
the notes to White's 1 1 th move.
Another example of White omit­
ting :e t (which does lend support to
the e-pawn) was N ogueiras-Shirov,
Moscow OL 1 994. After 10 � e3 b6
1 1 Wd2 �b7 1 2 �h6 :feS 1 3 �xg7
<j;xg7 1 4 dxe5 dxe5 1 5 :fd l :adS
1 6 We3 l!Jc5 1 7 a5 l!Je6 1 S �fl c5
B lack's strong grip in the centre en­
abled him to claim an edge.
10.•. :es (D)
An example of B lack delaying
this move is Martinovsky-Gallagher,
Geneva 1 995 : 10 b6 1 1 �e3 �b7
••.
1 2 Wd2 a6 1 3 dxe5 (my experience
of this variation is that White is usu­
ally panicked into this exchange
once B lack has lined up his queen­
side pawns from a6 to d6; not only
does he then have to contend with
. . . b5 , but he must also watch out for
. . . exd4 followed by . . . c5 ) 1 3 . . . dxe5
1 4 We2 b5 1 5 l!Jd2 :feS 1 6 f3 l!Jh5
The Kingside Fianchetto 161
1 7 Wf2 �fS ( 1 7 . . . f5 ! ?) t s :ad t
l!Jg7 1 9 l!Jb3 b4 20 l!Jbt c5 2 1 lDld2
�c6 22 a5 l!Je6 +.
11 �e3
An important alternative is 1 1 aS
:b8! (certainly not 1 1 . . .exd4 12 l!Jxd4
l!Jc5 1 3 �f4 l!Jh5 ? as 1 4 lDdb5 !
cxb5 1 5 �xd6 is crushing; this trick
has occurred several times) 12 �e3
bS 13 axb6 axb6 14 dS (the only try
for the advantage) 14 �b7 (better
•.•
than 1 4 . . . cxd5 ? ! which needlessly
ceded control over b5 in Rashkov­
sky-Schulz, Berlin 1 99 1 ; after 1 5
exd5 ltJc5 1 6 �fl ! �b7 1 7 b4 ltJce4
1 S l!Jb5 Wes 1 9 c4 White had a clear
advantage; 14 . . . b5 may be worth ex­
amining though) 15 dxc6 �xc6 16
l!Jd2 bS and now we have a couple of
examples:
1 ) The game Romanishin-C.Han­
sen, Groningen 1 99 1 continued 17
:a7 :b7 l S :xb7 Wxb7 1 9 l!Jb3 ? !
( 1 9 b4 d5 20 exd5 l!Jxd5 2 1 l!Jxd5
� xd5 22 �xd5 Wxd5 23 c4 ! bxc4
24 l!Jxc4 i s given as = by Stohl)
t 9 . . . b4 20 ltJd5 (20 l!Ja5 Wc7 2 1
l!Jxc6 bxc3) 20. . .l!Jxd5 2 1 exd5 �a4
22 l!Ja5 Wc7 23 lDc6 e4 ! 24 �f4 f5
25 b3 �xc6 26 dxc6 Wxc6 27 Wxd6
162 The Kingside Fianchetto
'iWxd6 2S i.xd6 i.c3 29 :dt llJe5 30
i.c5? (30 i.xe5 i.xe5 3 1 i.fl :e?
would leave White a little worse in a
defensible position) 30 . . . :as 3 1 :d5
:at + 32 �h2 :c 1 and White found
himself in a lost position as 33 i.xb4
i.xb4 34 :xe5 :xc2 35 �g 1 i.c5 is
hopeless.
2) 17 b4 ltJb6 (Stohl recommends
1 7 ... i.a8 but I'm not sure what Black
has achieved after l S :a3) 18 i. xb6
and now :
2a) Kurajica-Cramling, Debre­
cen 1 992 continued 18 'iWxb6 1 9
.•.
ltJb3 .id? 20 :e3 :as 2 1 :a5 ! :xa5
22 bxa5 'iWc7 23 ltJa2 i.e6 24 :c3
'iWa7 25 ltJb4 with some advantage
for White.
2b) I would prefer 18 :xb6!?••.
which feels like the more harmoni­
ous recapture. For example, 1 9 ltJb 3
.i d? 20 :e3 :cs 2 1 :a5 i.e6 ! ? 22
i.fl (22 :xb5 :c6; 22 ltJxb5 'iWxc2)
22 . . . i.h6 23 :f3 'iWe7 and White will
not be picking up the b-pawn free of
charge.
11•.• b6
12 'iWd2
12 dS i.b7 1 3 lLJ<l2 cxd5 14 exd5
a6 1 5 g4 :acS 16 ltJfl h5 1 7 g5 ltJh7
1 S ltJg 3 lLJc5 1 9 ltJge4 :bs ! intend­
ing . .. i.b7-cS, was quite satisfactory
for Black in Bernard-Stangl, B ad
Wildbad 1990. The position closely
resembles a S icilian Najdorf.
12.•• i. b7 (D)
I believe that Black,s rock-like
positio n gives him at least equal
chances. The problem for Whi te lies
in his knight on f3 . In the g3 Pirc i t
normally stands on the superior e2-
square which provides him with a
couple of aggressive options which
are not available to him here. One
would have been to increase the
pressure on the centre with a quick f4
and another would be to aim for a
kingside attack with g4 and ltJg3,
followed by g5 or f4.
With his knight on f3, Espig can
find nothing better than to exchange
dark-squared bishops, not with the
intention of starting a kingside attack
but simply to give his congested
pieces a little more room.
13 i.h6
13 dS cxd5 14 exd5 a6 gives Black
a good S icilian position, whilst 13
dxeS does little to help White.
13 .•. a6
14 :adl :ad8
15 i.xgl �xgl
16 dxeS dxeS
17 'iWe3 bS
Not 17 lLJcS? l S a5 !, but now
•••
Black is ready to bring his knight to
e6 via f8. White also feels the time is
ripe for a knight manoeuvre.
18 lLJd2 ltJf8
19 ltJb3 liJe6
20 tiJcs
White is worried abo ut Black
playing . . . b4 and . . . c5 , but the text
 The Kingside Fianchetto 163

allows Black's king' s knight to du­ 25 1Wb2 would save a temp o as

plic ate the movements of his queen­ Black would probably still continue
side colleague. Once it arri ves on e6 with ...c5.
its superiority over White's knight 25 ••• cs
on c3 will be evident. 26 axbS axbS
20 ••• li.JxcS 27 bxcS 1WxcS
21 1WxcS li.Jd7 28 'ih>2
22 1Wa3 28 1Wc3 1Wxc3 29 li.Jxc3 li.Jd4 is
also very unpleasant for White.
28 ••• :xdl
29 :xdl :cS
30 c3 1Wc4!
Once White has covered d4 the
black knight needs new employ­
ment. The text clears a route to the
queenside for it whilst the attack on
the e-pawn also forces a weakening
of the white kingside.
31 f3 li.Jcs
32 1Wd2 li.Ja4
33 :bl iLa6
34 JLn 1WcS+!
Again the queen vacates a square
for the black knight. Taking the pawn
would have been premature, e.g.
34 ltJxc3 35 li.Jxc3 1Wxc3 36 1Wxc3
. ••

23 b4!? :xc3 37 iLxb5 iLc8 (37 . .. iLxb5 38

A slightly controversial move;
White believes it will be in his inter­
est to exchange as many pawns as
possible on the queenside as wel l as
h oping that his queen can become
active on the long diagonal.
23 ••• li.Je6
24 li.Je2 <it;gS
A useful prophylactic measure
which was mainly inspired by my
opponent's time pressure; it's much
harder to play quickly against such
moves than against ones which cre­
ate concrete threats.
25 Wal?!
:xb5 :xf3 39 <if;g2 =) 38 ii.fl ! when
38 . .. :xf3 39 :b8 :xg3+ 40 <if;f2
11c3 4 1 iLa6 is likely to end in a
draw.
35 <it; g2 li.Jb6
36 li.Jgl ltJc4
37 iLxc4 1Wxc4 (D)
At first I was a little reluctant to
exchange off my heroic knight but
one has to look at what is left on the
board rather than what is going off it.
Black's advantage consists of his
safer king, the weak white pawn on
c3 and, most importantly, the dread­
ful white knight. It can't even go to
164 The Kingside Fianchetto

50 %:.bl "ifxc3!
5 1 :lxb5 "ife3
0-1

Game 26
Rivas Khalifman
-

Dos Hermanas 1993

1 d4 ltJf6
2 ltJf3 g6
3 g3 �g7
e2 because of . . . b4. The only factor 4 �g2 0-0
in White's favour is the reduced 5 0-0 d6
amount of ma terial which allows 6 %:.el
him to hope for a draw. The remaining 6th move alterna­
38 %:.al :lc6 tives are:
39 h4 h5 1 ) 6 ltJbd2, intending e4, when
40 �f2? ! "ifc5+! Black has a couple of ways to pre­
B lack takes the opportunity to pare . . . e5 .
transfer his rook to the d-file. l a) 6... ltJbd7 7 e4 eS (D) and
41 �g2 :ld6 now:
42 'ifcl �g7
43 ltJh3
Or 43 ltJe2 'ifc4 ! when the knight
is forced back to g l . Now that the
white pieces have been coaxed onto
inferior squares Black is ready to
start another assault on the enemy c­
pawn.
43 ..• :lc6!
Threatening . . . b4 .
44 "ifel �c8
45 %:.bl �d7!
46 :lb3 "ifc4 l al ) 8 dxeS dxe5 (S . . . ltJxe5 will
47 "ifdl :la6! transpose to ' 1 b2 l ' ) 9 b3 :es 1 0
48 ltJll :la2 �b 2 b6 l l ltJc4 �b7 1 2 ltJfd2 ( 1 2
By some accurate manoeuvring ltJfxe5 is also equal) 1 2. . ."ife7 1 3 a4
Black has succeeded in infiltrating to a6 1 4 %:.e l :ladS 1 5 "ife2 h5 with a
the seventh rank. His attack is now comfortable game for Black, Todor­
decisive, one nice point being that 49 cevic-Lautier, Palma l 9S9.
"ifd5 is refuted by 49 . . . � h3+ ! . l a2) 8 c3 b6 9 %:.e l :es 1 0 "ifc2
49 �gl �e6 �b7 (the position is the same as line

' l ' in the note to White's 8th move
except for the fact that a4 and a5
were flicked in there) 1 1 dxe5 lt:Jxe5
1 2 lt:Jxe5 1:lxe5 ! ? 1 3 ttJc4 ?! .ixe4 1 4
.ixe4 1:lxe4 1 5 1:txe4 lt:Jxe4 16 'ifxe4
d5 1 7 'iff3 ( 17 'ife2 dxc4 1 8 'ifxc4
'ifd 1 + 1 9 'iffl 'ifc2 favours Black)
1 7 . . . dxc4 1 8 .ig5 'ife8 1 9 .:td l c6 +
Frendzas-Moutousis, Peresteri 1994.
1 b) 6 lt:Jc6 and now:
.•.
1 bl ) 7 lt:Jc4 (a suggestion by An­
dersson) 7 . . . .ie6 ! ? 8 lt:Je3 (8 b3?!
'ifc8 9 .:te l .ixc4 10 bxc4 lt:Ja5 1 1
'ifd3 c5 is given as + by Shabalov)
8 . . . .id7 9 c4 (9 d5 lt:Jb4 10 c4 c6 is
another game) 9 . . . e5 10 d5 lt:Je7 1 1
c5 lt:Je8 1 2 lt:Jc4 f5 with an unclear
position, Shabalov-W. Watso n , Bel­
grade 1988.
1 b2) 7 e4 eS (7 . . . lt:Jd7 !? 8 c3 e5 is
an idea of Nigel Davies) and now:
1 b2 l) 8 dxeS lt:Jxe5 (since the
knight on c6 will not be well placed
after a subsequent c3 by White it
makes sense to exchange it) 9 lt:Jxe5
dxe5 1 0 'ife2 'ife7 1 1 1:ldl b6 1 2 b3
a5 1 3 a4 .ia6 1 4 'ife l 1:tfd8 1 5 .ia3
'ife6 1 6 lt:Jfl .if8 = Filip-Fischer,
Stockholm 1 962.
1 b22) White has been reluctant
to play 8 c3 here, perhaps because of
8 . . . exd4 ! ? (Davies's 8 . . .lt:Jd7 and
8 . . . .ig4, hoping to get White to
block the centre, are reasonable al­
ternatives) 9 cxd4 .ig4. A possible
continuation is 1 0 d5 lt:Je5 1 1 h3
lt:Jxf3+ 12 lt:Jxf3 .id7 1 3 .:te l 1:le8 1 4
e 5 ( l 4 lt:Jd4? allows 1 4 . . . lt:Jxd5 ! and
14 'ifc2 c6 is fine for Black) 1 4 ...dxe5
1 5 lt:Jxe5 .if5 when B lack has a
comfortable game as 1 6 g4 can be
met by 1 6 ... .ie4 ! .
The Kingside Fianchetto 165
lb3) 7 c3 eS and now (D):
l b3 1 ) 8 'ifc2?! (8 e4 is variation
' l b22' ) 8 . . . exd4 ! 9 lt:Jxd4 (9 cxd4 d5
is at least equal for Black) 9 . . . lt:Je5 ! ?
(attempting to mix i t u p in a crucial
last-round game, but 9 . .. lt:Jxd4 1 0
cxd4 d5 is the sensible way to play)
10 e4 c5 l l lt:Je2 'ife7 1 2 h3 :es 1 3
c4 lt:Jc6 1 4 lt:Jc3 lt:Jd4 1 5 'ifd3 .ie6
with good play for Black, Kurz-Gal­
lagher, Baden 1 996.
1 b32) 8 dxeS lt:Jxe5 9 lt:Jxe5 dxe5
1 0 e4 'ife7 1 1 a4 1:ld8 1 2 'ifc2 b6 1 3
.:te l .ia6 1 4 .ifl .ib7 1 5 a5 .ih6 !
with a slight initiative for Black,
Giertz-Gallagher, Vill ars 1 995 . The
rest of the game is worth a glance as
it demonstrates that White cannot
draw ' to order' by playing an early
dxe5 : 1 6 f3 1:ld7 1 7 .ih3 'ifc5 + ! 1 8
�g2 1:tdd8 1 9 lt:Jb3 'iff8 (B lack's
play may look suspect but the idea
was to lure White's pieces onto infe­
rior squares) 20 �g l .ixc l 2 1 'ifxc l
.ia6 22 .ifl .ixf 1 23 1:txfl 'ife8 !
(the queen will be very active on b5 )
24 .:td l 'ifb5 25 1:txd8+ 1:lxd8 26
'ifc2 lt:Je8 ! (the knight heads for c4)
27 axb6 axb6 28 .:td 1 1:txd 1 + 29
'ifxd l lt:Jd6 (White had assumed the
166 The Kingside Fianchetto
draw was very close but this was the
position I had been aiming for as
Black is simply winning) 30 �f2
ltJc4 3 1 Wc2 (the only move to save
the b-pawn . . . ) 3 l ... Wa4 ! ( ... but White
is now caught in a terminal pin) 32
�e2 ltJa5 ! 33 ltJd4 Wc4+ 34 1i'd3
exd4 and Black soon won.
2) 6 a4! ? (D)
An interesting idea which has be­
come quite popular recently. Before
deciding on his piece placement
White wishes to discover the shape
of the queenside . I am proposing
that B lack simply ignores White's
offensive. Firstly, because the Pirc
type positions that we are aiming for
are less promising once Black has
played . . . a5, and secondly, I don't
believe that a5 is such a strong posi­
tional threat that we sh ould save
White from wasting a further tempo
on playing it. 6 ltJbd7 7 as (consis­
•••
tent; instead 7 ltJc3 c5 S e4 a6 9 a5
Wc7 1 0 :Z.e l cxd4 1 1 ltJxd4 �5 1 2
h3 ltJc4 1 3 ltJb3 e5 ! ? 1 4 g4 .ie6 led
to an unclear Sicilian position in the
game Polugaevsky-J . Polgar, Hast­
ings 1 992/3 whilst 7 . . . e5 S e4 would
be similar to the previous game)
7.. c6 (7 ... e5 is worth examinin g as S
.
dxe5 dxe5 9 a6 e4 1 0 ltJg5 We7 1 1
ltJc3 ltJc5 looks OK for Black, whilst
7 . . . :Z.bS S ltJc3 c5 9 e4 cxd4 10 ltJxd4
a6 1 1 :Z.el ltJe5 1 2 b3 tt:Jc6 1 3 ltJxc6
bxc6 1 4 .ib2 c5 was roughly equal
in Hergott-Bologan, Biel 1 993) 8
ltJbd.2 (S ltJc3 is more common and
should also transpose to Game 25 )
8 eS 9 e4 and now:
...
2a) 9 Wc7 10 c3 :es 1 1 :Z.e l .:tbs
•••
1 2 ltJc4 b5 ( l 2 . . . exd4 1 3 Wxd4 !) 1 3
axb6 ltJxb6 1 4 ltJa5 c5 1 5 d5 .id7 1 6
.ifl was slightly better for White in
Khuzman-Gallagher, Antwerp 1 993.
2b) Perhaps Black could have ex­
changed in the centre before White
had the chance to support it with c3.
After 9 exd4!? 10 ltJxd4 ltJc5 the
•••
pressure on e4 will make it difficult
for White to complete his develop­
ment without playing b4 (f3 would
be even more suspect as it would
make the . . . d5 break more appeal­
ing). After, for example 1 1 :Z.e 1 :Z.eS
1 2 b4 ltJe6 1 3 ltJxe6 .ixe6 14 .ib2
d5 ! ? (by no means the only way to
handle the position) 15 e5 ltJd7
B lack has a comfortable game. His
plan will be to break up the white
centre with .. .f6, whilst 16 a6 is not a
serious worry as after 1 6 .. b6 1 7
.
ltJb3 c5 ! 1 S .ixd5 .ixd5 1 9 Wxd5
there is the resource 1 9 . . . ltJxe5 ! .
6 ••.
ltJbd7
Again Black is heading for the fa­
vourable version of the g3 Pirc that
we saw in the previous game.
6 ...ltJc6 is played quite frequently
but I believe that 7 d5 ltJb4 S e4 gives
White good chances of an edge . The
most interesting alternative to the

text is 6 cS which, depending on
•.•
White 's reply, can lead to p ositions
resemblin g the Sicilian, the B enoni,
the English or the King's Indian !
Let's take a brief look:
1) 7 c4 li.Jc6 S li.Jc3 (S d5) S . . . cxd4
9 li.Jxd4 li.Jxd4 1 0 'ifxd4 .ie6 1 1 'ifd3
:cs 1 2 li.Jd5 li.Jxd5 1 3 cxd5 .id7
was level in Andersson-Christian­
sen, Moscow IZ 1 9S 3 . In effect,
White's Ae l left him a tempo down
on one of the main lines of the Sym­
metrical English.
2) 7 dxcS dxc5 S 'if xdS :Z.xdS 9
c3 (9 li.Je5 li.Ja6 1 0 c3 li.Jd5 1 1 li.Jc4
.ie6 1 2 li.Jba3 Ad7 1 3 Abl AadS
was also fine for B lack in Anders­
son-Quinteros, Mar del Plata 19S 1 )
9 . . . li.Jd5 ! ? 1 0 Ad 1 li.Jc6 1 1 li.Jg5 e6 1 2
li.Je4 b6 1 3 .ig5 Af8 14 li.Ja3 h6 1 5
li.Jf6+ .ixf6 1 6 .ixf6 .ia6 1 7 c4
li.Jxf6 1 S .ixc6 :Z.acS 1 9 .ig2 AfdS
20 e3 <iii> f8 2 1 h3 <iti>e7 22 f4 :Z.xd 1 +
112- 1'2 Nogueiras-Ivanchuk, Moscow
1 990.
3) 7 dS b5 ! ? S e4 .ib7 9 c4 bxc4
1 0 li.Jfd2 li.Jfd7 1 1 li.Jxc4 li.Je5 1 2
li.Jba3 li.Jbd7 1 3 �3 .ia6 14 .ifl 'ifa5
1 5 <iti>g2 ( 1 5 .id2 'ifxd2) 1 5 . . . AabS
and Black had B enko Gambit style
play without being a pawn down in
Panno-J .Polgar, Aruba 1 992.
4) 7 e4 cxd4 S li.Jxd4 leads to a
pretty tame variation of the Sicilian
Dragon . One example is Hoffman­
Zapata, Seville 1 992 which contin­
ued S . . . .ig4 (presumably S ... li.Jc6 is
met by 9 li.Jxc6 bxc6 1 0 e5) 9 f3 .id7
10 a4 llJc6 1 1 li.Jb3 li.Je5 1 2 llJc3 AcS
1 3 <iii>h l 'ifc7 14 li.Jd4 a6 with a dou­
ble-edged position.
7 e4 es
The Kingside Fianchetto 167
8 a4
8 li.Jc3 is Game 25 , which leaves 8
c3 as the most important alternative.
After 8 :Z.e8 9 li.Jbd2 b6 10 a4 aS
•..
( now it would be uncomfortable to
allow a5) there is (D):
1 ) 1 1 'ifc2 .ib7 1 2 d5 (the pres­
sure on the e-pawn makes decent al­
ternatives hard to find) 1 2 . . . c6 ! 1 3
dxc6 .ixc6 (Black's pawns may look
weak bu t the point is that White
won ' t be able to prevent . . . d5 ) 14
li.Jc4 'ifc7 1 5 li.Jfd2 li.Jc5 1 6 b3 d5 17
exd5 .ixd5 with an edge for B lack,
Ztiger-Gallagher, B ad Ragaz 1 994.
2) 1 1 li.Jc4 occurred in the game
Ye Rongguang-Domingues, Cuba
1 992 where Black refrained from the
critical 1 1 exd4:
...
2a) Perhaps this was on account
of 12 li.Jxd4 .ib7 13 e5? ! ( 1 3 .ig5 is
better with a level game), but then
1 3 . . . .ixg2 1 4 exf6 :Z.xe l + 15 'ifxe l
.id5 ! 1 6 fxg7 (or 1 6 li.Je3 li.Jxf6)
1 6 . . . .ixc4 17 .ih6 .id5 ( 1 S li.Jc6
was threatened) should be good for
Black.
2b) The alternative recapture 12
cxd4 is also not particularly promis­
ing for White. After 1 2 . . . :Z.xe4 1 3
168 The Kingside Fianchetto

.:.xe4 l!Jxe4 14 l!Jg5 the solid 14 dS

••. The alternative way to deal with
1 5 l!Jxe4 dxe4 16 i.xe4 .:.bs 1 7 i.f4 the threat of ...l!Jg4, 13 h3, would also
l!Jf6 looks about level, whilst the have been met by . . .'fle7 and . . . l!Jc5
sharp 14 lDxgS 1 5 i.xa8 i.a6 1 6
••. when White would probably end u p
i.d5 l!Jh3+ 1 7 <iiig 2 l!Jf6 is certainly having to play f3 anyway.
worth looking into. 13 •.. 'fle7
8 ••. .:es 14 'fle2 tlJcs
9 dxeS?! 15 aS llJe6
This exchange, here and in similar 16 a6?!
positions, rarely brings White any White should have just played 1 6
benefit and normally just relieves c3 instead of driving the bishop onto
any pressure Black may have been a more active diagonal and awarding
experiencing. It is, however, not un­ himself a weak a-pawn.
usual for White to play dxe5 as the 16 i.c8
players who opt for the systems in 17 c3 h4 !
this chapter are often trying to play 18 l!Jc2 lLJhS
without the slightest risk. 9 tlJc3 c6 19 lDf1
would again transpose to Espig-Gal­ 19 'flf2 c5 ! also leaves Black with
lagher. a considerable space advantage.
9 •.. dxeS 19 cS
10 l!Ja3? ! 20 .:.edl (D)
And this is not a good follow-up
as Black will now be able to take
over the initiative by attacking e4 be­
fore White can get at e5 . Better was
10 l0c3 c6 with equality.
10 b6 !
1 1 i.e3 i.b7
12 l!Jd2 hS! (D)

20 ... ltJef4!
Excellent judgement from Black
who gives up a piece for purely posi­
tional reasons.
2 1 gxf4
2 1 i.xf4 exf4 22 g4 l!Jf6 is pretty
horrible for White and 2 1 'fld2 i.e6 !
22 gxf4 .:.ad8 is no better than the
13 f3 game.
 The Kingside Fianchetto 169

21 ••• exf4
22 �cl
22 �f2 h3 23 �h l Wg5+.
22 •.• h3
23 i.. b l
If that thing on h l tried to describe
itself as a bishop it would find itself
in contravention of the Trade De­
scriptions Act. Of course the per­
manent incarceration of this piece
is what B l ack's combination is all
about. 2S ••• .:xdl!
23 • .• �e6! 29 �xdl �b3
There is no need to rush, for ex­ 30 .:aJ � xc2
ample 23 . . . 'it'g5 + 24 ltlg3 ltlxg3 25 31 c4 �rs
hxg3 Wxg3+ 26 �fl would allow 32 �cl?!
White an easier ride. 32 b4 could have bee n played im­
24 ltld2 .:adS mediately against which Khali fman
2s �n .:d7 planned 32 ... .:es .
26 :et :eds 32 .•. :es
27 es 33 b4 cxb4
White tries to free his shackles, 34 .:b3 WcS
but passing may have been a health­ 3S �d2 .:xeS
ier option. 36 �xb4 Wc7
21 •.. .:ds 37 'it'dl Wxc4+
2S .:a4 (D) 3S �gl .:ds
White plans to meet 28 . . . .:xe5 0-1
with 29 .:e4, but now Black has a lit­ A fine performance from Black
tle combination to recuperate some who punished White for his passive
of his material. play.
1 3 The Veresov
The Veresov has never been very
popular at grandmaster level, except
for a brief period in the early 1 9 80s
when it was employed with some
success by Tony Miles. At club level,
though, it has always had its support­
ers as, like the Trompowsky and sev­
eral other systems in this book, it
avoids all the main lines and forces
Black to do battle on what is likely to
be unfamiliar territory.
As a young player I remember be­
ing taught never to block the c-pawn
in Queen's Pawn openings (later I
learnt about not blocking the f-pawn
in King's Pawn openings but that's
another story) as it then becomes
very difficult to create active play. As
a consequence of this lack of space
on the queenside White nearly al­
ways plays for e4 in the Veresov.
There are two main ways of doing
so: Grune 27 concentrates on the
risky 4 f3 (after the recommended
3 . . . liJbd7), signalling White's inten­
tion to play e4 as soon as possible.
This way of building up a strong cen­
tre would be the ideal plan for White,
except for some cunning tactics at
Black's disposal.
Game 28 deals with the solid
liJf3, where White intends to de­
velop his kingside before turning his
attention to opening the centre. The
minor 4th move options are exam­
ined in Game 27.
Game 27
Alburt - Tal
USSR Ch (Baku) 1972
1 d4 liJf6
2 liJcJ d5
3 .i.gS liJbd7
Black has a large number of play­
able alternatives (3 . . . c6 and 3 . . . .i.f5
to name a couple) but I am going to
concentrate on the solid 3 . . . liJbd7,
partly because of its solidity (Ver­
esov players tend to be a tricky, care­
free bunch) and partly because I
know it better than any of the other
lines.
4 f3
White dreams of constructing a
proud centre but ignoring one's de­
velopment is a risky business. The
sensible 4 liJf3 is the subject of the
next game whilst rarer alternatives
are examined below:
1 ) 4 e4 (obviously very similar to
the Blackmar-Diemer) 4 . .. liJxe4 5
liJxe4 dxe4 6 .i.c4 h6 7 .i.h4 liJf6 8
f3 'ii'd6 9 c3 .i.e6 (it's hard to believe
that White has anything for his
pawn) 1 0 'ii'a4+ c6 1 1 .i.xe6 'ii'xe6
1 2 0-0-0 liJd5 1 3 Ae 1 e3 1 4 .i.g 3 b5
1 5 'ii'c2 'ii'd7 1 6 liJe2 e6 17 liJf4 .i.e7
1 8 liJd3 ( 1 8 liJxd5 cxd.5 would win
back the pawn but leave White posi­
tionally lost as Black's minority at­
tack is going to become a full-scale
assault on the white king) 1 8 ... 0-0 1 9

i.f4 c5 20 dxc5 ltJxf4 2 1 ltJxf4
i.xc5 with a decisive advantage for
Black, G.Mohr-Lobron, Ljubljana
Vidmar mem 1 989.
2) 4 'ifd3 (D) and now:
2a) 4 h6 5 i.f4 c6 6 ltJf3 e6 7 a3
.•.
b5 8 ltJe5 ltJxe5 9 i.xe5 b4 1 0 axb4
i.xb4 1 1 'ifg3 ltJe4 1 1 2 'ifxg7 l:lf8
1 3 i.f4 'iV b6 1 4 i.d2 i.xc3 1 5 bxc3
'if b2 16 l:lc 1 a5 ! left White unable to
stop the bionic a-pawn in Porper­
Smirin, Tel Aviv 1 99 1 . An early . . . c6
seems a good idea as if White castles
long (clearly one of the main objec­
tives behind 4 'ifd3 ) as Black will
have an automatic attack by means
of . . . b5 .
2b) 4 c5! ? is also a pretty natu­
..•
ral reaction. I have had a couple of
pleasant experiences with this line:
2 b l ) P.Moore-Gallagher, Jersey
1984 continued 5 ltJf3 cxd4 6 'ifxd4
e5 !? 7 lhxe5 i.c5 8 'ifa4 'ifb6 9 0-0-0
d4 10 ltJc4 'ife6 l l lhb5 0-0 12 liJc7
'iff5 1 3 i.xf6 b5 ! (I don't remember
anything about this game except for
feeling pleased about this move) 14
ltJxb5 ( 1 4 'ifxb5 l:lb8 also saves the
rook) 1 4 . . . 'ifxf6 (Black has excel­
lent attacking chances in return for
The Veresov 1 71
his material investment) 15 e3 dxe3
1 6 fxe3 l:.b8 1 7 i.d3 ltJe5 1 8 l:.hfl
ltJxd3+ 1 9 cxd3 'ifc6 ! 20 liJc3 'ifxg2
2 1 ltJe4? i.xe3+ 22 �bl i.d4 23 b3
i.e6 24 ltJcd2 l:lfd8 25 l:lc 1 f5 0- 1 .
2b2) 5 e4 cxd4 6 'ifxd4 (Bellin
suggests 6 i.xf6 ltJxf6 7 ltJxd5 ltJxd5
8 'iV b5+, but surely the ending aris­
ing after 8 . . . i.d7 9 'ifxd5 i.c6 1 0
'ifxd8+ l:lxd8 i s good for Black?)
6 ... e5 7 'ifa4 d4 8 rtJd5 (8 i.xf6 lti'xf6
9 ltJd5 'ifd8 10 f4 would transpose to
'3a' in the note to Black's 4th move)
8 . . . i.e7 9 i.xf6 i.xf6 10 i.b5 0-0 1 1
i.xd7 i.xd7 1 2 ltJxf6+ gxf6 1 3 'ifa3
l:lc8 14 0-0-0 l:lc6 (the black king,
the one with the gaping holes around
it, is perfectly safe as White has
nothing to menace it with, whilst the
white king, the one with the plentiful
pawn cover, is about to be subjected
to a heavily co-ordinated attack from
the black artillery) 1 5 ltJe2 'ifc7 16
�bl l:lc8 1 7 'ifg3+ �f8 18 c3 i.e6
1 9 a3 l:l b 6 and the game dido' t last
much longer, Richmond-Gallagher,
Nottingham 1 987.
3 ) 4 e3 normally transp oses to
the next game after a subsequent
ltJf3, but there is one independent line
involving an early f4 by White, e.g.
4 g6 5 i.d3 i. g7 6 'ifd2 0-0 7 f4 ! ?
..•
c5! 8 ltJf3 b6 (Black plans . . . i. b7
and . . . ltJe4) 9 ltJe5 i. b7 10 l:ldl ( 10
0-0 ltJe4 ! 1 1 i.xe4 dxe4 12 ltJxd7
'ifxd7 1 3 ltJe2 i.a6 ! wins a pawn as
14 c3 loses to 14 ... h6 1 5 .li.h4 'ifg4 ! )
and now:
3a) Ermenkov-Grivas, Sofia 1 986
continued 10 llJeS 1 1 0-0 ltJd6 1 2
.••
i.h4 l:.c8 1 3 ltJb5 ! ltJxb5 14 i.xb5
ltJxe5 15 fxe5 with rough equality.
172 The Veresov
3 b) It may be possi ble for Black
to try 10 cxd4 1 1 exd4 tlJe4 12
.••
tlJxe4 dxe4 1 3 .i.c4 tlJxe5 and, after
either recapture, 14 . . . 'ii'c7.
Now we return to the main line af­
ter 4 f3 (D).
4 •.. c6
Again Black has the choice be­
tween . . . c6 and . . . c5. The former so­
lidifies d5, provides the option of
playing . . . b5 and opens a path for the
queen to b6 or a5 from where it can
harass the white queenside. The lat­
ter, which is considered belo w, is a
more direct attempt to punish White
for taJcing liberties such as 4 f3 . After
4 c5 White has three options:
••.
1 ) 5 e3 (rather inconsistent) 5 . . .e6
6 a3 (White fears . . .cxd4 followed by
. . . .i.b4) 6 . . . a6 7 'ii'd2 b5 8 tlJh3 'ii'a5
9 tlJe4 'ii'c7 1 0 .i.f4 'ii'c6 1 1 tlJxf6+
tlJxf6 with an active game for Black,
Sibilio-Gallagher, Chiasso 1990.
2) 5 dxc5 'ii'a5 (5 . .. e6 is not men­
tioned by theory but it deserves se­
rious consideration, e . g. 6 e4 .i.xc5 7
exd5 'ii'b 6 8 tlJa4 'ii'a5 + 9 c3 .i.xg 1
1 0 .:t.xg l and now 1 0 . . . b5 is a very
risky way to win a piece, but the al­
ternative 1 0 . . . tlJxd5 is much safer;
perhaps the critical line is 6 b4 b6 7
e4) 6 .i.xt6 ttJxf6 7 'ii'd4 (7 e4 should
be met by 7 . . . e6) 7 e5 ! ? (a similar
••.
sacrifice to the one seen above in
Moore-Gallagher) 8 'ii'x e5+ .i.e6 9
e4 .i.xc5 (D) and now:
2a) White was destroyed in the
game Wockenfuss-Timman, Bad Lau­
terberg 1977 after 1 0 .i.b5 +? ! (an
unfortunate square for the bishop)
10 . . . �f8 1 1 0-0-0 .i.e3+ 12 �b l d4
1 3 'ii'd 6+ �g8 14 b4 'ii'a 3 1 5 tlJd5
tlJxd5 16 exd5 .i.f5 17 tlJe2 a5 1 8
tlJxd4 ax b4 19 .i.c4 .i.xd4 ! 20 .:t.xd4
.i.xc2+ ! 2 1 �xc2 b3+ 0-- 1 .
2b) 10 0-0-0 0-0 ( 10 . . . 0-0-0 !?) 1 1
exd5 .i.xd5 and now White has a
choice of captures on d5 :
2bl ) 12 .:t.xd5 tlJxd5 1 3 'ii'x d5
.:t.ad8 14 'ii'b 3 when several sources
give the fascinating line 1 4 .i.eJ+ •.•
1 5 �b l 'ii'x c3 ! ?, which I believe
leads to a draw, but no-one mentions
the simple 14 .i.xgl! 15 :Xg l 'ii'g5+
.. .
16 �b 1 'ii'e 3 when White can resign.
2b2) 12 tlJxd5 tlJxd5 13 'ii'xd5 ?
.i.e3+ 14 �b l .:t.ad8 ! wins for Black,
but 13 .:t.xd5 'ii'xa2 is the critical line.
A possible continuation: 14 .:t.xc5
'ii'a l + ( 14 . . . .:t.fe8 1 5 .:t.a5 !) 1 5 �d2

l:.fe8 ! 1 6 'iVg3 'iVxfl with an ongoing
attack for Black.
3) 5 e4 cxd4 leads to a further
branch :
3a) 6 'iVxd4 e5 7 'iVa4 d4 8 i.xf6
(8 llJd5 �e7 is favourable for Black)
8 . . . 'iVxf6 (8 . . . gxf6) 9 ltJd5 'iVdS 1 0 f4
�c5 I 1 4Jf3 0-0 1 2 0-0-0 a6 1 3 fxe5
4Jxe5 ! 1 4 4Jxd4 �g4 1 5 l:.d2 'iVh4
with a very dangerous initiative for
Black, Shteinberg-Anka, Balaton­
bereny 1 993.
3b) 6 �xf6 (D) when:
3bl ) 6 dxc3 7 �xc3 dxe4 8 fxe4
...
is supposed to be better for White.
The game D.McDonald-Gallagher,
Hastings 1 99 1 , though, was quickly
decided in Black's favour: 8 . . . e6 9
4Jf3 f6 ! 1 0 �c4 ( 1 0 'iVd2) I O . . 'iVb6
.
1 1 ltJd4 ltJe5 1 2 � b5+ �f7 1 3 'iVe2
a6 1 4 �a4 �b4 ! 1 5 0-0 �c5 1 6
l:.ad l l:.d8 1 7 'iVd2 lbc4 1 8 'iVd3 llJe5
1 9 'iVd2 ltJc4 20 'iVd3 l:.xd4 ! 2 1
� xd4 ltJxb2 ! 22 �e8+ (22 �xc5
'iVxc5+ 23 'iVd4 4Jxa4 ! ) 22 . . . <&t>xe8
2 3 .i xc5 'iVxc5+ 24 'iVd4 ltJa4 ! with
a decisive material advantage.
3b2) 6 4Jxf6 is more respect­
•••
able. After 7 'iVxd4 dxe4 8 �b5+ (8
'iVxd8+ �xd8 9 0-0-o+ �c7 10 �c4
The Veresov 1 73
e6 1 1 ltJxe4 ltJxe4 1 2 fxe4 is usually
assessed as = but I would have
thought that Black's bishops count
for something) 8 . . .�d7 9 0-0-0 � xb5
I 0 ltJxb5 'iV xd4 1 1 l:.xd4 e5 and now
12 ltJc7 + �e7 1 3 llkl5+ is equ al ac­
cording to Alburt whilst 12 l:.c4 l:.d8
is more difficult to judge.
5 e4
5 'iVd2, preparing to castle queen­
side, can be met by 5 'iVa5 when the
•.•
game is liable to transpose into the
note to White's 6th move after 6 e4,
or perhaps by 5 h6 6 �h4 e6 when
•••
7 e4? fails to 7 ... ltJxe4 ! .
5 .. .
dxe4
6 fxe4 (D)
In the game Mestel-Webb, Bir­
mingham 1 97 5, White tried 6 'iVd2,
but after 6 . . .'iVaS 7 �xf6 (7 fxe4 e5 8
dxe5 ltJxe5 9 0-0-0 �e6 is also fa­
vourable for Black, but 7 ltJxe4
would probably maintain the equi­
librium) 7 . . . ltJxf6 8 fxe4 e5 9 dxe5
ltJg4 1 0 �e2 ltJxe5 I 1 ltJf3 �e7 1 2
0-0 ltJxf3+ 1 3 l:.xf3 �e6 Black had a
clear advantage .
6 ••• e5 !
O f course Black dido' t concede
the centre on his previous move
1 74 The Veresov
without having something concrete
in mind. 6 'iVb6! ? is also possible
•..
but after 7 iVd2 Black should play
7 . . . e5 rather than grabbing the b-
pawn.
7 dxeS
7 d5 .i.c5 and 7 �3 exd4 8 4Jxd4
.i.b4 9 4Jf5 0-0 I O .i.d3 lLJe5 1 1
.i.xf6 iV xf6 1 2 0-0 .i.xf5 1 3 l:txf5
iVe7, S chiller-Ligterink, Reykj avik
1 986, are both good for Black.
7
..• iVaS
8 exf6? !
Although the text ensures that
White will remain a pawn up for the
time being, Black is allowed to
dominate the dark squares. There are
a couple of alternatives:
I ) 8 .i.xf6 gxf6 (D) and now:
l a) 9 exf6 4Jxf6 10 'iVd4 .i.g7 1 1
0-0-0 0-0 1 2 iVa4 ? iVxa4 1 3 4Jxa4
4Jxe4 was the ridiculous continu­
ation of Philippe-Kennefick, Haifa
OL 1 976. In view of the threat of
. . . 4Jd5 White probably has to play
something like 1 2 llJge2, which can
be met by 1 2 . . . .i.e6 intending to
play a rook to d8, or 12 iVd2 when
12 .i.e6 is sensible but the insane
•••
12 4J xe4 ! ? should not be ruled out
••.
of contention as after I 3 llJxe4 iVxa2
14 iVf4 iVal + 1 5 �d2 iVxb2 White's
king is very exposed and Black's a­
pawn could easily cost White a con­
siderable amount of material.
I b) 9 e6 fxe6 I 0 .i.c4 (perhaps
White had intended I 0 iVg4 but after
I 0 . . . 4Je5 1 1 iVh5+ �e7 he has to
lose further time with his queen in
view of the threatened 1 2 . . . 4Jd3+)
I O . . . .i.b4 I I 4Je2 4Je5 1 2 .i.b 3 l:tg8
( 1 2 . . . .i.d7, intending . . . 0-0-0, is a
suggested improvement by Gufeld)
1 3 a3 ! .i.xc3+ ( 1 3 . . . .i.c5) 1 4 llJxc3
l:txg2 1 5 iVh5+ l:tg6 1 6 iVh3 ! (not
1 6 iVxh7 4Jf3+ 1 7 �f2 iVg5 ! 1 8
�xf3 l:th6 ! ) 1 6 . . . llJg4 1 7 0-0-0 ! 4Jf2
1 8 iVxh7 iVg5+ 1 9 �bl l:tg7 20
iVh8+ :lg8 21 iVh7 l:tg7 (Black
deemed 2 1 . . . llJxd I 22 l:txd I iVg7 23
iVh5+ 'iVg6 24 iVc5 to be too risky)
22 iVh8+ :lg8 1h- 1h Rossetto-Gu­
feld, Camaguey 1974.
2) 8 llJf3 llJxe4 9 .i.d2 4Jxd2 1 0
iVxd2 .i.b4 1 1 0-0-0 0-0 12 a 3 .i.xc3
1 3 iVxc3 'iVxc3 14 bxc3 is consid­
ered to be an equal ending by ECO.
White has some activity to compen­
sate for his wrecked paw n structure
but I feel that the long-term chances
must be with Black.
8 ... 'iVxg5
9 fxg7 .i.xg7
10 iVd2
10 4Jf3 iVe3 + 1 1 .i.e2 .i.xc 3+ 1 2
bxc3 iVxc3+ 1 3 4Jd2 4Je5 1 4 0-0
.i.e6 15 4Jf3 l:td8 1 6 iVe I 4Jxf3+ 1 7
.i.xf3 'iVxe l 1 8 :lfxe l l:td2 1 9 :le2
l:txe2 20 .i.xe2 �e7 2 1 �f2 :lg8 22
h4 �d6 was a nearly hopeless end­
ing for White, Elina-Chiburdanidze
USSR 1976.
 The Veresov 1 75

10 ••• 'ifxd2+ 19 llJxe5 was equally good.

..•

In an ideal world Black would 20 �f2 �f6

probably choose to keep the queens 2 1 :e3 llJc5
on but he is not willing to waste time 22 l::tael �d7! (D)
avoiding an exchange as his initia­ Tai, not surprisingly, avoided the
tive and control of the dark squares trap 22 f4? 23 l::txe6! llJxe6 24 �f5
••.

will persist into the ending. � d7 25 llJe4 ! after which White

1 1 �xd2 llJc5 wins material.
12 �d3 � e6
13 llJf3 0-0-0
14 �e2 b5
Black could have played 14 J:theS
•.

at once but he prefers to seize some

additional space on the queenside.
15 a3 a5 (DJ

·-
�-·-
- -
�� ,
w • • •,.,
,,

B i B.i.B B
• .& .
. . �
�� � �
��
f%{
% 23 llJxb5?
• -�· •
,, •

White 's position was poor but

�
�i � ��
� ... -� ­
t.t._) $�
there was no need for immediate ca­
pitulation.
·u
� � -�-
� ��� , 23 f4!
a • • m .:
•..

24 :es
24 l::txe6 llJxd3+ 25 cxd 3 l::txe6 is
16 h3 no improvement.
White probably arrived at this 24 • • • llJxd3+
strange move after examining some­ 25 cxd3 cxb5
thing like 16 l::th dl l::th e8 17 �2 26 :xb5 l::tb8
�g4, which does indeed look good 2 7 llJe5+ �d6
for Black. 28 l::txaS �h4+
16 •.• 0-1
17 l::th dl f5!
18 e5 Game 28
Forced, as 18 exf5 �xf5+ 1 9 �f2 Miles - Speelman
�xd3 20 cxd3 llJxd3+ and 18 llJg5 London 1982
�xc3 1 9 bxc3 fxe4 are both excel­
lent for Black. 1 d4 llJf6
18 .•. ttJd7! 2 llJc3 d5
19 l::te l �x e5 3 �g5 llJbd7
1 76 The Veresov
4 ttlf3 g6 (D)
Here, too, Black has a big choice,
but the King's Indian approach
seems the most logical to recom­
mend to King's Indian players.
5 e3
There are a couple of alternative
ideas:
1) 5 'ti'd2 (intending i.h6) and
now:
l a) 5 h6 6 i.f4 c6 7 0-0-0 i.g7
.••
8 h3? (better is 8 ttle5 with about
equal chances) 8 . . . l1Je4 ! 9 ttlxe4 dxe4
1 0 ttlh2 f5 1 1 h4 ttlb6 with adv an­
tage to Black, Zorigt-Olsson, Lugano
OL 1968.
lb) 5 ttle4! ? looks interesting.
•..
2) 5 'ti'd3 (aiming for a quick e4)
5 ... i. g7 6 e4 dxe4 7 ttlxe4 0-0 and
now:
2a) 8 ttlxf6+ ttlxf6 9 i.e2 c5 ! 1 0
dxc5 'ti'a5 + 1 1 c3 'ti' xc5 1 2 0-0 i.e6
1 3 'ti'd4 'ti'a5 1 4 a3 h6 ! 1 5 i.xf6 ! (15
i.f4 ttld5 ! +) 15 ... i.xf6 1 6 'ti'e3 i.g7
17 ttld4 i.d5 was very satisfactory
for Black in Smyslov -Gufeld, New
York 1989.
2b) 8 0-0-0 ttlxe4 9 'ti'xe4 ttlf6 10
i.xf6 (White was probably fright­
ened· of 1 0 . .. 'ti'd5 if he moved his
queen) 10 . . . i. xf6 1 1 ltJe5 i.e6 (I
prefer the more flexible 1 1 . .. c6) 1 2
f4 c6 1 3 h4 'ti'd5 1 4 'ti'xd5 cxd5 1 5
i.d3 i. xe5 16 dxe5 i.g4 with an
equal game, Mestrovic-Brenjo, Yu­
goslav Ch 199 1 .
5 ..• i. g7
6 i.d3
As White usually aims to play e4
as soon as possible this is the most
natural move.
6 i. e2 is the only serious alterna­
tive. After 6 . . 0-0 7 0-0 b6 8 l005
.
i.b7 9 i.f3?! ttle4 10 ttlxd7 ttlxc 3
1 1 'ti'd2 'ti'xd7 12 'ti'xc3 c5 1 3 'ti'd2
f6 ! 14 i.h4 cxd4 1 5 exd4 e5 Black
had the better game in Miagmas­
uren-Browne, Lucerne OL 1 982. 9
f4 has been suggested as an improve­
ment, upon which 9 . . . ttle8 !? is worth
considering.
6 .•. 0-0
7 0-0 c5 (D)
Better than 7 . . . b6 or 7 . . . c6 which
appear from time to time.
8 l:tel
Or:
1) 8 i.xf6? ttlxf6 9 dxc5 'ti'a5 10
ttlb5 a6 1 1 ttlbd4 ( 1 1 b4 'ti'xb4 1 2
ttlc7 l:ta7 !) 1 1 . . . 'ti'xc5 and with his

pair of bishops and strong centre
Black can already claim a sizeable
plus, Traudes-Gallagher, Liechten­
stein 1996.
2) 8 ttJeS (a more serious alterna­
tive but Black still seems to be able
to get a good game) 8 cxd4 9 exd4
•••
llJxeS 10 dxe5 llJg4 1 1 .i.e2 d4! 12
llJbS ( 1 2 .i.xg4 dxc3 1 3 bxc3 .i.xe5
1 4 .i.f3 flic7 ! is good for Black ac­
cording to Browne) 12 llJxeS 13 ••.
llJxd4 flib6! 14 c3 ( 14 .i.xe7 l:te8 15
.i. a3 l:td8 1 6 c3 llJc6 is an edge for
Black) 14 flixb2 1 5 .i.xe7 l:te8 1 6
••.
.i.b4 aS (D ) and now:
2a) Peters-Browne, USA Ch 1 98 1
continued 17 f/ib3 flixb3 1 8 axb3
.i.g4 ! 19 .i.c5 ! .i.xe2 20 llJxe2 llJf3+
2 1 gxf3 l:txe2 22 l:tfc 1 l:tb2 23 l:ta3
l:tc8 ! 24 l:txa5 l:txb3 25 c4 .i.b2 ! 26
l:tbl b6 27 :b5? ! (27 l:ta2 was the
best chance) 27 . . . :Xb5 28 cxb5 l:txc5
29 :Xb2 <itiif8 with a winning ending
for Black.
2b) 17 l:tbl is supposed to be a
significant improvement since after
1 7 . . . flixa2 1 8 l:tal Black is supposed
to avoid 1 8 . . . f/id5 ( 1 8 . . . f/ib2 =) on
account of 19 llJb5 . I don' t really un­
derstand this as after 1 9 . . . 'ti'xd l 20
The Veresov 1 77
.i.xd 1 (20 l:tfxd1 is even worse as the
bishop is en prise after 20 . . . llJc6 !)
20 . . . .i.d7 ! (20. . . llJc6 ! ?) 21 llJc7 axb4
22 l:txa8 l:tx a8 2 3 llJxa8 bxc3 White
still has a lot of work to do before he
can claim equality.
8 ... b6
The text has been Black's most
common choice but it can easily lead
to mass simplification. If you are
looking for a sharper struggle then
8 h6 9 .i.h4 (9 .i.f4 - but nobody
..•
plays it) 9 e6 looks worth a try. For
.•.
example:
1 ) 10 e4? cxd4 l l llJxd4 flib6 ! 12
llJb3 dxe4 1 3 llJxe4? ( 1 3 .i.xf6 +)
1 3 . . .llJxe4 14 l:txe4 f5 15 l:te3 g5 and
Black picked up a bishop for very lit­
tle, Schumacher-Gallagher, Liech­
tenstein 1990.
2) 10 h3 f!ib6 1 1 l:tbl a6 with a
couple of examples:
2a) 12 .i.n flic6 ! (preparing to
advance on the queenside) 1 3 �e5
llJxe5 1 4 dxe5 tDd7 15 f4 b5 16 f/id2
.i. b7 1 7 llJd l f6 1 8 exf6 llJxf6 1 9
.i. d 3 <ifiih 7 20 llJf2 e5 ! 2 1 fxe5 llJd7
22 c 3 �xe5 with an excellent posi­
tion for Black, Mariasin-Vorotnikov,
USSR 1 976.
2b) 12 'ti'd2 �h5 ( 1 2. . . flic6, as
above, looks better) 1 3 g4 llJhf6 1 4
b4? ! (Smith and Hall, in The Veresov
Attack, Chess Digest 1 99 4, claim
that 1 4 h4 gives White the advan­
tage; the only problem with this is
that there is already a bishop on h4 l ;
after 1 4 .i.g3, though, it's hard to see
what Black has gained by playing
. . . llJh5) 14 . . . cxb4 15 llJa4 flic6 1 6
flixb4 b5 l 7 llJc3 llJe4 1 8 .i.xe4 dxe4
1 9 llJd2 f5 20 gxf5 exf5 2 1 a4 g5 ! 22
1 78 The Veresov

axb5 'ii'g6 ! 23 i. g3 f4 with a clear i.c4 l:tc7 20 'ii'd6 'ii'xd6 21 l:txd6 a4

advantage to Black, Veresov-Kots, 1h- 1h Veresov-Shagalovich, Byelo­
USSR Cup 1 970. russian Ch 1 957 .
9 e4 dxe4 3 b ) White is not forced to ex­
10 l2Jxe4 i.b7 change on f6 as after 13 l2Jxc5 bxc5
1 1 l2Jxf6+ 1 4 l2Jb3 'ii'd5, there is the saving re­
1 1 c3 is a more critical move. Af­ source 1 5 'ii f3 ! when 15 'ii'xf3 1 6
•••

ter 1 1 cxd4 12 l2Jxd4 (D) there is:

••. gxf3 i.xf3 1 7 l:txe7 , 15••• 'ii'xb3 1 6
axb 3 i.xf3 1 7 gxf3, 1 5 'ii'xg5 1 6
.••

'ii'xb7 , 1 5 c4 1 6 i.xc4 ! and finally

..•

15 'ii'd7 1 6 l2Jxc5 ! all seem to be in

•..

White's favour. Black should prob­

ably play 14 'ii'c7 with a reasonable
•..

game.
11 ..• exf6
12 i.114
12 i.e3 f5 also gives Black good
play.
12 ... i.xf3!
13 'ii'xf3 cxd4
1) 12 l:tc8?! 1 3 'ii'e2 l2Je5 (the
••• 14 lladl
alternative 1 3 . . . l2Jc5 is probably bet­ 14 i.b5 is well by 14 . . . l:tac8 15
ter) 1 4 i.c2 l2Jc4 1 5 l:tad l 'ii'c7 1 6 lle2 l:tc5 ! .
l2Jb5 'fib s 17 i.xf6! i.xf6 1 8 l2Jxf6+ 14 ..• l%c8 (D)
exf6 19 i.b3 ltJe5 20 f4 lbc6 2 1 l2Jd6 Obviously not 14 ltJes? 15 l:txe5.
...

l:tc7 22 'ii'e8 ! l:txe8 23 l:txe8+ 'ii'xe8 White now came to the conclusion
24 l2Jxe8 l:te7 25 lLJd6 i.a8 26 l2Jxt7 ! that drastic action was required to
</;g7 27 lLJd8 ltJxd8 28 l:txd8 i.c6 29 avoid ending up in an inferior posi­
</;f2 +- Miles-Andersson, London tion.
1 982.
2) 12 tbxe4 13 i.xe4 i.xe4 14
••.

l:txe4 l2Jf6 15 l:te 1 'iid5 l 6 l2Jf3 'ii'b7,

Plaskett-Hazai, Maribor 1985 , is re­
puted to be fractionally better for
White.
3 ) 12 ltJcS (probably best) and
.••

now:
3a) 13 i. xf6 exf6 14 l2Jxc5 bxc5
1 5 l2Jb3 'ii' b6 (Black's position is dy­
namic enough to withstand the defi­
ciencies in his pawn structure) 1 6
'ii'e 2 f5 1 7 l:tad l f4 1 8 'ii'e7 l:tac8 1 9 15 i.a6! :Xc2
 The Veresov 1 79

16 Axd4 ttJes 23 Axes 1/2-1/2

17 'ii'd l f/Jc7 After a long sequence of forced
18 f4! tlJg4! moves the players agreed to a draw.
19 'ii'xg4 fS Black's extra pawn is meaningless
20 'ii'd l flies but it would still have been tempting
21 i.fl lhf2! to try a few more moves starting with
22 Ac4! Ad2+ 23 . i.d4+.
..
1 4 The Ba rry Attack

The Barry Attack is characterised by

the moves I d4 li:Jf6 2 li:Jf3 g6 3 li:Jc3
d5 (Pirc players can of course play
3 . . . i.g7 or 3 ... d6) 4 i.f4. The main
drawback, from White's point of
view, is that blocking the c-pawn , as
in its close relation the Veresov, can
easily lead to a lack of space on the
queenside.
The variation enjoyed a brief spell
of popularity a few years ago when
White won several games by march­ 5 'i'd.2, intending i.h6, does tend
ing his h-pawn up the board. This to invite 5 li:Je4 which seems to give
•.•

plan was most effective when Black Black a comfortable game. After 6
developed quietly so I am recom­ li:Jxe4 dxe4 we have a couple of ex­
mending a system of defence based amples:
on an early . . . c5 which rules out such I) 7 liJgS h6 ! ? (there is also noth­
crude behaviour as the centre will be ing wrong with 7 . . . 'i'xd4 = ) 8 li:Jxe4
too tense for wing attacks. The Barry g5 9 i.e5 f6 I 0 i.g3 f5 1 1 li:Jxg5 !
Attack has never acquired a great hxg5 1 2 'i'xg5 •xd4 ( 1 2 . . . 0-0 1 3
standing in the chess world and in 0-0-0 would be risky for Black) 1 3
fact earned its name as even its main 'i'g6+ 'ifi>f8 1 4 :.d I 'i'b4+ 1 5 c3
practitioners (including Grandmas­ i.xc3+ 16 bxc3 •xc3+ 17 :.d2 •c1 +
ters Hebden , Hodgson and Nor­ 1h- 1h Dolmatov-Gavrikov, Sverd­
wood) considered it to be a load of lovsk 1 984.
old Barry. 2) 7 !£Jes i.e6 8 e3 tt:xl 7 9 ltJc4
0-0 10 i.e2 i.xc4 ! (a well-timed ex­
Game 29 change which frees Black's game)
Josephs - Hebden 1 1 i. xc4 e5 1 2 dxe5 li:Jxe5 1 3 i.b3
Sheffield 1 991 a5 14 a4 li:Jd7 ! 1 5 0-0-0 li:Jc5 1 6
'i'xd8 :.fxd8 1 7 i.g5 :es 1 8 :.d5
1 d4 li:Jf6 i.f8 19 :.hd I li:Jxb3+ 20 cxb3 i.d6
2 li:Jf3 g6 w ith equality, Yusupov-Kasparov,
3 liJcJ dS Belfort 1 988.
4 i.f4 i.g7 (D) 5 ..• 0-0
5 e3 6 i.e2 cS! ?

Systems with an early . . . i.g4 are
also not bad but I like this temporary
pawn sacrifice which gives Black the
chance to take over the initiative on
the queenside. Another point in its
favour is that it has been the choice
of leading 'Barry ' exponent Mark
Hebden whenever he has been faced
by his own weapon.
7 dxcS
7 !DeS has also been tried, one ex­
ample being Rogers-J .Pol gar, Brno
1 99 1 which continued 7 . . . !Dc6 8 0-0
i.f5 9 dxc5 (White could find noth­
ing better than this capture as 9 i.f3
cxd4 1 0 exd4 !De4 looks pleasant for
B lack and other moves are not very
constructive) 9 . . .'ii'a5 ! 1 0 !Db5 ( 1 0
!Dxd5 ? !Dxd5 1 1 Wxd5 !Dxe5 1 2
i.xe5 :fd8 loses a piece) 1 0 . . . !De4 !
l l !Dxc6 (taking on d5 would still lose
material) 1 l . . . bxc6 1 2 tLJ<l4 Wxc5 1 3
!Dxf5 gxf5 14 c3 e5 and Black's
enormous centre gave her the advan­
tage.
7 ••• !Dbd7 (D)
7 Was has been played more
•.•
frequently, but in my view it is more
logical to aim for . . . !Dxc5 than
. . . Wxc5 .
The Barry Attack 181
8 0-0
8 fDxdS !Dxd5 9 'ii' xd5 i.xb2 1 0
: b l Wa5+ will cost White his right
to castle as 1 1 !Dd2 !Df6 followed by
1 2 . . . i.c3 would leave him in an
awkward pin . After 1 1 �fl , Wock­
enfuss-Hebden, Ostend 1 992 contin­
ued l l . . . i.g7 12 i.b5 ? ! ( 1 2 :b5 can
be met by 1 2 . . . Wc3 , but the text
looks even worse) 1 2 . . .!Df6 1 3 'ii' b 3
!De4 1 4 Wb4 Wxa2 ! 15 i.d3 a5 16
'ii' b6 !Dxc5 � 17 i.e5 !Dxd3 1 8 cxd3
f6 19 i.al e5 20 'ii'b 3 + Wxb3 2 1
.:.xb3 b5 ! and Black soon won.
I haven't seen any examples of 8
!DbS but one way of meeting this
would be to play 8 . . . a6 9 !Dc7 :a?
(9 . . . e5 is tempting, but after 1 0 i.xe5
!Dxe5 1 1 !Dxa8 !Dxf3 + 12 i.xf3
'ii'a5+ 1 3 c3 Wxc5 14 'ii'd4 ! the white
knight will escape) 1 0 !Dxd5 !Dxd5
1 1 Wxd5 i.xb2 1 2 :b l 'ii'a5+ with
similar play to the note above.
8 ... !DxcS
9 i.eS
White pins his hopes on restrain­
ing the black centre by blockading it
on the dark squares, at least until he
has had time to play c2-c4. A diffi­
cult task !
The slow 9 h3 was featured in the
game Izeta-Khali fman, Dos Herma­
nas 1993, with the idea of preventing
the annoying . . . i.g4 . However, after
9 . . . b6 1 0 !Db5 i.b7 1 1 i.e5 a6 1 2
!Dbd4 !Dcd7 1 3 i.h2 :cs (prevent­
in g c2-c4) 1 4 c3 !De4, with . . . e5 to
follow, Black had already assumed
control .
9 .
.. i.g4!?
This shows good understanding
of the position. Once the knight on f3
182 The Barry Attack
has disappeared White will have lit­
tle hope of controlling the centre.
9 llJcd7 is also quite interesting
•••
and has in fact been played more
often than the text. After 10 i.d4
'flc7 (D) there is:
1) 11 lllbS ? t 'flb8 12 c4 (the point
behind lll b5 ) 1 2 . . . dxc4 ! 1 3 i.xc4 a6
and I can ' t see how White can avoid
material loss, e.g. 14 llla 3 ( 14 lllc 3
e5) 14 . . . e5 ! 1 5 i.c3 b5 and White
has nothing better than the sad 1 6
i.b4.
2) 11 lLlxdS lllxd5 12 i.xg7 lbxe3
1 3 fxe 3 �xg7 14 'fld4+ f6 followed
by . . . e5 should be quite good for
Black although White may have
some slight attacking chances.
10 h3
On 10 i.xf6 i.xf6 1 1 'flxdS:
1 ) ll i.xc3 is probably enough
•.•
for equality after 12 'flxd8 :xd8 1 3
bxc3 lll a4 but 1 2 'fixes i.xb2 1 3
:abl i.f6 ( 1 3 . . b6? 14 'iib4) 14 :xb7
. 1 7 . . . fxe6 is not bad for Black while
:cs ( 1 4 . . . 'flc8 1 5 'flxb7 :fxc8 1 6
i.d3 is probably tenable) 1 5 'Wb4 !
i.xf3 1 6 i.xf3 :xc2 17 :xa7 leaves
Black fighting for a draw.
2) 1 1. 'flb6 is more to the point.
..
After 12 :abl ( 1 2 llxi4 i.xe2 1 3
llldxe2 { 1 3 lllcxe2 is rather similar }
1 3 . . . :fd8 1 4 'flc4 :ac8 gives Black
good play, as does 12 'flc4 ! ? i.e6 1 3
llld 5 i.xd5 1 4 'flxd5 :fd8 1 5 'flc4
ltac8) 12 i.e6! ? 13 'fld2 Jhd8 (D)
.•.
White has several tries :
2a) 14 llld4 i.c8 ! (the threat of
1 5 . . . e5 is surprisingly awkward to
meet whilst the reason for playing
1 3 . . . :ad8, as oppo sed to the more
natural 1 3 ... :fd8, is now revealed: in
the latter case the rooks would have
been disconnected after the bishop
retreat) 15 i.f3 ( 1 5 'fie 1 +) 1 5 . . . e5
1 6 llld5 :xd5 1 7 i.xd5 exd4 1 8
cxd4 llle6 1 9 c3 'fla5 ! 20 i.xe6 (oth­
erwise White will lose his d-pawn ,
e.g. 20 i.e4 i.xd4 2 1 b4 'fig5 ! )
20 . . . i.xe6 and Black's bishop pair is
in no way inferior to the rook and
pawns.
2b) 14 'flcl i.xc3 15 bxc3 'fla5
1 6 llld4 'flxc3 17 lllx e6 and now
1 7 . . . lll x e6 1 8 :xb7 :d2 also de-
serves consideration.
2c) 14 'flel i.f5 1 5 'flc l ( 1 5 i.d l
is very passive) 1 5 . . . i.xc 3 1 6 bxc3
'fla5 and again Black has no prob­
lems.
 The Barry Attack 183

2d) 1 4 i. d3 i.c4 1 5 lbe 1 and

now 15...i.xc3 16 1Vxc3 ( 1 6 bxc3
1Va5) 16 . . .i.xa2 wins the pawn back
with a roughly level g ame, while
pressure-increasing moves such as
15 ...:d7!? deserve consideration.
Now we return to the main line af­
ter 10 h3 (D):

14 ... 1Vf6
15 1Vd2
15 e4 is still met by 1 5 . . . d4, e.g.
1 6 li:Jb5 e5 17 ll:Jc7 :ac8 18 li:Jd5
ifd6 19 c3 lbe6 with a fine game for
Black as he w ill be soon able to ex­
change off the strong knight on d5 .
White can easily drift into a very bad
position, for example 20 :d2 li:Jb6
10 •.• i.xf3 2 1 :rd 1 li:Jxd5 22 exd5 dxc3 23 bxc3
1 1 i.xf3 e6 lLJc5 would be positionally lost.
All schoolchildren (well, at least 15 . . . li:Jb6
Russian schoolchildren) are taught 16 1Vd4 1Vxd4
not to block their c-pawn in queen's 17 exd4 ll:Jca4
pawn openings and this position is a 18 ll:Jxa4 ll:Jxa4
classic example; White needs to play 19 :bl b5 ( D)
c2-c4 to activate his pieces and chal­
lenge the black centre but with his .I�,,w
��� - �
& ·
�-• -
�
:?! 0 &
knight on c3 this will take too long to w

arrange. ,� � •. �-
� Ri �i8
,

12 1Ve2 li:Jfd7 !
Exchanging White's most active -i�i- •
piece is a goo d idea, especially as �
�/v::� -� '/* �
.�ft,;;: :u
� ?T �..
/•
� w%.·
-:j:,w@#. ,,
Black's queen will be able to replace
•�•. fl..tE LS
the bishop as guardian of the long di­
agonal.
LS Pd LS ,.
� LS
�
%. ,
13 i.xg7 <:Ji/xg7 � • •:
14 :ad l (D)
White hopes to equalise by play­ Black has the advantage as he has
ing e4, which would of course have something to undertaket namely a
been met by 14 . .. d4 if played at once. minority attack on the queenside,
184 The Barry Attack

whilst White has no active ideas of 26 �e2 g5 27 h4 h6 28 hxg5 hxg5 29

his own. With the centre blocked l:.h l l:.b8 30 l:.h7+ �f6 3 1 a3 a5 32
Black's knight is also the superior f4 gxf4 33 :n b4 34 l:.xf4+ �e7 35
minor piece. i.g6 l:.f8 36 cxb4 axb4 37 axb4 l:.c4
. .

The remaining moves were 20 c3 38 i.d3 l:.xb4 39 g4 e5 40 dxe5 ltxf4

ltfc8 2 1 i.d 1 llJb6 22 i.e2 a6 23 4 1 exd6+ �xd6 42 g5 �e5 43 l:.h5
.:Jc 1 l:.c6 24 �f 1 ltJc8 25 i.d3 lD<l6 �d4 O� l .
1 5 Blackmar-Diemer Gambit
I must confess that I had assumed
this Gambit to be only playable
against l . d5, as after 1 d4 tLlf6 2
. .
tLlc3 d5 3 e4 Black can simply play
3 . . . tLlxe4. When Gary Lane's recent
book, The Blackmar-Diemer Gambit
(B atsford 1 995 ) arrived the first
th ing I learnt was that this is called
the Hubsch Gambit, while the sec­
ond thing I learnt was that the
Hubsch Gambit is not so bad and
that Black can probably only obtain
an equal game . As we are looking
for more than this, the Blackmar­
Diemer has to be accepted . After
3 . . . dxe4 4 f3 exf3 5 tLlxf3 Black has
a wide choice but my vote goes to
5 . . . e6, the solid Euwe Defence.
The material below is not in the
usual complete game format as I
have been unable to locate a game
that suits my purposes. In Lane 's
book, for example, Blackmar, Die­
mer and various others (of whom the
most prominent is Diebert, who
probably employed the Gambit so
frequently as an attempt to have it
renamed the Blackmar-Diebert Gam­
bit perhaps, or more fittingly the Die­
mer-Diebert Gambit) score crushing
victories against all-comers. At first I
was a little intimid ated by this and
the thought of being bogged down in
my final chapter for weeks was not a
pleasant one. I even wondered if I
might have to abandon this whole
project, bowing to the superiority of
the Blackmar-Diemer Gambit. Hap­
pily, this nightmare scenario was
averted as a close examination of the
fine print in Lane's book revealed a
number of areas where Black could
fight for the advantage.
1 d4 tLlf6
2 tLlc3 d5
3 e4 dxe4
4 f3 exf3
5 tLlxf3
In return for his central pawn
White has received a tempo for de­
velopment and some vague attacking
chances due to the semi-open f-file ­
not really enough compensation but
Black must still take care as inaccu­
racies can be swiftly punished in the
Blackmar-Diemer.
5 'Yixf3, known as the Ryder
Gambit, generously offers Black a
second central pawn and he should
not hesitate before snapping it up:
5 ... 'Yixd4 6 i.e3 'Yig4 ! (gaining an
important tempo; 6 . . . 'Yib4? ! 7 0-0-0
i.g4? 8 tLlb5 ! ! should definitely be
avoided) 7 'Yif2 e5 (D) and now:
1 ) 8 i. e2 should be answered by
8 . . . 'Yif5 and 9 . . . i.b4 according to
Lane.
2) 8 i.d3 i.b4 9 tLle2 e4 10 i.c4
i.e6 1 l i.xe6 'Yixe6 was fairly hope­
less for White in Prins-Schneider,
corr 1 989.
186 Blackma.r-Diemer Gambit
3) 8 a3 (prevents the annoying
. . . i.b4, but if White is reduced to
this . . . ) 8 liJc6 9 liJf3 i.d6 1 0 0-0-0.
. . .
Lane make s a rather half-hearted
attempt to claim compensation for
White here but he obviously believes
that Black is much better. His main
game now concentrates on the ri­
diculous 10 a6; 10 .'ii'e6, returning
•.. ..
the queen from enemy territory
looks more to the point as 1 1 liJg5
ile7 1 2 i.c4 0-0 is nothing to worry
about.
4) Hodgson suggests 8 liJf3 fol­
lowed by 0-0-0 as being more in the
spirit of the opening :
4a) 8 e4 is perhaps a little opti­
•••
mistic. After 9 liJd4 (9 liJe5 iif5 1 0
iixf5 i.xf5 1 1 i.c4 i.e6 i s good
enough) 9 ... liJc6 ! ? 10 i.e2 liJxd4 1 1
i.xg4 ? liJxg4 1 2 iid2 liJxe3 1 3 :c l
liJdf5 White i s lost but he should
play 11 i.xd4 with some compensa­
tion. Black can of course play some­
thing else on his 9 th move.
4b) 8 liJc6 and 8 .i.d6 are very
•.• ••
playable but Black may end up a
tempo down on variation ' 3 ' . Per­
haps 8 i.b4 is best as after 9 liJxe5
.••
ile4 1 0 liJc4 i.xc3+ 1 1 bxc3 i.e6 !
(better than 1 1 . . . liJg4 1 2 ilf4 ! with
some chances for White) White has
nothing for the pawn. 9 0-0-0 is
probably his best chance.
5 .. . e6 (D)
This seems to be the most reliable
set-up. Black calmly develops his
pieces before seeking counterplay in
the centre with . . . c5 . It is true that
this locks in the queen's bishop, but
one can't have everything in life, not
e ven against the Blackmar-Diemer
Gambit.
6 i.gS
On 6 i.d3 (the bishop is better
here than on c4 as White usually tar­
gets h7) Black can simply continue
with 6 .. . i.e7, or take the opportunity
to play an immediate . . . c5, for exam­
ple: 6 cS 7 i.e3 cxd4 (7 . . . liJc6 8
•..
d xc5 iia5 also looks possible) 8
liJxd4 i.e7 9 0-0 0-0 1 0 ilf3 ( 1 0
ile l liJg4) 1 0 . . . liJbd7 ( 1 0 . . . e5 ! ?) . If
one were to stick a white pawn on
f2 I would still be happy with the
black position, but I suppose Black­
mar-Diemer fans may argue that it is
precisely because this pawn is miss­
ing that White has good attacking
chances. I' m not sure how he should
continue the attack though. If 1 1

'ii'h 3 then 1 1 . . . ltJc5 is good as 1 2
ltxf6 liJxd3 ! is not in White 's favour.
Other moves are also liable to be met
by . . . ltJe5 or . . . ltJc5 exchanging off
the dangerous bishop. Perhap s Black
will still have to soak up a little pres­
sure but he has an extra pawn and no
weaknesses.
6 ltJeS, a suggestion of Diemer,
also deserves a brief mention. After
6 . . . ltJbd7 7 'ii'f3 i.e7 8 i.g5 0-0 9
i.d3, Decleir-Viaene, Belgium 1 988,
Black should play 9 . . .c5 (in the game
he opted for the passive 9 . . . c6) after
which White will be hard pressed to
avoid unfavourable exchanges. 1 0
'ii'h 3 should, o f course, be met by
10 g6 rather than 10 h6.
••• •••
6
.• . i.e1 (D)
Euwe's original idea was to play
6 cS here but after 7 i.xf6 ! gxf6
•••
(7 ... 'ii'x f6 8 i.b5+ i.d7 9 0-0 is very
risky) 8 d5 e5 (I don' t see anything
better) 9 i.c4 White has real posi­
tional compensation.
White usually chooses between:
A: 7 'ii'd.2 , which can be followed
by 0-0-0; and
B : 7 i. d3 , usually followed b y
0-0.
Blackmar-Diemer Gambit 187
A)
7 'ii'd.2 0-0
8 i.d3
8 0-0-0 is not mentioned in Lane's
book but 8 . . .c5 still looks like a good
reply.
8 •.. cs
9 'ii'f4
A standard queen manoeuvre in
the B lackmar-Diemer; this case ap­
paers to be even more tricky than
usual .
9 .•. cxd4! (D)
In Diemer-Anon, France 1 957,
Black wasn't careful and got blown
away: 9 liJdS? 1 0 i.xh7+ ! �xh7
. .•
1 1 'ii'h 4+ �g8 1 2 i.xe7 'ii'xe7 (or
1 2 . . . ltJxe7 losing after 1 3 ltJg5 lle8
1 4 'ii'h 7+ �f8 1 5 'ii'h5 ! ltJg6 1 6 llfl )
1 3 ltJg5 ltJf6 14 ltJce4 ltJbd7 1 5 0-0 !
lle8 1 6 llxf6 ! ltJxf6 17 llfl ! 1-0.
10 'i'h4!
Definitely White 's best chance.
Others:
1) 10 ltJxd4 liJh5 ! when 11 i.xe7
ltJxf4, 1 1 'ii'e4 f5 and 1 1 'ii'h4 i.xg5
1 2 'ii'xh5 g6 13 'ii'g 4 e5 are all hope­
less for White.
2) 10 0-0-0 dxc3 ! 1 1 i.xh7+
�xh7 12 llxd8 cxb2+ 13 �bl llxd8
188 Blackmar-Diemer Gambit

and the queen has cost White too

much material.
3) 10 ltJe 4 ltJd5 ! 1 1 'ii'h4 i.xg5
12 ltJexg5 'ii'a5+ ! (not 1 2 . . . h6?? 1 3
i.h7 + �h8 1 4 ltJxf7+ !) 1 3 �e2 h6
with advantage to Black.
10 ••. dxc3 ! ?
A little bit o f fantasy. There are
two alternatives, one good and one
bad.
1 ) 10 h6 1 1 i.xh6! (naturally)
•••

and now: pieces (very important for minor

1 a) 1 1 dxc3 1 2 i.xg7 ! �xg7
•.. pieces battling against a queen). The
1 3 'ii'g 5+ �h8 14 'ii' h 6+ �g8 and outcome will hinge on the fate of the
White has at least a perpetual and b2-pawn and on whether Black can
probably more after 15 0-0-0. activate his potentially powerful
l b) 11 gxh6 1 2 'ii'xh6 'ii'a5 1 3
... queen's bishop. One idea for Black is
ltJg5 dxc3 and again White has a per­ to continue with . . . :d5 and . . . i.d7,
petual with 14 i.h7+ and perhaps whilst another (especially if White
more, e.g. 14 0-0 ltJbd7 1 5 :r3 cxb2 plays 1 6 :rd 1 , intending d4 and
1 6 :an 'ii'b 6+ 17 �hl bl 'ii' 1 8 :h3 ! :xb2) is to play . . . ltJd4 as an ex­
and mate follows shortly. change of knights would almost cer­
2) 10 g6! is the most solid con­
.. • tainly be in Black's favour.
tinuation, making it very difficult for
White to sacrifice anythi ng success­ B)
fully. 1 1 ltJe4 should be met by 7 i.d3 (D)
1 1 . . . ltJd5 ! when the possibilities of
. . . f6 and . . . f5 should enable him to
defend his kingside with ease.
1 1 i.xf6 'ii'xd3!
12 cxd3 i.xf6
13 'ii'c 4 cxb2
14 :ht (D)
Black's defensive combination
has netted him two bishops and three
pawns for the queen - a rough ma­
terial equality. After something such
as 1 4 . . . ltJc6 1 5 0-0 ( 1 5 d4 :d8 1 6
:xb2 ltJxd4 i s dangerous for White) 7 .. . cS
1 5 . . . :ds it is clear that Black has a Or:
very compact position with several 1 ) 7 ltJbd7 8 0-0 c5 9 dxc5
••.

secure posts in the centre for his ltJxc5 10 i.b5+ .i.d7 1 1 i.xf6 .i.xf6
 Blackmar-Diemer Gambit 189

1 2 1We2 and, according to Lane, this

position is assessed as = by Leise­
bein. I woul d have thought that
White is completely lost! I have no­
ticed a common trend among Black­
mar-Diemer analysts; once there is
no attack and the position looks
rather balanced they tend to assess
the game as = , conveniently forget­
tin g the fact that they are a pawn
down. I think that 8 1Wd2 is better as
8 . c5 9 0-0-0 looks quite unclear to
.. Now White has to agree to the ex­
me. change of queens or donate a second
2 ) 7 li:Jc6!? 8 1Wd2 (the inferior
••• pawn. There is little doubt as to
8 a3 is often played when 8 ... li:Jd5 which path Blackmar-Diemer fans
looks very good for B lack) 8 . . . li:Jb4 will foil ow.
(Black is willing to waste some time 13 li:Jd4 1Wxb2
in order to exchange off White's most 14 liJce2 ltJes
dangerous attacking piece) 9 0-0-0 15 a4
li:Jxd3+ 1 0 1Wxd3 0-0 1 1 h4 c5 (other Threatening to trap the queen
moves lead to a passive situation) 1 2 with 1 6 :fbl .
i.xf6 (perhaps White can improve 15 •.. 'iffb6
upon this ) 1 2 . . . i.xf6 1 3 dxc5 1Wxd 3 16 :xr6
14 :xd3 i.xc3 1 5 :xc3 f6 ! followed An attempt to confuse the issue in
by . . . e5 with a better ending for a lost position.
Black. 16 ... gxf6!
8 dxc5 1Wa5 The ' ! ' may seem to be a little ex­
8 li:Jbd7 allows 9 b4 whilst the
.•• cessive but in Sneiders-Breunig,
immediate capture 8 i.xc5 is un­
••• Corr 1 970- 1 Black was bluffed out
trustworthy. and played 16. li:Jxd3?. After 17
..

9 o ..o :ffl f6 1 8 i.e3 liJc5 ( 1 8 ... li:Je5, but it

9 1Wd2 can be met by 9 . . . li:Jbd7. looks dodgy) 1 9 li:Jb3 ! 1Wc6 (l 9 . . e5.

9 •.. 1Wxc5+ 20 1Wf2 ! ) 20 i.xc5 i.xc5 2 1 1Wh5 +

10 �bl li:Jbd7 (D) he had lost a piece and the game.
B lack is in no great rush to castle 17 i.xf6 1Wd8!
as there are other useful things to do, 18 :n
s uch as developing the queenside. What else?
1 1 lWel a6 18 ••. i.x f6
With what B lack has in mind it 19 :xr6 li:Jg4!
wil l be essential to deprive White of The game is over. After 20 lt:Jxe6,
the b5 -square. the simplest is 20 . . . 1Wxf6 2 1 li:Jc7 +
12 1'b4 'iffb4! �d8.
Index of Variations

A: King's Indian: White avoids 8 .i.e3 37

the main lines 8 .i.g5 43

1 d4 lbr6 5 lDf3 0-0

2 c4 g6 6 .i.e3 e5 7 dxe5 dxe5 8 11fxd8
3 lbc3 .i.g7 .:.xd8 9 lbd5 1 04
4 e4 6 .i.e2 e5:
7 d5 a5 8 h3 47
4 .i. g5 72 7 dxe5 dxe5 8 11fxd8 .:.xd8 85
9 .i.g5 .:es:
4 lDf3: 1 0 0-0-0 86
4 . . 0-0:
. 10 lDd5 lbxd5 1 1 cxd5
5 e3 d6 1 02 c6 1 2 .i.c4 cxd5 1 3 .i.xd5 lbd7 87
5 i.g5:
5 . . . c5 72 5 .i. e2 o..o:
5 . . . d6 74 6 .i.e3 106
4 . . . d6 6 g4 106
5 e3 109 6 h4 107
5 .i.g5 74: 6 .i.g5 lDa6 53 :
5 . . . c6 74 7 .i.f3 54
5 .0-0
. . 7 f4 54
6 e4 75 7 lDf3 59
6 e3: 7 h4 59
6 . . . c5 77 7 11fd2 65
6 . . . lDbd7 75
5 .i.g5 h6 6 .i.h4 c5 82
4 ... d6 :
5 .i.d3 0-0 6 lbge2 9 1 :
5 h3 0-0 3 1 : 6 ... c5 9 1
6 .i.g5 3 1 6 . . . e5 9 1
6 .i.e3 3 1 6 . . . lbc6 7 0-0 92
6 lDf3 36 6 . . . e5 :
7 dxe5 37 5 lbge2 97 :
7 d5 : 5 . . . c6 97
7 . . . lDhS 37 5 . . 0-0 6 lDg3 eS 7 d5 98
.

7 . . . lba6: 5 . . . a6 6 lDg3 c6 99
 Index of Variations 191

5 f4 0-0 6 lLlf3 lL>a6 10 4 f3 :

7 i.e3 10 4 . . .c5 172
7 c5 10 4 . . .c6 5 e4 dxe4 6 fxe4
7 e5 20 e5 173
7 i.d3 20
7 i.e2 e5 2 lLlf3 g6:
8 0-0 1 1
8 dxe5 1 1 3 b4 109
8 fxe5 16 3 g3 i.g7 4 i.g2 0-0 5 0-0 d6 157 :
6 4Jc 3 1 57
6 b3 157
B: 1 d4: White avoids the King's 6 :el 164
Indian 6 lLlbd2 1 64
6 a4 166
1 d4 lLlf6 3 lLlc3 d5 4 i.f4 j.g7:
5 'ti'd2 180
2 i. g5 1 1 1 2 ltJe4 :
••• 5 e3 180 5 . . . 0-0 6 i.e2 c5:
3 i.h4 1 1 2 7 lLJe5 1 8 1
3 h4 1 15 7 dxc5 lLJbd7 18 1
3 i.f4 c5 1 2 1 : 3 i.g5 1 37 3 . . .i.g7 4 lLJbd2 0-0:
4 d5 'ii'b 6 121 5 e3 1 37
4 f3 'ii'a5 + 5 c3 lLlf6 1 27 5 e4 d5 138
5 c3:
2 lLJc3 d5 5 . . . d5 142
3 e4 dxe4 4 f3 exf3: 5 . . . d6 6 e4 c5 143
-
5 'ii'xf3 1 85 3 i.f4 148 3 . . . i.g7 :
5 lLJxf3 1 85 5 . . . e6: 4 c 3 153
6 i.d3 186 4 lLJbd2 0-0 5 e4 148
6 lLJe5 187 4 e3 d6 5 h3 (5 i.e2 148) 5 . . . 0-0:
6 i.g5 i.e7 : 6 c4 149
7 i.d3 1 88 6 i.c4 lLlbd7 7 c3 153
7 'ii'd 2 1 87 6 c3 153
. 3 i.g5 lLJbd7 170: 6 i.e2 lLlbd7 149:
4 lLJf3 g6 176 7 c4 149
4 e4 170 7 0-0 'ii'e8 (7 . lDe4 1 50):
. .

4 'ti'd3 17 1 8 c4 150
4 e3 17 1 8 c3 1 53