The Tarrasch Defence: Move by

Move
by Sam Collins

Contents
About this Publication
Bibliography
Foreword
Introduction
Structural Introduction

1. 9 Bg5 cxd4
2. 9 Bg5 c4
3. 9 Bg5 Be6
4. 8 dxc5/9 dxc5
5. Other 8th/9th move options
6. Symmetrical Tarrasch
7. Tarrasch sidelines
8. Réti set-ups
9. In place of an epilogue...
 Back to Contents Page
The Tarrasch Defence: Move by Move
This series provides an ideal platform to study chess openings. By continually challenging the reader to answer
probing questions throughout the book, the Move by Move format greatly encourages the learning and practising of
vital skills just as much as the traditional assimilation of opening knowledge. Carefully selected questions and
answers are designed to keep you actively involved and allow you to monitor your progress as you learn. This is an
excellent way to study any chess opening and at the same time improve your general chess skills and knowledge.
The Tarrasch is an ambitious defence to the Queen's Gambit. Black's concept is an aggressive one – he willingly
accepts an isolated queen's pawn but in return seizes space in the centre and gains freedom of development for all
his pieces. In this book, International Master and Tarrasch expert Sam Collins takes an in-depth look at this popular
opening and its many variations. Using illustrative games, Collins highlights the typical plans and tactics for both
sides, presents repertoire options for Black and provides answers to all the key questions.
· Essential guidance and training in the Tarrasch Defence
· Utilizes an ideal approach to chess study
· Includes repertoire options for Black

Sam Collins is an International Master with two Grandmaster norms, and a former Irish and Japanese Champion.
He has represented Ireland at seven Olympiads, winning an individual gold medal at Bled 2002. He has a wealth of
teaching and writing experience, and has produced many books, DVDs and magazine articles on chess.

Publisher Information

The Tarrasch Defence: Move by Move

by Sam Collins

First published in 2013 by Gloucester Publishers plc (formerly Everyman Publishers plc)
Northburgh House, 10 Northburgh Street, London EC1V 0AT.
Copyright © 2013 Sam Collins
The right of Sam Collins to be identified as the author of this work has been asserted in accordance with the
Copyrights, Designs and Patents Act 1988.
All rights reserved. No part of this work may be reproduced, or transmitted in any form or by any means, without
prior permission of the publisher.

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library.
All other sales enquiries should be directed to Everyman Chess, Northburgh House, 10 Northburgh Street, London
EC1V 0AT
tel: 020 7253 7887; fax: 020 7490 3708
email: info@everymanchess.com
website: www.everymanchess.com
Everyman is the registered trade mark of Random House Inc. and is used in this work under licence from Random
House Inc.

EVERYMAN CHESS SERIES (formerly Cadogan Chess)

Chief Advisor: Byron Jacobs
Commissioning Editor: John Emms
Assistant Editor: Richard Palliser

Back to Contents Page

 Back to Contents Page
Bibliography
1 d4 Volume One, Boris Avrukh (Quality Chess 2008)
Chess Duels: My Games with the World Champions, Yasser Seirawan (Everyman 2010)
Garry Kasparov's Greatest Chess Games Volume 1, Igor Stohl (Gambit 2005)
Garry Kasparov on My Great Predecessors Volumes 1-5, Garry Kasparov (Everyman 2003-06)
Garry Kasparov Part I: 1973-1985, Garry Kasparov (Everyman 2011)
I Play Against Pieces, Svetozar Gligoric (Batsford 2002)
Kasparov vs Karpov 1975-1985, Garry Kasparov (Everyman 2008)
My Best Games, Anatoly Karpov (Edition Olms 2007)
My Best Games Vol.2: Games with Black, Victor Korchnoi (Edition Olms 2001)
My Most Memorable Games, Boris Gelfand (Edition Olms 2005)
Opening Preparation, Mark Dvoretsky & Artur Yusupov (Batsford 1994)
Paul Keres: The Quest for Perfection, Paul Keres and John Nunn (Batsford 1997)
Revolutionize Your Chess, Viktor Moskalenko (New in Chess 2009)
Secrets of Grandmaster Chess, John Nunn (Batsford 1997)
The English Opening: Volume Two, Mihail Marin (Quality Chess 2010)
The Kaufman Repertoire for Black and White, Larry Kaufman (New in Chess 2012)
The Modern Réti, Alexander Delchev (Chess Stars 2012)
The Tarrasch Defence, Jacob Aagaard and Nikolaos Ntirlis (Quality Chess 2011)
Winning Chess Middlegames, Ivan Sokolov (New In Chess 2008)

Websites
www.chesspublishing.com
www.chessbase.com

Software/Databases
ChessBase 12
Megabase 2013

Magazines
The London Review of Books

Back to Contents Page

 Back to Contents Page
Foreword
Move by Move is a series of opening books which uses a question-and-answer format. One of our main aims of the
series is to replicate – as much as possible – lessons between chess teachers and students.
All the way through, readers will be challenged to answer searching questions, to test their skills in chess openings
and indeed in other key aspects of the game. It's our firm belief that practising your skills like this is an excellent
way to study chess openings, and to study chess in general.
Many thanks go to all those who have been kind enough to offer inspiration, advice and assistance in the creation
of Move by Move . We're really excited by this series and hope that readers will share our enthusiasm.

John Emms
Everyman Chess

Back to Contents Page

 Back to Contents Page
Introduction

First Thoughts
The Advantage of Playing White
Fashion
Begging the question: What Happens When You Don't Define Your Terms

First Thoughts

An excellent potted history of the Tarrasch is set out by one of its exponents in the 1970s, GM John Nunn:
"...the Tarrasch, a defence which has always been considered an uphill struggle for Black, but which becomes
fashionable for short periods when some great player attempts to revive and improve it. Spassky used it for a while
in the 1960s, and more recently Kasparov adopted it twice in his first match with Karpov. Unfortunately, he lost on
both occasions and then largely abandoned it. I played it occasionally in the seventies, but after some poor results I
gave it up. My experience indicated that Black's winning chances were slim since White could usually draw even if
he played badly, while strong opponents would generally win! Today [Nunn was writing in 1997 - SC] only a few
players are willing to adopt it, Murray Chandler being one of this vanishing breed."
Dr. Nunn's comments represent the consensus view and are certainly an accurate representation of the historical
development of the Tarrasch before 2000. They throw up a lot of interesting material for discussion.

The Advantage of Playing White

On one analysis, White drawing if he plays badly and generally winning if he is a stronger player, is something
which happens in every opening. So what Nunn is describing isn't so surprising. A look at the statistics from
Megabase 2013 (admittedly a rough and ready analysis since I haven't corrected for rating) bears this out. In the
position after 1 d4 d5 2 c4 e6 3 Nc3, against 3...c5 White scores 54.8% from 8788 games (a lower percentage than
he achieves after 3...Nf6 or 3...Be7). In the position after 1 d4 d5 2 c4 e6 3 Nf3 (while this signals the Catalan, it is
also specifically an anti-Tarrasch move order since White has more options when he hasn't committed his knight to
c3), White scores 60.6% from 6484 games against 3...c5 (which is 0.1% less than he makes against 3...Nf6).
Looking at my own games provided interesting food for thought. I have played the Tarrasch against all kinds of
opponents, including some under 2000 and many strong GMs: Korchnoi, Khenkin, Baburin (twice), Pert, Bischoff,
San Segundo, Hebden, Burmakin, Socko. From what I can piece together, in 22 games with the line I scored 10
points. From a pure comparison of player's ratings, my expected score was 11 points. Given that White tends to
score approximately 55% in most openings, this result is pretty acceptable. The statistics are skewed in White's
favour by some other factors: first, if one looks at the opening positions I got from those games, the results ought to
have been even better (losses to Baburin, Pert, Khenkin and Socko occurred in equal or very playable
middlegames); second, in certain games (in particular against Bischoff) I horribly mishandled the opening and
genuinely played an awful game.
Accordingly, my own limited experience provides no reason to doubt the efficacy of the Tarrasch.

Fashion

Nunn is quite right to point to Kasparov's two losses with the Tarrasch in his first World Championship match
against Karpov damaging the opening's popularity in quite a permanent way.
Looking objectively at these two games, in the first (Game 7) Kasparov easily equalized and only lost due to a late
blunder. In the second (Game 9), against the best chess technician in history, Kasparov blundered on move 46 and
lost in 70 moves. To put it mildly, this was not a result of the opening. If openings were shelved because arising
endgames were lost against Karpov in the mid-80s, there wouldn't be a lot of openings played anymore.
It's quite hard to find someone who justifiably dropped the Tarrasch. For instance, Alexander Grischuk was the hero
of this opening in the early 2000s (and laid the foundation for many of the modern main lines, in particular through
his excellent use of 9 Bg5 c4!). However, his loss to Gelfand (Russian Team Championship 2004) seems to have
been Grischuk's last game with the Tarrasch at a classical time control (admittedly he gave the line a spin at the
Amber tournament in 2006, scoring ½/2 against Van Wely and Ivanchuk). So did Gelfand show a good antidote
against the Tarrasch? Actually, no. The game followed the same first 13 moves asSargissian-Halkias (covered
later in this volume) where, as you will see, 13...Nxd4! is a clean equalizer.
Moreover, let's see the endgame which resulted in Gelfand-Grischuk:
How did Gelfand know how to exploit his advantage here? Well, let's ask him!
"...I was very optimistic about the double-rook ending. In my first encounter with Anatoly Karpov in Linares 1991 I
had a similar ending, in which with Black I was defending a better pawn structure (my pawn was on g6, not on f6)
and where I was confident that we would soon agree a draw. But the 12th world champion taught me a memorable
lesson, allowing me no chances in such a harmless-looking position. My hope of exchanging my a- and c-pawns for
the a-pawn proved to be a mere illusion."
Let's have a look at the position Gelfand is referring to:

And the opening in that game was a Grünfeld, which nobody (including Gelfand) ditched as a result of the game.
Moreover, just how bad is this endgame? I've found the following examples of this or similar endgames arising from
the Tarrasch (many following the same line as in Gelfand-Grischuk):

V.Borovikov-W.Pereira Neto, Nettetal 2004 (½-½ in 60)

G.Pitl-E.Lie, Gausdal 2007 (½-½ in 38)
Wang Hao-S.Iuldachev, Asian Team Championship 2008 (1-0 in 64)
Zhou Jianchao-V.Akobian, Beijing (blitz) 2008 (½-½ in 66 - as Black is a rook up in the final position I'm not sure
about this result!)
Mi.Hoffmann-K.Kiik, European Cup, Ohrid 2009 (1-0 in 39)
A.Matthiesen-Kar.Rasmussen, Aarhus 2010 (½-½ in 61)
Zhu Chen-E.Kovalevskaya, FIDE Grand Prix, Nalchik 2011 (½-½ in 44)
I.Kragelj-T.Suc, Trieste 2012 (½-½ in 34)

So what looks like a forced win in the hands of Karpov or Gelfand starts looking pretty unclear at lower levels (even
"normal" GM level). Moreover, this endgame is close to the maximum White can hope for in the 9 Bg5 c4 line
(admittedly, as Gelfand observes, if White can keep the queens on Black may have some additional problems to
solve, as in S.Mamedyarov-V.Akobian, Gibraltar 2012) and ways to avoid it have been found, including 13...Nxd4 in
the Gelfand-Grischuk move order and other alternatives, such as Aagaard and Ntirlis' recommendation of an early
...h7-h6, not to mention the ninth move alternatives. So let's not throw our toys out of the pram just yet.
Looking at current representatives, it's true to say that the Tarrasch doesn't have too many regular GM advocates
(the patronage by Grischuk, T.L.Petrosian and Akobian appears to have cooled off recently). That said, there are
some players who have played it with success over the last few years. Occasional GM practitioners with excellent
results include Jacob Aagaard, Milos Perunovic, Ray Robson and Julio Sadorra, while Vlad Jianu, Lucian Miron,
Ivan Sokolov, Aaron Summerscale, Robin Swinkels and Branko Tadic have all played it several times with good
results. Alexander Berelowitsch, Goran Todorovic and the rapidly-improving Spanish GM David Larino Nieto are
probably the most consistent GM practitioners at the moment, with good to excellent results. Ekaterina
Kovalevskaya and Natalia Zhukova seem to play it all the time, again with good results. Admittedly, the Tarrasch
doesn't have a regular super-GM advocate at the moment and, accordingly, it is unlikely to become as popular as
when Spassky and Kasparov were beating everyone with it. But this is rather to our advantage, since playing an
opening at the height of fashion tends to lead to one's opponents being better prepared and more experienced in
opening at the height of fashion tends to lead to one's opponents being better prepared and more experienced in
the resulting structures.

Begging the question: What Happens When You Don't Define Your Terms
- suggested title for popular science book (London Review of Books, March 2011)

Let's see what we're talking about (View in Game Format):

1 d4
White occupies the centre. Classical theory would suggest that, if it were his turn to move again, 2 e4 would be his
choice.
1...d5
Black responds in kind, stopping e2-e4 and gaining a foothold in the centre.
2 c4
White chips away at the black central pawn. While everybody knows that this isn't a real gambit, after 2...dxc4 3 Nf3
Nf6 4 e3 White can be made to work a little to regain the pawn following 4...Be6!? or Romanov's favourite 4...b5.
2...e6
This is an almost perfect move (holding the centre and accelerating kingside development), with only one
drawback: namely that the c8-bishop is blocked inside its pawns. Of course, in the Tarrasch this drawback is
quickly resolved when pawns are traded on d5.
3 Nc3
The choice between this move and 3 Nf3 is a matter of taste and the overall repertoire of the White player. 3 Nc3 is
necessary for players who use the Exchange Variation against the Queen's Gambit Declined. 3 Nf3, on the other
hand, is indicated for Catalan fans.
Regarding the Tarrasch, it is fair to say that 3 Nf3 is slightly more flexible for White since he gains certain options
based on delaying the development of his queen's knight (whereas Nf3 is going to be a necessary move fairly early
in almost every variation). In addition, 3 Nc3 allows a pretty interesting gambit continuation which we'll see in a few
moves.
3...c5

The move which defines the Tarrasch. Black plays the move he wants to play (striking at the white centre, gaining
space, preparing a guilt-free development of the knight to c6 etc). The only drawback from a principled viewpoint is
that his d-pawn is going to become isolated...
4 cxd5
...hence this is the most principled continuation. White doesn't have to go for this immediately, or at all, and the
alternatives will be discussed in plenty of detail later in the book.
4...exd5
Taking the pawn back and keeping a pawn in the middle of the board. Moreover, the c8-bishop is looking at a clear
diagonal (which it actually gets to use, as may be compared with, for instance, many lines of the QGD Exchange
Variation).
While a lot of chess writers point out that the Stonewall Dutch is one of the very few openings where Black gains a
space advantage, this is actually not correct. First, Black doesn't technically have a space advantage in the
Stonewall (a white pawn is on the fourth rank, just like the d5- and f5-pawns, and White can play Ne5, just like
Black can play ...Ne4). Second, the best regular example of Black gaining a space advantage is the Tarrasch. If the
white d-pawn gets eliminated (through d4xc5 or ...c5xd4), Black will have the only pawn on the fourth rank and,
moreover, the only pawn in the centre. In some lines this pawn even makes it to d4 which, while by no means
always in Black's favour, certainly marks out even more central space. The lines where the white d-pawn survives
are less common, and the most notable is the 9 Bg5 c4 variation where central space is equal but Black gets a
queenside space advantage.
None of the above has any real bearing on the assessment of this position, or of the whole opening. However, what
I'm trying to point out is that Black's concept is an aggressive one. In effect, he's trying to play as White, albeit with
a few tempi less. If White plays passively, Black will be able to mobilize sufficiently to justify his central space and
the commitment inherent in an IQP. Whether Black can justify his concept in the face of accurate white play remains
to be seen. But for the moment let's put the pawn in the middle. As Rear Admiral Grace Hopper put it, it's easier to
ask forgiveness than it is to get permission.
4...cxd4!? is the gambit I mentioned a few moves ago.

Black gets compensation for sure. How much is the question...

5 Nf3
No real prizes for this one. Looking at the position I'm not even sure what other sensible moves exist, though I'm
sure that hasn't stopped people writing books and articles on them.
5...Nc6
A rule of thumb I tend to follow in the Tarrasch is that, given the option, I develop this knight before the king's
knight. One reason is because the queen's knight doesn't have a better square (as mentioned above, after ...c7-c5
there's no reason not to play ...Nc6), whereas the king's knight, in a small number of cases, might want to come to
e7 or to avoid an early pin by Bg5, as is well illustrated by 6 Bg5 Be7 7 Bxe7 Ngxe7.
6 g3

An invention of the great Akiba Rubinstein. The fact that his many victories from this position were achieved against
players displaying as much resistance as wet tissue paper doesn't change the assessment of this idea as ahead of
its time. Today it remains the critical test of the Tarrasch, for the simple reason that the d5-pawn is now in
considerable danger of just being taken and Black needs to find a way to avoid this happening without lapsing into
passivity.
6...Nf6
Knights before bishops. The Tarrasch is a classical opening and rewards players who put their pieces quickly on
good squares. At least this is my theory behind Jacob Aagaard's excellent successes with the line, when many
other titled players get caught up in more convoluted development schemes.
7 Bg2 Be7
The bishops go to their best squares. Obviously d6 isn't a great square since it interferes with the defence of the
d5-pawn and doesn't take any sting out of Bg5.
8 0-0 0-0
It's pretty rare for the kings to go elsewhere in the Tarrasch - maybe in some of the d4xc5 endgame lines, but
otherwise there's just too much happening on the c-file for either king to feel safe after long castling.
9 Bg5
From a principled standpoint, a perfect move. White develops his final minor piece to a good square and creates an
immediate threat of some combination of Bxf6 and d4xc5. Black has to respond somehow, by either 9...cxd4 (the
classical main line), 9...c4 (the trendy modern main line) or 9...Be6 (the discarded line). Notwithstanding what
modern theoreticians may suggest, having analysed these lines I think all three are perfectly playable - which,
incidentally, leads to a pretty high analytical overhead for White if he wants to play 9 Bg5, and may explain in part
his preference for other systems such as d4xc5 (at move 9 or earlier).
We'll get into specifics in a minute. In the meantime, I wish you best of luck with the Tarrasch!

Finally, I'd like to thank John Emms, Jonathan Tait and Byron Jacobs for their considerable assistance with this
book.

Sam Collins
Dublin,
October 2013

Back to Contents Page

 D34
Gelfand,B
Grischuk,A
Russian Team Championship 2004
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 c4 10.e5 e6 11.b3 a5 12.d2
ad8 13.bxc4 b4 14.xc6 bxc6 15.fc1
dxc4 16.xf6 gxf6 17.e3 c5 18.d5 xc3
19.xc3 xd5 20.xd5 xd5 21.c2 c8
22.xc4 d2 23.c1 xc2 24.1xc2 c6
25.g2 a6 26.a4 h6 27.f3 f8 28.e2
a5 29.b2 d6 30.b8+ e7 31.h4 b6
32.c8 d7 33.h8 b4 34.4xh6 b6
35.6h7 xa4 36.xf7+ c6 37.xf6+ b5
38.xb6+ axb6 39.h4 a2+ 40.f3 c4
41.h5 a7 42.c8 f7+ 43.e2 b4 44.g4
h7 45.f4 b5
1-0
Intro
[Sam Collins]
1.d4 White occupies the centre. Classical
theory would suggest that, if it were his turn to
move again, 2 e4 would be his choice. d5
Black responds in kind, stopping e2-e4 and
gaining a foothold in the centre. 2.c4
White chips away at the black central pawn.
[ While everybody knows that this isn't a real
gambit, after 2.c4 dxc4 3.f3 f6 4.e3
White can be made to work a little to regain
the pawn following e6!? ( or Romanov's
favourite 4...b5 . )]
2...e6 This is an almost perfect move (holding
the centre and accelerating kingside
development), with only one drawback:
namely that the c8-bishop is blocked inside its
pawns. Of course, in the Tarrasch this
drawback is quickly resolved when pawns are
traded on d5.
3.c3 The choice between this move and 3
Nf3 is a matter of taste and the overall
repertoire of the White player. 3 Nc3 is
necessary for players who use the Exchange
Variation against the Queen's Gambit
Declined.
[ 3.f3 , on the other hand, is indicated for
Catalan fans.
Regarding the Tarrasch, it is fair to say that
3 Nf3 is slightly more flexible for White since
he gains certain options based on delaying
the development of his queen's knight
(whereas Nf3 is going to be a necessary
move fairly early in almost every variation).
In addition, 3 Nc3 allows a pretty
interesting gambit continuation which we'll
see in a few moves. ]
3...c5 The move which defines the Tarrasch.
Black plays the move he wants to play
(striking at the white centre, gaining space,
preparing a guilt-free development of the
knight to c6 etc). The only drawback from a
principled viewpoint is that his d-pawn is
going to become isolated...
4.cxd5 ...hence this is the most principled
continuation. White doesn't have to go for this
immediately, or at all, and the alternatives will
be discussed in plenty of detail later in the
book.
exd5 Taking the pawn back and keeping a
pawn in the middle of the board. Moreover,
the c8-bishop is looking at a clear diagonal
(which it actually gets to use, as may be
compared with, for instance, many lines of the
QGD Exchange Variation).
While a lot of chess writers point out that the
Stonewall Dutch is one of the very few
openings where Black gains a space
advantage, this is actually not correct. First,
Black doesn't technically have a space
advantage in the Stonewall (a white pawn is
on the fourth rank, just like the d5- and f5-
pawns, and White can play Ne5, just like
Black can play ...Ne4). Second, the best
regular example of Black gaining a space
advantage is the Tarrasch. If the white d-
pawn gets eliminated (through d4xc5 or ...
c5xd4), Black will have the only pawn on the
fourth rank and, moreover, the only pawn in
the centre. In some lines this pawn even
makes it to d4 which, while by no means
always in Black's favour, certainly marks out
even more central space. The lines where the
white d-pawn survives are less common, and
the most notable is the 9 Bg5 c4 variation
where central space is equal but Black gets a
queenside space advantage.
None of the above has any real bearing on
the assessment of this position, or of the
whole opening. However, what I'm trying to
point out is that Black's concept is an
aggressive one. In effect, he's trying to play
as White, albeit with a few tempi less. If White
plays passively, Black will be able to mobilize
sufficiently to justify his central space and the
commitment inherent in an IQP. Whether
Black can justify his concept in the face of
accurate white play remains to be seen. But threat of some combination of Bxf6 and d4xc5.
for the moment let's put the pawn in the --
middle. As Rear Admiral Grace Hopper put it, [ Black has to respond somehow, by either
it's easier to ask forgiveness than it is to get 9...cxd4 (the classical main line) ]
permission. [ 9...c4 (the trendy modern main line) ]
[ 4...cxd4!? is the gambit I mentioned a few [ or 9...e6 (the discarded line).
moves ago. Black gets compensation for Notwithstanding what modern theoreticians
sure. How much is the question... ] may suggest, having analysed these lines I
5.f3 No real prizes for this one. Looking at think all three are perfectly playable – which,
the position I'm not even sure what other incidentally, leads to a pretty high
sensible moves exist, though I'm sure that analytical overhead for White if he wants to
hasn't stopped people writing books and play 9 Bg5, and may explain in part his
articles on them. preference for other systems such as d4xc5
c6 A rule of thumb I tend to follow in the (at move 9 or earlier).
Tarrasch is that, given the option, I develop We'll get into specifics in a minute. In the
this knight before the king's knight. One meantime, I wish you best of luck with the
reason is because the queen's knight doesn't Tarrasch! ]
have a better square (as mentioned above,
after ...c7-c5 there's no reason not to play ...
Nc6),
[ whereas the king's knight, in a small
number of cases, might want to come to e7
or to avoid an early pin by Bg5, as is well
illustrated by 5...c6 6.g5 e7 7.xe7
gxe7 . ]
6.g3 An invention of the great Akiba
Rubinstein. The fact that his many victories
from this position were achieved against
players displaying as much resistance as wet
tissue paper doesn't change the assessment
of this idea as ahead of its time. Today it
remains the critical test of the Tarrasch, for
the simple reason that the d5-pawn is now in
considerable danger of just being taken and
Black needs to find a way to avoid this
happening without lapsing into passivity.
f6 Knights before bishops. The Tarrasch is
a classical opening and rewards players who
put their pieces quickly on good squares. At
least this is my theory behind Jacob
Aagaard's excellent successes with the line,
when many other titled players get caught up
in more convoluted development schemes.
7.g2 e7 The bishops go to their best
squares. Obviously d6 isn't a great square
since it interferes with the defence of the d5-
pawn and doesn't take any sting out of Bg5.
8.0-0 0-0 It's pretty rare for the kings to go
elsewhere in the Tarrasch – maybe in some of
the d4xc5 endgame lines, but otherwise
there's just too much happening on the c-file
for either king to feel safe after long castling.
9.g5 From a principled standpoint, a perfect
move. White develops his final minor piece to
a good square and creates an immediate
 Back to Contents Page
Structural Introduction

Structure No. 1
Structure No. 2
Structure No. 3
Structure No. 4
Structure No. 5
Structure No. 6
Isolated Queen's Pawn

The main characteristic of an opening is the structures to which it leads, and for no opening is this more true than
the Tarrasch. The perceived (and actual) inferiority of Black's structure compared to White's is the only reason for
not playing an opening with such great development and, ordinarily, a central space advantage.
In this section I introduce some of the main structures to which the Tarrasch leads. This is not a comprehensive
overview. Nor do I deal with the intricacies of each structure - entire books have been devoted to the isolated
queen's pawn alone.
For those learning the Tarrasch, there are two great additional resources. First, look up games in other openings
which feature the IQP. The ideas are applicable in all IQP positions regardless of the initial sequence. Indeed, in
some of the positions in the Symmetrical Tarrasch, after poor play by White we end up in known variations where
White has an IQP, with colours reversed.
Second, study specialist resources on the IQP and related structures. A real classic in this field isWinning Pawn
Structures by Alex Baburin, which focuses on the IQP and the related cases of hanging pawns and the isolated
pawn couple. In addition, Winning Chess Middlegames by Ivan Sokolov is an absolute must for advanced players -
while it deals with other pawn structures too, Sokolov's knowledge and handling of IQP positions is exemplary.
Indeed, some of his games are featured in this text, with both colours.

Structure No. 1

This is the basic Tarrasch structure, most commonly arising from the traditional main line 9 Bg5 cxd4, as well as
some of the 9 dxc5 lines. White has fianchettoed his king's bishop, which puts immediate pressure on the d5-pawn
(ordinarily this pawn is securely blockaded and cannot advance).
One of the things I realized when researching this book is how even experienced Tarrasch practitioners don't
always realize just how vulnerable the d5-pawn is, especially at an early stage. White's c3-knight and g2-bishop
already target the IQP, and the pressure can be increased with Qb3 (or Qa4) and Rd1. This, when combined with
the fact that the black queen is normally on the d-file (d8 or d7), frequently leads to central shots where White takes
on d5.
By far the most common way in which Black can (and does) endanger his IQP is with ...Bg4 and, after h2-h3,
retreating the bishop to h5. The bishop then seems active, targeting the e2-pawn, and is immune from exchange.
My conclusion (after a lot of work) is that, ordinarily, the bishop is just misplaced on h5 in this structure. The
pressure on the e2-pawn is of minor significance (White's knight on c3 protects e2 and, if necessary, e2-e3 can be
played), but of more significance is that the bishop cannot trade its counterpart on g2 and, importantly, cannot drop
back to e6 to protect the pawn. Often a sequence involving g3-g4 and Nxd5 can clip a pawn, even following a
temporary piece sacrifice. Of course, if White plays g3-g4 there may be some compensation in the form of
weakened squares around White's king. However, just be sure, when putting the bishop on h5, that you can hold
d5, or at least extract sufficient concessions to form valuable compensation. If not - and in the majority of cases,
Black doesn't get enough if d5 falls - retreat the bishop to f5 (if available) or e6.
The position is rich and normally the structure is changed into one of those discussed in the next few pages.
The position is rich and normally the structure is changed into one of those discussed in the next few pages.
If White keeps the structure as it is (and the choice to change the structure is entirely White's, since Black can't
force piece trades on c6 or e6), then develop your pieces in the middle and see what happens. The knights will be
well placed on c6 and f6 - the f6-knight normally needs to stay put to hold d5, while the c6-knight can sometimes
seek activity by ...Na5-c4 or ...Nb4 followed by ...a7-a5. The light-squared bishop is well placed anywhere on the
c3-h8 diagonal and, if a trade of light-squared bishops can be arranged, this is normally in Black's favour. Black's
dark-squared bishop has a less clearly defined role in the early stages, and often is simply tucked out of the way
after ...Re8 and ...Bf8. Black's rooks are well placed on any of the c-, d- and e-files - rooks on c8 and e8 seem the
most common. The black queen is a difficult piece to handle, and finding a good square for her often signals the
solution to Black's opening problems; ...Qd7 followed by ...Bh3 is one common way, and sometimes ...Qb6 is seen
instead to put pressure on the queenside.

Structure No. 2

This is the most significant transformation of the basic Tarrasch structure, which occurs when White trades knights
on c6. Of course the change brings certain benefits to Black - the d5-pawn is protected and the half-open b-file is
available for a major piece.
When White goes into this structure he always has a lot of control over d4 and c5. Accordingly, while Black could,
in theory, go into a hanging pawns position with ...c6-c5, in the Tarrasch this happens very rarely.
White's plan is based on blockading the c- and d-pawns and, ultimately, winning one of Black's queenside
weaknesses. Note, however, that c6 is more difficult to attack than d5 (for instance, the bishop on g2 doesn't hit the
pawn, so White is normally limited to a frontal attack on the half open c-file; in addition, a white knight or bishop
frequently occupies the c5-outpost and, while such a piece is well placed, it shields the pawn from attack).
Therefore Black has time to try to build up his counterplay and the risk of an early disaster is reduced, although
Black needs to take care to avoid falling into passivity. Normal methods include seeking counterplay on the
half-open b- and e-files, while exchanging White's light-squared bishop with ...Bh3 is always a good plan, with
potential for a kingside attack.

Structure No. 3

This is a structure which commonly arises when Black plays ...Be6, White plays Nxe6 and Black recaptures with his
f-pawn.
At first glance it seems that White must have a significant advantage here, with two bishops and a superior pawn
structure. In fact Black gets major trumps of his own. First, his pawn structure is improved - the d5-pawn is now
solidly defended and his rook will be active on the half-open f-file. Black's central control is now extremely strong,
with more space and a 2-1 central majority. In addition, the pawns on d5 and e6 have a limiting effect on the white
bishop on g2, which is no longer attacking an isolated pawn on d5. This is a similar concept to that in the Stonewall
Dutch, where a bishop on g2 is restricted by the opponent's light-squared central pawn chain. Incidentally, this also
informs why an ...e6-e5 advance is extremely risky, since the g2-bishop would come powerfully into play.
Note that, by contrast to the previous structure, it is very difficult for White to attack the e6-pawn, which is not on a
half-open file.
It is very important for the assessment of these structures that White cannot favourably make an e2-e4 break. In
order for Black to be okay, White either should not be able to play e2-e4, or Black should be able to meet with
e2-e4 with ...d5-d4, keeping the position closed. By contrast, ...d5xe4 would be a strategic disaster, opening lines
for the white bishop on g2 and exposing the weak pawn on e6.
I can't be sure, but I suspect that a reluctance to go into this structure is the main reason that many players retreat
their bishop to h5 after White plays h2-h3 in the main lines after 9 Bg5 cxd4. As noted in my discussion above, I
really don't like putting the bishop on h5 in most positions, but you can only avoid it if you are happy to let White
capture on e6.
This structure was defended with great success by Kasparov in his Candidates matches. He had no problems in
game two of his 1983 match with Korchnoi, or game two of his 1984 match with Smyslov, and won a great game in
this structure in game twelve against Smyslov.
The following comment of Sokolov, which will arise in the annotated games, is worth noting here: "It is good to note
that per definition Black is never worried about 13 Nxe6 fxe6, since the loss of the bishop pair is fully compensated
for by the improvement of his pawn structure". I agree with this assessment and believe that a key part of handling
the Tarrasch well is being ready to go for this structure as Black.

Structure No. 4

QUESTION: Didn't we look at this already?

ANSWER: Not really. The structure we examined previously is the main Tarrasch position with a black pawn on d5,
as normally arises from the 9 Bg5 cxd4 variations. In those positions White has far too firm a grip on d4 for Black to
be able to play ...d5-d4 himself. The structure currently on the board is characteristic of a number of lines, including
most of the major continuations after 9 dxc5. Advancing the pawn to d4 affects the plans for both sides in several
profound ways. On d4 the pawn is safer from attack than it was on d5 - for instance, the g2-bishop isn't hitting the
pawn. The fact that Black has seized so much central space means that it can be awkward for White to arrange an
immediate attack on the pawn. Rather he will rely on other advantages - the g2-bishop's range of activity is
extended, hitting the b7-pawn, and a knight on d3 would be perfectly placed.
While the e2-pawn is rarely a real weakness in the 9 Bg5 cxd4 lines, here it is fixed on e2 (playing e2-e3 will lead,
at best, to an exchange of pawns) and its normal defender, the c3-knight, has been driven away from its natural
defensive square.
The plans for both sides are markedly different than with a black pawn on d5, so be sure not to play on autopilot.

Structure No. 5

This structure is characteristic of the Symmetrical Tarrasch. The major difference compared to lines we have
looked at so far is that White has not gone for the more aggressive option of fianchettoing his king's bishop, but
looked at so far is that White has not gone for the more aggressive option of fianchettoing his king's bishop, but
instead puts his bishop either on e2 or on b5 (intending to take a knight on c6).
Black is normally more comfortable in these lines than with a bishop on g2. The d5-pawn is under less immediate
pressure and, if White makes a couple of inferior moves, it is quite common to transpose (with reversed colours)
into other openings in which White gets a favourable IQP position. The pawn still being on g2 means that Black can
consider more direct kingside attacking methods, such as forming a queen and bishop battery on the h2-b8
diagonal. Indeed, the ideal IQP formation - pawn on a6, knights on c6 and f6, bishops on c7 and h3 (after White has
been forced to play g2-g3), queen on d6, rooks on d8 and e8 - is established by Black in a number of the examples
we'll examine in the chapter on the Symmetrical Tarrasch. If Black gets these moves in, White will be the one who
needs to equalize.
One effect of the pawn on e3 is that White's dark-squared bishop can't easily develop to g5, and instead tends to
go to b2 (after b2-b3 or a2-a3 and b2-b4). While the bishop is well placed here, it is at risk of being irrelevant if
Black succeeds in creating a strong kingside attack. In general White has fewer kingside defenders than in the
fianchetto lines, and Black must seek to exploit this.

Structure No. 6

This structure is characteristic of the 9 Bg5 c4 line as well as the Swedish Variation (featuring an early ...c5-c4,
...Bb4 and ...Nge7). Black gains additional queenside space and prepares, if allowed, a massive advance with
...b7-b5-b4. In addition, the d5-pawn isn't as weak as it would be if it sat on the end of a half-open d-file.
As against this, playing ...c5-c4 takes all pressure off the white centre. Thus e2-e4 becomes an idea, as does Ne5
and, if White wishes, Nxc6. We will also see examples where White seizes more kingside space with f2-f4.
The positions in this structure tend to become quite sharp. White's most reliable method is to seek to destroy the
black pawn chain by Ne5xc6 and b2-b3. Black, for his part after ...b7xc6, tries to arrange counterplay with ...c6-c5.
The game is unusual and unbalanced and I think that becoming closely acquainted with this structure significantly
enhances a Tarrasch player's level, especially since a number of quieter lines (for instance, after an early e2-e3 by
White) can also be met by ...c5-c4.

Isolated Queen's Pawn

One of the benefits of studying IQP positions is that they arise in such diverse contexts. As Ivan Sokolov observes:
"Isolated pawn structures are arguably the structures that arise from the most different openings (Tarrasch Defence
and Semi-Tarrasch, Queen's Gambit Accepted, Queen's Gambit Declined, Nimzo-Indian Defence, Meran Variation,
Ragozin Variation, Petroff Defence, etc) and are therefore very important positions to understand, regardless of the
opening preferences a player may have."
There are many examples of players who have managed to exploit transpositions and structural similarities to play
a wide range of openings confidently. For instance, a number of GMs are happy to defend an IQP with Black after 1
e4 e6 2 d4 d5 3 Nd2 c5 4 exd5 exd5, a line developed by Korchnoi and Mikhail Gurevich amongst others. Bartosz
Socko also feels comfortable in this line with Black and, in a game against Gawain Jones in the German Bundesliga
2012, he answered 1 e4 c5 2 c3 with 2...e6 3 d4 d5 4 exd5 exd5, obtaining a very similar position and making a
comfortable draw.
In the Politiken Cup 2013, I was interested to see Tarrasch enthusiast IM Jonathan Carlstedt obtaining positions
similar to his favourite opening in two consecutive rounds. Here he is in round six:

J.Carlstedt-J.Brorsen, Helsingor 2013

In round seven, the following happened:

J.Glud-J.Carlstedt, Helsingor 2013

I hope that these examples have demonstrated how universal Tarrasch-type positions are, and how much is to be
I hope that these examples have demonstrated how universal Tarrasch-type positions are, and how much is to be
gained by learning to handle them well.

Back to Contents Page

 D94 chances before the game ended in a draw at
Carlstedt,J move 75.
Brorsen,J
Helsingor 2013
[Sam Collins]

1.c4 c5 2.e3 f6 3.d4 g6 4.c3 g7

5.f3 0-0 6.e2 d5 7.0-0 Carlstedt has
used this set-up in several games as White,
scoring heavily.
[ After, for instance, 7.0-0 cxd4 8.exd4 c6
we have a Tarrasch with colours reversed
and an extra move for White. While I don't
believe this line is a serious attempt at an
opening advantage, Carlstedt is able to draw
on his experience with the Tarrasch and
plays such positions well. ]
The game continued:
7...dxc4 8.xc4 cxd4 9.exd4 g4 10.d5
bd7 11.e1 b6 12.b3 c8 13.h3 xf3
14.xf3 . Black hasn't handled the opening in
the most principled way and White has a
comfortable advantage, going on to win in 39
moves.

A20
Glud,J
Carlstedt,J
Helsingor 2013
[Sam Collins]

1.c4 e5
[ As we will see, the Tarrasch is sufficiently
universal that it can be used against the
English: 1...e6 followed by 2...d5 and ...c7-
c5. Carlstedt has played this way as Black
a number of times, but decides to surprise
his opponent (an extremely strong IM from
Denmark who made a GM norm in this
tournament). ]
2.g3 f6 3.g2 c6 4.d4 exd4 5.xd4 d5
6.f3 e7 7.0-0 0-0 8.cxd5 cxd5 9.c3
c6 10.a4 It's easy to see the similarities of
this position to the d4xc5 lines in the Tarrasch;
[ indeed, 10.d1?! c5?! would be a direct
transposition. Carlstedt has a lot of
experience in this structure and effortlessly
obtains a good game. ]
10...b6 11.b5 c5 12.xb6 axb6
13.f4 h6 14.a3 g5 15.d2 d4 16.b5
e4 17.b4 e7 18.ad1 d8 . Black is very
comfortable here. He went on to sacrifice a
pawn and actually obtained decent winning
 Back to Contents Page
Chapter One

9 Bg5 cxd4
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6

In this chapter we discuss the classical main line of the Tarrasch. I have to admit that Black's position looks a little
loose, and he needs to play accurately in order to keep the d5-pawn alive without making too many concessions.
During a conversation with an elite player, he informed me that Black's position looked "stupid" to him. Certainly
when we compare the position to other IQP positions (especially those obtained by White), here Black seems
several tempi down on his standard set-up. Moreover, White has completed development and managed to
fianchetto his king's bishop, which is the most aggressive anti-IQP approach (trying, as Larsen suggested, not to
blockade the pawn but to capture it). I think Black is okay here, but he needs to be accurate over the next 5-10
moves to avoid falling into a passive and prospectless position.
11 Be3
By far the main continuation, which takes up the bulk of the chapter. White reinforces his control over d4. The
bishop did its job on g5 by forcing Black to release the central tension.
11 Bf4 is an annoying sideline in which Black doesn't win many games but does lose a few. I've covered two
approaches for Black: the sound (Li Shilong-Braun) and the enterprising (Socko-Petrosian). These games also
cover 11 Bxf6 which Black should be happy to see appear on the board.
11...Re8
This was Kasparov's choice and accordingly is the only move most players are aware of in this position.
11...Bg4 is a great alternative which has scored 75% in world championship matches. Recent analysis (as
demonstrated in practice) by Sokolov has shown that this doesn't deserve to be a sideline. SeeMovsesian-Sokolov
for the details.
12 Rc1

The modern main line. White gets another piece into the game and prepares to strangle Black on the dark-squares
after Nxc6 and Na4 in some order.
12 Qb3 gets a game of its own for two reasons. First, this was Karpov's weapon against Kasparov in their first
world championship match, and the 2/2 he achieved basically buried the Tarrasch for a generation. Accordingly
Black should make a point of being well prepared here.
Second, this is the recommendation in Lars Schandorff's recent White repertoire series, so Tarrasch players can
anticipate to be tested here. 12 Qb3 is actually a line in which Black has a great new(ish) idea, as demonstrated in
Babula-Orsag.
12 Qa4 and White's 12th move alternatives are serious moves which have been played by serious players. Bu
Xiangzhi is one of them (see Bu Xiangzhi-Li Shilong).
12...Bf8
Completing the regrouping commenced with 11...Re8. This can't be a bad move but it's fair to say that White's
responses have been well mapped out. Sokolov-Petrosian is a nice demonstration of White's chances.
12...Bg4, just like on move 11, is a good alternative. See Kasimdzhanov-Berelowitsch.

Key Notes

1. 9 Bg5 cxd4 leads to some of the strategically richest middlegames in the Tarrasch. Black needs a lot of
knowledge to handle these positions well and avoid falling into passivity. In the main line after 12...Re8, White has
a very wide choice, with several continuations which have been endorsed by leading GMs. Black's treatment differs
for each of them.
2. In the lines with 12 Rc1 or 12 Nxc6, White tends to define the structure at an early stage. He will aim to get a
knight to c5 and solidly blockade Black's c- and d-pawns. Black should seek counterplay on the half-open b- and
e-files.
3. Early queen moves by White allow Black to create active counterplay, for instance 12 Qb3 Na5 or 12 Qa4 Bd7
13 Rad1 Nb4! 14 Qb3 a5!. I have not seen a convincing route to an advantage for White in these lines.
4. An underrated idea is ...Bg4, either on move 11 or after 11...Re8 12 Rc1. Black has been doing well here of late.
5. After 11 Bf4, Black needs to adjust his plans. He has a choice between playing a pretty balanced endgame, as in
Li Shilong-Braun, or fishing in troubled waters, as in Socko-Petrosian.

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 cxd4 10 Nxd4 h6 11 Be3
11 Bf4 Bg4 12 h3
12...Be6 – Li Shilong-Braun
12...Bh5 – Socko-Petrosian
11...Re8
11...Bg4 – Movsesian-Sokolov
12 Rc1
12 Qb3 – Babula-Orsag
12 Qa4 – Bu Xiangzhi-Li Shilong
12...Bf8 – Sokolov-Petrosian
12...Bg4 – Kasimdzhanov-Berelowitsch

Back to Contents Page

9 Bg5 cxd4
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.g5 0-0
9.0-0 cxd4 10.xd4 h6 In this chapter we
discuss the classical main line of the Tarrasch.
I have to admit that Black's position looks a
little loose, and he needs to play accurately in
order to keep the d5-pawn alive without
making too many concessions. During a
conversation with an elite player, he informed
me that Black's position looked "stupid" to him.
Certainly when we compare the position to
other IQP positions (especially those obtained
by White), here Black seems several tempi
down on his standard set-up. Moreover, White
has completed development and managed to
fianchetto his king's bishop, which is the most
aggressive anti-IQP approach (trying, as
Larsen suggested, not to blockade the pawn
but to capture it). I think Black is okay here,
but he needs to be accurate over the next
5-10 moves to avoid falling into a passive and
prospectless position.
11.e3 By far the main continuation, which
takes up the bulk of the chapter. White
reinforces his control over d4. The bishop did
its job on g5 by forcing Black to release the
central tension.
[ 11.f4 is an annoying sideline in which
Black doesn't win many games but does
lose a few. I've covered two approaches for
Black: the sound (Li Shilong-Braun, Game
6) and the enterprising (Socko-Petrosian,
Game 7). ]
[ These games also cover 11.xf6
which Black should be happy to see appear
on the board. ]
11...e8 This was Kasparov's choice and
accordingly is the only move most players are
aware of in this position.
[ 11...g4 is a great alternative which has
scored 75% in world championship matches.
Recent analysis (as demonstrated in
practice) by Sokolov has shown that this
doesn't deserve to be a sideline. See
Movsesian-Sokolov (Game 5) for the
details. ]
12.c1 The modern main line. White gets
another piece into the game and prepares to
strangle Black on the dark-squares after Nxc6
and Na4 in some order.
[ 12.b3 gets a game of its own for two
reasons. First, this was Karpov's weapon
against Kasparov in their first world
championship match, and the 2/2 he
achieved basically buried the Tarrasch for
a generation. Accordingly Black should
make a point of being well prepared here.
Second, this is the recommendation in
Lars Schandorff's recent White repertoire
series, so Tarrasch players can anticipate to
be tested here. 12 Qb3 is actually a line in
which Black has a great new(ish) idea, as
demonstrated in Babula-Orsag (Game 3). ]
[ 12.a4 and White's 12th move alternatives
are serious moves which have been played
by serious players. Bu Xiangzhi is one of
them (see Game 4). ]
12...f8 Completing the regrouping
commenced with 11...Re8. This can't be a bad
move but it's fair to say that White's
responses have been well mapped out.
Sokolov-Petrosian (Game 1) is a nice
demonstration of White's chances.
[ 12...g4 , just like on move 11, is a good
alternative. See Kasimdzhanov-Berelowitsch
(Game 2). ]
D34
Sokolov,I
Petrosian,TL
1: European Cup, Kemer 2007
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 cxd4 10.xd4 h6 11.e3 e8
The traditional main line, which has fallen out
of favour in recent years as Black has
experimented with different approaches, in
particular 9...c4. John Nunn makes the
following comment: "Technically, we have a
standard type of IQP position in which Black
will look for kingside counterplay by
occupying the half-open e-file and perhaps
the outpost at e4. White for his part has a
choice of two plans: he may continue to
blockade d4 and build up pressure on the
weak pawn, or he may exchange knights on
c6 and transfer his attack to the backward c-
pawn and the blockade point c5."
12.c1 This is by far the most popular move
and the main line. White develops his rook
and takes aim at the knight on c6. Right now
the rook occupies an open file and, in the
event that White takes on c6, the rook will be
perfectly placed exerting pressure on the c6-
pawn.
[ The immediate 12.xc6 is rarely played
since it gives Black more information – now
he knows that he will be playing with an
Isolated Pawn Couple rather than an IQP
and can arrange his pieces accordingly. An
interesting example from one of the great
Tarrasch experts of the 80s: bxc6 13.d4
h7!? (aiming for g5 from where the knight
supports ...Bh3 and can drop back to e6,
which is a great square for a knight in this
structure) 14.c1 g5 15.a4 e6!?
( simply 15...d7 looks equal ) 16.xc6 d7
17.c1 xd4 18.xd4 f6 19.d1
( Black is not worse in the endgame after
19.xd5 xa4 20.xa8 xa8 21.xa8
xa8 ) 19...b8 20.b3 xa4 21.bxa4 d4
and, while the computers love White here,
his extra doubled a-pawn isn't felt in this
opposite-coloured bishop position and a
draw was soon agreed in Y.Razuvaev-J.
Nunn, London Lloyds Bank 1983 (½-½ in
29). ]
12...f8 13.xc6 This move is the most
popular, though perhaps the trend has
recently shifted to 13 Na4. In some respects
13 Nxc6 is open to the same criticism as 12
Nxc6, namely that it shows White's hand too
early.
[ 13.a4 is a variation on the same theme –
White is still likely to capture on c6, but
brings his knight to c5 directly, which can
create additional problems by hitting the b7-
pawn. Black has a wide choice:
A) 13...d7 14.c5 and then:
A1) 14...a5 15.b3 c8 16.xd7
xd7 17.d3 c6 18.xc6 bxc6
19.fd1 b7 20.c5 xc5 21.xc5
e7 22.c2 a3 23.e3 led to a classic
win for White in V.Kramnik-M.Illescas
Cordoba, Linares 1994 (1-0 in 38). Again
Sokolov's comments are instructive:
"Black is facing a very difficult defence.
Due to his better pawn structure, White
has a lasting advantage, while Black
does not have dynamic activity to
nearly compensate for the positional
problems related to his weak pawns. It
is important to note that transitions from
an isolated pawn to two weak hanging
pawns are a very common way to
combat the isolated pawn. Contrary to
parallel hanging pawns in the centre,
such weak connected pawns are by
definition not mobile, so there are no
tactics that can be based on this mobility.
";
A2) 14...xd4 15.xd4 ( 15.xd4 c6
16.d3 d6 17.c2 c7 18.h4
led to a typical exchange sacrifice in B.
Galstian-T.L.Petrosian, Stepanakert
2005: xe3! 19.fxe3 and now b6
would have left Black with full
compensation ) 15...c6 16.d3 d6
17.e3 e4 18.a3 e7 19.e1 a6
20.h4 ad8 and Black had a solid
position in R.Kasimdzhanov-S.Iuldachev,
Tashkent 2009 (½-½ in 51).;
A3) 14...xc5 15.xc5 e7
is interesting; for instance, 16.c1 e5!
17.d1 h5 with some attacking
chances in Se.Ivanov-A.Lugovoi, St.
Petersburg 2005 (½-½ in 44).;
A4) 14...--;
B) 13...a5 .
EXERCISE: Can White win a pawn by
taking on c6?
ANSWER: Yes he can, and this is a key
point of the entire line: A. Yes he can, and
this is a key point of the entire line.
14.xc6 bxc6 15.xc6! (it has been
established that this is basically
unplayable for Black, so 13...Qa5 is
dubious) d7 16.d2 b4 ( 16...d8
17.c1 xa4 18.xa4 xe2 19.c3
e4 20.d4 g5 21.f3 1-0 was a rapid
defeat for a strong player in A.Lupor-S.
Galdunts, Bad Wörishofen 2008; while
16...b5 is met by 17.xf6! gxf6 18.c3
xb2 19.xd5 with extremely strong
compensation for White ) 17.c5 xa4
18.xa4 xa4 19.xb4 xe2 20.a5
with a very difficult endgame for Black, as
in A.Greenfeld-M.Illescas Cordoba,
Thessaloniki Olympiad 1988 (½-½ in 42).;
C) . Some recent games have suggested
that Black can take advantage of White's
omission of 13 Nxc6 with 13...e5!?
14.b5 and now:
C1) 14...c4 15.d4 g4 16.ac3
e5 17.xa7
C1a) 17...c6?! 18.xc6? ( however,
18.xf6! followed by Nxd5 was a more
critical test ) 18...bxc6 19.h3 f5
20.a3 d7 21.f4 c5 was agreed
drawn in M.Sharif-V.Akobian, Abu
Dhabi 2000 – Black has full
compensation for the pawn, with
pressure on the e-pawn and Benko-
style play on the queenside; in addition,
 his central control is excellent, like in
a Blumenfeld gambit;
C1b) 17...d7 18.ab5 h3 19.b3
ac8 . Here White went on to lose in D.
Anderton-O.Rodriguez Vargas,
European Senior Team Championship,
Dresden 2008 (0-1 in 41), but he
would have been winning after
20.xd5 .;
C2) 14...c6 15.c5 ( 15.f4!
is slightly better for White ) 15...g4
16.e1 xc5 17.xc5 b6 18.d6
e7 was level in T.Nyback-J.Ehlvest,
Harjumaa (rapid) 2010 (½-½ in 34). The
jury's still out on this 13...Ne5 idea.;
C3) 14...--;
D) 13...-- ]
13...bxc6 14.a4 d7 15.c5 xc5
16.xc5 QUESTION: What's your
assessment of this position?
ANSWER: It seems to me that White has an
edge. I found Ivan Sokolov's comments on
this position to be extremely instructive:
"White's strategy is rather clear. The dark-
squared bishops have been exchanged, the
knight on c5 is a strong, dominant piece and
the black pawn on c6 is a newly created
weakness. Black has to create counterplay in
order to achieve a dynamic balance. However,
without the dark-squared bishops and with the
white knight on c5, counterplay is not easy to
find. It should also be mentioned that almost
all further piece exchanges favour White."
g4 17.e1 a5 By far the most popular
move here, activating the queen and targeting
a2.
[ Other moves such as 17...b6 have been
tried, but I think the text is best. ]
18.h3 Now Black faces a difficult decision of
where to retreat the bishop. h5 Keeping an
eye on the e2-pawn, but allowing the
manoeuvre which White executes in the game.
[ 18...f5 has been tested more often:
A) 19.d4 ab8 20.a3
A1) an attempted improvement over
20...b5 21.b3 e4 ( Stohl suggests
21...a5! ) 22.b4 a5 23.xe4 xe4
24.c5 b6 25.bxa5 and White went
on to win in G.Kasparov-M.Illescas
Cordoba, Linares 1994;
A2) 20...b6 21.b3 ( 21.f4 e4
22.b3 bd8 23.b4 e7 24.d2 xg2
was agreed drawn in A.Bratchenko-A.
Bezgodov, St. Petersburg 2000 )
21...e4 22.xe4 dxe4 23.h2 e3!?
24.fxe3 bd8 25.f4 g6 with a fine
position for Black in M.Dziuba-V.Genba,
Pardubice 2008 (0-1 in 32).;
B) 19.a4 d2 20.f4 xf4 21.gxf4
ab8 22.b3 is an ending which has been
defended by Akobian a couple of times.
Then b5 ( or 22...g5!? 23.fxg5 hxg5
24.c3 g7 25.f1 h5 26.e4 g6
27.exd5 xe1+ 28.xe1 f4 29.f1
cxd5 and, if anyone, Black was better in V.
Filippov-V.Akobian, Las Vegas 2002; while
22...b4 23.e3 g5 24.a3 b6 25.fxg5
hxg5 was level in T.Enhbat-V.Akobian,
Connecticut 2004 ) 23.e3 g5 24.a3
was Al.David-F.Manca, Cento 2011, and
now I like a5! , keeping a target on b3.;
C) 19.-- ]
19.d3 g6
[ Sokolov suggests 19...xa2 20.xc6 g6
21.b4 ( 21.f4 e4! ) 21...xb2 22.xd5
xd5 23.xd5 h7 24.e4 ac8
"and a draw should be the most likely
outcome". ]
20.f4 e4
[ Again Sokolov suggests the dynamic
20...xa2 21.xg6 fxg6 22.xc6 xb2
, as after 23.xd5+ xd5 24.xd5+ h8
25.xg6 a5! , the a-pawn should give Black
enough compensation to draw. ]
21.xc6 xg2 22.xg2 d4 23.c1 d5
[ 23...xa2 24.c4! is similar to the next
note, since Sokolov points out that xb2??
fails to 25.c7 f8 26.g6 . ]
24.d3 b4
[ Sokolov gives 24...xa2 25.c4
with advantage for White in view of his
superior pawn structure. ]
25.xb4 xb4 26.c4 b7+ 27.g1 d7
28.f4 ad8 29.d1! xh3 30.cxd4
xd4 31.xd4 a6
[ Not 31...xe2?? 32.d8+ h7 33.d3+
, winning a rook. ]
32.e3 White has a sound extra pawn. A
consideration of the endgame would take us
too far afield, but we can conclude that the
opening and middlegame have been a
success for White.
c8 33.d7 xd7 34.xd7 c2 35.b7
g5 36.a4 a5 37.b3 g4 38.b5 b2 39.f1
g7 40.xa5 xb3 41.f5 b2 42.e1
a2 43.f4 h5 44.d1 g6 45.c1 f5
46.b1 d2 47.a5 d8 48.c2 a8
49.a4 g5 50.d3 h4 51.gxh4+ xh4
52.a6 h3 53.a7 g2 54.a2 f3 55.d4
f4 56.exf4 xf4 57.a3
1-0
D34
Kasimdzhanov,R
Berelowitsch,A
2: German League 2011
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.c3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 cxd4 10.xd4 h6 11.e3 e8
12.c1 g4 While this move has not been
historically as popular as 12...Bf8, it seems to
be the recent trend. It is notable that the
move also has Kasparov's endorsement, who
writes: "Apparently 12...Bg4 13 h3 Be6 is
better than 12...Bf8, as, for example, in the
simultaneous game Züger-Kasparov
(Switzerland 1987) and in many later games."
13.h3 e6 QUESTION: Can't White just win
the bishop pair by taking on e6?
ANSWER: Yes he can, but this trade gives
Black certain advantages too. Black's
structure is improved – his d5-pawn is
strongly supported and he has the half-open f-
file which can be used by his rook. Black's
central control is excellent. White would love
to play e2-e4 to crack open the position for
his bishops but this isn't so easy to achieve.
We'll see later in the notes to Bu Xiangzhi-Li
Shilong (Game 4) that Garry Kasparov
successfully defended this structure as Black
in his Candidates matches against Smyslov
and Korchnoi.
14.xc6 Scherbakov notes: "This approach
can hardly pose problems for Black."
Nevertheless, 14 Nxc6 is the most popular
move in the position.
[ The immediate 14.xe6 fxe6 15.f4
is an ambitious approach, playing for a
central bind.
However, pushing the f-pawn in this fashion
means that Black will always have
counterplay; for instance, h8 ( or 15...a5
16.d2 ad8 17.a4 b4 18.xa5 xa5
19.fd1 xc3 20.bxc3 e5 21.c2
and a draw was agreed in F.Levin-A.
Berelowitsch, Düsseldorf 2011;
Stohl suggests that 15...b4!? 16.f2 e7
"followed by ...Rad8 leads to an unclear
position, as the scope of White's bishops
is limited" ) 16.f2 b4 17.b3 e7 18.a3
c5 19.xc5 xc5+ 20.h2 e3
with good play for Black in Z.Rahman-S.
Iuldachev, Asian Championship, Doha
2003. ]
[ 14.a4 d7 has also been extensively
tested:
A) 15.h2 a6 and now:
A1) 16.fd1 b5 17.c2 ac8
has been well defended by Jeroen
Bosch: 18.b1 ( 18.xe6 fxe6 19.b1
d6 and, if anyone, Black was better in J.
Timman-J.Bosch, Dutch Championship,
Hilversum 2006 ) 18...f8 ( but White's
exchanging operation could have been
anticipated with the more active
18...d6! , when Black is fine ) 19.xe6
fxe6 20.e4! was slightly better for
White in C.Matamoros Franco-J.Bosch,
Deizisau 2010 (½-½ in 82).;
A2) 16.xc6 bxc6 17.fd1 ab8
18.d2 f5 19.a3 e6 20.f4 bc8
and Black had no problems in J.
Djordjevski-G.M.Todorovic, Skopje 2011.
The game is a good example of how
Black's kingside build-up can become
threatening if White is not accurate:
21.g4?! g6 22.g3 h5! 23.f3?!
(White is trying to hold his kingside
structure, needlessly weakened by g3-
g4?!, but is creating a lot of dark-square
weaknesses in the process) d7
24.e1? b6! 25.xa6 c4 26.d4
xb2 ( 26...h4 27.f4 a8 28.b7
c5 29.dd1 d6 is even stronger,
when the white queen is trapped ) 27.e4
h4 28.f4 g5 29.e5 xf4+
and White resigned.;
A3) 16.--;
B) . It is interesting that, after getting less
than nothing with White in this line against
Bosch in round two of the Dutch
Championship, Timman decided to try the
black pieces in round 8! J.Werle-J.Timman,
Hilversum 2006, saw 15.g4 d6 16.xe6
fxe6 17.f4 ad8 18.cd1 e7 19.f2
c5 20.e3 f8 with a tense position in
which Black's chances were no worse
(½-½ in 60).;
C) 15.-- ]
[ 14.c2 d7 15.xe6 fxe6 was solid for
Black in J.Hjartarson-M.Illescas Cordoba,
Linares 1988 (½-½ in 88). ]
[ 14.a4 was played in K.Chernyshov-G.M.
Todorovic, Budapest 2002 (½-½ in 21), and
now xd4 15.xd4 c8 gives Black good
counterplay. ]
 [ 14.h2 is the most flexible continuation,
with a normal opening edge for White: d7
15.c2 ( 15.b3 ac8 16.fd1
was slightly better for White in L.Van Wely-
M.Magomedov, Yerevan Olympiad 1996,
and after the game's e5?? Ivan Sokolov
notes that White could have won on the spot
with 17.xe6 fxe6 18.xd5!! exd5
19.xd5 xd5 20.xd5+ h8 21.e6 )
15...ac8 16.fd1 ed8 ( or 16...c5
17.a4 b6 18.c2 ed8 19.d4
and a draw was agreed in D.Komarov-R.
Ponomariov, Donetsk 1998 ) 17.a3 d6
18.xc6 bxc6 19.d4 h7 20.b4 h5
21.h4 a5 with a draw in R.Ruck-G.M.
Todorovic, Topola 2004. ]
14...bxc6 15.a4 d7 16.h2
[ Not 16.c5? xc5 17.xc5 xh3
and White was a pawn down in B.
Khotenashvili-N.Batsiashvili, European
Rapid Championship, Gaziantep 2012 (0-1
in 33). ]
16...f5 17.c5 ab8
[ Instead: 17...e4 18.xe7 xg2 19.xg2
xe7 20.e3 e4+ 21.h2 ac8
was equal in Y.Seirawan-J.Piket, Merrillville
1997. ]
[ 17...d8!? , preserving the important dark-
squared bishop, was Kasparov's choice in
his simultaneous game (via a different move
order), which continued 18.d4 e4
and then:
A) 19.f3 g5 20.c5 e7? ( 20...c8!
held the balance ) 21.e4 dxe4 22.fxe4
g6?! was B.Züger-G.Kasparov, clock
simul, Switzerland 1987, when the champ
would have been on the ropes after 23.h4!
.;
B) 19.e3 d6 ( 19...e7 20.c5
was M.Gurevich-R.Ponomariov, Belfort
1998, and now xc5 21.xc5 d8!?
again retains the dark-squared bishops
and promises interesting counterplay; note
that Black can defend his c6-pawn with the
active ...Re6, building up on the e-file
and preparing to swing to the kingside )
20.c5 xc5 21.xc5 d7 22.a4 e4
23.fd1 b7 24.b3 xg2 25.xg2 e6
and Black was fine in I.Cheparinov-G.M.
Todorovic, Pancevo 2003 (½-½ in 54).
The exchange of the light-squared bishops
has weakened the white king, while Black
has met the pressure on c6 in a
convenient way, activating his rook with ...
Re6!. This rook holds the c6-pawn and is
well placed on the e-file, while it retains
options of moving to f6, g6 or (after ...h6-
h5) h6 to menace the white king.;
C) 19.-- ]
18.b3 d6 19.e3
[ 19.xd6 xd6 20.e3 d7 21.d4
was A.Greenfeld-A.Frois, Novi Sad
Olympiad 1990, and now a5 would have
retained the balance. ]
EXERCISE: Is there a way for Black to
undermine White's kingside pawn structure?
19...e4
[ ANSWER: Instead of this sensible move,
Scherbakov suggests the typical 19...h5!?
20.xd6 xd6 21.d4 e4 22.c5 e5
23.c3 e7 "and the idea ...h5-h4 gives
Black sufficient counter chances." ]
20.xd6 xg2 A little too compliant.
[ Again Scherbakov suggests 20...xd6
21.d4 h5 . ]
21.xg2 xd6 22.d4 e4 23.c5
White has an edge, but Black remains solid.
a5 24.fd1
[ QUESTION: What do you think of White's
chances in the rook and pawn endgame
after 24.xe4 xe4 25.c5 xc5
26.xc5 ?
ANSWER: This is a typical transformation of
the position and one which we need to know
how to assess properly. In general, Black
should not have any problems in the rook
and pawn endgames arising from the
Tarrasch. However, he has to play actively.
Here the move which fits the bill is a4!
, eliminating the queenside. After 27.xc6
axb3 28.axb3 xb3 , even if White wins the
d5-pawn we would have a theoretically
drawish 4 vs. 3 position on the kingside
(Black would prefer to trade one pair of
rooks before entering this structure).
Scherbakov analyses 29.c8+ h7 30.d1
e5 31.c5 b2 32.dxd5 xe3 33.c7
, concluding: "White wins the f7-pawn and
can certainly play for a win with four rooks
on the board, but Black should be able to
defend." ]
The rest is not so relevant for our purposes.
24...g5 25.e4 b4 26.c4 e6 27.xe6
fxe6 28.dc1 e5 29.d2 xc4 30.xc4
b8 31.a4 b5 32.xa5 xa5 33.xa5
dxe4 34.a4 f7 35.a7+ f6 36.a5 d5
37.h2 xb3 38.b6 a4 39.h4 e3
40.xe3 xa5 41.f3+ e7 42.xc6 a7
43.g2 d4 44.g6 f8 45.f5+ e7
46.g6 f8 47.e6 e4+ 48.h2 f3
49.g1 e4 50.d5 h5 51.e5
½-½
D34
Babula,V
Orsag,M
3: Czech Championship, Ostrava 2010
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.c3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 cxd4 10.xd4 h6 11.e3 e8
12.b3 a5 13.c2 g4 14.f5 c8
[ 14...b4 is the modern main line, after
which 15.d4!? xc3 16.xc3 xe2
and now 17.d1 ( or 17.d3 gives White
interesting compensation for the pawn. )]
15.d4
[ In the seventh game of the 1984/85 match
Karpov opted for 15.xe7+ . After xe7!
16.ad1 e8! Kasparov reached a position
he had analysed before the match, but had
taken an hour and a half on the clock, which
cost him dearly later in the game: 17.h3
( Black has several good responses to
17.xd5 , including Kasparov's suggestion
of b5 18.a3 h3 "and the threat of ...Rxe3
forces White to seek a way to maintain the
balance" ) 17...h5 18.xd5 ( 18.f5 xc3
19.bxc3 xe2 20.d4 e6 was agreed
drawn in P.Harikrishna-S.Iuldachev,
Commonwealth Championship, Sangli
2000 ) 18...g6 19.c1 xd5 20.xd5
c4 21.d4 ec7 22.b3 b6 23.e5
d7 24.e3 f6 25.c5 xc5 26.xc5
xh3 and Black has a comfortable game,
losing only because of a late blunder in A.
Karpov-G.Kasparov, Moscow 1984 (1-0 in
42). ]
15...c5 16.xc5
[ Not 16.e3?! e4 17.h4? xd4 18.exd4
b5! with a clear advantage for Black in S.
Savchenko-V.Ivanchuk, Odessa-Istanbul
(rapid) 2006 (0-1 in 32). ]
16...xc5 17.e3
[ 17.d4 c6! is suggested by Kasparov. ]
17...d4! This is the move! Kasparov gives
detailed analysis proving equality and
observes:
"Unfortunately, I rejected the energetic 17...
d4! – it seemed to me that the tactical
features of the resulting position were
advantageous to White. This is clear evidence
of the general slump in my play, caused by
the serious failures at the start of the 1984/85
match: after all, the obvious breakthrough in
the centre was both in the spirit of the
Tarrasch Defence, and in my style!"
[ 17...e6 and now:
A) 18.ad1 d7 ( the passive 18...c8
was Kasparov's choice in his ninth game
against Karpov in Moscow 1984/85; the
game, and its knight vs. bishop endgame,
is an absolute classic and one of the
most analysed games in chess history – I
encourage everyone to look it up in a
database or, ideally, to examine
Kasparov's commentary )
A1) Kasparov suggests 19.b4! c7 ( or
19...cc8 20.bxa5 d4 21.d3 xc3
22.xc3 dxc3 23.xc3 and White has
an extra pawn ) 20.cxd5 xd5 21.b1
with a clear advantage for White;
A2) 19.d3?! c4 20.xc4
was A.Lastin-A.Bezgodov, Russian
Championship, Moscow 1999, and now
xc4! 21.e3 ec8 22.xd5 xd5
23.xd5 xd5 24.xd5 xd5 25.xd5
c2 leaves Black with excellent drawing
chances in the rook and pawn endgame
in view of his activity.;
B) 18.fd1! (more accurate than 18 Rad1,
according to Kasparov) c4
was S.Shipov-A.Bezgodov, Russian
Championship, Elista 2001 (½-½ in 29),
and now Kasparov suggests 19.exd5!
xd5 20.xd5 xd5 21.e4! ( 21.b4!
is even stronger ) 21...f6 ( 21...a5
is a better chance, with some hope of a
blockade after 22.exd5 d6 ) 22.xd5
xd5 23.exd5 "with an extra pawn, since
in the event of xb2 24.xb2 xb2
25.f1! Black has a difficult endgame.";
C) 18.-- ]
[ 17...h5 is "interesting, but hardly
sufficient", according to Kasparov:
A) 18.a4 c6 19.fd1 b6 20.b3
xb3 21.axb3 was S.Temirbaev-A.
Bezgodov, Russian Cup, Samara 2002,
and now Kasparov suggests d4! 22.xc6
( or 22.a4 b5 ) 22...xe3! 23.fxe3
dxc3! 24.xb7 cxb2 25.ab1 c2
"with excellent compensation for the
exchange".;
B) 18.ad1 c8 was S.Iskusnyh-A.
Bezgodov, Russian Team Championship
2005, and now Kasparov suggests
19.fe1! b5 20.b1! g6 21.a1
"hiding the queen in the corner and
 nevertheless picking up the d-pawn (or
winning the exchange: xe3 22.fxe3 e6
23.e4 ).";
C) 18.-- ]
18.ad1
[ 18.fd1 is the main alternative: c6
19.xc6 xc6 20.xg4 xg4 21.e3 e5
22.exd4 f3+ 23.g2 xd4 24.a4 d6
25.xa7
A) 25...c8!? 26.xd4? ( or 26.f3 ed8
is suggested by Kasparov, with full
compensation for the pawn ) 26...a6!;
B) 25...d7 26.f3 c2 and Black had
more than enough for the pawn in H.
Boshku-B.Ivanovic, Nea Makri 1990 (0-1 in
34). ]
18...c6
[ 18...c4 19.xg4 xg4 was about equal
in L.Mkrtchian-R.Pokorna, Khanty-Mansiysk
Olympiad 2010. ]
19.xg4 xg4 20.e3 ge5 21.b3
EXERCISE: How can Black justify his play?
ANSWER: d3! This novelty was suggested by
Kasparov (in a book published 24 years after
the match!). His assessment of both the move,
and the impact of his 9th matchgame against
Karpov, is fascinating:
"To all appearances, 17...d4 gives Black
sufficient counterplay and is the best response
to Karpov's plan. But this move has still not
won general recognition. The point is that our
first match was not properly analysed, and my
defeats in the 7th and 9th games created the
impression of being opening failures, after
which the Tarrasch Defence lost its popularity
and the development of its theory was
retarded."
Black sacrifices a pawn, but the passer on d3
provides at least sufficient compensation.
22.xb7 a5 23.xc6 xc6 24.d7
[ Kasparov gives 24.xd3 b8 25.b4 xb4
26.xb4 xb4 27.b1 xd3 28.xb8+
h7 29.e4 ( or 29.e2 c2 ) 29...c1+
30.g2 f5! 31.d6 c2 with equality. ]
24...d8 25.g4 e5 26.e4 c4 27.g2
xc3?!
[ 27...d2! would have left Black on top. Black
still won this game against a good GM,
which is yet another advertisement for the
benefits of familiarizing oneself with the
classics! ]
28.bxc3 d2 29.e4 xc3 30.b1 d3
31.e7 c8 32.fd1 c2 33.d7 e5
34.d6 e4 35.b4 f3+ 36.h1 d4+
37.g1 c1 38.b8+ h7 39.b1 e2+
40.f1 h1+
0-1
D34
Bu Xiangzhi
Li Shilong
4: Chinese Championship, Xinghua 2013
[Sam Collins]
1.f3 c5 2.c4 f6 3.c3 e6 4.g3 c6
5.g2 d5 6.cxd5 exd5 7.d4 e7 8.0-0 0-0
9.g5 cxd4 10.xd4 h6 11.e3 e8
In this game we will deal with White's less
popular 12th moves. Many of these are
serious attempts for an advantage, and
Black's response varies depending on White's
set-up, so good preparation is required.
12.a4 By creating a cheap threat on c6,
White gains time for Rad1. In addition, he
hopes that the black bishop will be passively
placed on d7 (and interfere with the defence
of the d5-pawn).
[ 12.c2 was another approach, used by
Beliavsky in the sixth game of his
Candidates match against Kasparov in
Moscow 1983. After g4 , White again has
a wide choice:
A) 13.fd1 f8 ( 13...d7! is Kasparov's
suggested improvement, intending to meet
14 Nb3 with 14...Rad8; instead 14.f3?!
h5 15.f2 g6 16.a4?! c5
was already better for Black in A.Graf-G.
Ginsburg, Nuremberg 2006 ) 14.ac1
( 14.xc6 bxc6 15.d4 h5 was level in
M.Krasenkow-M.Illescas Cordoba,
European Team Championship, Batumi
1999 ) 14...c8 15.xc6 ( Kasparov
gives 15.a4! as stronger ) 15...bxc6
16.d4 ( 16.xa7?! c5! – Kasparov )
, was A.Beliavsky-G.Kasparov, Candidates
(6th matchgame), Moscow 1983, and now
16...a5! (Kasparov) would have held the
balance.;
B) . Beliavsky tried the other rook some 22
years after his game with Kasparov:
13.ad1 d7 14.f3 ( 14.b3 e6
15.c5 xc5 16.xc5 was a better try for
an advantage, as in E.Alekseev-V.Yemelin,
St. Petersburg 2006 ) 14...h5 15.f2
g6 16.a4 a6 17.xc6 bxc6 18.d2
e6 and Black had no problems in A.
Beliavsky-T.L.Petrosian, Moscow 2005
(½-½ in 32).;
C) 13.fe1 d7 14.ac1 ac8 15.xc6
 bxc6 16.d4 was Zhao Jun-Tu Hoang
Thong, Asian Championship, Cebu 2007,
and now f5 17.d2 a5 would have kept
chances level.;
D) 13.h3 has been tried by several GMs
recently, giving Black a choice of where to
put the bishop:
D1) 13...d7 14.ad1 c8 15.h2
f8 ( 15...c5 16.xc6 xe3 17.b4
d4 18.b1! was slightly better for White
in Zhang Zhong-S.Iuldachev, Manila
2008; 1-0 in 36 ) 16.xc6 bxc6 17.c1
e5! 18.d4 h5 with good attacking
chances in R.Chytilek-K.Malinovsky,
Czech League 2013 (0-1 in 34).;
D2) 13...e6 14.ad1 c8?!
( instead of 14...Qc8?!, 14...c8
looks natural and good ) 15.h2 f8
16.xe6 fxe6 17.f4 and White had a
favourable version of a standard set-up
in H.Stefansson-A.Pashikian, European
Championship, Antalya 2004 (1-0 in 42),
since Black's pieces are relatively
passive.;
D3) 13...h5 14.ad1 b4
was S.Volkov-S.Guliev, Dubai 2011, and
now 15.f5! would have given White a
slight advantage.;
D4) 13...--;
E) 13.-- ]
[ 12.a3 was a move played five times
against Garry Kasparov in Candidates
matches – once by Korchnoi and four times
by Smyslov. The results of those games
(four draws and one win for Black) helped
drive the popularity of the Tarrasch in the
1980s. The move is obviously quite slow
(the point seems to be to control b4 and
allow Qb3-a2 in some lines, though Black
could be forgiven for not quaking in his
boots at either of these ideas) and Black has
a wide choice:
A) 12...f8 13.c1 ( or 13.b3 a5
14.a2 e6 15.ad1 ½-½ D.Lekic-B.
Tadic, Montenegrin Team Championship
2007 ) 13...g4 14.b3 a5 15.a2
e6 16.fd1 c6 17.b1 d7 18.xe6
fxe6 19.h1 f7 20.f4 ad8 and a draw
was agreed in S.Bromberger-A.Lugovoi,
Gausdal 2004.;
B) 12...g4 13.b3 ( 13.h3!?
seems more typical ) 13...a5 14.a2
c4 15.c1 c8 and Black was very
comfortable in A.Yusupov-V.Akobian,
Philadelphia 2002 (½-½ in 41).;
C) 12...e6 was Kasparov's selected
defence, used in his Candidates matches
against Korchnoi and (four times!) against
Smyslov:
C1) 13.b3 d7 14.xe6 fxe6
15.ad1 d6! 16.c1 h8 17.a4
e7 18.e3 a6 19.h4 ac8
was comfortable for Black in V.Korchnoi-
G.Kasparov, Candidates (2nd
matchgame), London 1983 (½-½ in
31).;
C2) 13.h1 was tested three times in
the Smyslov-Kasparov match.
The following comments are based on
Kasparov's notes:
C2a) 13...d7?! 14.xe6 fxe6 15.f4!
ed8 16.g1 ac8 17.a4 h8
18.ad1 e8 19.e4 with a White
initiative in V.Smyslov-G.Kasparov,
Candidates (2nd matchgame), Vilnius
1984 (½-½ in 41).;
C2b) 13...g4! was Kasparov's
improvement, achieving comfortable
play after 14.f3 ( 14.h3 d7! )
14...h5 15.g1 ( 15.xc6 bxc6
16.a4 c8! 17.d4 e6 18.c1
d7 was also comfortable for Black in
the 10th game of the match; ½-½ in
38 ) 15...d7! 16.a4 c5 17.ad1
b6 18.fe1 g6! and a draw was
soon agreed in V.Smyslov-G.Kasparov,
Candidates (8th matchgame), Vilnius
1984 (½-½ in 28).;
C2c) 13...--;
C3) . Finally, 13.xe6 fxe6 14.a4
c8 15.ad1 h8 16.h1
was V.Smyslov-G.Kasparov, Candidates
(12th matchgame), Vilnius 1984 (0-1 in
40).
EXERCISE: Kasparov now played 16...
a6. Can you do better?
ANSWER: Kasparov recommends a5!
giving the following variations:
C3a) 17.xa7 c4 , when White is in
danger of losing his bishop, and gains
some compensation (but not more)
after 18.b3 ( or 18.b5 b6 19.a4
d7 20.e4 c7 21.exd5 xa7
22.dxe6 f6 ) 18...b6 19.e4 c7
20.b5 b7! .;
C3b) 17.f4 c4 ("Black gains an
important tempo and the entire rhythm
of White's attack is disrupted" –
Kasparov) 18.c1 b6! with good
play for Black.;
 C3c) 17.--;
C4) 13.--;
D) 12...-- ]
12...d7 13.ad1
[ Here 13.fd1 c5 14.e6!? fxe6
15.xc5 wins the bishop pair but improves
Black's structure, which should lead to
balanced play; for instance, b6 16.d6
a5 17.d4 c4 18.e5 c8
with equality in Wang Yue-T.L.Petrosian,
Tiayuan 2005 (1-0 in 49). ]
[ 13.xd5 xd5 14.xd5 b4 15.b3
xd5 16.xd5 h3 17.xd8 axd8
18.fd1 f6 gave Black full compensation
in G.Kaidanov-R.Robson, St. Louis 2011
(½-½ in 30). ]
EXERCISE: How can Black activate his
position?
ANSWER: 13...b4 14.b3 a5!
"Black holds on by tactical means, threatening
by ...a5-a4 to drive the queen from its active
position". (Kasparov)
I really like this plan, gaining space and
stability on the queenside. The b4-knight is a
monster which defends the d5-pawn and
almost traps the white queen, and White will
need to spend time getting rid of it.
15.a4 The most principled move, cementing
the b5-square, stopping Black's queenside
advance in its tracks and safeguarding the
queen on its active post. The b4-knight is
happy about the situation though.
[ 15.xd5? bxd5 16.xd5 xd5
17.xd5 h3! , winning the exchange, has
happened several times, including R.
Vaganian-B.Ivkov, USSR-Yugoslavia match,
Odessa 1975. ]
[ In the stem game Beliavsky allowed the
pawn to advance: 15.d2 a4 16.d1 a3
17.b1?! ( 17.bxa3 xa3 was also fine for
Black in Y.Korsunsky-J.Arizmendi Martinez,
Tanta 1998, ½-½ in 120; Kasparov suggests
17.b3! , giving several variations to
demonstrate White's advantage, e.g. a5
18.bxa3 xa3 19.xd5 bxd5 20.xd5
xd5 21.xd5 "winning a pawn and forcing
the opponent to fight only for a draw" )
17...f8 and now:
A) 18.b3 c8 19.c1 f5 20.a1
axb2 ( 20...c6 is level ) 21.xb2 xe3!?
22.fxe3 ( 22.xd5! xb3 23.axb3 b8
24.xf6+ gxf6 25.xf6 , with some
advantage, was a better attempt ) 22...e6
and Black enjoyed full compensation for
the exchange in a typical position in V.
Kunin-E.Weinzettl, Oberwart 2006.;
B) 18.bxa3 xa3 19.b2 a8
was A.Beliavsky-G.Kasparov, Candidates
(2nd matchgame), Moscow 1983 (0-1 in
38), and now Kasparov suggests 20.c2
xc2 21.xc2 with a level game. Instead,
Beliavsky played passively and Kasparov
went on to score a famous victory.;
C) 18.-- ]
15...c8 16.c2 Hurrying to exchange the
powerful b4-knight.
[ While White has tried other moves, I don't
believe he has serious chances for an
advantage if he leaves the b4-knight on the
board. For instance, 16.c1 c5 17.h3?
b6 18.fd1? runs into a standard
combination: xe3! 19.fxe3 d6! 20.f1
xg3 21.f3 g5 ( 21...e5 22.d1 h7
also led to a decisive attack in L.Ljubojevic-
J.Bosch, Dutch Team Championship 1999;
0-1 in 41 ) 22.cb5 e4 and Black's attack
was too strong in F.Elsness-B.Byklum,
Norwegian Team Championship 2009 (0-1
in 36). ]
EXERCISE: How would you continue here with
Black?
16...xc2 Not bad, but Black had a more
incisive option.
[ ANSWER: Kasparov suggests 16...b5!
as "the most interesting reply"; for instance,
17.xb4 bxa4 18.xa4 xb4 19.c3
( White lost material after 19.b6? xe3!
in V.Korchnoi-G.Kasparov, Herzeg Novi blitz
1983 ) 19...xc3 20.bxc3 a4 "with good
counterplay" (Kasparov), an assessment
which has been borne out by a number of
games: 21.b2 ( or 21.a2 e7
22.fe1?! , V.Majorovas-J.Ehlvest, USSR
1983, when xc3 23.d4 a3 is better for
Black ) 21...e7 22.fe1 was A.Lugovoi-A.
Lastin, St. Petersburg 2002, and now a3!
23.a1 b5 24.d4 e4 would have left
White struggling to equalise. ]
The game continuation gives White chances of
an advantage.
17.b6! A nice intermezzo, as otherwise
White has nothing.
[ 17.xc2 b4 18.d4 e4 19.b3 xc3
20.xc3 c6 was agreed drawn in Wu
Shaobin-C.Gokhale, Singapore 2002. ]
17...xa4! 18.xa4 d6 19.e3
[ 19.xa5 a6 20.c3 b5 21.xc2 bxa4
22.d3 xd3 23.xd3 e4 24.e5 f6
25.xf6 xf6 was level in E.Ghaem
Maghami-K.Yang, Subic Bay 2009 (½-½ in
 80). The white pawn on b2 is as much of a
target as anything in the black camp. ]
19...b4?
[ 19...b4 20.xb4 axb4 21.xd5 xd5
22.xd5 b3 should give Black sufficient
counterplay against b2 to hold, although I
prefer White. ]
[ 19...d8 , trying to exchange the strong b6-
bishop, also looks like a better defensive
attempt. ]
20.xa5 a8 21.xb4 xb4 22.xb4
xb4 23.b6 a2 24.b3
[ I don't think that the endgame after
24.xd5 xd5 25.xd5 xb2 26.b1
xb1 27.xb1 f8 28.xb7 is a piece of
cake for Black either with rooks on the
board. ]
24...d4 25.exd4?!
[ 25.xd4 c5 26.a4 gives more winning
chances than the text. ]
25...ee2 Black has now rustled up serious
drawing chances and held without much
trouble. The last few moves cannot be
recommended from a theoretical perspective
though and, in particular, 16...b5! seems
more promising.
26.f3 eb2 27.d5 xd5 28.xd5 b6
29.c1 c2 30.e4 xc1 31.xc1 g6
32.c7 e1 33.f4 f2+ 34.h1 f5
35.d5+ h8 36.f7 g5 37.d5 c5
38.e6 a1+ 39.g2 a2+ 40.f1 gxf4
41.gxf4 f2+ 42.e1 xf4 43.e2 h4
44.xf5 xh2+ 45.d3 f2 46.e6 f4
47.c6 g7 48.c3 f6 49.c8+ e5
50.xh6
½-½
D34
Movsesian,S
Sokolov,I
5: Sarajevo 2009
[Sam Collins]
1.c4 e6 2.g3 d5 3.g2 f6 4.f3 c5
5.0-0 c6 6.cxd5 exd5 7.d4 e7 8.c3
0-0 9.g5 cxd4 10.xd4 h6 11.e3 g4
This move is played about three times less
frequently than the most popular move, 11...
Re8, but has an impressive pedigree. It has
been used by Spassky (several times, and in
World Championship matches) and Korchnoi.
Modern GMs who have tried this more than
once include Ivan Sokolov (who gave it a
good write-up in his wonderful book Winning
Chess Middlegames), Daan Brandenburg and
Michael Feygin. The heroes of the line,
however, are Spanish GM David Larino Nieto
and German GM Alexander Berelowitsch, both
of whom have used this move extensively.
12.a4 A logical reply. Since the black bishop
has already committed itself to g4, Black
needs to look for something other than ...Bd7
(which worked well in Bu Xiangzhi-Li Shilong
above).
[ 12.h3 e6 and now:
A) . Sokolov makes the following
instructive comment: "It is good to note
that per definition Black is never worried
about 13.xe6 fxe6 , since the loss of
the bishop pair is fully compensated for by
the improvement of his pawn structure".
This was tested against me in a recent
game: 14.c2 c8 15.ad1 and now in
V.Burmakin-S.Collins, Riga 2013, I played
the careless d6? and immediately
panicked about the consequences of 16
Bxh6!, which my opponent missed but
would indeed have been very strong. Later
I even gained a winning position before
messing up in time trouble (½-½ in 38).
( From a theoretical perspective, 15...e8!
is much stronger, for instance 16.f4 b5
with counterplay. );
B) 13.c1 d7 14.h2 and then:
B1) 14...e5 (Sokolov's
recommendation) 15.b3 ( after 15.a4
xa4 16.xa4 c4 17.f4 g5! 18.b3
d7 , Black was not worse in B.Spassky-
V.Korchnoi, USSR Championship,
Moscow 1957; ½-½ in 30 ) 15...fc8
16.d2 was A.Czebe-A.Berelowitsch,
Haarlem 2007, and now a3!?
is interesting, when White can either
agree to a repetition with 16 Rb1 Bb4 17
Rbc1 Ba3 or play a complex position
after 17.f4!? ( 17.c2 c6 also looks
fine for Black, who has good prospects
of gaining control of the c-file thanks to
the bishop on a3 ) 17...xc1 18.xc1
c6 19.f5 xd4 20.xd4 xf5
21.xf6 gxf6 with a mess.;
B2) 14...ac8 15.xc6 bxc6 16.a4
was what worried Sokolov, but
Berelowitsch seems to have no
problems: fe8 ( 16...f5 looks fine too,
when 17.c5 can be met by d6 )
17.c5 xc5 18.xc5 f5 19.e1
e4 (Black has equalized and the game
comes to a logical conclusion) 20.d4
 g5 21.h4 e6 22.a4 xc5
23.xc5 e4 24.h3 f5 25.g2 e4
26.h3 f5 27.g2 e4 28.h3 f5
with a draw by repetition in S.Siebrecht-
A.Berelowitsch, Dutch Team
Championship 2013.;
B3) 14...--;
C) 13.-- ]
[ 12.b3 e6 and now 13.c1 ( 13.c5
xc5 14.xc5 e7 15.xe6 fxe6
was harmless in M.Dambacher-A.
Berelowitsch, Dutch Team Championship
2007; ½-½ in 31; 13.-- ) 13...e8 was tested
in two games of the Petrosian-Spassky
World Championship match, Moscow 1969:
A) . In Game 2 Petrosian opted for 14.b5
d7 15.5d4 h3 16.xc6 bxc6
17.d3 xg2 18.xg2 a5!
("with equality" – Kasparov) 19.c2 a4
20.d2 ( 20.c5 xc5 is a favourable
swap for Black, according to Sokolov )
20...b7 21.g1 ac8 22.fc1 d7!
"transferring the knight to e5 and preparing
...c6-c5 at the same time. Black was better.
" (Sokolov) The game was drawn in 61
moves.;
B) . Kasparov suggests 14.c5 xc5
15.xc5 as leading to a small plus for
White, giving the following variations: a5
( or 15...e4 16.xe4 dxe4 17.a4
d5 18.xe4 xa2 19.xa2 xa2
20.f3 ) 16.d4 xd4 ( or 16...e4 17.a3
ad8 18.e3 ) 17.xd4 ac8 18.e3 c4
19.d3 ec8 20.cd1 8c5 21.a3 .;
C) 14.e1 was Petrosian's choice in
Game 4, which continued d7 15.c5
( again Kasparov suggests 15.c5!? )
15...ac8 16.xe7 xe7 17.e3
and, as noted by Sokolov, "As a result of
the exchange of the dark-squared bishops,
White has good control of the dark
squares in the centre (d4 and c5) and also
enough space to manoeuvre his pieces.
White should be slightly better here."
Nevertheless, Black went on to win in 41
moves, which doubtless contributed to
the Tarrasch's popularity (again, showing
how fickle chess fashion is, since the
result had nothing to do with the
opening!).;
D) 14.-- ]
EXERCISE: How can Black activate his
position on the queenside?
ANSWER: 12...a5! Starting a knight tour
which will hit White's bishop (...Nc4) and
queen (...Nb6). While I don't like knights on
b6 generally, here it has the plus side of
securely defending the d5-pawn and not
blocking Black's counterplay on the c-file. In
addition, since it is difficult for White to play
b2-b3 because of pressure on the c3-knight,
a knight on b6 will always have options of
coming to c4.
[ 12...d7 is not bad and is actually the most
popular move in the position, but the
stronger guys have gone 12...Na5 and I
think it's a better move. ]
13.ad1 c4 14.c1 b6!
[ As Sokolov notes in his book, this move is
more precise than 14...c8 15.c2 d8
16.b3 e5 17.b2 d7 18.f3 h5 19.f5
, when White was better in T.V.Petrosian-B.
Spassky, World Championship (12th
matchgame), Moscow 1969 (½-½ in 49). ]
15.b5 Keeping the queen active and trying
to prevent ...Qd7.
[ 15.c2 d7 16.f3 looks odd to me,
despite having been played by a number of
GMs.
A) 16...h3 17.xh3 xh3 18.f5
( 18.f5 xf5 19.xf5 is better, as in A.
Matnadze-L.Hodova, European Junior
Championships 2001 ) 18...c5+ 19.h1
was Y.Vovk-D.Brandenburg, Dieren 2011,
and Black would have had an excellent
game after fe8! .;
B) 16...h5 17.f4 ( 17.f5 e8
was G.Timoscenko-Z.Sturua, Pavlodar
1982, which Sokolov assesses as better
for Black ) 17...ac8 18.f5 d6 19.d3
, H.Tikkanen-S.Brynell, Swedish Team
Championship 2010, and now e5!
, threatening ...Bxd4+ followed by ...Rxc3
and ...Bxe2, would have left Black on top.;
C) 16...-- ]
[ 15.b3 and now:
A) 15...d7 maintains the bishop on the
h3-c8 diagonal and seems the natural
response to White's refusal to play Qb5.
Then 16.fe1 ( instead, 16.a4
was met by the instructive ac8 17.f4
c4! 18.f3 b4 19.a2 e6
with equality in J.Cuenca Jimenez-D.
Larino Nieto, Spanish Junior
Championships 2003; and the endgame
after 16.b5 ac8! looks acceptable for
Black, who has better development to
compensate his potentially weak d5-
pawn; but 16.e3 gives some pressure,
since c5 17.e6! led to a White
 advantage in L.Vadasz-P.Szilagyi,
Hungarian Championship, Budapest
1974 ) 16...ac8 ( or 16...fe8 and if
17.a4 , as in M.Horn-M.Breutigam,
German League 2010, then ac8!
with equality looks like the most active set-
up for Black ) 17.db5 c4 18.a3
( but in this line 18.xd5! is critical: b4
19.xf7+ xf7 20.xd7 xd7 21.d1
xb5 22.a3 a4 23.xa4 xa4
and Black has enough material for the
queen, but after 24.b4! his rook is
sidelined and White is much better )
, was D.Larino Nieto, Madrid 2007, when
18...d4 19.e3 e6 20.exd4 xc3
21.xc3 xb5 22.xb7 bd5
would have given Black a good game.;
B) 15...c8 16.h3 and now:
B1) 16...h5 17.f5 g6 18.e4 xf5
19.exf5 d7 was A.Gupta-P.Konguvel,
Paleochora 2009, and now White could
simply have played 20.xd5
with a clear advantage.;
B2) 16...d7 is my preference. This is
actually the position after 17 Qb3 in
Movsesian-Sokolov with White to play!
Very surprisingly, White doesn't seem to
have a good way to exploit his free move.
For instance, 17.g4
( White's alternatives on move 17 are no
more impressive: 17.f4 , as in R.
Disconzi da Silva-M.V.Santos,
Guarulhos 2005, could be met by a5!?
18.a3 a4 19.a2 e8 with decent play
for Black; while 17.d3 c4 18.d1
c8 19.h2 b4 was comfortable for
Black in De.Abel-D.Larino Nieto, World
Junior Championships, Chalkidiki 2003
(0-1 in 34); finally, Sokolov gives 17.h2
c4! with good play for Black )
, as in E.Cordova-M.V.Santos, Havana
2007, can be logically met by 17...d6!
, taking control of the dark squares.
After 18.e3 c4 19.xd5 xd5
20.xd5 h4 21.xc4 xh3 22.f4
xg4+ 23.g3 xc4 24.xc4 xg3
25.fxg3 xg3+ , the game ends in
perpetual.;
B3) 16...--;
C) 15...-- ]
15...c8! The most solid move, developing a
piece to a good square.
[ After the more aggressive continuation
15...a6 16.d3 d7 , then 17.b3 fe8
18.b2 reaches a position which has been
played five times by Larino Nieto, scoring +2
=3 -0! In addition, he is the only player to
have played this way as Black, at least
according to my database!
A) 18...c5 19.c1 ac8 ( 19...h3
20.fd1 e5 was a more aggressive
approach in E.Llobel Cortell-D.Larino
Nieto, La Roda 2005, and ½-½ in 47, but it
would not have looked convincing after the
simple 21.e3 ) 20.e3 ( 20.fd1 h3
21.h1 g4 22.e3 e5 was agreed
drawn in A.Franco Alonso-D.Larino Nieto,
Elgoibar 2008 ) 20...h3 21.ce2 xg2
22.xg2 e4 23.f4 c7 ( 23...e7
24.f3 c5 25.d2 f6 26.b4 c7
was P.San Segundo Carrillo-D.Larino
Nieto, Spanish Championship, Ceuta 2008,
but here 27.h5! would have given White
a decisive advantage ) 24.f3 d6 25.a4
and Black's position looked ugly in V.
Andrejic-D.Larino Nieto, European
Championship, Budva 2009, though he
went on to win (0-1 in 38).;
B) 18...ad8 19.c1 h3 20.fd1 d6
21.e3 e5 22.ce2 e4 23.h1 e7
24.f3 f6 25.a4 bd7 26.d2 b8
27.f2 h5 28.e1 h4 29.f4 xf4
30.exf4 d6 31.g2 xg2 32.xg2
hxg3 33.hxg3 xe1 34.xe1 e8 35.f5
f8 was played in P.San Segundo Carrillo-
D.Larino Nieto, Spanish Championship,
Ceuta 2008 (0-1 in 45). The notation starts
going a little crazy over the next few
moves, which are irrelevant in any event,
since instead of the game's 36 Rxe8,
White could have gained a decisive
advantage with 36.h1 , when the
difference in piece activity is too great.;
C) 18...-- ]
16.h3 d7 17.b3
[ Another very strong player was easily
neutralized after 17.d3 e8 18.b3 b4
19.b2 e7 20.e3 e5 21.ce2 e4
22.a1 e7 and Black's activity fully
compensated for his IQP in R.
Kasimdzhanov-A.Berelowitsch, German
League 2006 (½-½ in 47). ]
17...c4 18.c2 c8 19.e3 c5
20.exd5 fxd5 21.xd5 xd5 22.xd5
xh3 23.xb7 Movsesian grabs a pawn.
[ 23.fe1 b6 looks equal too. ]
23...c7 24.f4 a5 25.e3 g5!
[ 25...xf1?? loses decisive material after
26.xe7 , since h3 27.b4! xb4 28.d6
is curtains. ]
26.xg5 xg5 27.g2 xg2 28.xg2
xa2 29.b4
½-½
D34
Li Shilong
Braun,Ar
6: Wijk aan Zee 2008
[Sam Collins]
1.d4 e6 2.f3 d5 3.c4 c5 4.cxd5 exd5
5.c3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 cxd4 10.xd4 h6 11.f4 A rather
annoying sideline. The bishop is active on f4
and puts some early pressure on Black's
position. In particular, White has the e2-e4
break in many lines (because the pawn isn't
blocked by a bishop on e3). The tried and
tested response is 11...Bg4 12 h3 Be6,
usually leading to a tenable endgame which,
admittedly, is not a lot of fun for Black. He
does have more enterprising alternatives, but
Black needs to be careful to hold his position
together since, after Qd4 and Rfd1, White will
already be creating strong central pressure.
After writing the above comments, I found
Korchnoi's notes to his game against Lobron
(discussed below), which describe the perfect
attitude for handling the endgame in this line:
"Here 11 Be3 is most often played. The
bishop move to f4 is also known in theory. I
think that here my opponent committed his
first inaccuracy. The bishop move to e3 leads
to a complicated middlegame, whereas after
Bf4 White intends to gain an advantage in the
endgame. And it is the endgame that I play
strongest of all!"
A great approach, and while it is admittedly
adopted by one of the strongest endgame
players in history, it is worth striving to be
able to approach endgames (in this and other
lines) with the same attitude.
[ By contrast, 11.xf6 xf6 leads to very
comfortable positions for Black.
Y.Seirawan-G.Kasparov, Niksic 1983,
continued 12.b3 ( 12.xc6 bxc6
is an excellent hanging pawn position for
Black, who will easily develop with ...Re8
and ...Rb8, while White struggles to put
any pressure on the c6- and d5-pawns;
12.e3 is the most popular move, but xd4
13.exd4 e6 leads to equality, for instance
14.b3 xd4 15.xd5 xc3 16.xe6
d4 17.d5 f6 18.f3 xb2 19.ad1
xf3 20.xf3 ad8 with total equality in B.
Thorfinnsson-A.Pashikian, European
Championship, Antalya 2004; ½-½ in 47 )
12...d4 13.e4 marked as "dubious" by
Seirawan in his annotations; ( 13.d5
can be comfortably met by e5! ) 13...e7
14.c1 b6 15.ec5 "The 'big idea' is
revealed. Clearly, the plan is to stop the c8-
bishop from developing thanks to the long
diagonal pressure. Also, simply to snap off
the d4-pawn, sometimes at the costs of the
g2-bishop." – Seirawan. ( a more recent
game saw an impressive activation of
Black's queen's rook after 15.d2 a5!
16.bc5 a4 17.d3 a5! 18.c2 e6
and Black was better in M.Al Sayed-E.
Senador, Olongapo City 2010 ) 15...d8
A) Seirawan had prepared 16.d3 b4
17.e4 xc5 ( in addition, Black has
other options such as 17...g5 18.cd1
c6 , and it isn't clear what White has
achieved ) 18.xc5 xa2 19.c4 xb2
20.b1 which he thought was good for
White, but decided against it at the last
moment – in fact Houdini thinks this
position is equal;
B) 16.c4?! xc5! 17.xc5 ( "Already I
have to fight for equality. While it is typical
in this line to play 17.xc5 e6 18.xc6
, giving up my pride and joy, Black isn't
obliged to recapture my bishop. Instead,
Black has xb3! 19.c1 xa2
when he is a pawn to the good. The text
gives Black a smaller plus." – Seirawan )
17...xb2 18.c2 xc2 19.xc2 b8
20.b2 d6 21.d1 b6 22.b3 b7
( it was also interesting to try 22...d3!?
23.xd3 xd3 24.exd3 d7 when the
question is whether the d3-pawn will prove
to be a strong passed pawn or a weak
isolated pawn ) 23.bd2 bd8 24.f1
("Although I will win back my pawn by
capturing the c6-knight and then the d4-
pawn, the resulting position of rook and
bishop versus rook and knight will favour
Black, and I'll be forced to fight for the
draw. Black will have the sunny side of
the torture." – Seirawan) a6 25.xc6
xc6 26.xd4 c5 27.b3 xd2
28.xd2 c7 29.d8+ h7 30.e1 c4
31.d2 g6 32.c1 g7 33.a3 f6
34.e3 e7 35.d4 d7 36.c3 xd4
37.xd4 b5 38.d3 xd3 39.xd3
d6 and here, instead of the drawing 40
Kd4!, Seirawan blundered with 40.e4?
 and went on to lose a famous endgame
(0-1 in 54). ]
11...g4 12.h3
[ 12.a4 as at move 13, can be met by b6
. ( Instead, 12...d7 13.fd1 fd8
14.ac1 a6? 15.xd5! xd5 16.xd5
xd4 17.xd4 was clearly better for White
in M.Kanep-J.Ehlvest, Harjumaa (rapid)
2010. )]
12...e6 The main line. This variation (as will
be seen in the game and the notes) often
gives rise to an endgame (after White takes
on c6 and plays e2-e4) where White is slightly
better in view of his two bishops but, in
practice, Black has held the balance without
too much trouble.
Black's 11th and 12th move alternatives will
be covered in the next game.
13.xe6 This is the most direct, leading
directly to the endgame.
[ 13.cb5 d7 14.xe6 fxe6 15.c1 e5
16.d2 ad8 left Black with no problems in
N.Dzagnidze-A.Gara, French Team
Championship 2013 (½-½ in 25). ]
[ 13.c1 c8 14.xe6 fxe6 15.e4 d4
16.e5 dxc3 17.exf6 xf6 18.bxc3 xd1
19.fxd1 fd8 also leads to a slightly better
endgame for White which Black ought to
hold, as in I.Sokolov-V.Akobian, Calatrava
(rapid) 2007 (½-½ in 45) and A.Greenfeld-
S.Iuldachev, Calicut 2007 (½-½ in 62).
Sokolov's assessment of his game against
Akobian was: "In the slightly better endgame
that resulted after 11 Bf4, I never had any
real winning chances and the game ended
in a dull draw." We will get an idea of how
Black should handle these positions in the
main game. ]
13...fxe6 14.e4! The next few moves are
forced. As noted in the Structural Introduction,
taking on e4 would be a disaster. d4 15.e5
dxc3 16.exf6 xf6 17.bxc3 xd1
[ The alternative is 17...xc3 . As explained
by Korchnoi: "The game is proceeding along
a well-trodden theoretical path. There were
also games in which Black avoided the
capture on c3 and in the end equalized. But
I thought: White already has the advantage
of the two bishops, and if Black has to suffer,
then it should at least be for some material."
White now has a choice:
A) 18.c1 f6 19.xc6 ( 19.b3 d4
and White has nothing better than
retreating to d1 ) 19...bxc6 20.b3 e5
21.xe5 xe5 22.xc6 ( 22.fe1 d5
equalized in M.Wiedenkeller-Al.Poulsen,
Copenhagen 1986 ) 22...ae8 and White
has nothing, according to Korchnoi.;
B) 18.b1 xd1 ( 18...f7 19.xb7
xb7 20.xc6 xd1 21.xd1 ab8
22.xb8 was agreed drawn in V.Anand-
Comp Fritz 6, Frankfurt rapid 1999 )
19.fxd1 and now:
B1) 19...f7 20.xb7 ( 20.xc6 bxc6
21.bc1 e5 22.e3 d4 23.xd4
exd4 24.xd4 b8 25.c2 c7
was drawn in U.Krause-A.Ilgner,
Kappeln 1990; but here 20.bc1; or
20.d6 were better tries for an edge, so
I prefer the move order Fritz used
against Anand, given above ) 20...xb7
21.xc6 ab8 22.xb8 and a draw was
agreed in W.Paschall-G.Kallai,
Hungarian Team Championship 2012.;
B2) 19...ad8 is also quite sound:
B2a) 20.e3 d4 21.xd4 xd4
22.xb7 e2+ 23.g2 c3 24.xd8
xd8 leads to a level endgame, as in D.
Hergott-R.Alonso Garcia, Linares,
Holland 1992 (½-½ in 50) and J.Jurek-
D.Florea, Olomouc 2002 (½-½ in 50).;
B2b) . After 20.dc1 Korchnoi gives
d2 21.xd2 xd2 22.xb7 d4
23.cc7 fxf2 24.xg7+ ( 24.e4
f5 25.xf5 g2+ and Black has
perpetual ) 24...f8 with equality,
which is confirmed by the computer:
25.h1 f6 , followed by checks by
the rook on the d-file.;
B2c) 20.xd8 xd8 and now:
B2c1) 21.xb7 xb7 22.xb7
d4! is level.;
B2c2) 21.d1 c6 leads to a
further branch:
B2c21) 22.e3 b4 23.d7 f7
24.xc6 bxc6 25.xa7 xa7
26.xa7 f7 was roughly level in
D.Yevseev-A.Kirusha, Gatchina
2001. White has an outside
passed pawn but Black has a
passed c-pawn and he really
shouldn't lose this position.;
B2c22) 22.d7 f7 23.d6 e7
24.xc6 bxc6 25.xc6 d4
with equality in A.Panin-R.Bataev,
Murom 2000 (½-½ in 42).;
B2c23) 22.d6 e5 23.e3 d4
24.d7 xe3 25.fxe3
(this endgame is assessed by
Gurevich as slightly better for
 White) a5 26.d5+ h7
(as Gurevich notes, Black plans ...
b7-b5 and ...Nc4 with equality)
27.e7 was M.Gurevich-J.Bosch,
German League 2000. Now after
Gurevich's suggestion of d8
28.e4 ( 28.e4+ g8 29.xe5
c6 is okay for Black ) , I like
28...d6! , intending to activate the
rook via b6. The position is quite
unbalanced and, frankly, I still
prefer White, but I think this is fully
playable for Black.;
B2c24) 22.--;
B2c3) 21.e3! ("This subtle move,
as I later learned, was an innovation:
it forces an important weakening of
Black's pawns on the queenside." –
Korchnoi) b6 ("Now the black knight
at d8 will not have a post at c6, and
much time will be required to bring it
into play." – Korchnoi) 22.d1
was played in E.Lobron-V.Korchnoi,
Bad Homburg 1998. Korchnoi
assesses the position as follows:
"White has definite compensation for
the pawn. His rook will invade
Black's position, and the two bishops
will have their say." ( Korchnoi notes
that 22.e4 has been played in
subsequent games, but not ones
which appear in my database; I
suggest continuing to bring the
knight into play with f7 23.c1
d6 24.d3 e5 with a complex
fight ahead ) . The game continued
22...f6 ( the attempt to force an
immediate draw with 22...f7
23.d7 e5 24.xa7 c4 25.c7
xe3 26.fxe3 d2 fails to 22 Rc7!
Bb2 23 f4! – Korchnoi ) 23.e4 f7
24.g6 b7 25.f4 f7 26.c1
d7 27.c8+ d8 when Korchnoi
gives the following assessment:
"White's advantage is obvious –
Black is obliged to make forced
moves. But perhaps White's success
is only temporary. If Black can
regroup, he will evict the bishop from
g6 and neutralize the white rook that
has invaded his position."
Korchnoi has analysed the rest of
the game in detail in Informator and,
in particular, his wonderful best
games collection, the second volume
of which is a fantastic lesson in how
to win games with Black (often
featuring apparently dull or passive
positions such as this one). I can't
recommend it highly enough. For our
purposes, this dull but playable
position gives White full
compensation but no more, so we
shouldn't be afraid of it.;
B2c4) 21.--;
B2d) 20.--;
B3) 19...--;
C) 18.-- ]
18.fxd1 fd8
[ 18...xc3 has been played in several
games, but if Black wants to go for this it is
more accurate to play it on move 17 (as
Korchnoi did), when Rc1 can be met by ...
Qf6. Here 19.ac1 gives White strong
pressure. ]
19.ab1
[ 19.e3 xc3 20.ab1 was M.Dlugy-M.
Vucic, New York Open 1992, and now I like
d4! 21.xb7 ( 21.xd4 xd4 22.xb7
ab8 looks level too ) 21...xe3 22.xd8+
xd8 23.fxe3 d1+ 24.f2 d2+ 25.f1
a5 26.xa7 c4 with enough counterplay
to hold. ]
19...xd1+ 20.xd1 d8 21.xd8+
[ The position after 21.c1 d3 22.b1 d7
23.c4 d4 has been defended twice by
Czech IM Milan Orsag: 24.b5 ( 24.h4 f7
25.e1 was R.Akesson-M.Orsag, Aaland-
Stockholm 1997, and now I suggest b4!?
26.h3 xf2+ 27.xf2 d3+ 28.e2
xe1 29.xe1 d3 , when Black is not
worse in the endgame ) 24...d8 25.h4 b6
and Black held without difficulty in J.Nilssen-
M.Orsag, Istanbul Olympiad 2012 (½-½ in
40). ]
21...xd8 22.d2
[ Golod gives 22.e3 xc3 ( but 22...b6
is more solid, keeping the position more
closed ) 23.xa7 as slightly better for
White. ]
22...f7 23.f1 c6 24.e2 e7 25.e4
d6 26.h4 b6 Black's position is solid and, in
the game, he drew without major problems. I'd
be the first to admit that such endgames aren't
a great deal of fun for Black, but sometimes
we have to play positions like this. In general,
the entire endgame seems much closer to a
draw than to a serious White advantage.
27.f4 e7 28.h5 e5 29.g4 exf4 30.xf4+
e5 31.d2 d5 32.d3 b5 33.f3 a5
34.e1 b6 35.f2 a4 36.d4 c5+
37.xc5+
½-½
D34
Socko,M
Petrosian,TL
7: Gibraltar 2008
[Sam Collins]
In this game I've compiled a few ideas for
Black should he want to avoid the endgame
featured in the last game. These are
enterprising but have to come with a health
warning since White seems to have at least an
edge in all lines.
1.d4 d5 2.c4 e6 3.c3 c5 4.f3 c6
5.cxd5 exd5 6.g3 f6 7.g2 e7 8.g5
cxd4 9.xd4 0-0 10.0-0 h6 11.f4 g4
[ Black has tried a lot of other moves here:
11...b6 and now:
A) . A game from the Women's World Blitz
Championship demonstrates a number of
Black's attacking ideas in the Tarrasch:
12.xc6 bxc6 13.a4 a6 14.c1 g4
15.e1 fe8 16.h3 h5 17.e3? b4
18.c3 xe3! (this typical sacrifice works
perfectly in this position) 19.fxe3 b6
20.d3 g6 21.d2 e8 22.a3 xe3
23.h1 e4! 24.xe4 xe4+ 25.h2
xg3! 26.f4 g2+ 27.h1 xc3
( 27...g1+ forced mate ) 28.xc3 xe2+
and White resigned in Zhao Xue-V.Gunina,
Batumi 2012.;
B) 12.b3 d8 ( or 12...d4 13.a4 b4
14.xc6 bxc6 15.xd4 ) 13.c1 d4
14.a4 was Y.Pelletier-G.Meins, Biel
2009, and now Black could improve with
b5 15.ac5 d5 with decent play.;
C) . The forced line after 12.e3 xb2
13.xd5 xd5 14.xd5 h3 has scored
heavily for White following 15.b1 a3
16.xb7 xd4 17.xd4 xf1 18.xf1
, as in M.Gurevich-A.Berelowitsch, Tanta
1997.;
D) 12.-- ]
[ 11...e8 is another possibility:
A) 12.cb5 , as in I.Sokolov-G.M.
Todorovic, Niksic 1991, can be met by
b6 13.c7 ( 13.c7 a6 14.b3
xd4 15.xd4 c5 is fine for Black )
13...xd4 14.e5! c6 15.xa8 d8
16.c7 xe5 17.xe8 xe8 18.xd5
xd5 19.xd5 c6 with three minor
pieces for two rooks and a pawn.;
B) 12.c1 g4 13.h3 h5 has been
tested in a number of GM clashes:
B1) 14.a4 c8 ( not 14...c5?
15.xc6 bxc6 16.g4! g6 17.xc6
and White was clearly better in L.
Fressinet-G.M.Todorovic, European
Championship, Budva 2009 ) 15.xc6
( 15.e3 might be a better try for an
advantage ) 15...bxc6 16.g4 g6
17.xa7 was Val.Popov-V.Yemelin, St.
Petersburg 2005, and now d6
would have left Black with full
compensation for the pawn.;
B2) 14.b3 b4 15.g4 xc3 16.bxc3
g6 17.c4 e4 and Black was not
worse in D.Yevseev-A.Lugovoi, St.
Petersburg 2005 (½-½ in 41).;
B3) 14.g4 g6 15.cb5 b6 16.c7
a6 17.a4 , with an edge for White, was
tested in B.Lajthajm-G.M.Todorovic,
Serbian Team Championship 2012 (½-½
in 37).;
B4) 14.--;
C) 12.-- ]
[ 11...a6 has also been tried:
A) 12.xc6 bxc6 13.e4 d4 14.e5 dxc3
15.exf6 xf6 16.bxc3 e6 17.d6 e7
18.xe7 and a draw was agreed in R.
Vidonyak-G.Ginsburg, Nuremberg 2011.;
B) 12.c1 e8 13.a4 d7 and then:
B1) 14.a3 f8 15.e1 c8 16.xc6
bxc6 17.e3 e4 18.c5 and a draw
was agreed in J.Lembak-K.Malinovsky,
Slovakian League 2013. Play could have
continued xc5 19.xc5 xc5
20.xc5 a5 , followed by ...Qb6 with
counterplay, ( or 20...b8 21.b4 b5!?
. );
B2) 14.xc6 xc6 15.b3 a5 16.fd1
f8 17.e3 was slightly better for White
in A.Aleksandrov-E.Magerramov, Dubai
2001 (1-0 in 57).;
B3) 14.--;
C) 12.-- ]
12.h3 h5 An enterprising move. Having
played over several hundred Tarrasch games
while preparing this book, I can say that very
often players seem to like retreating their
bishop to h5 after h3, presumably with a view
to pressuring the e2-pawn. In fact the
pressure against e2 rarely becomes a
significant factor, but the fact that the bishop
can no longer defend the d5-pawn renders
Black's position tactically shaky. Whether he
can get away with it depends on the specific
features of a given position – in general you
should spend a couple of minutes before you
play ...Bh5 in the Tarrasch and try to make
sure that your d-pawn isn't dropping.
Here the move makes some sense since,
without a bishop on e3 defending the d4-
knight, Qb3 isn't available.
[ Instead, 12...d7 can lead to a forcing
sequence after 13.h2 b6 14.e3 xb2
15.xd5 xd5 16.xd5 ad8 17.b1
a3 18.xb7! (improving on the 18 Qc1 of J.
Lautier-A.Grischuk, Cannes rapid 2001; ½-½
in 29), and White went on to win in K.
Sasikiran-A.Lugovoi, Moscow 2004 (1-0 in
41). Whatever the objective merits of this
line, it seems rather impractical to go in for
a long forced variation, where a single
mistake might be fatal, in a basically
harmless sideline like 11 Bf4. ]
13.c1
[ Many other moves have been tested
against T.L.Petrosian. While the games
were, admittedly, mainly played at rapid or
blitz time controls, they are instructive in
demonstrating Black's attacking potential:
13.b3 e8 14.g4 g6 15.xd5 xd5
16.xd5 f6 and Black had equalized in V.
Golod-T.L.Petrosian, Dos Hermanas (online
blitz) 2007, since he is about to regain his
pawn. ]
[ 13.a4 is dangerous.
As indicated in my note to Black's 12th,
White aims simply to attack and win the d5-
pawn: d7?! ( 13...b6! 14.f5 c5!
is sounder, as in R.Backelin-L.Johannesson,
Gausdal 1990; 0-1 in 54 ) 14.fd1
(already it is unclear how Black can
neutralize the pressure) fd8 15.ac1 g5
16.e3 ac8 was L.Fressinet-T.L.Petrosian,
France-Armenia rapid match, Paris 2009,
and now both 17.g4 ( and 17.xd5
would have led to a decisive advantage for
White. )]
[ 13.f5 g6 14.xe7+ xe7 and now:
A) 15.xd5 xd5 16.xd5 ad8
17.b5?! ( 17.c4 d4 18.h2 e4
gives Black interesting compensation; for
instance, 19.fd1 xg2 20.xd4 xd4
21.xd4 c6 22.d6 d8 23.xg7+
xg7 24.xe7 d2 with a likely draw in
view of the pieces remaining and Black's
activity ) 17...d4! was already better for
Black in V.Golod-T.L.Petrosian, Dos
Hermanas (online blitz) 2007.;
B) 15.e4 d4 16.d5 xd5 17.exd5
b4 18.a4 ( 18.xd4!? c2 19.d6
d7 20.d5 xa1 21.xa1 gives White
enough compensation for the exchange )
, was G.Vojinovic-G.M.Todorovic, Skopje
2011, and now 18...c2 19.ad1 d3
would have offered Black counterplay
based on his strong d-pawn.;
C) 15.-- ]
13...d7 14.a4 ac8
[ After 14...fd8 15.fd1 a6 16.xc6
, a draw was agreed in G.Gutman-T.L.
Petrosian, Cappelle la Grande 2008. ]
[ But 14...a6 looks interesting; for instance,
15.fd1 b5 16.b3 a5 17.c2 g6
with good play for Black. ]
15.fd1 White has achieved an optimal set-up
and the bishop on h5 just looks really
misplaced. fd8
[ 15...xd4 16.xd4 b5 17.b3
was better for White in N.Kosintseva-L.
Mkrtchian, Krasnoturinsk 2006 (1-0 in 61). ]
16.xc6 Giving Black some major respite by
solving his biggest tactical problem.
[ Instead, 16.g4! would have led to a clear
advantage for White. 16...Bg6 17 Nxd5! is a
typical winning shot, so Black would have to
go in for xd4 17.xd7 xd7 18.xd4
g6 19.xd5 xc1+ 20.xc1
, when White is a solid pawn up and Black's
drawing chances don't look so great. ]
[ Even the immediate 16.xd5 xd5
17.xd5 xh3 ( 17...xd5 18.xc6 wins )
, leaves Black struggling to stay in the game
after 18.g2 g4 19.f3 h3 20.xc6
bxc6 21.xa7 . ]
16...bxc6 17.e3
[ 17.g4 g6 18.e4 would have led to
unclear play after d4 . ]
17...b8 Black has now equalized. The rest is
not so relevant since we have covered the
most important theoretical moments. White
lost the game in the end, which was partially a
result of weaknesses she later incurred on
the light squares, but more a result of playing
against a strong and inventive GM.
18.d2 b4 19.a5 e8 20.d4 db8
21.b3 4b7 22.c5 g6 23.e3 d7
24.xe7 xe7 25.e2 e5 26.d4 b6
27.c3 c8 28.f4 d7 29.xd5 f6
30.g2 e4 31.xe4 xe4 32.c5 d7
33.h2 b5 34.c4 d5 35.f2 c5
36.e2 e8 37.c3 h5 38.f1 f5 39.h4
xe3 40.d1 e7 41.d5 xh4+ 42.gxh4
xh4+ 43.g1 e4 44.g2 xg2 45.a6
h1+ 46.f2 f3+
0-1
 Back to Contents Page
Chapter Two

9 Bg5 c4
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 c4

This caused something of a stir when Alexander Grischuk started beating up various 2600+ GMs with it in the early
2000s. Certainly the arising positions are unusual and unbalanced, and if White is not well prepared he can quickly
find himself worse. The same mystery elite GM I mentioned in the 9...cxd4 introduction confessed he had found
"nothing" against 9...c4 after several hours' work. More recently, Aagaard and Ntirlis devoted over 100 pages of
computer analysis to proving equality in these lines (generally featuring an early ...h7-h6 by Black). I don't wish to
parrot this work so my focus is on other lines which also seem very playable.
10 Ne5
This is almost universally played.
10...Be6 11 b3

This move, and the closely related 11 Nxc6, are, in my view, the best responses to 9...c4. White sets about trying to
dismantle the black pawn chain. Often complex endgames result where White has a superior structure but Black
has activity and counterplay (as well as the less noble drawing methods of relying on simplified positions with the
pawns on one side, and the peaceful properties of rook endgames). SeeSargissian-Halkias for the details.
White's alternatives at moves 10 and 11 are covered inOnischuk-Akobian. Of these the most challenging is 11 f4.
The mystery elite player just made a face when I suggested that this was dangerous for Black, so I suppose we
have a convincing antidote...

Key Notes

1. 9 Bg5 c4 leads to fundamentally different positions to the Classical Tarrasch line 9...cxd4. Black has found a lot
of resources here and White needs detailed preparation to have a chance at obtaining an advantage.
2. In Sargissian-Halkias White uses the safest and most classical plan, aiming to break the black queenside and
central pawn chain. Often an endgame results where White has a superior structure but Black has a lot of activity
and, on occasion, relies on the drawish properties of rook and pawn endgames.
3. Onischuk-Akobian shows a more ambitious approach by White, playing for a kingside space advantage and a
2-0 central majority. While this plan is very dangerous, we can see from the game and notes that Black always has
counterplay against the white centre and, sometimes, the white king.

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 c4 10 Ne5 Be6 11 b3 –
Sargissian-Halkias
Sargissian-Halkias
11 f4 – Onischuk-Akobian

Back to Contents Page

9 Bg5 c4
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 c4 This caused something of a stir
when Alexander Grischuk started beating up
various 2600+ GMs with it in the early 2000s.
Certainly the arising positions are unusual
and unbalanced, and if White is not well
prepared he can quickly find himself worse.
[ The same mystery elite GM I mentioned in
the 9...cxd4 introduction confessed he had
found "nothing" against 9...c4 after several
hours' work. More recently, Aagaard and
Ntirlis devoted over 100 pages of
computer analysis to proving equality in
these lines (generally featuring an early ...h7-
h6 by Black). I don't wish to parrot this work
so my focus is on other lines which also
seem very playable. ]
10.e5 This is almost universally played. e6
11.b3 This move, and the closely related 11
Nxc6, are, in my view, the best responses to
9...c4. White sets about trying to dismantle
the black pawn chain. Often complex
endgames result where White has a superior
structure but Black has activity and
counterplay (as well as the less noble drawing
methods of relying on simplified positions with
the pawns on one side, and the peaceful
properties of rook endgames). See
Sargissian-Halkias (Game 8) for the details.
[ White's alternatives at moves 10 and 11
are covered in Onischuk-Akobian (Game 9).
Of these the most challenging is 11.f4
. The mystery elite player just made a face
when I suggested that this was dangerous
for Black, so I suppose we have a
convincing antidote... ]
D34
Sargissian,G
Halkias,S
8: European Championship, Aix-les-Bains
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 c4 10.e5 e6 In this game we will
examine White's main attempts, 11 b3 and 11
Nxc6 (which often transpose). White aims to
break up the black pawn chain on c4 and d5
(often a pawn comes to c6 too).
11.b3
[ The alternative is 11.xc6 bxc6 and then:
A) 12.a4 c8 13.b3 a5 14.e3 fd8
15.xf6 has been tried by a couple of
strong GMs. gxf6! ( better than 15...xf6
16.c5 with a queenside bind in V.
Laznicka-A.Givon, European Cup, Eilat
2012; 1-0 in 26 ) 16.h5 c5 17.dxc5
xc5 18.xc5 xc5 19.fd1 c3
with an unclear game where Black enjoys
strong counterplay with the c-pawn, as in
H.Koneru-N.Zhukova, FIDE World Cup,
Khanty-Mansiysk 2012 (0-1 in 37).;
B) 12.b3 and now:
B1) 12...cxb3 13.axb3 h6 14.xf6
xf6 15.a4 f5 16.d2 d6 17.c5
gave White some edge in P.H.Nielsen-I.
Rogers, Turin Olympiad 2006 (1-0 in
55).;
B2) 12...h6 is an idea deeply analysed
by Aagaard and Ntrilis. To take just one
line: 13.xf6 xf6 14.bxc4 dxc4 15.e3
a5 16.c1 ac8 17.a4 xa4
18.xa4 e7 19.b2 ( 19.b1 c7
20.c3 d8 21.fd1 a5 22.f3 b4
looked balanced in Al.Ramirez-R.Skytte,
Cappelle la Grande 2012 ) 19...a3
20.c2 xb2 21.xb2 c7 22.c1
was tested in A.Rasmussen-J.Aagaard,
Helsingor 2012. Now Jacob played g5
, ( forgetting his own recommendation
22...g6 , which was given a couple of
pages of analysis in his book with
detailed lines ending in equality. );
B3) 12...a5
B3a) 13.d2 ad8 transposes to 13
Nxc6 bxc6 in the main game,
( although Black might consider other
options, such as 13...a6!? );
B3b) 13.c2 fd8 ( 13...ad8
seems less logical, since the rooks
should aim for the queenside; after
14.fd1 b4 15.a4 h6?! 16.xf6
gxf6 17.bxc4 dxc4 , Ax.Smith-E.Berg,
Swedish Championship blitz play-off,
Falun 2012, White could have taken a
free pawn with 18.xc6 ) 14.fd1
ac8 15.a4 c5 16.d2 b5
( 16...c7 17.f4 a5 18.xc5
xc5 19.dxc5 xc5 20.e5 d4
21.xf6 gxf6 is a playable alternative,
as in I.Sokolov-B.Tadic, Bosnian Team
Championship 2009; ½-½ in 30 )
17.bxc4 cxd4! has led to a couple of
excellent Black wins: 18.xh7+ ( or
 18.cxb5? xc2 19.e1 e4 20.f1
f6 21.a3 b8 22.ab1 c3
23.xc3 dxc3 24.bc1 xc1
25.xc1 c8 and White resigned in M.
Rodshtein-A.Kolev, Albena 2011 )
18...xh7 19.cxb5 e4 20.a5 d7
21.ac1 xc1 22.xc1 b7 23.b1
d7 and now, rather than 24.f3
with an unclear game,
( White blundered with 24.e3?? xb5
25.xb5 xb5 26.b2 a3 27.d1
e2 (White can save the piece, but at
too high a cost) 28.f3 d6 29.f2
c4 30.e1 dxe3 31.h3 d4
32.f4 d3 33.xe2 d2 34.xd2
exd2 35.c3 e3 and White
resigned in E.Bacrot-A.Grischuk, FIDE
Rapid Grand Prix, Dubai 2002. );
B4) 12...--;
C) 12.-- ]
11...a5 This is the move we will focus on.
[ 11...h6 is recommended by Aagaard and
Ntrilis, which normally transposes into the
lines after 11 Nxc6 bxc6 12 b3 h6 above. ]
12.d2
[ 12.d2!? was a fresh attempt from a
former FIDE world champion: b4
( 12...b6!? looks logical, trying to exploit
the lack of the defence of the d4-pawn; for
instance, 13.bxc4 xd4 14.cxd5 xd5
15.xd5 xd5 and Black must be close to
equality ) 13.xc6 bxc6 14.c2 a6
15.xd5! xd5 16.bxc4 with some
advantage for White in R.Kasimdzhanov-S.
Citak, Turkish Team Championship 2010
(½-½ in 38). ]
12...ad8 Black brings a rook to the d-file. He
chooses this rook not for strategic purposes
(while one rook certainly belongs on d8, its
colleague would be better off on c8 or b8
rather than e8), but to avoid White being able
to take with the bishop on c6 with gain of time
after b3xc4 and Nxc6.
[ After 12...b4 13.xc6 bxc6 14.xf6
gxf6 15.fc1 we reach a major position
which has been tested in several
encounters between world-class players:
ad8 ( after 15...ac8 16.bxc4 dxc4 17.a3
xc3 18.xc3 xc3 19.xc3 fd8 20.e3
c5 21.d5 xd5 22.xd5 xd5 23.xc4
b8 , White has the better endgame, as
reflected in Wang Hao-S.Iuldachev, Asian
Team Championship, Vishakapatnam 2008,
and Mi.Hoffmann-K.Kiik, European Cup,
Ohrid 2009 ) 16.bxc4 dxc4 17.e3
and now Black has a choice:
A) 17...c5 18.d5! xc3 19.xc3 xd5
20.xd5 xd5 21.c2! c8 ( or 21...b8
22.xc4 b4 23.b1 xb1+ 24.xb1
b6 25.c2 and White went on to win in
S.Mamedyarov-V.Akobian, Gibraltar 2012;
1-0 in 104 ) 22.xc4 d2 23.c1 xc2
24.1xc2 led to an absolutely model
endgame performance in B.Gelfand-A.
Grischuk, Russian Team Championship
2004 (1-0 in 45). Anyone wishing to study
this ending in detail is referred to
Gelfand's wonderful best games collection
for instructive coverage.;
B) 17...f5!? 18.xc6 f4 is a much more
aggressive approach. The only problem is
that, as far as I can see, it doesn't quite
work: 19.a3 fxe3 20.xe3 xc3 21.xc3
d6 was W.So-V.Akobian, Wijk aan Zee
2010, and now 22.d5!? xd5 23.xd5
xd5 24.e2 would leave White a pawn
up with good winning chances.;
C) 17...-- ]
13.bxc4 This move allows a shot which was
thought to equalize completely. In this game
Sargissian has a new idea, but it seems
rather tame.
[ 13.fd1?! xe5 14.dxe5 d4! was already
good for Black in H.Tikkanen-E.Berg,
Gothenburg 2012. ]
[ Otherwise White mostly plays 13.xc6
bxc6 and then:
A) 14.a4 xd2 15.xd2
was M.Brodsky-V.Ikonnikov, Le Touquet
2007, and now Black took on b3, but I
think f5 would have equalized.;
B) 14.c2 leaves White a tempo down on
a line covered in the notes to 11 Nxc6
above. However, as noted there, with the
queen on c2 the most logical locations for
the black rooks are d8 and c8, so Black is
perhaps slightly disorganized here.
Nevertheless, after h6 15.d2 a6
16.e3 c5 17.dxc5 xc5 , Black had at
least equalized in Zhao Xue-E.
Kovalevskaya, Women's World Blitz
Championship, Batumi 2012.;
C) 14.bxc4 dxc4 15.xc6 b4 16.ac1
, as in M.Gurevich-S.Galdunts, French
Team Championship 2006, is well met by
b6 17.d5 xd5!! 18.xd5 xd5
19.e3 b7 20.b1 a8 , regaining the
material with a decent game.;
D) 14.fd1 b4 15.xf6 gxf6 16.dc1
leads to another position where Black is a
 whole tempo up on a known line (see the
notes to 12...Bb4). After c5 (a logical
attempt to make use of the extra move)
( instead 16...a3 17.d1 b4 18.dc1
a3 was agreed drawn in Bu Xiangzhi-
Wang Rui, Beijing 2005, but White could
have played on with 19.c2 f5 20.e4
, when I haven't found a way for Black to
equalize ) 17.bxc4 dxc4 18.d5 xc3
( 18...xd5 19.xd5 xc3 was Zhu Chen-
E.Kovalevskaya, FIDE Grand Prix, Nalchik
2011, and now I like 20.xc3 xc3
21.xf7+ xf7 22.xc3 with a difficult
endgame for Black ) 19.xc3 xc3
20.xc3 xd5 21.xd5 xd5 22.xc4
b8 , a typical endgame arose in A.
Matthiesen-Kar.Rasmussen, Aarhus 2010.
Black's extra tempo isn't keenly felt but it
has enabled him to start his counterplay
earlier with ...Rb8. All the same, White
still seems a bit better.;
E) 14.-- ]
EXERCISE: Black to play. Try and find a way
to destroy White's centre.
ANSWER: 13...xd4! A great shot which
seems to equalize.
[ 13...b4 transposes into the note to Black's
12th (after Nxc6 and Bxf6) and was in fact
the move order used in several of the
games featured there. ]
14.cxd5!? A new attempt by the Armenian
Super-GM.
[ Instead, 14.xd4 dxc4 15.e3 c5
16.f4 xc3 17.xf6 gxf6 18.xf6
leads to an interesting position, but
interesting only for Black! Now, rather than
d4?! as in V.Malakhatko-M.Meinhardt,
Paris 2005 (1-0 in 39), ( Black has two better
options: forcing an immediate draw with
18...d6 (when White has nothing better
than delivering perpetual from g5 and f6);
or playing for a win by 18...e7!? 19.xe7
xe5 with an unclear game, but one where
the black c-pawn looks like a major trump. )]
14...b4
[ 14...xc3 15.xc3 xe2+ 16.h1 xc3
17.dxe6 fxe6 18.xb7 gives White
something to play for, with the bishop pair
and a superior structure. ]
15.xf6 gxf6 16.xd4 xc3 17.c4
White's only attempt for advantage.
[ 17.h4 xd5 18.xf6 leads to a draw
after xg2 19.g5+ h8 20.f6+ . ]
17...xd4 18.xa5 xa1 19.dxe6 c3
20.e7 xa5 21.exd8 xd8 22.b1
A nice touch, trying to interfere with Black's
control of c7. b6 23.c1 d7 Given that
Sargissian's day job is analysing Marshall
endgames with Levon Aronian, he must have
had some view on his winning chances here.
Certainly White can never lose and can play
for a long time. However, it is hard to see the
final result being anything other than a draw –
a result which would be immediately agreed
after an exchange of rooks.
24.h3 e7 25.e3 b4 26.c4 c5
27.f5 h6 28.g2 e5 29.g4+ f8
30.c2 d5 31.f1 d2 32.c4 d5
33.e2 h5 34.e4 e5 35.d3 h5
36.a4 a5 37.f4 g7 38.h4 d5 39.a4
d8 40.f5 d7 41.g4 e7 42.f4 c7
43.d5 f8 44.b5 c2+ 45.f3 c3
46.d7 a3 47.e4 c5 48.f5 xe3
49.xf6 e7+ 50.f5 xh4 51.b7 e6
52.c4 f6+ 53.e4 g3 54.f5 h4
55.d5
½-½
D34
Onischuk,A
Akobian,V
9: Merida 2008
[Sam Collins]
1.c4 e6 2.d4 d5 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 8.c3 0-0
9.g5 c4 10.e5 e6 11.f4 This is perhaps
the most ambitious move at White's disposal,
but Black seems to get decent counterplay.
Moving pawns in front of your own king means
that you need to be accurate for the rest of
the game.
[ 11.e3 seems to give Black several ways to
equalize:
A) 11...d7 12.xc6 bxc6 13.xe7
xe7 leaves Black with a very solid pawn
chain and no problems: 14.b3 ( 14.a4
ab8 15.c2 b4 16.b3 fb8
gave Black excellent queenside
counterplay in H.Stefansson-V.Akobian,
Lubbock 2008; 0-1 in 37 ) 14...b6
15.c2 fc8 16.fc1 ab8 17.bxc4
( 17.ab1 was V.Burmakin-E.Kanter,
Dubai 2013, and now I like g6 , preparing
...Bf5 ) 17...xc4 18.a4 g6 19.f1 b4
20.c5 was A.Adly-V.Kotronias, World
Team Championship, Bursa 2010, when
Black can hold the balance with b2
21.c3 f6 22.c2 xc2 23.xc2 f5
 24.c3 d6 , reaching a comfortable
Carlsbad structure – the black knight is
excellently placed on d6, as known from
the classic games in the Queen's Gambit
Declined Exchange Variation: M.Bobotsov-
T.V.Petrosian, Lugano Olympiad 1968
(0-1 in 41), and L.Portisch-G.Kasparov,
World Cup, Skellefteå 1989 (0-1 in 62).;
B) 11...h6 12.xf6 xf6 and now:
B1) . A great demonstration of Black's
chances is 13.xc6 bxc6 14.c2 b8
15.fd1 a5 16.a4?! xa4 17.xa4
b4 with excellent queenside play for
Black in A.Beliavsky-I.Jelen, Slovenian
Championship, Sentjur 2013.;
B2) 13.f4 e7 14.d2 ( 14.b3
, as in D.Flores-P.Barrionuevo, Moron
2012, can be met simply by c8 )
14...a5 15.c1 ad8 and Black had
no cause for complaint in A.Schenk-M.
Petr, German League 2012 (½-½ in
45).;
B3) 13.--;
C) 11...-- ]
[ The principled 11.e4? is in fact a serious
mistake.
EXERCISE: How should Black respond?
ANSWER: xe4! 12.xc6 bxc6 13.xe7
xc3 14.xd8 xd1 15.e7 ( 15.axd1
axd8 simply leaves Black a pawn up )
15...xb2! 16.xf8 xf8 (in this position
Black has two pawns for the exchange and
a serious advantage: the protected passed
pawn on c4 is a huge asset, and the white
rooks struggle for scope) 17.fc1 b8
18.f1 ( or 18.c3 a5 19.a3 a4 20.f4
e7 21.f2 d6 22.e3 c5 23.f3 e8
24.dxc5+ xc5 and White resigned in J.
Burnett-P.Kiriakov, British League 2005 )
18...e7 19.c3 a5 20.f3 d6
( the immediate 20...c5 21.dxc5 d4
was even stronger ) 21.b1 f5 22.bc1
c5 23.dxc5+ xc5 and Black was winning
in H.Callam-S.Collins, European Cup,
Rogaska Slatina 2011, since White has no
long-term defence to the plan of carefully
pushing the c- and d-pawns. ]
11...g4 This is the consistent choice of the
strongest players with Black.
[ Alternatively, 11...h6!? is again
recommended by Aagaard and Ntirlis and
has scored well. ]
12.xg4
[ 12.xe7 xe7 13.d2 doesn't pose
serious problems for Black; for instance, f6
14.xg4 xg4 15.b3 cxb3 16.fb1 d7
17.xb3 was about level in M.Krysztofiak-K.
Miton, Polish Championship, Warsaw 2011
(½-½ in 24). ]
12...xg4 13.xd5
[ As Scherbakov notes, the problem with
13.xd5 is f6! (as yet unplayed)
( all games which reached this position
proceeded with 13...xg5?! ) 14.h4
( 14.h3! d7 15.xe7+ xe7 16.d5
is better, although after c5+ 17.h2 fxg5
18.dxc6 xc6 Black is certainly not worse )
14...g5 15.h3 d7 16.xe7+ xe7 17.d5
and now Black is better after, for instance,
gxh4 18.dxc6 c5+ . ]
13...xg5 14.fxg5 xg5 15.f4! White now
has a 2-0 central majority but Black can get
rapid counterplay down the central files. ad8
[ 15...e6 is a common alternative, but I
prefer to bring a rook into play than move a
well-placed bishop. ]
16.d2 h3!? This has become established
as the main line. Black threatens the d5-
bishop and aims to provoke e2-e4, after which
the white king has even less pawn cover. The
position is sharp but I think Black has a good
share of the chances.
17.e4 e7 This looks sensible.
[ 17...e6 is also possible (since 18 Bxe6,
trying to win material, fails to 18...Nxd4!),
but after 18.af1 Black probably has
nothing better than e7 anyway. ]
18.f2
[ 18.d1! is a good move, overprotecting the
key d4-pawn. After e6 19.xc6 bxc6
20.d5 ( 20.a4 is interesting, when Black
can try f6 21.c5 f7 ) 20...b7 21.f2
d7 , White had some advantage in N.
Javanbakht-A.Ismagambetov, Al Ain 2012
(½-½ in 52). ]
18...e6 19.f1 A brave decision, leaving the
dangerous knight on the board.
[ 19.xc6 bxc6 was sounder, though Black
will always have counterplay against the
white d-pawn – the black pawn moving from
b7 to c6 has taken a lot of dynamism out of
the white centre. ]
19...b4! 20.xe6 EXERCISE: Black to play!
xe6?
[ ANSWER: 20...fxe6! would leave White
struggling to hold his centre together. ]
21.d5 Attacking the queen and hitting the a7-
pawn. Black should probably try 21...Qb6,
though the endgame is no picnic. In the game
he sacrificed a pawn for activity, but it wasn't
enough.
e7 22.xa7 d3 23.f5 g6 24.f6
xb2 25.d4 d3 26.xc4 e5 27.d4
d6 28.xd6 xd6 29.b5 d7 30.xe5
xb5 31.d4 c8 32.d6 d7 33.d1 h5
34.g2 h4 35.d3 h3+ 36.g1 a4
37.d7 d8 38.d2 c6 39.f2 b6+
40.f3 f6+ 41.e3 g5+ 42.e2 g4+
43.f2 e6 44.d5 f6+ 45.e3 b6+
46.f3 g1 47.g4 f1 48.d3 g1
49.xh3 g7 50.d4+ f6 51.xg1
1-0
 Back to Contents Page
Chapter Three

9 Bg5 Be6
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 Be6

Black develops a piece while protecting a central pawn, a concept known to be good from the moves 1 e4 e5 2 Nf3
Nc6 (and some other openings). So why does nobody play this anymore? Well, because everyone "knows" that the
line is refuted by a long forcing sequence, resulting in an endgame which Spraggett lost against Yusupov in their
Candidates match in 1989. (The same endgame was played by Spassky to make a very easy draw against
Petrosian in their 1969 World Championship match, but maybe people have forgotten about that.)
The various nuances which might affect the above assessment - Yusupov is a great player, especially strong in the
endgame, and played almost perfectly, which included deploying a strong novelty; Spraggett's defence could have
been considerably improved; and most of us aren't facing World Championship Candidates when deciding what line
to play in the next club match - don't appear to have been sufficiently influential to rescue this line. So I hope that
Kunte-Aravindh is some good PR for this line, as well as a demonstration that the technical "conversion" of this
heavily simplified position is no walk in the park for a normal GM. Incidentally, White is not forced to go into this
endgame (I'm not sure if I'd do so myself, given the chance) and his alternatives are also covered.

Key Notes

1. 9 Bg5 Be6 is considerably better than its reputation!

2. If White chooses to go for the forced line, resulting in an endgame, Black needs to be patient to hold his slightly
inferior position. It isn't obvious to me that Black's game is any worse than those enthusiastically defended by the
world elite on a daily basis.
3. White has a range of alternatives on move 11 which keep more tension. Black should be familiar with the main
themes from the 9 Bg5 cxd4 variation, as similar positions often arise here.

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5 5 Nf3 Nc6 6 g3 Nf6 7 Bg2 Be7 8 0-0 0-0 9 Bg5 Be6 – Kunte-Aravindh

Back to Contents Page

9.Bg5 Be6
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.g5 e6 . Black develops a piece while
protecting a central pawn, a concept known to
be good from the moves 1 e4 e5 2 Nf3 Nc6
(and some other openings). So why does
nobody play this anymore? Well, because
everyone "knows" that the line is refuted by a
long forcing sequence, resulting in an
endgame which Spraggett lost against
Yusupov in their Candidates match in 1989.
(The same endgame was played by Spassky
to make a very easy draw against Petrosian
in their 1969 World Championship match, but
maybe people have forgotten about that.)
The various nuances which might affect the
above assessment – Yusupov is a great
player, especially strong in the endgame, and
played almost perfectly, which included
deploying a strong novelty; Spraggett's
defence could have been considerably
improved; and most of us aren't facing World
Championship Candidates when deciding
what line to play in the next club match –
don't appear to have been sufficiently
influential to rescue this line. So I hope that
Kunte-Aravindh (Game 10) is some good PR
for this line, as well as a demonstration that
the technical "conversion" of this heavily
simplified position is no walk in the park for a
normal GM. Incidentally, White is not forced to
go into this endgame (I'm not sure if I'd do so
myself, given the chance) and his alternatives
are also covered.
D34
Kunte,A
Aravindh,C
10: New Delhi 2013
[Sam Collins]
1.f3 d5 2.c4 e6 3.g3 f6 4.g2 e7
5.d4 0-0 6.0-0 c5 7.cxd5 exd5 8.c3 c6
9.g5 e6 QUESTION: What are the pros
and cons of this move?
ANSWER: Pros: Black develops his last minor
piece and defends the crucial d5-pawn, thus
taking the sting out of White's threats of Bxf6
and d4xc5 (in some order), followed by
capturing on d5. In so doing, Black maintains
the tension, rather than resolving it with 9...c4
or 9...cxd4. As a general principle,
maintaining the tension is often the most
ambitious continuation, since resolving it
always comes with concessions (for instance,
in this particular position 9...c4 takes the
pressure off the white d4-pawn, while 9...cxd4
brings the white knight to a strong post on d4
and unleashes the g2-bishop on the long
diagonal). However, the big question when
seeking to maintain the tension is whether a
player can get away with it or not! Whether
Black can get away with such ambition is
something we will discuss in the further course
of the game.
Cons: The bishop on e6 could be classed as
passive, in that it has a defensive function (as
opposed to a more active deployment on g4,
from where it would increase the pressure on
White's d4-pawn). In addition, the bishop can
be a target for exchange – if White gets his
knight to d4, Black will need to consider the
consequences of Nxe6. This exchange, as we
have seen, is by no means only in White's
favour – he gains the two bishops, but Black's
centre is strengthened and he gets the half-
open f-file to aid his attack.
The last con in the position was hinted at
above: maintaining the tension always begs
the question as to whether a player can get
away with it. Here, White has a liquidating
sequence which leads to an endgame in which
he has scored 75% over a large number of
games – although, as we will see, I believe
this position is quite playable for Black.
10.dxc5
[ Australian GM Ian Rogers notes that,
although the Encyclopaedia of Chess
Openings gives White a small advantage
after 10.c1 , c4! leaves the rook
misplaced on c1. This looks like an
improved version for Black of the 9...c4
variation, since ...Be6 is necessary whereas
Rc1, although sometimes useful, is not
invariably required. ]
10...xc5 11.xf6 Starting a forced
sequence resulting in a slightly favourable
endgame for White.
[ However, there are several alternatives:
11.c1 e7 ( 11...b6?! 12.b3!
, threatening Bxf6 and Nxd5, leads to a
White advantage, as in the stem game A.
Rubinstein-Em.Lasker, Berlin 1918, and in
V.Korchnoi-J.Piket, 3rd matchgame,
Nijmegen 1993 ) 12.d4 h6 13.e3 d7
14.a4 fc8 ( or 14...h3 , as in V.
 Korchnoi-J.Nunn, Hastings 1975/76, "with
an acceptable game" – Kasparov ) 15.fd1
g4? was Kir.Georgiev-E.Berg, Gibraltar
2008, and now 16.xe6! would have won
material, but 15...Bh3 would have been
fine. ]
[ 11.e1 d4 12.xf6 xf6 13.e4 e7
14.xc5 xc5 led to success for White in a
game between the inventors of both the
Tarrasch and the g3-system against it: A.
Rubinstein-S.Tarrasch, Carlsbad 1923 (1-0
in 45). Nevertheless, Black looks fully equal
here. We will see more examples of how to
handle this structure in the chapter on 9
dxc5. ]
[ 11.a4 was played in M.Tal-P.Keres, Bled/
Zagreb/Belgrade Candidates 1959,
A) , where Keres won a famous game
after 11...b6 12.xb6 axb6 .;
B) . On 11...e7
B1) Kasparov also gives the variations
12.d4 xd4 13.xd4 a5 14.fc1
( or 14.c3 ac8! and Black is fine )
14...fc8;
B2) 12.e3 , Keres suggested that
"White could have created strong
pressure on the c5-square and secured
an enduring initiative without any risk". In
fact this position more commonly arises
with White to move (in the variation 9
dxc5 Bxc5 10 Na4 Be7 11 Be3 Be6), so
it is not surprising that Black, a whole
tempo up on a known variation, has few
difficulties here. . After e8 ( 12...g4
gets Kasparov's endorsement: 13.c1
e8 14.c5 xc5 15.xc5 e4
16.e3 d7 with level chances in E.
Geller-B.Spassky, 7th matchgame, Riga
1965; ½-½ in 41 ) 13.d4 g4 14.xc6
bxc6 15.c1 d7 16.e1 and Black
had no problems in L.Guidarelli-D.
Brandenburg, Vlissingen 2010 (½-½ in
31).;
C) 11...-- ]
11...xf6 12.xd5 Neither player has a
decent way to avoid the following forcing line.
xb2 13.c7 ad8 14.c1 xc1
15.axc1 e7
[ Alternatively, 15...b6 16.xe6 fxe6
is an endgame which has been extensively
defended by German GM Lutz Espig.
According to my database, he has made a
remarkable +2 =9 -0, which is rather
amazing since Black should never win such
a position! The most famous example here
is the 16th game of the 1969 Petrosian-
Spassky World Championship match in
Moscow, where after 17.c4
( several commentators, including Yusupov,
have suggested that White retains better
chances of an edge with 17.e3!?; or
17.h3!? ) 17...d4! 18.xd4 xd4
19.xd4 xd4 , the trade of a pair of rooks
and knights had considerably eased Black's
defence. The game concluded 20.e3 c5
21.d1 f7 22.e4 g6 23.h4 c7 24.g2
e7 25.f3 g7 26.d2 b4 27.c2
xc2 28.xc2 f6 29.d3 h6 30.e4 g5
31.h5 e1 and a draw was agreed. ]
16.xe6 fxe6 QUESTION: What's your
assessment of this endgame?
ANSWER: One thing is clear – Black can't be
better! White has the superior pawn structure
which guarantees him a small but enduring
advantage. The stats from this position are
strongly in White's favour, suggesting that his
position is considerably easier to play.
On the other hand, the game has been
considerably simplified. While Black,
admittedly, has a weakness on e6 his position
is otherwise quite solid, and the "principle of
two weaknesses" suggests that White will
need to create another target before
generating serious winning chances. The
presence of opposite-coloured bishops is also
a potentially significant factor: if the rooks and
knights are traded the players would agree to
an immediate draw. However, it is important
not to overestimate the drawing tendencies of
the opposite-coloured bishops – if White
manages to generate an initiative on the light
squares it will be hard for Black to defend.
Black has no f7-pawn, and if the pawn on e6
is lost or moved to e5, Black will be severely
weakened on the a2-g8 diagonal which the
white bishop could likely occupy (for instance,
Bd5, anchored by a white pawn on e4, is
possible). Conversely, White would need to be
very careless to allow Black similar chances
on the a7-g1 diagonal, since white pawns on
f2 and e3 would stunt the black bishop's
activity.
We should also mention that White has a
kingside majority and might consider
mobilizing it with f2-f4 and e2-e4, although,
for the reasons given above, I would be very
careful to check that Black can't generate play
on the dark-squares before moving my f-pawn.
17.c4! QUESTION: Why is this a strong
move?
ANSWER: Nimzowitsch defined a strong
positional move as one which advances your
own plans while preventing your opponent's.
This move ticks both boxes. The rook is well
placed on c4, where it can move along the
fourth rank to target the black pawns (by Re4,
for instance). The way is also cleared for
White's king's rook to come to b1 or c1. In the
meantime, the rook prevents one of Black's
ideas, namely further simplification with ...Nd4,
after which he would be one trade closer to a
drawn opposite-coloured bishop endgame.
The most famous game to feature 17 Rc4! is
probably Yusupov-Spraggett, Candidates (3rd
matchgame), Quebec 1989, which ended in a
win for White after 42 moves. However,
Yusupov has tremendous technique even by
GM standards, and the line hardly deserves to
be consigned to the dustbin just because he
won a good game. Chess fashion is rather
fickle and variations are often discarded
simply because, for instance, Karpov won a
precise technical game. Positions which are
unplayable against the elite can often be used
with success against lesser mortals and even,
as in this game, against good grandmasters.
f6 18.h3
[ Instead: 18.b1 is one of the most popular
moves in this position. However, after d7
( 18...d6!? is also possible since the pawn
on b7 is poisoned. Black has good
equalizing chances here, as for instance in
A.Beliavsky-M.Illescas Cordoba, Linares
1988 (½-½ in 50) ) 19.h4 , Black can follow
Spassky's idea from his match against
Petrosian and simplify with d4!?
, as in L.Ftacnik-N.Minev, Bucharest 1978. ]
[ 18.e3! was the novelty Yusupov unleashed
against Spraggett in Game 3 of their 1989
Candidates match. Yusupov explains the
idea behind the move as follows: "Realizing
the disadvantages of exchanging rooks,
White doesn't hurry to play Rb1. In fact, it is
not yet clear whether this move is needed at
all (the rook may turn out to be more useful
on c1). The move played, which I had
prepared to use against Illescas [in Linares
1988 – SC], proved to be a theoretical
novelty, although hardly anything could be
more logical – White takes control of the d4-
square, on which unwelcome exchanges
could otherwise take place."
The game continued d6 19.h4 h6 20.e4
fd8 ( a more recent attempt also ended
badly for Black: 20...b5 21.c1 a6 22.h3
f7 23.xe6 xe6 24.xe6+ xe6
25.xc6+ d5 26.xa6 and White was
winning in S.Brunello-D.Contin, Italian
Championship, Martina Franca 2007; 1-0 in
46 ) 21.h3 f7 22.g2 e8 23.c1 e7
24.c2 b6? this move significantly weakens
the c6-knight, and it is indicative of how
difficult it is to play such passive positions
that someone of Spraggett's class would
make such a mistake; ( instead, Rogers
suggests 24...g6 , intending to mark time
with ...Kg7-f7 and challenge White to find a
way to break through ) 25.f4! (creating an
immediate threat of Rxc6 and Ne5+) g6
(Rogers says that moving the king to e8 or
g8 would have been better, although White
is still on top after 26 g4) 26.g4! a1
( Yusupov gives 26...e5 27.g5 hxg5
28.hxg5 d3 29.gxf6 xf4+ 30.exf4
with a clear advantage to White, for
instance gxf6?! 31.f5+! ) 27.c1 b2
28.c2 a1 29.a4 e5 30.xe5+ xe5
31.f8 (White now has a devastating attack
on the black king) dd7 32.f4 c7 33.d2
c3 34.d6 h7 35.g5 hxg5 36.hxg5
b4
A) Yusupov notes that 37.g6+!
was immediately decisive: h6 ( or
37...xg6 38.f5+ ) 38.f5!;
B) 37.dd8 g6 ( 37...g6 is better )
38.f3 f7 39.h8 e5 40.g4 exf4
41.d5 fxe3+ 42.g3 and Black resigned
in view of imminent mate. ]
18...d6 19.b1 f7 20.e4 e7 21.a4 g6
22.e3 g7 23.f1 d8 That's some solid
defence of e6! As mentioned in previous notes,
White needs to generate another target to
create genuine winning chances.
24.h4 h6! There is no need to allow the white
knight access to g5. 25.c4 c6 26.e4
d8 27.c1 c6 28.xc6 bxc6 Rather a
committal decision, albeit one which worked
out well in the game.
[ I don't see anything wrong with the
straightforward 28...xc6 . Every trade
favours Black and he still seems rock-solid
here, even if White can play on with no
risk. ]
29.d3 f7 30.c4 c7 EXERCISE: How
can White improve his position?
31.d4?! As will be seen in the game, this
allows a pretty clear equalizing sequence.
[ ANSWER: Pretty much any neutral move
would be okay, but White had a couple of
ideas to maintain the pressure: 31.g4 g5
 32.hxg5 hxg5 33.c4 and White is advertisement for the solidity of this system.
obviously better here, although in view of ½-½
the simplified position and the drawish
properties of the remaining pieces, my
money would be on Black holding on for a
draw. ]
[ 31.c5!? , taking control of some useful
squares on the fifth rank and preparing Be4.
]
[ The immediate 31.e4 isn't so convincing
because of c5 , and Black is fine. ]
EXERCISE: How can Black bring the position
to safety?
ANSWER: Part of the technique of defending
passive, slightly worse positions is being able
to accurately calculate when you can
transform the position, for instance through a
trade or a bid for counterplay.
31...xd4! Black is alert and realizes that he
can either get counterplay or trade off his
weak e6-pawn.
[ Admittedly, in view of the fork by the knight
on d4, Black didn't have very much choice
since 31...d8 was not attractive. ]
32.exd4
[ 32.xd4 is the move White wants to play,
but then Black gets counterplay with the c-
pawn; i.e. c5! 33.e4 f6 and Black is fine,
given his ideas of ...g6-g5, ...Nd6 and, if
appropriate, ...c5-c4 (though this last move
needs to be considered carefully, since the
additional threats created by pushing the
passed pawn must be balanced against the
pawn's increasing vulnerability and the
weakening of the d4-square). White has
nothing better than the repetition available
after 34.f4+ g7 35.e4 f6 . ]
32...e5 33.e4 The sensible move.
[ After 33.dxe5 xe5 34.c3 xd3
35.xd3 c5 , if anyone has chances in the
resulting rook and pawn endgame, it's not
White. ]
33...exd4 34.xc6
[ If White wanted to play on, his last chance
was 34.xd4 . The computer proposes that
White still has an edge, though I think Black
isn't worse at all, given his passed pawn and
the simplified nature of the position. After g5
35.hxg5 hxg5 White has no reason to claim
any advantage. ]
34...xc6 35.xc6 g5! Trading off more
pawns brings the position closer to safety.
36.hxg5 hxg5 37.f4 gxf4 38.gxf4
. A very solid draw with Black, despite the
150+ rating point gap, and a good
 Back to Contents Page
Chapter Four

8 dxc5/9 dxc5
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7

Systems with d4xc5 at move 8 or 9 have a number of attractions from White's perspective. White defines the
structure and saddles Black with an IQP, and accordingly does not need to worry about alternative structures such
as those which arose in the 9 Bg5 c4 line. Similarly, Black's choice is somewhat limited and so the theoretical
workload required to play 9 dxc5 is much less than that involved in 9 Bg5. White can take on c5 at any stage after
move three.
8 Nc3
The alternative 8 dxc5 Bxc5 can result in a transposition after 9 Nc3 0-0, but also has independent significance
after White's alternatives such as 9 a3, which is covered in Manolache-Jianu.
8...0-0 9 dxc5 Bxc5
9...d4!? is an enterprising gambit continuation which I have tried before and might try again. See
Chatalbashev-Sokolov for the details.
10 Bg5
This was historically the main continuation but now, following recommendations of authors like Avrukh, you are just
as likely to face an alternative here such as 10 Na4 or 10 a3. See Akopian-Jianu for the details.
10...d4 11 Bxf6
11 Ne4 (and the similar 11 Bxf6 Qxf6 12 Ne4) are examined inBaburin-Collins.
11...Qxf6 12 Nd5

This is the old main line and is covered in Mulyar-Perunovic, the first game below.

Key Notes

1. Capturing on c5 is a principled decision and forces Black into an IQP structure.

2. In the main line with 10 Bg5 d4 11 Bxf6 Qxf6 12 Nd5, White hopes to obtain a slight advantage after solidly
blockading the d4-pawn and playing on the queenside. Black has a wide choice as to how to respond and, in
particular, whether to retreat his bishop to b6, a7 or f8.
3. 11 Ne4 or 11 Bxf6 Bxf6 12 Ne4 leads to quiet and balanced positions. Black should be sure to be well prepared
here since, as my game with Alex Baburin demonstrates, thoughtless play will lead to a passive and inferior position.
4. 10 Na4 is a recently popular try. I like Jianu's simple solution of ...Bg4 and taking on f3, which has scored well
4. 10 Na4 is a recently popular try. I like Jianu's simple solution of ...Bg4 and taking on f3, which has scored well
against very strong players.
5. a2-a3 on move 9 or 10 has the idea of queenside expansion with b2-b4; here ...Ne4! is a good response, even
as a pawn sacrifice.
6. I'm not sure of the objective merits of the gambit variation inChatalbashev-Sokolov. From a practical perspective,
however, the line is fully playable and gives Black dangerous compensation.

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 Nc3

8 dxc5 Bxc5 9 a3 – Manolache-Jianu
8...0-0 9 dxc5 Bxc5
9...d4!? – Chatalbashev-Sokolov
10 Bg5
10 Na4 or 10 a3 – Akopian-Jianu
10...d4 11 Bxf6
11 Ne4 – Baburin-Collins
11...Qxf6 12 Nd5 – Mulyar-Perunovic

Back to Contents Page

8 dxc5/9 dxc5
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7
Systems with d4xc5 at move 8 or 9 have a
number of attractions from White's perspective.
White defines the structure and saddles Black
with an IQP, and accordingly does not need
to worry about alternative structures such as
those which arose in the 9 Bg5 c4 line.
Similarly, Black's choice is somewhat limited
and so the theoretical workload required to
play 9 dxc5 is much less than that involved in
9 Bg5. White can take on c5 at any stage
after move three.
8.c3
[ The alternative 8.dxc5 xc5 can result in a
transposition after 9.c3 ( but also has
independent significance after White's
alternatives such as 9.a3 , which is covered
in Manolache-Jianu (Game 15) ) 9...0-0 . ]
8...0-0 9.dxc5 xc5
[ 9...d4!? is an enterprising gambit
continuation which I have tried before and
might try again. See Chatalbashev-Sokolov
(Game 14) for the details. ]
10.g5
[ This was historically the main continuation
but now, following recommendations of
authors like Avrukh, you are just as likely to
face an alternative here such as 10.a4 ]
[ or 10.a3 . See Akopian-Jianu (Game 13)
for the details. ]
10...d4 11.xf6
[ 11.e4 ]
[ and the similar 11.xf6 xf6 12.e4
are examined in Baburin-Collins (Game
12). ]
11...xf6 12.d5 . This is the old main line
and is covered in Mulyar-Perunovic, the first
game below.
D34
Mulyar,M
Perunovic,M
11: Chicago 2012
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 8.dxc5
xc5 9.c3 0-0 10.g5 d4 11.xf6 xf6
12.d5 d8
[ 12...f5!? is a pet line of Narciso Dublan
and has been adopted by current Tarrasch
exponents like Aagaard, Swinkels and
Carlstedt. However, I think 12...Qd8 is
more reliable. ]
13.d2 An idea of Jan Timman, as far as I'm
aware. White intends Rc1, putting the
QUESTION to the c5-bishop. Black has a
choice of three squares – f8, b6 and a7. Each
option seems fully playable to me.
e8 The rook is always good on this square
and particularly so in this line, since there is
no knight on c3 to hold the e2-pawn. Tactics
against e2 are much more common here than
in the 9 Bg5 variations.
[ 13...a6 14.c1 a7 is the main alternative,
after which White has a wide choice:
A) 15.b3 d6 16.d2 e8 17.f4
f5 with level chances in J.Plachetka-J.
Nunn, European Team Championship,
Skara 1980 (½-½ in 39).;
B) 15.c4 g4 16.d2 c8 17.h3 e6
18.f4 f5 19.a4 was P.Wells-J.Cooper,
British League 2006, and now the simple
e8 20.b4 d7 equalizes.;
C) 15.a3 g4 16.f4 c8 17.d3 e8
18.f3 d7 and Black had no problems in
E.Matsuura-C.Umetsubo, Maringa 2012
(½-½ in 41).;
D) 15.b4 g4 16.e4 e8 17.e1 e5!
18.f4 e7 and White was struggling to
meet all the threats in J.Leiva Rodriguez-C.
J. Ruiz, Lima 2013 (0-1 in 40).;
E) 15.e4 h8 16.f4 e7 17.d3
f5 18.ec5 ac8 19.d2 fe8
20.fe1 d8 21.b4 a5 and a draw was
agreed in R.Khusnutdinov-L.D.Nisipeanu,
Pardubice 2012.;
F) 15.f4 is the main line and has scored
well for White in practice. However, the
computer thinks this position is completely
equal:
F1) 15...g4
F1a) 16.e1 c8 17.e4
was V.Epishin-G.Schnur, Werther
2000, when b8 18.c5 xf4
19.gxf4 e7 20.a3 a5 leaves Black
no worse – White also has to take care
of his own weaknesses here;
F1b) , instead, 16.f3 xf3 17.xf3
has been played twice by Petr Haba –
in both games Black opted for the
pseudo-active a5 ; ( a beginner
would put rooks on the open files with
17...e8 18.b3 d7 19.fd1 ac8
which, fortunately, completely
 equalizes );
F1c) 16.h3 f5 17.b3 b8 18.g4
( 18.a3 was P.Brodowski-M.Chlost,
Olomouc 2010, and now it seems
Black can get away with g5!? 19.d3
e8 20.fe1 e7 21.xe7 xe7
with no problems in the endgame )
18...g5 19.g3 e6 20.xe6 fxe6
21.e4 b5 22.b3 e5 23.h1
f4 and Black was no worse in R.
Kasimdzhanov-A.Mallahi, Asian
Championship, Tehran 1998.;
F2) 15...e7 led to a draw by repetition
after 16.d5 d8 17.f4 e7
in A.Moiseenko-R.Ponomariov,
Ukrainian Team Championship 1998.;
F3) . The prophylactic 15...b8
has been played numerous times and
should be fine for Black. However, I
prefer 15...Bg4 since the queen's rook
might find a better square on c8 or d8.;
F4) 15...--;
G) 15.-- ]
14.c1
[ One of my happiest memories with the
Tarrasch proceeded 14.b3 f8 15.f4
g4 16.xc6 bxc6 17.xd4 xe2
( 17...xd4 18.xd4 g5 19.h3 g7
20.ad1 c8 21.h5 is level ) 18.xe2
xe2 19.c3 b6 20.ad1 ae8 21.d7
8e7 22.d4?! ( 22.d6 held the balance )
22...g6 23.fd1 h5! (White's position is
unpleasant since Black intends 24...Bg7; I
turned down a draw somewhere around
here but was surprised by how quickly
White's game went downhill) 24.f3 xb2
25.d8 ee2 26.1d7 xf2 27.xf8+
xf8 28.xf7+ g8 29.xf2 xf2 30.xf2
e3 and Black soon won in A.Baburin-S.
Collins, Dun Laoghaire 2010 (0-1 in 35). ]
14...b6
[ 14...f8 is also fully playable and has been
endorsed by Grischuk amongst others. ]
15.c4
[ 15.e1 is another logical move:
A) 15...g4 16.f4 ( 16.c4 a5
transposes to the main game ) 16...d3!?
17.xd3 was tested in two online games L.
Van Wely-T.L.Petrosian in the European
Blitz Championship 2004. Game 3
proceeded xe2 ( an improvement over
the first game's 17...xe2 18.xe2 xd3
when White was clearly better; but
17...d4! 18.f1 e7 seems stronger,
with a good game for Black ) 18.xe2
xd3 19.xe8+ xe8 20.c4 xd1+
21.xd1 and White enjoyed a small edge
in the endgame (though 0-1 in 51).;
B) . In a game in California I mixed up my
move order with 15...a5?! 16.a3! e6
17.f4 g4 18.h3 ( 18.b4 c7 19.d3
seems better, with an edge for White,
although this should be playable for
Black ) 18...d3? ( simply 18...f5
would have led to a normal position )
19.hxg4 dxe2 20.xe2 xe2 21.xe2
xd2 22.xd2 xd2 .
QUESTION: How do you assess this
endgame?
ANSWER: In my analysis I had assumed
that the endgame, with equal (and
reduced) material and opposite-coloured
bishops, must be level. In fact Black is in
for a difficult defence, since all the white
pieces are more active than their black
counterparts. This game was a salutary
lesson for me in the Tarrasch. When you
trade the isolated d-pawn, there is a
natural tendency to relax since Black's
position has no formal weaknesses
anymore. However, it is vital to check
whether the opponent's activity can be
neutralized. In particular, and especially
with a white bishop on g2, the b7-pawn is
a key target which can be difficult to
defend. D.Zilberstein-S.Collins, Berkeley
2011, continued 23.c2 d8 ( 23...xf4
24.gxf4 c8 was a better try ) 24.xc6
bxc6 25.xc6 g6 26.a6 d7 27.f1
with an extra pawn and a clear advantage
for White. I eventually lost a painful rook
endgame (1-0 in 65).;
C) 15...-- ]
15...g4 16.e1 a5
[ 16...c5 is an interesting alternative,
preserving the dark-squared bishop after
17.a3 a6 18.b4 a7 . Now 19.d2 c8
20.a4 e6 ( 20...h5! 21.f4 h4 22.d3
e6 gave counterplay in D.Yevseev-V.
Yemelin, St. Petersburg 2006; ½-½ in 49 )
21.f4 xc4 22.xc4 e5 23.xc8 xc8
24.c1 d7 is a logical sequence leading to
a position which has been successfully
defended by a couple of Tarrasch experts:
25.c2 d3 ( 25...b8 26.h3 b5 27.axb5
axb5 was L.Ftacnik-A.Berelowitsch, German
League 2003, ½-½ in 50, but now 28.c5
would have been better for White ) 26.exd3
( and here 26.xd3! xd3 27.xd3 xa4
28.c7 keeps more pressure ) 26...d4
 27.h3 g6 . This position is tough for White
to win, especially against Grischuk, and a
draw eventually resulted in L.Van Wely-A.
Grischuk, Enghien les Bains 2001 (½-½ in
53). Nevertheless, Black was under
pressure in these games after 20...Be6, and
I prefer Yemelin's more purposeful approach
with 20...h5!. ]
17.xa5 xa5 EXERCISE: How can White
fight for an advantage?
ANSWER: 18.b4! This shot is a key resource.
The white pieces gain a lot of activity and
Black needs to be careful not to lose his
whole queenside. xb4
[ 18...xa2? was refuted the first time it was
played, which hasn't stopped this error
being repeated in over a dozen games since.
A.Chernin-G.Giorgadze, Tallinn 1984,
continued 19.c7 d3 20.xd3 ad8
21.d5! xe2 22.h3!! xe1+ ( Scherbakov
gives 22...e5 23.c3 xd5 24.xd5
which wins, since Black has no time to
recapture in view of his porous back rank )
23.xe1 e6 24.e4 xd5 25.xd5 b2
26.f4! xb4 27.xb4 xb4 28.d1!
c6 29.xe6 fxe6 30.d7 and after a
series of incredibly accurate moves, White
went on to win the endgame (1-0 in 49). ]
19.d2 Perhaps not the most dangerous
move, though Black needs to be well prepared
here.
[ 19.xd4 xd5
A) 20.xd5 xd5 21.xd5 ad8
22.xb7 xe2 23.xe2 xe2 led to easy
draws in A.Miles-P.Konguvel, Sakthi 1996
(½-½ in 28) and E.Prokopchuk-V.Potkin,
Russian Cup, Moscow 1999 (½-½ in 38);
B) , but 20.xg4 leads to a position which
most GM commentators seems to agree is
good for White.
Milov: "This position is slightly better for
White due to the superiority of bishop over
knight. [The] knight is not stable in the
centre and soon will have to retreat."
Atalik: "Although theory is counting on this
position like it is equal, in reality White is
clearly better due to the better piece and
the liquidability of [the] queen's flank
pawns."
Black usually plays ad8 ( 20...f6
21.c4 ab8 led to a quick draw in E.
Ghaem Maghami-S.Mamedyarov, Baku
2003, ½-½ in 26, but this approach is
passive and White must be better here )
B1) 21.f3 f6 22.c4 b6 23.ed1
g6 led to draws in E.Inarkiev-V.Potkin,
Elista 2001 (½-½ in 39) and I.Saric-O.
Zierke, Pula 2007 (½-½ in 44).;
B2) 21.ed1 f6 ( not 21...xa2??
22.d4 with a fatal pin; while Milov
claims an edge for White after 21...c3
22.xd8 xd8 23.c4 )
B2a) 22.c4 b5 23.c2 is a position
which should certainly be tenable for
Black, but isn't much fun. Many strong
GMs have happily gone for this as
White and have caused a lot of
problems. For example: h6 ( 23...h5
24.f3 h4 25.g2 hxg3 26.hxg3
xd1 27.xd1 was better for White in
V.Milov-M.Haag, Suncoast 1999; 1-0
in 37 ) 24.f3 xd1+ ( here 24...a6
25.xd8 xd8 26.g2 a3 27.h4
g6 28.c7 d6 29.xd6 xd6
30.e3 g7 31.c2 d7 32.e2
b6 was agreed drawn in J.Nogueiras
Santiago-R.Vera Gonzalez Quevedo,
Linares, Holland 1993, but White
would have been clearly better had he
avoided the queen exchange with 29
Qb7 or 29 Qa7 ) 25.xd1 d8 ( or
25...b4 26.d6 d8 27.c6 e5
28.g2 and Black was in for a difficult
defence in S.Halkias-V.Akobian, World
Junior Championships, Yerevan 2000 )
26.c1! (as explained by Atalik in a
different variation, White should keep
the rooks on in this position) b4
( 26...d2 is worth considering )
27.c7 xc7 28.xc7 d7 29.c8+
h7 30.c5 d1+ 31.g2 a1
32.c2 g5 with equality in R.
Kasimdzhanov-J.Bosch, Hoogeveen
1999.;
B2b) . The direct 22.xd8 xd8
23.c8 gave White an extra pawn
after h6? in J.M.Hodgson-A.
Summerscale, French Team
Championship 1999 (1-0 in 95) ( and
23...g6? in S.Atalik-V.Akobian, Los
Angeles 2003 (1-0 in 26).; Instead,
23...b5 is better, keeping equal
material, though I still prefer White in
this position since (amongst other
reasons) I prefer to capture queens
when they land on my back rank! );
B2c) 22.-- and now:;
B3) 21.c4 gives Black a choice:
B3a) 21...g6 was successful in a
recent game. 22.b3 ( 22.ed1
 is a better attempt ) 22...b4 23.f1
xa2 24.a1 d2 25.xb7 b6
26.xb6 axb6 27.ab1 exe2
28.xe2 xe2 29.xb6 and a draw
was agreed in L.Ftacnik-P.Dias,
Portuguese Team Championship
2012.;
B3b) 21...b5 22.b3 f6 23.c6
e5 and now:
B3b1) 24.ed1 d2
( Atalik assesses 24...xd1+
25.xd1 h6 26.f3 c5 27.g2
as good for White ) 25.xd2 xd2
26.d1 xe2 27.d8+ e8
28.g2 h6 ( Atalik gives the nice
variation 28...a5 29.c3 e6
30.d7 e7 31.c8 e4+ 32.h3
f8 33.c5! g8 34.xe7 xe7
35.xe8+ and wins ) 29.a8
and White was better in K.Spraggett-
K.Gentes, Canadian Championship,
Brantford 1999 (1-0 in 41).;
B3b2) 24.b2 c5 25.f3
was S.Atalik-J.Prizant, Vladimir 2002,
and now Atalik notes that dc8!
would have equalized, since Black
has no problems after all the rooks
are exchanged: 26.g2 xc1
27.xc1 xc1 28.xc1 g6
with level chances.;
B3b3) 24.--;
B3c) 21...--;
B4) 21.-- and then:;
C) 20.-- ]
19...c6 20.xa5 By far the most popular,
but probably not the best.
[ Instead, 20.g5!? is an underrated idea:
d3 21.f6+ h8 ( after 21...f8
, as in A.Fominyh-B.Miljanic, Budapest 1990,
White should have interposed 22.xh7+
g8 23.f6+ f8 before trading queens )
22.xa5 xa5 23.xe8 d2 24.d6 e6
25.a3 d8 26.xb7 xb7 27.xb7
left White with an extra pawn in M.
Hörstmann-J.Blauert, German League 2004.
Black has reasonable saving chances after
dxe1+ 28.xe1 d2 29.a6 a2 30.d1
g5 31.d3 , but only White can try to win
this position. ]
20...xa5 21.c7
[ 21.f3 d3 22.ed1? ( after 22.a1 d7
a draw was agreed in A.Yusupov-M.
Petursson, Reykjavik 1985 – a pretty easy
draw with Black against the future world
number three ) 22...dxe2 23.e1 ac8
24.c7 e6 25.e7+ xe7 26.xe7 f8
27.xe6 fxe6 28.xe2 with an extra pawn
for Black in A.Dreev-B.Ivanovic, Manila
Interzonal 1990 (½-½ in 50). ]
EXERCISE: How can Black stay in the game?
ANSWER: 21...d3! The only move, but fully
sufficient. 22.exd3
[ Alternatively: Ftacnik gives 22.b1 d2!
with counterplay. Then 23.ed1 xe2
24.xd2 ad8 25.bb2 xd2 26.xd2
c4 27.xe8 ( after 27.d4 b8
White should bail out with 28.e4 d3
29.d4 e2 ) 27...xd2 was drawn in K.
Spraggett-M.Leski, San Francisco 1987. ]
[ 22.xe8 d2 23.d6 has also been played.
After d8 24.xb7 xb7 25.xb7 xe2
26.ed1 dxc1 27.xc1 , the players can
shake hands with a clear conscience. ]
22...xe1+ 23.xe1 d8 24.h3
[ 24.d4 f8 25.d5 was played in T.Fiedler-
N.Michaelsen, German League 2002, when
the cleanest draw is c8! 26.d6 c4
27.xb7 b8 28.e4 xd6 29.xg4 xb7
30.d4 xc7 31.xd6 ]
[ while after 24.b5 g6 25.xa7 xd3
Black was at least equal in A.Frois-M.
Illescas Cordoba, Saragossa 1991 (½-½ in
47). ]
24...d7 The ending is level. In the game
White loses the thread and goes on to lose.
25.d4 f8 26.d5?! c4! 27.f1 d6
28.e6+ xe6 29.dxe6 e8 30.c1 xe6
White has decent drawing chances despite his
pawn deficit, but didn't manage to hold on.
31.g2 e2 32.a4 e8 33.b1 b6 34.a5
bxa5 35.b8+ d7 36.a8 e1+ 37.f1
c8 38.g2 xf1
0-1
D34
Baburin,A
Collins,S
12: Bunratty 2011
[Sam Collins]
1.d4 e6 2.c4 d5 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 8.c3 0-0
9.dxc5 xc5 10.g5 d4 11.e4
This is a quiet sideline, but one which has
appealed to a number of GMs who are simply
looking for a normal, playable game against
the Tarrasch. Being able to handle the quiet,
level, simplified positions arising in this
variation is a key skill for Tarrasch players,
since if Black is inaccurate then he might end
up worse in view of his compromised structure.
Accordingly we will devote some attention to
how Black should play here.
[ 11.xf6 xf6 12.e4 e7 13.xc5
xc5 is another quiet way for White to play:
A) 14.b3 e8 15.fe1 e7 16.ac1
a5 17.a3 g4 18.h3 h5 19.g4 g6
20.h4 d8 21.ed1 g5 22.xg6
hxg6 23.d3 e5 24.g3 f6
was fine for Black in A.Onischuk-V.
Akobian, Montreal 2009 (½-½ in 54).;
B) 14.c1 b6 15.d1 d8 16.c2 h6
17.d2 e6 18.b3 d7 19.ad1 ad8
was balanced in N.Vasovski-G.M.
Todorovic, Skopje 2011 (0-1 in 57).;
C) 14.d2 was tried against me in a very
recent game: d8 ( perhaps 14...g4
15.fd1 ad8 is simpler, as in P.Bailet-V.
Jianu, Rennes 2013; ½-½ in 40 ) 15.fd1
e6 16.f4 h6 17.a3 a5 18.ac1 b6
19.d2 d5 ( 19...a4 is sensible ) 20.e1
f5 21.e4 and now, instead of the d8
of B.Socko-S.Collins, Riga 2013 (which
still left me with a decent position),
( I should have played 21...b5
– the b2-pawn is falling and White needs
to generate something; for instance,
22.d3 f5 23.f4 xd3 24.xd3 d8
25.e4 xb2 26.f3 f6 27.f5 e5
28.h7+ f8 and White's compensation
is only enough for a draw. );
D) 14.c1 b6 and now:
D1) 15.c2 h6 16.fd1 ( 16.a3
, as in O.Cvitan-T.L.Petrosian,
European Championship, Warsaw 2005,
is best met by e6 17.b4 a5
with equality; while 16.d2 e6 17.c4
a6 18.b3 fd8 was already equal in D.
De Vreugt-A.Berelowitsch, Bussum 2008,
½-½ in 40 ) 16...e6 ( 16...e8 17.d2
e6 18.b3 ad8 was level in B.
Kurajica-H.Velchev, European
Championship, Plovdiv 2010; ½-½ in
46 ) 17.b3
D1a) or 17...ad8 18.d2 b4
19.d1 a5 20.h3 ( while 20.xd4
xa2 is just equal ) , as in F.Berkes-A.
Graf, Leipzig 2002, and now 20...a4!
21.bxa4 xa2 22.b1 a7
gives Black good counterplay;
D1b) 17...fd8 18.d2 a5 ( or
18...ac8 19.b2 c7 20.e1 d5!
21.xd5 xd5 22.f3 cd7
and Black was very solid in B.Lajthajm-
R.Sertic, Paracin 2011; ½-½ in 65 )
19.h3 ac8 20.b2 c7 and Black
was fine in I.Radulov-A.Cioara, Sunny
Beach 2012.;
D2) 15.d2 e6 16.b3 h6 17.fd1
ad8 18.b2 d7 19.d2 fd8
20.cd1 a5 and Black had established
his ideal defensive set-up, with no
problems in R.Barcenilla-J.Sadorra,
Philippines Championship, Boracay
Island 2012 (½-½ in 65).;
D3) 15.--;
E) 14.a3 e8 ( alternatively, 14...f5
15.c1 b6 16.d2 fe8 17.c4 d8
18.b4 d7 19.e1 h3 20.xh3 xh3
21.d3 ad8 was fine for Black in B.
Kurajica-D.Larino Nieto, Malaga 2009, 0-1
in 62; or 14...d8 15.e1 e7 16.d3
g4 17.e1 d6 18.c1 e8
and Black had no problems in N.Kalesis-S.
Halkias, Greek Team Cup 2004, ½-½ in
27 ) 15.b4 b6 16.d3 g6 17.fd1 f5
18.d2 was A.Astaneh Lopez-S.Collins,
Bunratty 2012.
EXERCISE: How should Black continue?
ANSWER: Instead of the premature d3?
, after which I had to struggle to make a
draw (½-½ in 65), ( I could have obtained
a pleasant position with the simple
18...ad8 .
At the risk of repeating myself, the
Tarrasch is a classical opening and
responds well to classical ideas like rapidly
and actively developing pieces. If you look
at Aagaard's games, this is basically all
he's doing, which is sufficient for a 2700
performance. 18...d3? is a typical error, a
result of looking for a tactical continuation
to justify Black's inferior structure, when
in fact the justification for my structure is
that I have tons more space in the centre
and all my pieces are more active than
their counterparts. );
F) 14.-- ]
11...e7 12.xf6
[ 12.xf6+ xf6 13.xf6 ( 13.d2 h6
14.xf6 xf6 15.fd1 d8 16.b4 a6
17.b2 f5 was very comfortable for Black
in J.Murey-S.Vysochin, Cappelle la Grande
2003; 0-1 in 39 ) 13...xf6 . White can't
even pretend to have an advantage in this
position. 14.b4 ( or 14.e1 e8 15.d3
f5 16.c1 e4 17.xe4 xe4 18.c5
e7 19.c2 ae8 and Black was at least
equal in A.Wirig-M.Kazhgaleyev, French
 Team Championship 2003; 0-1 in 27 )
14...d8 15.d2 g4 16.h3 ( spotting the
threat: 16.ac1?? d3 and White resigned
in A.Wohl-M.Kanep, Gibraltar 2011 )
16...e6 17.fd1 c4 18.b2
was R.Swinkels-A.Berelowitsch, Dutch
Team Championship 2013, and now Black
has several good ways to play – I like b5
, cementing Black's control over c4, when he
is certainly not worse. ]
12...xf6 13.e1 White has a wide choice
here. In all lines Black should remain
confident in his position and try to develop his
remaining pieces.
[ 13.c1 e8 14.e1 f5 15.c5 b6
16.ed3 ac8 17.a3 ("Two bishops and
active pieces slightly favour Black in this
position; White has no active plan." –
Ftacnik) g5 18.c4 b5 19.b3 e7
20.c1 was P.Nikolic-A.Grischuk, French
Team Championship 2003, and now Ftacnik
gives e5 21.a4 b6 22.xe5 xc5
23.f4 e6 24.xc5 xc5 25.d3 e7
as good for Black. ]
[ 13.c5 b6 14.d3 e8 15.d2 f5
16.ac1 ad8 shows Black quickly reaching
an optimal set-up with full development. I
like the idea of putting the bishop directly on
f5, from where it is well placed and has
ideas of trading on d3 or trading the g2-
bishop after ...Be4. Then 17.fd1 h6 18.b3
e7 19.f4 c6 20.d3 e7 21.f4
c6 22.d3 e7 23.f4 was agreed
drawn in I.Enchev-M.Nikolov, Sofia 2010,
but in fact g5! 24.h5 ( or 24.d3 d5
followed by ...Nc3 ) 24...h8 leaves Black on
top – taking on g5 doesn't work and
otherwise Black will tidy up with ...Bg6 and ...
Nd5. ]
[ 13.d2 f5 14.xf6+ xf6 15.f4 fe8
is equal and has led to draws in several
games between good players; for instance,
16.h4 e4 17.xf6 gxf6 18.e3 ad8
19.ad1 dxe3 and a draw was agreed in C.
Foisor-Al.David, Milan 2009 – ( although it
may have been more accurate to play 19...f5
. )]
[ The forcing 13.b4 needs to be carefully
considered. After xb4 ( 13...a6 14.a4
xb4 15.xf6+ xf6 16.xd4 is similar )
14.xf6+ xf6 15.xd4 ( or 15.xd4 d8
16.a4 a5 17.e3 d7 18.b3 ab8
and Black had no problems in P.Prohaszka-
J.Horvath, Austrian League 2012; ½-½ in
31 ) 15...e7 , White has liquidated Black's
only weakness and the position is drawish.
Nevertheless, Black needs to be careful,
since he has not yet completed his
development. The transition to such
positions can be psychologically
uncomfortable for Tarrasch players and it is
important to remain attentive:
A) 16.b2 c6 17.fd1 ( or if 17.a4
, as in P.Maletin-A.Belozerov, Tomsk 2009,
then and now d8 followed by 18...Be6
leads to equal play ) 17...f5 18.d4
xd4 19.xd4 was N.Vitiugov-V.Yemelin,
Tallinn (rapid) 2006, and now the active
ae8 20.e3 b6 levels the chances, since
Black has solved the problem of his b7-
weakness.;
B) 16.e5 xe5 17.xe5 leads to a
further branch:
B1) . After 17...e6?! 18.xb7 ab8
19.f3 , Black didn't have enough for the
pawn in J.Hammer-V.Akobian, Lubbock
2009 (1-0 in 46). It is interesting to see
the mindset of Akobian (a Tarrasch
expert), preferring active counterplay to
passive defence. In this particular
position, however, Black could have
equalized with more cautious play.;
B2) 17...e8 18.fd1 f5 19.f3
( alternatively, 19.f4 , as in A.Pashikian-
J.Bejtovic, European Championship,
Plovdiv 2012, could have been well met
by f6 20.c4 xe2 21.xb7 b8
with no problems; for instance, 22.d6
c2 23.d4 a5 and Black's pieces are
active and sec ) , was D.Fridman-M.
Feygin, Belgian Team Championship
2009, and now 19...c6 was the most
accurate, with equality.;
B3) 17...--;
C) 16.-- ]
13...e8
[ 13...f5 14.d3 xe4!? 15.xe4
was agreed drawn in M.Palac-R.Zelcic,
Croatian Championship, Split 2008. I don't
think this is entirely necessary, but trading
down to opposite-coloured bishops is worth
bearing in mind – as is well known, an
advantage of the two bishops is that you
can always trade one of them. ]
14.d3 e7 15.c1 b6
[ 15...f8 16.e1 b6 17.d2 f5
18.ec5 ac8 looks like the ideal set-up.
After 19.h3 g6 , a draw was agreed in M.
Suba-R.Vazquez Igarza, Collado Villalba
2008. ( Black had other options too, such as
 19...e5 when his position is slightly
preferred by the engines. )]
16.d2
[ A.Belozerov-A.Shomoev, Russian Cup,
Tomsk 2002, saw 16.a3 g4 17.h3 h5
It's always a tough decision where to retreat
this bishop: while h5 always looks the most
active (targeting e2), in fact the bishop can
end up slightly passive on g6 (after g3-g4).
In addition, Nf4 is likely to be annoying. In
the circumstances 17...Bf5 seems like a
more active approach. Black plans simply ...
h7-h6 and ...Bf8 with a harmonious position
in which his chances are no worse. Instead,
after 17...Bh5 the game continued 18.b4
f8?! (a little passive) ( 18...a5
is a standard way to seek to break down the
white queenside – here it seems dubious in
view of 19.f4 g6 20.c5 and White is
better since axb4 21.d5 a7 22.xb7
looks threatening, but after ab8 23.xc6
xb7 24.axb4 xb4 25.c4 f8 26.xd4
xd4 27.xd4 , Black has decent drawing
chances with all the pawns on one side of
the board; alternatively, 18...ac8 19.f4
g6 20.c5 xc5 21.xc5 e4
also keeps White's advantage within
manageable limits ) 19.f4 g6 20.c5
xc5 21.bxc5 a6 22.xc6 bxc6 23.xd4
xa3 24.a1 and White had the better
prospects (1-0 in 43). However, I've outlined
how the defence could have been
improved. ]
16...h6 17.ec5 a5
[ It was interesting to consider 17...g5!?
18.f4 e7 , trying to loosen the white
position. ]
18.a3 d6 A logical move, preparing ...Ne5.
[ Again 18...g5!? 19.f4 e7 creates some
weaknesses and seems more enterprising. ]
19.c2 e5 20.f4 c6 21.d5
[ EXERCISE: Assess the endgame after
21.d5 xc5 22.xc5 xc5 23.c7 xe2
24.xa8 f8 . ]
ANSWER: Black has a pawn (the d4-passer)
and the bishop pair to compensate for his
exchange deficit. Black really doesn't seem at
all worse here and I think I made the mistake
of trusting my GM opponent.
21...d8?! This compliant move lets White
get an edge. 22.fe1 e5 23.d3 e8
24.b3 a4 25.b6 a5? Carelessly
conceding control of b6.
[ 25...a6 26.xd8 xd8 looks like a level
endgame. ]
26.xd8 xd8 27.b6 Now the queenside
pressure effected by the Nb6, Bg2 and Rc1 is
unpleasant. Nevertheless, Black still has a lot
of defensive resources. e6
[ 27...f5 was more active; for instance,
28.xc6 bxc6 29.xc6 a6 with a better
version of the game continuation. ]
28.xc6 bxc6 29.xc6 a6 30.ec1 g5!
Even after a number of inferior decisions by
Black, the resulting position is far from
unplayable. Black has the bishop pair to
compensate his pawn deficit, and Houdini
assesses the position as only slightly better
for White.
31.f1 g7 32.e1 f8
[ 32...e7 was slightly more active. ]
33.d2 e8 34.c5 xc5 35.1xc5 b8
I had placed my hopes on the double pin, but
White can deal with this without too many
problems.
[ 35...g4 was more enterprising: 36.f3 h3
37.c4 xc6 38.xc6 f1 39.e4 dxe3+
40.xe3 d8+ 41.e1 d3 with drawing
chances. ]
36.b5 a7 37.b4 ab7 38.d3 b3
39.xd4?
[ Houdini recommends 39.h4 with a clear
advantage for White. ]
EXERCISE: Find a way for Black to stay in the
game.
39...e7?? Certainly not this!
[ ANSWER: 39...d8+ 40.c3 d1
gives Black active play for the two pawn
deficit and it's actually not even clear that
White is better here. ]
40.e4 d8+ 41.e3 . The game score ends
here, and I presume I lost after some time
scramble.
1-0
D34
Akopian,V
Jianu,V
13: European Championship, Legnica 2013
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 8.dxc5
xc5 9.c3 0-0 Here we will look at a
couple of lines recommended in recent
repertoire books. Accordingly, these lines are
important since they are very likely to arise in
practice.
10.a4 This is Boris Avrukh's selection.
[ 10.a3 is recommended by Larry Kaufman
in his repertoire book. Analysis by Aagaard
and Ntirlis showed that Black needs to be
very careful here. They recommended the
enterprising pawn sacrifice e4!
(not mentioned by Kaufman) which has
subsequently been tested in several games
between strong players:
A) 11.xd5 e6 12.c3 ( 12.e3
was tested in E.Vorobiov-T.Suc, Trieste
2012, and now Black should have played
f5! 13.xd8 axd8 14.b4 d4
with full compensation thanks to his much
more active pieces ) 12...xc3 13.bxc3
xd1 ( or 13...e8 14.a4 a5
15.xa5 xa5 16.d4 c4 17.e3
ac8 18.f3 a6 and it was hard to find
a way for White to make progress in A.
Greenfeld-R.Swinkels, Amsterdam 2013 –
that Black went on to win can't be
considered a result of the opening, but
does perhaps suggest that White's
technical task is far from straightforward )
14.xd1 fd8 ( 14...ad8 15.b2 a5
16.d4 b3 17.e1 a4 18.e3 c4
19.c1 b5 and Black had a dream Nimzo-
Indian position – even with a pawn less –
and full compensation in E.Bacrot-A.
Delchev, European Championship, Plovdiv
2012; ½-½ in 31 ) 15.f4 xd1+ 16.xd1
xa3 17.d4 d8 18.xc6 bxc6 19.a1
c5 20.xc6 c8 21.d4 was A.Giri-M.
Kopylov, German League 2012, and now
d7 would have kept full compensation for
the pawn. I don't see how White can make
progress while avoiding drawn opposite-
coloured bishop endgames.;
B) 11.xd5 xc3 12.xc5 xe2+
13.h1 xc1 14.axc1 f6 15.b4
( while after 15.c3 e8 16.b5
, as in I.Tokarev-S.Bezgodova, Kazan
2010, Aagaard gives h6 17.d1 b8
as leading to equality )
, was R.Khusnutdinov-R.Bulatov, Pavlodar
2012, and now Black equalizes with
15...g4 .;
C) 11.c2 xc3 12.xc3 b6
( 12...e7 13.d4 xd4 14.xd4 e6
15.f4 f6 16.e5 xe5 17.xe5
gave White an edge in K.Lahno-N.
Zhukova, Ukrainian Women's
Championship, Kharkov 2012 ) 13.b4 f5
( 13...d4 is also playable: 14.d2 e7
15.b2 d8 16.fd1 e6 17.b5 a5
18.b4 xb4 19.axb4 b3 20.a3
ac8 , when Black had no problems and
went on to score an upset in A.
Rasmussen-B.Byklum, Stockholm 2012;
0-1 in 45 ) 14.b2 d4 ( 14...f6
also deserves attention ) 15.d2 e4
16.a4 a6 17.b5 a5 18.b4 axb5
19.axb5 d5 20.xd4 xf3 21.xb6
xg2 22.fd1 c6 ( 22...e6 23.c5
b3 is okay, but I prefer Moranda's
treatment ) 23.xf8+ ( 23.bxc6 xa1
24.xa1 h3 25.f3 xc6 is balanced )
23...xf8 24.xd5 xd5 25.bxc6 xc6
and the long forcing sequence led to a
drawn endgame in V.Potkin-W.Moranda,
European Championship, Legnica 2013
(½-½ in 41).;
D) 11.-- ]
10...e7 Retreats to d6 and b6 have also
been tested, but the text seems the most
logical and is by far the most popular. 11.e3
While (as will appear) I don't agree with his
assessment of certain key variations,
Avrukh's explanation of the strategic plans
here is crystal clear: "In general we have clear
rules about how to play either side of a
position with an isolated pawn. In this specific
case exchanging certain minor pieces would
clearly favour White. The most desirable
exchanges are: trading dark-squared bishops
and a pair of knights (f3 for c6), which would
help White to establish full control over the d4-
square. Obviously, the endgame positions are
always better for the side playing against the
isolani. On the other hand, Black should strive
for activity keeping as many pieces as
possible on the board."
[ After 11.f4 g4 12.c1 e8 13.a3 e4
14.b4 c8 15.d3 f6 , Black was
perfectly mobilized in R.Salvador-J.Sadorra,
Philippines Championship, Boracay Island
2012 (0-1, 36). ]
11...e8 One of three popular moves here.
[ 11...e4 was tested in a game between
two young American GMs: 12.c1 e8
13.d4 d7 ( 13...f6 14.xc6 bxc6
15.xc6 d4 16.f4 didn't give Black quite
enough for the pawn in V.Iordachescu-Y.
Kuzubov, European Championship,
Kusadasi 2006 ) 14.b3 ( Avrukh's
suggestion of 14.c3 xc3 15.xc3
is probably a better try ) 14...f5 15.ac5
xc5 16.xc5 d7 17.e1 h3 18.h1
h5 19.e3 h4 20.c5 xc5 21.xc5
ad8 and Black had no problems in A.
Ramirez Alvarez-R.Robson, US
 Championship, St. Louis 2012 (0-1 in 63). ]
[ 11...g4 12.c1 ( 12.h3 f5 13.c1 c8
14.d4 e4 15.xc6 bxc6 16.xa7 xg2
17.xg2 e4 looked like a comfortable
version of a standard sacrifice in V.Artemiev-
A.Nadanian, World Blitz Championship,
Khanty-Mansiysk 2013 – Black has potential
play down three half open files and a
reasonable game )
A) 12...c8 13.h3 ( or 13.c5 xc5
14.xc5 e8 15.h3 h5 16.g4 g6
17.e3 e4 18.a4?! h5! 19.g5 d7
and Black was much better in Y.Shulman-
V.Akobian, blitz match, Upper Lake 2009;
0-1 in 27 ) 13...h5 14.c5 xc5
15.xc5 with some advantage for White in
E.L'Ami-V.Akobian, Wijk aan Zee 2010
(1-0 in 52).;
B) 12...d7 13.c5 xc5 14.xc5 fe8
15.e1 ( 15.g5 f5 16.xf6 xf6
was level in D.Fridman-A.Berelowitsch,
Dutch Team Championship 2011; ½-½ in
36 ) 15...h6 16.d4 h3 17.xc6 bxc6
18.a4 xg2 19.xg2 ac8
looked reasonable for Black in D.Antic-V.
Kotronias, Greek Team Championship
2011 (0-1 in 55).;
C) , and now, as well as 12...e8
(transposing into the main game), Black
has a choice between two good moves: ]
12.c1 g4
[ 12...f5 is a pet variation of Larino Nieto.
He tested the position after 13.d4 xd4
14.xd4 e4 15.e3 b6 16.c3 d7
17.e2 ad8 twice against strong
opposition in Don Benito 2012 and had no
problems, even using the same plan in both
games:
A) 18.fd1 g4 19.f3 xc3 20.xc3
(Black could also retreat to f5 or e6 since
the g2-bishop is passive, but trying to
weaken the white king is logical) h3
21.xh3 xh3 22.c7 d7 23.xd7
xd7 24.c1 h5!? 25.g2 d6 26.d3
b8 was L.Rubio Mejia-D.Larino Nieto,
Don Benito 2012 (½-½ in 62). White has a
hard job co-ordinating an attack on the c-
file and minding his king at the same
time.;
B) 18.a3 g4 19.f3 xc3 20.xc3 h3
21.fc1 xg2 22.xg2 d6 23.b4 h5!?
was A.Alonso Rosell-D.Larino Nieto, Don
Benito 2012 (½-½ in 57). Although
computers like this for White, I think Black
is not worse in a practical game, since his
king is safer and White has no entry points
on the c-file. ( 23...-- );
C) 18.-- ]
13.h3
[ Instead: 13.c5 d7 14.e1 ad8
15.d4 was M.Ragger-S.Brenjo, Bosnian
Team Championship 2011, and now c8
16.xe7 xe7 17.xc6 xc6 would have
been equal. ]
[ 13.c3 d7 14.a4 h6 15.fd1
was P.Tregubov-V.Akobian, FIDE World
Cup, Khanty-Mansiysk 2009, and now b4
( or 15...ad8 leads to good play for
Black. )]
[ 13.d4 d7 14.xc6 ( 14.f3 h5 15.f2
ac8 16.e1 e5 17.c3 c5 18.a4
was agreed drawn in R.Markus-S.Halkias,
Serbian Team Championship 2010; while
14.c5 xc5 15.xc5 was A.Dreev-V.
Yemelin, Tallinn rapid 2009, and now I like
h3 , exchanging White's best minor piece )
14...bxc6 leads to a further branch:
A) 15.d4 ac8 ( 15...ab8 16.e1
b4 17.c3 d6 18.a3 was M.Kekelidze-
V.Akobian, Las Vegas 2010; now Akobian
launched his h-pawn forward, but e5
is solid and equalizes ) 16.f3 h3 17.d2
xg2 18.xg2 e6 19.c2 d7
20.fc1 h5! with good kingside
counterplay in Y.Drozdovskij-K.Kiik,
Tromsø 2009 (0-1 in 42). In particular,
Black gets good compensation if White
grabs on a7.;
B) 15.e1 b4 16.d2 d6 17.c5
f5 18.b7 was M.Ragger-M.Kopylov,
German League 2013, and now
Scherbakov recommends e5!
which leads to good play for Black.;
C) 15.-- ]
[ 13.c5 is Avrukh's suggestion.
EXERCISE: Find a good response for Black.
ANSWER: b6! is one of the first
"novelties" I found on preparing the Tarrasch
a few years ago (if you count Houdini's first
choice as a novelty). Certainly the opposition
of the black queen and the e3-bishop
creates an odd impression but White simply
doesn't have any useful discoveries:
A) 14.a4 a6 15.a3 ad8 16.e1
e4 17.d4 e5 18.f3? ( 18.b3
is equal ) 18...d7 19.fxe4 xa4
and Black was clearly better in Mi.
Novkovic-V.Jianu, European
Championship, Plovdiv 2012 (0-1 in 37).;
B) 14.d3 d4 15.g5 ( 15.xc6 bxc6
 16.xd4 c5! is not worse for Black )
15...h6 16.f4 d5 17.d2 ad8
and Black was very comfortable in Ki.
Georgiev-B.Byklum, European Cup, Eilat
2012.;
C) . The position after 14.b3 a6 15.a3
ac8 has been extensively tested:
C1) 16.d3 xd3 17.exd3 h6
was agreed drawn in A.Ushenina-N.
Zhukova, Ukrainian Women's
Championship, Kharkov 2012.;
C2) 16.e1 h6 17.h3 ( after 17.bd4
, as in P.Jaracz-J.Asendorf, German
League 2012, I like a5 , aiming for
c4 ) 17...f5 18.c5 e4 19.e3 xc5
20.xc5 b6 21.b4 a5 22.bxa5
was also agreed drawn in D.Reinderman-
R.Swinkels, Dutch Championship,
Amsterdam 2012. With Black I might be
tempted to play on here, since his pieces
are very active and he runs no real risk
of losing once the white queenside is
liquidated.;
C3) 16.h3 h5 ( normally I prefer to
retreat the bishop to f5, and here too
16...f5 looks logical ) 17.e1 h6
18.d3 xd3 19.exd3
with a symmetrical and equal endgame
in V.Laznicka-V.Jianu, European
Championship, Plovdiv 2012.;
C4) 16.--;
D) 14.-- ]
QUESTION: Where should Black retreat his
bishop?
ANSWER: He shouldn't!
13...xf3!? A rather striking approach which
has been played with success by Vlad Jianu
in several games. Black gives his opponent
the bishop pair but manages to organize his
own position fully. Retreats to e6 and f5 have
been tested too, but we will concentrate on
the text.
14.xf3
[ 14.exf3 d4!? ( 14...d7 15.c5 xc5
16.xc5 h6 17.d3 ad8 was also equal
in A.Giri-V.Jianu, European Championship,
Plovdiv 2012 ) 15.g5 d3 16.c3 a5
17.xf6 xf6 18.xd3 ad8! 19.c3 xc3
20.bxc3 xa2 and Black had equalized in O.
Barbosa-Wei Yi, Jakarta 2012 (½-½ in 31). ]
14...d7 15.g2
[ 15.g2 ad8 16.c5 xc5 17.xc5 e4
18.d4 f5 19.e3 c8 20.c3 h5!?
21.f3 xf3 22.xf3 xc3 23.bxc3 ( after
23.xc3 d4 24.exd4 xd4 25.xh5 xc3
26.bxc3 e2+ 27.xe2 xe2
, Black should be able to hold the endgame
due to his active rook and the white
queenside weaknesses ) 23...e5 24.c4
was R.Wojtaszek-A.Delchev, French Team
Championship 2013, and now simply dxc4
25.xc4 g6 looks solid enough for Black. ]
15...ad8 16.c2
[ 16.c5 xc5 17.xc5
was D.Arngrimsson-V.Jianu, Arad 2012, and
now Black can create interesting play with
e4 18.c2 g5 19.g4 e6 20.fd1
g6 , followed by ...f7-f5 or ...h7-h5. Black has
some (albeit faint) attacking hopes, whereas
it's not quite clear how White should
continue. ]
16...e4 17.xe4 dxe4 18.fd1 e6
19.xd8 xd8 20.c4 xc4 21.xc4 f5
The endgame is equal. While this is no
guarantee of a favourable result against
Akopian, Jianu managed to draw with ease,
playing actively and even sacrificing a piece.
22.c5 b5! 23.c1 d4! 24.b7 d7
25.a5 xe2 26.c8+ f7 27.c2 d4
28.xd4 xd4 29.c7 d2! 30.c6 e6!
31.xe7 e3 32.f1 xf2+ 33.e1 xb2
34.xa7 g5 35.g8 f4 36.e7+ d5
37.gxf4 gxf4 38.f6+ d4 39.d7+ c4
40.c7+ d3 41.d7+ c4 42.c7+ d3
43.d7+ c4
½-½
D34
Chatalbashev,B
Sokolov,I
14: Jakarta 2012
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 f6 7.g2 e7 8.0-0 0-0
9.dxc5 d4!? 10.a4 Otherwise Black
captures on c5 with no problems. f5
QUESTION: What does Black have for the
sacrificed pawn?
ANSWER: Frankly, Black's compensation is
not completely obvious, and it's far from clear
that it's sufficient to balance the pawn deficit.
First and foremost, the white knight is rather
out of play on a4 and will take several moves
to get back into the game. Black wants to
continue with moves like ...Be4, ...Qd5, ...
Rfe8 and ...Rad8 with an impressively
centralized position, and possible tactical
shots with ...d4-d3, ...h7-h6 and ...g7-g5 also
come into consideration.
White doesn't need to do much, since if he
simply consolidates he will be a pawn up.
However, one subtlety of the position should
be noted: Black's only weakness is on b7
which is actually a bit awkward to defend after
Qb3; but if White plays the most natural plan
of consolidating c5 and bringing his knight
back into play (a2-a3, b2-b4 and Nb2), then
the b7-pawn can no longer be attacked.
11.b4!? A dynamic approach. White returns
the pawn in order to activate his rook on the b-
file. Giving White b2-b4 "for free" (without a2-
a3) would be too much of a concession, so
Black needs to take the pawn, but can play
an intermezzo first.
[ 11.f4 e4 12.c1 gives Black a wide
choice:
A) 12...e8 and now:
A1) 13.a3 d5 14.b4 ad8 15.e1
( 15.b5 a5 16.c7 b3 17.xd8
xd8 is an important tactical point, with
equality in E.Vegh-D.H.Toth, Hungarian
Team Championship 2007; but 15.b2
g5 16.d2 g4 17.e1 xg2 18.xg2
e4 was good for White in M.Stojanovic-
A.Profumo, Lugano 2006, since Black
had given a lot of squares away, in
addition to his pawn ) 15...h6 16.b5
( 16.b2 looks more sanguine ) 16...a5
17.c7 b3 18.xd8 xd8 19.c2
xc2 20.xc2 xc5 21.xc5 xc5
and Black had a good Tarrasch-type
position in A.Barsov-E.Magerramov, Abu
Dhabi 1999 (0-1 in 40).;
A2) 13.b4 xb4 14.xd4 , as in M.
Hörstmann-R.Östreich, Bad Lauterberg
1994, should be met by xa2 15.cd1
xd4 16.xd4 ac8
with an approximately level endgame.;
A3) 13.b3 and now, after the
ridiculous a5? ( instead, 13...d7
14.e3 ad8 was more sensible, with
decent compensation ) 14.b4 c6
15.xb7 d5 16.fd1! , I lost without a
fight in L.Cernousek-S.Collins, Teplice
2011.;
A4) . Houdini's recommendation of
13.e3! doesn't seem to have been tested
yet.;
A5) 13.--;
B) 12...h6 13.a3 g5 has been essayed
by Ekaterina Kovalevskaya, whose
Tarrasch expertise has been praised by
none other than Viktor Bologan. After
14.d2 d5 15.b4 e5 , it is rather hard
to believe that Black enjoys full
compensation here, but the position is
messy and Black's practical chances
shouldn't be underestimated. While White
has basically consolidated on the
queenside (her pawn chain is strong and
Nb2-d3/c4 is in the pipeline), Black doesn't
have to worry about her b7-weakness
which, as will be seen, features in some
other lines:
B1) 16.h4 h7 17.c6 b6 18.c3!
was much better for White in A.
Galliamova-E.Kovalevskaya, FIDE
World Cup, Khanty-Mansiysk 2012 (1-0
in 44). Since the g5-pawn wasn't really
threatened, Black could improve with
16...Ng6, though it remains to be seen
whether she has enough
compensation.;
B2) 16.b2 d3 ( simply continuing the
build-up with 16...fe8 17.d3 xf3+
18.xf3 ad8 looks sensible – I don't
see how White is going to unravel )
17.xe5 ( 17.c3! would have given
Black serious problems ) 17...xg2
18.bxd3 ( 18.e1 h3 19.f3 dxe2
20.xe2 might be a better try for an
advantage, though players who go for
this line should be delighted at the long-
term weakness of the light squares on
the white kingside ) 18...xf1 19.xf1
e4 and White had good compensation
for the exchange, but Black had no
reason to complain in V.Cmilyte-E.
Kovalevskaya, European Cup, Eilat
2012 (½-½ in 56).;
B3) 16.--;
C) 12...d7 is a speciality of Serbian GM
Goran M Todorovic (not to be confused
with Serbian IM Goran N Todorovic). Black
simply puts his pieces in the centre and
challenges White to break out. One
feature of this line is that Black hangs on
to his b7-pawn, which rather holds his
position together. After 13.a3 ad8 14.b4
, there are several examples:
C1) 14...fe8 15.b2 ( or 15.e1
xg2 16.xg2 d5 and Black had
pleasant compensation in V.Radonjanin-
G.M.Todorovic, Serbian Team
Championship 2005; 0-1 in 37 ) 15...d5
16.d2 f5 17.b3 a6 18.d3 f6
19.f4? was a typical blunder in S.
Krivoshey-G.M.Todorovic, Dos
 Hermanas (online blitz) 2004. Now the
game's 19...d3 was alright (0-1 in 49),
but xf4! 20.gxf4 ( or 20.xf4 d3 )
20...e6 would have been even better.;
C2) . The speculative 14...d5 15.d2
e6 16.e1 f6 17.b5 fe8 18.bxc6
xc6 19.b2 didn't give Black enough
for the piece in Ki.Georgiev-G.M.
Todorovic, Topola 2004 (1-0 in 40).
Black has improvements such as 17...
Nce7 and 16...a6; but earlier White has
the annoying 15 Bd6, when I don't think
Black has enough compensation.;
C3) 14...--;
D) 12...d5 13.b3 h5 seems dubious
to me. ( 13...xb3 (trying to bail out)
14.axb3 , as in I.Zugic-I.Seitaj, Calvia
Olympiad 2004, and now d5 15.d6
fe8 16.xe7 xe7 leaves Black with an
uphill struggle to draw )
D1) 14.xb7! looks very strong: d3
(Black doesn't have to play this, but it's
one of the main ideas of the whole line
and it's not clear what else he would play
to justify his two-pawn deficit) 15.fe1
( Houdini prefers 15.d2 ) 15...dxe2
16.xe2 ( again 16.d2! was stronger )
16...xf3 17.xf3 xf3 18.xe7 e4
19.xe4 xe4 20.c3 f3
with an edge for White in D.Fridman-S.
Halkias, European Team Championship,
Novi Sad 2009 (1-0 in 52).;
D2) 14.fd1 ad8 ( after 14...d3
15.xd3! xd3 16.exd3 , White had
two pawns for the exchange plus much
better pieces in J.Gonzalez Valero-S.
Pozo Vera, Spanish Team
Championship 2009; note how difficult it
is to defend the b7-pawn ) 15.d6?
( Ftacnik gives 15.xb7 fe8 16.g5
xg2 17.xg2 d5 18.xc6 xg5
19.xg5 xg5 20.xd4 xc1 21.xd5
as clearly better for White ) 15...b5!
taking advantage of the bishop on d6;
16.xe7 ( either capture of the black b-
pawn – 16.cxb6; or 16.xb5
– drops a piece ) 16...bxa4 17.xa4
xe7
D2a) or 18.xd4 xg2 19.xg2
ed5 20.b3 fe8 ; Ftacnik gives
21.f3 g5 , ( but modern engines
spot 21...g4! immediately: 22.h3
de3+! 23.fxe3 xe3+ 24.g1
xf3 25.exf3 xd1 26.xd1 e5
with a big advantage );
D2b) 18.xd4 xd4 19.xd4 xg2
20.xg2 d5+ 21.f3 g5
was slightly better for Black in S.
Matveeva-A.Lugovoi, Solin 2005 (½-½
in 58).;
D3) 14.g5 featured in one of the most
exciting games I ever played. Having
ruined an excellent shot at a GM norm
by losing with White to Sune Berg
Hansen in the previous round, I needed
to win with Black against a strong player
to keep a realistic norm shot. I got some
excellent advice (unfortunately not fit for
publication) before the game from
Aussie GM Dave Smerdon, and selected
this gambit variation. B.Gundavaa-S.
Collins, Khanty-Mansiysk Olympiad 2010,
continued g4 15.h3 xg2 16.xg2
xf2!? ( 16...xg5 is sounder, when
Black is at least not worse ) 17.xb7?
( 17.xf2 xg5 18.d5 h6 19.h4
is balanced ) 17...xg5 18.xc6 xf4
19.gxf4 xh3+! 20.xf2 ae8
(starting a caveman attack a full knight
down) 21.f3 h4+ 22.g2 e3
23.h1 e7 24.h5 e4+ 25.f1
. Now I went on to win a crazy game
after
D3a) However, 25...d3! was a better
move: 26.c4 ( I didn't want to let
White "bail out" with 26.xh7+
, but in fact xh7 27.xh7 xh7
is obviously very good for Black –
clearly a little too much adrenaline! )
26...dxe2+ 27.f2 e1+ 28.xe1
xc4 29.xe3 xf4+ 30.f3 xa4
and, while the game isn't over yet
(especially in view of White's c-pawn),
the three connected passers give
Black cause for optimism.;
D3b) 25...xf4+ (0-1 in 101).;
D4) 14.--;
E) 12...-- ]
11...d3! This seems best.
[ 11...xb4 12.xd4 e4 13.xe4 xe4
was Z.Li-I.Seitaj, Istanbul Olympiad 2012,
and now 14.f5! returns the pawn with a
much better endgame after xc5 15.xc5
xc5 16.e3 xd1 17.fxd1 . ]
12.e3
[ 12.a3 dxe2 13.xe2 d3 14.b2 xf1
15.xf1 was unclear in Zhu Chen-E.
Kovalevskaya, FIDE Grand Prix, Shenzhen
2011 (½-½ in 29). ]
12...xb4
 [ 12...b5 13.cxb6 axb6 targets the a4-knight
in an attempt to fuel Black's counterplay.
Unfortunately, 14.b5 b4 15.d4 e4
16.b2 seems better for White. ]
13.d4 e4 14.xe4 xe4 15.b1 xa2?
[ 15...a5 gives good equalizing chances. ]
16.xd3 g5 17.a3 a5
[ 17...f3+ 18.h1 d5 19.f5!
keeps White on top. ]
18.b6 a6 19.fd1 b4 20.xb4 axb4
21.c6 f3+ 22.h1 a2 23.xd8 xd8
24.g2 bxc6 25.d7?? A horrible mistake,
throwing away his advantage and then some.
[ Instead, after 25.xf3 e7 26.d7 c8
27.xb4 , Sokolov could only beg for a
draw. ]
EXERCISE: Black to play and win!
ANSWER: 25...d2! The knight blocks the d-
file and shields the d8-bishop from the rook on
d1. 26.xf8 xf8 27.xb4 e7
White should still hold this, but it's unpleasant
– and obviously nothing to do with the
opening. 28.d4 a5 29.g4 g6 30.g5 c3
31.h4 e6 32.c1 e5 33.b4 f5
34.b7 e4 35.xf7+ e6 36.f3 xg5
37.f8 e7 38.c8 e4 39.f3 xf2
40.f1 h3 41.e4 xh2 42.xc6 d2
43.f3 g5+ 44.g4 g2+ 45.h4 h5
0-1
D30
Manolache,M
Jianu,V
15: Baia Sprie 2012
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 8.dxc5
xc5 9.a3 This is another attempt which has
been played against me by a strong GM. 0-0
[ Aagaard and Ntirlis recommend 9...e4
. We have yet another example of Jacob
slightly varying from his own
recommendation and getting a good result:
A) 10.fd2 (another novelty suggested by
Aagaard and Ntirlis which has
subsequently been tested in practice) f5!
( after 10...f5 11.xe4 xe4 12.xe4
dxe4 13.c2 e7 14.c3 , Black had
some problems in V.Georgiev-M.
Meinhardt, German League 2012; 1-0 in
51 ) 11.c3 e6 12.b3 b6 13.a4
d4 14.d2 xd2 15.xd2 0-0 16.c1
d5 and Black had equalized in W.Ju-N.
Zhukova, FIDE World Cup, Khanty-
Mansiysk 2012.;
B) 10.b4 b6 ( here 10...e7 11.b2
f6 12.xf6 xf6 13.a2 0-0 14.c2
d8 with equality in Y.Kruppa-V.Potkin, St.
Petersburg 2000, was the route suggested
in Jacob's book ) 11.b2 0-0 12.c3
xc3 13.xc3 d4 14.b2 a6 15.d2
e6 16.f4 c4 17.fe1 e8 18.f1 a5
19.ad1 ( 19.ac1 is another try, but
after c7 20.d2 b5 Black's pieces look
too well placed for him to be worse; a
natural line might continue 21.e3 axb4
22.axb4 dxe3 23.xe3 xd2 24.xd2
xe3 25.fxe3 xb4 26.xc4 bxc4
27.xc4 with what should be a drawn
endgame ) 19...b3 20.d3 c4 21.dd1
b3 22.d3 c4 23.dd1 and a draw
was agreed in L.Johannessen-J.Aagaard,
European Team Championship, Porto
Carras 2011. Another example of Jacob's
exemplary handling of the Tarrasch –
nothing fancy, just putting his pieces on
good squares.;
C) 10.-- ]
10.b4 b6 11.c3 h6 12.b2 e8 13.c1
[ 13.a4 c7 looks okay for Black. ]
13...a6
[ 13...f5 was my choice when I reached this
position in N.Pert-S.Collins, British League
2012. For the moment I find a good plan,
bringing my pieces to decent squares. Nick's
idea of Na4xb6 will weaken his control of the
e4-square. The game continued 14.a4
e4 15.xb6 axb6 , when I seem to
remember being dissatisfied with my
position, but unfortunately I had forgotten
my chess history.
A similar position (same structure and with
White having the bishop pair) was reached
in M.Tal-P.Keres, Bled/Zagreb/Belgrade
Candidates 1959 (0-1 in 79). Although it
looks as though White must be better, in fact
it's hard for him to find an active plan. My
bishop on e4 is the best minor piece on the
board and if it's traded for its white
counterpart, Black will always have some
attacking chances. Also the e2-pawn is
weak and doesn't want to move since this
would weaken a bunch of light squares.
After 16.h4
A) , and then 16...e5 17.xe4 xe4!
, the knight is threatening to come to c4,
while if White trades on e5 the position
should be level since White has his own
weaknesses on a3 and e2.;
B) . Alternatively, 16...xg2 17.xg2
e7 followed by ...Rad8 is a sound
continuation, with equal chances.;
C) . Instead, I played 16...e7?
, which indicates a mistaken plan.
One of the things to be careful of when
playing the Tarrasch is structural changes
and, in particular, a diminished sense of
danger regarding lines where White
"corrects" Black's pawn structure. To take
a basic example, if the d5- and e2-pawns
get traded in a normal Tarrasch position,
we have a symmetrical and open position
in which the better mobilized side will have
the better chances. Often the better
mobilized side will be White, a natural
consequence of getting to move first.
Similarly here, I barely gave any thought to
Nick sorting out my pawn structure by
taking on e4. However, in the resulting
position, the bishop on b2 is extremely
powerful (it can never be blocked by ...d5-
d4, for instance) and White's weakness on
e2 is covered. The dark-squared bishop
can combine well with the knight on h4,
particularly if this knight gets into f5.
After 17.xe4! , I could in fact have kept
the game unclear with
C1) So 17...dxe4 18.b3 ad8
(more playing by rote) ( 18...e6
was the only attempt to keep the knight
on h4 ) 19.f5 (White is already clearly
better) e6 20.xe6 xe6
( the computer recommends the pawn
sac with 20...fxe6 21.xh6+ h7 )
21.fd1 (around here I was probably
looking with pride at my half-decent
pawn structure, not noticing how White's
pieces have become several times more
active than their counterparts) b8
22.d6 f8 23.b5 e7 24.xe6 fxe6
25.xe7 xe7 26.c7+ d6 27.xg7
and White is now winning and converted
without too much hassle (1-0 in 46).;
C2) 17...xe4 here, since the capture
on f6 isn't as dangerous as it looks –
though this is still worse than my options
on the previous move, and when I
played 16...Qe7 I was of course planning
to take back on e4 with the pawn (to the
extent that I examined this line at all –
as outlined above, I probably devoted
more time to continuations which left my
IQP on the board).;
D) 16...-- ]
14.a4 a7 15.c5 The knight transfer to
c5 brings no advantage after Jianu's accurate
reaction. e4! 16.c2 xc5
[ 16...a5!? is a good alternative. ]
17.bxc5 e7?! A rather speculative pawn
sacrifice.
[ After 17...f5 18.d2 e4 19.h4 xg2
20.xg2 d4 21.xd4 xd4 22.xd4 c7
, Black regains his pawn with equality. ]
18.xd5 e6 19.d2 ad8 20.c3 f8
21.a1?
[ 21.e4! would keep White on top. ]
21...f5 22.d2 xc5 23.xd8 xd8
Black has fully equalized. The rest of the
game sees Jianu convincingly outplaying a
good GM. 24.b2 d7 25.e5 xe5
26.xe5 e6 27.a4 b6 28.c6 e7
29.b2 c4 30.f3 d7 31.c2 b5
32.axb5 axb5 33.b2 d8 34.f5 f8
35.a1
[ 35.e4 was a better move, when Black
maintains equality with g6 36.e5 f6!
, since he is certainly not worse after
37.xf6 xf6 38.xf6 f7 and 39...Bxe2. ]
35...d1+ 36.g2 xa1 37.xa1 e7
38.e5 g6 39.f4 g7 40.xg7 xg7
Black is now solidly better. White's defence
can be improved, but in the game he didn't
last much longer. 41.d4+ f6 42.d2 e5
43.h4 h7 44.g4 c5 45.a5 e7
46.h3 b4 47.h5 gxh5 48.gxh5 e6+
49.g3 g5+ 50.xg5 hxg5 51.e4+ h6
52.f4 f6 53.b1 xh5 54.e4 b3 55.d3
b2 56.e5 fxe5 57.fxg5 xg5
0-1
 Back to Contents Page
Chapter Five

Other 8th/9th Move Options

(View in Game Format)

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7

After the previous chapters I think we're pretty well prepared for this position. White has a couple of logical
alternatives which we'll discuss.
8 b3
This queenside fianchetto, and the closely related 8 Nc3 0-0 9 b3, are the subject of Giri-Swinkels.
8 Be3 has been the traditional choice of the legendary Victor Korchnoi. See how I got on inKorchnoi-Collins.

Key Notes

1. The double fianchetto with b2-b3 is a playable way to get a complex position but is not a serious try for an
advantage. Black was doing just fine for a long time in Giri-Swinkels.
2. Korchnoi's Be3 leads to positions which have more in common with 9 Bg5 c4 than with the 9 Bg5 cxd4 lines. If
Black can get a knight to c4 he tends to be doing fine.

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5 exd5 5 g3 Nc6 6 Bg2 Nf6 7 0-0 Be7 8 b3 – Giri-Swinkels

8 Be3 – Korchnoi-Collins

Back to Contents Page

Other 8th/9th Move Options
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 After the
previous chapters I think we're pretty well
prepared for this position. White has a couple
of logical alternatives which we'll discuss. 8.b3
[ This queenside fianchetto, and the closely
related 8.c3 0-0 9.b3 , are the subject of
Giri-Swinkels (Game 16). ]
[ 8.e3 has been the traditional choice of
the legendary Victor Korchnoi. See how I got
on in Korchnoi-Collins (Game 17). ]
D34
Giri,A
Swinkels,R
16: Dutch Championship, Amsterdam 2012
[Sam Collins]
1.c4 e6 2.g3 d5 3.d4 c5 4.cxd5 exd5
5.f3 c6 6.g2 f6 7.0-0 e7 8.b3
[ White can also start with 8.c3 0-0 9.b3
. This was played against me in a
tournament where I made my second GM
norm. After e4 10.b2 f6 , White needs
to deal with the pressure on d4.
A) 11.a4 b6 has proven resilient for
Black; for instance, 12.c1 e8 13.dxc5
xb2 14.xb2 bxc5 15.d2 f6
16.a4 ( 16.xe4 dxe4 17.xc5 g4
and ...Rad8 promises reasonable
compensation ) 16...a6 17.xe4 dxe4
18.d2 ad8 19.e3 d4 gave Black
good chances in J.Speelman-T.Luther,
German League 2002 (0-1 in 33).;
B) 11.e3 g4 12.a4 cxd4 13.xd4
xd4 14.exd4 b5 15.b2 ( 15.c5
is more active ) 15...b6 16.d3 xf3
17.xf3 xd4 18.xe4 dxe4 19.e2
with a decisive advantage for Black, which
I eventually (and following several major
mistakes) converted in R.Goetzee-S.
Collins, Tromsø 2010 (0-1 in 43).;
C) 11.-- ]
8...0-0 9.b2 White has not attempted to put
any immediate pressure on the d5-pawn, so
Black has a wide choice. e8
[ Alternatively: 9...e6 10.c3 b6 11.c1
c8 is solid, but a little passive, and White
has a good score here (64% in the
database). ]
[ 9...g4 is one of the more aggressive
moves, immediately pressurizing d4.
A) 10.e3?! is entirely misconceived: cxd4
11.exd4 d7 12.bd2 e4 and Black's
pieces were much better placed in the
symmetrical pawn structure arising in I.
Teodorescu-G.Timoshenko, Rumanian
Team Championship 2009 (0-1 in 77). The
b2-b3 and Bb2 idea makes no sense in
this structure – the dark-squared bishop is
biting on its own pawn, and the c3-square
is weak.;
B) 10.dxc5 xc5 11.c3 a6! 12.c1
a7 and Black had comfortably
established his bishop on its ideal diagonal.
This idea of ...a7-a6 and ...Ba7 (or, with
White, a2-a3 and Ba2) arises very often in
IQP positions. In addition to placing the
bishop on a great square where it is safe,
targets important pawns and squares on
the g1-a7 diagonal, and doesn't get in the
way of Black's pieces, the set-up has an
important benefit in preventing White's
Nc3-b5-d4 manoeuvre. I already prefer
Black, who scores very well from this
position (62% in the database); GM
Spartak Vysochin actually has 4/4 here as
Black.;
C) 10.e5 and now Black has a choice:
C1) 10...h5 11.dxc5 xc5 12.d3
d6 13.d2 e8 14.f3 leads to an
interesting position where the slightly
unusual feature is the white knight on d3.
On one hand, this knight shields the d5-
pawn from attack (much as a knight
would on d4); but it is also controlling
important dark-squares and limiting the
activity of Black's pieces. Probably the
position is dynamically balanced, as
suggested by a couple of examples: g6
( 14...e7 15.e1 ad8 16.c1 g6
17.d4 b4 18.xb4 xb4 19.f1 a6
was level in W.Ju-Zhao Xue, FIDE
Grand Prix, Shenzhen 2011 ) 15.c1
e7 16.e1 e4 17.f4!? xf4
18.gxf4 ad8 19.e3 led to an
interesting struggle in J.Tompa-D.
Berczes, Hungarian Team
Championship 2012. Black has
weakened the white kingside – at the
cost of (probably) his best minor piece.
If Aronian was handling the white pieces
the result could have been quite different,
but in the game White rapidly allowed
his GM opponent a devastating
initiative: h5 20.c2?! (the rook is
 misplaced on this square) e6 21.a3?
(White is justifiably worried about ...Nb4
shots, but it was better to simply move
the rook back) g6! 22.f1 and now
a5 23.b4 c4 would have left Black
completely on top.;
C2) 10...cxd4 11.xg4 xg4 12.e3
f6 13.exd4 , and in this symmetrical
pawn structure the two bishops are not
really felt – in fact, White needs to be
careful that his bishop on b2 is not
locked out of the game. After e8
14.c3 f8 15.d3 d7 16.ae1
xe1 17.xe1 d8 18.c1 h6 19.f4
b4 20.d1 d6 21.e3 a3 22.f4
d6 23.e3 a3 24.f4 a draw was
agreed in M.Gurevich-L.Aronian,
European Team Championship, Heraklio
2007.;
C3) 10...--;
D) 10.-- ]
[ By far the most popular move is 9...e4
and now:
A) 10.c3 transposes to the 8 Nc3 0-0 9
b3 variation in the first note to this game.;
B) 10.bd2 f6 11.e3 is a tame set-up.
Here g4 ( instead, I like 11...f5!?
as in B.Svendsen-S.Brynell, Gausdal 2005,
with active play for Black (0-1 in 23) ) 12.h3
might be a little better for White.;
C) 10.dxc5!? xc5 is critical:
C1) 11.c3 xc3 12.xc3 d4 13.b2
g4 leaves Black comfortable. After
14.c1 b6 15.h3 f5 16.d2 d7
and a draw was agreed in I.Morovic
Fernandez-A.Bachmann Schiavo,
Asuncion 2010.;
C2) 11.e3 has scored well for White.
C2a) . Black has been suffering after
11...g4 12.c3! ; for instance, a5
( 12...xc3 13.xc3 d7 14.d2
fd8 was A.Galkin-S.Kozhuharov,
Turkish Team Championship 2010,
and now 15.xg7! would have been
good for White in view of the forced
variation xf3 16.xf3 f5 17.b2
xf3 18.c3 d4 19.exd4 xc3
20.xc3 with an extra pawn ) 13.xd5
ad8 14.c2 and Black didn't
demonstrate sufficient compensation
in V.Zvjaginsev-Xu Jun, FIDE World
Championship, Moscow 2001 (1-0 in
22).;
C2b) . Perhaps Black should try
11...b6!? 12.e1 ( 12.xd5? xf2!!
is excellent for Black ) 12...d8 13.c3
xc3 14.xc3 g4 with equality, as
in Mi.Marin-M.Thesing, Rumanian
Team Championship 2009 (½-½ in
47).;
C2c) 11...--;
C3) 11.--;
D) 10.-- ]
10.c3 g4 11.dxc5 Drawing the bishop to
the c5-square where it will be vulnerable after
White plays Rc1.
[ By contrast, the immediate 11.c1
seems to save some time for Black: cxd4
12.xd4 d7 13.e1 ad8 and Black had
established his desired set-up in B.
Chatalbashev-Y.Solodovnichenko, Val
Thorens 2008 (0-1 in 32). ]
11...xc5 12.c1 f8 Keeping the bishop
out of harm's way looks logical.
[ However, an alternative approach was tried
in a rapid game between two top GMs:
12...b4!? 13.h3 h5 14.a3 xc3
15.xc3 e4 16.b2 b6 17.g4 g6
18.d4 ac8 with comfortable play for
Black in R.Mamedov-S.Ganguly, Moscow
(rapid) 2013. ]
13.h3 Probably a useful move to include.
[ Trying to do without it by 13.e1 d7
14.d4 allows Black to threaten the white
king with h3 15.h1 (otherwise the
kingside light squares would be weak, but
now the white king is short of breathing
space) ad8 with a dream IQP position for
Black in Z.Ilic-T.Willemze, Haarlem 2012
(0-1 in 32). ]
13...h5 14.b5 e4 15.bd4 c8
QUESTION: How would you assess this
position?
ANSWER: Black has every reason to be
happy. He has active development with all the
pieces still on the board (so a long way from a
depressing IQP endgame) and White doesn't
have any pressure on the d5-pawn. In addition,
while the blockade on d4 looks very solid,
Black has some pressure – an exchange of
rooks on c1 could draw away a white defender,
while the bishop on h5 and the rook on e8
combine well to pressurize the e2-pawn. In
short, a full and complex middlegame is in
prospect, where the better player will win. In
the game the 200+ rating point gap proves too
much.
16.c2
[ White can also opt for a more direct
approach: 16.xc6 bxc6 17.e5 a5
 ( perhaps Black should try the more solid
17...f6 18.d3 d6 19.f4 g6
20.xg6 xg6 with approximately level
chances ) 18.xc6 xa2 was played in H.
Hoang Canh-Tu Hoang Thong, Subic Bay
2009, and now White took immediately on
d5, missing the strong intermezzo 19.d4!
followed by 20 Qxd5(+) with a clear
advantage. ]
16...b4!? A good move, aiming for
queenside counterplay.
[ Houdini likes the bizarre-looking 16...f6!?
. The idea seems to be to overprotect g7
(and so allow ...Nxd4 in some lines) while
allowing ...Bf7 to cover the d5-pawn if
needed. ]
17.xc8 xc8 18.a3 a6?! Aiming to put
the knight on the tempting c5-square, but it
turns out not to be doing much there.
[ The natural 18...c6 seems perfectly
acceptable for Black. As so often in the
Tarrasch, the obvious, centralizing move
proves to be best. ]
19.c1
[ 19.g4 g6 20.h4 , followed by taking on
g6 and playing b3-b4, gives White some
advantage. ]
19...ac5 20.h4 e7
[ Again 20...f6 is Houdini's recommendation,
for reasons which will become clear on the
next move. ]
21.df5! Giri has managed to obtain co-
ordination in an unusual way. Now the
weakness of g7 is apparent. xh4
Allowing a nice tactic.
[ Swallowing his pride and retreating with
21...f8 would have kept White's advantage
within manageable bounds. ]
EXERCISE: Find a shot for White.
ANSWER: 22.xg7! That's a big pawn to
lose. g5
[ 22...xe2 23.xe8 g5 24.f6+ xf6
25.xf6 is similar – White has a strong
initiative since his bishops are so powerful. ]
23.d1 g6 24.xe8 xe8 25.xd5
Black's position collapses. e6 26.d1
xd5 27.xd5 e7 28.b4 a4 29.a1
ec3 30.d7 f6 31.f1 b5 32.xa7 g7
33.g4 h6 34.c6 c2 35.a6 d4 36.e3
d3+ 37.e1 f6 38.h4 c4 39.g5 hxg5
40.hxg5 xg5 41.xa4
1-0
D30
Korchnoi,V
Collins,S
17: San Sebastian 2011
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.cxd5 exd5
5.g3 c6 6.g2 f6 7.0-0 e7 8.e3
A favourite of Korchnoi, who explains the
ideas as follows: "One of the best moves in
this position is 8 Be3, and if 8...c4 9 b3 cxb3
10 Qxb3, when the d5-pawn is weak and still
requires defending. In addition, White has
pressure on the half-open b-file. " Korchnoi
goes on to note that, while he has not won all
his games with this setup, he has always had
an advantage in the middlegame.
Looking ahead, I can say that this passage
perfectly describes the current game!
c4
[ 8...cxd4 9.xd4 0-0 10.c3 leads to a
position from the 9 Bg5 cxd4 variation, but
with the black h-pawn on h7 rather than h6.
As ...h7-h6 is generally a useful move, I
can't recommend this line for Black. ]
9.b3
[ 9.e5 0-0 10.b3 ( 10.xc6 bxc6 11.b3
a5 12.bxc4 dxc4 was fine for Black in J.
Glud-R.Alonso Garcia, Catalonian Team
Championship 2012, since 13.xc6 h3
14.xa8 xa8 gives full compensation for
the pawn; White can't hold on to the
exchange with 15.e1? as g4!
leads to a very strong attack ) 10...cxb3
11.xb3 transposes to the main game,
( while 11.axb3 b4 12.d3 f5 13.xb4
xb4 was very comfortable for Black in J.
Glud-S.Brynell, Danish Team Championship
2010 (½-½ in 31). )]
9...cxb3 10.xb3 0-0 11.e5
[ 11.c1 h6 12.e5 a5 13.a4 e6
14.d2 c4 15.xc4 dxc4 16.e3 c8
17.b5 b8 18.e4 a6 19.b2 b5
was level in V.Epishin-V.Akobian,
Minneapolis 2005 (½-½ in 32). ]
11...a5
[ Instead: The computer's preference
11...d6 has been tested in a game: 12.c3
xe5 13.dxe5 xe5 14.ad1 xc3
15.xc3 e6 16.d4 b6 and White had
full compensation in Se.Ivanov-C.Jepson,
Swedish Team Championship 2012 (1-0 in
54) – of course, he can also just take on f6
and d5 with an endgame which should
probably be drawn. ]
 [ 11...b6 has a great lineage: 12.c1
( Kasparov gives 12.xb6 axb6 13.c1
b4 14.c3 e6 15.a3 c6 16.b5
fc8 as unclear ) 12...xb3 13.axb3 b4
with an endgame which Black went on to win
in V.Korchnoi-G.Kasparov, Candidates
(6th matchgame), London 1983 (0-1 in 78).
Worth investigating if you don't like the
main line, but I think White should be a bit
better in this endgame (his rooks are doing a
good job pounding the queenside), so I
prefer 11...Na5. ]
12.d3
[ 12.a4
A) 12...a6 13.d2 c4 14.xc4 b5
15.a5 bxc4 16.xd8 xd8 17.a5 d7
( but 17...e8! equalizes ) 18.c3 b7
19.ab1 is some old analysis of
Kasparov's, which the great man thought
led to equality – modern computers prefer
White here;
B) 12...e6
B1) 13.d2 c4 14.xc4 dxc4 15.e3
was J.Norri-K.Kiik, Finnish Team
Championship 2013. Now Black played
c8 , seeking to trade light-squared
bishops with ...Bh3. ( The simple
15...d5 looks fine too. );
B2) 13.c1 was S.Baumegger-P.Pisk,
Austrian League 2013. Here Black
played b6 , ( but he could have been
more principled: 13...a6! , preparing ...
Nc4 with a good game. Note that, with
the queen on a4, this can be played
immediately since Nxc4 can be met by
the intermezzo ...b7-b5. Otherwise, just
play ...b7-b5 and then ...Nc4. );
B3) 13.-- and now: ]
12...e6 13.c3 c8 14.ab1! A strong
move, preventing Black's idea of planting his
knight on c4.
[ Instead, 14.g5 c4 15.ab1 d6
16.bc1 a5 17.d2 d8 18.g5
was drawn in Y.Kruppa-V.Potkin, Kiev
2001. ]
14...g6 15.d2 c4 This is the move Black
wants to play, but I didn't assess the resulting
position properly.
[ The alternatives are sounder: 15...a6
16.a4 ( 16.a4 seems more principled )
16...b5 17.c5 f5 18.bc1 , as in L.
Ftacnik-V.Vojtek, Slovakian Team
Championship 2012, could be well met by
c4 19.xc4 when either capture on c4 is
fine for Black. I prefer bxc4! , keeping a
stake in the centre and, provided we can
prevent White from playing e2-e4, giving
Black every reason to play for a win. ]
[ 15...b6 is more solid. Then 16.b5 c4
17.xc4 xc4 is equal, since 18.xa7 d7
19.b5 fc8 gives Black full compensation
for the pawn. ]
16.xc4 xc4 EXERCISE: How should
White play?
ANSWER: 17.xb7! This move, which is
based on an exchange sacrifice, was
confidently cracked out by Korchnoi and
Black's position immediately becomes critical.
b4 18.xb4! xb4 19.g5 Now we can
see White's compensation. He already has a
pawn for the exchange, plus a much better
structure. The dark squares around my king
are just horrible, and the pin of the f6-knight is
extremely annoying.
Still, we have to make a move I suppose...
c4 20.f4? Going for the knockout punch.
This was the move I had feared but, in fact, it
squanders White's advantage.
[ The straightforward 20.xd5? doesn't work
either: xd5 21.xd5 xd5 22.xf6 f5!
and Black defends his king while preparing
his counterplay with ...Rc2. ]
[ Instead, the calm 20.d1 would have kept
some advantage for White; for instance, e8
21.xd5 xd5 22.xf6 xf6 23.xd5 c7
and White has two pawns for the exchange
plus a monster bishop on d5. Black should
be able to draw with careful defence (an
exchange sacrifice on d5, going into a
possibly tenable pawn down position with
major pieces, might be a resource), but
White is having all the fun. ]
EXERCISE: Find a way for Black not to lose on
the spot.
ANSWER: 20...h6!! Probably the best move
I've ever played, especially considering the
calibre of my opponent and how the position
turns 180 degrees.
[ 20...a5 21.f5! is not playable for Black:
xc3 22.xc3 xc3 23.fxe6 e4
( 23...fxe6 24.xf6 followed by Bh3 is
horrible ) 24.xe4 dxe4 25.d5 wins for
White in view of fxe6 26.d6 . ]
[ Partial credit for 20...e8!? . White is better
after 21.f5! ( 21.h4 is dangerous and
perhaps stronger ) 21...xf5 ( 21...gxf5
22.h4 again looks horrible for Black ) 22.g4
xc3 23.xf6 c2 24.xc2 xc2 25.xd8
xd8 26.f2 is a slightly favourable
endgame for White, but Black should hold. ]
21.xh6 39.g2 b6 40.a2 b3 41.h3 a3
[ 21.h4 a5 22.f5 xc3 23.xc3 xc3 42.c2 c3 43.a2 f6 44.h5 b4 45.g2
24.fxe6 g5! is the difference compared to b2 46.a1 a2 47.h6 cc2
20...Qa5. After 25.e7 e8 26.xg5 hxg5 0-1
27.xf6 xe7 28.f2 c2 , Black is the
one with winning chances. ]
21...b6 This, of course, is the point
(otherwise Black's last move would have been
poor). The counterplay on the a7-g1 diagonal
based on ...Rxd4 and, if necessary, ...Ng4 is
actually quite difficult to deal with.
22.h3?? Taking away the g4-square and so
preparing Qe3 in response to ...Rxd4, but it
wastes crucial time.
[ 22.e3?? also loses to fc8 with perfect co-
ordination – see how the white queen can
never reach the dark squares on the
kingside. ]
[ 22.b1 is logical, aiming for a queen trade,
but after xd4 23.xb6 xd2 , the black
rook has broken through and the white king
is still uncomfortable. On 24.d6
I was planning to continue in dynamic style
with g4!? ( 24...c2! is even stronger )
25.xf8 d4 , when the mating net forces
White to search for a draw with 26.xd4
( after 26.h3 dxc3 27.hxg4 xf8
, the passed c-pawn looks too strong )
26...xd4 , which will be difficult to achieve
since he is losing a pawn after 27.a3
xa2! . ]
[ 22.xf8! xd4 23.e1! was the only
defence, which I'm pretty sure I had missed.
Luckily Black has more than enough
compensation after c4+ 24.e3 xf8
in view of his extremely active pieces and
the continuing weakness of the white king,
but White should have gone for this. ]
22...fc8!
[ Actually 22...xd4 23.e3 e8
was winning too, but my move is stronger. ]
23.h2 xd4 The computer prefers taking
on c3 (or even the bizarre 23...Bf5!?!?), but I
was delighted to trade into an endgame
where White's queenside will be demolished
by my rooks.
24.xd4 xd4 25.b5 a4 26.f5
The only attempt at counterplay, but it's not
sufficient. gxf5 27.d6 c2 28.xf5 xf5
29.xf5 xe2 The threat of doubling on the
seventh means that Black saves his knight.
30.g1 axa2 31.f1 e6 Now only basic
care is required to bring home the full point.
32.f4 a5 33.e5 e4 34.h4 g6 35.f3
a4 36.d3 d2 37.a3 e6 38.f4 d4
 Back to Contents Page
Chapter Six

Symmetrical Tarrasch
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nc3 c5 4 e3 Nf6 5 Nf3 Nc6

The Symmetrical Tarrasch is fundamentally different to the fianchetto lines we have just discussed at length. White
abandons hope of creating early pressure on d5, since his light-square bishop no longer attacks the pawn, while its
dark-squared counterpart can't come to g5 anymore. Indeed, it's no longer clear if there will be any IQP to attack,
and in fact White is often the side which ends up with the IQP in this line (albeit not in the variations I recommend
for Black).
The black side of this position (in contrast to the regular Tarrasch) is regularly endorsed by 2700+ GMs via a
Symmetrical English move order (for instance, 1 c4 c5 2 Nf3 Nc6 3 Nc3 Nf6 4 e3, when Black often plays 4...e6 5
d4 d5). Accordingly we are on slightly firmer ground from a theoretical standpoint here. My main concern with
Black's position in the previous lines was hanging on to the d5-pawn without making major concessions, and so I
(and most Tarrasch players, I expect) automatically feel more comfortable when the white pawn lands on e3.
Nevertheless, this is a serious line and demands serious work.
Black has a wide choice from a structural perspective, and many lines involve early ...c5xd4 or ...d5xc4 trades. My
choice is more consistent with the Tarrasch, keeping the central tension for as long as possible and inviting White
to give us an IQP. This choice is only partially motivated by principle; if it were otherwise, then all sorts of c4xd5
and e2-e3 move orders could throw us out of book.
White has many reasonable moves here, of which the mysterious-looking...
6 a3
...is by far the most popular. Control of the b4-square is useful either if White gets an IQP (after ...c5xd4, e3xd4
...d5xc4) or if White expands on the queenside with d4xc5 and b2-b4. SeeAronian-Melkumyan for some of the
latest developments.
White's many alternatives, including the popular and principled 6 cxd5, are covered inCarlsen-Radjabov. As
always, don't let the result of the game fool you into thinking that Magnus got anything out of the opening.

Key Notes

1. The Symmetrical Tarrasch is a completely different concept to Rubinstein's kingside fianchetto. The resulting
positions have more in common with other IQPs (such as those from the Nimzo-Indian), since White's bishop is
developed more passively, normally to e2, and doesn't put immediate pressure on the d5-pawn.
2. 6 a3 is a move which has been popular for a long time. The equalizing approach inAronian-Melkumyan looks
very reliable - if White has anything here, it's not much. All the lines where Black gets an IQP look more than fine for
the second player.
3. 6 cxd5 exd5 leaves Black with an IQP and good chances. If, as inCarlsen-Radjabov, White captures on c6, a
different structure arises, but one where Black is solid and has good chances of counterplay.

1 d4 d5 2 c4 e6 3 Nc3 c5 4 e3 Nf6 5 Nf3 Nc6 6 a3 – Aronian-Melkumyan

6 cxd5 – Carlsen-Radjabov

Back to Contents Page

Symmetrical Tarrasch
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.e3 f6 5.f3
c6 The Symmetrical Tarrasch is
fundamentally different to the fianchetto lines
we have just discussed at length. White
abandons hope of creating early pressure on
d5, since his light-square bishop no longer
attacks the pawn, while its dark-squared
counterpart can't come to g5 anymore. Indeed,
it's no longer clear if there will be any IQP to
attack, and in fact White is often the side
which ends up with the IQP in this line (albeit
not in the variations I recommend for Black).
The black side of this position (in contrast to
the regular Tarrasch) is regularly endorsed by
2700+ GMs via a Symmetrical English move
order (for instance, 1 c4 c5 2 Nf3 Nc6 3 Nc3
Nf6 4 e3, when Black often plays 4...e6 5 d4
d5). Accordingly we are on slightly firmer
ground from a theoretical standpoint here. My
main concern with Black's position in the
previous lines was hanging on to the d5-pawn
without making major concessions, and so I
(and most Tarrasch players, I expect)
automatically feel more comfortable when the
white pawn lands on e3. Nevertheless, this is
a serious line and demands serious work.
Black has a wide choice from a structural
perspective, and many lines involve early ...
c5xd4 or ...d5xc4 trades. My choice is more
consistent with the Tarrasch, keeping the
central tension for as long as possible and
inviting White to give us an IQP. This choice
is only partially motivated by principle; if it
were otherwise, then all sorts of c4xd5 and e2-
e3 move orders could throw us out of book.
White has many reasonable moves here, of
which the mysterious-looking...
6.a3 ...is by far the most popular. Control of
the b4-square is useful either if White gets an
IQP (after ...c5xd4, e3xd4 ...d5xc4) or if White
expands on the queenside with d4xc5 and b2-
b4. See Aronian-Melkumyan (Game 18) for
some of the latest developments.
[ White's many alternatives, including the
popular and principled 6.cxd5 , are covered
in Carlsen-Radjabov (Game 19). As always,
don't let the result of the game fool you
into thinking that Magnus got anything out of
the opening. ]
A35
Aronian,L
Melkumyan,H
18: German League 2013
[Sam Collins]
1.c4 c5 2.f3 c6 3.c3 f6 4.e3 e6
5.d4 d5 6.a3 A useful waiting move. White
controls the b4-square and, in some lines,
prepares d4xc5 and b2-b4 followed by Bb2.
In addition, by keeping his bishop on f1, he
ensures that ...d5xc4 can be met by Bxc4
without loss of time.
a6 This is my recommendation, as it seems
most consistent with the main ideas of the
Tarrasch. Black is still happy to play with an
IQP.
[ By contrast, 6...cxd4 7.exd4 e7
leads to a position where White could end
up with an IQP himself – although 8.c5!?
has become fashionable, gaining queenside
space and playing for a bind. ]
7.dxc5 White prepares to expand on the
queenside.
[ Alternatively: 7.d3 is, on one level, a
ridiculous move, for a reason which will
become apparent; i.e. dxc4 8.xc4 b5
. If you want ideas on how to play this
position, check the main game, since
White's loss of tempo means that you now
have the same position as after Aronian's
9th move, but with colours reversed! Still, it
can be confusing playing a white opening
with the black pieces, so here are a couple
of examples: 9.a2 b7 10.0-0 c7
11.e2 d8 12.d1 e7 13.dxc5 ( 13.d5
exd5 14.xd5 xd5 15.xd5 was agreed
drawn in D.Lemos-K.Mekhitarian, Villa
Martelli 2012 ) 13...xd1+ 14.xd1 e5
15.e1 xc5 was good for Black in A.
Beliavsky-R.Ruck, Slovenian Team
Championship 2012 (½-½ in 60). ]
[ 7.h3 cxd4 ( 7...h6 8.d3 e7 9.0-0 0-0
was also decent in S.Ernst-C.Daly, Dublin
2012 ) 8.exd4 dxc4 9.xc4 e7 10.0-0
0-0 , as in R.Kempinski-K.Piorun, Polish
Championship, Chorzow 2013, looks like a
comfortable IQP position for Black, since h2-
h3 isn't especially useful. ]
[ Keeping the tension with the rare 7.b3
is interesting and can lead to numerous
different pawn structures depending on
whether Black captures on c4 and/or d4. I
suggest the simple e7 followed by castling,
after which a complex middlegame results
 where Black has a good share of the
chances.. ]
[ 7.cxd5 exd5 8.e2 ( 8.e5 d6 9.xc6
bxc6 10.dxc5 xc5 was at least not worse
for Black in D.Lintchevski-A.Smirnov, Kazan
2013 ) 8...c4 ( 8...e7 9.dxc5 xc5 10.0-0
0-0 11.c2 d6 12.d1 e6 was equal in
Zhao Xue-B.Kovanova, Russia-China rapid
match, St. Petersburg 2012 ) 9.e5 d6
10.f4 was L.Lenic-J.Bejtovic, European
Championship, Plovdiv 2012, and now a5!
, aiming for the b3-square, is good for
Black. ]
7...xc5 8.b4 a7 Again Black has a choice
of bishop retreats.
[ 8...b6 seems less logical to me since
Black can no longer play ...b7-b5 easily. All
the same, 9.b2 dxc4 10.xd8+ xd8
11.xc4 d7 12.e2 c8 13.d3 0-0
14.ac1 e7 led to a pretty level endgame
in Z.Ribli-J.Pinter, Austrian League 2010
(½-½ in 32). ]
[ 8...d6 9.b2 and now:
A) 9...a5 was tried a couple of times by
Akobian at the 2008 World Mind Sports
Games, but with poor results: 10.c5 axb4
( 10...e7 11.b5 b8 12.a4
was already good for White in J.Markos-V.
Akobian, Beijing rapid 2008; 1-0 in 55 )
11.axb4 xa1 12.xa1 e7 13.a3 0-0
14.b5 d7 15.0-0 (White has an edge
due to his queenside space advantage)
b6?! ( 15...c7 at once is preferable )
16.a4 c7 was A.Korobov-V.Akobian,
Beijing blitz 2008, and now, instead of
17.d4 , ( White should have played
17.xc6 xc6 18.b5 with a clear
advantage. );
B) 9...0-0 10.cxd5 exd5 11.e2 e6
12.0-0 and Black has a decent position.
For example, c8 ( 12...e7 13.c1
fe8 14.h3 ad8 15.d4 e5
was fine for Black in R.Laxman-N.Short,
Commonwealth Championship, Gauteng
2011; 0-1 in 45 ) 13.c1 e7 ( 13...b8
14.a4 e7 15.c5 e4 16.d3 xc5
17.xc5 a5 18.b5 axb4 19.axb4 f6
20.d4 was agreed drawn in V.Potkin-E.
Bacrot, Moscow 2009 ) 14.d3 fd8
15.fd1 b5 and Black had his full share of
the chances in L.Le Quang-Y.Pelletier,
French Team Championship 2012 (½-½ in
55).;
C) 9...-- ]
9.b2
[ 9.cxd5 exd5 10.b5 ( 10.e2 0-0 11.0-0
d4 12.exd4 xd4 13.xd4 xd4
14.xd4 xd4 is equal, S.Mamedyarov-A.
Istratescu, European Team Championship,
Gothenburg 2005; ½-½ in 23 ) 10...d4
( also possible is 10...axb5 11.xb5 b6
, intending ...Ba5+ ) 11.bxc6 dxc3 12.xd8+
xd8 13.e5 c7 14.c4 e8 15.f4
( 15.cxb7 xb7 16.xf7 was better, but
Black is still for choice after e4 ) 15...bxc6
16.xf7 e7 17.a4 g4 18.xg4 xg4
with a clear advantage for Black in A.
Aleksandrov-V.Gashimov, Warsaw (rapid)
2009 (0-1 in 64). ]
9...0-0 10.c2 The best move, but if you think
White is messing around while Black is
developing pieces and seizing space in the
middle, you're basically right.
[ 10.e2 dxc4 11.xc4 xd1+ 12.xd1
d8 13.e2 ( 13.xd8+ xd8 14.e2 b5
15.d3 b7 was level in A.Onischuk-G.
Kaidanov, US Championship, St. Louis
2012 ) 13...d7 14.d3 ac8 15.d2 e8
16.e4 xe4 17.xe4 xd2+ 18.xd2
e7 19.c1 xc1 20.xc1 b5+ 21.d1
a4+ 22.e2 b5+ 23.d1 a4+
24.e2 and a draw was agreed in Zhou
Jianchao-Bu Xiangzhi, Chinese
Championship, Xinghua 2013. ]
[ The following game is a great
demonstration of the attacking potential of
the IQP: 10.cxd5 exd5 (this already looks
good for Black since White can't
conveniently stop ...d5-d4) 11.a4?
( Moskalenko suggests 11.e2 d4 12.exd4
xd4 13.xd4 xd4 14.0-0 e6
, limiting Black to an edge ) .
EXERCISE: How should Black continue?
ANSWER: 11...d4! (a thematic
breakthrough, opening the e-file while the
white king is stuck in the centre) 12.exd4
( Moskalenko suggests 12.xd4 xd4
13.xd4 xd4 14.xd4 xd4 15.exd4
e8+ 16.e2 g4 with some initiative for
Black ) 12...g4 13.e2 xf3! (eliminating
a defender of the d4-pawn and the e-file)
14.xf3 e8+
A) Moskalenko notes that 15.e2 fails to
xd4 16.xd4 xd4 17.c1 b5 ( and
17...xf2+ 18.xf2 e4+ 19.g1 g5
, with a decisive attack, is even stronger );
B) 15.f1 xd4
B1) perhaps White should have tried
16.c5 xc5 17.bxc5 xf3 18.xd8
( but not 18.xf3?? e4! and wins –
 Moskalenko ) 18...exd8 19.gxf3
with some saving chances;
B2) 16.xd4 xd4 17.c1 ( 17.a2
was better, though Black is still on top
after e4 ) 17...e4 (the black pieces
dominate and it's hard to offer White
any constructive advice) 18.xe4 xe4
19.g3 e6 20.c2 d5 21.g1 ae8
22.c5 d6 23.g4 h5 24.f4 xc5
25.xc5 e1+! and White resigned in S.
Halkias-V.Moskalenko, Chalkida 1998. ]
10...d4 The immediate central breakthrough
has proven very reliable in recent games,
though Black's choice here is a matter of
taste.
[ 10...e7 (chosen by Anand against
Aronian in the 2011 Amber rapid event)
11.d1 ( 11.d3 dxc4 12.xc4
was successful in both T.Nyback-D.Tjiam,
European Cup, Plovdiv 2010, and O.Cvitan-
O.Kurmann, Swiss Team Championship
2012, but for some reason in both games
Black avoided the natural b5 13.d3 b7
followed by ...Rac8 with equality; 11.e2 h6
12.0-0 dxc4 13.xc4 b5 14.a2 b7
was nothing for White in R.Vasquez
Schroder-M.Cornejo, Cochabamba 2013,
0-1 in 45 ) 11...d8 12.e2 and, frankly, it's
hard to see how White can hope for an
advantage playing in such a way.
A) 12...dxc4 13.xd8+ xd8 14.e4
( 14.xc4 b5 15.d3 b7 was equal in T.
Georgescu-D.Dumitrache, Cap Aurora
2013; ½-½ in 31 ) 14...xe4 15.xe4
d7 16.xc4 c8 and Black had no
problems in K.Miton-B.Macieja, Polish
Championship, Warsaw 2012 (½-½ in
31).;
B) 12...d4 13.exd4 xd4 14.xd4 xd4
15.0-0 e5 16.a4 e6 17.c5!
gave White an edge in L.Aronian-V.Anand,
Monte Carlo (rapid) 2011 (1-0 in 59).;
C) 12...h6?! is a little slow: 13.0-0 dxc4
14.xd8+ xd8 15.e4 e8 16.xc4
and White had the initiative in H.Banikas-
J.Borisek, Istanbul Olympiad 2012 (½-½ in
64).;
D) 12...-- ]
[ 10...dxc4 11.xc4 e7 12.a2 d7
13.0-0 ac8 14.ad1 fd8 also looked
solid and sound in N.Vitiugov-K.Sakaev,
FIDE World Cup, Khanty-Mansiysk 2007
(0-1 in 92). ]
11.exd4 xd4 12.xd4 xd4!
[ 12...xd4 is less precise. 13.d3 h4
14.g3 h6 15.e4 xe4 16.xe4
and Black's pieces were poorly placed in T.
Nyback-J.M.Degraeve, German League
2009 (1-0 in 39). ]
13.e2 e5 14.0-0 e6 15.a4
[ 15.ad1 c7 16.c5 ad8 17.f3 h6
18.fe1 fe8 didn't offer White much in T.
Sanikidze-E.Bacrot, Baden-Baden 2013
(½-½ in 49). ]
15...xb2 Instead of this move, Black can try
to keep the tension.
[ 15...c8 16.ad1 e8 17.c5 e7
18.xd4 exd4 19.xd4 didn't give him
enough for the pawn in V.Tkachiev-T.
Markowski, European Blitz Championship,
Warsaw 2012 ]
[ but 15...c7 16.xd4 exd4 17.h3 ad8
was fine for Black in L.Le Quang-Zhou Weiqi,
Asian Cup, Zaozhuang 2012 (½-½ in 47). ]
16.xb2 c7 17.c5 f5 18.fe1 fe8
19.ad1 a5 20.f1 axb4 21.axb4 b6
22.b3 h6 White has nothing here. 23.h3 e4
24.c5 bxc5 25.xc5 b6 26.d4 ad8
27.ed1 xd4 28.xd4 c8 29.a6 b8
30.b5 g6 31.d6 a7 32.d4 h7
33.a4 e8 34.c5 b8 35.b4 b6
36.e5 a8 37.d4 b8 38.d6 a7
39.d4 c7 40.d7 xd7 41.xd7 c1+
42.d1 c3 43.d4 c1+ 44.h2 f4+
45.g3 f3 46.e3 f5 47.d6 e5 48.f4
b2 49.e3 b4 50.d5 c4 51.d2 f1
52.g4 e8 53.h4 h5 54.d5 e6 55.g5
c4 56.c5 b4 57.d5 c4 58.c5 b4
59.d5
½-½
A35
Carlsen,Ma
Radjabov,T
19: Sandnes 2013
[Sam Collins]
1.c4 c5 2.f3 c6 3.c3 f6 4.e3 e6
5.d4 d5 6.cxd5 The most critical move.
White's attempts to delay the development of
his king's bishop just let Black develop in
comfort, as we saw in the last game, so
releasing the central tension is the natural
way of trying to make use of White's extra
tempo.
exd5! The natural reaction for a Tarrasch
player.
[ 6...xd5 is a position from the Semi-
Tarrasch which is a completely different
 opening, where White often ends up with an
IQP himself. ]
7.b5 By putting pressure on the c6-knight
White discourages ...c5xd4. The resulting
positions are popular with strong GMs who
presumably think that White has chances of
an edge, but I really don't see why Black
should have anything to worry about.
[ Instead, 7.e2 has been tested in over a
thousand games. Going into similar
positions to the main game with 7...Bd6 (or
7...Be7) 8 dxc5, although fully playable,
leaves White's bishop better placed on e2
than on b5. I would suggest instead cxd4
8.xd4 d6 as a simple and reliable
solution.
A) 9.f3 0-0 10.0-0 e5 is fine for
Black:
A1) 11.d3 g4 ( 11...c7 12.xc6?!
bxc6 13.e4 dxe4 14.xe4 xe4
15.xe4 f6 and Black was already
better in T.Banusz-A.Volokitin,
European Championship, Rijeka 2010 )
12.xg4 xg4 13.f3 d6 14.h3
xc3 15.xc3 f6 was equal in Z.Ribli-
J.Borisek, Slovenian Team
Championship 2008 (½-½ in 60).;
A2) 11.e2?! a6 12.f3 c7 13.b3
e8 14.b2 d6 15.g3 h3 16.e1
ad8 gave Black an ideal IQP formation
in G.Izsak-P.Tregubov, German League
2013 – in fact, with White's loss of a
tempo (Be2-f3-e2 spent two extra moves,
compared to one extra move on ...Bd6-
e5-c7), I think we have directly
transposed into an IQP position
commonly arising from the 2...d5 line of
the c3-Sicilian, and in which the side with
the IQP scores heavily.;
A3) 11.--;
B) 9.0-0 0-0 is logical.
B1) 10.e1 e8 11.xc6? bxc6 12.b3
was K.Drozdowski-M.Parligras,
European Championship, Legnica 2013,
and now c7 13.g3 h5! would have
left Black clearly better since the h-pawn
is untouchable. As often happens in IQP
positions, after a couple of passive
moves White now seems to be playing
"as Black", with less space and facing a
strong initiative.;
B2) . The danger inherent in these
positions for White is well illustrated by
the following game: 10.d3 e8
11.xc6? (why does everyone keep
taking on c6 in these positions? – the
black pawn structure is considerably
improved and the d4-knight can no
longer drop back to cover the kingside)
bxc6 12.b3 g4 ( 12...c7 13.g3 d7
, aiming for h3, looks dangerous as
well ) 13.xg4 strategic capitulation;
( White had to try 13.h3 ) 13...xg4
14.b2 e6? ( 14...h4 15.f4 e6
was much stronger ) 15.f4 ( 15.e2
kept chances to defend ) 15...e7
16.ae1 c5 17.c1 g6! 18.d2 f5
19.a6 xe3+ and White resigned in J.
Lopez Martinez-A.Fier, Sabadell 2010.;
B3) 10.b3 xd4 11.xd4 e8 12.b2
e5 13.d2 was D.Khismatullin-S.Ionov,
Russian Team Championship 2010, and
now Black could have equalized
comfortably with g4 14.xg4 xg4
15.h3 f6 16.fd1 c8 17.d3 a5
.;
B4) 10.--;
C) 9.-- ]
7...d6 8.0-0 0-0 9.dxc5 xc5 10.b3 g4
11.b2 QUESTION: How do you assess this
position?
ANSWER: I think any Tarrasch player should
be very happy with a position like this. He is
fully developed, all the pieces are on the
board, and the d5-pawn isn't under serious
pressure. Black's challenge is to mobilize his
major pieces, and if he can generate a
kingside initiative, so much the better.
a6 Black has tried a range of moves in this
position. We will only focus on a few of the
more logical ones.
[ 11...d6 has been extensively tested. We
will confine ourselves to one example: 12.h3
h5 13.e2 a6 14.h4 xe2 15.xe2
e8 16.c1 e4 17.f5 e5 18.xe5
xe5 was level in V.Kramnik-J.Polgar, Baku
(rapid) 2010 (½-½ in 67). ]
[ The most popular line is 11...c8 12.c1
and then:
A) 12...a6 13.e2 ( here 13.xd5!
seems good for White; for instance, xd5
14.xd5 xd5 15.xc5 cb4 16.xc8
xc8 17.c4 and Black needs to prove
his compensation for the pawn ) 13...a7
14.h3 h5 15.h4 g6 16.xg6 hxg6
17.f3 d4 18.exd4 xd4 19.a4 xb2
20.xb2 d4 21.xc8 xf3+ 22.xf3
xc8 was level in V.Topalov-S.Karjakin,
Stavanger (blitz) 2013 (0-1 in 64).;
B) 12...d6 has also been tested in
grandmaster games:
B1) 13.e2 a6 14.d4 ( 14.h3 xf3
15.xf3 fd8 16.c2 was Z.Ribli-A.
Adorjan, Riga Interzonal playoff,
Budapest 1979, and now e5
would have equalized ) 14...xd4
15.xg4 xg4 16.xg4 e5 17.fd1
xh2+ 18.f1 e6 19.f3 e5
20.a3 d6 21.xd6 xd6 and a draw
was agreed in J.Speelman-L.Portisch,
World Cup, Brussels 1988.;
B2) 13.a4 a3 14.xa3 xa3
15.xc6 xf3 16.gxf3 xc6 17.xc6
bxc6 was equal in T.Hillarp Persson-S.
Brynell, Swedish Team Championship
2006 (½-½ in 56).;
B3) 13.--;
C) 12...d6 leaves White with a wide
choice:
C1) 13.e2 xf3 14.gxf3 e5!
was level in S.Kalinitschew-M.Thesing,
German League 2003, and B.Golubovic-
B.Vuckovic, European Championship,
Rijeka 2010. A draw by repetition after
15.a3 d6 16.b2 e5 would be a
logical finish.;
C2) 13.xc6 bxc6 14.e2 xf3
15.gxf3 e8 16.g3 left Black with no
problems in I.Rozum-D.Ayupov, Kazan
2013 (1-0 in 60).;
C3) 13.h3 h5 ( 13...e6 , avoiding
the exchange of light-squared bishops,
is also sensible, as in Cu.Hansen-B.
Gulko, Malmö 2001 ) 14.e2 a6 15.h4
xe2 16.xe2 e5 17.xe5 xe5
18.d4 e8 19.f5 xc1 20.xc1
d7 21.fg3 h6 with a typical position
in B.Grachev-N.Chadaev, Russian Team
Championship 2013. White has
managed to exchange several pieces
and definitely enjoys an edge (1-0 in
55).;
C4) 13.e2 is more critical. Now Black
has tried a number of moves but we will
focus on the caveman approach b8
which has been endorsed by Kasparov
and Spraggett.
C4a) 14.h3 h5 15.h4 (a typical
idea, aiming to trade Black's light-
squared bishop)
C4a1) 15...xe2 16.xe2 e8
17.f5 d7 18.eg3 e5
basically equalized in M.Voiska-I.
Farago, Bolzano 2000; White was
actually a bit worse after 19.a3?!
( while on something like 19.xe5
xe5 , if White does have an
advantage, it is minimal );
C4a2) 15...d6 16.g3 xe2
( 16...g6 17.b5 e6 18.xf6
xf6 19.xg6 hxg6
was Y.Seirawan-N.De Firmian, US
Championship, Key West 1994,
when White could have simply taken
the pawn ) 17.xe2 (this position is
assessed as better for White by both
Ftacnik and Curt Hansen) fd8
( 17...e6 18.g2 e5 19.xe5
xe5 20.f3 e4 21.c3 e6
22.d3 e5 23.xe5 xe5
24.c2 was beginning to look like a
trademark Karpovian squeeze in A.
Karpov-J.Lautier, Monte Carlo
blindfold rapid 2000; 1-0 in 93 )
18.d4 e5 19.xc8 xc8
20.df3 e6 21.g5 d7 22.f4
g6 23.xg6 hxg6 24.d3 a6
25.d4 e4 26.xe4 dxe4
27.xe4 xh3 28.g2 g4
29.f3 h3 30.g2 g4 31.f3
h3 and a draw was agreed in V.
Zvjaginsev-C.Lutz, Essen 2002.;
C4b) 14.e1 e8 15.g3 h5!?
showed a typical response to g2-g3,
increasing Black's control over g4 and
preparing to soften up the white
kingside pawns with ...h5-h4. Black
went on to lose a mammoth game in T.
V.Petrosian-B.Ivanovic, Bar 1980 (1-0
in 123), but it had nothing to do with
the opening.;
C4c) 14.b5 and now:
C4c1) 14...e4 15.bd4 ( 15.fd4
xe2 16.xe2 was D.Svetushkin-D.
Hummel, Canarias online blitz 2004,
when g5 would have been active
and strong ) 15...e8 16.h3 xf3
17.xf3 d6 18.d3 , and now
Black went on to win with the direct
g5!? in J.Sunye Neto-G.Kasparov,
World Student Team Championship,
Graz 1981. ( Instead, 18...cd8
with equality is sounder; while the
pawn sacrifice 18...d4!? 19.exd4
cd8 with full compensation is
interesting as well. );
C4c2) 14...e8 has been endorsed
by Tarrasch specialists Murray
Chandler and Kevin Spraggett:
15.d3 ( or 15.bd4 d6 16.g3
 h3 17.e1 e4 with good play for
Black in T.Markowski-K.Spraggett,
Cannes 1994 ) 15...e4 16.bd4
d6 17.g3 h5 18.xc6 bxc6
19.h4 h3 20.fd1 h6 21.f3
g5 22.g2 h4 and, while Black
went on to lose in A.Karpov-M.
Chandler, Bath 1983 (1-0 in 36), this
again had more to do with the
strength of his opponent than with
any defect in his position. Black has
active and promising play here.;
C4c3) 14...--;
C4d) 14.--;
C5) 13.--;
D) 12...-- ]
12.xc6 Changing the structure and, with it,
the plans. Taking the knight means the ...d5-
d4 advance is not really an issue for White
anymore. He aims to build pressure on the c6-
pawn after Rc1, and restrain any freeing ...c6-
c5 break.
[ Alternatively, White can keep the pawn
structure as it is with 12.e2 , after which
a7 is logical to get the bishop out of the
way.
A) 13.d4 d7 14.xc6 xc6 15.f3
was R.Loncar-L.Totsky, Zadar 2000, and
now Black retained the tension with d6
, which was okay (½-½ in 37),
( but had he wanted to he could have
equalized immediately with 15...d4
16.exd4 xd4 17.xc6 bxc6 18.a4 c5
. );
B) 13.h3 e6 14.c1 d6 15.d3
( 15.d3 ad8 16.fd1 b8 17.b1
e4 18.a3 b4 19.d4 a5 20.c3
was A.Ornstein-S.Brynell, Swedish
Championship, Haparanda 1994, when
xc3 21.xc3 c8 would have retained
the balance ) 15...ad8 16.e2 e4
17.ed4 e5? ( instead, 17...xd4
18.xd4 xd4 19.xd4 d7 is just an
edge for White ) 18.xe5 xe5 19.c6!
xb2 20.xd8 xd8 21.xe4
was clearly better for White in L.Portisch-
G.Ligterink, Wijk aan Zee 1985 (1-0 in
29).;
C) 13.c1 d6 and now:
C1) 14.d4 e5? ( Ftacnik gives
14...xd4 15.xg4 e5 16.h3 d4
17.exd4 xd4 with equality ) 15.f4!
xe2 16.fxe5 xd1 17.exd6 g4
18.xf6!! xd4 19.exd4 gxf6 20.xd5
and the d-pawn was too strong in V.
Filippov-E.Romanov, Russian Team
Championship 2005 (1-0 in 37).;
C2) 14.h3 h5 (retreating to e6 or d7
would have avoided the exchange, with
a balanced game) 15.h4!? g6
16.xg6 hxg6 17.f3 ad8 18.e2
was J.Horvath-J.Borisek, Meissen 2013,
and now simply fe8 would have kept
level chances.;
C3) 14.--;
D) 13.-- ]
12...bxc6 13.c1 a7
[ 13...d6 has been more popular in
practice: 14.e2 ( moving the knight away
from the centre with 14.a4 inevitably gives
Black some attacking chances; c8 15.d3
was J.Sunye Neto-L.Christiansen, World
Student Team Championship, Mexico 1980,
when simply completing development with
e8 would have left Black with a fine game )
A) 14...xf3 (a logical attempt to punish
Ne2) 15.gxf3 c8 16.h1 ( 16.d3
d7 17.g3 has high-level endorsement,
albeit in a simul: G.Kasparov-Jo.Wilson,
London simultaneous 2003; 1-0 in 56 )
16...e8 17.g1 g6! (very logical, dealing
with the pressure on the g-file, taking f5
under control, and preparing ...Bd6-f8-g7)
18.d2 f8 19.b4 a5! 20.a3
(moving the b-pawn forward – to b5 or a5
– allows 20...c5 with active counterplay)
axb4 21.axb4 g7 and Black was fine in
M.Ashley-N.De Firmian, New York 1996
(0-1 in 75).;
B) 14...c8 15.g3 e8 16.d3 e4
( 16...a5 is sensible too, as in I.V.Ivanov-M.
Ashley, New York 1994; ½-½ in 21 )
B1) here 17.c2 xg3 18.hxg3 e6
19.xa6 a8 20.d3 xa2? ( but
20...c5 brings about a position where
White's extra pawn isn't really felt – see
the rest of this note for more examples )
21.f6! was A.Moiseenko-P.Leito, Bled
Olympiad 2002, 1-0 in 45;
B2) 17.d2 xg3 18.hxg3 h5 19.f3
( 19.xa6 d7 gives full positional
compensation for the pawn; White's only
active plan is to push the a-pawn, which
will leave him with queenside
weaknesses – even Houdini wants to
repeat with 20.a4 f6 21.d3 f5
22.e2 g4 23.d3 ) 19...xf3
20.gxf3 g5 21.f4 g4 22.g2 h4
23.h1 was F.Peralta-R.Swinkels, Wijk
aan Zee 2013 (1-0 in 32), but Black
 could have played 19...Qd7! with that Magnus goes on to win does not suggest
equality since, as noted before, taking that he got anything in the opening. In fact,
on a6 leaves Black with full positional this whole game strongly reminds me of
compensation.; Carlsen's crucial win against Radjabov in the
C) 14...-- and: ] 2013 London Candidates Tournament.
14.e2 Keeping the knight within touching 28.f1 a4 29.bxa4 d4 30.a5 a4 31.d2
distance of the kingside and d4 seems f8 32.d3 f6 33.b2 xa5 34.c4 a4
sensible. 35.c2 e7 36.e2 e6 37.d3 d5
[ 14.a4 is more ambitious, but I would be 38.a3 h5 39.h3 h4 40.c1 g6 41.c2 g5
happy to see such a move with Black: d6 42.c1 a6 43.e1 b8 44.e7 f4
15.c4?! e6 ( 15...xf3 16.gxf3 ad8 45.c3 f5 46.b3 g4 47.a4 gxh3 48.gxh3
17.c1 c5 was good too ) 16.c5 c8 g6 49.a5 g1 50.a6 b1+ 51.c3 c1+
17.c1 e4 18.b4 ( 18.d3 is equal ) 52.d3 d1+ 53.e2 a1 54.b6+ d6
18...xc5! 19.bxc5 f6! changed the pawn 55.g7 c6 56.g6+ b5 57.d5 e5
structure in Black's favour in U.Nielsen-G. 58.b6+ c4 59.e3+ c3 60.f4 d4
Sax, Vejle 1994 (0-1 in 46). The c6-pawn is 61.xf5 c4 62.c6 h1 63.d6 h2+
protected by White's c5-pawn, and the d5- 64.f3 d3 65.xc4 xh3+ 66.g4 h1
pawn is a rock. Black's last move takes 67.a4 f2 68.a3+
control of the e5-square and blunts the b2- 1-0
bishop on the long diagonal. After a few
more moves Black's position got even
better: 20.d4 xf3 21.gxf3 g5 22.g4
b8 23.c3 b5! and White struggled to
cover all his weaknesses. ]
14...d6 15.e5 e7 16.ed4 xf3?!
After this slightly passive move White has
chances of an edge.
[ Instead, 16...c5! seems to work well:
17.c6 e8 18.xf6 xc6 19.e7!? fc8
and the hanging pawns are alive. White
seems to have no route to an advantage;
for instance, 20.e5 ( 20.h3 f5 21.d4
d7 22.xf5 xf5 looks comfortable for
Black ) 20...xd1 21.xc6 xb3 22.axb3
( 22.xa7 xa7 23.xc5 xc5 24.xc5
c7 25.d6 c6 26.axb3 xd6
is another drawn rook and pawn endgame )
22...xc6 23.fd1 c7 24.d6 b7
25.xc5 xc5 26.xc5 xb3 27.cxd5 g6
with a draw. ]
17.xf3 fc8 18.d3 a5
[ 18...e4 was another option, which White
could have avoided by taking on f6 a move
earlier. ]
19.xf6 xf6 20.c2 d8 21.fc1 c5
22.e4
[ 22.b5 was tempting. Then h6 ( 22...b6
23.e5! is good for White ) 23.xa5 ( after
23.h3 d4 24.exd4 cxd4 Black has decent
counterplay ) 23...b6 24.b5 a5 25.e2
a3 is probably better for White, but it won't
be easy for him to mobilize his a-pawn. ]
22...g6 23.e1 dxe4 24.xe4 xe4
25.xe4 d1+ 26.e1 xe1+ 27.xe1 d8
This endgame is equal. As so often, the fact
 Back to Contents Page
Chapter Seven

Tarrasch Sidelines
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nf3 c5

As I hope has become clear by now, the Tarrasch is a principled opening and White should not be able to hope for
too much by avoiding the main lines. That said, main lines change over time and this chapter contains one very
serious attempt.
This is probably a good time to mention that, when confronted by a sideline (in the Tarrasch or elsewhere), please
do NOT rush your response. Sidelines by their nature have slightly offbeat ideas which, if well neutralized, can
leave Black with a pleasant game. However, often on seeing an early Bg5 or e2-e3 in the Tarrasch (I speak from
bitter experience here, in particular from my game against Bischoff) the natural tendency is to relax and play a quick
move. Black still needs to be precise in the sidelines as in every other Tarrasch variation - his structure is
fundamentally more vulnerable than White's and we don't have the luxury of second-rate responses. Apologies if
this seems blatantly obvious, but I have personally scored worse against sidelines than against main lines in the
Tarrasch, and want to save you from a similar experience.
Coincidentally, all the games in this section are played by Jacob Aagaard. Jacob has a great feel for the Tarrasch
and, in particular, is comfortable developing his pieces to good squares rather than searching for ways to "justify"
his compromised structure. The sidelines tried here have not troubled him.
4 cxd5
4 g3 Nf6 5 Bg2 Nc6 6 0-0 Be7 7 dxc5 Bxc5 leaves Black with a big pawn in the middle of the board, and no
weaknesses. Black should not be worse here and Schandorff-Aagaard shows why.
4...exd5 5 Nc3 Nc6

6 dxc5!?
This move has become one of White's most serious and popular attempts, following Aagaard and Ntirlis' analysis
which showed that Black needs to be precise to equalize. See L'Ami-Aagaard for a good example.
The less threatening 6 Bg5, together with 6 Bf4 and 6 e3, is covered in Rewitz-Aagaard.

Key Notes

1. Against all Tarrasch sidelines, take your time and handle the first few moves carefully. Some of these variations
have dangerous ideas and Black needs to respond accurately.
2. 6 dxc5 is probably the most dangerous sideline, going straight for positions with an open centre where White
relies on his bishop pair. Nevertheless, Black has excellent development and can expect to equalize with accurate
relies on his bishop pair. Nevertheless, Black has excellent development and can expect to equalize with accurate
play.
3. 6 Bg5 Be7 7 Bxe7 Ngxe7 is a quiet sideline which shouldn't trouble Black. After 6 e3, 6...c4! gives Black a great
version of the 9 Bg5 c4 variations.
4. Delaying the capture on d5, as in Schandorff-Aagaard, would inconvenience Black if he had to take on c4 or
fianchetto his queen's bishop, but he can solve his problems in a more direct way by simple development. If White
insists on not taking on d5, Black can seize central space with ...e6-e5; then if White takes on d5, the resulting IQP
position is fine for Black.

1 d4 d5 2 c4 e6 3 Nf3 c5 4 cxd5
4 g3 Nf6 5 Bg2 Nc6 6 0-0 Be7 7 dxc5 Bxc5 – Schandorff-Aagaard
4...exd5 5 Nc3 Nc6 6 dxc5!? – L'Ami-Aagaard
6 Bg5 – Rewitz-Aagaard

Back to Contents Page

Tarrasch Sidelines
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 As I hope has
become clear by now, the Tarrasch is a
principled opening and White should not be
able to hope for too much by avoiding the
main lines. That said, main lines change over
time and this chapter contains one very
serious attempt.
This is probably a good time to mention that,
when confronted by a sideline (in the Tarrasch
or elsewhere), please do NOT rush your
response. Sidelines by their nature have
slightly offbeat ideas which, if well neutralized,
can leave Black with a pleasant game.
However, often on seeing an early Bg5 or e2-
e3 in the Tarrasch (I speak from bitter
experience here, in particular from my game
against Bischoff) the natural tendency is to
relax and play a quick move. Black still needs
to be precise in the sidelines as in every other
Tarrasch variation – his structure is
fundamentally more vulnerable than White's
and we don't have the luxury of second-rate
responses. Apologies if this seems blatantly
obvious, but I have personally scored worse
against sidelines than against main lines in
the Tarrasch, and want to save you from a
similar experience.
Coincidentally, all the games in this section
are played by Jacob Aagaard. Jacob has a
great feel for the Tarrasch and, in particular,
is comfortable developing his pieces to good
squares rather than searching for ways to
"justify" his compromised structure. The
sidelines tried here have not troubled him.
4.cxd5
[ 4.g3 f6 5.g2 c6 6.0-0 e7 7.dxc5
xc5 leaves Black with a big pawn in the
middle of the board, and no weaknesses.
Black should not be worse here and
Schandorff-Aagaard (Game 22) shows
why. ]
4...exd5 5.c3 c6 6.dxc5!? This move
has become one of White's most serious and
popular attempts, following Aagaard and
Ntirlis' analysis which showed that Black
needs to be precise to equalize. See L'Ami-
Aagaard (Game 20) for a good example.
[ The less threatening 6.g5 ]
[ together with 6.f4 ]
[ and 6.e3 , is covered in Rewitz-Aagaard
(Game 21). ]
D32
L'Ami,E
Aagaard,J
20: Helsingor 2012
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.dxc5!? This was believed to be
a harmless sideline until the monumental
analytical work done by Aagaard and Ntirlis
demonstrated that Black's traditional
responses led to difficult positions. In fairness,
the move had already been discovered by
Krasenkow, who tends to score 70-80%
against a lot of Black's main defences to 1 d4
and so is a guy worth following.
d4 Seizing space in the centre is the most
principled response.
[ 6...f6 leaves White with some advantage
after
A) whereas the most popular 7.g5
should be met with the excellent move
xc5! (not yet played by anyone over 2000,
as far as I can see) with very comfortable
play for Black, since 8.xf6 xf6 9.xd5
( the complications after 9.xd5 xb2
are in Black's favour ) 9...b4 10.c1 e6
11.e4 0-0 leaves White struggling to
equalize. If he tries to hang on to his extra
material with 12.a3 , then xc3+ 13.xc3
ad8 and White will need to show
something special to get out of the
opening in one piece.;
B) 7.e3! , ]
7.a4
[ 7.e4 was tried several times by Ivan
Ivanisevic, but he seems to have shelved it
after the following defeat: f5 ( 7...d5
8.d6+ xd6 9.cxd6 xd6 was also
playable in I.Ivanisevic-G.M.Todorovic,
Serbian Team Championship 2005 ) 8.g3
g6 ( an earlier game saw 8...g4 9.h3
xf3 10.exf3 xc5 11.a3 f6 12.d3 0-0
13.0-0 , I.Ivanisevic-G.M.Todorovic, Serbian
Team Championship 2004, and I prefer
White ) 9.h4 h6 10.h5 h7 11.e4 xc5
12.d3 was I.Ivanisevic-R.Zelcic, Cannes
2006, and now apart from d6 (0-1 in 81),
( simply 12...ge7 13.0-0 0-0 looks fine
for Black. )]
7...xc5 Definitely the safest move.
[ The alternatives are considerably more
complicated but should lead to some
advantage for White, it seems: 7...g4 8.a3
e7 9.h3 h5 10.b4 and White was en
 route to a rapid victory in A.Dreev-V.Yemelin,
Tomsk 2006 (1-0 in 17). It's hard to believe
that Black has enough for the pawn here. ]
[ 7...b5 8.cxb6 axb6 9.e3 b4+ 10.d2
is clearly better for White, as most recently
demonstrated in F.Elsness-E.Lie,
Norwegian Championship, Lillehammer
2013 (1-0 in 47). ]
[ I don't trust 7...f6 8.e3 g4 9.e2
either, since Black will struggle to regain
his pawn. ]
8.xc5 a5+ Now White has a major
decision to make. 9.d2
[ 9.d2 xc5 is another branch:
A) 10.e3 leaves Black with a pleasant
choice:
A1) 10...f6!? is an interesting gambit,
if not 100% sound: 11.xd4 ( or
11.exd4 d5 12.e2 0-0 13.0-0 e8
with some compensation – but probably
not enough – in M.Misojcic-N.Somborski,
Vrnjacka Banja 2008 ) 11...0-0 12.e2
( here 12.c3! looks like a better shot )
12...e4 13.c2 a5+ 14.d2 xd2
15.xd2 xd2+ 16.xd2 d8 17.f3
xd4 18.exd4 xd4+ and a draw was
soon agreed in G.Serper-V.Akobian,
Reno 2004 (½-½ in 26).;
A2) 10...dxe3 11.xe3+ xe3+
12.xe3 e6 ( 12...f6 13.c4 e6
14.xe6 was agreed drawn in R.Ris-A.
Demuth, Wijk aan Zee 2012, but this
looks like an edge for White ) 13.b5
( or 13.d4 xd4 14.xd4 f6
15.b5+ f7 16.c1 e7 17.a3 ac8
18.0-0 a6 19.e2 xc1 20.xc1 d8
with approximate equality in M.Narciso
Dublan-J.Gros Fernandez, Catalonian
Team Championship 1995; ½-½ in 61 )
13...ge7 14.d4 d5 15.0-0 0-0
16.fd1 fd8 17.h3 h6 18.b3 a6
19.f1 ac8 and Black's superior piece
placement balanced the bishop pair in V.
Ivanchuk-A.Grischuk, Monte Carlo
(blindfold rapid) 2006 (½-½ in 59).;
A3) 10...--;
B) 10.a3 is more ambitious, but Black
gets good compensation by sacrificing the
d4-pawn: ge7 11.b4 b6 12.b2 0-0
. This novelty was suggested by Aagaard
and Ntirlis, and played in a game which
followed their analysis for a long time:
13.xd4 xd4 14.xd4 c6 15.b2
( Aagaard and Ntrilis give 15.c5 xc5
16.bxc5 a5 with good compensation for
Black ) 15...a5 16.b5 a4 17.e3 a5
18.e2 b3 19.d1 a5+ 20.f1 e6
21.h4 (White is the first to play a move not
analysed by Aagaard and Ntirlis) fd8
22.d4 xd4 23.exd4 ac8 and Black
had more than enough compensation in F.
Urkedal-I.Jelen, Pula 2013.;
C) 10.-- ]
9...xc5 10.c1
[ Instead: 10.e3 should be met by dxe3
11.xe3 b4+ 12.d2 ( or 12.d2 xb2 )
12...xd2+ with a balanced endgame in E.
Grünfeld-S.Tarrasch, Teplitz-Schönau 1922
(½-½ in 26). ]
[ 10.b4!? xb4 11.c1 d6 12.e3
was an interesting sacrifice in F.Graf-J.
Krassowizkij, Hanover 2012, when e7
13.xd4 0-0 seems to be safest response,
with good equalizing chances. ]
10...b6
[ 10...d6 was another successful Tarrasch
outing for the German IM in J.Timman-J.
Carlstedt, Helsingor 2012 (½-½ in 25), but
he would have been suffering had Timman
found 11.c3! . ]
11.e3 f6
[ 11...dxe3?! is now too compliant and has
scored badly in practice; for instance,
12.xe3 xb2 ( not 12...b4+ 13.d2!
xb2? 14.c3 and wins ) 13.c4 b4+
14.d2 with a strong initiative for the pawn
in E.L'Ami-T.Willemze, Haarlem 2012 (1-0 in
24). It tends to be advisable to develop your
own pieces rather than the opponent's. ]
12.c4
[ After 12.a4 , another recent game
followed Aagaard and Ntirlis' analysis for
quite some time: 0-0 13.xd4 xb2
14.xc6 bxc6 15.c2 b8 16.c4 xc2
17.xc2 b1+ ( Aagaard and Ntirlis
recommend 17...f5 ) 18.c1 xc1+
19.xc1 e6 20.xe6 fxe6 21.f3 d5
22.e4 ( Aagaard and Ntirlis had analysed
22.d2 b8 23.e4 b4 24.a3 a2
25.c2 xc1 26.xc1 f7 with a very
slight edge to White in the rook ending )
22...c3 23.e3 d8 24.0-0 a6 25.h4
f7 26.h2 d3 27.f4 d1 28.e1 c5
29.e2 c4 , when White had had enough
excitement and decided to force a draw with
30.c2 c3 31.d2 d3 32.c2 c3
33.d2 d3 in A.Indjic-S,Brenjo, Banja
Koviljaca 2013. ]
12...dxe3 Yet another novelty from the
Aagaard/Ntirlis laboratory.
 [ After 12...0-0 13.0-0 g4 14.exd4 ad8
15.c3 , Black had problems restoring
material parity in C.Sandipan-M.Vucicevic,
Lugano 2013 (1-0 in 31). ]
13.xe3 b4+ 14.d2
[ 14.f1 e6 15.xe6 fxe6 16.e2 a5
17.g5 xa2 18.b5?! ( 18.b4 xe2+
19.xe2 e5 , intending 20...Nd4+, is equal )
18...0-0 19.xb7 ac8 20.g1
was M.Krasenkow-J.Koch, French Team
Championship 2012, and now g4
would have given Black a serious initiative. ]
14...e7+ 15.e2 0-0 16.0-0 d8
[ Departing from his own recommendation of
16...e6 . Perhaps he should have stuck to
this, since in the game he gets into trouble. ]
17.e1 e4 18.b5 g4 19.xc6 bxc6
20.e2? Losing his way.
[ 20.c2! would have left Black on the
ropes. ]
20...e8 21.e3 d7 22.c3 xc3
. White no longer has any advantage and a
draw is a fair result.
[ For instance, 22...xc3 23.xc3 xf3
24.xf3 d2! levels the chances. ]
½-½
D32
Rewitz,P
Aagaard,J
21: Danish Team Championship 2011
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g5 This move does have
some point – after all, trading dark-squared
bishops should be in White's favour in most
IQP positions. Having lost a bad game in this
variation (see the note with 8...0-0 below), I
know how important it is to retain
concentration when an 'innocuous sideline' is
used; Black's development is accelerated but
he needs to avoid ending up with no
counterplay in an inferior structure.
[ 6.f4 f6 7.e3 is also well met by c4!
, as played by Tarrasch himself. White
doesn't have much here; for instance, 8.e2
b4 9.0-0 the most popular move but Black
has no problems after it; ( instead, 9.e5
e4 10.xc6 bxc6 11.a4 b6 12.c1
was Vl.Georgiev-T.L.Petrosian, Chicago
2009, ½-½ in 43; while 9.d2 a5 10.c2
0-0 11.0-0 g6 12.g5 f5 13.c1 e7
gave Black a comfortable game in P.
Harikrishna-I.Sokolov, Sarajevo 2009 )
9...0-0 10.c2 ( 10.e5 xc3 11.bxc3
e4 was equal in V.Tkachiev-V.Yemelin,
Russian Team Championship 2006 )
10...g4 11.fc1 e8 12.h3 h5 13.b5
c8 14.a4 a6 15.a7 xa7 16.xb4
b5 17.g4 g6 18.e5 a5 19.e1 c6
and Black had promising queenside play in
M.Meyer-A.Berelowitsch, German League
2011 (0-1 in 48). ]
[ 6.e3 gives rise to Symmetrical Tarrasch
positions where White has shown his hand
by capturing on d5 early. For instance, after
f6 7.b5 d6 we have transposed into
Game 19 (Carlsen-Radjabov) in the previous
chapter. ]
6...e7 7.xe7 gxe7 8.e3 By taking
control of the d4-square, White prepares to
capture on c5.
[ The immediate 8.dxc5 is met by d4 9.e4
( 9.b1 0-0 10.g3 f5 11.g2 a5+
12.d2 xc5 13.0-0 e6 14.a3 fd8
15.fd1 ac8 was comfortable for Black in
S.Cicak-S.Galdunts, German League 2003 )
9...0-0 and White needs to be a little careful:
A) 10.g3 d5 11.ed2 xc5 12.g2
e6 13.0-0 ( if 13.b3 b6 14.0-0 fd8
15.e1 b4 16.d3 , as in S.Collas-G.M.
Todorovic, Skopje 2011, I like ac8
which covers c5 and gives Black good
play ) 13...b5 ( on a principled basis I
prefer 13...fd8 followed by ...Rac8 –
Black hasn't completed development yet )
14.b3 xb3 ( again 14...fd8
would be my preference ) 15.xb3 xe2
16.fe1 a6 17.f1 a5 18.xb7 ab8
19.a6 was T.Polak-F.Ljubicic, Split 2008,
and now xa6 20.xa6 xb2 21.xd4
g6 would have led to a draw.;
B) 10.ed2 e8 11.g3 is asking for
trouble.
EXERCISE: How can Black take
advantage of White's lack of
development?
ANSWER: d3! (of course!) 12.a3
( not an attractive move, but White was
already in a bad way: 12.e4 b4
is much better for Black; while after 12.c1
d5 White will do well to get out of the
opening in one piece ) 12...dxe2 13.xe2
h3 14.g5 ( 14.c4 is a better try, but
Black's initiative continues in the endgame
after f5! 15.xd8 axd8 ) 14...g2-+
15.g1 f5 16.xg2 xg5 (a knight is
coming to d4 and White is losing decisive
 amounts of material) 17.b4 cd4 18.f4
f6 19.c4 ad8! (nice attacking
technique, bringing in the final piece)
20.b1 c2+ (other wins were easier but
it doesn't matter at this point) 21.xc2
xa1+ 22.f2 d4 and White resigned
in V.Bernadskiy-K.Malinovsky, Krakow
2013.;
C) 10.-- ]
QUESTION: How would you respond with
Black?
ANSWER: To a certain extent, what you play
here should be informed by your approach
after 9 Bg5. If you delight in the IQP positions
after 9...cxd4, you can get similar structures
here by taking on d4 or castling. If you go for
9...c4, you should push your pawn here too.
8...c4 In general, I think this is the best move.
The 9 Bg5 c4 lines require White to play very
purposefully in order to break the black d5/c4
pawn chain and, if White doesn't achieve this
goal, he can easily end up worse. The current
position seems to me like a very favourable
version of this structure for Black – there is no
annoying bishop on g5 threatening to disrupt
the kingside structure by taking on f6, while
White's light-squared bishop is not actively
supporting the e4-advance from g2, but rather
will sit passively on e2.
[ Instead: I played 8...cxd4 9.xd4 0-0
a couple of years ago.
The game is a decent example of how Black
can gradually overtake the initiative if White
is careless: 10.e2 e6 11.0-0 b6
12.a4 a5 13.a3 ( 13.xe6 fxe6
14.c1 was better, when I prefer White )
13...xd4 14.exd4 ( 14.xd4 c6 15.d1
d4! is fine for Black ) 14...ad8 15.b4
(bringing a rook to an open file seems more
purposeful) c7 16.d2 f5 17.c5 c8
18.ac1 e7 19.g4 d6 20.fe1 h4
21.f3 b5 22.b3 xa3 23.d3 c4
24.xd5 xd5 25.xc4 e6 (an unusual
structure has been reached; Black certainly
doesn't have any problems here, but nor
should White) 26.c5 ( 26.c7
is more active ) 26...g6 27.xd5 xd5
(now the black bishop is dominant) 28.c5
d8 29.e5? g7? (letting White back in
the game) ( 29...c6 was stronger, since
30.d5? xb4 doesn't work for White ) 30.f3?
(in trying to restrict the bishop on d5, White
severely weakens his king's position and
creates some tactical weaknesses) ( 30.e3
would have held the balance ) 30...b6 31.g3
(the white knight doesn't have a good retreat,
since going to e4 allows the black queen to
capture on b4 after a check on e1) h6
32.e4 ( 32.b3 was a better try, but Black
is now well on top after xb3 33.xb3
c1+ 34.f2 d2+ 35.f1 c8! )
32...c1+ 33.g2 (losing immediately,
though alternatives failed to save White
either) c4! – an elegant finish: mate on f1
is looming if the queen moves, so White
resigned in L.Klement-S.Collins, Teplice
2011. ]
[ 8...0-0 9.dxc5 a5 ( not 9...g4? 10.e2
a5? 11.a4 xa4 12.xa4 and I had
no compensation for the pawn in K.Bischoff-
S.Collins, Dun Laoghaire 2010; 1-0 in 29 –
this is a classic example of Black, thinking
he needs to "justify" his compromised
pawn structure, trying to be too clever )
10.e2 xc5 11.0-0 and Black has a
comfortable game.
Here's a great example of how a Tarrasch
expert can play such positions for a win: e6
( 11...d8 12.c1 f5 was also reasonable
in H.Teske-L.B.Hansen, German League
2000; ½-½ in 25 ) 12.b5 f5 13.fd4
fxd4 14.xd4 ac8 15.c1 b4
16.xe6 ( 16.xc6 xc6 17.xc6 bxc6
is level as well ) 16...fxe6 17.g4
(White has created some imbalance by
capturing on e6, but – as always – this
gives Black chances too, based on his
strong centre and potential play on the f-file)
f6 18.c2 cf8 19.f4?! (White loses
patience and creates permanent
weaknesses; simply holding the position was
more appropriate) b6 ( the immediate
19...d4 was fine but, interestingly, Akobian
prefers to prepare this move ) 20.e2 h8
21.ff2 d4 (while this position is level
according to the computer, Black's position
is much easier to play) 22.b3? was played
in Y.Lapshun-V.Akobian, Philadelphia 2002,
( the counter-intuitive 22.exd4
was apparently the best move ) , and now
22...xb3 23.axb3 dxe3 24.xe3 e5
would have been good for Black. ]
9.e2 0-0
[ 9...b8 10.0-0 b5 has also been tested in
a number of games, with good results for
Black. My personal preference is to
complete development and offer a trade of b-
pawns, as Jacob does in this game. ]
10.0-0
[ 10.e5 b5 11.xc6 xc6 12.f3 e6
 13.a3 a5 14.0-0 ( here 14.xb5 b8
15.a4 b4 16.0-0 gives Black full
compensation, or simply a draw after f5
17.c1 a2 18.a1 b4 ) 14...b8
15.e1 d6 16.d2 f5 and Black had
more space at little cost in J.Svatos-S.
Svoboda, Czech League 2013 (½-½ in 44). ]
EXERCISE: How can Black get going on the
queenside?
ANSWER: 10...b5!
[ Black can also prepare this advance with
10...b8 but, although this has been played
by Tarrasch experts like Akobian, I don't
think it's necessary, since the trade of b-
pawns favours Black provided he can stop
White playing e3-e4. ]
11.xb5
[ The more sedate 11.b3 , as played in M.
Bartel-M.Petr, Czech League 2012, can be
well met by a5 12.c1 d8 . Again Black
seems to have a great version of the 9 Bg5
c4 lines. ]
11...b8 12.c3 xb2 QUESTION: Who's
better here?
ANSWER: I already prefer Black in this
position, since he has a protected passed
pawn and it's entirely unclear what trumps
White has. In such a structure there's no
pressure on d5, so the chances for White to
break the d5/c4 pawn chain (his favourite plan
in the 9 Bg5 c4 line) are close to zero.
Note the bishops. It's a trite point, but while
Black technically has a "bad bishop" (since
his central pawns are on the same colour)
and White has a "good bishop", just looking at
the position reveals that the e2-bishop is
passive and completely restricted by the black
pawn chain, while its counterpart has bright
prospects on f5, or even just e6, holding the
key d5-pawn. If White were able to simplify
and get his bishop to f3 the assessment might
change, but for the foreseeable future the
black bishop is going to have more fun. This
concept of a formally "good" bishop actually
being no better than its counterpart, since it is
restricted by the opponent's pawns, is known
from several openings, in particular the
Stonewall Dutch.
13.a4
[ Two alternatives have been tried against
the German IM Jonathan Carlstedt, who is
a faithful Tarrasch practitioner and has very
good results with the opening. Indeed, one
of his biggest upsets (a miniature win with
Black against Milov) occurred in this
position: 13.c1 b4 14.d1 f5 15.h4
e6 16.g3 a5 17.g2 fb8 ( since e3-
e4 is only White's idea, it is worth using
prophylactic measures against it: 17...d8
18.f4 f5 keeps White bottled up;
and even 17...f5!? looks like a promising
idea ) 18.f4 f5 19.d2 4b7 20.f3
b4 21.e4 dxe4 22.xe4 bd5 23.xd5
xd5 (through inspired play Milov has
managed to achieve the e3-e4 break, only to
make an amazing mistake...) 24.c5?? b1
. Whoops. White resigned in V.Milov-J.
Carlstedt, Odense 2011. ]
[ 13.e5 a5 ( 13...xe5 14.dxe5 f5
is good too ) 14.xc6 xc6 15.c1 b6
16.d1 f5 17.f3 d8 18.d2 d3
19.h3 f5! and Black was better against a
strong GM in R.Buhmann-J.Carlstedt,
Deizisau 2010 (½-½ in 25). ]
13...b4 14.c5 h6
[ 14...d6 is also strong, since 15.e4
can be met by g6 with some advantage
for Black. ]
15.h3
[ White would like to play 15.e4 , but he is
insufficiently prepared for the opening of the
game after dxe4 ( or 15...g4 ) 16.xe4
g4 . ]
15...f5 16.c1 a5 QUESTION: How
would you assess this position?
ANSWER: White is busted. He has one good
piece (the knight on c5) which is not co-
ordinating with any of his other pieces. In the
meantime, Black is preparing to penetrate
down the b-file. White decides on a rather
desperate kingside break, though it is already
hard to offer him good advice.
17.g4 h7 18.g5 hxg5 19.xg5 f5
20.g4 g6
[ 20...d8 was more accurate, taking aim at
the knight on g5. ]
21.d1 d8 22.f3 d6 23.h4 g7
24.g2 Jacob can now improve the position
as he chooses but goes instead for a trap.
c3!?
[ 24...h8 is good, just attacking on the
kingside. ]
25.ac1? Falling for the trap, but White's
position was already terrible.
EXERCISE: Black to play and win.
ANSWER: xg4 26.xg4 xc5!
Oh dear. I tend to resign when I'm a piece
down for nothing against a GM but I suppose
not everyone feels the same way. In fairness
Jacob soon returns the piece (probably
informed by his own sense of humour rather
than the objective merits of the move) and
takes all White's pawns instead.
27.a3 c4 28.f4 xd4 29.exd4 xd4
30.xd4+ xd4 31.xc3 xh4 32.f4 f5
33.c7 e3+ 34.g3 g4+ 35.f3 xg5
36.xe3 g3+ 37.d4 xa3 38.f5 a4+
39.xd5 a5+ 40.e4 e8+ 41.d4 d8+
42.e4 e8+ 43.d4 xf5 44.xf5 gxf5
45.xa7 g6 46.d3 f4 47.a4 g5
48.a7 f6
0-1
E06
Schandorff,L
Aagaard,J
22: Danish Championship, Odense 2011
[Sam Collins]
1.d4 d5 2.c4 e6 3.f3 c5 4.g3 f6 5.g2
c6 6.0-0 e7 7.dxc5 xc5 8.a3
This is another approach (which, again, can
arise from multiple move orders). White aims
for b2-b4 without taking on d5, thus keeping
the bishop on c8 bottled up, but this quiet
approach allows Black to complete
development comfortably. If White takes on d5,
Black will have the option of favourable
Tarrasch positions where White has
weakened key squares on the c-file. If White
does not take on d5, practice has shown that
Black has sufficient defensive resources, as
illustrated by the present game.
0-0 9.b4 b6 It is more common to retreat
the bishop to e7,
[ but 9...b6 is more active and the move I
prefer. ]
10.b2 e7 11.bd2 d8 12.cxd5
This is White's last chance to create an IQP.
[ Instead, 12.c2 allows Black to seize the
centre with e5 . White is not worse here but
his play is a little too subtle for me.
A) 13.cxd5 xd5 and now:
A1) 14.c4 e4 15.xb6 xb6
16.d2 was R.Buhmann-R.Strohhäker,
Böblingen 2006, when f5 would have
been quite playable for Black.;
A2) 14.e3 a6 15.e4 f5 16.h4
e6 17.ac1 ac8 was very
comfortable for Black in E.Bricard-S.
Galdunts, French Team Championship
2006 (0-1 in 52).;
A3) 14.b5 d4 15.xd4 exd4 16.b3
g4 17.xd4 e3! 18.fxe3 xe3+
19.h1 xd4 20.xd4 xd4 21.xb7
e8 with at least equal chances for
Black in J.Magem Badals-M.Illescas
Cordoba, Leon 1990 (½-½ in 29).;
A4) 14.fd1
A4a) the immediate 14...e6
is also possible, since 15.xe5??
( and 15.xe5? d4 16.d3 f6
is good for Black: ...Nb3 is a threat )
15...xe5 16.xe5 d4 wins;
A4b) 14...f6 15.c4 e6 16.e4 c7
17.xb6 axb6 18.f1
was O.Romanishin-S.Silva, Figueira
da Foz 2008, and now e8!
was good, aiming to bring the knight to
d6 and (after ...b6-b5) perhaps to c4.;
A5) 14.ad1 f6 15.c4 e6 16.e4
c7 17.xb6 axb6 18.e2 b5
19.xd8+ xd8 20.d1 c4
was level in O.Romanishin-M.Meinhardt,
Hockenheim 2006 (½-½ in 41).;
A6) 14.--;
B) 13.b5 d4 14.xd4 exd4 15.c5
xc5 16.b3 b6 17.xd4 g4 18.fe1
ac8 19.b2 was A.Khalifman-A.
Grischuk, German League 2003. It is
remarkable to see how Grischuk, from an
equal position against a world-class player
and former FIDE world champion,
manages to whip up a dangerous attack
and win the game in 10 moves.
EXERCISE: How should Black start to
build an initiative?
ANSWER: h5! – always a standard move
against the kingside fianchetto. Grischuk
wants to put his pawn on h4, from where it
can either weaken the white structure (...
h4xg3) or threaten the white king's
breathing space (...h4-h3). Here Ribli
recommends
B1) but Khalifman makes a couple of
careless moves: 20.a4 h4 21.a5 xd4
22.xd4 e5 23.ab1?! ( 23.ac1
maintains the balance ) 23...e4 24.f3
xb2! 25.xb2 h3! 26.f1 d4
27.d1?? (an outright blunder).
EXERCISE: Black to play and win.
( 27.e5 e6 28.d3 is still only
slightly better for Black ) .
ANSWER: 27...d3! (a standard
Tarrasch breakthrough: White sheds
material) 28.b4 ( or 28.xd3 xd3
29.exd3 xf3 30.dxe4 c1 , followed
by ...Bg2 ) 28...xf3 29.exf3 c3!
and White resigned, in view of 30.xd3
 e2+ 31.xe2 c1+ 32.f1 xd3
(Ribli).;
B2) 20.h3 as equal,;
C) 13.-- ]
12...exd5 Black has no problems. His position
is active, all the pieces are still on the board,
he has several targets (the e2-pawn, the c4-
outpost, and possibilities of undermining the
white queenside with ...a7-a5), and there is
zero pressure on the d5-pawn.
13.b3 e4 14.b5 a5 15.xa5 xa5
16.d4 f6 17.a4 b6 18.d4 d7
19.fc1 e8 20.b4 ac8
[ 20...xb5 21.xb6 axb6 22.d4
regains the pawn, though the position
remains equal. ]
21.a4 xc1+ 22.xc1 c8 23.a3 f7
24.h3 . The players agreed a draw in this
balanced position. I really like Jacob's play in
this game – simple development, putting his
pieces on good squares. People often get too
clever in the Tarrasch and end up putting
their pieces on the wrong squares.
½-½
 Back to Contents Page
Chapter Eight

Réti Set-ups
(View in Game Format)

1 Nf3 d5 2 c4 e6 3 g3 c5

One of the attractions of the Tarrasch is that it can be played against everything except 1 e4. Come to think of it,
Tarrasch's "refutation" (3...c5) of his own variation (1 e4 e6 2 d4 d5 3 Nd2) does bear a strong stylistic resemblance
to some lines of our opening, but I'll leave coverage of this line to more experienced French players.
White can play a broad range of flank openings and we can't possibly cover them all here. Fortunately we don't
need to, since many of these openings have a remarkable tendency to reach the same position:
4 Bg2 Nf6 5 0-0 Nc6 6 b3 Be7 7 e3 0-0 8 Bb2 b6 9 Nc3 Bb7

This position is covered in Reinderman-Irwanto and knowledge of this game should be sufficient against most
English/Réti set-ups. If White goes more offbeat, your first few moves involved gaining central control and
developing pieces, so I suggest you keep doing that.

Key Notes

1. One of the great advantages of the Tarrasch is that it can be played against everything except 1 e4. I suggest
setting up the d5/c5/e6 formation as soon as possible, when the best White has is a transposition into regular
Tarrasch lines.
2. In the main Réti line (as in Reinderman-Irwanto), Black has a number of options which might appeal to Tarrasch
players. I have recommended two lines: 10 cxd5 Nxd5 11 Nxd5 exd5 12 d4 a5!? and 12...Ba6 13 Re1 c4!?. In both,
I rather like Black's active and purposeful queenside play.

1 Nf3 d5 2 c4 e6 3 g3 c5 4 Bg2 Nf6 5 0-0 Nc6 6 b3 Be7 7 e3 0-0 8 Bb2 b6 9 Nc3 Bb7 – Reinderman-Irwanto

Back to Contents Page

Réti Set-ups
[Sam Collins]
1.f3 d5 2.c4 e6 3.g3 c5 One of the
attractions of the Tarrasch is that it can be
played against everything except 1 e4. Come
to think of it, Tarrasch's "refutation" (3...c5)
of his own variation (1 e4 e6 2 d4 d5 3 Nd2)
does bear a strong stylistic resemblance to
some lines of our opening, but I'll leave
coverage of this line to more experienced
French players.
White can play a broad range of flank
openings and we can't possibly cover them all
here. Fortunately we don't need to, since
many of these openings have a remarkable
tendency to reach the same position:
4.g2 f6 5.0-0 c6 6.b3 e7 7.e3 0-0
8.b2 b6 9.c3 b7 . This position is
covered in Reinderman-Irwanto (Game 23)
and knowledge of this game should be
sufficient against most English/Réti set-ups. If
White goes more offbeat, your first few
moves involved gaining central control and
developing pieces, so I suggest you keep
doing that.
A14
Reinderman,D
Irwanto,S
23: Jakarta 2011
[Sam Collins]
1.f3 The games in this section began
through various move orders.
[ For example, Kramnik-Jones, covered in
the notes to Black's 10th, in fact began
1.f3 c5 2.b3 d5 3.e3 f6 4.b2 e6
5.g3 c6 6.g2 e7 7.0-0 0-0 8.c4 b6
9.c3 b7 . Gawain was aware that this
was a known position (albeit one not in his
repertoire, which normally features a
kingside fianchetto for Black, such as in
the Sicilian Dragon and King's Indian
defences in which he is a leading expert)
and was spending a lot of time in the
opening trying to find a way to deviate.
The fact that he didn't find such a way, to
me, indicates a couple of points about this
variation. The first is that Black has
developed his pieces in a natural way, to
active squares, and has seized space in the
centre. This fully conforms to opening
principles and such a logical approach
simply cannot be bad. The second is that
Black's set-up is universal. Unlike fianchetto
set-ups (where the 1 Nf3 c5 2 b3 move
order is actually a bit of a pain), to say
nothing of lines like the Benko which
require White to play a very specific series
of moves in order for Black to reach his
desired formation, the Tarrasch can be
played against basically everything except 1
e4. This is of great benefit since it means
we have fewer move order issues after 1 c4
and 1 Nf3, and also means that we can gain
more experience in the Tarrasch structures
since they will arise in more of our games. ]
1...f6 QUESTION: This move certainly isn't
bad, but is it the best choice for a Tarrasch
player?
ANSWER: I don't think so. If Black is
comfortable with a Tarrasch set-up (...d7-d5,
...c7-c5, ...e7-e6), he is best to get these
moves on the board as soon as possible.
Since both 1...c5 and 1...e6 allow
transpositions to major king's pawn openings
after 2 e4, this suggests that 1...d5 is the best
option.
[ QUESTION: But what is the problem with
1...f6 ?
ANSWER: Let's see a possible
continuation: 2.d4 d5 3.c4 e6 4.c3
, and now if Black wants to get Tarrasch-
type positions he needs to play c5
here, but this is a Semi-Tarrasch.
QUESTION: Can we get back to the
positions we've studied?
ANSWER: Well, yes if White plays 5.e3
(which is a Symmetrical Tarrasch), but
White's main option 5 cxd5 creates more
problems. The normal Semi-Tarrasch move
is to recapture with the knight, which leads
to entirely different positions from those
we've looked at after both of White's main
moves (namely 6 e4 and 6 e3). ( Thus Black,
if he wants to play a regular Tarrasch, needs
to meet 5.cxd5 with exd5 .
But now we've been tricked into a position
which could arise after 1 d4 d5 2 c4 e6 3
Nc3 c5 4 cxd5 exd5 5 Nf3 where, instead of
playing the recommended 5...Nc6 (when 6
Bg5 can be met by 6...Be7), we have
developed our king's knight first. Black is
not lost here, or even much worse, but
White makes a nice score with the pin 6.g5
and Black needs to learn a different
defensive set-up.
One of the main points of 1 Nf3 is to trick
 the opponent into positions outside his
repertoire. It is always worth spending a
few minutes on the initial moves in order to
make sure we have the transpositions
covered. )]
[ My game with Socko from Riga 2013
(covered in the notes to Game 12) followed
a more usual move order: 1...d5 2.g3 c5
(since there is no pressure on the centre,
there is no reason for Black to commit to ...
e7-e6 yet) 3.g2 c6 . Now Black is
basically threatening 4...e5, so 4.d4
is logical, and then e6 5.0-0 f6 6.c4
. Here Black can transpose into Catalan
lines with e7 ( but the most logical choice
for a Tarrasch player is the simple 6...dxc4 )
. After a few minutes' thought, my opponent
found nothing better than 7.dxc5 xc5
8.cxd5 exd5 9.g5 0-0 10.c3 d4
with a transposition into Chapter Four. ]
2.c4 e6 3.g3 d5 4.g2 e7 5.0-0 0-0
6.b3 b6 As noted above, a true Tarrasch
devotee will set up the c5/d5/e6 formation
earlier to avoid getting tricked into positions
from other openings.
7.b2 b7 8.e3 c5 9.c3 c6 10.cxd5
The only serious move.
[ 10.d3 is sometimes played just to get a
position; for instance, d7 11.e2 ad8
12.fd1 d4 13.exd4 xd4 14.xd4 cxd4
15.b5 xg2 16.xg2 c5 17.b4 xb4
18.xd4 e7 with equality in V.Gashimov-
Wang Yue, FIDE Grand Prix, Astrakhan
2010 (½-½ in 62). ]
[ 10.d4 dxc4 equalizes on the spot, since
Black is even better after 11.bxc4?! a5!
with a rapid attack on c4. ]
10...xd5 This is the main line.
[ The move I wanted to make work was
10...exd5 , which is the most consistent
Tarrasch response, keeping pieces on the
board and not being afraid of having a
potential weakness on d5. Unfortunately,
after 11.d4 White seems to have the better
chances:
A) 11...e8 12.c1 f8 ( 12...e4
13.dxc5 xc3 14.xc3 bxc5 is similar to
11...Ne4 below and looks a little better for
White )
A1) 13.e2 e4 14.f4 a6 15.e1
c4 16.bxc4 xc4 17.a3 c8
was balanced in N.Vitiugov-D.Frolyanov,
Russian Rapid Championship, Olginka
2011.;
A2) 13.dxc5 bxc5 14.a4 e4
15.d2 a6 16.xe4 dxe4 17.xd8
axd8 18.fd1 f5 and Black had
sufficient activity to compensate for his
queenside weaknesses in G.Welling-C.
Sandipan, Gibraltar 2008 (½-½ in 49).;
A3) 13.e1 c8 ( 13...d7 14.dxc5
xc5 15.a4 b4 16.e2 e4
17.ec2 ac8 18.f1 ed8 19.b5
is already good for White, and after
d6?? 20.a3 Black resigned
immediately in V.Papin-A.Rychagov,
Voronezh 2012 ) 14.h3 b8 15.e2
e4 ( 15...c4 looks okay for Black )
16.f4 cxd4 17.xd4 xd4 18.xd4
was better for White in V.Ivanchuk-L.
Ljubojevic, Monte Carlo (rapid) 2002 (1-0
in 54).;
A4) 13.-- and now:;
B) 11...c8 12.c1 ( after 12.e2 a6
13.e1 c4 14.bxc4 xc4 15.c3 b5
16.d2 b4 17.a4 a5 , Black had
active queenside play in J.Cuenca
Jimenez-R.Ponomariov, Spanish Team
Championship 2009; 0-1 in 54 ) 12...a6
13.e1 cxd4 14.xd4 was slightly better
for White in V.Akopian-V.Iakemov,
European Championship, Rijeka 2010 (1-0
in 34).;
C) 11...e4 12.dxc5 xc3 13.xc3
bxc5 14.e2 was V.Kramnik-G.Jones,
London 2012.
As Kramnik explained in the press
conference after the game, the knight
belongs on d7 in such positions, from
where it can come to f6 and keep
everything defended. Regarding White's
ideas, Kramnik mentioned three: Qb2,
Qb5, and Rd1.
The game continued e8 ( Gawain had
overlooked that after 14...b6 15.b2!
he is forced to play the weakening ...f7-f6;
and 14...d7 15.b2! is similar;
while after the logical 14...d6 15.fd1
fd8 , Kramnik thought that the excellent
16.e1! would put the d5-pawn under
intolerable pressure, e.g. e6 17.d3!
and 18 Nf4 is threatened ) 15.fd1 f8
C1) or 16.e1 e7 17.d3
intending Nf4. b6 ( or if 17...g6 then
18.h4! and h4-h5 is coming ) 18.f4
ad8 , when Kramnik noted that Qg4,
Qh5 and Qe1, threatening Ba5, were all
promising;
C2) 16.b5 b6 ( 16...b8 17.e1
a8 18.a4 is clearly better for White )
 17.xb6 axb6 18.xd5 going for a
forcing line; ( 18.e1 e7 19.d3
ed8 20.a4! is also quite pleasant for
White according to Kramnik ) 18...d4
19.xd4 xd5 20.xd5 cxd4 21.xd4
(as Kramnik noted after the game, if
Black could trade a pair of rooks he
would get good drawing chances; but
with rooks on the board, White has
excellent winning prospects) a5!
a good move which sets a trap;
( 21...ab8 is met by 22.a4! ) 22.e4
( the tempting 22.c4 runs into b5!
23.b6 a3 , when 24.xb5?? loses to
b8 ) 22...c5 ( Kramnik analysed
22...b5 23.c1 xa2 24.c7 e7
25.c8 , intending Bc5, as winning for
White ) 23.c3 a3! ( 23...a7 24.a4
c7 25.d2 is easier for White ) 24.b2
a7 25.a4 c7 26.a2! an excellent
move overlooked by Gawain, defending
against ...Bxf2+ and keeping rooks on;
( and stronger than 26.c1
Kramnik suggested ee7!? , trying to
take on c7 with the rook after 27.g2
b4 , when the black bishop is more
active ) 26...h6 ( or 26...b4 27.d4
c1+ 28.g2 c5 29.xc5 xc5
30.b4 c7 31.a5 bxa5 32.bxa5 a7
33.a6 and White wins by slowly bringing
his king to the queenside ) 27.g2 h7
28.f4 ( Kramnik later suggested the
"pensioner style" 28.g4 , killing all
counterplay ) 28...f6 29.f3 d7 30.a5
bxa5 31.xa5 c8 32.b5 d6
( Kramnik gave 32...a7 as a better
drawing chance ) 33.b7 b6 34.f7
f8 35.c7 d6 36.d7 a6 37.d4
b8 38.c5 e8 39.g4 a2 40.h4
c8 41.b4 a3 42.h5 c3 43.d4
3c7 44.xc7 xc7 45.f5 d6 46.b5
c1 47.b6 b8 48.e6 h1 49.c5
and Black resigned.;
D) 11...-- ]
11.xd5 exd5
[ The main line is 11...xd5 , after which
12.d4 leaves Black with a number of
continuations but no route to easy equality.
In fact I had this position myself with Black
against an English Opening specialist: a5
13.a3 fe8 14.dxc5 xc5 15.xd5 xd5
16.xc5 bxc5 17.ac1 c4 18.bxc4 xc4
19.e5 xf1 20.xa8 a6 21.e4 c8
22.xc8+ xc8 23.f4 and a draw was
agreed in A.Kosten-S.Collins, Kilkenny 2009.
However, there are of course improvements
and, in general, the positions after 12 d4
strike me as slightly inferior and passive
for Black, which is not what Tarrasch
players normally want. ]
12.d4 a5!? This move has not been played
very often but has been tried by a number of
players for whom I have a lot of respect – in
particular, Andrei Sokolov, who would simply
not play an illogical variation with either colour.
Black takes advantage of White's slightly
passive approach (after all, Black is not
behind in development and his central pawns
are not under any immediate pressure) to
generate counterplay on the a-file. Having
examined all the games played here my
conclusion is that the line is quite playable,
and White can easily get into trouble if he
doesn't know what he's doing.
[ Another useful and little-played idea is
12...a6 13.e1 c4!? . Black aims for an
unusual structure where he has good
chances of developing active queenside
counterplay by pushing his pawns and
bringing a knight to c4.
A) after 14.c3!? c8 15.d2 d6
16.e5 rather than the cxb3 of I.Csom-L.
Jakobetz, Hungarian Team Championship
1993, ( Marin proposes the novelty 16...b5
17.a3 cxb3 18.eb1 with the
assessment: "White has the more
compact structure and an active position.
It should be noted that the c3-bishop
prevents the strong manoeuvre ...Na5-c4
which, if it were possible, would
immediately turn the tables." );
B) 14.bxc4 xc4 15.d2 ( 15.a3 c8
16.d2 a5 17.c3 e2 18.xe2
xc3 was J.Sadorra-M.Kravtsiv, Beijing
blitz 2008, and now 19.b5 would have
been a little better for White so perhaps
Black should try 17...Bb5 ) 15...a5
16.c3 b5 17.xc4 xc4 18.b1 b8
19.h5 ( Kosten suggests 19.e4
, opening the position for the bishops )
19...b4 20.a1 b5 21.e4 a3 22.e2
was A.Kosten-K.Landa, French Team
Championship 2008, and now Kosten
suggests xb1 23.xb5 c3 24.xc3
bxc3 , when the opposite-coloured bishops
make the position quite drawish. ]
13.dxc5 White has tried a bunch of moves
here without demonstrating a clear route to
an edge. In general, it seems logical to
respond to a move like ...a7-a5 by opening
the centre, since Black's last move did nothing
to protect his d5-pawn. However, it turns out
that Black has sufficient dynamic resources to
meet this direct approach.
[ 13.d2 a4 14.fd1 axb3 15.axb3 xa1
16.xa1 a8 17.c3 d8 18.b2 f8
19.e1 cxd4 20.xd4 xd4 21.xd4
a5 22.d3 b5 23.b2 d6 24.f4
xf4 25.gxf4 h6 was level in J.Hickl-A.
Sokolov, Swiss Team Championship 2007
(½-½ in 48). The pawn on d5 is a target but
it will be difficult for White to go after it
without exposing his king, especially now
that he no longer has a g-pawn. ]
[ 13.e1 a4 14.bxa4 a5 15.b1 c4
16.dxc5 xb2 17.xb2 xc5 was slightly
better for White in M.Sorokin-E.Romanov,
Russian Team Championship 2006, but
Black could have kept the balance by
simply developing with 13...Qd7. ]
[ 13.e5 xe5 14.dxe5 d7 15.h5 a4
16.ad1 ( 16.fd1 b5! is also preferable
for Black ) 16...axb3 17.axb3 a2 18.c3
e6 gave Black the advantage in A.Srebrnic-
I.Bulmaga, European Team Championship,
Porto Carras 2011 (0-1 in 59). ]
[ 13.b1 a4 14.d1 axb3 15.dxc5?
( 15.axb3 is equal ) 15...bxc5 ( why not take
the pawn with 15...bxa2? ) 16.axb3 b4
17.xa8 xa8 and Black had an active
position with strong hanging pawns in P.
Blatny-J.Pelikian, Groningen 1997 (½-½ in
42). ]
[ 13.a4 , as in C.Mena Crespo-Z.Corrales
Jimenez, Havana 2009, concedes the b4-
square without a fight; i.e. b4
with a wonderful position for Black. ]
[ 13.c1 a6 14.e1 b4 15.e5 f6
16.a3 fxe5 17.axb4 cxd4 18.exd4 e4
19.bxa5 bxa5 20.f3 b4 21.c3 d3
was agreed drawn in J.Forgacs-L.Sapi,
Hungarian Team Championship 1988, but
Black is slightly better in the final position
and should have played on. ]
13...bxc5 14.h4!? White goes for a forcing
approach, radically altering the character of
the position. Black can't allow the knight to
come to f5 so the following sequence is
mandatory.
[ Alternatively, White can try a quieter build-
up with 14.e1 a6 15.d3
A) , Delchev analyses 15...c8 16.e1 c4
17.f4 b4 18.xd5 xe1 19.xe1
( and in fact in this line 19.f6+!
wins on the spot ) 19...cxb3 20.axb3 e8
21.d1 with great compensation for the
exchange;
B) ; instead, I like the thematic 15...a4!
, creating queenside counterplay –
depending on White's response, the pawn
can capture on b3 or advance to a3;
C) 15...b4 16.xb4 axb4
was E.Pigusov-I.Lysyj, Russian Team
Championship, and now White should
have tried 17.xd5 xd5 18.xd5 ad8
19.fd1 e2 20.e1 xd5 21.xe2 a8
with some winning chances in view of his
extra pawn. ]
14...xh4 15.gxh4 xh4 16.xd5 d8!
[ 16...ad8 was tried in S.Makarichev-A.
Panchenko, Moscow 1981, but here simply
17.xc5 would have left Black with
insufficient compensation for the pawn. ]
EXERCISE: Does Black have more than a
draw after 17 Qxc5?
17.e5 Wisely trading the queens, though
White has nothing in the endgame.
[ ANSWER: He certainly does: 17.xc5??
allows more than perpetual: xg2 18.xg2
g4+ 19.h1 f3+ 20.g1 a6! 21.e5
g6+ 22.g3 e6 with a decisive attack. ]
17...f6 18.g3 Keeping the queens on (for
instance, with 18 Qf5) seems risky to me.
Queen and knight often co-operate better
than queen and bishop and, while the knight
on d8 doesn't look threatening at the moment,
when it comes into play White will regret
having traded off his g-pawn and light-
squared bishop.
xg3 While I extolled the dangers of the
middlegame for White in the last note, trading
queens is Black's best option here since the
queen on g3 is very active and covers the
king very well.
[ 18...h5 19.ad1 xg2 20.xg2
seems like an edge for White. ]
19.hxg3 xg2 20.xg2 e6
[ 20...a4 , as in L.Timurova-S.Bezgodova,
Russian Women's Championship, Elista
2000, is slightly less accurate since White
can open lines for the b2-bishop with 21.a3
e6 22.b4 . ]
21.a3 fb8 Black has no problems in this
endgame and, in fact, I slightly prefer his
position since he has the pawn break ...a5-a4,
whereas White has no active play and
struggles to make use of his bishop on a3.
22.ab1 b5 23.fc1 a4 24.b4 ab8
25.c4 f5 26.b2 f7 27.g4 fxg4 28.xg4
8b7 29.c4 b8 30.f4 g6 31.f3 h5
32.e4 8b7 33.g3 g7 34.f3 f5
35.c4 e7 36.e2 cxb4 37.xb4 d7
38.a3 bb7 39.c3 d1 40.ec2 bd7
41.c1 xc1 42.xc1 d3 43.c3 xc3
44.xc3 e6 45.e4 d6 46.b4 e8
47.c3 d6 48.b4 e8 49.c3
½-½
 Back to Contents Page
Chapter Nine

In Place of an Epilogue
(View in Game Format)

1 d4 d5 2 c4 e6 3 Nc3 c5

As a bonus for getting this far, I've decided to offer coverage of two lines which, while arising from a Tarrasch move
order, are actually different openings. Both give rise to dynamic play which I think would suit Tarrasch players well
and, while I've only tried one of these lines on one occasion, it resulted in a nice win. So here we are:
4 cxd5 exd5
4...cxd4 is an aggressive gambit continuation. Enough for the pawn? Maybe. SeeWang Yue-Wang Chen.
5 Nf3 Nc6 6 g3
Of course we're familiar with this position. The new concept (actually an old and forgotten concept) I'd like to cover
is the Swedish Variation...
6...c4

...followed by a quick ...Bb4 and ...Nge7. See Parligras-Miron.

There are a lot of lines in the closed openings which resemble the Tarrasch, and I encourage you to look them out
and test them. A good understanding of Tarrasch structures just makes someone a better and more dangerous
player. For example, Ivan Sokolov (whose model games with Black and White have featured in this book) has given
1 d4 d5 2 c4 e6 3 Nf3 Nf6 4 Nc3 c5 (the Semi-Tarrasch, which can lead to quite different positions to the regular
Tarrasch) 5 cxd5 cxd4!? a whirl twice (so far) in 2013. Clearly, in order to play this, one needs to feel comfortable
with an IQP. As it turns out, neither of his games resulted in an IQP for Black, either against L'Ami at Wijk aan Zee
in January (½-½ in 84) or against Jakovenko in the French Team Championship in June (0-1 in 34). But for any 1
d4 d5 players, Sokolov's games are compulsory viewing, so keep an eye on him.

Key Notes

1. The gambit variation 4...cxd4!? leads to interesting complications. Wang Chen has demonstrated how enduring
Black's development-based compensation can be, even against very strong GMs, and despite White's lack of
weaknesses.
2. The Swedish Variation, 6...c4, leads to positions analogous to the 9 Bg5 c4 lines. As there, White needs to play
actively in order to avoid falling into a worse position. I think this is a perfect surprise weapon, especially for players
who like 9 Bg5 c4.

1 d4 d5 2 c4 e6 3 Nc3 c5 4 cxd5 exd5

4...cxd4 – Wang Yue-Wang Chen
 4...cxd4 – Wang Yue-Wang Chen
5 Nf3 Nc6 6 g3 c4 – Parligras-Miron

Back to Contents Page

In Place of an Epilogue
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 As a bonus for
getting this far, I've decided to offer coverage
of two lines which, while arising from a
Tarrasch move order, are actually different
openings. Both give rise to dynamic play
which I think would suit Tarrasch players well
and, while I've only tried one of these lines on
one occasion, it resulted in a nice win. So
here we are:
4.cxd5 exd5
[ 4...cxd4 is an aggressive gambit
continuation. Enough for the pawn? Maybe.
See Wang Yue-Wang Chen (Game 24). ]
5.f3 c6 6.g3 Of course we're familiar with
this position. The new concept (actually an
old and forgotten concept) I'd like to cover is
the Swedish Variation...
c4 ...followed by a quick ...Bb4 and ...Nge7.
See Parligras-Miron (Game 25).
There are a lot of lines in the closed openings
which resemble the Tarrasch, and I encourage
you to look them out and test them. A good
understanding of Tarrasch structures just
makes someone a better and more dangerous
player. For example, Ivan Sokolov (whose
model games with Black and White have
featured in this book) has given 1 d4 d5 2 c4
e6 3 Nf3 Nf6 4 Nc3 c5 (the Semi-Tarrasch,
which can lead to quite different positions to
the regular Tarrasch) 5 cxd5 cxd4!? a whirl
twice (so far) in 2013. Clearly, in order to play
this, one needs to feel comfortable with an
IQP. As it turns out, neither of his games
resulted in an IQP for Black, either against
L'Ami at Wijk aan Zee in January (½-½ in 84)
or against Jakovenko in the French Team
Championship in June (0-1 in 34). But for any
1 d4 d5 players, Sokolov's games are
compulsory viewing, so keep an eye on him.
D32
Wang Yue
Wang Chen
24: Chinese Championship, Xinghua 2012
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 cxd4
This dynamic gambit is a speciality of the very
strong Chinese IM Wang Chen, who has used
it to good effect against a number of Chinese
Super-GMs. It has been used several times by
the very strong Romanian IM Miron as well. I
expect that at least one of these players will
be a GM by the time this book goes to print,
so you could do worse than follow their
games!
5.a4+! The most precise move order.
[ 5.xd4 c6 gives Black an extra option
(see note 'b1'):
A) 6.a4 exd5 7.f3 d4 8.e4 ( or
8.b5 b4+ 9.d2 xd2+ 10.xd2
ge7 11.g3 0-0 with decent play for
Black in D.Harika-P.Cramling, Beijing blitz
2012 ) 8...f5 9.g3 e6 10.e4 a6
11.d3 b4+ 12.d2 xd2+ 13.xd2
f6 14.0-0 0-0 with good play for Black in
V.Malakhov-Wang Chen, Chinese Team
Championship 2008 (½-½ in 41).;
B) 6.d1 exd5 and now:
B1) 7.xd5 allows Black to transpose
into the main game with d7 ,
( but it is also interesting for him to try
the endgame after 7...e6!? 8.xd8+
xd8 . Black has a number of threats
including ...Nb4 and White has to be
careful. This position has little
theoretical significance since Black can
elect to play the main line after 7...Bd7 in
any event, so 5 Qxd4 (unless followed
by the relatively harmless 6 Qa4) just
seems to be an inaccurate move order. );
B2) 7.e3 was played against me once:
f6 8.f3 c5 ( if Black's bishop is
aiming for c7 then 8...d6 seems to be
a quicker route: 9.e2 0-0 10.0-0 g4
11.b5 e7 12.b3 d7 13.b2
and a draw was agreed in a complex
position in B.Posedaru-G.M.Todorovic,
Belgrade 2012 ) 9.e2 0-0 10.0-0 a6
11.a3 b6 12.b4 c7 13.b2 d6
14.c1 e8 15.g3 h3 16.e1 ad8
(this looks like an ideal IQP formation
and White had to try to disrupt Black's
game with 17 b5) 17.d4?! xd4
18.xd4 b6 19.h4 d7 20.f3 c6
21.ed1 e6 ( already 21...d4
works well ) 22.e2 .
EXERCISE: Find a thematic
breakthrough for Black.
ANSWER: d4!
B2a) I was expecting 23.xc6 bxc6
( I had seen the variation 23...dxe3
24.xd8 exf2+ 25.f1 xd8
26.xb6! xb6 27.d4 and White is
better ) 24.xd4 a2 with only a
slight advantage to Black;
 B2b) 23.xc6? dxe3 , and now the
key point is that after 24.xe8 ( 24.f3
xd1+ 25.xd1 xc6 is good for
Black; instead, White opted for 24.d4
, but after exf2+ 25.g2 xd4
26.xd4 bxc6 27.xf6 xd1 28.xd1
e1 , my French IM opponent
resigned in Nguyen Chi-Minh-S.Collins
Vandoeuvres 2013 ) 24...exf2+ 25.f1
( or 25.g2 xe2 ) 25...xe8!
White has no defence; e.g. 26.d2
g4 27.c5 e4 and wins.;
B3) 7.--;
C) 6.-- ]
5...d7 6.xd4 exd5 7.xd5 c6 8.f3
[ White can also delay the development of
this knight: 8.e3 f6 9.d1 c5 10.a3!?
(a similar idea to the main game) e7
11.f3 and now Black needs to decide
where to place his king: 0-0-0 ( 11...0-0
was the choice of the talented American GM
Ray Robson – after 12.e2 fd8 13.0-0
ac8 14.b4 d6 15.b5 b8 16.b2
e4 17.bd4 g4 18.c1 h5 19.g3 a6
20.e1 xd4 21.xd4 g5 22.xc8
xc8 23.h4 e4 24.g2 g6 25.f3
and a draw was agreed in A.Sharma-R.
Robson, Las Vegas 2012, but White is just a
pawn up for nothing and the result reflects
the rating difference and not the position )
12.c2 b8 can lead to aggressive play
with opposite side castling. A recent upset
occurred in the following game: 13.e2 g5
14.b4 b6 15.b2 c8 16.0-0 g4 17.h4
hg8 18.ad1 c7 19.b5 e5 20.b1
xb2 21.xb2 e4 22.d3 e5
23.xe4? ( 23.xa7! is very good for
White ) 23...xb5 24.f5 f6 25.d4
xf1 26.xf1 c4 and Black went on to win
in M.Turov-I.Karimov, Tashkent 2013 (0-1 in
34). ]
[ 8.g5 f6 9.d2 h6 10.h4 g5 11.g3
, as in I.Ivanisevic-M.M.Ivanov, Cannes
2006, can be met in a range of ways.
Personally I like the look of the endgame
after f5!? 12.xd8+ xd8 with some
initiative for Black. ]
8...f6 9.d1
[ Instead: Wang Chen has also defended
Black's cause after 9.d3 b6 10.a3 e7
11.e3 0-0 12.e2 fd8 13.0-0 ac8
when he again enjoyed the rapid and
comfortable development characteristic of
this line. After 14.d1 a5 15.c2 b3!
A) 16.xb3 xb3 17.a2 ( or 17.b1
f5 ) 17...e6! wins material;
B) 16.b1 b6 17.d4 d6 18.b4 c4
19.b3 c7 20.g3 a6 21.xc4 xc4
22.xc4 xc4 23.b2 e5 , Black's
bishop pair game him reasonable
compensation in Zhou Jianchao-Wang
Chen, Chinese Team Championship 2011
(½-½ in 64). ]
[ 9.b3 looks more natural, but in fact
leaves Black with a number of reasonable
responses since the threat to the b7-pawn
is illusory.
A) 9...c5 10.e3 0-0 11.e2 e6
12.a4 a6 13.0-0 b5 14.c2 b4
15.b1 c8 16.d1 e7 17.g5
( Houdini likes White after 17.h3
, but I don't clearly see how he is going to
unravel ) 17...c4 18.a3 xe2 19.xe2
fd8! with full compensation in Ki.
Georgiev-N.Sedlak, Montenegrin Team
Championship 2010 (½-½ in 30).;
B) 9...e6 is interesting too:
B1) 10.xb7 b4 forces a draw; for
instance, 11.b5+ ( or 11.d4 b8
12.xa7 a8 13.b7 b8 etc )
11...d7 12.e5+ e6 13.b5+ d7;
B2) 10.a4 a6 11.e3 b5 12.d1
c7 13.g5 e7 with advantage to
Black who has gained several tempi on
the main line, Ge.Grigore-L.Miron,
Bucharest 2011 (0-1 in 28).;
C) 9...-- ]
9...c5
[ Black has tried other moves, such as
9...b4 ]
[ and 9...a5 , but there seems no reason to
avoid putting the bishop on its most active
square. ]
10.a3!? A fresh and interesting attempt to
pose Black some problems, and advocated by
Kaufman in his repertoire book.
[ The traditional main line runs 10.e3 e7
11.e2 0-0-0 12.0-0 and now:
A) . I like the prophylactic move 12...b8!
– the black king needs to step off the c-file
in any event, so Black takes care of this at
once while retaining all his aggressive
options.
A1) . Now White has not yet tried 13.b4
in practice, perhaps because Black's
prospects seem better than with 12...g5
13 b4, as after xb4 14.b2
Black isn't weak on the a1-h8 diagonal,
( while 14.b5 only hits a well-guarded
pawn on a7. However, 13 b4 may still be
 worth investigating, since the quieter
moves seem to give Black good play. );
A2) . For instance, 13.a3 g5 14.b4 g4!
15.bxc5 ( 15.d2 b6 was agreed
drawn in D.Berczes-G.M.Todorovic,
Budapest 2012, which suggests that
White didn't like his position too much;
15.d4 , as in B.Dubessay-L.Miron,
Fourmies 2013, looks like a better try )
15...gxf3 16.xf3 e5 , when Black is
temporarily two pawns down but this is
fully compensated for by his initiative –
in addition, the position is much harder
to play with White, who immediately
goes wrong) 17.e2? h3! 18.b3
hg8 19.g3 xf1 20.xf1 ( 20.xf1
xc5 looks a little better for Black too )
20...f3+ 21.h1 ( 21.g2 improves,
but after h4+ 22.g1 xc5
White doesn't have enough for the
exchange ) 21...g4 with a material
advantage and an attack in J.
Richardson-R.Pert, London 2011 (0-1 in
32).;
A3) . How quickly White's game can go
downhill is demonstrated by the following
game: 13.d2 g5 14.b4 xb4 15.b5
c5 16.c3 g4 17.fd4 e4 18.b2
hg8 19.xc6+ xc6 20.d4 xd4
21.xd4 c3 22.xc3 e4 and White
resigned in P.Amoyal-L.Miron, Saint-Lo
2012 – not model play by White, it has to
be said!;
A4) 13.--;
B) The direct 12...g5 launches into
complications with pawn storms and
opposite-side castling. It seems well
established that White's best approach is
not to nurse his extra pawn but to strive to
open files on the queenside with the
aggressive 13.b4! . White has scored
heavily after xb4 ( Black hasn't done any
better with 13...g4 14.d4 )
B1) 14.b3 is the most popular move,
but Wang Chen has demonstrated
Black's resources here: b8 15.b1
e6 16.a4 f5 17.b2? ( Krasenkow
suggests 17.e4 with unclear play )
17...g4 18.b5 xb1 19.xb1 gxf3
20.xf3 e4 21.xh8 xh8 22.xa7
xa7 23.xb4 c8 and White didn't
have enough for the piece in Yu Yangwi-
Wang Chen, Qinhuangdao 2011 (0-1 in
66).;
B2) 14.b2 g4 15.d4 b8!
has also been tested, with promising
play for Black in an unclear position,
for instance V.Shishkin-T.Fodor, Baia
Sprie 2012 (0-1 in 69).;
B3) 14.b5!? is an aggressive recent
attempt: e4 15.c2 b8 16.b2
he8 17.e5!
B3a) Krasenkow gives 17...f5 18.ad1
xe5 19.c7+ a8 20.xe5 xe5
21.xd7 xc7 ( or 21...xd7
22.xd7 f6 23.f7 f8 24.c4 )
22.xc7+ b8 23.xh7 e7
24.xe7 xe7 25.e6 d2 26.c4
with a slight advantage for White in
view of his extra pawn, though Black
retains drawing chances with his active
pieces and opposite-coloured
bishops;
B3b) 17...xe5 18.xe4 c5
19.fd1 f6 20.ac1 c6 21.f5
xd1+ 22.xd1 a6 23.d4
and Black had regained his material,
but the much more active white pieces
(with outposts on d4 and f5) left White
clearly better in I.Ivanisevic-S.Sulskis,
European Team Championship, Porto
Carras 2011 (1-0 in 61).;
B4) 14.-- and now:;
C) 12...-- ]
10...0-0
[ Wang Chen deviates from his seventh
round game against the very strong GM
Zhao Jun, which saw 10...e7 11.g5
( 11.e3 would transpose to the note with 8
e3 above ) 11...h6 12.xf6 xf6 13.c1
b6 14.e3 d8 ( Kaufman analyses
14...0-0-0 15.b5 b8 16.e2
and Black has very little for the pawn"; but
14...d8 is better and more consistent with
how Wang Chen handles the position )
15.c2 f5 16.a4 0-0 17.e2 .
QUESTION: What does Black have for his
pawn deficit?
ANSWER: Black has two pluses in this
position. First, he has the bishop pair, which
can never be underestimated, especially in
an open position such as this. In addition to
their inherent ability to create and support
active counterplay, the bishop pair often
gives rise to good drawing chances in other
ways – for instance, if the queens, rooks and
a pair of knights were exchanged, if Black
could trade his light-squared bishop for the
remaining white knight, then opposite-
coloured bishops would be left on the board,
 which exhibit strong drawing tendencies in
heavily simplified positions.
Second, he has a lead in development, and
his pieces are more actively placed.
However, he needs to act quickly since
White is ready to castle and neutralize
Black's development advantage.
QUESTION: How would you continue with
Black?
ANSWER: g5! – I really like this move!
Because White has no weaknesses, it is
hard to see how Black makes progress after
more normal continuations; for instance,
after 17...Rfe8 18 0-0 Black has no
immediate threats and White is preparing to
fight for control of the d-file. Wang Chen's
aggressive pawn move aims to disturb the
white knight on f3, with a number of benefits:
1. White's kingside loses a key defender.
2. Black gains access to the d2-square,
which dramatically enhances the value of
his control of the d-file.
3. Black's knight can occupy a strong square
on e5.
It should be noted that White is still
perfectly fine in this position – as mentioned
before, he has no weaknesses and his
development is about to be completed. But
Wang Chen's aggressive play means that he
keeps sufficient compensation for the pawn.
The game continued 18.0-0 g4 19.e1
e5 20.e4 g6 21.g3 e6 22.f4
g5! 23.e4 ( Black is happy to play the
endgame after 23.xg5+ hxg5 , as the
white pieces are in a bind and, in addition
to other plans, Black can consider mobilizing
his kingside pawns with ...f7-f5-f4 ) 23...d5
24.f4 e6 and if I were White, I would be
tempted to take the repetition here. Instead,
he played on with 25.e4 , but after xf4
26.exf4 g6 27.g3 f5 Black had full
compensation and went on to win in Zhao
Jun-Wang Chen, Chinese Championship,
Xinghua 2012 (0-1 in 48). ]
11.g5
[ Kaufman advocates 11.f4 b6 12.e3
. Frankly, after fd8 this looks like Black's
normal compensation for the pawn in this
line, namely better development. In the
absence of white weaknesses I would
suggest that Black has about half a pawn's
worth of compensation, with some
additional practical chances based on his
position being easier to play. Certainly
Wang Chen's games have demonstrated
that these positions are not easy to handle,
even for very strong GMs. ]
11...b6 12.e3 e7
[ Not 12...xb2? 13.xf6 , since gxf6?
14.a4 wins. ]
13.d3 g4
[ The b-pawn is poisoned, because of the
standard idea 13...xb2?? 14.a4
, trapping the black queen. ]
14.c2 h6 15.h4 fd8
[ Perhaps Black should look for an
improvement around here. For instance, the
immediate 15...xf3 was interesting:
16.gxf3 e5 17.e2 ac8 18.0-0 g6
19.g3 h5!? attempting to provoke a further
weakness in the white kingside structure. ]
16.0-0 xf3 17.gxf3 d5 18.xd5 xd5
19.g3 White now has a clear advantage: his
two bishops control a lot of squares and
prevent Black from asserting that he has real
compensation for his pawn deficit. The only
"weakness" Black might point to is the
doubled f-pawns in front of the white king, but
these are hard to attack and, in fact, cover a
number of useful squares.
f6 20.ab1 ad8 21.fd1 h5 22.h7+
h8 23.xd5 xd5 24.e4 d7 25.h3 h4
26.f4 a5 27.d1 xd1+ 28.xd1 g5
29.h2 g7 30.d7 xb2 31.xb7 e5
32.xc6 . I think this gambit is underrated,
and it seems several top Chinese GMs might
now agree with me!
1-0
D33
Parligras,M
Miron,L
25: Rumanian Championship, Sarata Monte
[Sam Collins]
1.d4 d5 2.c4 e6 3.c3 c5 4.cxd5 exd5
5.f3 c6 6.g3 c4 The Swedish Variation.
Here's an introduction from Tarrasch expert,
Svetozar Gligoric. "The so-called Swedish
variation of the Tarrasch Defence, where
Black relaxes the tension in the centre,
creates a pawn majority on the queenside
and closes the d-file in order not to be left
with an isolated d-pawn. There is a price for
this – some delay with the development of the
pieces."
7.g2
[ Gligoric notes that the double pawn
sacrifice with 7.e4 dxe4 8.g5 xd4
 is risky:
A) 9.f4 b4 10.c2 ( not 10.gxe4??
xe4+ and White resigned in S.Khalilbeili-
S.Furman, Tbilisi 1956 – no comment
required! ) 10...g4 and Black was well on
top in V.Winz-G.Stahlberg, Mar del Plata
1941 (0-1 in 40).;
B) 9.e3 xd1+ 10.xd1 h6
( Scherbakov suggests 10...g4!? , since
11.xc4 xd1 12.xf7 is much better for
Black ) 11.d5 b4+ 12.xb4 xb4
13.xe4 c2+ 14.d2 xe3 15.xe3
e6 16.xc4 xc4 17.d6+ e7
18.xc4 f6 with a level endgame in A.
Zaitsev-V.Mikenas, USSR Championship,
Yerevan 1962 (½-½ in 29).;
C) 9.-- ]
7...b4 8.0-0 ge7 9.e4 This principled
break (now or in the next few moves) is
White's only real attempt at an advantage.
[ The immediate 9.a3 often transposes after
9...Ba5 10 e4 dxe4 11 Nxe4 0-0, but here
Black can try xc3 10.bxc3 0-0 11.d2
e6 12.e1 b5 13.b1 b8 14.e4 a5
15.f1 dxe4 16.xe4 d5 with an active
position in A.Onischuk-V.Akobian, Merida
2008 (½-½ in 43). ]
[ White can insert 9.e5 0-0 10.xc6 bxc6
before breaking with 11.e4 , but then Black
has a number of sound continuations; for
instance, e6 12.a3 xc3 13.bxc3 dxe4
14.xe4 d5 with a solid position in P.
Cooksey-A.Kosten, British League 2004
(½-½ in 46). ]
9...dxe4
[ Black can also let White capture on d5 after
9...0-0 , but I prefer the text. ]
10.xe4 0-0 11.a3
[ 11.c2 looks critical; for instance, g4
12.xc4 xf3 13.xf3 xd4 was Nguyen
Thi Thanh An-N.Zhukova, World Team
Championship, Mardin 2011, and now,
rather than taking on d4, all queen retreats
promise White a slight advantage. ]
11...a5 12.a4 b6 13.e3
[ Instead: 13.xc4 was Gligoric's choice. He
commented: "This complying with the
opponent's intentions is actually the result
of the better arrangement of White's pieces."
S.Gligoric-Bo.Kostic, Yugoslav
Championship, Ljubljana 1947 (1-0 in 81),
continued xd4 ( modern engines instantly
find 13...e6! , followed by 14...Nxd4 with
equality ) 14.xd4 xd4 15.e2
– "The key move, after which it is clear
that Black's position is difficult. The black
queen is exposed and the b7 square is
weak" (Gligoric). ]
[ 13.g5 h6 14.xe7 xe7 15.d5 xe4
16.dxc6 xc6 ( 16...bxc6 17.d2
was M.Euwe-G.Stoltz, Zaandam/
Wormerveer 1946, and now Euwe suggests
e2 with equality ) 17.xc6 bxc6 18.e5
a6 19.xc6 ae8 and Black had no
problems in D.Hergott-O.Rodriguez Vargas,
Manila Olympiad 1992 (½-½ in 27). ]
13...f5
[ Alternatively: 13...e6 14.c5 ( 14.ad1
d5 15.eg5 was L.Bernal Moro-O.
Rodriguez Vargas, Spanish Championship,
Lleida 1991, when f5! 16.xc4 c8
would have given Black interesting
compensation – he plans to take on e3 and
build up on the e-pawn, trying to open the
game for his unopposed dark-squared
bishop ) 14...xc5 15.dxc5 was V.Korchnoi-
O.Rodriguez Vargas, Salamanca 1991, and
now I think f5 equalizes. Gligoric believed
that this had to be preceded by ...h7-h6 in
view of 16.g5 ( instead 16.f4
is a better try, but then e7 (targeting c5)
maintains the balance ) , but in fact 16...d5
is completely fine for Black. ]
[ 13...f5 14.c5 xc5 15.dxc5 d3
16.fd1 f5 17.f4 ( but here the
intermezzo 17.g5! f6 18.f4 would have
been better for White ) 17...f6 was also
fine for Black in J.Gomez Esteban-O.
Rodriguez Vargas, Salamanca 1991 (½-½ in
36). ]
14.xc4 e7 15.c5!? Returning the pawn
is a practical choice,
[ but White can play more ambitiously:
15.d5!? leads to complicated play, where it
seems that Black's chances are no worse
after xe3 16.fxe3 f5 17.fd2 e5! . ]
[ 15.d3 d8 16.c3
A) the d-pawn is poisoned: 16...cxd4?
17.xd4 xd4 18.d5 e5 ( or
18...xd5 19.xd5 xe3 20.fxe3
with insufficient compensation for the
exchange; note that Black still needs to
develop his queenside pieces ) 19.f4;
B) 16...e6! gives Black decent
compensation for the pawn. . Perhaps
White's best is to return it immediately; i.e.
17.ad1 cxd4 18.xd4 xd4 19.xd4
xd4 20.e3 with an edge in view of his
dominant g2-bishop. ]
15...cxd4 16.xd4 xd4 17.b4
This is a critical position, and one of a type knight on c5 for the moment, but possibilities
which frequently occurs in the Tarrasch after of ...a7-a5 or ...Bc7 (or both) mean that
the white e-pawn and the black d-pawn have only Black has winning chances here, even
been traded. From a structural point of view, if White should be able to hold. ]
this suggests that Black should have ½-½
equalized, since he no longer has any formal
weaknesses and the pawn structure is
completely symmetrical. However, Black
needs to be extremely careful over the next
few moves. White, by virtue of moving first,
has a lead in development and, in particular,
his bishop on g2 is much stronger than its
counterpart on c8. Black needs to find a
decent way to develop his c8-bishop without
losing the b7-pawn.
EXERCISE: What would you play here with
Black?
xf3+ Going in for a slightly speculative
sacrifice; however, I think it would have been
simplest to take on c5 immediately.
[ ANSWER: 17...xc5! 18.bxc5 ( 18.xc5
xc5 19.bxc5 b3 collects the c5-pawn,
though after 20.ab1 xc5 21.d4
Black will have to return the pawn – on b7 –
in order to untangle ) 18...xf3+ 19.xf3
e6 20.fe1 f6 21.d3 ac8 22.e3
e7 and the b7 and c5-pawns will be
exchanged, with a draw. ]
18.xf3 b8
[ It is too late for 18...xc5 in view of
19.xc5! with advantage. ]
19.fe1 f6 20.d5 This is now slightly
unpleasant for Black, who needs to find a way
to complete development. Miron comes up
with an enterprising pawn sacrifice which
seems to me to level the chances.
h3
[ 20...h5!? was logical too. ]
21.a2 be8 22.ae2 xe2 23.xe2 g6
24.xb7 d8
[ 24...c3 is more active, since the
endgame after 25.e3 xe3 26.xe3
should be an easy hold for Black. ]
25.e3
[ 25.c1 retained some advantage. ]
25...g7 26.g2 f5 27.h4 d3 28.f4
xa3 29.e8 a1+ 30.h2 d1 31.h5 g5
32.e3 d8 33.e5 h6
[ The game score in my database shows that
a draw agreed was after 33...h6 34.f4??
, which I find hard to believe for obvious
reasons! Instead, I guess the players shook
hands here. Even now I would play on with
Black, who has the bishop pair if nothing
else. The bishop on b6 is contained by the