James Schuyler

the
dark knight
system
a repertoire with 1 Nc6
...

EVERYMAN CHESS
Gloucester Publishers plc www.everymanchess.com
First published in 2013 by Gloucester Publishers Limited, North burgh House,
10 Northburgh Street, London EC1V OAT

Copyright© 2013 James Schuyler

The right of James Schuyler 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 publication m ay be reproduced, stored in a

retrieval system or transmitted in any form or by any m eans, electronic,
electrostatic, magnetic tape, photocopying, recording or otherwise, without prior
permission of the publisher.

British Library Cataloguing-in-Publication Data

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

I S B N : 978 1 8 5744 995 2

Distributed in North America by The Globe Pequot Press, P.O Box 480,
246 Goose Lane, Guilford, CT 0643 7-0480.

All other sales enquiries should be directed to Everym an Chess, Northburgh House,
10 Northburgh Street, London EC1V OAT
tel : 020 7 2 5 3 7887 fax: 020 7490 3 708
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

Chief advisor: Byron Jacobs
Commissioning editor: John Emms
Assistant editor: Richard Palliser

Typeset and edited by First Rank Publishing, Brighton.

Cover design by Horatio Monteverde.
Printed and bound in Great Britain by Clays, Bung ay, Suffolk
 Abo ut the Author

James Schuyler is a F I D E Master. H e was Nevada State Champion in 2007 and won
the Virginia State Champion ship in both 2011 and 2012. He has been teaching
chess for over 2 5 years.
 Conte nts

Introduction 7
Weak Colour Complex 13

Section One: 1 d4lbc6 15

1 2 lLlf3 18
2 2 C4 34
3 2 dS 43

Section Two: 1 e4lbc6 55

4 2 d4 57
5 2 lLlf3 92
6 2 lbc3 114

Section Three: others 119

7 1 c4 lbc6 120
8 1lL:lf3 lLlc6 127
9 Others 129
10 Miscellaneous Topics 132

Illustrative Games 134

Index of Variations 213
Index of Games 222
 I nt rod uction
1 ...
ctJc6 and the Kevitz System
Why another repertoire book on 1 ...l2Jc6
- ? Didn't one come out just a few years
ago? As it turn s out, the subject matter
here is completely different, as are the
types of positions reached. 1...lZ:'lc6 isn't
really an opening yet, but the starting
point for many openings. This book
generally concerns itself with Black's
plans for using 1...lZ:'lc6 to force ... e7-e5,
as played by Tony Miles, for instance.
This idea is properly known as the
Kevitz System. (Wisnewski's repertoire
book on ... l2Jc6 was all about the
Nimzowitsch and Chigorin Defences, in
which Black plays l...ds or 2 .. ds.)
. okay so there will be no need to aban­
The Dark Knight System?
For reasons I will lay out, I will b e rec­
ommending a fianchetto of the king's
bishop if White resists the ...e7-e5 ad­
vance. These are in fact the most com­
mon positions, and they have a much
different feel from a typical Kevitz Sys­
tem. Furthermore, I believe that the fi­
anchetto is a substantial improvement
over the commonly played moves, and
therefore a new name is in order. Since
it is a black knight venturing out from a
dark square to initiate a strategy of
dark-square control... need I go on ?
Does the Dark
Knight System work?
lt works wonderfully, and in two ways.
Firstly, it can throw White on his own
devices as early as move one ! When
forced to improvise, even titled players
can play shockingly weak moves or ex­
pend their time and energy in the
opening. Secondly, the opening is fully
sound and playable against all calibre
of opposition . Black is fundamentally
don the repertoire just because your
opponents are no longer surprised.
Furthermore, this book is intended to
leave you a step or two ahead of even
very well-prepared opponents.
I would like to take the opportunity
to say that, in general, the quality of
play in Dark Knight variations has been
low for both colours. In many common
positions, as early as moves six, five,
four, and three (!) the unquestionably
best moves have been rarely or never
7
Th e Dark Kn ight System
played! To some extent this is under­
standable in an "unorthodox" opening,
particularly for White, who h as more
important things to worry about. lt is
less understandable for Black, who can
h ardly be surprised by his own open­
ing. However, from Black's standpoint,
this is highly correctable - and what
better opening to use than one in
which there is a long history of incom­
petence by the opponent, even at the
GM level ? I suggest that Black's practi­
cal results {which are by no m eans bad)
can be substantially improved.
Who plays it?
As far as I know, nobody plays the Dark
Knight in exactly the fashion I will be
recommending, but many strong play­
ers use large parts of the repertoire,
and most of the bits and pieces have
been tested in high-level encounters.
In spite of the relative obscurity of
the Kevitz System, it should be noted
that it was one of Miles's regular
weapon s again st both 1 e4 and 1 d4,
and IM Zvonimir Mestrovic plays it fre­
quently - they each have hun dreds of
1 ... ltlc6 (with the idea of ... e7-e5) games
to their credit.
lt h as also seen use by GMs Bogol­
jubow, Mikenas, Short, Hoi, Lazic, Veli­
mirovic, Benjamin, Gausel, Svidler,
Huang Thong Tu, Hart, A.Sokolov, Sul­
skis, Godena, lzeta Txabarri, Gulko,
Klinger, Rogers, Olafsson, Tolnai,
Art.Minasian, Ubilava, Sadler, Anand,
Dizdarevic, Gonzales, Speelman, Mohr,
8
Bachmann, Zarnicki, Gelashvili, Leko,
Johansen, Shkuro, Rohde, Karpatchev,
Ermenkov, Bezgodov, and de facto GM
Nikolaevsky - not to mention IMs Kjeld­
sen, Cvetkovic, Przewoznik, Vlassov,
Danailov, Tarlev, Barle {frequently),
Z. Nikolic, Vujadinovic, Wohl, Sommer­
bauer, Mascara, Matikozian, O' Donnell,
Eid, Ambrus, Kos, Bus, and presumably
many others that I h ave missed.
So, as we can see, not only are
strong players willing to play these po­
sitions (and against other strong play­
ers) they do so over and over, in some
cases without any expectation of sur­
prising their opponents. This says a lot
about the hidden consensus as to the
merits of the opening among those in
the know.
Coverage
This is a repertoire book, but I am not
adhering slavishly to the concept. Side­
lines for Black are presented if they are
useful or enlightening. One situation
that sometimes comes up is that a
main line, while objectively fine for
Black, offers very few winning chances.
In this case, I will try to offer an alter­
native which makes it more practical to
play for a win, normally with substan­
tial additional risk (otherwise it would
have been chosen as the main line).
Transpositions to other openings
are obviously frequent, but I will not
abandon the reader just because we
h ave reached a position that happen s
to be known by a different name. I will

mention transpositions when available
and cover the transpositions that I rec­
ommend.
That being said, it is not simple to
fit a whole Black repertoire into one
volume, and decisions needed to be
made about what to devote space to.
Except for here, I will not waste space
expressing the wish that I h ad more
space. However, if certain positions
receive light treatment, this is gener­
ally the reason . When deciding what to
focus on, I weighed both frequency and
danger, only intentionally ignoring
White moves that are both rare and
weak. Besides, space aside, I see no
point bogging down the reader with
information he won't need.
Transpositions? Aargh !
Why would anyone want t o learn inde­
pendent Dark Knight and Kevitz posi­
tions when they are just going to have
to learn regular (transpositional) open­
ings on top of it? One part of the an­
swer is that a player may greatly enjoy
the non-transpositional positions, and
these are reached frequently. Another
important part is that White normally
has to give up valuable options in order
to enter the tran sposition . For instance,
in the Pirc reached through the Dark
Knight System, White can only play the
Classical Variation which, though fairly
popular, is just not very challen ging for
Black. Admittedly, Black's knight
reaches the slightly unusual square c6,
and does so unusually early, but I will
In troduction
demonstrate that this is not a problem.
With White's options limited and Black
committed to this sideline, the study
material is relatively small.
To continue, a player who plays 1 e4
es must typically learn the Ruy Lopez,
Two Knights, Scotch, King's Gambit,
Vienna, and other sidelines. Compared
to this, the Scotch reached via the Dark
Knight is a light workload, not particu­
larly dangerous, and not a popular
choice for White. Therefore, play the
Dark Knight System still.
Oh, the humanity!
I am admittedly human and, further­
more, fallible, but I will refrain from
continuously hedging in the text (e.g.
"If my analysis holds up, it seems to me
that perhaps Black may indeed h ave
the better practical chances, though
this idea is untried and further investi­
gation is needed"). If there are particu­
lar doubts about conclusions, the nor­
mal solution is not to express them,
but to rectify them.
Hopefully I am far less fallible with
the help of chess engines, especially
Houdini (whom I sometimes refer to
affectionately as "Mr. H"). Everything
presented is computer-checked, which
offers the reader substantial protection
when relying on the analysis. However,
I have only used long computer­
generated variations when absolutely
necessary; i.e. there are no relevant
human games to draw from, and the
positions aren't settling down into
9
Th e Dark Kn ight System
something that can be understood and
assessed. In other words, fairly often .
I am inevitably prone to error when I
quote statistics, or when I say that a
move is new. These statements are nec­
essarily based on games I have access to.
I will try to avoid saying, "according to
my database" every time, since that
should be taken as a given. And I apolo­
gize in advance to the true originators
for such errors in attribution.
Untested? (*gasp*!)
In opening books, untested - or lightly
tested - moves are typically treated like
embarrassing relatives, introduced
quickly for propriety's sake and then
shuffled off to somewhere they won't
bother anyone. Admittedly it is far eas­
ier to discuss and analyse moves that
have been played repeatedly by GMs,
but ultimately moves need to stand on
their own merits, and we should not
shy away from a little work in order to
play better chess. Besides, isn't it good
to catch our opponents unprepared?
As for enemy novelties, it is also
sensible to be ready, especially if it is a
computer novelty. After all, if "my"
Houdini says a move is best, my oppo­
nent's will too, and I will soon be facing
this move at the board.
Who?
"I" is me, James Martin Schuyler. "You"
is you, the reader. "We" is n ot the royal
we - it is me and you, the reader. "Our"
opening is the Dark Knight System. I
10
am nobody i n particular. My qualifica­
tion for writing this book is the fact
that I wrote the excellent book you are
now holding in your hands.
Assessments
Chess writers will often tell you that
your understanding of a position is
more important than the objective as­
sessment. No doubt this is true, but this
is not a good reason to be unconcerned
with assessments. An objectively poor
position will require a great deal of
preparation and understanding in order
to be worth playing. Also, what if your
opponent happens to understand it
too?! Wouldn't it be better to take the
time to understand a sound position
instead of a questionable one?
I will try to convey as much of my
understanding as possible, but I am
also extremely concerned with the ob­
jective quality of the position (to the
extent that it is possible to determine
it). I do not want to place us one or two
inaccuracies away from an extremely
difficult position, nor do I want our op­
ponent to have the luxury of one or
two inaccuracies and still retain
chances for an advantage.
If you are not concerned with as­
sessments, simply ignore them, or cross
out the words and write in crayon,
" Black is okay". I do not find this useful,
but it is sufficient for many and true as
far as it goes - if the position were not
extremely playable, it would not be in
the book.

Houdini is not the final arbiter of
anything - especially since it is people
who must play the positions - but he is
a far stronger player than I am, and he is
nothing if not objective, so when look­
ing for what passes for the truth, his
assessments carry considerable weight.
When his opinions have not made sense
to me, I have looked deeper. Typically, I
have become convinced, but sometimes
I am able to convince him - rarely do we
continue to disagree.
Assessments in this book are in­
tended to apply to n arrow ranges.
"Equal" corresponds to an advantage
for one player of n o more than 0.09
pawns. "Comfortably equal" is the
more pleasant half of that range. "Tiny
advantage", "tiny edge", or "slightly
better" is an advantage of 0.10 to 0.17
pawns, while "nearly equal" would be a
similar disadvantage. In most chess
works, such positions are simply la­
belled as equal, but I believe that there
is far too big a difference between
+0.15 and -0.15 (two to three inaccura­
cies or even two to three tempi in many
positions) to let it go without mention.
An "edge" or "small advantage" is be­
tween 0.18 and 0.25 pawns. In other
works, such positions are often called
"approximately equal" or ±1=. I under­
stand that the style of assessment I am
using implies a degree of precision that
is difficult to attain, but I would rather
strive for precision and risk falling
short than strive for vagueness in the
hopes of evading criticism.
In troduction
I h ave not found it necessary to in­
clude in the repertoire positions worse
than a quarter pawn disadvantage, but
they are not uncommon in the notes. I
have not tried to be as precise in my
descriptions of theoretically unimpor­
tant positions, but the unadorned
words "advantage" and "better" mean
approximately 0.26 to 0.39 pawn s,
while 0.4 to 0.6 is a "comfortable ad­
vantage" and more would be "clearly
better" or some such, while more than
one pawn would be "nearly winning".
Personal history (with 1 ctJc6)
...
My love affair with 1 ... tLlc6 goes back to
the late '8os, and my trusty old Batsford
Chess Openings. I had owned it for
some time before I came across a single
line by Bogoljubow concerning the
amazing 1 d4 tt:Jc6 !?. Should White
"take the bait" and try to play a kind of
mirrored Alekhine's, a wonderfully in­
teresting position may be reached: 2 d5
tt:Je5 3 f4 tt:Jg6 4 e4 e5 5 f5 (??) 'i1Yh4+ 6
�d2 'ifxe4(?) 7 fxg6 'ti'xd5+ 8 �e1
'fkxd1+ 9 'it>xd1 hxg6.
11
Th e Dark Kn ight System
Although labelled as unclear, Black's
compensation seemed tremendous to
me, with three premium pawns and
the half-open h-file for a small knight.
My shoddy but practical analysis con­
firmed this: after 10 lt:Jc3 ?! c6 11 lt:Jf3
f6 ! 12 .id3 ? ! lt:Je7 13 .id2 dS 14 .ie2
lt:Jfs
12
White will b e lucky t o survive, even
should he find a defence to 1S ... e4 and
16 ... lt:Jg 3 17 l:tg 1 .ics. This is, in fact,
what many players tend to do as
White. Importantly, after the correct 10
c4! I still preferred Black.
Two of Bogo's opponents were kind
enough to allow s ... 'ifh4+! . Alas, after
26 years of 1 ... lt:Jc6, I have yet to bring
this variation to the board during a
tournament game. (And now I never
will. Even if White plays into it, I will be
obligated to correct Black's sixth move.
More on this in Chapter Three.) On the
plus side, I have yet to encounter any
real opening difficulties against any
calibre of opposition.
Therefore, play the Dark Knight Sys­
tem !
 Weak Colour Complex
There is no way to play chess well while
adhering to a single idea, or even two
or three ideas - the game is far too
complicated - but I have noticed that,
in the Dark Knight System, one concept
assumes far greater than n ormal im­
portance, and that is the n otion of the
weak colour complex. In many of the
high-level games that Black wins, it is
by taking advantage of White's weak­
nesses on the dark squares. I would
assume that most readers are familiar
with the idea of a weak colour complex,
but since it is especially important in
the DKS, I will prattle on about it any­
way.
In some positions, a player is more
likely to have problems because of
weaknesses on a single colour. The con­
ditions:
t Most of the player's pawns are
on a single colour (at least in a
certain area of the board). N atu­
rally, the weak colour complex
will occur on the opposite colour.
t The player is missing the bishop
that he would need to guard the
weak colour.
t The player's king is in the vicinity
of the weak squares.
t The player's opponent still h as
the bishop that can infiltrate on
the weak colour.
A player will usually experience
problems if three of the conditions are
met. Furthermore, if you notice two
conditions in your opponent's position,
it is worth seeing if you can aggravate
his situation.
A few more observations:
t Obviously, if a player's opponent
has no access to the "weak"
squares, there is no weak colour
complex, regardless of what
other conditions are present.
t A knight is a handy piece for the
invader to h ave, because it al­
lows him to extend the attack to
the other colour. (A knight sit­
ting on a weak dark square at­
tacks light squares).
t A space advantage is no protec­
tion against a weak colour com­
plex - it can even be a vulner­
ability.
13
 t Weak colour complexes are com­
mon in fianchetto openings - for
the opponent of the player that
fianchettos, that is. This is be­
cause of how each player is likely
to set up their pawns, and also
because the fianchettoed bishop
(the one that is likely to be infil­
trating the enemy weaknesses)
14
has extra protection against be­
ing traded off for a knight.
The weak colour complex will not
come up often in the theoretical sec­
tion because it does not usually appear
in a full-blown form until the middle­
game, but the idea permeates the
g ames section.
 Section On e
1 d 4 ti:Jc6
The Dark Knight first caught my at­
tention as a defence to 1 d4. Although
playing it against 1 e4 requires very
little additional knowledge, Black may
find it useful that his 1 d4 opponents
are unlikely to know the ins and outs of
the Scotch and the Pirc. Besides, since it
is not so easy (for most of us) to meet 1
d4, it is especially nice to find an effec­
tive defence. lt even neutralizes the
London, Trompowsky, and other white
"easy" systems.
Apart from 2 e4 which is covered via
1 e4 in Section Two, White has three
main moves:
2 t2Jf3 - Chapter One
2 c4 - Chapter Two
2 ds - Chapter Three
Others:
a) 2 e3 e5 (or 2 ... d6 and 3 ... g6, to keep
things interesting - White's pawn on
e3 makes a poor impression in this
King's Indian type of position) 3 t2Jf3
sees White try to play a French with an
extra move, but this move order gives
Black a few good options. 3 ... e4 4 t2Jfd2
f5 5 c4 t2Jf6 6 t2Jc3 �e7 ! transposes to
line B2 in Chapter Seven. Also possible
is 3 ... exd4 4 exd4 d5 (an Exchange
French) to dry up the game. Instead, 3
c4 transposes to A2; whereas 3 d5 is
schizophrenic nonsen se: 3 ...t2Jce7 4 c4
d6 5 t2Jc3 f5 6 �d3 l2Jf6 7 �c2 g6 gave
Black an extremely comfortable version
of a King's Indian in J.Paasikangas
Tella-T. Lindqvist, Finnish Team Cham­
pion ship 1996 (see Game 1).
b) 2 t2Jc3 e5 will soon transpose, af­
ter 3 d5 t2Jce7 4 e4 or 3 dxe5 l2Jxe5 4 e4,
15
Th e Dark Kn ight System
into positions considered in Chapter
Four (see 4 l2Jc3 in lines A and B respec­
tively).
c) 2 c3 e5 3 e4- see 3 c3 at the be­
ginning of Chapter Four.
d) 2 i.g 5 ? ! looks like an attempt to
outdo Black in the weird department,
though it does prevent ... e7-e5 for now.
Black must not shy away from the bi­
zarre: 2 .. .f6 ! (breaking the pin and chal­
lenging White to prove that the bishop
is well placed on the rim) 3 �h4 d5 !
(this is not normal for us, but White
h as given us space and tim e and cen­
tre, so we do not wish to be disturbed
by a belated d4-d5) 4 e3 (4 c4? ! e 5 ! 5
dxe5 �b4+ 6 lbc3 d4 7 a3 �e7 8 l2Jd5
fxe5 9 lbxe7 lbgxe7 is just good for
Black) 4 ...l2Jh6 ! 5 lbc3 (5 c4 e5 or 5 �d3
e 5 ! 6 dxe5 l2Jxe5 7 lbc3 c6 8 l2Jf3 l2Jxd3+
9 'ii'x d3 l2Jf5) 5 ... l2Jf5 6 �g 3 and we
h ave a few sound choices {6 ... e6, 6 ... g6),
but the point of our play is to pursue
the bishop, so let's go!
6 ...h 5 ! ? 7 i.e2 g 6 8 lbf3 �e6 (time to
take a break; 8 ... h4 9 i.f4 g 5 10 l2Jxg 5
fxg 5 11 �h 5+ �d7 12 �xg 5 is pretty
16
dangerous) 9 i.f4 i.f7 10 h 4 a6 1 1 o-o
'*'id7 and 12 ... 0-0-0 is equal.
e) 2 f4! ? is another radical way to
prevent 2 ... e5. In the very limited prac­
tice, Black h as done well with 2 ... d5 3
l2Jf3 i.g4 4 e3 f6 ! ?; e.g. 5 c4 e 5 ! 6 cxd5
�xd5 7 lbc3 �b4 and White was al­
ready worse in C. Depasquale-A.Ker,
New Zealand Champion ship 2001;
while after 5 �b5 'it'd6 6 o-o a6 7 ii.xc6
�xc6 8 c3 l2Jh 6 9 l2Jbd2 0-0-0 10 'i!Ve1
i.f5 11 �e2 e6 12 l:.e1 i.e4 13 c4 il.b4
14 cxd5 exd5 1 5 a3 i.xf3 16 'ii'xf3 i.xd2
17 i.xd2 f5, Black h ad continuously
maintained his grip on e4 and went on
to win with his good knight versus bad
bishop in J.Vialatte-F.Giroux, Paris 2006
(see Gam e 2). Of course, White also h as
3 e3 i.f5 4 �d3 e6! 5 l2Jf3 l2Jf6 6 o-o,
and although White h as no advantage,
he h as reached a Stonewall position
with which he is presum ably comfort­
able.
Nothing wrong with any of that, but
taking the opponent out of his comfort
zone is one of the things this book is
about, and for this purpose the brand­
spanking-new, 1-just-made-it-up,
2 .. .f5 ! ? fits the bill. No sane person
would play the Stonewall against the
Dutch - White gives up the e4-square
without getting the e5-square in re­
turn - so only 3 d5 can be critical. But
will the Stonewall player be h appy in
the resulting positions? He m ay not
play 3 d5 at all, but here's what h ap­
pens if he does: 3 d5 l2Jb4! 4 a3 l2Ja6 5
l2Jc3 l2Jf6 6 l2Jf3 l2Jc5 ! 7 e3 (or 7 b4 l2Jce4
 1 d4 lLlc6

8 i.b2 as 9 bS e6 10 dxe6 i.cs ! 11

lLlxe4 lLlxe4 12 i.d4 d6 13 e3 i.xe6 14
i.d3 Wke7 with a bizarre position that
slightly favours Black) 7 ... e6 8 dxe6
lLlxe6 (8 ... dS ! ?) 9 i.d3 g6 10 o-o dS and
while Mr. H calls it equal, I would be far
more comfortable sitting behind the
black pieces.
Seeing how the black army con­
verges upon the e4-square, White may
rethink his decision to evict the lLlb4,
but leaving it there is not convenient Mr. H likes White here, but what do
either, since White must fortify the dS­ you think? Without further ado, the
pawn and he can no longer play i.d3; not sane Stonewall Attack vs. Dark
e.g. 4 c4 lLlf6 S lLlc3 g 6 6 g 3 i.g7 7 i.g 2 Knight Dutch, which I can't wait to see
0-0 8 lLlf3 e6 is equal. happen in real life: 1 d4 lLlc6 2 f4 fS !
Alternatively, White may try to ad­ (I'm giving this move an upgrade as of
dress the weak e4-square by placing a now) 3 lLlf3 e6 4 e3 lLlf6 s ..td3 b6! 6 0-0
black pawn there: 4 e4! ? fxe4 S a3 (oth­ i.b7 7 a3 (if White doesn't play this
erwise s ... c6!) s ... lLla6 6 lLlc3 lLlf6 7 i.e3 soon, ... lLlb4 is going to be extremely
c6 ! ? (7 ... g6 leads to more "normal" po­ annoying) 7 ... lLle7 (7 ... g 6 ! ?) 8 c4 lLlc8 !
sitions) 8 i.xa6 bxa6 9 dxc6 dS 10 with ... ..te7 ... o-o ... lLld6 coming, and a
lLlge2 e6. wonderfully fun equal position.

17
 Chapte r On e

1 d4 lt:Jc6 2 lt:Jf3

This is the most common reply, and a find a convenient opportunity to relo­
very logical one for a 1 d4 player. White cate for some time, leaving the c-pawn
stops Black's planned 2 ... e5, while try­ out of play.
ing to maintain a familiar position White h as:
(unlike 2 d5).
2 d6
••• A: 3 C4 1 9
B: 3 ds 22
(: 3 .if4 26
D: 3 g3 30
E: 3 .tgs 33

There may be nothing wrong with
2 .. d5, but with 2 ... d6 Black continues to
. 3 ... e5 4 e4, which is now some sort of
fight for the e5-square. Also, after 2 ... d6
Black's tt::l c 6 will usually be presented
with an excuse to move soon, freeing
the c-pawn to join in the battle for the
centre. While the knight is n ot exactly
glued to the board in Chigorin-type
positions (i.e. after 2 ... d5), it may not
Instead:
a) 3 e4 is covered via 1 e4 - see
Chapter Five.
b) 3 tt::l c 3 tt::lf6 4 e4 also reaches
Chapter Five.
c) 3 e3 can be met by 3 ... g6.
d) 3 h 3 ! ? may be a trick to induce
Philidor. 4 ... exd4 5 tt::l xd4 g 6 ! ? is not
really bad, but 3 ... tt::lf6 4 tt::l c 3 g 6 5 e4
will transpose into lines we are more
familiar with - see line Cl in Chapter
Five again.
e) 3 c3 was used to good effect in
B.Kurajica-Z.Mestrovic, Bosnian Team

18

Championship 2003, continuing 3 ... e5
4 e4 l2Jf6 5 i.d3 (5 i.b5 ! ?) 5 ... i.e7 6
-2lbd2 and White went on to win . In­
stead, 3 ... l2Jf6 4 e4! g6 tran sposes to 4
c3 g6 at the beginning of Chapter Five.
A: 3 C4 g6!
Until I started researching this book,
always played 3 ... e 5 ? ! here, but if
White follows up correctly, 4 d5! will
lead to unpleasant positions for Black
(though this is not widely known). The
text move is in keeping with the idea of
provoking d4-d5 while leaving the a1-
h8 diagonal open, a common theme in
the Dark Knight.
4d5!
White should play this n ow, while
Black is mid-fianchetto, or it will be l ess
effective.
4 l2Jc3 i.g7 5 e4 i.g4 6 i.e3 e5 7 d5
..'2id4 8 i.e2 i.xf3 9 i.xf3 c5 is a com­
fortable variation of the Modern De­
fence in which Black has outscored
White, though the game should be
equal after 10 dxc6 bxc6 11 o-o l2Je7 12
c5! 0-0. Another option is 9 ... l2Je7 ! ?, and
1 d4 ltJc6 2 lDf3
if 10 i.xd4? ! exd4 11 l2Je2 c5 12 dxc6?!
i.xc6 13 l2Jxd4?! then 13 .. .'it'a5+ 14 �f1
'ii'c 5 regains the pawn with a clear ad­
vantage.
On 5 d5 Black could tran spose to
our main line with 5 ... l2Jb8, though
5 ... l2Je5 ! ? 6 l2Jxe5 i.xe5 7 e4 l2Jf6 ! 8 i.d3
o-o 9 l2Je2? ! (9 0-0 c6! 10 h 3 ! improves,
when White retains an edge) 9 ... l2Jd7 is
more fun. This was actually played in
M.Tratar-M.Srebrnic, Slovenian Cham­
pionship 2010, and A.lpatov­
R.Antoniewski, German League 2011.
Black h as equalized, reached a fascinat­
ing new position and, furthermore,
went on to win both times (see Games
3 and 4).
4 l2Jb8 5 l2Jc3 i. g7 6 e 4 l2Jf6 1 i. e 2 o-o
..•
This is now an obscure King's Indian
variation that can arise via 1 d4 l2Jf6 2
c4 g 6 3 l2Jc3 i.g7 4 e4 d6 5 l2Jf3 o-o 6
i.e2 l2Jc6 ! ? (6 ... e5 is "normal") 7 d5
l2Jb8. lt is also .. .
Position One
White's position certainly is large
and, indeed, he has the advantage, but
19
The Dark Kn ight System

Black has counterplay on the dark We will need to look carefully at:
squares. One important thing to realize
is that both ... e7-e5 and ... c7-c5 are poor Al: 8 h3 20
at this stage in the game (though they A2: 8 0-0 21
are frequently played). Both m oves take
squares away from Black that he will Al: 8 h3l2Jbd7
enjoy using, while neither m ove puts
any pressure on White's centre, or
anywhere else for that matter. If that's
not clear enough, notice that ... e7-e5 or
... c7-c5 will place Black two tempi down
in a main line King's Indian (Petrosian
System) or Benoni. Meanwhile, ... c7-c6
or ... e7-e6 actually does pressure
White's centre while reserving the cs­
and eS-squares for Black's pieces. Even
so, there's no rush, since White has no
convenient pawn break. Because Black With h 2-h 3 included, it becomes
has good control over when and how possible to start con sidering ... e7-e5,
much the board opens up, it makes though the immediate 8 ... es 9 i.g s has
sense for him to wait for a particularly scored 100% for White.
good opportunity. 9..ie3
Since Black's play is on the dark Logically, White takes measures
squares, White has a space advantage, against 9 ...l2Jcs.
and the game is not open, the trade 9...l2Jcsl
... �g4xf3 suggests itself - but in spite
of the favourable factors, it is still no
bargain to part with the bishop pair. In
any case, Black is not the one in control
of the trade, since White could have
played h 2-h 3 at any point going back to
move five. Indeed, GM Neverov and IM
Bonin did choose 5 h 3 when confronted
with this situation . Personally, that
would please me, as it costs White a
tempo and saves me from a difficult
decision. If you prefer ... �g4, play it as But we play it anyway! White's dark
soon as White plays ..te2. bishop is way too valuable to trade.

20

10 'ii'c 2
In Y.Balashov-G .Kuzmin, USSR
:hampionship, Vilnius 1980, White
ulayed 10 e5 tt:'lfd7 (10 ... tt:'lfe4! ?) 11 exd6
=xd6 12 �d4 tt:'lf6 and the g ame soon
oetered out to a draw (see Game S).
10 �xc 5 ? ! dxc5 11 e5?! (11 0-0 e 5 ! )
' s no good because of 11 . . .tt:'ld7 12 e6
"xe6 13 dxe6 tt:'le5 14 'ii'b 3 tt:'lc6 1 5 o-o
.i.xe6 16 'ifxb7 tt:'ld4 with the better
3ame for Black.
1o ... as 11 o-o tt:'lfd71
Black's main idea is to play 12 ... e5.
This resembles Yates's plan in line D of
:his chapter. Black improves his pieces
and clamps down on the dark squares
�fore engaging in pawn play. White
should also be concerned about the
!Xlsitional damage he could suffer after
for instance) 12 .l:r.ad1 �xc3 ! ? 13 bxc3,
:hough this is obviously a double-
edged sword.
Now if 12 tt:'ld4 e5 13 tt:'ldb 5 f5 with
good play, or 12 �d4 �xd4 13 tt:'lxd4 e5
14 dxe6 fxe6 1 5l:tad1 e 5 16 tt:'lf3 b6 17
:fe1 �b7 18 �f1 �g 7 19 .l:r.e3 tt:'le6
with equality. Black may consider
1 d4 ltJc6 2 tt:'lf3
....:xf3 at some point. As in many simi­
lar cases, White hurts his own position
by trading off Black's fianchettoed
bishop.
A2: 8 o-o a s
8 . . .�g4! ? is a logical alternative.
9 tt:'ld4
9 tt:'le1 is also played, when 9 ... tt:'la6
10 tt:'ld3 b6 11 �e3 tt:'lc5 or 11 ... tt:'ld7 is
similar to the main line. 9 h 3 is seen as
well, with a likely transposition to A1;
e.g. after 9 ... tt:'lfd7 10 i.e3 tt:'la6 and
11 ...tt:'lac5.
9 tt:'la6
...
9 ... e 5 ? ! is more common, but this
move of Stefano Rosselli del Turco's is
the most accurate. Piece play before
pawn play in this variation !
10 �e3 tt:'lcs 11 f3 es 1 12 tt:'lb3 tt:'lfd7
The two-time Italian Champion's
12 ...b6 13 'ii'd 2 tt:'le8? ! 14 l:!.ae1 f5 1 5
exf5 tt:'lxb3 16 axb3 gxfs 17 f4 was not
very successful in S.Flohr-S.Rosselli del
Turco, Zurich 1934.
13 'i'd2 b6
White may be slightly better be-
21
Th e Dark Kn ight System
cause of his extra space, but he has no
pawn play.
Meanwhile, we can advance .. .f7-f5
at any time; we are also threatening
14 ... lt:'lxb3 1 5 axb3 lt:'lc5, which will force
White to play an awkward m ove to de­
fend the b-pawn (or else give up his
good bishop).
B: 3 dS
The time t o play this was m ove two.
Now White will h ave great difficulty
finding an advantage.
3 ...lt:'les
The most accurate reply, though there
is a tendency for the position s to dry up
22
early. I f Black must play for a win,
3 ... lt:'lb8 is better, probably transposing
to line A above.
4 lt:'lxes
If White delays this capture, pre­
sumably with 4 e4, Black plays 4 ...ltJf6 5
lt:'lc3 ltJxf3 6 'ii'xf3 g 6 and the fianchet­
toed bishop will enjoy its open diago­
nal, while White's queen will soon need
to move again (but to no particular
effect); as for example in R.Fischer­
J.Schuyler, Richmond 2008 (see Game
6}.
Sometimes in blitz White tries to
avoid trading my "problem" knight by
playing 4 ltJd4, but White's knight is no
better off after 4... c5. If 5 dxc6 ltJxc6 6
e4, we are in a Sicilian with Black hav­
ing slipped in an extra move while
White's attention was diverted. Focus
on the board, White!
4... dxes 5 e4
This is nearly automatic, but not
obligatory - in L.Altounian-J .Schuyler,
Las Vegas 2008, White played 5 c4 e6 6
lt:'lc3 lt:'lf6 7 g 3 ! ? (see Game 7).
s ...lt:'lf6!
 1 d4 liJc6 2 liJf3

The move Black wants and needs to White is doing well with 7 f4! .
play is ... e7-e6, but after s . . .e 6 ? ? Black is
already lost! Then 6 i..b S+! i.. d 7 7 dxe6!
l.xbs 8 'ifh s ! was brutal in J .Bonin­
J.Schuyler, New York 1988 (see Game
8). I shouldn't feel too badly, I suppose
- Mestrovic has made this blunder, and
Miles made it twice! ! (Since Miles was
�iles, he actually lost neither game.)
White now has:

81: 6li'Jc3 23
82: 6 i.. b S + 24 7 i.. b S+ i.. d 7 8 dxe6 i..x bs!
Previously, both Jonathan Speelman
6 i..d 3 is less common and less logi­ and Emmanuel Bricard have tried
cal. H.Keskar-J.Schuyler, Norfolk 2008, 8 .. .fxe6, each securing a draw - though
continued 6 ... e6 7 c4 i.. c s 8 o-o o-o 9 if you see the gam es, you may not be so
::X3 exds 10 cxds, when I set about eager to repeat their methods.
blockading the d-pawn with 10 ...li'Je8 Instead, after 8 ... i.. xbs, White h as
11 i..e 3 i..x e3 12 fxe3 i.. d 7 13 Yi'f3 cs 14 several paths to the endgame, but al­
ifg 3 f6 15 .Uad1li'Jd6 and Black is a bit most no chance of extracting anything
better already. The game concluded 16 from the position.
b3 bS 17 'ifh4 �as 18 �cl b4 19 li'Ja4
=ac8 20 l::tf3 c4 21 .iifl cxb3 22 .:!xc8
.:.Xc8 23 axb3 i.. xa4 24 bxa4 b3 25 'ii'g 4
b2 26 Yi'e6+ 'it>h8 2 7 llxf6 'i¥d8 0-1.
Three years later, H.Keskar-
J .Schuyler, Hampton 2011, went 7 dxe6
l.xe6 with comfortable equality, albeit
with a long struggle for a win (see
Game 9).

81: 6li'Jc3 e6!

This is still the move Black needs to
play - and now he actually can, as my
computer explained to me a few years
ago. This is of great theoretical impor­
tance because following 6 ... a6 or 6 ... g 6,
gli'Jxbs
White should capture neither the
queen n or the f-pawn, though this is
not simple for him to figure out.
a) 9 'i!fxd8+ :xd8 10 liJxbS h as been

23
Th e Dark Kn ig h t System
played four times, but it is slightly
weaker than the text. Then 9 .. .fxe6 ! 11
tbxc7+ ci;f7 12 ttJbs .l:.c8 ! (better than
12 ... ttJxe4?!, as in I. Kreitner-H.Stenzel,
Long Island 1997) 13 tbc3 i.b4 14 i.d2
.i.xc3 15 i.xc3 ttJxe4 16 .i.xe s :xc2 is
equal .
b) 9 exf7+ costs a move (as opposed
to waiting for .. .f7xe6) and improves
Black's king position, so it is n ot a good
idea. 9 ... ci;xf7 10 ttJxbs (or 10 'ii'x d8+
.l::tx d8 11 ttJxbs ttJxe4 12 f3 a6 13 fxes
axbs and Black has slightly the better
of the probable draw) 10 .. .'ilt'xd1+ 1 1
ci;xd1 ttJxe4 12 ci;e2 c 6 13 tb c 3 ttJxc3 1 4
bxc3 is roughly equal again.
9 1\Vxd1+ 10 ci;xd1 0-0-0+ 11 �e2 a6
•••
12 tbc3 i.b4 13 ttJds ttJxds 14 exd s
l:txd s 15 exf7
White has a nominal edge because
of the isolated e-pawn, but the game is
all but drawn. Still, let's be careful and
avoid any mishaps by taking the a7-g 1
diagonal immediately before White's
bishop entrenches itself on e3: i.e.
1 S ... i.cs 16 i.e3 i.xe3 17 ci;xe3 .t1f8 18
.l:!.ad1, and now let's centralize the king :
24
18 ...l:.d6 19 ci;e4 .l:!.xf7 20 f3 ci;d7 - al­
though White can't win, he can try to
lose if he likes with 2 1 �xes l:.e7+ 22
ci;f4 .:!.e2.
82: 6 i.bS+ Ji.d7
6 ... tbd7?, as in E. Bukic-Z.Mestrovic,
Belgrade 1978, is an experiment that
should not b e repeated. Had White
seen 7 'i!Vh s ! our hero would h ave
found himself a pawn down for noth-
in g.
7 'ili'e2
7 'ii'd 3 is rarely played: 7 ... a6 8 i.xd7
'iix d7 9 o-o (D.Haessel-J .Schuyler, Paw­
tucket 2008, continued 9 tbc3 e6 10
i.g s i.b4 11 o-o-o o-o-o 12 f3 'ii'e 7 13
'iic4 h 6 with equality, though there
was still some play, and I went on to
win - see Game 10) 9 ... e6 10 c4 i.e7 1 1
tbc3 is obviously similar t o the main
line. White's queen has some extra op­
tion s, but his d-pawn is pinned. These
differences are important enough to
change Black's best method of coun­
terplay: 11 ...b s ! 12 �d1 b4 13 ttJe2 o-o
14 i.g s as with a slight edge for White .

One important point of Black's
queenside expan sion is that it secures
the c5-square for his bishop, ensuring
that it won't get shut out of play (as it
would if White were allowed to seize
space on the queenside with a2-a3 and
b2-b4}. White's knight h as also been
taken out of contact with the impor­
tant d5-square. Notice that if White
ever plays d5xe6, and for some reason
Black doesn't feel like recapturing with
a piece, .. .f7xe6 is position ally sound
because the f-file is valuable and the
e6-pawn controls critical squares.
Hold on ! Couldn't 12 ...b4 have been
prevented? Indeed, 12 a3 is possible for
White and not a bad move, but 1 1 ...b5
was not played with only 12 ...b4 in
mind. After 12 a3 o-o 13 .l:!.d1 exd5 14
cxd5 l:.fd8, Black's idea is to pl ay 1 5 ... c5
and, if White doesn't capture, 16 ...tt:Je8
and 17 ... ti:Jd6. White's edge is tiny.
7 a6
...
I'm not crazy about 7 ... g 6 ! ? with
that silly pawn sitting on e5, but it does
avoid spending a tempo on 7 ... a6, and
Mestrovic is 2-0 with it, which suggests
it is worth a try in a must-win game.
Check out D. Rasic-Z.Mestrovic, Croatian
Team Championship 2001, and J.Barle­
Z.Mestrovic, Slovenian Championship
1997, in the games section (Games 11
and 12).
8 i.xd7 'i¥xd7 9 o-o
Somehow White h as done well with
9 i.g 5, though it is not a move that
should cause problems. In P.Staniszew­
ski-H.Kaulfuss, Darmstadt 1996, Black
1 d4 tLlc6 2 tLlf3
certainly had the right idea and was a
bit better following 9 ... e6 10 i.. xf6 gxf6
11 dxe6 fxe6 (11 .. .'iWxe6 is also good} 12
o-o?! (12 'irh 5+ 'i¥f7) 12 ... 0-o-o 13 'ii'c4
l:tg 8 14 ti:Jc3, even if he went on to lose
after the passive 14 ...l:te8? ! (here
14.. .'�i'b8 15 .l:r.fd1 i.d6 16 b4 f5 was bet­
ter).
For White, the most accurate con­
tinuation may be 10 ti:Jc3 i.b4 11 0-0
i.xc3 12 bxc3 exd5 13 i.. xf6 gxf6 14
l:tfd1 'i�Vh5 15 'ii'f3 o-o-o, though he has
nothing to show for it - instead he
should have been accurate on move
three!
Alternatively, the untried 9 ... h6!? is
playable, although in this case 10 i.xf6
exf6 11 o-o f5 12 exf5 'ii'xf5 13 ti:Jc3
i.d6 14 tt:Je4 o-o 15 c4 is a tiny edge for
White.
9 e6 10 c4 i.e7!
...
Allowing Black to castle short. In­
stead 10 ... i.c5 ? ! 11 i.g 5 spells trouble.
11 ti:Jc3 o-o 12 1:td1 exds 13 cxds
White's d5-pawn is currently an as­
set, and Black h as two possible ways to
neutralize it. First, he can attack it di-
25
Th e Dark Kn ight System
redly with ... c7-c6. This is definitely
worth considering in some similar po­
sitions, but it doesn't work so well here;
e.g. 13 ... c6?! 14 i.. g s ! .:fd8 15 i..xf6
i..xf6 16 dxc6 'ii'x c6 and the eternal
knight sets in with 17 lt:lds.
If Black tries to prepare this with
... h7-h 6 (in fact not a bad move at all) it
is unlikely he will be fully ready for
14 ... c6 anyway; e.g. 13 ...h 6 14 a3 (14
i.. e 3 lt:lg4!) 14 ... c6 1 5 i..e 3 cxds 16
lt:lxds lt:lxds 17 l:!.xds ii'e6 18 :ad1 and
while Black's position is playable, it is
very dull, and White has a small but
clear advantage.
The second plan is very appropriate
here, which is a timely ... lt:le8 and
... lt:ld6. This idea is useful in many
variations of the Dark Knight, but espe­
cially the ones starting with 3 d5. The
point of the knight transfer is fourfold:
the knight is safe, as White's own d­
pawn shields it from attack; the knight
is active - centralized and controlling
the important e4-, fs-, c4-, and b7-
squares; the knight blockades the
strong d-pawn, so Black does not need
to worry about an eventual d5-d6 by
White; and finally, the knight has
cleared itself from the f-file, so it is now
possible (and usually desirable) for
Black to play .. .f7-fS. Coincidentally, the
.. .f7-f5 break also does (at least) four
things: frees Black's rook(s), isolates
White's d-pawn (or pressures White's
e-pawn), clears the second rank for
easy defence of the c7- and g 7-pawns,
and g ains space.
26
Therefore, 13 . . ..l:!.ad8 1 4 i.. g 5 lt:le8 15
i..x e7 'ifxe7 16 .l:f.ac1 lt:ld6 17 b3 fs and
with all Black has accomplished, per­
haps White should resign? Unfortu­
nately, chess is not quite that simple,
but Black can now start fighting on
equal terms.
C: 3 i..f4
White plays the London System,
which stops the ... e7-e5 break for now,
but the bishop bites a granite pawn on
d6.
3 lt:lf6!
...
To be honest, I 've always played
3 ... i.. g 4?! here, intending to force
through ... e7-e5 one way or another (4
c4 eS!; 4 e3 e S ! ; or 4 lt:lbd2 lt:lxd4! 5
lt:lxd4 es). However, in researching this
book, I found that 4 dS ! lt:lb8 gives
White a large advantage. Although this
h as only been played once in my data­
base and never against me, I believe it
is a bad idea to play moves one knows
to be poor, h owever unlikely it might
be to encounter the refutation. This
kind of "hope chess" is bad for one's
 1 d4 ltJc6 2 lLlf3

confidence and psychology - one is no s �e2 � g7

longer in control of the g ame.
As for 3 ... �g4?!, it is not actually sur­
prising that this move is questionable -
there are virtually no cases in the Dark
Knight where the bishop goes to g4
early. With Black's pawns set up on dark
squares, it is costly to trade the light­
squared bishop for a knight, and the
bishop rarely has anywhere decent to
retreat to. Should you wish to ignore my
advice and take your chances, I will
point out that 4 d5 ! is White's only good White has:
move, and it goes so far against the
grain for a typical London System player C1: 6 0-0 2 7
that you, too, may never encounter it. C2: 6 h3 28
4e3
If 4 d5?!, then 4 ... e 5 ! 5 dxc6 exf4 6 C1: 6 o-o 0-0 7 h3
cxb7 �xb7 7 �d4 d5 ! 8 �xf4 l2Je4! 9 c3
l.d6 10 �cl o-o 11 e3 c5 and, with
12 .. d4! coming, White's tiny material
.

advantage does not make up for all of

his pathetic grovelling . In stead, 5 �g 5
ti'Je7 6 c4 (or 6 l2Jc3 h 6 7 �xf6 gxf6 8 e4
a6 9 �d3 f5) 6 ... l2Je4! 7 �d2 (7 �h4?!
c6!) 7 ... g 6 is equal .
4 g6
...

4 ... l2Jh 5 has done well in practice;

for instance, 5 �g 5 h6 6 �h4 g5 7 �g 3
�xg 3 8 hxg 3 �g7 reaches a position
Black is normally happy to have even
with a tempo less. However, White
once again h as the annoying novelty 5
d5! l2Jxf4 6 exf4 l2Jb8 7 lLlc3 and Black's
knight excursions have placed him too
far behind in development, a situation
he will h ave trouble fixing; e.g. 7 ... g6 8
'ii'd4! .
Black is finally ready to play 7 ...l2Jh 5,
so White prepares a retreat, but this is
still a slow move. Instead:
a) 7 c3 l2Jh 5 8 �g 5 h6 9 �h4 g5 10
l2Je1 l2Jf6 11 �g 3 l2Je4 12 l2Jd2 l2Jxg 3 13
hxg 3 e5 14 l2Jd3 �e6 15 �f3 d5 is
equal. Of course, White wasn't playing
for an advantage anyway, but this isn't
the position he wanted either.

27
Th e Dark Kn ig h t System

b) 7 c4 e S ! 8 �g S (or 8 dxe s lL:\e4 9 spon se, and Black i s comfortably equal.

h 3 dxes 10 �h 2 �fs with an edge) g...l:te8
8 ... h6 9 �h4 gS 10 �g 3 exd4 11 lL:\xd4
lL:\xd4 12 exd4 lL:\e4 13 lZlc3 .l:f.e8 and
Black is a little better.
7 ...e51
After this unplayed novelty, White is
the one who must be careful to keep
the balance. Furthermore, in doing so,
he may not be able to maintain his
typically comfortable London structure.
8 �h2
8 dxes gives up more of the centre
for no gain. Black's point is 8 ... lL:\e4! and Simple chess! Black h as no problems
White's extra e-pawn is pinned to the at all. Moves worth considering in the
b2-pawn. 9 'ilVdS ? ! dxes 10 'ii'x e4 exf4 near future are ... ..tfs ( ... �e6), ... lL:\e4
11lL:lc3 fxe3 just makes matters worse ( ...lL:lds), ... �h 6, .. .'ii'd7, ... :b8, ...b7-bS. If
for White. lL:lbd2, then ... a7-a5 is a good idea,
8 ...exd41 while if lZlc3, then ... a7-a6 and ... b7-b5
A second annoyance for th e London is effective.
System player, who is hoping for ... es­
e4 at some point, after which White C2: 6 h3 0-0 7 c3
h as excellent long-term prospects be­
cause of the h2-bishop, which makes it
h ard for Black to generate meaningful
play on the kingside, whereas White
h as no difficulties making progress on
the queen side, due in large part to that
same piece. lt is not h ard to see why
that bishop appeals to so many players!
However, we will be sure to m aintain
control over it.
gexd4
9 lL:\xd4 may be slightly preferable This slows White's queenside play
(because the f3-square will be useful to a crawl, but it does avoid problems
for the e2-bishop), but White players on the long diagonal.
do not seem to consider such m oves. In 7 c4?! is a bad idea. Black was al­
any case, 9 ...l:te8 is still the correct re- ready better following 7 ... e s ! 8 �h 2

28

exd4 9 exd4 lt:le4! in K.Gunasekaran­
K.Akshayraj, Dhaka 2005, and went on
to win after 10 lt:lc3 ? (but if 10 i..f4!
.:e8 11 o-o g 5 ! 12 i.. e 3 g4! 13 hxg4
8g 3 ! 14 .l.:.e1 lt:lxe2 15 .l:!.xe2 i.. x g4 with
a big edge to Black) 10 ... lt:lg 5 ! 11 o-o
.:::.xf 3 + 12 i..xf3 lt:lxd4 with a free pawn.
7 lt:ld7
...
I shall soon try the untried 7 ... a5 ! ?. I
do like the idea of avoiding the ... e7-e5
break for the time being, as it makes
contact with the enemy where he is
already fortified. Black's plan is 8 0-0 a4
9 lt:lbd2 a3 10 b4 lt:ld5 11 'ii'c 1 lt:lxf4 12
exf4 - equal according to my com­
puter, but this does not even vaguely
resemble the position White was hop­
ing to play. Notice how we have rid
ourselves of the London bishop, while
starting to soften the long diagonal for
our own dark-squared bishop.
Naturally, White could stop the a­
pawn with 8 a4, in which case we go to
plan B: 8 ... lt:ld5 9 i..h 2 e5 10 o-o exd4
(the mini-operation succeeds - White
must give up the b4- or f4-square) 11
exd4 i..h 6 ! 12 .U.e1l:te8 13 lt:lbd2 lt:lf4 14
1 d4 lbc6 2 lbf3
i..xf4 i.. xf4 1 5 'ifh3 d5 16 i.. d 3 .l:txd1 17
.:.xd1 lt:le7 and although White is n ot
worse, he is once again without his fa­
vourite bishop. I've played the London
System quite a lot, so I know exactly
how annoying these ideas can be .
8 0-0 es 9 i.. h 2 'ii'e 7
This is the main line - which shows
that every once in a while a main line is
actually good! Interestingly enough,
Houdini prefers White, but extensive
practice shows the opposite. Black
normally continues with 10 .. .f5,
11 .. .'�i'h8, and then looks for a good op­
portunity to shut out White's London
bishop with .. .f5-f4!. Perhaps Mr. H is
underestimating the problem of the
h2-bishop. Sometimes Black prepares
.. .f5-f4 with ... e5-e4 and ... g 6-g 5 (or just
... g6-g 5). Meanwhile, White shoves the
a- and b-pawns. Sooner or later the c6-
knight gets kicked and usually re­
routes itself to the f7-square via d8.
These ideas come to life in
P.B.Pedersen-D.Bekker Jensen, Danish
Team Championship 2008, and
R.Valenti-V.Tkachiev, Corsica (rapid)
29
Th e Dark Kn ight System
1997 (see Games 13 and 14); whereas
V.Golod-E.Sutovsky, Netanya (rapid)
2009 (Game 15) shows Black, a strong
GM, fail utterly to contain th e London
bishop.
0: 3 g3
A subtle (read: boring) move.
Mestrovic has responded 3 ... e 5 ! ? here
five times, drawing all corners from
expert to GM. This is a good bet for
Black theoretically, since the endgame
after 4 dxe5 ttJxe5 5 ttJxe5 dxe5 6 'ii'x d8
'it>xd8 gives only a tiny edge for White.
Perhaps your opponent will turn away
from this Mutually Assured Dullness,
but I would as soon not give him the
opportunity. Thus, the usual solution :
3 g6!?
..•
When we must look at:
01: 4 d5 30
02: 4 i.g 2 31
01: 4 d5
If White is going to kick the knight,
he should do it now - though, accord-
30
i n g to m y database, 4 d5 has never ac­
tually been played.
4...ttJb 8
I have selected this retreat over
other options in many variations, and
the more I think about it, the more I
like it. Like General MacArthur, the
knight shall return, likely settling on
the newly soft c5-square. Incidentally,
if ... ttJb4 is played, it is with a similar
idea: ... a7-a5, ... ttJa6 and ... ttJc5.
5 i.g2 ..t g 7 6 0-0 ttJf6
I also like 6 ... e 5 ! ? 7 dxe6 (7 e4 ttJd7 8
c4 tbe7 is likely to transpose to D2; e.g.
9 tbc3 0-0) 7 .. .fxe6 8 e4 ttJh 6 ! 9 c4 tDf7.
7 c4 o-o 8 ttJc3
This line was topical in the 1920s (!)
with Frederick Yates seen frequently
behind the black pieces again st the
best players of his day, while Alekhine
and Grunfeld championed White. Rich­
ard Reti played both colours. Yates was
able to defeat Reti, Kmoch, and
Alekhine, the last of these games win­
ning a brilliancy prize at Carlsbad 1923.
8 ttJbd7
...
The most common move, 8 ... e5?!,

didn't work in the 1920s and still
doesn't work in the 21st century. Even
with the centre closed, Black will lose
too many tempi with his knights in
order to play .. .f7-fS. Another common
move, 8 ... c s ? ! , is equally illogical and
un successful.
The board resembles Position One,
with White's fianchetto not particu­
larly helpful to him. Black should de­
velop and establish his pieces on the
dark squares before initiating pawn
play.
9 h3
Prophylaxis against ...tbg4 and
... tt:iges.
9 ... a5 10 �e 3
Or 10 e4 tt:ics 11 'ii'c 2 and now, with
the knight anchored on the cs-square,
Black is ready for 11 ... es (12 dxe6 �xe6
=) 12 ... tt:ifd7 and 13 .. .fs.
10 ...tt:ics
Black intends 11 ... tt:ife4 with near
equality. 11 tt:ld4 �d7 transposes to
H .Kmoch -F.Yates, Hastings 1927/28, a
beautiful demon stration by Black of
how to build an attack (see Game 16).
1 d4 tt:ic6 2 tt:lf3
02: 4 �g2 �g7 5 0-0
5 ds is playable now and at any
point, but it was most forcing on move
four, when Black was obligated to play
... tt:ib8. Now Black also h as s ...tt:ies,
s ... tt:lb4! ?, and s ...tt:ias (!) as in
A.Galliamova-M.Krasen kow, Koszalin
1997 (see Game 17), though transpos­
ing to Dl with s ... tt:ib8 is obviously sim­
plest.
s .. tt:if6 6 c4 o-o 1 tt:ic3
.
By transposition we have reached
the fi anchetto variation of the King's
Indian Defence. There is an obscure but
logical sideline for Black that h as been
played successfully by strong players.
31
Th e Dark Kn ight System
7 ...tbd71?
This prepares both ...e7-e5 and .. .f7-
fs, and holds up White's c4-c5 break.
The immediate 7 ... e5 is frequently
played, but White has done very well
with 8 ds tbe7 9 c 5 ! .
8 e4
8 d5 tba5 (or 8 ...tbce5 9 tbxe5 tbxe5
10 'iib 3 and now 10 ... c5, 10 ... a5 or
10 ... b6 is level) 9 'ifa4 c6 10 �g5 tbc5 11
'ifb4 tba6 12 'i'a4 tLlc5 i s a draw.
s . es
. .
9 d5
These other moves are almost as
common :
a) 9 �g 5 �f6 ! ? (9 .. .f6) 10 �xf6 tbxf6
32
11 h 3 (11 ds lbb8!) 11.J:te8 12 a3 (12 ds
lbb8!) was A.Kotov-A.Lein, USSR Team
Championship 1962; then 12 ... exd4 12
tbxd4 a6 is a tiny edge for White.
b) 9 �e3 ? ! exd4 10 tbxd4 tbde5 1 1
tbxc6 bxc6 12 'ii'a4 cs and Black is bet­
ter; e.g. 13 tLld5 c6 14 tbc3 �b8 ! 15
l:.ad1 (15 b3 tbd3 !) 1 S ... l:tb4 16 'it'c2
tbxc4 with a big advantage.
g tbe7 10 tbe1 fs 11 tbd3 h6 12 f4
•..
This is not a very good move, but it's
what White h as been playing for. Black
must try to prove that White's position
is overextended.
12 ...exf4 13 tbxf4
13 ..txf4 lbb6 ! 14 'i'b3 fxe4 1 5 tbxe4
tbf5 is equal.
13 ... tbes 14 'ii'b 3 fxe4 15 tbxe4
15 �xe4 g 5 16 tbe6 l:txfl+ 17 �xf1
�xe6 18 dxe6 c6 gives Black the advan­
tage, since if 19 'i'xb7?! tbxc4 20 �xc6?
l:lb8 2 1 'ii'a 6 'ii'f8+ 22 'it>g 2 �d4, White's
exposed king loses him the game.
1s ... gs 16 tLlh s ::txf1+ 17 �x11 �h s
Things are still complicated, but
Black is slightly better due t o White's
looser position .
 1 d4 t'Llc6 2 t'Llf3

E: 3 i.gs?! deliberately provoked these moves,

only to find that his knight has no­
where to go. He will also find his i.g 3
awkwardly placed.
6 t'Llg1!
Limiting the damage. Others are
clearly worse.
a) 6 d5 gxf3 7 dxc6 bxc6 8 exf3 .l:tb8 !
with some advantage t o Black i n a
complicated position.
b) 6 t'Llh4?! e6! threatens 7 ...i.e7,
trapping the knight, which explains the
Preventing ... e7-e5 for now, this following contortion s: 7 h 3 h 5 8 e4 i.e7
move has given Black plenty of practi­ 9 d5 i.xh4 10 i.xh4 Vi'xh4 11 dxc6 g 3
cal problems, but objectively it is poor 1 2 f3 bxc6 and Black pockets a pawn.
to allow Black to play ... h 7-h 6 and ... g7- 6 ... i. g 7 7 e3 h S 8 h 3 h4 9 i.h2 t'Llh6 10
g5 for free with White's knight already hxg4 t2Jxg4 11 t'Llc3 es
committed to f3.
3 ... h6 4 i.h4
4 i.f4 g5 5 i.g 3 comes to the same
thing; 5 i.d2 g4 win s the d-pawn;
while after 5 i.c1 White may die of
shame. As a matter of fact, my engine
recommends both 5 i.c1 and 4 i.c1,
which certainly makes clear what it
thinks of this whole 3 i.g 5 fiasco.
4 . g s s i.. g 3 g4!
..

Dare I say it? This logical move is

another strong, unplayed novelty Black has an edge in this bizarre po-
which fully turns the tables. White has sition .

33
 Chapte r Two

1 d4 t2Jc6 2 c4

T o m y mind, this is already a l ax move. Personally, I like 3 ... i.b4+ 4 tt::l c 3

White puts up no resistance to Black's i.xc3 5 bxc3 d6, playing a kind of
logical follow-up. Presumably some Nimzo-lndian with ... e7-e5 in one go.
players are hoping for a Chigorin (after After 6 i.d3 f5 ! or 6 tt::lf3 e4 7 tt::l d 2 f5,
2 ... d5). the game can also be thought of as a
2 es
... reversed Grand Prix Attack (cf 1 e4 c5 2
tt::lc 3 tt::l c 6 3 f4 e6 4 tt::lf3 d5 5 i.b5). Of
course, 4 �d2 is also possible: 4... exd4 5
i.xb4 (5 exd4? 'ii'e 7+! wins the d-pawn)
5 ... tt::lxb4 6 exd4 tt::lf6 (not 6 ...'ii'e 7+ 7
i.e2 'ii'e 4? ! 8 'it>fl ! and Black's queen is
worse than White's king) 7 ti::lf3 d5 is at
least equal - the exchange of bishops
will help Black in the coming i sol ated
queen pawn position . If instead 7 a3
tt::l c 6 8 d5 ? ! , then 8 . . 'ii'e 7+! 9 i.e2 tt::l e 5
.

and Black is better, as White must fig­

3 dS
Instead:
a) The lame 3 e3 offers Black a few
methods: 3 ... exd4 4 exd4 d5 5 tt::lf3 tt::lf6
is an equal Exchange French position.
3 ... d5 ! ? is a kind of Chigorin/Albin
which has been played repeatedly by
many GMs - with poor results, how­
ever.
ure out some way to develop and
guard the c4-pawn.
b) 3 dxe5 tt::l x e5 (3 ... d6 ! ?) cannot be
dangerous either. Black's position re­
sembles a Budapest Gambit, but with­
out any of the inconvenience normally
associated with recovering the e-pawn;
e.g. 4 e3 tt::lf6 5 tt::l c 3 i.b4 6 i.d2 0-0 7
i.e2 c6! ? was fine and worked out well

34

for Black in R.Aghasaryan-A.Chibukh­
chian, Kajaran 2011 (see Game 18). And
4 e4? ! is an especially bad idea: after
4 ... ..ic5 White is already worse and
must be very careful; e.g. 5 ..ie2 'ifh4!
or 5 lLlf3 tt:Jg4 6 tt:Jd4 d5 ! ? 7 cxd5 't!Vf6 8
.1e3 tt:Jxe3 9 fxe3 'ifh4+ and White will
be needing both his chess resources
and his sense of humour as he plays 10
1td2.
c) 3 lLlf3 transposes to 1 c4 tt:Jc6 2
.'Z::lf3 e5 3 d4, covered at the beginning
of Chapter Seven .
3 ..ib4+!
...
3 ...tt:Jce7 is played 90% of the time,
and with excellent results, but if White
really understands what's going on,
Black will be forced to play positions I
cannot recommend; e.g. 4 lLlf3 ! tt:Jg6 5
M! ..ib4+ 6 tt:Jbd2 h 5 7 g3 tt:Jf6 8 tt:Jg 5 ! ?
and Black doesn't have much t o look
forward to - White has all the squares
and will soon have the bishop pair unless
Black makes a pathetic retreat. Black also
needs to worry (after 5 tt:Jc3 lLlf6) about
all of the "dynamic" tries mentioned in
Richard Palliser's 2005 book on the
Tango, some of which pose questions to
which Black has yet to find answers.
(Whenever Black reaches the Tango via
the Dark Knight, White is already com­
mitted with a pawn on e4, which takes
away all "dynamic" tries, leaving "classi­
cal" tries, which Palliser rightly considers
to be pleasant for Black.)
Attempts to transpose to a King's
Indian with 4 lLlf3 d6 do not bring hap­
piness either, because (to make a long
1 d4 lLlc6 2 c4
story short) White gets to break on the
queen side early with 6 c5 or 7 c5.
Instead, with 3 ... ..ib4+, Black devel­
ops his bishop before it gets obstructed
by ... tt:Je7 or ... d7-d6, incidentally solv­
ing his space issues and defusing
White's h2-h4-h 5 ideas. What could be
more logical ?
White blocks with :
A: 4 ..id2 3 6
B: 4 liJd2 40
4 tt:Jc3 ? ! is legal and it doesn't lose
material. So much for its positive
points. 4 ... tt:Jce7 5 �c2 (otherwise
5 ... SLXC3+) 5 ... a5 6 a3 SLXC3+ 7 't!Vxc3 d6
8 e4 f5 gives Black a comfortable posi­
tion typical of this chapter. Our devel­
opment is simple and we can consider
clamping down on the queenside with
... a5-a4 when we feel we can spare the
time. 9 �g 3 tt:Jg6 10 exf5 ..ixf5 11 h4
�e7 12 tt:Je2 tt:Jf6 13 h5 tt:Jf8 14 tt:Jc3
tt:J8d7 1 5 �g4 ..ic2 ! might make White
feel like a tough guy, but it does not
lead to an advantage.
35
Th e Dark Kn ight System
A: 4 ..id2 i.xd2+
5 'ii'xd2
Alternatively, 5 t2Jxd2 t2Jce7 6 d6 ! ? (or
6 e4 d6 7 ..id3 t2Jf6 ! - in this particular
position it is too costly to play 7 .. .f5?!,
activating both White's ..id3 and his
ridiculous t2Jd2; instead, Black plans
... o-o, ...t2Jg6, ...'ii'e 7, ...tLlh 5, and/or ... a7-
a5, ... b7-b6, ... ..id7) 6 ... cxd6 7 t2Je4 'ii'a 5+
8 'i'd2 'i!Vxd2 9 �xd2 tLlf5 10 g4 l2Jh4 11
t2Jxd6+ <:Ji;e7 12 c5 b6 13 b4 l2Jf6 14 t2Jf3
t2Jxf3+ 15 exf3 t2Je8 is equal.
s ...t2Jce 7
A1: 6 d6!? 3 6
A2: 6 t2Jc3 38
36
Here 6 �g 5 t2Jg 6 7 'i!Vxd8 <:Ji;xd8 8
t2Jc3 d6 is just equal, though not with­
out play.
A1: 6 d6!?
6 cxd6
••.
A.Hoffman-A.Femandez, Mar del
Plata 1996, went 6 ... t2Jc6? ! 7 t2Jc3 cxd6 8
tLlb5 t2Jf6 9 t2Jxd6+, and gives a good
example of what Black must avoid (see
Game 19).
7 'i!Vxd6
7 t2Jc3 t2Jf6 (or 7 ... d5 8 t2Jxd5 t2Jxd5 9
'ii'x d5 'ii'e 7!) 8 t2Jf3 d5 9 cxd5 (or 9 t2Jxe5
d6 10 t2Jf3 ..ie6) 9 ... d6 10 e4 o-o is no
problem for Black.
7 t2Jf6! 8 t2Jc3
..•
8 'ii'x e5 frees Black's d-pawn and
does nothing to address White's devel­
opment - he is still four {!) moves away
from castling kingside and the queen­
side is not a safe place: 8 ... 0-o 9 t2Jc3 d5 !
10 cxd5 (10 e3 ..ie6 11 t2Jf3 t2Jg6 12
'ii'd4 dxc4 13 'ii'x d8 .:taxd8 is equal; or
10 t2Jf3 t2Jc6 11 'ii'f4 'ii'a 5 12 cxd5 t2Jxd5
13 'ii'd2 t2Jxc3 ! 14 'ii'x c3 l2Jb4 15 l2Jd4
lld8 with considerable pressure; or 10

o-o-o i.. d 7 ! 11 lbf3 :res 12 cxd5? :es 13
"ii'd4 lbexd5 14 'it>b1 lbxc3+ 1 5 bxc3
-'Lle4, winning) 10 ...lbexd5 11 lbxd5
-'Llxd5 12 a3 (this sad move is necessary
to prevent 12 .. .'i!Va5+; castling just loses
after 12 0-0-0?? i.. e 6, with 13 ...l:tc8+ 14
ittb 1 lbc3+ coming) 12 ...l:te8 13 'fi'd4
(after thirteen moves, White's kingside
is hilarious - the most plausible expla­
nation is that he rolled very bad dice in
his Chaturanga game) 13 ... b 5 ! 14 lbf3
..i..b 7 15 e3 "fic7 16 i.. e 2 l:tad8 gives
Black plenty for the pawn.
8 0-0
...
Which brings us to:
Position Two
This position has never occurred,
but it is the obvious way to disturb
Black's easy play after 3 ... i..b4+, so I
think it will be contested many times in
the future. Furthermore, Black's meth­
ods of counterplay are tricky and there­
fore require special attention.
The first time I saw this position was
on an analysis board about 25 years
ago, when my teacher was trying to ex-
1 d4 lbc6 2 c4
plain to me why people didn't play
3 ... i..b4+. lt seemed obvious at the time
that Black's position was bad - after all,
the d-pawn is hopelessly backward. How
can Black possibly evict White's queen
and achieve the ... d7-d5 advance?
Often we can't, but there are other
options and, depending on how White
continues, it is actually possible to
"play around" White's queen and leave
the d7-pawn alone for the foreseeable
future. After all, it is securely guarded
and Black's pieces do have other ways
to develop. This idea is demon strated
by the variation 9 e4 lbc6 10 lbf3 'ii'a 5 !
11 o-o-o (11 i.. d 3 :es 12 i.. c 2 b6 13 a3
'ii'c 5 or 12 o-o J::r. e 6 13 'fi'a3 'ti'xa3 14
bxa3 and White can stop bragging
about his superior structure) 1 1 ... a6 12
'it>b1 .:te8 13 a3 b5 14 cxb5 axb5 15
i..xb5 i.. a 6.
Black's whole army is activated and
the d-pawn is neither an obstruction to
Black nor a target for White's counter­
play. Black's compensation is more than
sufficient. Notice the ... a7-a6, ... b7-b5
idea, which develops Black's bishop,
37
Th e Dark Kn ight System
weakens White's control over d5, and
opens lines against White's king. lt is
powerful enough that it can sometimes
be used even when White has not cas­
tled queenside (and ... b7-b5 can occa­
sionally be played without ... a7-a6).
Was White's play too co-operative?
Presumably 9 e4 is the culprit, provid­
ing a target for Black's ... t2Jf6 and
.. ."i¥a5. lt also leaves a hole on the d4-
square which could turn into a long­
term problem, although it did stop
... l2Jf5, a useful move for Black. There­
fore 9 e3 I:te8 10 t2Jf3 l2Jf5 11 11t'd2 d6 !
(11 ...b 5 ! ? is a great try, but 12 t2Jxb5
i.b7 13 i.e2 l2Je4 14 'i!i'c2 'ii'a 5+ 1 5 t2Jc3
t2Jxc3 16 bxc3 t2Jd6 17 J:txd7 .l::t ab8 is a
little better for White) 12 i.e2 (12 l:.d1
'ilt'b6 13 .ie2 i.e6 14 o-o h6 or
14 ... .l:tad8 ! ?; or 12 e4 l2Je7 13 0-0-0 'iVb6
14 'ii'x d6 'ii'xf2 15 t2Jxe5 and 15 ... i.f5,
1 5 ... i.g4 ! ?, 15 ... i.e6 ! ?, or 1 5 ... t2Jc6 ! ?
with equal chances i n all cases) 12 ...b6
13 e4 l2Je7 14 o-o .ib7 1 5 'ii'd 3.
By playing for ... d6-d5 (with the
queenside fianchetto) Black h as man­
aged to provoke e3-e4 again, which is
38
once again a target, and again leaves
the dark squares weak. On the other
hand, if Black doesn't find a plan, he
may find himself statically worse with
his backward d-pawn, despite White's
"bad" bishop. lt is Black's idle l2Je7 that
will save the day by repositioning to
target those soft dark squares: 15 ... t2Jg 6
16 g 3 t2Jf8 ! 17 l:tad1 l2Je6 ! 18 iVxd6 l2Jd4
(or 18 .. ."ifxd6 19 l:txd6 l2Jc5) 19 'ii'x d8
t2Jxe2+ 20 t2Jxe2 .l:.axd8, which is at
least equal for Black.
Can White save a tempo by keeping
his e-pawn flexible? Not unless he
wants an e-pawn shoved up his king's
file: 9 t2Jf3 .:r.e8 (or the wild 9 ... e4 10
l2Jd4 e3 11 fxe3 b 5 ! 12 t2Jdxb5 a6 13
tbc7 .l:.a7 14 l2J7d5 t2Jexd5 15 cxd5 llb7
with excellent play) 10 l:.d1 l2Jf5 11 'i¥a3
(11 �d2 e4 12 'iic 2 d 5 ! with advantage)
11 ... e4 12 l2Jd4 t2Jxd4 13 .:txd4 e3 and
although Black's d-pawn remain s,
White now has a matching one on the
lovely e2-square, and h as at least as
much to worry about.
A2: 6 t2Jc3 d6

7 e4
7 lL'lf3 fs 8 g 3 was V. Rao-J.Schuyler,
New York 1986 (see below).
7 .. .fs
7 ...lL'lf6 is surely playable, hoping for
a better opportunity to play .. .f7-f5, but
I prefer this active move in spite of a
few down sides; i.e. weakening the e6-
square, and opening the game for
White's lousy bishop. Black's rook will
soon be enjoying the f-file, and the lL'le7
will gain access to d4 via the fS-square.
After 7 .. .fs, there is only one game
in my database, H .Titz-C.Barlocco,
Dresden 2004:
8 exfs .txfs 9 .td3 lL'lf6 10 lL'lge2!
1 d4 lL'Ic6 2 c4
This knight covers the soft f4-
square. Instead, lL'lf3 is vulnerable to
... .tg4 and will be loose if White finds it
necessary to play g2-g3.
10...0-0 11 o-o .txd3 12 'it'xd3 lL'lhs
This typical move usually provokes
White into playing g2-g3, a long-term
weakness.
13 g 3 'ii'd 7 14 f3 a6 15 �ad1 .l:.ae8 16
lL'le4 h6
An important move: lL'lg s-e6 must
be prevented.
17 cs lL'lf6 18 lL'l2c3
There is nothing wrong with this,
but knights that control e4 are not
permitted to control d4 as well !
18 ... lL'lfs! 19 b4 lL'lxe4 20 lL'lxe4 gS
This is intended to discourage
White from playing f3-f4, which would
undermine Black's knight as it arrives
on d4.
21 '>t>g2 lL'ld4
The position has been equal since
move six, but somehow Black contrived
to win in 75 moves (see Game 20).
lt is now my great displeasure to
39
Th e Dark Kn ight System
show V. Rao-J.Schuyler, New York 1986. I
had not yet started to use 1 ... li:Jc6 regu­
larly, but it seemed like the perfect
choice again st the straight-laced,
booked-up senior master. We pick up
after 1 d4 li:Jc6 2 c4 e5 3 d5 i..b 4+ 4 i..d 2
i..x d2+ 5 �xd2 li:Jce7 6 li:Jc3 d6 7 li:Jf3 f5.
8 g3!? li:Jf6 9 e4 h6
Generally a useful move, preventing
li:Jg 5 and making ... g 7-g 5 possible. Here
Black prepares to play 10 ... fxe4.
10 exf5 i..xf5 11 i.. g 2 o-o 12 o-o
I was justifiably h appy with my po­
sition. Houdini prefers Black and sug­
gests ...�d7, ... i..h 7, .. JU7, .. Jlaf8. How­
ever, I was a 15-year-old expert, there-
fore .. .
12 ... g5??! 13 h4 g4 14 li:Jh2 'it>h7 15 f3
gxf3 16 l::txf3 i.. g6?! 17 l:taf1 li:Jeg 8 18
g4 'ike7 19 h5 i..e 8 20 'iVd3+ 'it>g7 21
li:Je4?!
I h ave been barely hanging on, but
there is a light at the end of the tunnel
since I will be in good shape if I can
reach an endgame. Also, it seems that
White's attack is not simple to play.
21 ....l:tf7 22 li:Jg3 i.. d 7 23 'ii'e 3 b6 24 g 5?!
40
hxg5 25 'ii'xg 5+ 'it>h8 2 6 li:Jf5 i.xf5 2 7
.l:txf5 .l:iaf8 28 h 6 'ii'e 8
29 'ii'g 6??
White, short on time and frustrated
about being unable to break through,
commits a h orrible blunder, allowing
his queen and rook to be forked. In­
stead, after 29 l:txf6 ! ? li:Jxf6 30 l:i.xf6
"fle7 3 1 ki.xf7 'ii'xg 5 3 2 l:lxf8+ 'it>h 7 a
crazy endgam e arises - presumably
White has some advantage.
29 li:Jh7???
.•.
From completely winning to com­
pletely losing in one move, as we ap­
proach the time control. In my haste I
both overlooked the fork and the fact
that 29 "flg6 attacks f7. White's posi­
tion would have disintegrated com­
pletely after 29 ...li:Je7. If you think I'm
over this after just 26 years, you would
be wrong . I resigned shortly.
B: 4 li:Jd2
Considering the pawn structure and
the closed position, it will be fine for us
to trade off our dark-squared bishop
for a knight.

Besides, White's knight would have
controlled the important e4-square, a
job for which White's dark-squared
bishop is uniquely unsuited. I beat fu­
ture-GM Jesse Kraai (then a senior
master) in this variation, though the
game doesn't survive. Apparently I am
not one of those awesome people who
remember every game they h ave ever
played.
4 tLlce7
...
5 a3
Just about everyone plays this, but
White has more challenging moves:
a) 5 lLlf3 and now Black can't play
s ... d6?? 6 'ii'a4+, so 5 ... .i.xd2+ is nearly
1 d4 lbc6 2 c4
forced: 6 .i.xd2 (don't worry about 6
�xd2 d6 7 'ii'g 5 tLlg6 8 �xd8+ �xd8 9
e4 h 6 ! 10 .i.d3 tLl8e7 11 o-o fs with
equal chances) 6 ... d6 7 e4 f5 8 exf5 lLlf6
9 .i.e2 o-o 10 o-o lLlxf5 ! with interest in
... e5-e4, or ... a7-a5, ...tLld7, ... tLlc5, ( ... b7-
b6), while ...h 7-h 6 is generally very use-
ful as well. Houdini claim s that all
roads lead to equality, but the fact that
he likes 11 .i.c1 a whole bunch seems
like a bad omen for White.
b) 5 'ili'a4 ! ? was played in B.Avrukh­
I.Ben Menachem, I sraeli Team Champi­
onship 1999 - an annoying move be­
cause it pins the d-pawn, complicating
the defence of es (if Black has to play
...lbg 6 in order to guard the pawn, it
defeats the purpose of 3 ... .i.b4+). The
game continuation 5 ... c5 6 a3 (6 lLlf3 ! ?)
6 ... .i.xd2+ 7 .i.xd2 lLlf6 8 lt:\f3 e4
(8 ... 0-0 ! ?; 8 ... lLle4! ?) 9 lt:\h4 o-o 10 d6
was pleasant for White.
Although Black has ways to try to
make 5 ... cs work, I prefer 5 ... a5 6 a3
(after 6 lLlf3 e4 7 a3 .i.xd2+ 8 tLlxd2 f5 9
g3 lLlf6 10 lLlb3 0-0 1 1 .i.h 3 d6 12 lt:\d4
.i.d7 1 3 �c2 a4 14 0-0 �e8 White's
bishops barely register as an asset;
15 ... b5 ! ? is likely to follow) 6 ... ..txd2+ 7
.i.xd2 lLlf6 8 lt:\f3 tLle4! 9 .i.e3 o-o 10
'Wi'c2 (10 tLlxes lLlf5 11 .i.f4 d6 12 lLlf3 g 5
1 3 ..tc1 'ii'f6 1 4 h 3 h 5 or 1 4... lLlh4 gives
Black considerable pressure for the
pawn) 10 ...tLlf6 1 1 tLlxe5 d6 12 lLlf3 (or
12 tLld3 .l:r.e8 13 .i.d2 .i.fs 14 e3 c6 1 5
dxc6 tLlxc6 and 16 . . .d 5 ) 12 . . .c 6 ! 13 dxc6
.i.f5 14 �a4 bxc6 with full compen sa­
tion; e.g. 15 .l:.d1 .l:.e8 16 .i.d4 lt:\e4 17 e3
41
Th e Dark Kn ight System
c5 18 .i.c3 lt:'Jxc3 19 bxc3 lt:'Jg6 20 .i.e2
l:tb8 21 o-o lt:'Jf4 or 21 ... .i.e4 ! ?, or 15
.i.d4 lt:'Je4 16 g4! ? .i.g6 17 h4 h 6 with
just a big mess.
For the cowardly it is n ot strictly
necessary to sacrifice a pawn : 8 ... e4
(instead of 8 ... lt:'Je4!) 9 lt:'Jd4 0-0 is play­
able, though White has a small advan­
tage after 10 'i¥c2 (or 10 e3 d6 11 .i.e2
.i.g4 12 f3, or 10 g3 c5 11 lt:'Jb5 lt:'Jf5 12
e3 b6 intending 13 ....i.a6, 14 ... .i.xb5
and 1 5 ... lt:'Jd6) 10 ... d6 1 1 g 3 c6 12 dxc6
ltJxc6 13 lt:'Jxc6 bxc6 14 .i.g2 d5 15 cxd5
cxd5 16 .i.c3.
c) 5 'ii'c 2 ! ? f5 6 lt:'Jf3 .i.xd2 7 lt:'Jxd2 d6
8 e4 lt:'Jf6 9 .i.d3 was tried in S.Ariste Cas­
tano-J.Salgado Gonzalez, Saragossa
1998, at which point Black panicked and
played 9 .. .f4?, a decision which led to
difficulties since he lacked counterplay
against White's big queenside (10 c5! is
best). Instead, 9 ... 0-0! 10 exf5 'i!Ve8 is ab­
solutely fine for Black; e.g. 11 f3 'iWh 5 12
g4?! 'ir'h4+ 13 'ito>d1 h5 14 g5 'ifxg 5 15
lt:'Je4 'ifxf5 16 lt:'Jg 5 e4!, or 11 h3 'ifh 5 12
g4?! ltJxg4 13 .i.e2 .i.xf5 14 �d1 'iWh4 15
.i.xg4 .i.xg4 16 'ifxg4 'ifxf2+ 17 'ito>d1 lt:'Jf5
18 lt:'Je4 'i¥d4+ and the suffering of
White's king is far greater than our
small material investment.
s ... .i.xd2+ 6 .i.xd2
6 'i¥xd2 d6 7 ii'g 5 is pointless be­
cause of 7 ... h 6 ! 8 'i!Vxg7 ? ! lt:'Jg6 9 lt:'Jf3
(White should probably prefer 9 h4!
lt:'Jf6 10 h5 �h7 11 �xh 7 lt:'Jxh 7 12 hxg6
fxg6 13 l:txh6 lt:'Jf8 14 lt:'Jf3, when 14 ... c6
42
o r 1 4.. JWd7 gives Bl ack slightly the bet­
ter chances) 9 ... lt:'Jf6 10 lt:'Jg 5 "ille 7 11 g4
lt:'Jxg4 12 lt:'Je4 .i.f5 13 lt:'Jg 3 .id7 14 .ih 3
0-0-0 15 .i.xg4 and now in D.Justo­
P.German, Buenos Aires 1995, Bl ack
spoiled some fine work with
1 5 ... .i.xg4?! 16 lt:'Je4 .i.f5 17 lt:'Jf6 which
was approximately equal, whereas
15 ... lldg 8 ! 16 .i.xd7+ �xd7 17 lt:'Jf5
Uxg 7 18 lt:'Jxe7 lt:'Jxe7 would have left
him with chances to convert his advan­
tage.
6 ... d6 7 e4 fs 8 exfs
Since the main line offers White
nothing, it makes sense that in
M.Gurevich-D.Zoler, Antwerp 1998,
White looked for and found something
else - 8 'ti'h 5 + ! ? - and was immediately
rewarded by 8 .. 5.fi>f8 ? ! , after which
Black's static king provided a nice tar­
get for Gurevich's king side storm (see
Game 21). Instead, 8 ... g6 9 'ii'h 4 fxe4 10
f3 exf3 11 lt:'Jxf3 lt:'Jf5 12 �xd8 �xd8 13
.i.d3 lt:'Jf6 leaves White with enough for
the pawn, but no more than that; e.g.
14 o-o .l:!.f8 15 .U.ae1 'ito>e7 (15 ... 'iit e 8 ! ?,
1 5 ... a5 ! ?) and now Mr. H sees nothing
better than 16 .ia5 'it>d7 17 .i.d2 'ito>e7,
repeating position.
s ... .i.xfs 9 lt:'Je2 lt:'Jf6 10 lt:'Jg3
Either 10 ... .ig6, as in E.Arlandi­
M.Lanzani, San Marino 1998 (see Game
22), or 10 ... 0-o, as in G.Grigore­
P.Brochet, Creon 1999 (see Game 23), is
good enough for equality, though in
the latter g am e Black contrived to lose.
 C h a pter Th ree
1 d4 'Dc6 2 dS
White takes the bull by the horns - a
surprisingly uncommon reaction. lt is
worth noting that this is how Miles
played against his own specialty when
he faced Zvonimir Mestrovic.
2 ttJes
•.•
White normally chooses between :
A: 3 e4 43
8: 3 f4 46
Others:
a) After 3 lLlf3 Black may wish to try
3 ....�:Jxf3 4 exf3 e5 (or 4... e6 or 4 ... g6),
but I don't care for it - White has de-
velopment, space, and open lines in
exchange for his anti-positional cap­
ture. Simply 3 ... d6 transposes to 1 d4
tt:Jc6 2 tLlf3 d6 3 d5 tLle5 in Chapter One.
b) 3 i.f4 tt:Jg6 4 i.g 3 - does the
bishop dominate the knight, or does
the knight dominate the bishop? In the
blitz g ame E.Bacrot-C.Bauer, Ajaccio
2007, it looked like the latter after
4... tt:Jf6 5 tt:Jc3 e5 6 e4?! i.b4 7 i.d3 d6 8
f3 tLlh 5 9 .if2 c6! 10 dxc6 bxc6 11 g 3
i.a5 12 f4?! tLlf6 13 f5 tLle7 1 4 tt:Jge2
tLlg4, when Black h as a huge advantage
(though he went on to lose). Bauer's
method of meeting 3 i.f4 should be
copied exactly. 6 dxe6 is an improve­
ment for White, but Black has nothing
to fear after 6 .. .fxe6 and 7 ...i.b4.
c) 3 tt:Jc3 e6 will soon transpose to
other lines in this chapter; e.g. 4 dxe6
fxe6 5 e4 (see line A) or 4 f4 tt:Jg6 5 dxe6
fxe6 6 e4 (see 6 tLlc3!? in line B).
A: 3 e4 e6
Mestrovic played 3 ... d6!? 4 f4 lLld7 5
tt:Jc3 c6 and went on to draw with GM
43
Th e Dark Kn ight System
Drasko (Game 24). Black has also scored
well with 3 ... lZ:lg6?!, which can transpose
into lines considered below, but his po­
sition is highly suspect after 4 h4!.
4 dxe6!
4 f4? ! is seen here most often and
seemed fine the first 5 5 times it was
used, but on the 56th, the German
master Lutz Diebl played 4 ... exd5 ! and
drew with GM Gyimesi in the
Bundesliga. Bravo! If White takes the
knight, Black draws by perpetual : 5
fxe5 'ili'h4+ 6 'it>e2 'Ml 5 + 7 'it>d2 'ii'h 6+ 8
'iii> c 3 'ifc6+ 9 'it>d3 'ifa6+ etc.
If 6 g 3 ?, as Gyimesi actually played,
then 6 ... 'ifxe4+ 7 'ife2 'ifxh 1 8 lZ:lf3 b6 9
lZ:lc3 and now, rather than Diebl's pre­
mature 9 ... SLa6?, the preliminary 9 ... c6!
maintains Black's nearly winning ad­
vantage (10 ... SLa6 is still coming to ex­
tricate the queen).
Meanwhile, White doesn't need to
go in for Gyimesi's contortion s (Gyim­
n astics?) to avoid a forced draw; he can
play 5 exd5 or 5 'ifxd5, although he
cannot h ope for an opening advantage
after such concessions. The best re-
44
sponse is 5 lZ:lc3! lZ:lg 6 6 'i!Vxd5 lZ:lf6 7
i¥d3 SLc5 8 SLe3 'Wie7 9 o-o-o o-o 10 lZ:lf3
with a small advantage for White.
Diebl s's 4 ... exd5 was tested again in
V.Erdos-R.Rapport, Hungarian Team
Champion ship 2012, to produce an­
other entertaining draw (see Game 2 5).
If it is Black who is keen to avoid the
draw, he needs to forego 4 ... exd5 and
pl ay 4 ... lZ:lg 6, transposing to position s
considered in line B below.
4 fxe6
...
This recapture is certainly dynamic,
although the somewhat exposed posi­
tion of Black's king requires careful
treatment.
Many players prefer 4 ... dxe6 5
'ifxd8+ 'it>xd8, but Black's results have
been poor - pretty much draws and
losses. Nonetheless, Short m ade it look
easy to defend in S.Gordon-N.Short,
British Championship 2011, so his
m ethod could certainly be tried (see
Game 26); and M.Gurevich-M.Rohde,
Philadelphia {blitz) 1989, shows that it
is possible for Black to win if White
overextends (see Game 27).

5 t'Llc3
This flexible move makes it h ard for
Black to activate his f8-bishop, since
s .. ..ic5?? loses to 6 'ifh 5 + and 5 ... i..b 4?
.
i s met by 6 'ili'd4!, targeting Black's
loose bits on b4, e5, and g 7.
Alternatively:
a) 5 f4 t'Llg6 is line B below. 5 ... t'Llf7 ! ?
and 5 ... t'Llc6 ! ? are fine too, but no better
than the text.
b) 5 t'Llf3 t'Llxf3 6 'ii'xf3 'i!Vf6 is already
equal according to Kalinin, but 7 'i!Vg 3
will gain some advantage - White's
queen is active while Black's is m ainly
awkward. Therefore, just 5 ... t'Llf7 which
i s similar to the main line (and trans­
poses after 6 t'Llc3 b6).
c) 5 ..if4 is untried but should lead
White to a normal plus; e.g. 5 ...t'Llg6
(5 ... t'Llf7 ! ?) 6 ..ig 3 ..ic5 7 t'Llc3 a6 8 t'Llf3
t'Llh 6 9 'i!Vd2 o-o 10 o-o-o d6 11 h4 b5.
d) 5 ..ie3 ? ! t'Llf6 6 t'Llc3 ..ib4 7 ..id4?!
t'Llc6 8 a3 ..ia5 9 e 5 t'Llxd4 10 'ii'xd4 t'Lld5
is a pleasant position for Black.
s ... b61
The fastest way for Black to mobi­
lize. The fianchettoed bishop is active
1 d4 t'Llc6 2 d5
and occupies a diagonal which may
soon be lengthened by White's e4-e5 .
5 . . .t'Llc6 ! ?, a s recommended by Rybka 3 ,
is also possible - after all, the knight
retreats sooner or later, and c6 is not a
bad square.
6 t'Llf3
V.Burmakin-J.Ulko, Moscow 1995,
continued 6 ..if4 t'Llg6 ? ! 7 ..ig3 ..ib7 8
h4 h 5 ? ! , and after the simple 9 t'Llf3
White would h ave been much better.
Instead, 6 ... t'Llf7 7 t'Llf3 ..ib4! 8 ..id3
..ixc3 9 bxc3 ..ib7 10 o-o t'Llf6 11 l::t e 1
o-o 12 e 5 t'Lld5 13 ..id2 c5 14 'iVe2 c4! 15
..ixc4 llc8 16 i..b 3 'i!Vc7 is only a bit bet­
ter for White.
6 ... t'Llf71
lt is a bad idea to activate White's
queen with 6 ...t'Llxf3+, as tried by
B.Savchenko.
7 ..ic4?1
Instead:
a) 7 ..if4 transposes to 6 ..if4 t'Llf7 7
t'Llf3 above.
b) 7 ..id3 is stronger - at this stage
White is more likely to play e5 than
Black is. After 7 ... ..ib7 8 o-o t'Llf6 9 'ii'e 2
45
Th e Dark Kn ight System
..ib4 10 l:te1 lt:\g4! 1 1 h3 lt:\ge 5 12 lt:\xe5
lt:\xe5 13 'iih 5+ lt:\f7 14 ..if4 i.. x c3 1 5
bxc3 g 6 16 'ii'g 4 'ir'e7 White h as a small
advantage, though he has long-term
concerns about his pawn structure.
Black still has a useful choice as to
where he should castle. 17 ... 'it'cs is
usually a good move.
7 ... ..ib7 8 iie2 a6
This useful little move prevents lt:\b5
and ..ia6, while preparing ... b6-b5-b4.
9 ..ib3 i.. b4
Here 9 ... lt:\f6 ? ! was played in C.Crouch­
A.Karpatchev, Cappelle la Grande,
1993, starting complications which
objectively favour White. The straight-
46
forward 9 ... ..ib4 can also lead to com­
plication s - e.g. 10 o-o lt:\f6 11 l:i.d1 fiie 7
12 e5 ..ixc3 13 bxc3 lt:\e4 14 .l:!.d4 lt:\xc3
15 'ifd3 lt:\d5 - but here Bl ack is fine.
8: 3 f4 lLlg6
4 e4
Other moves:
a) 4 h4 is an interesting attempt to
take advantage of Black's inflexible
knight. In the expert section of the
1985 New York Open, I fell for White's
trap and played 4... e5? 5 h S ! lt:\xf4 ! ? 6
e3, though I went on to win an ugly
miniature with 6 ... 'it'g 5 ! 7 iif3 lt:\xds 8
'ii'x d5 iig 3+ 9 'iii>d 1 d6 10 ..ibS+ 'it>d8 11
'ii'xf7 lt:\f6 12 lt:\f3 ?? \i'xg 2 13 :f1 ..ig4
14 ..ie2 'ii'xf1+! 15 i.. xf1 ..ixf3+ 16 ..ie2
..tds (trapping White's queen) o-1. Hi­
larious!
Instead, 4 ... e6! 5 h 5 lt:\6e7 6 c4! ? (6
dxe6 fxe6 7 e4 d5 transposes to 6 h4 d5
7 h5 lt:\6e7 in the notes to Position
Three below) 6 ... lt:\f6 7 lt:\c3 lt:\f5 8 'ii'd 3
(if 8 dxe6 fxe6 9 g4, then 9 ... lt:\xg4! ? 10
e4 lt:\fe3 or 10 ... lt:\fh 6 is possible, but
relatively simplest is 9 ...lt:\g 3 10 l:th 3

tbxf1 11 g 5 tbxh 5 12 .l:txh 5 tbg 3 13 .:i.h 3
tbf5 14 e4 tbd6 1 5 tbf3 tbxc4 16 tbh4
l:i.g 8 17 'ifh 5+ g 6 18 'i!Vxh 7 .U.g 7 19 'ifh6
l:i.g 8 with a draw) 8 ... i.c5 ! 9 e4 tbg4 is
fine for Black, according to Mr. H,
though there are some crazy variations
to con sider:
a1) 10 exf5 tbf2 11 'ifg 3 tbxh 1 12
'ii'xg7 �4+ 13 'it>d2 (or 13 'it>d1 'ili'g4+!
14 'i¥xg4 tbf2+) 13 ... l:i.f8 14 tbge2 tbf2
15 lbb5 i.b6 16 d6 Vi'xh 5 17 fxe6 tbe4+
18 'it>c2 tbxd6 is in Black's favour.
a2) 10 dxe6 tbf2 11 'i!Vd5 (not 11
exf7+? 'iii>f8 12 'ii'd 5 tbxh 1) 1 1...d6 12
·ifxf5 i.xe6 13 't!Vxc5 dxc5 14 Wxf2
.ii. xc4 with an unbalanced but roughly
equal position.
a3) 10 tbd1 tbfh6 11 dxe6 f5! 12 i.e2
'ife7! with excellent play for the pawn.
b) 4 f5 overextends: 4 ...tbe5 5 i.f4
d6 6 e4 g 6 ! (in practice, Black h as
played 6 ... tbf6 ? ! or 6 ... e6? but it is best
to challenge White's space immedi­
ately) 7 tbf3 i.g 7 8 i.b5+ i.d7 9 Si.xe5
i.xe5 10 i.xd7+ Vi'xd7 1 1 tbxe5 dxe5 12
o-o tbf6 with equality; e.g. 13 'ifd3 c6
14 c4 cxd5 1 5 cxd5 .:.c8 16 tbc3 o-o.
1 d4 tbc6 2 d5
c) 4 lDf3 is legal and was in fact the
move order for Onischuk-Shkuro men­
tioned below (see note 'e' to Position
Three). 4 ... e6 5 dxe6 fxe6 6 e4 trans­
poses to the main line, while 5 c4 i.c5
will not bring White any happiness.
4 es
...
4... e6 may tran spose after 5 fxe6, or
it may turn into a kind of mirrored
Alekhine, which usually works badly for
White because of the weak a7-g 1 di­
agonal; e.g. 5 c4? ! exd5 6 cxd5 i.c5 or 5
tbf3 exd5 6 exd5 ? ! i.c5 . However, 5
tbc3 ! exd5 6 1\Vxd 5 ! tbf6 7 'ifd3 i.c5 8
i.e3 Vi'e7 9 o-o-o o-o 10 tbf3 is better
for White (though Black is still okay).
5 dxe6!
White's only good move.
a) 5 f5 ? could lead to the position
mentioned in the introduction to this
book after 5 ...'ifh4+ 6 'it>d2 1Vxe4 7 fxg 6
'iVxd5+ 8 'it>e1 'ii'x d1+ 9 'it>xd1 hxg6, as­
sessed as unclear by Bogoljubow. The
endgame is favourable to Black, but
that is a moot point because 6 ...tbf6 ! is
even stronger - as one of my students,
Matthew Shih, was kind enough to
47
The Dark Kn ight System
point out to me last year.
a1) 7 fxg6 ? tt:Jxe4+ 8 �e2 'i¥f2+ 9
�d3 tt:Jcs+ 10 �c3 lt:Ja4+ 11 �b3 'iib 6+
12 �c4 'ii'a 6+! 13 �b3 tt:Jcs+ 14 'iiic 3
lt:Je4+ 1 S �b3 'i¥b6+ 16 �c4 'i¥h4+ 17
�d3 lt:Jf2+ wins White's queen.
a2) 7 lt:Jc3 �b4 8 fxg 6 lt:Jxe4+ 9 �e2
'i¥f2+ 10 �d3 fs (10 ... �xc3 11 'iitx e4!
�as 12 c3 hxg6 13 �d3 d6 14 �c4 i.fs
1 S 'i¥e2 'ii'h 6 may be better, but it's far
more complicated, as Black will still be
down material for some time) 11 tt:Jxe4
�d4+ 12 �e2 'ili'xe4+ 13 i.e3 f4 14 'i¥d3
'ili'xe3+ also leaves Black much better.
a3) 7 'i¥f3 lt:Jxe4+ 8 �e2 lt:Jf4+ 9
i.xf4 'i¥xf4 10 Vxf4 exf4 11 �f3 tt:Jf6 12
lt:Jc3 �b4 13 tt:Jge2 (13 .l:te1+ 'iiif8 14
i.c4 b S ! 1 S i.xbs �b7) 13 ... d6 14 �xf4
h S ! is relatively best, but still very good
for Black.
b) S lt:Jf3 ? is apparently tempting (it
h as been played several times), but af­
ter S ... exf4 6 lt:Jc3 �CS 7 i.d3 'ife7 ! 8
'i¥e2 d6 9 lt:Ja4 (9 �d2? ! a6 ! is even
worse) 9 ... �b6 10 �d2 tt:Jf6 White has
far too little for the pawn.
c) S 'ii'f3 exf4 6 lt:Jc3 (6 i.xf4 tt:Jxf4 7
48
'i¥xf4 "ii'f6 8 "ir'xf6 tt:Jxf6 leaves Black
with a superior pawn structure and
White with an acute shortage of dark­
squared bishops) 6 ... i.cs 7 i.xf4 tt:Jxf4 8
'ifxf4 �f6 9 'iifg 3 d6 10 i.bs+ �f8 1 1
lt:Jf3 'ili'g6 and Black can look forward to
a long and pleasant endgame.
d) s lt:Je2 (or s lt:Jh 3) s ... exf4 6 tt:Jxf4
�d6! already puts the enemy kingside
under pressure: 7 tt:Jxg6?! hxg6 8 'iff3
�4+ 9 �d1 lt:Jf6 10 i.d3 i.es is obvi­
ously not satisfactory for White, but 7
'iif3 tt:Jf6 8 lt:Jc3 0-0 9 i.d3 leaves Black
with several good ideas, the simplest
being 9 ..�es (9 ... c6! ? 9 ...tt:Jes ! ?) 10 o-o
.
d6 11 h3 cs, when Black's activity and
strong es-point give him the advantage.
s fxe6
...
The endgame after s ... dxe6 6 'i¥xd8+
'iitx d8 is playable in theory, but with
Black's slightly misplaced g6-knight, it
is less appealing than the similar end­
ing in line A (without 3 f4 tt:Jg 6). In
practice, Black's results are quite poor.
Instead, S .. .fxe6 brings us to:
Position Three

Not to put too fine a point on it, but
the players who have reached this posi­
tion with the black pieces need collec­
tively to have their heads examined.
White's main asset is his powerful
pawn duo on e4 and f4. Black's main
asset is his central pawn majority. One
need notice only one of these two
things in order to come up with the
correct plan (or at least the correct
sixth move) for Black. In fact, I'm not
even going to insult my readers by say­
ing it out loud, so if you still don't
know, see Wood-Penrose below and
then read Pawn Power in Chess by Hans
Kmoch.
Instead of taking the opportunity to
strike in the centre, Black h as generally
been seduced by 6 ... �c5, occupying the
a7-g 1 diagonal, presumably to stop
White from castling. N aturally this is
less important than the central battle
and, what's worse, it doesn't even
work. If White finds the bishop trouble­
some, he can trade it off with a timely
tt:Ja4 or 'ii'e 2 and �e3 . To add insult to
injury, White is usually better off cas­
tled long anyway. Sometimes 7 ... �c5 is
a good idea (or 7 ...�b4+ 8 c3 �c5), but
if there is a knight on c3, it is much bet­
ter to put pressure on White's centre
with 7 ... �b4, which prepares ...li:Jf6-e4.
Black's e4-knight can be a very an­
noying piece. As we see in the following
analysis, White's light squares are usu­
ally too weak (because he h as h ad to
play g 2-g3) to allow him to eliminate
the knight comfortably with .i.d3 and
1 d4 li:Jc6 2 ds
�xe4, even if this wins a pawn .
Let's get to the analysis.
6 li:Jf3
Other moves:
a) 6 �d3 is less popular and less
successful. Then Black has done fine
with 6 ...�c5, but 6 ... d5! is more accu­
rate, as in B.Wood-J. Penrose, Southend
1957(!),
which continued 7 e5 li:Jh 6 8 .i.e3 ? ! ,
and n o w not the game's 8 . . .li:Jf5 ? ! 9
�xf5 exf5 10 li:Jf3 .i.e7? ! (1o ... c 5 ! ) 11 c4
which is good for White, but instead
8 ... li:Jh4! 9 'ii'e 2 c 5 ! and Black is better.
No better is 7 li:Jc3 �b4 8 �d2 li:Jh6
9 li:Jf3 o-o 10 g3, when Black gently
plays 10 ... e5 ! ! and lets White try to
work out the details with his king in
the centre. The correct solution is 11
li:Jxd5 �xd2+ 12 'ii'x d2 c6 13 li:Je3 exf4
14 �c4+ 'it>h8 1 5 'ii'x d8 :xd8 16 gxf4
li:Jxf4 17 :g1 b5 with equality. Accord­
ing to Houdini, 11 f5 dxe4 12 li:Jxe4
.i.xd2 13 'ii'x d2 �xf5 14 o-o-o 'ilie7 is
also equal, but to me it looks a lot like
White is down a pawn for nothing. For­
tunately, it's not really our problem.
49
Th e Dark Kn ight System
b) 6 i.e3 is seldom played - al­
though it prevents ... i.cs, White's im­
portant dark-squared bishop i s vulner­
able:
6 ... ds (of course) 7 g 3 (7 lt:Jc3 i.b4 8
�d3 i.xc3 9 'ii'x c3 lt:Jf6 10 0-0-0 0-0 11
exds exds 12 lL'lf3 i.g4 is equal) 7 ... cs! 8
lt:Jd2 'ii'b 6 9 l:tb1 (sad, but there's really
nothing better: 9 fs exfs 10 exds lt:Jf6
11 'ii'e 2 i.e7 12 lt:Jc4 �a6 13 d6 b S ! or
13 i.xcs o-o 14 d6 i.d8 1 5 d7 i.xd7 16
i.xf8 lt:Jxf8 gives Black m ore than
enough for a small exchange; while 9
i.d3 i.e7 10 lt:Jgf3 lt:Jh6 11 'ii'e 2 lt:Jg4 12
exds exds 13 i.g 1 o-o 14 o-o-o i.d6 15
h 3 lt:Jf6 is fine for Black) 9 ...i.d7 10
lt:Jgf3 lt:Jf6 11 es lt:Jg4 12 i.g 1 "ikc7 13
c4! d4 14 i.d3 i.c6 1 5 h 3 lt:Jh 6 16 i.f2
i.e7 17 0-0 0-0 18 b4 b6 19 lt:Je4.
White's play makes an excellent im­
pression, but here Black has 19 ... tt:Jxf4!
20 gxf4 l:txf4 2 1 �e2 tt:Jfs 22 .:tb3 .l:r.f8,
which reverses the initiative at the cost
of a small material investment. Hou­
dini calls it equal, but if I h ad the
choice, I'd sit behind the black pieces.
c) 6 g3 dS 7 lL'lf3 transposes to 6 lt:Jf3
50
dS 7 g 3 below.
d) 6 h4! ? has never been played, but
it's a venomous move. The tactical jus­
tification is 6 ... lt:Jxh4? 7 'ii'g 4! i.e7
(7 ... lt:Jg 6 8 .l:i.xh 7) 8 'ii'xg 7 with a large
advantage for White. Correct is the
anti-shocker 6 ... ds even though 7 h s
lt:J6e7 blocks the f8-bishop. After 8 lL'lf3
lt:Jc6 9 l2Jc3 i.b4 10 i.d2 lt:Jf6 11 es
i.xc3 12 i.xc3 lt:Je4 13 i.d3 lt:Jxc3 14
bxc3 Black gets out of Dodge with
14 ... 'ife7 15 tt:Jg s i.d7 16 i.xh 7 o-o-o,
when White has space and a pawn, but
is badly overextended with nowhere
for his king. I will utter a naughty word:
unclear.
e) The rare 6 lt:Jc3 ! ? is logical, fight­
ing for the dS-square, but Black forces
... d7-d5 anyway with 6 ...i.b4! 7 lt:Je2
(other moves, such as 7 lL'lf3 and 7 i.d3,
transpose elsewhere) 7 ... ds 8 'i¥d3 c6 9
i.d2 lt:Jf6 (after 9 ... dxe4 10 "ikxd8 'ii;x d8
11 a3 i.as White is a little better in the
endgame) 10 es lt:Jg4 11 h 3 lt:Jh 6 12
o-o-o o-o 13 g4 bs 14 wb1 i.cs 1 5 i.g 2
as and although White h ad a head
start in the race, it is difficult for him to
advance further. Then 16 lt:Jd4 ! ? 'ifb6
17 tt:Jxc6 i.b7 18 lt:Jxds exds 19 lt:Jxas
lt:Jxf4 20 i.xf4 'ii'x as 21 i.xdS+ r;i;h8 22
i.xh 6 gxh 6 is certainly complicated,
but not unfavourable to Black.
The text move, 6 lL'lf3, is by far the
most common - apparently with good
reason since White h as won the last
five g ames in a row from this position,
most notably A.Onischuk-I.Shkuro,
Ukrainian Team Championship 2009,

which is a perfect example of what
Black must avoid (see Game 28). Clearly
we need some new and improved
ideas.
6 ds!
...
This move is part of my original
an alysis that dates back to 1986, and
in spired the above variations. In
twenty-five opportunities, Black has
somehow failed to try this, so we ana­
lyse in a vacuum. For most players, the
following lines need hardly be memo­
rized, but offer an excellent opportu­
nity to get acquainted with the wide
variety of plans for both sides.
7 tLlc3
a) The first thing I realized 26 years
ago was that 7 f5 ? ! is not a problem :
7 ... dxe4 8 'ir'xd8+ 'ititxd8 9 tZ::l g 5 exf5 10
tLlf7+ 'itite8 1 1 tZ::l xh 8 tZ::lxh8 12 tLlc3 c6 13
.ic4 (13 g4! ?) 13 ... tZ::lf6 14 .ie3 .id6 is a
bit better for Black. As it turns out,
7 ... exf5 is also okay: 8 exd5 tZ::lf6 9 tLlc3
.ib4 10 'ii'e 2+ 'ifitf7 ! 11 tLlg 5 + 'ititg8 12
.id2 ..txc3 13 ..txc3 tZ::l x d5 14 .id2 h6 1 5
tLlf3 'itith7 16 0-0-0 .Ue8 and i t is White
who has the hard job of proving full
1 d4 tLlc6 2 dS
compen sation.
Notice that Black isn't actually
threatening 7 ... dxe4, so White has an
array of options:
b) 7 e5 may not be best, but it is cer­
tainly critical.
White plan s tLlg 5 and tZ::l xh 7, an ar­
gument he will try to enhance with h4-
h 5, .id3, or 'ifh 5 . This plan is indeed
dangerous, especially if Black is castled
on that side of the board. However,
Black does not skip his turns, and as
long as White is pursuing this plan he
is neither developing quickly nor at­
tending to his own king . For instance,
7 ... tLlh 6 ! 8 g3 (or 8 .ie3 tZ::l g 4 9 'ir'd2
tZ::lx e3 10 'it'xe3 'ir'e7 ! 11 tZ::lc 3 'ii'h4) 8 ... c5
9 tLlg 5 .ie7 10 tZ::l xh 7 (after 10 .ib5+
'ifitf8 White has to worry about both
11 ... c4 and 1 1 ... tZ::l x e5) 10 ...tLlf5 11 'ii'h 5
'ifitf7, when Black h as good compensa­
tion after White's failed attack; e.g. 12
tLlc3 'ititg8 13 tZ::lf6+ ..txf6 14 'i¥xg6 .ie7
15 .ig 2 c4 16 o-o l::th 6 17 'ii'g 4 'ii'h 6 + 18
'itith 1 .id7 and White is having trouble
with development, the centre, and his
king, which doesn't leave much to be
51
The Dark Kn ight System
h appy about. There is also 8 h 4 ! ? ..tcs 9
h S (or 9 ..td3 l2Jg4 when, according to
Houdini, White h as nothing better than
10 l2Jd4 l2Jh 6 11 l2Jf3, with a repetition)
9 ... lbe7 10 lLlg s lLlhfs 11 'ii'd 3 h 6 12 g4
l2Jd4 13 c3 l2Jdc6 14 l2Jf3 as ! with equal
chances. Black intends ...b7-b6, ... .i.a6
( ... ..tb7), ... �d7 (or ... dS-d4, .. .'�ds), and
will have the choice of which side to
castle. White may enjoy his space, but
may also find himself overextended.
Hold on : what's so great about
7 ... l2Jh 6 - ?
I used to h ave problems in this posi­
tion in blitz games because I didn't
know the proper arrangement for the
pieces - especially whether to play
7 ... -tcs or 7 ... cs, and also whether to
play ... l2Jh 6 or ...l2J8e7. As it turns out, it
is not yet clear whether it is the bishop
or the pawn that belongs on cs, so it
makes sense to wait on that decision,
but the knight is just about always best
on the h 6-square. All of ...l2Jg4, ...lLlfs,
and ... lLlf7 are useful options from
there, and it can hold up White's g- and
f-pawns. Just as important, developing
52
the knight to h 6 avoids a traffic jam on
the e7-square, which may be needed
for Black's other knight, not to mention
the bishop and queen . (You may now
resume your normal programming.)
c) The immediate 7 h4! ? is also pos­
sible, but Black is already fine after
7 ... .i.b4+! 8 c3 (not 8 .i.d2 ? dxe4! )
8 . . ..i. c s 9 h s l2J6e7; e . g . 10 b 4 .i.b6 11
c4! ? dxc4 12 'it'xd8+ 'it>xd8 13 .i.b2 (not
13 ttJg s ? ..ltd4!) 13 ... exf3 14 ..txg 7 lLlfs
15 ..txh8 f2+ 16 Wd2 l2Jg 3 17 cs l2Jxh 1
18 cxb6 axb6 19 lLlc3 lbe7, which is cer­
tainly no worse for Black.
d) 7 .i.d3 allows us to surprise White
with 7 ... dxe4! 8 ..txe4 'i!t'xd1+ 9 'it>xd1
l2Jf6 10 ..txg6 hxg6.
Although our pawns are vile, our
bishop pair and overall activity are
quite enough, particularly since the
enemy king is a bit loose; e.g. 11 l2Jbd2
..td6 12 l2Jc4 bs 13 lLlces .i.b7 14 'it>e2
..txes 15 fxes l2Jd7 and Black is more
comfortable because of White's bad
bishop; or 11 .i.e3 ? ! ..id6 12 l2Jc3 l2Jg4
13 We2 0-0! 14 g 3 ? ! (14 ..tc1 !) 14 ...b6!
and suddenly White is in big trouble; or

11 tLic3 i.. d 7 12 tLie5 (12 �e2 i..d 6! 13
tLie5 i.. xe5 14 fxe5 tLig4 nets a pawn)
12 ... 0-0-0! ! 13 tLif7 i..c 6+ 14 ttJxd8 �xd8
15 h4 i..x g 2 16 l:i.h 2 i..f3+ 17 �el i.. c 5
and how is White going to untangle
himself without shedding any mate­
rial ?
e) 7 c4! ? does force 7 ... dxe4, but
weaken s White's position as well : 8
"O'xd8 �xd8 9 ttJg 5 'ilo>e8 10 ttJxe4 tLif6
11 i.. d 3 b6 12 ttJbc3 ..ib7 13 o-o l:td8
and Black is comfortably equal.
f) 7 g 3
7 . . ...ic5 ( 7. . ...ib4+ 8 c3 ..i c 5 9 'ii'e 2
·iie7 10 tiJbd2 tLih6 1 1 tLig 5 o-o 12 h4
.:e8 is okay, too) 8 li'e2 (8 tiJc3 tiJf6 ! 9
e5 tLig4 10 tLid4 0-0 11 'ii'x g4 ..ixd4 12
�d2 c5! 13 o-o-o li'b6 with equal
chances) 8 ...tiJf6 9 e5 tLie4 10 ..ie3 ..ixe3
11 'ifxe3 tLie7 ! 12 i..d 3 i.. d 7 13 ttJbd2
�xd2 14 ttJxd2 (14 'ii'x d2 c5) 14 ... 0-0 1 5
o-o tLif5 and with 16 . . .'ii'e 7 (or 16 . . .b6)
and 17 ... c5 coming, Black h as sufficient
counterplay.
g) 7 exd5?! exd5 8 i..d 3 makes no
sense - it surrenders the centre and ac­
tivates Black's problem piece, the c8-
1 d4 tL\c6 2 dS
bishop. Unsurprisingly Black has many
ways to play: 8 ... ..ig4 9 o-o i.c5+ 10 �h l
tiJ8e7 11 h 3 i.xf3 12 'i!Vxf3 o-o seems
simplest, or Black can enter an equal
endgame with 8 .....ib4+ and 9 .. .'iVe7+.
7 ..ib4
...
8 g3
White usually finds it necessary to
play this sooner or later. Otherwise:
a) 8 i.d3 i.. x c3 9 bxc3 dxe4 10 i.. xe4
li'xdl+ 11 �xdl tiJf6 12 i.. d 3 o-o 13
.l:tel tL'lg4 14 i.. x g6 hxg6 15 h 3 tiJf6 16
tLie5 tLih 5 17 ttJxg6 l:!.f6 18 tLie7+ �f7 19
ttJxc8 ttJxf4! with equal chances.
b) 8 e5 tiJ8e7 (8 ... c5 isn't bad either)
9 g3 0-0 10 ..td3 c5 and Black's good
centre and rapid deployment ensure
that he will not be rolled up on the
kingside and that his chances are not
worse.
c) 8 'ii'd 3 ! ? i.. xc3+! 9 'ii'x c3 tiJf6 10 e5
tLie4 1 1 �a3 'ii'e 7 12 �xe7 ttJxe7 13
i..d 3 b6 14 i..x e4 dxe4 1 5 tLig 5 h6 16
ttJxe4 i.b7 17 tL'lc3 tLif5 18 o-o o-o-o and
Black h as the d-file, while White's cl­
bishop is a huge problem, a situation
which fully compensates for the pawn.
53
The Dark Kn ight System
8 lt:Jf6! 9 i.d2 0-0 10 es .txc3 11 i.xc3
...
lt:Je4 12 i.d3 b6!
White's c3-bishop is a silly piece - at
the moment Black is far better off
keeping the e4-knight and blocking the
other bishop.
a) 13 i.xe4 dxe4 14 tt:Jg s is pointless
54
because of 14 ... h6! 1 5 lt:Jxe4?! .tb7 16
"ife2 'ili'ds 17 lt:Jf2 tt:Jxf4! 18 gxf4 .l:r.xf4 19
l:tf1 (not 19 �d1?! .l:.xf2 ! ) 19 .. J:taf8 20
.td2 (still not 20 .l:td1? ! "ifxa2 ! 21 l:.d7
'i¥a4 22 l:txc7 i.g 2 and White is toast)
2o ... .l:.f3 21 a4 l:.8fS 22 o-o-o .l:txes 2 3
'i¥xf3 'ii'xf3 24 lt:Jd3 'i!Vds 2 5 tt:Jxes 'il'xes
26 .tc3 'ii'e 3+ 27 ct>b1 .tds 28 h4 gs
with some advantage to Black.
b) 13 'ii'e 2 i.b7 14 o-o-o 'ii'e 8 1 5
i.xe4 dxe4 16 tt:Jg s e3 17 l:.he1 h 6 1 8
lt:Jf3 lt:Je7 19 "ifxe3 tt:Jds with enough
play for the pawn; e.g. 20 'i¥d3 as 2 1
lt:Jd2 lt:Jb4! o r 2 1 lt:Jd4 a4 22 a 3 cS 2 3
tt:J e 2 bs.
c) 13 o-o lt:Je7 14 'i¥e2 i.b7
(14 ... tt:Jcs ? 15 .txh 7+ is too strong) 15
.l:.ad1 'ii'e 8 is equal.
 Section Two
1 e4 tt:Jc6
One might argue that it m akes
more sense to play the Dark Knight
ag ain st 1 e4 than 1 d4, because the
dangerous m ove 2 dS is n o l on g er
available. On the other h an d, White, as
a 1 e4 pl ayer, is more likely to be fa­
miliar with the possible tran spositions
to the Scotch and the Pirc. Personally, I
find the second argument to be m ore
compelling th an the first (for a decade
I played 1 d4 Nc6 with no plan for de­
fendin g the Scotch), but th e reader
can m ake up his own mind, or better
yet just pl ay 1...l2Jc6 ag ain st every­
thing !
2 d4 (including The Scotch) -
Chapter Four
2 lLlf3 (the Dark Knight Pirc) -
Chapter Five
2 tLlc3 - Chapter Six
Other moves:
a) 2 ..ic4 commits the bishop way
too early: 2 ... tLlf6 3 lLlc3 e6! 4 d3 ..ib4 5
.id2 d5 6 exd5 exd5 7 ..ib5 o-o 8 tLlge2
tLle 5 ! 9 a3 ..id6 10 ..ig 5 c6! was P.Cruz­
A.Kogan, Lisbon 2000. Black has a big
advantage, largely due to h aving shut
White's light-squared bishop out of
play.
b) 2 f4 is not as bad as it looks.
2 ... d5 ! 3 exd5 'ii'x d5 4 tLlc3 leads to a
bizarre kind of Scandin avian. Fortu­
nately, 4 .. .'ii' e 6+! is more awkward for
White than it is for Black; e.g. 5 ..ie2
tLld4! or 5 'ife2 tLlb4! or 5 tLlge2 tLld4! 6
d3 tLlf6 7 l2Je4 tLlf5 8 c3 g 6 (8 .. tLlxe4 9
.
'ii'a4+) with some advantage to Black in
all cases, according to Houdini.
3 e5 is a good Nimzowitsch Defence
55
Th e Dark Kn ig h t System
for Black since White's pawn does not
belong on f4 {at least until .. .f7-f6 is
played). In R.Barkman-S.Lejlic, Karls­
krona 1997, Black used this to good
effect: 3 ... tt:lh6 4 d4 .ltg4 5 .lte2 .ltxe2 6
tt:lxe2 tt:lf5 7 c3 e6 8 'ii'd 3 h 5 {8 .. .'ifh4+!)
9 tt:ld2? ! �4+ with advantage due to
White's light-square problem s - White
decided to pitch a pawn with 10 tt:lg 3
'ii'xf4 11 tt:lxf5 'ii'xf5 12 'ii'xf5 exf5, but
he had no compensation .
Alternatively, Black can throw a clog
in White's machinery with 3 ... g 5 ! ? 4 d4
gxf4 5 .ltxf4 .ltf5 6 c3 'i!Vd7 7 tt:lf3 h5 8
ie2 0-0-0 9 tt:lh4 ig4 10 tt:ld2 ih6 as
in H.Gohlke-S.Wiezer, Gorlitz 1985,
56
with a small advantage according to
New In Chess. Actually, I don't believe
Black is better, but the position is in­
teresting and Black has the makings of
a kingside attack. Wisnewski's recom­
mendation is 3 ... d4! ? - one idea is to
clear the d5-square for Bla�k's queen,
but I am n ot keen on relinquishing con­
trol over e4.
For those who are interested, 2 ... e5
3 tt:lf3 f5 ! ? is the Adelaide Counter­
Gambit, a wild and fascinating defence
to the King's Gambit, but truly outside
the scope of this book. Besides, there is
also 3 tt:lc3, with wildness that White is
presumably prepared for.
 C h a pte r Fo u r
1 e 4 'Dc6 2 d4 es
Now White has:
A: 3 dS 57
B: 3 dxes 73
C: 3 lLlf3 79 - The Scotch
On 3 c3, Black can transpose to a
Ponziani with 3 ... ll:Jf6 4 ll:Jf3; but 3 ... d5 4
dxe5 dxe4 5 'ii'x d8 ll:Jxd8 is simpler and
equal, while for a more interesting
endgame Black can play 4 ... i.e6 ! ? 5
exd5 ii'xd5 6 'ii'x d5 i.xd5 with full
compensation; e.g. 7 f4 f6 8 exf6 ll:Jxf6
9 C4 i.e6 10 lLlf3 0-0cO 1 1 ll:Jc3 lbb4
with a large advantage to Black.
A: 3 d s ll:Jce7
White normally chooses between:
Al: 4 ll:Jf3 58
A2: 4 C4?! 62
Others:
a) 4 f4? is a fairly common mistake at
amateur level. The apparently hard-to­
spot 4 ... ll:Jg6! transposes to 1 d4 li:lc6 2
d5 ll:Je5 3 f4 li:lg6 4 e4 e5 with the differ­
ence that White's only good move, 5
dxe6 en passant, is not permitted by the
rules. How unfortunate! White does not
have my sympathy though, since
57
Th e Dark Kn ig h t System
4... exf4?! 5 i..xf4 lt:lg6 is also lousy.
b) Miles h ad a wonderful answer to
4 .i.e3 in 4 .. .f5 ! ?, with which he drew
with Beliavsky and beat Campora (see
Games 29 and 30). The point is to take
advantage of the newly moved bishop,
which is vulnerable to both ... f5-f4 and
5 exf5 lt:lxf5. If Black does not play 4 .. .f5,
White may be able to stop the f8-dark
bishop from developing comfortably,
e.g. 4 ... lt:lg6 5 a3 ! ?, though 5 ... b6 is sat­
isfactory for Black.
c) 4 lt:lc3 lt:lg 6 5 Si.e3 lt:lf6 (5 ... .tb4!?)
6 a3 is very lightly tested. Then 6 ...b6 7
lt:lf3 .tc5 is fine, when 8 .txc5 bxc5 9
d6?! o-o and 10 ... .tb7 is at least equal.
d) 4 d6 has been played a few times
and Black usually just takes, which is fine
- but I prefer 4...lt:lg6, accelerating devel­
opment after 5 dxc7 'i!i'xc7; e.g. 6 lt:lc3
li.b4 7 lt:le2 (after 7 Si.d2?! .txc3 ! 8 i..xc3
lt:lf6 9 Si.d3 o-o 10 lt:le2 .l:.d8 and 11 ... d5,
Black is better) 7 ...lt:lf6 8 a3 .tc5 9 b4 .tb6
and Black is fine, since 10 lt:lb5?! runs
into 10...1i.xf2+ 11 �xf2 �6+.
A1: 4 lt:lf3 lt:lg6
58
4 ... d6 is recommended by De Fir­
mian in MC014 as a way to keep the
pl ay obscure. In my opinion, the reverse
is true: 4 ... d6 will tran spose to Pirc or
King's lndian -type positions, whereas
4 ... lt:lg6 maintains a unique "Kevitzian"
flavour. Furthermore, 4... d6? ! 5 c4! lt:lf6
6 lt:lc3 g 6 runs into 7 c 5 ! which is in­
deed a certain kind of King's Indian -
the bad kind.
After the text White h as two impor­
tant moves:
All: 5 h4! 59
A12: 5 .ie3 61
After other moves, Black is already a
statistical favourite, reaching positions
similar to line A2.
a) 5 lt:lc3 lt:lf6 6 Si.g 5 ? ! (this is wrong,
as usual; it was not too late for 6 h4!
h 5, though this is rarely played here)
6 ... h 6 ! (Black can wait, but putting the
question immediately places the most
pressure on White to follow through
with this ill-conceived exchange) 7
.txf6 't!Vxf6 8 g 3 .tc5 9 Si.h 3 a6 and
Black is already a bit better due to his
powerful dark-squared bishop - and if
White doesn't prepare the exchange of
light-squared bishops Black has a com­
fortable advantage.
b) 5 Si.d3 .tc5 6 o-o lt:lf6 7 lt:lc3 c6?! 8
lt:la4 li.e7 was Y. Dembo-R.Goldin, Petah
Tiqwa 1996, which Black went on to
win - even so he should have preserved
the bishop on the diagonal with 7 ... a6
or 7 ... a5.

A11: 5 h4!
This is the way to start posing prob­
lems for Black and his knight on g6.
s hs
...
Black is forced to weaken his king­
side or accept an extremely cramped
position. But let's not forget that White
has al so weakened his kingside.
6 i.gs
Alternatively:
a) 6 ttlc3 ttlf6 7 ttlg 5 ! ? (7 i.g 5 trans­
poses to the main line) 7 ... i.b4! 8 i.e2
d6 9 'iid 3 i.d7 10 a3, and instead of
10 ... i.xc3+ as in M.Kravtsiv-K.Tarlev,
Evpatoria 2007 (see Gam e 3 1}, I prefer
10 ... .1i.c5, retaining the bishop for now;
e.g. 11 g3 o-o 12 o-o ttlg4 with a tiny
edge to White. Black intends ...'iie 7 and
.. .f7-f6 to evict the knight, and probably
... a7-a6 and ...b7-b5.
Riskier is 7 ... .1i.c5 ! ? 8 d6 ! cxd6 9 .fi.c4
o-o 10 o-o b 5 ! 11 ttlxb5 ii.a6 12 a4 l::tc 8
13 ii.e2 d5 14 exd5 ttlxh4 1 5 ii.xh 5 e4 -
Black has solved the problem of his
weak pawns by sacrificing them all and
now has reasonable play for the pawn
minus, but White had many other tries.
1 e4 tbc6 2 d4 e5
b) 6 g3 .i.c5 7 i.g 5 f6 ! ? (7 ... ttlf6 8
.i.h 3 d6 9 ii.xc8 �xc8 10 i.xf6 gxf6 is
simpler - Black will play .. .f6-f5 at his
convenience, with near equality) 8 i.d2
d6 9 ttlc3 .i.d7 10 i.e2 ttl6e7 was a little
better for White in H .Meissner-A.Miles,
European Cup, Slough 1997, though
Miles m ethodically ground out the win
in an instructive ending (see Game 3 2).
c) 6 ii.e2 ttlf6 7 ttlc3 a6 8 .i.g 5 ii.c5 9
o-o d6 is already equal since Black h as
quickly established his bishop in the
ideal position . Black will normally fol­
low with 10 .. .'ii'd7, breaking the pin
and intending ... lt::l g 4 or .. .'ii' g 4.
6 ...ttlf6 7 ttlc3
7 ...ii.b4
7 ... i.c5 is a possible alternative.
Then 8 ttla4?! ..tb4+ 9 c3 ..te7 10 ..txf6
..txf6 11 d6 was A.lvanov-J. Benjamin,
US Championship, Parsippany 1996,
which lvanov won, but Black's inaccu­
racies are yet to come. Indeed, after
11 ... cxd6 12 g 3 ! ? d5 ? ! 13 'ili'xd5 d6? 14
..tb5+ 'iii>f8 White was much better, but
Benjamin could h ave played 12 ...b6 13
c4 (13 'i!Vxd6? ! ..tb7 14 'iid 3 ..tc6 15 b3
59
Th e Dark Kn ight System
d5 is great for Black) 13 ... i.b7 14 tt:'lc3
i.e7, followed by ... tt:'lf8-e6 - Black's
extra pawn may not be much of an as­
set, but he is not suffering for h aving it.
However, the correct respon se to
7 ... ..tc5, namely 8 tt:'ld2 !, puts Black un­
der some pressure.
8 tt:'ld2 !
Instead:
a) 8 a3 i.xc3+ 9 bxc3 c6 10 c4 d6 11
tt:'ld2 'i¥a5 was L.Christiansen-J . Ben­
jamin, U S Championship 2000 (see
Game 3 3). Black h ad nothing to com­
plain about and went on to win .
b ) 8 i.e2 allows u s t o preserve the
bishop with 8 ... d6 9 o-o i.d7 10 a3 i.c5
11 'ii'd 3 a6 12 g3 i.h 3 13 .:fd1 and
break the annoying pin by 13 ...'ii'c 8 (or
13 ...'ii'b 8! ?), when we're just about
equal. (This queen manoeuvre should
be kept in mind throughout this sec­
tion .) 14 ... tt:'lg4 is likely, targetin g the f2-
square.
The text move is more challenging.
8 d6!
...
In E.Morten sen-C. Hoi, Ostrava 1992,
Black played 8 ... c6 9 i.e2 (9 g 3 ! ?)
60
9 ... i.xc3 10 bxc3 cxd5 11 i.xh 5 (11
exd5 ! ?) 11...tt:'Jf4 12 i.f3 tt:'le6 13 i.xf6
'ifxf6 14 exd5 tt:'lc5 15 g 3 (15 h 5 ! ?)
1 5 ... d6 with only a tiny disadvantage,
but White missed some good chances
along the way.
After the text, we're in uncharted
territory again, which is just how we
like it. Right? ! Right.
a) 9 i.b5+ is certainly n ot a big deal
after 9 ... i.d7 10 i.xd7+ 'i¥xd7. Sure, we
liked that bishop, but White has solved
our space problem and helped us break
the annoying pin on our f6-knight.
Then 11 i.xf6 gxf6 12 'iif3 'i¥g4
(12 ... i.xc3 13 'iVxc3 f5 14 g3 o-o-o is
okay, too) 13 'ifxg4 (or 13 'ifxf6 i.xc3 14
bxc3 tt:'lf4! 15 tt:'lf3 'i¥xg 2 16 tt:'lg 5 .l:.f8 17
o-o-o 'ili'xf2 18 tt:'lh7 tt:'le2+ 19 'ifi>d2 'ii'xf6
20 tt:'lxf6+ 'ifi>e7 2 1 tt:'lxh 5 �h 8 22 'iti>xe2
.l:i.xh 5 23 l'!dg 1 with a draw, presuma­
bly, though only Black can pretend to
play for a win) 13 ... hxg4 14 g3 i.xc3 1 5
bxc3 tt:'le7 16 c 4 f5 is level.
b) 9 i.e2 tt:'lf4! (as in many similar
positions, White does not want to take
this knight, especially since his bishop

is strong on the g 5-square) 10 ii.b5+
�8 (this time 10 ...1i.d7 doesn't work
well - White will gain the very useful
g2-g3 with tempo) 11 0-0 ii.g4 12 f3 (or
12 �e1 Ji.d7 ! - the situation has
ch anged again already; the bishop ex­
change now brings equality: 13 ii.xd7
·o-xd7 and Black will follow with
14 .. .'�g4 if possible, otherwise 14 ... c6,
or similarly 13 Ji.d3 ! ? c6) 1 2 ... c6! 13
..i.a4 ii.xc3 14 bxc3 ..td7 15 dxc6 ii.xc6
16 Ji.xc6 �6+ 17 'it?h 1 �xc6 18 Ji.xf4
exf4 19 lbb3 d5 20 e5 lbd7 2 1 l:le1 .:e8
22 "ikd4 .:i.h 6 is roughly equal.
There are many other ways the
game could go, and White does h ave
chances for an advantage, but the pre­
ceding lines give a good indication of
Black's resources.
A12: s ii.e3
White doesn't actually play this very
often, but Black's attacking ideas in
this variation are too important and
widely applicable to relegate to a foot­
note. Study carefully. Don't worry, it's
fun and easy.
1 e4 lbc6 2 d4 es
s lDf6 6 1i.d3
..•
Not 6 lbbd2 c6! 7 c4? ! (this is n o
good, but the alternative is t o give up
the centre) 7 ... lbg4! 8 Ji.g 5 ? �6 9
�e2 ? ! ii.c5 and White needs C PR.
6 .lbg41 7 Ji.d2
..
Or 7 ii.g 5 1i.e7 8 1i.d2 1i.c5, tran spos­
ing.
1 ...ii.cs s o-o as 9 h 3 lDf6
White h as gotten h 2-h 3 for free, but
free is still way too expensive - he must
be extremely careful because of his
weakened kingside:
a) 10 iie1 d6 11 ii.xa5 ? ! ii.xh 3 ! 12
gxh3 'ii'd 7 13 �h 2 lDf4 14 lDg 1 lbg4+!
15 �h 1 (not 15 hxg4?? 'ii'x g4 and ...�g 2
mate) 1 5 ... tbxf2+ 16 .l::t xf2 1i.xf2 17 'ii'xf2
l:lxa5 18 'ii'f3 o-o 19 lbc3 f5 20 exf5
lbxd3 2 1 'ii'x d3 ltxf5 and Black is better
because of White's weak king and
awkward knights.
b) 10 a4 0-0 11 lba3 d6 12 �e1 (or
12 lbc4 c6! 13 dxc6 d5 ! 14 exd5?! e4! )
12 . . .ttJh 5 13 ii. e 3 lbgf4! 1 4 Ji.xc5 �f6 ! !
15 ..te3 ..txh 3 1 6 ..txf4 lbxf4 1 7 lDh 2
Ji.xg 2 18 �e3 Ji.xf1 and even though
White survives, Black maintains some
61
The Dark Kn ight System
material and position al pluses.
Note that in many similar position s
White voluntarily spends a tempo on
h 2-h 3 (?) - a move which contributes
greatly to his downfall. Note, too, that
Black's attack with ...lL:lh s, ... l2Jgf4, ... �f6
is so strong that he can sacrifice a full
piece to accelerate it. One free tempo
and the g ame would h ave been over.
A2: 4 C4?!
This weak move is almost as com­
mon as 4 lLlf3 . lt wastes a tempo in a
critical position, obstructs the fl­
bishop, and leaves a big hole on the d4-
square. In exchange, White h as forti­
fied dS, but the price is too high - he
has already squandered his first move
advantage, and if he plays at all indif­
ferently he will soon stand worse.
White imagines he will eventually be
playing c4-cS, but this is difficult to
achieve.
4 l2Jg6
...
4 ... l2Jf6 is similar, when S l2Jc3 l2Jg6
transposes to a position often reached
through the Tango. However, 6 h4! is a
62
dangerous move here, and Black h as
both objective and practical problems.
The text move order makes it difficult
for White to play an early h2-h4 and so
avoids this troublesome line.
White has:
A21: 5 l2Jc3 62
A22: 5 i.e3 64
A2 3 : 5 lLlf3 67
A24: 5 i.d 3 68
A2 5 : 5 g3 70
A26: 5 a 3 71
By the way, I hate all these nested
letters and numbers which make the
book look like the awful, tedious out­
lines we had to write in seventh grade
Social Studies. lt makes my eyes glaze
over just looking at it, and probably
yours, too, so just read about S l2Jc3
and S �e3 and then skip to line B. The
rest doesn't much happen and it's
pretty obvious anyway. Maybe some­
day, when you need to look something
up, you'll thank me that the material is
so well organized. I take cash and
cheques.
A21: 5 tLlc3 �cS!
If s ... l2Jf6 then 6 h4! and Black is
back on less solid ground.
6 lLlf3 l2Jf6 7 .i.e2
By now 7 h4? is not possible be­
cause it does not address the threat of
7 ...l2Jg4!.
For 7 �d3, see A24, n ote 'b' to White's
seventh move.

1 0-0
...
Black will be playing ... d7-d6 and
usually ... a7-aS soon, but this move or-
der is the most accurate since 7. . .d6? ! is
met by 8 b4 (8 ... il.xb4?? 9 �a4+).
7 ... as is sometimes played, but
White is actually quite far from being
able to play lt:Ja4 because the e4-pawn
would hang. Holding off on ... a7-aS
makes sense because:
1. Black may change his mind and
play ... a7-a6.
2. Sometimes Black can prevent
!Lla4 with ... il.d7 instead.
3. lt is possible to allow the bishop
to be traded off under some circum­
stances - I have recently noticed that
tempi sometimes matter in chess.
s o-o d6
8 ... as 9 lt:Je1 d6 (which could just as
easily have been 8 .. d6 9 lt:Je1 as) 10
. because it's often necessary in similar
li:Jd3 il.d4! was W.Weisser-L.Trumpp,
German League 2003, in which White
quickly reached a difficult position and
got abused tactically (see Gam e 34).
9 'if'c2
In B.Perrusset-I.Moullier, Paris 200S,
1 e4 ltJc6 2 d4 es
White tried 9 a3 aS 10 i. d2 'iie7 1 1
�b1 .i. d7 12 b4? ! axb4 13 axb4 i. d4!,
but she was worse after the queenside
had opened and it was Black who was
successful in the end (see Game 3 S).
9 ... il.d7
This stops lt:Ja4 for now.
Instead, D. Baramidze-E.Griezne,
Baunatal 1999, continued 9 ... a6 10 a3
"ike7 11 b4 il.a7 12 li:Jd1, when Black got
to demonstrate the typical manoeuvre
12 ... lt:Jh s 13 lt:Je3 lt:Jgf4 which he used
to good effect (see Game 3 6).
10 a3 a s
Black is slightly better.
If 11 �b1, we will certainly fix
White's queenside pawns with 11 ... a4.
If 11 b3, we have available the slow
plan 1 1 ...'if'e7, 12 ...h6, 13 ...lt:Jh s,
14 ...lt:Jgf4, 1S ...�f6, etc, which I provide
positions - but here we just blast out
11 ... lt:Jh s ! 12 li:Jxes (12 g 3 is safer, but it
does create a weakness, and Black is
better after the simple 12 ...lt:Jf6, menac­
ing ... il.h 3 and/or ... lt:Jg4) 12 ... lt:Jxes 13
il.xh s �4 14 il.e2 fs 1 S g 3 'ii'h 3 16
63
The Dark Kn ig h t System
i.f4 �ae8 (or 16 ... g s ! ? 17 i.xes dxes 18
exfs .l:!.a6 with more than enough for
the pawn) 17 b4 (not 17 i.xe s ? .l:!.xes 18
i.f3 f4 19 i.g 2 'ii'h s 20 b4 i.h 3 2 1 bxcs
i.xg 2, winning) 17 ... axb4 18 axb4
i.xb4 with clearly the better game for
Black.
A22: 5 ..te3 tt::lf6
6 l2Jc3
Other moves:
a) 6 i.d3 b6! (6 ... tt::lh 4! ? 7 rio>f1 b6 8
g 3 tt::l g 6 9 Wg 2 i.cs is similar, but White
is certainly not suffering from his in­
ability to castle) 7 tt::l c 3 i.. c s (in V.Vilkov­
A.Provotorov, Kalug a 1996, the only
time this position has been reached,
Black played the highly inconsistent
7 ... i.b4?! and went on to lose) 8 tt::lf3
0-0 9 0-0 and there are many roads for
both players, but they all lead to equal
positions - the imminent and posi­
tionally favourable bishop trade bal­
ances White's space advantage; e.g.
9 ... as 10 'ii'd 2 'flle 7 1 1 .l:!.fe1 d6 12 .txcs
bxcs 13 g3 a4 and the game is still bal­
anced.
64
b) 6 f3 i.b4+ (6 ... b6 is still interest­
ing, but after 7 tt::l c 3 i.cs 8 i.xcs bxcs 9
d6 ! White has balanced the chances) 7
tt::l d 2 'ii'e 7 (7 ... b 6 ! ?). Miles scored 4-0
from this position, beating GM Kaida­
nov and IMs Shirazi and Langeweg; e.g.
8 g 3 o-o 9 ..th 3 c6 (or 9 ... as or 9 ... i.cs)
and Black is better, going on to win in
G.Kaidanov-A.Miles, Palma de Mallorca
1989 (see Game 37).
c) 6 i.g S ? ! is ridiculous here and in
all positions where Black can play ...h7-
h6. The bishop is way too valuable to
trade, and even when it's possible to
retreat, ... h 7-h 6 is always a useful move
for Black. So 6 ...h 6 ! 7 i.xf6 'fllxf6 and
White h as even more dark-square
problems than usual. Black's queen is
also very happy on the f6-square.
6 .tb4 7 f3
...
Against 7 'ftle2 or 7 i.. d 3 Black plays
exactly the same way.
7 i.xc3+
...
Black's results are excellent (+21 -12
=8) with this move, so there is little
need to worry about alternatives.
Sometimes Black plays 7 ...'ile7 with the
idea of trading the dark-squared bish­
ops, but this is time-consuming com­
pared to the text, and hasn't worked
well in real life.
Quoting Joel Benjamin: "This is im­
portant: do not hang your bishop with
7 ... d6 8 'ili'a4+." (Actually, according to
my database, 7 ... d6? ? h as been played
three times and White h as yet to play 8
'ili'a4+, even in the g am e where White
was a GM. I won't name the GM. Can

this possibly be right? ! Can you explain
yourself, Pablo?)
8 bxc3 d6 9 'if'd2!?
In fact, just about anything is better
than the lemon 9 c S ? ! that White
squeezes out most of the time, which
leads to:
Position Four
Apparently i t is just too tempting to
liquidate the weak c-pawn, but the
lines that open are all going to be
seized by Black; and while White's
weakness on c4 is no more, he h as an
even bigger problem on c3, which is
directly in the line of fire, sitting on
Black's half-open c-file. The games all
go the same way: 9 ... 0-0 10 i.d3 ti:Jd7 !
11 cxd6 cxd6 12 ti:Je2 'ili'as ! (or 12 .. .'ii'c 7 )
13 0-0 tt:Jcs, and ... i.d7/ ...llac8/ ...tt:la4,
or ... b7-b6/ ... i.a6/ ... Itac8 on the queen-
side, and after White is pacified there,
.. .f7-f5 on the kingside. S.Brudno­
J.Benjamin, Boston 2001, is a classic
example (see Game 38), whereas in
R.Mitchell-J.Schuyler, Bloomington
1991 (Game 39) Black focuses entirely
1 e4 tt:lc6 2 d4 es
on the queen side, a strategy which is
far less effective, as we will see.
Can things really be that simple for
Black? Not quite, but if White wants to
get any play, he pretty much needs to
go crazy like a Grandmaster (in I.Jelen­
E.Dizdarevic, Ljubljana 1992 - see
Game 40 - White went crazy like an I M,
which did not work), as in E. Rozentalis­
M.Lazic, Genoa 2004, which went 9 cs
o-o 10 cxd6 cxd6 11 i.d3
11 .. .'iic1 12 tt:Je2 ti:Jd7 13 h4! ? tt:Jcs
14 hs tt:Je7 15 g4! and while White may
not h ave been better, he certainly was
scary, and he did go on to win after
1S ... ti:Jxd3 ? ! 16 \li'xd3 bS 17 ti:Jg 3 as 18
�f2 i.a6 19 'iVd2 f6 20 tt:Jfs .l:.f7 21
lthg 1 .l:tc8 22 .l:.ac1 'i¥i>f8 23 'ili'h2 tt:Jxfs 24
gxfs 'iii'c4 25 'iVd2 h6 26 i.b6 'iVa4 27
�g 3 l:!b7 28 i.e3 I!.bc7?? 29 i.xh6 b4
30 �h4 bxc3 31 i.xg7+ �e8 32 'iVg 2
1-0 .
Notice that White's pawn storm had
less to do with blowing Black open than
controlling the fS-square, preparing it
for his knight. The square is especially
important since, had Black been able to
65
Th e Dark Kn ight System
break with .. .f7-fS, White's strategy
would h ave failed. N otice, too, that
Black was never losing - in fact, he was
never significantly worse - until
28 ... 1:!.bc7??. Instead 28 . .J::!.f7 ! (preparing
the escape 29 .. .'�e7 if necessary) would
have pre-empted White's sacrificial
attack and held the balance. White's
ideal position wasn't so dangerous af­
ter all !
Let's go back to move 11, set up the
proper defence without fear, and try to
play a little more quickly and actively
on the queen side: 11 ... 'iia s ! (or
11 ... lt:ld7 first) 12 lt:\e2 lt:ld7 13 h4 lt:le7
14 hs h6 1s g4 f6 16 'i!Vd2 lt:lcs 17 ..ic2
b6 18 lt:lg 3 (18 c4! 'ii'a6 is equal)
18 ... ..ia6 19 lt:lfs lt:lxfs 20 gxfs .l:.f7 2 1
llg 1 'ilo>f8, followed by 22 . . ..:!. c 8 with
pressure on the c3-pawn.
We now return to our regular pro­
gramming.
9 ... b6
Palliser concludes that there is no
point delaying castling, but I am find­
ing that 9 ... 0-0 10 h4! (intending
10 ... lt:lh s 11 ..if2 or 11 lt:\e2) is very
66
dangerous for Black, who must b e ex­
tremely resourceful in a dizzying array
of variations in order to avoid disaster.
(See for yourself - I won't bore you with
the details. Frankly, I wouldn't even
know where to start.) After the text
move, Black can bail out to the queen­
side if things get too hairy too quickly -
a surprising but valuable option !
10 h4
Not forced, obviously, but if this is
what White is about, he is likely to get
right to it.
If 10 ..id3 (or 10 lt:\e2, for that mat­
ter), Black could castle, h aving lost
nothing - but I would prefer to stay
flexible with 10 ... lt:ld7; for example, 10
lt:\e2 lLld7 11 h4 h6 12 h S lt:lgf8 ! (in
Z. Koczka-Zsu.Simon, Hungarian Team
Champion ship 2003, 12 ... lt:le7? was 1-o
in 30 moves) and 13 ...lt:lh 7 will be like
the main variation. If 10 ..id3 lt:ld7 11
lt:\e2, Black can still delay ... o-o in favour
of 1 1 ... lt:lcs ! . Can White keep waiting
before committing to either castling or
h 2-h4 - ?
10... h6
This stops White's pawn from going
to h6, which would leave Black with a
big cramp and weaknesses on the dark
squares.
11 g3
Someone will try 11 hS lt:lf8 ! (I just
love this move - the knight heads for
the open cs-square, whereas a knight
on e7 would be going nowhere for a
long time) 12 g4 lt:\6h 7 ! (preventing 13
g S and preparing an eventual ...lt:lg s !,
 1 e4 lDc6 2 d4 e5

while ...11i'h4+ is also sometimes useful)

13 Si.d3 liJd7 (Black is at least equal in
spite of White's space - his knights are
happy pieces, unlike White's sad bish­
ops) 14 'iii>f2 ttJcs 15 'iii> g 2 as 16 ltJe2
ttJg s etc. lt is even safe to castle king ­
side now.
11 ...liJd7
This tran sposes to A.Karpov-D.Chev­
allier, France 1993 (see Game 41). Ac­
cording to Palliser, this game was very
influential and popularized the varia­ We no longer have to worry about a
tion with Sl.e3 for White. If so, the positional squeeze by White, real or
game's true theoretical significance imagined. White's knight will soon be
was greatly overestimated - Black was established on the e4-square, but it has
fine well into the middlegame, and no forward movement and does not
even better at one point. His only real control important squares. One likely
problem was that he was playing Kar­ plan for us is ... ltJe7-c8-b6 to close the
pov! b-file and pressure White's weakness
12 liJh3 ttJcs 13 liJf2 on c4. We can also consider ... l:.ab8-b6
and .. J:Ub8. Black is clearly no worse
and I believe White has to be careful
that he is not punished for his overex­
tended pawns.

A23 : 5 liJf3 liJf6 6 liJc3 Sl.b4

So far we h ave followed the famous

game, and Black could certainly con­
tinue more or less like Chevallier (who
tried 13 .. .'i*'d7 here), but I would rather
play actively with 13 .. .fs ! ?; e.g. 14 exfs
i..xfs 15 Jl.xcs bxcs 16 i.. d 3 o-o 17 o-o
'ii'd 7. This is a difficult decision, especially

67
Th e Dark Kn ight System

since we will need to know the posi­
tions after 6 ... i.cs anyway because of
the different move orders White can
use (see line A2 1 for this). Although
Black's bishop is very strong on the a7-
This is not good, but i t is always
played, if it hasn't been played already.
Even Houdini likes it! Black's attacking
ideas should be familiar by now:
1o tt:Jhs
...

g 1 diagonal, the doubled c-pawn s are

also a big problem for White, and one
that cannot possibly be fixed. Not only
are they targets in the late middlegame
and endgame, they immobilize White's
queenside, making it difficult for him
to gen erate meaningful play on that
side of the board. Furthermore, White
is forced to play an awkward m ove like
7 i.d3 or 7 'ifc2 to defend his e-pawn.
Because it is possible to inflict the posi­
tional damage and still attack, I prefer Black intends ... tt:Jhf4 and ... lL'lh4, or
to double the pawns if possible, though ... lLJgf4 and .. J!Vf6. On the queenside, it
this is largely a matter of taste. will probably be necessary to play ... b7-
7 i.d3 b6 soon, while a2-a4 should be met by
7 i.d2 ? ! keeps the pawns from be­ ... a7-a5 to eliminate counterplay. In
ing doubled, but it is bad nonetheless: theory White should be okay, but in
7 ... i.xc3 8 i.xc3 tt:Jxe4 9 i.xes tt:Jxes 10 reality Black has won every g ame.
tt:Jxes 'ilif6 11 tbd3 o-o 12 i.e2 d6 13 o-o W.M.Buehl-J.Benjamin, Reno 1999 (see
i.fs and Black has a significant advan­ Game 42) is typical.
tage in activity.
A24: s i.d3 i.cs

7 i.xc3+ 8 bxc3 d6 9 o-o o-o 10 h3?!

...

68

6 tt::l c 3
With 6 tt::l e 2 ! ? tt::lf6 7 o-o White gets
his knight to e2 without spending a
tempo on h 2-h 3 (as compared to the
main line), but he will still find that he
has not found a route to an advantage:
1 ...0-0 (7 ... d6? ! 8 b4!) 8 tt::lb c3 a6 9 tt::l g 3
d 6 1 0 tt::lfs tt::lf4! 11 �e3 (not 11 �xf4?!
exf4 12 li'f3 ? ! g 6 ! 13 tt::lh 6+ �g 7 14
"ifxf4 tt::lh s 1 S 1id2 'i1Yh4! 16 b4 'iif4! 17
·ifxf4 tt::l xf4 winning, because the tt::lh 6
i s still trapped) 11 . . .tt::l d 7! 12 �c2 g6
with equal play.
6 tt::lf6
...
7 h3
Given an exclamation mark by some
sources, which to me seem s a little ex­
treme - the move is more like an ar­
guably necessary evil.
a) Admittedly 7 h3 is far superior to
7 tt::l g e2?? tt::l g 4 8 o-o? �4, winning.
b) 7 tt::lf3 isn't so great either. 7 ...0-0 8
o-o as (8 ... a6 is also reasonable, but the
text move is sound and Black is s-o with
it) 9 a3 d6 10 l:tb1 tt::lh s with a danger­
ous initiative in practice. In A.Hahn­
J.Bonin, New York (rapid) 2003, White
1 e4 tt:lc6 2 d4 e5
managed to block Black's favourite
bishop with 11 b4 axb4 12 axb4 �b6 13
tt::l a4 Ji.a7 14 �c2 tt::lhf4 1S cs, but Black's
attack was still strong, and he won a
topsy-turvy battle (see Game 43).
c) 7 �e3 b6! transposes to A22
above (the first note). 7 ... �xe3 isn't the
worst move ever, but it loses the f4-
square for Black's knights, which love
to settle there and checkmate White.
1 o-o 8 tt::l g e2
...
8 a6
...
Not the only good move. 8 ... as is
also worth considering - it slows down
White's b2-b4, though it does leave
Black's queenside less flexible and
weakens the bS-square.
K.Ellmauer-D.Huber, Schwarzach
2001, went 8 ... d6 (8 ... as is a better
move order) 9 o-o aS 10 �h 1 ? ! (10 tt::l g 3 )
1 o. . .tt::lh s 11 g 4 ? ? � 4 ? ! (11 . . .tt::lh f4!) 12
�g 2 �xg4 13 hxg4 'ti'xg4+ 14 tt::l g 3 ? (14
�h 2) 14 ... tt::lgf4+ 1 S �xf4 tt::l xf4+ 16
�g 1 'ti'xg 3+ o-1. Short and sweet.
I.Jelen-Z.Mestrovic, Slovenian Team
Championship 1996, saw 8 ...tt::lh s ! ? 9
g3?! (9 tt::l a4!), which is incorrectly as-
69
Th e Dark Kn ight System
sessed by De Firmian as better for White
- a case of annotating by result. The
game continued preposterously with
9 ...'ii'f6?! 10 .l:th2 d6? ! (10 .. .'i!Vd8! is equal)
11 ltJa4 �b4+?? 12 ltJec3 ! and Black's
h S-knight is toast.
Going back to move nine, it is true
that White has more space, but his po­
sition is very loose. Simply 9 ... a6 ! fol­
lowed by ... d7-d6 and ... c7-c6, and what
moves does White imagine that he is
going to be playing ? For example, 9 ... a6
10 o-o d6 11 Wh 2 liJf6 12 I;lb1 c6 13 b4
'ifd7 14 ltJg 1 �d4 15 ltJce2 il.. a 7 and
White's position is slowly degrading.
9 0-0 d6 10 ltJg31
otherwise, the nasty 10...ltJh S ! is
coming; e.g. 10 .l:!.b1 liJh S ! 11 b4 �a7 12
Wh 1 ltJhf4 and .. .'ii'h 4 or ...ltJh4 with
threats again st g2, h3, and f2.
(9 ... ltJh S ? ! was no good because of 10
ltJa4! �a7 11 d6 ! with a small edge and
a large disruption of Black's plans.)
The position after 10 ltJg3 has never
been reached, but we have m any rea­
sonable and sensible moves such as
10 ...�d7, 10 ....l:!.e8, or 10 ...h6 (with equal-
70
ity), and ...c7-c6 may be on the cards.
One interesting idea is ...ltJf4. One
thing we absolutely do not need to
worry about is �xf4 exf4 - White's
good bishop is too valuable to squan­
der in such a fashion, and our pawn on
f4 is strong, not weak. We will very
much enjoy the use of the es-square
for our knight, bishop, rook, or queen.
Even if ... g7-g 5 is necessary to defend
f4, Black's kingside will remain safe.
Another idea is ... �d4 and ... c7-c5,
anchoring the bishop. In this closed
position, we will not mind if our bishop
gets traded off, especially since we will
wind up with a protected passed pawn
on the d4-square (if the bishop is taken,
either recapture can be considered,
though ... esxd4 is the typical answer).
Naturally, White will not be eager to
play dsxc6 (en passant) because of the
loss of space and centre. 10 ...�d4 is not
the most accurate, but 10 ... h6 11 ltJa4
can definitely be met by 11 ... �d4 and
12 ... cs.
A25 : 5 g3

Taking the f4-square away from
Black's knight is sen sible, but this move
is slow, and the white bishop is not go­
ing to be active on g2.
s .tbf6
..
6 il.g2
If 6 tLlc3, we have that choice again
- to take the strong diagonal or to
double White's c-pawns. I prefer to
have a queenside target, particularly
since White has taken measures
against our kingside play. Also, White's
fianchetto leaves the c4-pawn without
protection. So 6 ... il.b4 7 il.g 2 il.xc3 8
bxc3 d6 9 tLle2 (or 9 h4 h 6 with a com­
fortably equal position very similar to
A.Karpov-D.Chevallier in A22 - if 10
h s ? ! then 10 ... tLlf8 ! with advantage)
9 ... 0-0 and Black has won every g ame.
White's plan to play f2-f4 is far too
weakening - after ... esxf4 g 3xf4, White
has problems on c4, e4, f4, g4, and h4.
While Black is waiting for f2-f4, he can
play ... b7-b6, ... tLld7, .. .tbcs, ...f7-f5 and
perhaps ... a7-a5 and ... il.a6. E.Schien­
dorfer-D.Recuero Guerra, Herceg Novi
2006, is a typical disaster for White (see
1 e4 tLlc6 2 d4 es
Game 44).
6 il.cs 7 tLlc3
...
The popular 7 tLle2 allows Black to
take the initiative immediately with
7 ... h S !, when 8 ... h4 cannot be stopped
because 8 h4? tLlg4 9 o-o? (or 9 l:i.f1
'ii'f6 ! ) 9 ... tLlxh4! is awful for White, as is
8 il.g s ? il.xf2+.
7 0-0
...
7 ... d6, 7 ... as and 7 ... hs are also in­
teresting, but castling is flexible and
leads into the amusing miniature
H.Titz-C.Rossi, Austrian Team Champi­
onship 2001: 8 tLlf3 d6 9 0-0 a6 10 'ii'd 3
bS 11 b3 il.d7 12 a4 b4 13 lL:ld1 'i!Vc8 14
il.e3 tLlxe4 15 tLlxes tLlxes 16 'ifxe4 il.fs
17 'i!Vh4 tLlg6 0-1. White resigned be­
cause 18 'ii'h s .tg4 19 'i!Vg s h6 traps the
queen .
As seen in this g ame, ... a7-a6 is of­
ten the best way to preserve the bishop
in this variation, because with White's
bishop fianchettoed, the advance ... b7-
bS is easy to achieve and likely to be
effective.
A26: 5 a3 tLlf6 6 tLlc3 il.cs
71
Th e Dark Kn ight System
7 .i.d3
Other moves:
a) 7 l2Jf3 ? lt:Jg4! or 7 l2Jge2 ? l2Jg4! is
terrible. Have you noticed that White
needs to be careful on the dark
squares? Eh?
b) I 've seen White play 7 h3 here, af­
ter which he should h ave his right to
play White permanently revoked - but
the fact that he always loses here has
as much to do with his obviously pas­
sive attitude as the defects of his posi­
tion. 7 ... d6 8 ltJf3 as 9 g3 o-o 10 .i.g 2 c6
was C.Baluta-A.Cioara, Bucharest 1996
(see Game 45), in which White, an FM,
got manhandled while Black demon­
strated all the available ideas ( ... c7-c6,
... b7-b5, ... l2Jg4, .. .f7-f5, etc).
c) 7 b4 certainly looks stupid and
has lost all the games so far, but in
analysis things are not so clear: 7 ... .i.d4
8 l2Jge2 l2Jxe4 {8 ... c s ! ? looks okay) 9
lbxe4 (not 9 l2Jxd4?? l2Jxc3 10 'ifd3 exd4
11 'ifxd4 'ii'e 7+, winning) 9 ... .i.xa1 10
d6 ! (stronger than 10 .i.e3 ? ! .i.d4 11
l2Jxd4 exd4, though after 12 'ifxd4! 0-0
13 d6 White still h as some play) 10 ... b6
72
11 h4 i.b7 12 dxc7 'ii'x c7 13 l2Jd6+ with
full compensation and more than one
way to pursue the attack. Against a
dangerous and prepared (GM) oppo­
nent, Black would probably do well to
avoid the whole mess with 8 ... cs, but
otherwise it makes sense to pocket the
material and let White try to figure out
how to prove compensation .
7 ... a s
7 . . .a 6 has done well in practice, but I
can't see the logic of allowing White to
accelerate his queenside play.
8 l:!.b1
In stead:
a) 8 l2Jge2?? is still terrible due to
8 ... l2Jg4 9 0-0?! 'iVh4.
b) 8 l2Ja4 is interesting, but not es­
pecially strong : 8 ...i.a7 9 cs d6 10 .i.e3
was R.Staudte-M. Roth, Chemnitz 1998,
when 10 ... l2Jh4! (instead of 10 ...l2Jg4? ! )
11 �fl l2Jg4! 12 i.d2 dxcs already
leaves White in poor condition. 9 d6 is
a little better, but 9 ... cxd6 and 10 ... ds
will be sufficient for easy equality.
8 ... d6 9 h3 0-0 10 b4 axb4 11 axb4 i.a7
Black was already slightly more corn-

fortable in Bu Xiangzhi-L.Christiansen,
Deizisau 2000 (see Game 46), and Bu's
12 g 3 ? ! should have worsened White's
situation after 12 ... c6 or 12 ... �d7 (as
opposed to Christiansen's 12 ... l2Je8?!).
Black, who was better throughout most
of the opening and early middlegame,
...lost.
8: 3 dxes ltJxes
White has:
81: 4 lL'lf3 73
8 2 : 4 f4 76
Instead, 4 lL'lc3 .tc5 5 lL'lf3 ? (5 f4 l2Jc6
transposes to line B22 below) 5 ... l2Jg4! 6
l2Jd4 is like a Two Knights Defence (1 e4
e5 2 lL'lf3 l2Jc6 3 �c4 lL'lf6 4 lL'lg 5 d5 5
exd5 l2Jxd5) with reversed colours, ex­
cept that Black's king's knight is at
home instead of his queen's knight. This
makes it harder to pressure the l2Jd4, so
the standard idea of 6 ...l2Jxf2 is less ef­
fective; e.g. 7 'iitxf2 'ii'f6+ 8 'iiie 3 lL'le7 9
lL'lb5 'iie 5 10 c3 f5 11 'iiid 2 fxe4 12 'iitc 2
o-o 13 g3 d5 14 �f4 does not give Black
1 e4 ltJc6 2 d4 es
enough compensation. However, Black
still has 6 ... d5 ! ?; e.g. 7 l2Jxd5 c6 8 f3 l2Je5
9 b4 �d6 10 lL'lc3 .txb4 11 .tb2 11i'h4+
12 g3 flie7 and White is very loose. Al­
ternatively, 6 ...'ifh4 is less ambitious but
much simpler: 7 g3 �f6 8 "ii'x g4 �xd4 9
l2Jd1 d5 ! (since the c8-bishop is de­
fended!) 10 'ii'e 2 dxe4 11 'ifxe4+ lL'le7
and Black is a bit better.
81: 4 lL'lf3 l2Jxf3 5 'iVxf3 lL'lf6!
The usual move is 5 ..."fif6, but then
White can either allow or avoid the
queen trade as he pleases, with a
pleasant position in either case. Al­
though the text move has had no seri­
ous trials, analysis demonstrates its
viability and none of the ideas previ­
ously tried h ave proven themselves
reliable. In most cases, we intend to
play like a Philidor, counting on the
knight trade to ease our space disad­
vantage, while trying to show that
White's queen is awkward on f3; and
there are other possibilities depending
on how White reacts. lt is not possible
to analyse exhaustively, but the follow-
73
Th e Dark Kn ight System

ing information should leave us a few

steps ahead of the opponent, who has
never even considered this position.

811: 6 lLlc3 74
812: 6 ..tc4 74
8 1 3 : 6 e s 75
814: 6 ..id 3 75
8 1 5 : 6 ..te2 75
816: 6 ..tgs 76

811: 6 lLlc3 ..tb4 7 ..td3 ..txc3+ 8 bxc3 6 .....td6

o-o 9 o-o d6 10 :e1 6 ... d6 ! ? is not bad, and keeps things
very Philidor-esque, but I prefer the
bishop to be more active if possible.
7 lLlc3 o-o 8 ..tgs
Or 8 o-o 'iie 7 9 i.f4 (9 i.g s ? ! ii'es !
1 0 'iig 3 'ifcs 1 1 'iifh 4 ! i.es ! is level)
9 ... ..txf4 10 'ii'xf4 d6 11 l::!. a d1 !:te8, in­
tending ... i.e6 or .. .'�i'es. The exchanges
have eased Black's space problem, and
White's edge is tiny.
8 ... c6 9 i.xf6 'ii'xf6 10 'ifxf6 gxf6 11
o-o-o ..tcs 12 f3 bs 13 ..td3 d6
10 i.g s h 6 11 i.xf6 'ilfxf6 12 'ii'xf6
gxf6 is an equal endgame.
After 10 .l:te1, White is a little better
in spite of his bad pawns. lt is tim e for
Black to reposition the knight with
10 ...lLld7, which steps out of the poten­
tial pin, prepares ... 'iif6, and eventually
... lLlcs or ...lLles, where it should have
g ood prospects. Black further intends
...b7-b6, ... i.b7, and ...l:.e8, targeting
White's e-pawn, and may eventually
try for .. .f7-fS. Black's powerful dark-squared
bishop is (nearly) enough to balance his
812: 6 i.c4 ugly pawn structure.

74
 1 e4 lZJc6 2 d4 es

813: 6 es 'ife7 On the surface, this looks bad for

Black, who is still four or more moves
away from completing his develop­
ment - but he controls the e6-square,
and White lacks the means to orches­
trate a breakthrough, so White's ad­
vantage is small; e.g. 11 'iii>b 1 .ifs
(11 ... h s ! ?) 12 lZ:Jb3 lZ:Jxb3 13 'ii'x b3 o-o-o
14 i.. d 3 i.. e 6! with no real problems.

814: 6 .id3 d6 7 o-o .ie7 8 lZ:Jc3 o-o 9

.if4
7 i..f4 9 tZ:Jds lZ:Jxds 10 exds i..f6 11 c3 l:te8
7 'ii'e 2 tZ:Jds 8 c4 'ii'b4+ 9 lZ:Jd2 lZ:Jf4 10 gives White very little.
'iVe4 lZ:Je6 is equal. 9 ....ig4 10 �g3 i.. e 6
7 ... c6 8 lZ:Jd2
Other moves lead to equality:
a) 8 'iVe2 tZ:Jds 9 i.. g 3 fs ! 10 c4 lZ:Jc7
11 lZ:Jc3 lZ:Je6 12 f4 d6 ! 13 o-o-o "iic 7 14
'iVc2 g6 15 Wb1 dxes 16 fxes i.. g 7 (or
16 ... i.. e 7 or 16 .. .f4 ! ?).
b) 8 lZ:Jc3 d6! 9 o-o-o dxes 10 l:!.e1
lZ:Jd7 11 i.. c4 f6 and Black keeps his ex­
tra pawn, although White has the ap­
propriate compensation. Black intends
... lZ:Jb6 (or ... tZ:Jcs), ... .id7 and ...o-o-o.
8 ds 9 o-o-o lZ:Jd7 10 'Wi'g3 tZ:Jcs
••• The threat to trade off White's good
bishop with 11 ...tZ:Jh s is annoying. As­
suming White defends against that,
Black should consider the candidate
moves ... c7-c6, ...�e8, ... lZ:Jd7 (or ... lZ:Jg4),
... tZ:Jes, ... lZ:Jxd3, ... i..h 4, ... i..f6, and ... 'ii'c 7
(or ...Ve7). White is only slightly better.

815: 6 i.. e 2 ds 7 es lZ:Je4 8 lZ:Jc3 lZ:Jxc3 9

'ii'x c3 c6 10 i..e 3 i.. e 7 11 f4 o-o
And with 12 .. .f6 ! coming, Black h as
equalized.

75
Th e Dark Kn ig h t System
816: 6 .tgs d6 1 tt:Jc3
As usual 7 .txf6? ! 'i!Vxf6 8 'i!i'xf6 gxf6
is equal.
1 ... .te1
In principle, we can already notice
that, compared to a regular Philidor,
White's .tg s and 'ti'f3 are awkwardly
placed, so White cannot expect much
from the opening; e.g. 8 h 3 0-0 9 0-o-o
c6 10 'ili'g 3 (10 es tt:Je8 doesn't do any­
thing, while after 10 .te2 �as 11 .td2
'ili'c7 12 �g 3 bS 13 .th 6 tt:Je8 Black is
certainly no worse) 10 ... �as 11 f4 .ie6
leads to a wild race with approximately
even chances.
76
82: 4 f4
This is far less popular than 4 4Jf3
among the strongest players, but gives
excellent results. Fortunately, it's easy
to see where Black has been going
wrong.
4... tt:Jc6
White has:
821: 5 .tc4 77
822: 5 4Jf3 78
Alternatively:
a) 5 tt:Jc3 .tcs 6 4Jf3 transposes to
line B22.
b) 5 .ie3 prevents s ... .tcs, but Black
is okay after s ... .tb4+ 6 c3 .tas
(R.Hilbner-V.Hort, German League
1984, see Game 47). Even better is the
unplayed novelty s ... d S ! which brings
equality; e.g. 6 exds tt:Jb4 7 .tbs+ .id7 8
.ixd7+ 'i!i'xd7 9 4Jf3 tt:Jxds 10 'i!i'e2 o-o-o
- come to think of it, White is much
worse here, so 8 1We2 tt:Jxc2+ 9 'ii'x c2
.ixbs 10 4Jc3 .ta6 11 4Jf3 .1i.d6 12 o-o-o
tt:Jf6 13 llhel o-o, when Black's bishop
pair balances White's space advantage.
 1 e 4 lb c 6 2 d4 e s

821: 5 .i.c4 ltJd5+ 2 3 s.t>c4 ltJe3+ 2 4 'it>c3 with a draw.

Preventing s ....i.cs because of 6 6 ... Jl.b4 7 ltJe2
.i.xf7+ �xf7 7 �dS+. Here 7 e5 d5 8 exf6 dxc4 9 'ii'e 2+?!
s . ..tt:lf6! 'it>f8 ! ? (9 ....i.e6 is also good) 10 .i.e3
'i¥xf6 11 0-0-0 .i.e6 was better for Black,
who went on to win in M.Orso­
G.Bordas, Budapest 2000 (see Game
48). Go Bordas ! 9 �xd8+ improves for
White, when 9 ... s.t>xd8 10 fxg 7 .:.g 8 11
.i.e3 .i.xc3 12 bxc3 l:i.xg7 13 'it>f2 .i.f5 14
litd1+ �c8 15 �d2 b6 is equal.
7 ...tLlxe4!
7 ... d6 is equal; Godena finds more.
8 .i.xf7+ <3;xf7 9 iid s+ 'it>f8 10 'ii'xe4 d s

This strong move, from the I M-GM

clash L.Shytaj-M.Godena, Italian Cham­
pionship, Marta Franca 2008, leads to
very sharp play, not unfavourable to
Black. Before this, s ... .i.b4+ was the
standard reply, as played by E.Pedersen,
Przewoznik, Miles and Hart, who scored
one draw between them ! (lt was Miles.)
6 lbc3
Shytaj's choice.
Only 6 eS is critical, but Black has A sharp reversal o f fortune has oc­
sufficient resources: 6 ... dS! 7 .i.b3 lbg4 8 curred in the centre, and this is far
.i.xd5 .i.b4+ 9 lbc3 (or 9 c3 .i.cs 10 more important than any minor incon­
.i.xc6+ bxc6 11 'ii'xd8+ 'it>xd8 12 ltJe2 venience suffered by Black's king .
.i.f2+ 13 'it?f1 .i.b6 14 ltJd4 f6 15 exf6 11 1i'f3 'ii'h 4+! 12 g3 .i.g4!
.i.a6+ 16 �g1 l;Ie8 17 .i.d2 :e4 18 fxg7 After seeing this, Godena is my new
�e7 19 g 3 .i.d3 with dangerous play - hero.
Black's light-squared bishop is an abso­ 13 'ii'd 3?
lute monster) 9 ... 0-0! 10 .i.xc6 'ii'x dl+ 11 White, who is suffering badly, may
'it?xd1 bxc6 12 'it?e2 f6 ! 13 h 3 ltJh6 14 as well take a pawn for his troubles. 13
ltJf3 ltJf5 15 lbe4 fxes 16 fxes .i.e6 17 'ii'x ds i s not clearly losing, whereas the
.i.d2 .i.xd2 18 �xd2 .l:.ad8+ 19 s.t>c3 lbe3 text move is.
20 ltJd4 .i.ds 21 ltJcs .i.xg2 22 l:thg 1 13 ..."ii' h s 14 o-o

77
Th e Dark Kn ig h t System
14...d4?!
14 ... .l::[e 8! allows less counterplay: 15
.i.e3 d4! 16 ll:\xd4 .l:txe3 17 'ji'xe3 ll:\xd4
18 'ili'd3 lLlf3+ 19 .l:.xf3 .i.xf3 .
15 fs ! dxc3 16 ll:\f4! 'jjff7 17 ll:\e6+ 'iitg8
18 bxc3 .i.d6 19 'ji'e4 ll:\es 20 'i!i'xb7 lieS
21 ll:\d4 cs 22 'iVa6 cxd4 23 'ii'x d6 ll:\f3+
24 'iitg 2 'ii'b 7
White has had enough. Did I men­
tion that Godena is my new h ero?
B22: s lLlf3
Rare at a high level, this lazy-looking
move is White's best, apparently doing
nothing to contest the a7-g 1 diagonal.
s ... .i.cs 6 lLlc3
78
6 ... d6
In stead:
a) 6 ... a6? ! 7 'iNe2 ! d6 8 .i.e3 .i.xe3 9
'iHxe3 ll:\f6 10 o-o-o is pleasant for
White.
b) 6 ... ll:\f6 ! ? leads to wild complica­
tions and is fully playable if you enjoy
such positions. Here are some sample
variations: 7 e s ! ll:\g4 8 .i.c4 (or 8 ll:\e4!
.i.b6 9 i.c4 ds 10 'i!Vxds 'ii'e 7
{10 ...'ii'x dS ! ?} 1 1 h 3 i.e6 12 �5 a6 13
'iWa4 .i.xc4 14 'ifxc4 ll:\e3 15 i.xe3 i.xe3
16 g3 0-0-0, when Black has compensa­
tion for most of a pawn) 8 ... d6 9 ll:\g s
(or 9 lLle4 i.e3 10 exd6 0-0 {10 ... i.xc 1 ! ?}
11 i.xe3 ll:\xe3 12 'ife2 ll:\xc4 13 'i!Vxc4
cxd6 14 o-o-o i.e6 15 'iih s dS 16 fS a6
17 'iic s i.xfs 18 l:.xds 'iie 7 19 ll:\d6 i.e6
20 �d2 i.xa2 ! 21 b3 'ii'f6 22 ll:\d4 as)
9 ...0-0 10 h3 ll:\h 6 11 ll:\a4 i.b6 12 ll:\xb6
axb6 13 o-o dxes 14 'i!Vxd8 l:txd8 1 5
fxes ll:\xes 16 i.b3 cs 17 .i.f4 c 4 with
equality.
7 ll:\a4! i.b6 8 ll:\xb6 axb6 9 i.d3 ll:\f6
9 ... ds transposes to S.Fedorchuk­
A.Miles, Ohrid 2001, which Miles won
(see Game 49), but the idea does not

merit the exclamation mark bestowed
by Kalinin - the simple text is best.
10 o-o o-o 11 .:!.e1
So far we have followed M.Heyne­
R.Vogel, Passau 1999, which actually
started as an Englund Gambit! (1 d4
es?!). Instead of 11 .. J:te8 (which could
be met by 12 es !), Black should h ave
played 11 .. .lbb4, an annoying attack on
White's bishop pair, which he can no
longer preserve (12 .ll c4 .ll e 6!). After 12
.lld2 'Llxd3 13 cxd3 Black's plan is to
mobilize his queenside majority with
13 ... cs, ... b6-bS-b4 and ... b7-b 5 (again !),
develop the bishop with ... .ll d 7-c6, and
restrain White's centre by ... .l:i.e8. lt is
also important to prevent .ll g s, so if
White plays f4-f5, then ...h 7-h6 is the
normal response. I cannot tell a lie -
White has a tiny edge. Notice that
13 ... cs 14 a4 b s ! is tactically possible
(and desirable) because 1 5 axbs l:!.xa1
16 �xa1 .ll d 7 recovers the pawn and
activates the bishop.
If White prevents ...'Llb4 with 11
.lld 2, the answer is 11 ... d s ! 12 es 'Lle4
13 .ll e 1 'Llcs 14 fs 'Llxd3 15 �xd3 �e7
1 e4 'Llc6 2 d4 es
16 .ll c 3 'ilfc5+ 17 �h 1 'ii'c4 with equal­
ity.
White has also tried 11 b3 ! ? .ll d7 12
.llb 2 'Llb4! 13 �d2 'Llxd3 14 cxd3 c5 15
f5 Si.c6 16 �g 5 h 6 17 Si.xf6 'ii'xf6 18
�xf6 gxf6, when Black had equalized in
J.R.Capablanca(!)-M.H.McGuire, New
Orlean s (simul) 1911. Black went on to
win, outmanoeuvring Capablanca with
his better minor and queenside pawn
majority. Outrageous !
C: 3 'Llf3 exd4
Now White h as:
C1: 4 'Llxd4 80 - the Scotch Game
C2: 4 .ll c4 88 - the Scotch Gambit
If 4 c3 (the Goring Gambit), we de­
cline with 4 ... d5 ! 5 exd5 'ii'x d5 6 cxd4
.llg 4 7 .ll e 2 .llb4+ 8 'Llc3 .ll xf3 9 Si.xf3
�c4 and White, who can't yet castle,
has done terribly after 10 'ilfb3 'ii'xb3 11
axb3 with vile pawns, or 10 Si.xc6+
bxc6 ! 1 1 1'ie2+ 'i!Vxe2+ 12 �xe2 0-0-0 13
.lle 3 'Lle7, when Black's king is safe and
guards his weaknesses. White should
79
Th e Dark Kn ight System
really be okay, but has had serious
problems in practice. One good idea for
Black is ... .tas-b6 to pressure White's
d4-pawn and shore up the queenside.
Cl: 4 lt:lxd4 .tcsl
We are allowed to play the normal
move on occasion . There's always time
for weirdness later.
White's main moves are:
C11: 5 i..e 3 81
(12: s lt:lxc6 85
First let's take out the trash :
a) s lt:lb3 ? ! leaves our bishop uncon­
tested on a strong diagonal : s ... .i.b6 6
a4 a6 7 lt:lc3.
The position is highly reminiscent of
a Caro-Kann Classical Variation (in mir­
ror image), but with a few advantages
for us. We have saved a move on ... c7-
c6 (i.e . .. .f7-f6 here), it is hard for White
to trade our strong dark-squared
bishop, and we will find castling short
to be far more efficient and effective
than castling long in the Caro (where
80
... �b8, ....l::!. c 8, ... c6-cs is time consum­
ing).
If anyone tells you to let White play
8 lt:lds, don't believe them : 7 ...lt:lf6 !
(7 ... lt:lge7 is good too, transposing to 6
lt:lc3 lt:lge7 below, but n ot the insipid
main variation with 7 ... d6?) 8 .tg s (8
.i.e2 is preferable) 8 ...h 6 9 .i.h4 d6 and
Black is better already.
Apparently Magnus Carlsen h as h ad
some good results with 6 'ii'e 2 d6 7 .i.e3
but, given the above comparison, it is
clear that this should not be danger­
ous. Black simply needs to keep in mind
the ...f7-f5 break: 7 ...lt:lge7! 8 lt:lc3 o-o 9
0-0-0 fs ! - a few people h ave noticed
this, and Black is +7 -4 =2 in this posi­
tion (most of the games reached by
transposition). Goh Wei Ming-F.Bellini,
Turin Olympiad 2006, is a nice win by
Black (see Game SO).
A cagier move order which is some­
times used is 6 lt:lc3, when it is not yet
clear whether White will be playing a2-
a4 or 'ii'e 2. lt is more important to be
prepared for the fashion able 'ii'e 2, so
6 ... lt:lge7 7 'ii'e 2 d6 8 .i.e3 transposes to

the previous paragraph, while 7 a4 a6 8
.tg s f6 9 i..h 4 o-o is absolutely fine for
Black, who intends ... d7-d6, ... tt:lg6 and
.. .f6-fS, or else ... d7-d6, ... i.. e 6, ... 'ili'd7 (or
...'ili'e8) and .. .f6-fS.
b) 5 lbfs ? ! performs well if Black
doesn't know his stuff (he usually
doesn't). Surprise the surpriser by
knowing this short variation : s ... d S ! 6
tt:lxg 7+ 'iiif8 7 tt:lh s 'ifh4 8 tt:lg 3 tLlf6 ! 9
i.. e 2 dxe4 and Black has a small advan­
tage with which he almost always
wins.
c) 5 c3?! indicates that White has
forgotten the move order. Don't lazily
transpose with s .. .'ii'f6? ! 6 i.. e 3, but
play s ...tt:lf6 ! in stead, with advantage.
C11: 5 i..e 3 'i¥f6
6 C3
White can try 6 tt:lbs here, but after
6 ... i.. x e3 7 fxe3 'jjVh 4+ (this intermezzo
forces a concession) 8 g3 'ii'd 8 9 'i¥g4
g S ! 10 tt:l1c3 tt:Jes 11 'i¥e2 d6 12 h 3 c6
13 lbd4 tt:lf6 14 o-o-o 'iie 7 15 'ii'f2 i.. e 6
16 i.. e 2 o-o-o Black was obviously fine
and went on to win a marathon g ame
1 e4 tt:lc6 2 d4 es
in P.Bontempi-O.Jovanic, Nova Gorica
2008 (see Game 5 1).
6 ...'i¥g6!?
From here the queen threatens e4,
pressures g2, and clears the f6-square
for the knight. Also, we steer clear of
the most heavily analyzed continua­
tion s. Black has done very well with this
move!
There are two main replies, and a
bunch of minor ones.
(111: 7 tt:ld2 82
c112: 1 tt:Jbs 84
Others:
a) The unlikely-looking 7 'i!Ve2 has
also been popular. White hopes for
7 ...ilixe4?? 8 tt:Jxc6 i..x e3 9 tt:ld4, win­
ning (this has yet to work, but hope
springs eternal). Instead, 7 ... tt:lxd4! 8
cxd4 (or 8 i..xd4 i.. xd4 9 cxd4 lbe7 10
tt:lc3 o-o 11 o-o-o c6 and with 12 ... ds
coming, Black is slightly better)
8 ... i..b4+ 9 i.. d 2 i..x d2+ 10 lbxd2 tt:le7
11 g 3 ! (to discourage ll...dS) 11...0-0 12
i.. g 2 �6 and White will have compen-
81
Th e Dark Kn ight System
sation for the pawn he is losing, but no
more than that.
b) 7 f3 a6 ! ? (this is a novelty - it's
time to put a stop to lZ:lbs once and for
all) 8 lZ:ld2 d6 9 'i¥c2 lLlge7 10 0-0-0 il.. a7
11 <it>b1 0-0 is equal; or 8 'ii'd 2 lZ:lge7,
tran sposing to S.Vajda-S.Skembris,
Naujac 1999, which continued 9 lZ:lc2
il..x e3 10 lZ:lxe3 d6 11 il.. e 2 and Black
found plenty of activity with 11 .. .fs,
going on to win (see Game 52).
c) Dembo and Palliser (D&P) like 7
'fif3 pretty well,
but after 7 ... d6 8 lZ:lbs i.. g 4 9 'ii'f4
i.. x e3 10 'i!Vxe3 .l:!.c8 I don't agree that
White is better; e.g. 11 f4 lLlf6 12 i..d 3
0-0 13 0-0 .l:tfe8 14 lZ:ld2 i.. d7 and Black
has no particular problems. 1 5 lZ:lxa7? !
is met by 1 S . . .lZ:lg4 16 f s 'ii'h s 17 'ii'g 3
lZ:lxa7 18 l::tf4! 'ii'xh 2+ ( 1 8. . .lZ:le s 19 l:th4
'i¥xh4 20 Vxh4 lZ:lxd3 2 1 f6 is less clear)
19 'ii'xh 2 lZ:lxh 2 20 'it>xh 2 lZ:lc6 with a
small but persistent endgam e advan­
tage for Black due to White's backward
e-pawn, weak es-square, and defensive
bishop.
d) D&P also like 7 i.. e 2, and it is in-
82
deed dangerous to try to win a pawn,
but in E.Stavropoulou-M. Ikonomo­
poulou, Athen s 2003, Black found an
excellent alternative: 7 ...lZ:lf6 ! 8 lLld2 (8
0-0 d6 9 'it>h 1 ii.d7 10 lLld2 o-o is equal)
8 ... ds ! 9 exds lZ:lxds 10 il..f3 lZ:lxe3 11
fxe3 lLleS ! with a clear advantage.
(111: 7 lLld2
This is pathetic, but it's played most
of the time, so I guess that makes it the
main line! I can't imagine why White
would play the Scotch if this is his plan
for dealing with 6 ... 'i¥g6. The following
draws significantly on Dembo and Pal­
liser.
7 lZ:lf6
...
8 'ii'f3
Other moves:
a) 8 f3 is more common, when Black
shoul d strike in the centre with 8 ... d S ! 9
il..b S ! (stronger than 9 �c2 ? ! i..xd4 10
cxd4 0-0 11 0-0-0 dxe4 12 fxe4 l:te8 13
a3 lZ:lxe4 14 i.. d 3 i..fs 15 dS lZ:ld6 16
i..xfs Vxfs and Black keeps an extra
pawn; or 9 lLlbS ? ! i..x e3 10 lZ:lxc7 'it>f8
11 lZ:lxa8 dxe4 12 'ii'e 2 'i¥h6 13 lZ:lxe4

lbxe4 14 fxe4 lbes and White needs
both to survive and to extricate the
knight, which is more hard than easy)
9 .. .'�xg2 10 l:!.g 1 �xh 2 11 J::t xg 7. Up un­
til now we've been following E. Berg­
I .Morovic Fernandez, European Cham­
pionship, Saint Vincent 2000, which
continued 11 ... 1i.d7 12 'ifl>3 �4+ 13
�d1 and D& P says favours Black, but
Houdini calls even, and Black did go on
to lose. Instead, I am recommending
11 ...�f8 ! ? 12 .l:.g 1 lbxd4 13 cxd4 1i.e7 14
'ifh3 c6 15 e s ! (after 15 1i.d3 dxe4 1 5
fxe4 'i!Vh4+ 16 1i.f2 'ii'f4 17 l:i.fl 1i.e6,
White's compensation is insufficient)
1S ... cxbs 16 exf6 'i!Vh4+ 17 �e2 'ii'xf6 18
'it'xds 'li'fs 19 'ii'e 4 l!Vxe4 20 fxe4 hs 2 1
lbf3 li.g4 with approximate equality i n
an unbalanced endgame.
b) 8 1i.e2 dS ! transposes to Stavro­
poulou-lkonomopoulou a few para­
graphs above. I must say it's interesting
that Black so rarely captures on e4 or
g2. However, 6 .. .'ikg 6 has other points
to it, and White can't leave those
pawns hanging forever.
c) 8 'ii'c 2? ! lbg4! 9 o-o-o lbxe3 10
fxe3 o-o 11 lb2f3 d6 12 lbfs J::te 8 and
Black is on top - D&P.
d) 8 �e2 lbg4 9 lbc2 lbxe3 10 tt:\xe3
lbe7 1 1 lbb3 1i.b6 12 lbfs 'ii'f6 13 tt:\xe7
'ii'x e7 14 g3 0-0 15 1i.g 2 d6 with a small
advantage in R.Roszkowski-A.Leniart,
Grodzisk Mazowiecki 2007 - D& P.
e) 8 lbfs (or 8 lbc2) 8 ... 1i.xe3 9 lbxe3
o-o 10 1i.d3 dS 1 1 exds �xd3 12 dxc6
l:re8 13 lbb3 'ii'b s 14 a4 occurred in
J.Hoogendoorn-A.Van de Oudeweeter-
1 e4 tt:\c6 2 d4 es
ing, Leeuwarden 2001, which was
equal after 14 .. J!kxc6 15 o-o (D&P), but
Black can keep some pressure with
14 ...'i¥a6 !, stubbornly denying White's
castling rights. After 15 �e2 �xc6 16
0-0 tt:\g4 17 tt:\d4 'iUd6 White will soon
have an isolated e-pawn and a some­
what worse position.
f) 8 f4 has worked very well for
White in practice, so be careful ! The
careful respon se is 8 ... lbxd4 9 cxd4 ..ltb4
10 fs 'ii'g 4 11 'ii'x g4 lbxg4 12 li.f4 ds 13
h3 lbf6 14 es lbe4 15 g4 g6 16 fxg 6
fxg6, which is slightly better for Black
because of his imminent f-file control,
though it was not enough to win in
A.Motylev-S.Gligoric, Yugoslav Team
Champion ship 2000 (see Game 5 3).
g) 8 h4 hs! is not helpful to White in
any variation .
s lbg4! g ttJfs
...
The unplayed 9 liJbS ! ? is a better try,
though after 9 ...lbxe3 10 fxe3 �d8, the
position is balanced. And imbalanced.
Go figure.
g lbxe3 10 lbxe3 o-o 11 lbds lbes 12
...
�g3
83
Th e Dark Kn ig h t System
So far this is B.Kharashkina­
O.Stjazhkina, St. Petersburg 2001, and
now instead of 12 ... lLJg4? !, Black should
h ave preferred 12 ...iVxg 3 13 hxg 3 c6,
which D&P call "at least equal", though
I think is safe to call a small advantage
for Black.
c112: 1 tt:Jbs!?
This is the only critical move, but it
is rarely played.
7 ... ..txe3 8 tt:Jxc7+
Just as often White wimps out with
8 fxe3 'lt>d8, when White's structural
problems are far more serious than
Black's king position . Black win s virtu­
ally every game; e.g. 9 tt:Jd2 tt:Jf6 10 'ii'f3
d6 1 1 h 3 l:.e8 12 o-o-o ..td7 and Black
can capture the e-pawn at his conven­
ience. IM Rathnakaran somehow got
caught behind the white pieces in this
variation and tried 10 'it'b3 tt:Jg4! ? 11
0-0-0? tt:Jf2 12 ..te2 tt:Jxh 1 (K.Rathna­
karan-P.Negi, Indian Championship,
Mangalore 2008), but the game is far
too silly to put in my book.
8 ... 'lt>d8 9 tt:Jxa8 i..f4! 10 'ii'f3 i.. h 6!
84
White's queen is lousy on the f3-
square - it can't reach the queenside,
it's vulnerable to ... i.. g 4, and makes it
impossible for White to guard his e­
pawn securely with f2-f3. The position
is still extremely complicated, but ap­
proximately equal.
Some more analysis follows, though
it is hardly practical to be exhaustive,
so here are some bits to keep in mind:
1. The material imbalance (two mi­
nor pieces for a rook, once the aS­
knight drops) is inherently the most
useful for Black in the middlegame -
the minor pieces are well suited for
attack. This means that Black should
not be eager for trades even though his
king is a bit loose.
2. If an endg ame is reached, it is far
better for Black to retain his single rook
than to trade it for one of White's two
rooks.
3 . Pawn exchanges favour White,
whose rooks h ave the most to g ain
from an open board.
4. Naturally, Black will try to keep
his bishop pair if possible. Even deep in
 1 e4 t'Dc6 2 d4 es

an endgame, two bishops are normally shows that in the endgames that are
equal to a rook and two pawns. normally reached, White sometimes
Black h ad a chance to show the wins and just about never loses. Yuck!
power of points #2 and #4 in Don't worry, it's all taken care of.
P.Hromada-L.Ostrowski, Moravian
Team Champion ship 2002 (see Game
54).
In D.Campora-V.Tkachiev, Biel 199S,
White tried 11 iifs 'ii'xfs 12 exfs b6 13
tLla3 i..b 7. As I just mentioned, a queen
trade generally favours White, but the
price was too high: a tempo, a crippled
pawn majority, and a weakened centre
position. Black had no problems after
pocketing the knight, and had winning
chances, although the game ended in a s 'iff6 6 11Vd2
...

draw (see Game S S). Apparently, 6 'iff3 ! ? is topical at the

Dembo and Palliser recommend 11 moment: 6 ...'iWxf3 7 gxf3 bxc6 8 i.. e 3
ie2 t'Df6 12 o-o t'Dxe4 13 i.. d 3 (end of i..x e3 9 fxe3
analysis), but Houdini greatly prefers
Black's position after 13 .. .fs. N aturally,
Black will play ...b7-b6 and ... i..b 7 at his
earliest convenience; e.g. 14 t'Da3 t'Des !
(14 ... t'Dd2 1 S 'ilt'ds i s not worth it) 1 S
·1Wh 3 (not 1 S 'ii'e 2? i..f4! 1 6 g 3 ? ! t'Dg s ! 17
f3 .i:!e8 with a brutal attack) 1 S ...b6 1S
:tad1 i..b 7 and it's still complicated,
but Black is trapping the knight, acti­
vating his pieces, generating threats
against the enemy king, and not being
checkmated, which adds up to an ex­ At first glance the position looks
cellent position . completely equal, which goes to show
you that sometimes first glances are
C12: 5 t'Dxc6 dead on . In practice White has a nag­
This move is popular among Ex­ ging edge (+13 -S =2 1), but that's only
tremely Boring GMs and people with because Black h as not found the cor­
no idea what's going on. (I'm thinking rect plan until this very moment: 9 ... d6 !
of taking it up myself.) Experience 10 t'Dc3 (or 10 l:tg1 g6) 10 ... t'Df6 ! (the

85
Th e Dark Kn ight System
novelty) 11 �f2 (or 11 o-o-o) 11 ...0-o 12
I:!.d1 I:!.e8.
The point is to immobilize White's
centre pawns and prepare ... lieS, from
which the rook has quite a pleasant
view! One obvious idea is ... .l:th s, while
under some circumstances ... d6-dS can
be played (especially if White has cas­
tled long). Black's activity is quite suffi­
cient to neutralize the practical advan­
tage White has been getting after
9 ... lZ:Ie7?!. This idea was inspired by
I.Grynfeld-A.Bisguier, Helsinki Olympiad
1952! (see Game 56), though Bisguier in
fact missed his chance to play ... l:Ies.
6 ...dxc6
86
7 l2JC3
Or:
a) 7 �f4 i.. e 6 8 lZ:Ic3 0-0-0 9 i.d3
h 6 ! ? - kingside expansion with 10 ... g s
will be useful whether the queen s are
traded or not. In any case, Black h as
activity in exch ange for his crippled
queenside majority. For the adventur­
ous, 9 ... h s ! ? is also possible: 10 o-o h4
11 h 3 lZ:Ie7 12 'ifxf6 gxf6 13 lZ:Ie2 l:.hg 8
with more weaknesses for more activ­
ity.
b) 7 i.. d 3 i.e6 and (to make a long
story short) Black gets to castle long,
with excellent development.
7 ... i.. d 4!
White intended 8 'ii'f4 to enter an
annoying endgame. The text move
stops this insidious and somnolent
plan and equalizes, according to
Dembo and Palliser. In practical play,
Black seem s to do even better.
8 i.. d 3 lZ:Ie7 9 o-o lZ:Ig6
Black's knight is taking the scenic
route to the g4-square, where it will be
dangerous to the white king. If Black
had tried to take a shortcut with
 1 e4 lL\c6 2 d4 es

8 ... l2Jh6, White would have put a stop

to it immediately with 9 h 3 ! .
10 �h1
Since the main variation offers noth­
ing, White sometimes tries 10 t2Je2 ! ?
which frees his position after 1 0...il.b6.
So far, taking the pawn has performed
badly for Black, but it is the critical
move, and I do not believe in letting
White off so easily: 10 ...i.xb2 11 il.xb2
�xb2 12 f4 'ika3 13 f5 t2Je5 14 1i'g 5 i.d7
is equal, as in J.Smeets-A.Beliavsky, Mar­ 11 il.e2 l2Jg4 12 l2Jd1 'i¥d6
ibor (rapid) 2004 (see Game 57). Even 12 ... i.e6 is fine too, intending to
better is 14 .. .'�f8 ! followed by 15 .. .f6, castle long.
when White does not have quite 13 i.xg4
enough for the pawn. If 15 f6? ! 'i¥c5+ 16 After 13 g 3 h 5 14 c3 i.b6 15 'ii'g 5
cJo>h 1 g6, the threat of 17 ... l2Jxd3 gives 11t'e5 1 6 f3 'i¥xg 5 1 7 il.xg 5 l2Je5, Black
Black time for ...i.d7 and ... 0-0-0. had the initiative in D.Pavasovic­
N.V.Pedersen, Bled Olympiad 2002; or if
13 f4, Black h as the pleasant choice
between 13 ...t2Jxh 2, with another per­
petual, and 13 ... h 5 ! ?, trying for more
(Dembo and Palliser).
13 ...i.xg4

10...ttJes
The knight is strong, but if White
tries to dislodge it with 11 f4, then
11...t2Jg4 12 'ii'e 1 (12 l2Jd1? ? t2Jxh 2)
12 ... l2Jxh 2 (in B.Sultimov-N .Pokazanjev,
Russia 2007, Black tried for and got
more with the risky 12 ... i.d7 ! ? - see 14 c3 i.f6 15 'it'xd6 cxd6 is about
Game 58) 13 cJo>xh 2 'iih 6+ 14 cJo>g 3 'i¥g6+ equal, while the common 14 'i¥g 5 'it'g6
is a draw by perpetual. is a little better for Black.

87
Th e Dark Kn ig h t System

C2: 4 Si.c4 compensation for the sacrificed mate­

The Scotch Gambit. rial .
4 ...lt:Jf6! 9 ... Si.e7 10 lt:Jxd4

The Two Knights Defence. I have 1o ...f5! 11 .l:If4

chosen sound sidelines for Black to
simplify the study material.
White has:
(21: 5 0-0 88
C22: 5 e5 89
Or s tt:Jg s d S ! 6 exds 'ii'e 7+ 7 �f1
tt:Jes 8 'ii'xd4 lt:Jxc4 9 'ifxc4 h 6 10 lt:Jf3
'iic s 11 'il'e2+ il.e7 12 c4 lt:Jxds
(12 ...b s ! ?) with an edge for Black.
Or 11 Si.h 6 ! ?, when Black can take a
draw with 1 1 .. .fxe4 12 Si.xg 7 .l:!.f8 13
'iih s+ .l:l.f7 14 .l:Ld1 'it'd6 1 S lt:JdbS 'ii'f4 16
tt:Jds 'ifxf2+ 17 �h 1 Si.d7 18 lt:Jf6+ ..txf6
19 lt:Jxc7+ �e7 20 lt:JdS+ c.ii> e 6 2 1 lt:Jc7+,
or play on with 11 ... 0-0 12 lt:Jxc6 bxc6
13 .l:.d4 'ife8 14 il.f4 Si.f6 with equal
chances.
11 ...0-0 12 lt:Jxc6 'il'xd1+ 13 lt:Jxd1 bxc6
14 .l::t c4 c5 15 Si.e3 �e8

C21: 5 o-o lt:Jxe4 6 .l:!.e1 d5 7 il.xd5 'ikxd 5

8 lt:Jc3 'iVdS!?
This is rare but sound. The idea is to
make it more difficult for White to re­
cover the d-pawn.
g l:txe4+
Just as common is 9 lt:Jxe4? ! Si.e7 10
il.g s f6 ! and White, down a pawn and
lacking his light-squared bishop usu­
ally panics with 11 lt:Jxf6+ gxf6 12 il.xf6 Black's bishops are strong enough
o-o 13 il.xe7 lbxe7, but there is no that he is n ot worse.

88

C22: s es lbg4!?
At first this seems insane because of 6
..txf7+ �xf7 7 lbg 5+, but as in so many
similar positions, Black's king position is
fine since White has traded off his most
powerful attacking piece (the light­
squared bishop). So 7 ...�g8 ! 8 'ii'x g4 (af­
ter 8 'ii'f3 ..tb4+! 9 c3 'ii'e 7 10 'ii'd 5+ �8
11 o-o lbgxe5 12 cxb4 lbxb4 13 'Wb3 h6
White has nowhere near enough for two
pawns) 8 ...h6! 9 lbf3 d6 10 'ii'e 4 (or 10 e6
'iif6) 10 ... dxe5 11 l2Jxe5 and now 11...'ikf6
12 lbxc6 bxc6 13 0-0 ..tf5 is simplest,
though Houdini likes 11 ...'ike7 12 f4
'ifh4+ 13 g3 'ii'f6 14 o-o ..tf5 15 'ii'd 5+
�h7 with a large advantage because
White has a stupid position - or, to be
more specific, due to Black's bishop pair
and White's weak light squares.
White al so has:
(221: 6 'ike2 89
(22 2 : 6 0-0 90
C221: 6 'ii'e 2
Most often played, but White strug­
gles to equalize in the main lines.
1 e4 lbc6 2 d4 e5
6 .. JWe7 7 ..tf4 f6! 8 exf6 'it'xe2+!
This is the most forcing.
9 ..txe2
9 �xe2?! is more common, when
the brand new, never been opened, still
in its original packaging 9 ... gxf6 ! gives
Black the advantage; e.g. 10 �e1 l2Jge5
11 �f1 d6 12 lbbd2 ..td7 13 lbe4 ..tg7,
and although Black's position is far
from perfect, White is h aving a hard
time recovering the pawn. Or 10 .l:.d1
lbge5 11 ..td5 d3+ 12 cxd3 l2Jxf3 13
..txf3 l2Jd4+ 14 �d2 lbxf3+ 1 5 gxf3 c6,
when White h as recovered the pawn,
but wishes he hadn't.
g ..lbxf6
.
89
Th e Dark Kn ig h t System
There is still chess to be played, but
the position is equal after (among oth-
ers) 10 ttJbd2 d6 11 ttJb3 d3 12 i..xd3
ttJb4 13 o-o-o ttJxd3 14 .l:!.xd3 h 6 and
1 S ... �f7, though Black has man aged to
stir things up a bit. Worse for White is
10 i.. x c7? ! d6 ! 11 .i..b s i..d 7 1 2 i..x c6
i.. x c6 13 ttJxd4 i.. x g 2 14 .l:.g 1, as in
R.Stranz-K.Neumeier, Austrian Team
Championship 2004, when Black
should have continued 14 ... i..h 3 with
an edge.
C222: 6 o-o i.. e 7
7 l:.e1
Or 7 i..f4 g S ! 8 ttJxg s (8 i.. g 3 h S ! 9 h 3
h4!, D.Von Wantoch Rekowski-J.Peric,
Tivat 2001, was a strong case for Black
- see Game 59) 8 ... dS! 9 e6 (9 exd6? !
i..xg s 1 0 dxc7 1Wf6 11 1We2+ <;t>f8
doesn't give White enough for the
piece) 9 ... i.. xg s 10 'i¥xg4 i..x e6 1 1 i..xg s
i.. x g4 12 i.. x d8 .Uxd8 and Black keeps a
pawn, for which White has n ot nearly
enough.
7 d6 8 exd6 cxd6!
...
Only 8 ... 1Wxd6 has been played, but
90
then 9 b3 ! o-o 10 .i.. a 3 'ii'd 8 11 i.. x e7
ttJxe7 12 'ifxd4 'ifxd4 13 ttJxd4 gives
White a good endgame.
Instead, the text move is very dy­
n amic - a main idea is to use the g4-
knight to harass White on the soft f2-
square. Let's analyze: 9 ttJxd4 (not 9
h 3 ? ! ttJge S and Black is better already)
9 ... 0-0 and n ow:
a) 10 h3 ds! 11 ttJxc6 (or 11 i..b 3 ? !
i.. c s ! 12 c3 ttJxf2 ! 13 'it>xf2 ttJxd4 14
cxd4 'iWh4+ 15 �f1 i.. xd4 16 i.. e 3 1Wf6+
17 'itg 1 i.. xb2 18 ttJd2 i.. x a1 19 ii'xa1
'ii'x a1 20 l;lxa1 l:.e8 with advantage -
quite a small one actually; the material
balance very strongly suggests trades
for Black, so if 21 'iii>f2 i.. e 6 22 ttJf3 ? !
then 22 . . .d4! ! makes progress) 1 1. . .bxc6
12 hxg4 (12 .i.. d 3 ttJf6 is equal) 12 ... dxc4
13 'ii'f3 i.. e 6! 14 1Wxc6 .i.. x g4! 15 1Wxc4
.i.. e 6 16 �e2 l:i.e8 17 ttJc3 i..f6 and the
board is so open that the bishop pair
offsets the missing pawn. White also
lag s in development.
b) 10 ttJc3 i..h 4! 11 g3 i..f6 12 ttJdbS
a6 13 ttJxd6 ttJxf2 !
Wow! it's complicated, but at least

you won't be seeing the position for
the first time, unlike your poor oppo­
nent. As it turns out, White has several
acceptable routes to a draw, but no
good way to play for a win :
b 1 } 1 4 t£:,xf7 'iVxd1 1 5 Ci:,xd1 Ci:,xd1
and White can repeat moves now with
16 Ci:,d6+ 'it>h 8 17 Cf:,f7+, or play 17 l:f.xd1
.id4+ 18 .l:!.xd4 (18 �h 1 .ig4! is trouble
for White, as is 18 �g 2 :f2+ 19 �h 1
.ig4) 18 ... Ci:,xd4 and repeat moves here
with 19 Cf:,f7+ 'it>g 8 20 Ci:,g s+ etc.
b2} 14 'i!Vds .i.d4!, when White h as a
1 e4 Ci:,c6 2 d4 e5
couple draws, but 15 .l:!.e8 ? ! is not one
of them : 1S ...'il¥xe8 16 Ci:,xe8 .i.e6 ! 17
� 5 .i.g4 18 t£:,f6+ .txf6 19 'iVds .l:tad8
20 Wxf2 .l::!. x ds 21 tt:Jxds .i.d4+ is better
for Black, who is more active, and
White's king is still a problem. Instead,
15 'ifxf7+ l:!.xf7 16 t£:,xf7 Ci:,d3+ 17 'iii>h 1
Cf:,f2+, and now i t i s Black who is best
advised to repeat moves; or 1 5 t£:,xf7
'Lld3+ 16 .ie3 'ifxds 17 .txds Ci:,xe1 18
.l:txe1 .i.fs 19 .ixd4 Ci:,xd4 20 .l:.e7 .i.xc2
21 l:txb7 Cf:,fs and the usual repetition
follows.
b3) 14 .txf7+ �h 8 15 �xf2 .id4+ 16
�g 2 Vi'xd6 17 .if4 Vies 18 Ci:,e4 'iffs and
the complications are not over, but
White's king position is far too fragile
for him to have serious thoughts about
winning - losing is quite attainable
though !
b4} 14 'it>xf2 ? ! doesn't make sense .
14 ... .i.d4+ and 1 S ... 'it'xd6 is not terrible
for White, but he h as nothing to com­
pensate for his loose king.
91
 C h a pter Five

1 e4 t2Jc6 2 tiJf3

is unlikely to keep the bishop pair and

he won't be able to stop Black from
shifting his pawn structure to light
squares. Nonetheless, 3 ... tiJf6 4 tiJc3 g6
5 d4 a6 is simpler, especially since we
will need to know this position anyway.
3 tiJf6
.••

With this move, White hopes to en­

ter his favourite double-king-pawn po­
sition, which he will be happy to do
after 2 ... e5. We are obligated to disap­
point him, and in the resulting Pirc po­
sition s White may wi sh he h ad not
committed his knight so soon - the
Austrian Attack (f2-f4) and the Argen­ 4 tiJc3
tinean Attack (f2-f3) are n o l onger Other moves:
available to him, nor are several other a) 4 c3 is possible and does not ac­
aggressive variations. tually drop a pawn : 4... ttJxe4?? 5 d5
2 d6 3 d4
... tiJb8 6 1Wa4+ picks up the e4-knight.
After the premature 3 ..tb5 we have However, 4 ... g6 i s fine for Black. Miles
an extremely rare situation where it is downed GMs Becerra Rivero and Zelcic
fine to play 3 . ..lt:Jf6 4 tiJc3 ..tg4 - White starting with this move (see Games 60

92

and 61). Mestrovic is also 2-0 with 4 ... g 6
here. T o continue: 5 .i.d3 .il.g 7 6 o-o o-o
7 h 3 e5 8 .l:le1.
Although Black h as tried many
things in hundreds of games, I am rec­
ommending 8 ... b6!, as played once in
I.Vasilevich-M.Allakhinova, Russian
Team Champion ship 2002. One point is
that, after 9 d5 ti:Je7, Black's pawn
holds up White's queenside play; while
9 .i.e3 exd4! 10 cxd4 tl:Jb4 picks up the
bishop pair: 1 1 ti:Jc3 ti:Jxd3 12 'ii'x d3 l;!e8
13 .l:!ad1 (or 13 llac1 c5) 13 ... .il.b7 14 d5
l:tc8 with only a small advantage to
White; ....l:.c8 helps prepare an eventual
... c7-c6, and pre-empts any White pres­
sure down the c-file.
The game continuation should not
have been too dangerous either: 9 'iWa4
.i.d7 10 .i.b5 'ii'e 8 11 d5 ti:Je7 (11 ...ti:Jb8 !)
12 .i.xd7 ti:Jxd7 (12 .. .'ii'xd7!) 13 'ii'c4 ti:Jc5
14 b4 b5 15 "i!t'e2 tl:Ja4 16 .i.d2 and now,
instead of 16 ... c5?! as played, 16 .. .f5 ! ! 17
ti:Jg 5 fxe4 18 ti:Je6 'ii'f7 19 tl:Jxc7 .l:!.ab8 20
ti:Je6 l:tfc8 21 'ifxe4 .i.f6! is good for Black,
who plans 22...l2Jb6 and 23 ... ti:Jbxd5.
White tried 8 ti:Ja3 in L.Rozman-
1 e4 ti:Jc6 2 tl:Jf3
J.Schuyler, Washington 2012: 8 ... d5! 9
.i.g 5 ? ! (9 exd5 ti:Jxd5 10 dxe5 ti:Jxe5 11
ti:Jxe5 .i.xe5 12 .l:le1 .i.g7 is equal)
9 ... dxe4 10 .i.xe4 exd4 11 .i.xc6? ! (11
cxd4 is still fairly level) 11 ... dxc3! with
great complications favouring Black
(see Game 62).
b) 4 ti:Jbd2 is similarly met by 4 ... g6.
Miles took down GMs Zapata and
Nijboer with this move (see Games 63
and 64). Given his successes with this
fianchetto, it is surprising to me that
Miles was never willing to play 4 ... g 6
against 4 ti:Jc3, even when other moves
were giving him trouble.
c) 4 .i.d3 doesn't look right because
Black can pick up the bishop pair with
4... tl:Jb4. My first inclination was that
this must equalize immediately, but
apparently tempi sometimes matter in
chess ! After 5 o-o ti:Jxd3 6 'ii'x d3 g6 7 c4
.il.g7 8 ti:Jc3 it will be difficult for Black
to disrupt White's huge centre. There­
fore, 4 ... g6 here too! All three of White's
fourth move sidelines may transpose
into each other.
4 g6!
...
93
The Dark Kn ight System

A: 5 .ib5 94
B: 5 d 5 9 6
C: Others (without d4-d 5) 102

A: 5 .ib5 a6

4 ... i.. g 4 is by far the most common

move, ensuring that Black reaches po­
sitions unique to 1...tt:Jc6, but Black has
both objective and practical problem s
after White's most accurate reply 5
i.. e 3. Then Miles always played 5 ... e6
and, when that stopped working, 5 ... a6, 6 i..xc6+
which was worse, while Mestrovic has 6 i.. a4 doesn't make much sense,
repeatedly but unsuccessfully tried to but it has been played several times,
uphold 5 ... e5. After 6 i..b 5, his record most significantly in M.Kozakov­
speaks for itself (+2 -9 =3). A.Zajarnyi, Lvov 1998, which continued
On the other hand, our standard fi­ 6 ...b5 7 i..b 3 i.. g 7 8 h3 o-o 9 o-o e6 10
anchetto with 4 ... g 6 is reliable, trans­ a3 i..b 7 11 1:te1 tt:Ja5 12 i.. a 2 c5 13 d5
posing to a Classical Pirc (1 e4 d6 2 d4
tt:Jf6 3 tt:Jc3 g6 4 liJf3) with the odd
4 ...tt:Jc6. There are many strong players
h appy to use this placement for the
knight; for instance, GMs Smirin and
Adorjan h ave played (4... i.. g 7 5 i.. e 2 o-o
6 0-0) 6 ... tt:Jc6 twenty times between
them, with seven wins and one loss. Of
course, 6 ... tt:Jc6 is not precisely 4 ...tt:Jc6,
but I intend to show that the differ­
ences are not critical.
White has many moves h ere, so I and now after 13 ... e 5 ? ! 14 b4! cxb4
will divide the material into three 15 axb4 'fkc7 16 bxa5 'i!Vxc3 17 .i.d2
groups: 'i!Vc7 18 c4! White eventually converted

94

his advantage. Instead, Black could
have kept the balance with 13 ... exd 5 !
14 lZ'lxd5 ..txd5 1 5 ..txd5 l2Jxd5 16 �xd5
lZ'lc4. The important feature of the posi­
tion from Black's standpoint is the ac­
tive g7-bishop.
The simple 8 ... lZ'la5, picking up the
bishop pair, also brings equality; e.g. 9
o-o ..tb7 10 e5 dxe5 11 dxe5 l2Jxb3 12
axb3 lZ'le4 13 'ii'x d8+ l:txd8 14 l2Jxe4
..txe4 15 llxa6 ..txf3 16 gxf3 ..txe5 and
if you're still awake after all these
trades, you'll notice that White's dis­
gusting pawn structure is balanced by
some extra rook activity.
Based on the Kozakov g ame, 6 ..ta4
is recommended (sort of) by Andrew
Greet in Beating Unusual Chess De­
fences: 1 e4, so it is a good idea to be
prepared.
6 bxc6 7 0-0 ..tg7
...
8 h3
White usually plays this move
sooner or later in the Dark Knight Pirc,
as he gets tired of worrying about
... ..tg4. Others:
a) 8 e 5 ! ? lZ'ld5 9 lZ'le4 o-o 10 l:te1 was
1 e4 l2Jc6 2 lZ'lf3
equal in O.Biti-G.Belamaric, Portoroz
2005. White's lunge has pushed into
where Black is strong and opened up
squares for the c8-bishop. Black should
play 10 ... a5 11 h 3 c5 ! ? 12 exd6 cxd6 13
dxc5 dxc5 14 lZ'lxc5 'fid6 1 5 lZ'le4 �c6 16
c3 a4 and in this open position, the
bishop pair and queen side pressure are
fully worth the sacrificed pawn.
b) 8 'ii'e 2 o-o 9 .:td1 ..tg4 (9 ... a5 !
equalizes) 10 h 3 ..txf3 11 'ii'xf3 lZ'ld7 12
..te3 e5 with a small edge for White in
E.Sveshnikov-T.Gelashvili, Cappelle la
Grande 2009, albeit one he was unable
to convert (see Game 65).
8 0-0 9 1::te 1 l::[ b 8 10 l:l.b1 lZ'ld7
...
u ..tgs?!
The strong White and Black players
to hold this position were focused on
the ... e7-e5 break when they should be
preparing for ... c6-c5.
11 ..te3 is a bit better, though it still
doesn 't stop 11 ... c 5 ! ; e.g. 12 dxc5 .l:txb 2 !
13 .l:!.xb2 ..txc3 1 4 .Ub3 ..txe1 1 5 'i!i'xe1
l2Jxc5 16 ..txc5 dxc5 equal, or 12 Vi'd2
cxd4 13 ..txd4 l2Jf6 14 e5 lZ'lh 5 15 exd6
cxd6 16 ..txg7 lZ'lxg 7 with equality.
95
Th e Dark Kn ight System
u cs!...
After 11...h6 12 .i.h4 g S 13 .i.g3 eS?!
(13 ... cs or 13 .. .fs ! ? is better) 14 dxes
dxes, as in B.Vuckovic-Z.Petronijevic
and D.Nestorovic-N.Ristic {from the
2003 Serbian Team Championship),
White enjoys a large structural advan­
tage which is not compensated for by
Black's inactive bishop pair.
Whereas 11 ... c s ! 12 es l:te8 13 exd6
cxd6 14 tt:Jds f6 15 dxcs tt:Jxcs sees
Black starting to get the better of it. Of
course, White could swallow his pride
and play 12 .i.e3, which is equal after
12 ... e6.
B: s d s tt:Jbs
As I have mentioned before, one of
the significant benefits of the fi­
anchetto is that 5 dS opens the diago­
n al for the g 7-bishop. lt would be a
shame to close it with s ...tt:Jes 6 tt:Jxes
dxes .
6 .i.e2
Other moves:
a) 6 i.. c4 isn't as stupid as it looks -
after all, when Black plays ... c7-c6
96
White will enjoy having the option to
open the a2-g 8 diagonal whenever he
wants. However, Black is allowed to
change his plan s too: 6 ... i.g 7 7 0-0 o-o
8 h3 (this is usually played by this
stage, on moves 5, 6, 7 or 8; otherwise
White has to worry about both ... i.. g 4
and ... tt:Jg4) 8 ... e s ! ? (this is a new move
intended to punish i.. c4 and h 2-h 3; if
White refuses to open the position, the
c4-bishop is atrocious - both passive
and hindering White's natural plan of
c2-c4-c5) 9 i.. e 3 (or 9 dxe6 .txe6 10
.i.xe6 fxe6 11 es dxes 12 �e2 tt:Jc6, and
with the knights jumping to the dS­
and d4-squares, Black h as full compen-
sation for his weakness on e6) 9 ... a6 10
b4 (if 10 a4 then 10 ... as and 11 ... tt:Ja6)
10 ... tt:Jh s 11 'it'd2 ti:Jd7 12 tt:Je2 "ile7 and
with .. .f7-f5 coming, Black has reached
a good King's lndian-type position .
Instead of 8 . . .e s ! ?, N .Praznik-A. Beli­
avsky, Bled 1999, continued 8 ... c6 9 a4
as 10 l:.e1 tt:Jfd7 11 i.. e 3 tt:Ja6. Though
White was better, Black played effec­
tively on the dark squares throughout
the game and went on to win, making
this a must-see (see Game 66).
b) 6 i.g s is in fact the most often
played. White intends long castling and
a quick attack, but despite the time
gained against Black's knight, this strat­
egy is questionable - Black's counter­
play with ... c7-c6 is very fast: 6 ... i..g 7 7
�d2 c6 8 i..h 6 (after 8 0-0-0, Black cou­
rageously castled in F.Lukez-S.Lejlic,
Rodeby 1998, and equalized with 8 .. 0-o .
9 i..h 6 i.. g 4 10 i..xg7 �xg7 11 i.e2
 1 e4 l'Llc6 2 l'Llf3

l'Llbd7 - White's "attack" is going no­ position to B2. Of course White could
where; e.g. 12 h3 i..xf3 13 i..xf3 'irb6 14 have played 6 h3, insisting on B1.
h4? ! tt:Jes, threatening ...l'Llc4) 8 ...i..xh6 ! 9 7 0-0 0-0
�xh6 'i!Vb6 10 o-o-o i.. g 4 is an interest­
ing position, not at all unfavourable for
Black; e.g. 11 l:.d2 i..xf3 12 gxf3 l'Llbd7 13
f4 cxds 14 exds .l:.c8 (14...0-o-o! ? and
15 ...<;i;>b8) with great interest in ....l:txc3;
or 11 i..e 2 'ii'xf2 12 .l:.hf1 'ii'c s and while
White is certainly well developed, he has
nothing concrete for the pawn. In
N.Sulava-M.Muse, Croatian Team
Championship 2002, White changed the
course of the game with 11 e s ? ! dxes 12
d6, but had Black spotted 12 ...l'Llbd7 13 White has:
dxe7 'ii'h4, followed by 14...'ili'xe7 and
15 ...0-0-0, White would have found him­ 81: 8 h 3 9 7
self without sufficient compensation. 82: Others 99
c) 6 i.. e 3 is very similar to 6 i.. g s,
into which it often transposes (i.e. after Or 8 a4 as ! - Black cannot allow
i..e 3-h6). One time that didn't happen himself to become further cramped.
was in K.Nemcova-F.Olafsson, Marian­ The insertion of the two a-pawn moves
ske Lazne 2008, which went 6 ... i.. g 7 7 is helpful to Black though, since it helps
'iWd2 c6 (by transposition) 8 h 3 b S ? ! him to establish knight outposts on the
( 8. . .0-o 9 i.. e 2 b s ! , a s i n J. Hjartarson­ cs- and b4-squares.
F.Olafsson, Reykjavik 1995, was a better
move order - see Game 68) 9 a3 ? ! (9 81: 8 h 3
dxc6 ! b4 10 i..b s ! is unpleasant for
Black) 9 ... a6 10 dxc6 l'Llxc6 1 1 i..d 3 o-o
12 o-o i..b 7, resultin g in an equal
Dragodorf type of position which Black
went on to win (see Game 67).
d) 6 h 3 ! is the most accurate move,
reaching line B1 below after 6 ... i.. g 7 7
i.. e 2 o-o 8 o-o.
6 i.. g 7
...

There is something to be said for

6 ... i.. g 4 (or 7 ... i.. g 4), with a likely trans-

97
The Dark Kn ig h t System
This is played most often, spending a
tempo to stop 8 ... .ig4. As we will see,
8 ... .ig4 was indeed Black's intention,
but a tempo is a tempo. There is some
disagreement about the merits of 8 h 3
- Alburt and Chemin adorn i t with an
{!), while Nunn says it is "not really
necessary". I think it is the best move in
the position .
s es!
...
Smirin and Gulko h ave each chosen
8 ... es three times, with an even (total)
score, while Finkel, Urban, and Gufeld
have four wins and three draws with
8 ... c6, but against this, Benjamin's plan
(9 a4 as 10 l!e1 tt:la6 11 .ixa6 ! - see the
note on 8 ... c6 in line B2) is very strong
for White.
For those interested in making ... c7-
c6 work, I suggest 8 ... a6 (or 8 ... b6) 9 a4
b6 10 .ie3 .ib7 11 'ii'd 2 c6 12 .l:!.ad1
'Wic7 13 .ih 6 tt:lbd7 as in A.Bachofner­
P.Hopman, Amsterdam 2006, when
White is only slightly better.
9 dxe6
If White does n ot play this m ove, he
winds up in a King's Indian type of po-
98
sition where his normal play (with c2-
c4-c5) is blocked by the c3-knight. This
problem is serious for White, more so
than Black's funny knight on b8 (which
is n ormally on the e7-square). Black will
play carefully for .. .f7-fS. Part of being
careful is considering ...h 7-h 6 to pre­
vent White's tt:lg s-e6.
g .ixe6
...
Black has caught up in development
and now has only to worry about a
small space disadvantage. If White is
not alert, Black will fix this with a quick
... d6-ds.
White has now only showed interest
in :
811: 10 i.gs 98
812: 10 tt:ld4 99
811: 10 .igs !
Dubious according t o Nunn, in h i s
1 9 8 9 The Complete Pirc, this h a s been
by far the more dangerous move in
practice.
10 h6
...

11 i.. e 3
In stead:
a) 11 i..h 4 tZ:lc6 12 'iid 2 gS 13 li.g 3
dS equalizes.
b) 11 i..f4 lL'lc6 12 'ikd2 gS 13 i..h 2
.ll e 8 1 4 l:tad1 tZ:ld7 offers White a tiny
edge, but typical dark-square play for
Black.
11 ... tZ:lc6 12 'i¥d2 'it>h7 13 .l:tad1 l:te8 14
.U.fe1 a6 15 a4
So far we have followed J.Piket­
B.Gulko, Amsterdam 1989, which con­
tinued 1 S ... i.. d 7 16 i.. c 4 i.. e 6 17 i.. e 2
'ife7 18 tZ:lds 'ild8 19 tZ:lc3, indicating
that nobody could come up with a plan
- this is already a threefold repetition,
1 e4 tLlc6 2 lDf3
but somehow the game continued and
Gulko contrived to lose, even after a
later repetition.
Alternatively, 1 S ...'i¥e7 is fine to test
the opponent for a draw (16 tZ:lds �d8
17 tZ:lc3), but if one wants to continue,
there is 1S ...tZ:ld7 ! ? 16 .tfl b6, intending
17 ... tZ:lcs, while ... lL'ldes is also a possi­
bility, as is the manoeuvre ...'i¥c8-b7.
White is a little better.
Y.Gruenfeld-I.Smirin, I sraeli Team
Championship 1997 (see Game 69),
shows what happens if White does not
play a2-a4 to contain Black's queenside
expansion - Black's counterplay was
more than sufficient.
812: 10 tZ:ld4 i.. d 7
The insertion of these moves makes
it easier for Black to pressure the e4-
pawn.
11 .l:r.e1 tZ:lc6
N ow White either loses time re­
treating the knight or agrees to an ex­
change that will ease Black's position .
Similar is 11 ...l:te8 12 i..f1 tZ:lc6 13 tZ:lb3
as 14 a3 a4 1 5 tZ:ld2, and instead of
99
Th e Dark Kn ight System
15 ...tt:\a5 ? ! 16 lLlf3, which was good for
White in K.Hulak-S.Marangunic, Yugo­
slavia 1977, Black should play 15 ... tt:\d4!
16 tt:\f3 tt:\xf3+ 17 'ii'xf3 i.c6 with an
equal position.
12 lLlf3 :es 13 i.c4 .i.e6 14 tt:\ds 'it'd7
15 c3 i.xds 16 exds .l:txe1+ 17 'ili'xe1
tt:\es ! 18 tt:\xes dxes
The position is equal, White's bish­
ops being offset by his poor develop­
ment and the inconvenience of guard­
ing the d5-pawn. If 19 'it'xe5, then after
19 ... l:te8 and 20 ...l:te1+, White will never
complete his development.
82: Others (besides8 h3)
1 00
If White is going to refrain from 8
h3, it makes very little difference which
move he chooses, but we'll take the
following as the main line:
8 i.e3
Alternatively:
a) 8 i.g 5 should be treated simi­
larly: 8 ... i.g4! 9 h 3 i.xf3 10 i.xf3 c6 11
'ii'd 2 lLlbd7 and Black follows with
...'ii'b 6 , ... a7-a5, ... tt:\c5 and ... tt:\fd7, with
typical dark-square play, even if White
is still slightly better.
White can al so try to do without h2-
h 3 . For in stance, 9 'ir'd2 c6 10 l:lfe1
tt:\bd7 transposes to L.Vajda-M.Marin,
Rumanian Champion ship, Bucharest
1998, and M.Kolosowski-Dan .Fraczek,
Legnica 2011, which both continued 11
dxc6 ! ? bxc6 12 tt:\d4 :c8? ! 13 f3 i.e6 14
tt:\xe6 fxe6 with advantage to White.
According to Houdini, Black can equal­
ize in this line with 12 ...'ii'b 6 ! 13 tt:\a4
'iWc7 14 i.xg4 tt:\xg4 15 i.xe7 d5 16 f4
(16 lLlf3 ? ! tt:\de 5 ! 17 i.xf8 lLlxf3+ 18
gxf3 'ili'xh 2+ 19 �f1 .l:.xf8 20 fxg4 'ili'h 1+
2 1 �e2 'ili'xe4+ 22 �1 'ir'xa4 is good for
Black) 16 ... l:tfe8 17 exd5 cxd5 18 h 3
lLl g e 5 19 dxe5 l:!.xe7.
b) 8 l:!.e1 i.g4 (of course) 9 h3 i.xf3
10 i.xf3 tt:\fd7! 11 i.e3 c6 was
D. Rogozenco-V.Nevednichy, Rumanian
Team Championship 2005, where
White went wrong immediately with
12 i.d4?!, initiating an exchange bene­
ficial to Black; e.g. 12 ...i.xd4 13 'ii'xd4
'ii'b 6 is equal. Without this mistake
White is a little better, but Black's mi­
nor pieces are all good, and he h as the

simple plan of ... a7-a5, ...t'Da6, ... t'Dacs,
... as-a4-a3 (if possible), and ...'i!Vb6 or
.. .'�C7.
8 ..tg4
...
Black has done very well with 8 ... c6,
but Joel Benjamin (who played 1...l'Dc6
frequently in the 1990s) warned m e
that Black's position is difficult after 9
a4! as 10 h 3 (otherwise 10 ... ..tg4)
10 ... l'Da6 11 ..txa6 ! (an idea I have not
been able to find in any published
games - the point is to stop Black's ac­
tive 1 1...l'Db4) 11...l:.xa6, when it is hard
to find a con structive plan for Black,
whereas White can still build; e.g. 12
'i!Vd2 .l:.a8 13 l':!.ad1 l:.e8 14 ..td4 'ilc7 15
.:tfe1. Black lacks space, development
and pawn play, while his only "asset",
the light-squared bishop, is more of a
problem than anything else. Indeed,
this type of position acts m ore closed
than open, in part because nobody
wants to relieve the tension between
the ds- and c6-pawns - for White to
trade would assist Black greatly in the
central battle, while if Black trades, he
has accessible weaknesses on the bs­
and e7-squares.
A possible antidote is 8 ... as ! ? 9 a4
tt:Ja6, when White shoul d be less eager
to snap off the knight. N otice that since
Black h as not yet played ... c7-c6, he can
later try ... e7-e6 or ... e7-e5 instead.
However, there are other moves to
worry about besides 9 a4.
8 ... ..tg4 is simplest, transposing to a
favourite line of the great Pirc expert
Alexander Chemin, who used it with
1 e4 t'Dc6 2 tbf3
great success in the 1990s (three
draws, three wins, all against GMs). The
normal m ove order to reach this posi­
tion is 1 e4 d6 2 d4 ttJf6 3 t'Dc3 g6 4 l'Df3
..tg7 5 ..te2 0-0 6 0-0 ..tg4 7 ..te3 t'Dc6 8
ds t'Db8. (8 ... ..txf3 is the traditional
main line, but it is not stronger.)
9 h3
White must play this sooner or later
unless he is intending to allow the ex­
change of light-squared bishops. White
should at least retain the bishop pair if
he is h oping to keep an advantage:
a) 9 t'Dd2 ..txe2 10 'ilxe2, and now
Chemin's recommendation 10 .. J:te8 11
f4 e6 has been tested only once, in
A.Czebe-N .Resika, Budapest 2000 -
Black, an FM, held the draw against the
GM.
b) 9 l'Dd4 ..txe2 10 'i/xe2 cs 11 l2Jf3
'iib 6 12 l1ab1 'ila6 13 'i/d2 l'Dg4 14 ..tg s
l:te8 with equality in B.Ch atalbashev­
M.Popchev, Cacak 1991 (see Game 70).
c) 9 t'Dg s ..txe2 10 'ilxe2 c6 11 .l:.ad1
'i/as 12 f4 'i/a6 13 'ilf3 t'Dbd7 with
equality. Black, with more experience
in this type of position, went on to win
1 01
Th e Dark Kn ight System
in R.Ziatdinov-A.Chemin, N ew York
Open 1998 (see Game 71).
d) 9 'ifd2 ("Here the 'threat' of .i.e3-
h 6 is a fiction because the exchan ge of
bishops is in Black's favour. But what
other plan can White try?" - Chemin)
9 ... c6 10 .l:.ad1 'ii'a s 11 a3 .:tc8 12 b4
'ilc7 13 .i.d4 lt:lbd7 14 l:lfe1 as was
equal in P.Blatny-A.Chemin, Pardubice
1993, and ended as a draw.
9 ... .i.xf3 10 .i.xf3 c6 11 a4 as
The GM to hold this position as
White found nothing better than 12
lt:lb1 ? ! . To me, this is a good indication
of the health of Black's position.
12 ...lt:lbd7 13 g 3 was V.Arbakov-A.Cher­
nin, Bern 1995, where Black can equal-
ize with 13 ... lt:les 14 .i.g 2 lt:\c4 15 .i.d4
lt:lh s or 13 ... lt:lcs 14 .i.xcs dxcs 1 5 dxc6
bxc6, as well as Chemin's 13 ... lt:lb6 14
.i.g 2 'ilc7.
Instead, in D.Primel-A.N owocien,
French Team Championship 2007,
White tried 12 'i/d2 'ilc7 13 :ad1 lt:\a6
14 l:lfe1 lt:lb4 15 .i.d4 lt:ld7 16 .i.xg7
�xg 7, again with equality; while
R.Kashtanov-A.Lugovoi, St. Petersburg
1 02
2000, saw 12 .i.e2 'ifc7 13 f4 lt:\a6 14
.i.c4 lt:ld7 15 'ii'e 2 lt:lb4 (1S ... .i.xc3 ! ?) 16
dxc6 bxc6 17 J:tad1, and now Black took
a break from his dark-square strategy
to play 17 ... e6! 18 i.b3 d S ! .
I f there is a way t o an advantage i n
this variation, White h as yet t o fi n d it.
C: Others (without d4-dS)
In these lines White chooses to play
a regular-ish Pirc in stead of trying to
challenge the correctness of Black's
early ... lt:\c6 with s ds or s .i.bs.
There are a few choices:
(1: s h 3 1 02
C2: s il.e2 106
C 3 : 5 .i.e 3 1 08
C4: 5 .i. c4 1 09
C S : S .i.g5 1 1 2
(1: s h 3
This is the most frequently played
move here. White does not want to be
bothered by ... i.g4 or ... lt:\g4. Neverthe­
less, a tempo is a tempo, however well
motivated.

s .i.g7 6 .i.e3
...
By far the most common, but also
seen are:
a) 6 .i.g s o-o 7 'iii'd 2, when both
7 ... a6 and the surprising 7 ... ds have
worked extremely well for Black.
a1) 7 ... ds 8 exds ltJxds 9 .i.h6 (9
o-o-o ltJxc3 10 'iii'x c3 'ii'd 6 11 .i.c4 .i.fs
12 l:the1 l:t.ad8 13 'i¥e3 ltJas 14 .i.d3
.i.e6 gives White a tiny edge) 9 ... ltJxc3
(9 ... .i.e6 ! is equal) 10 .i.xg 7 �xg 7 11
'ii'x c3 'if'ds 12 o-o-o .i.e6 13 b3 .i.fs and
Black converted White's advantage
(har!) in So.Polgar-J.Fries Nielsen, Ri­
mavska Sobota 1991 (see Game 72).
a2) 7 ... a6 8 o-o-o bs (8 ... ds ! ?) 9 a3 (9
.i.d3 .i.b7 10 �b1 l:e8 1 1 l:the1 e s 12
dS ltJe7 is slightly better for White)
9 ... l:tb8 10 .i.h6 b4 11 axb4 ltJxb4 12
.i.xg7 �xg 7 13 es ltJfds was equal in
M.Yilmazyerli-D.Arutinian, I stanbul
2007, though the stronger player
(Black) went on to win (see Game 73).
b) 6 .i.e2 i s not very consistent with
5 h 3 because the pin h as been pre­
vented already - White is normally
hoping to retain the option of .i.c4.
1 e4 ltJc6 2 lDj3
Then 6 ... 0-o 7 0-0 (7 .i.e3 tran sposes to
7 .i.e2 in the notes to the main line)
7 ... es (7 ... a6 ! ? is more combative, and
was tried successfully in N.Ryba­
J.Schuyler, Washington 2012 - see
Game 74) and now:
b1) 8 dxes
Hold on ! Seriously, how do we de­
cide how to recapture in this position
(and similar positions) ?
First of all, i t is safe t o assume that
if White took once on es he will take
again given the opportunity, so we will
wind up with the same pawn structure
in either case - the only difference be­
ing the c6- and f3-knights. Do we like
having them there or not? The main
factor if the knights are on is that
White can play ltJds without having to
worry about being evicted by ... c7-c6,
but sometimes Black's ...ltJd4 is useful,
too. If the knights are off, White has the
possibility to pressure Black's centre
and activate his rook with f2-f4, but
this also frees Black's g7-bishop. In
general, the ability to play ... c7-c6 is the
most important factor, so ...ltJxes is the
1 03
Tn e D a rk Kn ig h t Sys tem
normal choice. The reader would be
well advised to take special n ote of any
exceptions.
Here 8 ....�)xe5 9 lt::lx e5 dxe5 10 .ig 5
c6 is equal, when 11 .ic4 b5 allows
Black free expansion on the queen side.
b2) 8 .ie3 is a bad version of 7 .ie2
e5 in the main line - if White is castled
kingside he has no attack to compen­
sate for his troubles; i.e. 8 ... exd4 9 lt::lxd4
.l:te8 and White already lacks a comfort­
able way to defend his e-pawn. 10 lt::lxc6
bxc6 11 .if3 is most common, when
Black plays ... .ia6, ... lt::ld 7, and can con-
sider ...l:l.b8, ...'iib 8, ...lt::le 5, and/or ...lt::lb 6.
White is equal according to Mr. H, but
Black wins most of the games. There are
no worthy examples because White
can't seem to hold onto his pieces.
b3) 8 d5 lt::l e 7 9 .ie3 is similar to the
main line - except that White is not
fast enough with his pressure on the d­
file, so 9 ... c6 is already equal, and after
10 dxc6 bxc6 11 'iWd2 fllc 7 12 .l:.ad1 d5
13 exd5 lt::le xd5 14 lt::lx d5 ? ! cxd5 Black
took over the centre in N .Jhunjh­
nuwala-S.Gligoric, Lucerne Olympiad
1982 (see Game 75).
c) On 6 .ic4 Black could proceed
"normally" with 6 ... 0-0 (cf line C41 be­
low), but the immediate 6 ... lt::l x e4! is an
equalizer: 7 .ixf7+ �xf7 8 lt::l x e4 d5 9
lt::l c 5 'i¥d6, threatening 10 ... lt::l xd4 or
10 ... e5. White can sacrifice a pawn with
10 lt::ld 3, when 10 ...lt::l xd4 11 i.f4 lt::lxf3+
12 'ii'xf3 'ii'e 6+ 13 i.e5+ �g 8 14 'ii'g 3
.ixe5 1 5 lt::lx e5 'ii'a 6 leaves h i m with
enough compensation but no more. Or
1 04
Black can decline the gift: 10 ....l:tf8 11 c3
(11 .if4 'ii'e 6+ 12 .ie5 �g8 13 c3 lt::l x e5
14 lt::l dxe 5 'ir'd6 is equal) 11 .. .<�g8 12 o-o
h6 13 l:te1 g 5, though White is slightly
better here.
d) 6 Ji.b5 looks silly since Black can
castle out of the pin, but after 6 ... 0-0 7
o-o Black has nothing better than 7 ... a6
8 .ixc6 bxc6, transposing to line A
above .
6 0-0
...
7 'i¥d2
Or 7 .ie2 e5 (7 ... a6 is more aggres­
sive, but riskier) 8 dxe5 (other moves
transpose elsewhere: 8 d5 lt::l e 7 9 fild2
is 9 i.e2 in the main line, while 9 o-o
and 8 0-0 are respectively notes 'b3'
and 'b2' above) 8 ... dxe5 (Didn't I just
say this was wrong ? - this is one of
those exceptions; actually, 8 ...lt::lx e5 is
fine too, but the text move h as per­
formed much better, so why not?) 9 o-o
flle 7 with equal chances. The point is
that 10 lt::l d 5 is not dangerous because
of 1o ... lt::lx d5 11 exd5 lt::l d 4! 12 lle1 lLlf5 !
(chasing the more dangerous bishop)
and Black is comfortably equal.

7-.eS 8 dS
Others:
a) 8 0-0-0 exd4 9 lZJxd4 l!e8 10 f3 -
3lack wins nearly every game from this
!Xlsition. Indeed, White's pawns on h3
and f3 make a ludicrous impression (he
's essentially down a full tempo in a
Philidor Defence Larsen Variation),
though White should not actually be
worse. 10 ...lZJxd4 11 .ixd4 i.e6 12 g4 (12
l.f2 a6 13 'iii>b 1 bS 14 h4 cs! was about
equal in G.Bastrikov-E.Geller, Tashkent
19 S8 - see Game 76) 12 ... cs ! 13 .ie3
IUrXd6i'
waS .li'. (1 3 ... d S I. equa1·lZeS ) 14 'B
·;M;
l.h6! i.xh6 1S 'ii'xh6 i.xa2 16 l:f.xd6 lle6
is slightly better for White) 14...l2Jxe4! 1S
fxe4 .ixc3 16 'ifxcs ? i.xb2+ 17 'it>xb2
·«xa2+ 18 'iii>c 1 l:lac8 and White soon
called it quits in D.Bescos Cortes-S.Garza
Marco, San Jose 1998. 16 bxc3? 'ii'a 3+ 17
'it>d2 l:!.ad8 wouldn't have worked either;
instead after 16 i.d3, closing the d-file,
the game still continues, though Black is
clearly better.
b) 8 dxes seems like it is headed for
dullness, but things could get interest­
ing if nobody trades those queens; e.g.
8 ... l2Jxes 9 lZJxes dxes 10 o-o-o .ie6 11
g4 c6 12 g S (12 'ifxd8 .Ufxd8 13 �xd8
l:i.xd8 14 Si.xa7 .ih6+ 1S .ie3 .ixe3+ 16
fxe3 hs 17 g s lLih7 18 h4 f6 19 gxf6
lL\xf6 is equal despite the slight pawn
deficit: Black's activity and future
passed h-pawn are sufficient) 12 ...tZ:'ld7
13 h4 'it'as and the race is on. Black has
no reason to be pessimistic about his
prospects.
8 tZ:'le7
•.•
1 e4 lL\c6 2 lZJf3
8 ... l2Jb8 is also quite reasonable. The
e7-knight supports the ... c7-c6 break,
. .l (14 but with White likely to castle long and
form a battery on the d-file, the useful­
ness of this option is reduced.
9 0-0-0
Instead:
a) 9 .ih 6 is statistically the most
dangerous, scoring 6/7 for White, but
Black was heavily outrated and his play
quite uninspired. Artashes Minasian
shows us the way in D.Saduakassova­
Art.Minasian, Dubai 2011: 9 ... .id7 ! 10
.id3 (10 o-o-o? ! b S ! ) 10 ... c6 (or 1o ... bs
11 a3 aS 12 0-0 b4 intending ... C7-C6
with counterplay) 11 dxc6 .ixc6 (taking
firm control over the dS-square; back­
ward pawn ? what backward pawn ?) 12
.ixg 7 'it>xg 7 13 o-o "fic7 (13 ...bs ! ?) with
easy equality (see Game 77).
b) 9 Si.e2 .id7 (9 ... lZJxe4 ! ? 10 l2Jxe4 fs
11 lLic3 f4 12 o-o-o is slightly better for
White) 10 g4 (not 10 .ih 6 ? ! c6 ! 11 dxc6
.ixc6 with advantage) 10 ...b s ! 11 g S
(11 Si.d3 ? ! b4 12 tZ:'le2 c 6 13 dxc6 .ixc6
14 tZ:'lg 3 lZJxe4! is good for Black; while
11 a3 l:tb8 is equal, since 12 .ixa7? runs
1 05
Th e Dark Kn ight System
into 12 ... .l::t a 8 ! 13 j_e3 b4) 1 1 ... b4 12
gxf6 bxc3 13 �xc3 j_xf6 with a level
position.
c) 9 .U.d1 tt:'lh s 10 g4 lL'lf4 (this ag­
gression is called for because White
can no longer tuck his king away on the
queenside) 11 j_xf4 exf4 12 �xf4 f5 13
exfs gxfs 14 g s j_xc3+ 15 bxc3 .l:te8 16
j_e2 lbg6 with compensation .
d ) 9 g4 a 6 should transpose t o the
main line once White castles long.
9 a6 10 g4 bS 11 gS
...
11 ..td3 j_b7? ! 12 lbe2 ? ! (better 12
a3 or 12 gS with an edge) 12 ... c6?! was
equal in K.Haznedaroglu-T.Gelashvili,
Antalya 2009; but Black could have
blown it open with 11 ...b4 12 lbe2
ltJexd s ! 13 exd5 e4 14 i.xe4 lbxe4 15
'ifxb4 l:.e8, when White has a great
deal to worry about for his extra pawn.
In the actual game 12 ... ltJfxdS ! would
have been even stronger, giving Black a
comfortable advantage, but White
need not have allowed that.
11 lL'lhs
•••
A race scenario is developing in
which Black's chances are n o worse -
1 06
White has a hard time evicting the h s ­
knight, and will also fi n d i t hard to
profit even if the h-file opens. One con­
tinuation of many: 12 ..ti>b1 ..td7 13
lL'lh 2 ltJc8 ! ? 14 lL'lg4 lL'lb6 15 b3 b4 16
lbe2 as, still with approximately even
chances.
C2: 5 j_e2 i.g7 6 o-o o-o
Azmaiparashvili plays 6 ... i.g4 here,
making sure White can't change his
mind and play 7 ds and 8 h3. Black may
even have considered s ... ..tg4! ?. lt is n ot
clear which is the most accurate. Be­
sides, why should White change his
mind?
7 i.e3
White tries to make do without h2-
h 3 . I s he inviting trouble or saving a
tempo? The alternative is 7 j_g s h 6 and
then :
a) 8 ..th4 g 5 9 j_g 3 lL'lh s 10 ds lL'lxg 3
1 1 hxg3 lL'le5 is equal; e.g. 12 lbd4 c6 13
lL'lfs j_xfs 14 exf5 'iib 6 1 5 .l:.b1 lbd7 16
j_f3 l:Ife8 17 g4 lL'lf6.
b) 8 j_e3 lL'lg4 9 j_d2 eS 10 dS lL'le7
11 h3 lL'lf6 12 ..te3 lL'ld7 13 'i!Vd2 'it>h 7 14

lbe1 fs most closely resembles a King's
Indian, where White's attack will be
greatly delayed by his pawn stuck on
the C2-square.
c) 8 il..f4 lbg4! 9 h3 (if White does
not play h 2-h 3, the game will transpose
to note 'b' just above; e.g. 9 dS es 10
.i.d2 lbe7, or 9 .id2 es, or even 9 .ic1
es 10 ds etc) 9 ... es 10 dxes ttJgxes 11
lbxe s ? ! (11 i¥d2 is better, retaining a
tiny edge after 1 1 ...c.t>h 7; Black can con-
sider ... lbxf3+ and ...lbd4, or ... .i.e6, or
.. .f7-f5, or ... l:te8) 11 ... dxes was level in
A.Sakharov-A.Adorjan, Sochi 1976,
though Black went on to win a wild
game (see Game 78).
7 es!
•••
As usual, 7 ... a6 can be tried - the
main line is a bit drawish - but then 8
dS lbb8 9 a4 is an excellent answer.
Alternatively, 7 ... il..g 4 8 dS lbb8 trans­
poses to line B2 (8 ... .ixf3 9 .ixf3 lbes
10 �e2 c6 is a main line Classical Pirc
which will not be covered).
8 dxes
Or:
a) 8 dS lbe7 9 1i'd2 lbg4 10 �g s h 6
11 il..h 4 g s 12 .ig 3 f s 13 h 3 lbf6 1 4 exfs
lbxfs 15 .ih 2 'ii'e 8 with a tiny edge as
Black considers a kingside attack based
on his space advantage there, or the
... eS-e4-e3 lunge, or ...'ii'f7 and ... lbe7,
building pressure on White's d-pawn.
White is not well situated to use his
asset - the e4-square.
b) 8 i¥d2 (as usual, it is a bad idea
for White to try to maintain the tension
- this only works if Black is not h appy
1 e4 lbc6 2 lbf3
to take! ) 8 ... exd4 9 lbxd4 .l::!. e 8 10 f3 (10
lbxc6 bxc6 11 f3 dS tran sposes;
11 ...'i*'e7 is also fine) 10 ... d s ! 11 lbxc6
bxc6 12 llad1 'i¥e7 13 il.. d4 dxe4 14
fxe4 lbxe4 15 lbxe4 'ii'x e4 16 .ixg 7
c.t>xg 7 17 .if3 i¥e3+ 18 'it'xe3 .l:!.xe3 19
.ixc6 l:f.b8 20 �d8 �b6 with a level
endgame.
s lbxes
...
8 ... dxes is also playable. Z.Bratanov­
B.Chatalbashev, Bulgarian Champion­
ship 2004, continued 9 'ii'x d8 �xd8 10
il..c4 h6 11 h3 b6 12 lbds lbas ! 13
lbxf6+ �xf6 14 .ie2 .ib7 1 5 b4 lbc6 16
c3 as 17 a3 lbe7 18 lbd2 .ig S ! and
Black went on to win (see Game 79).
g lbxes dxes 10 'i!Vxd8 J:!.xd8
In practice this is equal, but White
can try to extract a little something:
a) By far the most common is 11
�fd1 (or 11 .laad1 - it doesn't much
matter) 11 ... �e6 12 lbbs, when the
new move 12 .. J:tdc8 ! keeps things level;
e.g. 1 3 lbxa7 .l:.xa7 14 �xa7 b6 1 5 .l:.d3
.l:Ia8 16 .l:.a3 .if8 17 :a6 �c8 18 :a4
.id7 etc.
b) 11 lbbs .id7 12 f3 looks scary,
107
� ,.., e :J a rtc Knig h t System
and both 12 ...b6 and 12 ... ..1i.xb5 give
White something to work with. The
novelty 12 ... a6 13 ti:Jxc7 llac8 14 i.b6
i.c6 1 5 li:Jxa6 l:!.d2 16 .itd3 bxa6 17 .ite3
llxd3 18 cxd3 is a tiny edge for White,
but the imbalances should provide
Black some winning chances as well.
c) 11 ..tc4 c6 12 ltad1 l:f.e8 1 3 a4 i..f8
14 f3 Wg 7 1 5 .l::t d 2 i..b 4 16 J::f.fd1 :e7 17
�f2 l:td7 and White is running out of
things to play for.
C3: 5 i.e3 i.g 7 6 'Yi'd2
Other moves are covered elsewhere:
6 d5 ti:Jb8, 6 h3 o-o, and 6 .lte2 0-0 7 o-o
were seen in lines B, C1 and C2 respec­
tively, while 6 ..li.c4 is C43 below.
6 0-ol
...
6 ... li:Jg4?! should not work here. 7
i..f4 e5 8 i.g 5 f6 9 i.h4 o-o 10 i.. c4+
�h 8 11 d5 ti:Je7 12 h 3 ti:Jh6, intending
13 ... ti:Jhg8, 14 ... h6 and 15 .. .f5 is satisfac­
tory for Black, but after 7 i.g 5 h6 8 i..h 4
the bishop chase is n ot payin g divi­
dends and White win s nearly every
game.
7 ..li.h6
1 08
7 h 3 is C 1 again, and 7 d5 ti:Jb8 8 h 3
c6 is i n the notes t o line B. Others:
a) 7 o-o-o h as not scored well, but
White's set-up is challenging - he h as
man aged to omit h 2-h 3 and stop ... e7-
e5, while 7 ... li:Jg4? ! 8 ..li.f4 e5 9 dxe5
ti:Jgxe5 10 ti:Jxe5 lt:\xe5 11 h4 is still not
convincing.
However, White h as committed his
king, so: 7 ... a6 ! 8 i.h 6 bS 9 ..li.xg 7 (9
..li.d3 allows 9 ... e 5 !) 9 ... �xg 7 and the
annoying threat of 10 ...b4 already
forces some sort of concession (such as
the weakening 10 a3 or an awkward
defence of the e-pawn). White is only
slightly better.
b) 7 i.e2 e5 (7 ... a6 8 d5! ti:Jb8 9 a4
was K.Wang-J.Schuyler, Washington
2012; I felt my a-pawn was misplaced,
and I did go on to lose, but I was n ot
without chances - see Game 80) 8 dxe5
(or 8 o-o-o li:Jg4 9 dxe5 dxe5 10 'Yi'xd8
ti:Jxd8 11 ti:Jd5 ti:Je6 with an even posi­
tion, while 10 'Yi'e1 ! ? li:Jd4 11 �b1 c6 12
..li.xd4 exd4 13 h3 'iif6 is unclear)
8 ... dxe 5 is equal. With White's queen
on d2 and his passive bishops, this is

not the time for the exchange varia­
tion, and 9 o-o-o 'Wie7 (or 9 .. .tL'lg4) 10
�g 5 �e6 11 lt:Jd5 ? ! i.xd5 12 exd5 .t'!.fd8
just makes matters worse for White.
c) 7 �c4 lt:Jg4 8 i.g 5 h6 9 i.h4 g5 10
�g 3 e5 11 d5 lt:Jd4 is about equal, but
not a very rational position - the tac­
tics would take pages. Instead,
7 ... i.g4!? keeps things under control;
e.g. 8 d5 i.xf3 9 gxf3 lt:Je5 10 i.e2 c5 11
o-o-o 'i!Va5 12 ..t?b1 l:tab8, or 8 o-o-o
tt:Jxe4 9 tt:Jxe4 d5 10 �d3 dxe4 11 �xe4
li'd7 12 d5 �xf3 13 �xf3 tt:Je5 14 �e2
lt:Jg4 15 �d4 e5 16 dxe6 'i!Vxe6 17 �xg 7
'it>xg 7 with near equality.
7 esl
...
Centre play beats wing play.
8 �xg7 '>t>xg7 9 0-0-0
After 9 d5 lt:Je7 10 o-o-o .:tb8 11 �d3
(if 11 h4 i.g4, or 11 '>t>b1 b5) 11 ... b5,
11 ... c6 or 11 ... �g4, Black h as full coun­
terplay.
9 �g4 10 dxes dxes
...
This position has led to three draws,
but Black h as a slight advantage based
on White's weak bishop. For instance,
11 'ili'e3 lt:Jd4 12 lt:Jb5 ? ! was F.Saez-
1 e4 ltJc6 2 lLlf3
A.Safranska, Grenoble 2003, when
12 ... c5 13 c3 'i!Va5 ! 14 lt:Ja3 (14 cxd4?
cxd4 15 'ii'a 3 .l:.ac8+ 16 Wb1 'ifxa3 17
tt:Jxa3 lt:Jxe4 18 .l:te1 lt:Jxf2 19 l:f.g 1 �xf3
20 gxf3 d3 21 :1xe5 d2 22 �e2 .l:tfe8 2 3
lt:Jc4 b5 is nearly winning) 1 4. . .�xf3 1 5
gxf3 lt:Je6 would h ave given Black a
large positional advantage.
C4: 5 �C4 � g 7
And n ow:
(41: 6 0-0 110
C42: 6 1i'e2 110
C43 : 6 � e 3 112
Instead:
a) 6 i.g 5 ? ! tt:Jxe4! 7 tt:Jxe4 (or 7
�xf7+ '>t>xf7 8 t2Jxe4 d5 9 lt:Jc5 lif8)
7 ... d5 8 c3 ! ? dxc4 9 d5 f5 ! 10 dxc6
'ili'xdl+ 11 ltxd1 fxe4 12 lt:Jd2 bxc6 13
tt:Jxe4 l:.b8 is good for Black.
b) 6 �f4 - As a novice, I remember
thinking that this was some kind of
ideal position for White. N owadays, it
looks like White is begging for trouble,
as the bishops are vulnerable and do
1 09
Th e Dark Kn ight System

not defend his centre. Indeed, my re­ only a little worse for White) 9 ... dxe4 10
search uncovered a large collection of i.. x e4 l2Jxd4 is clearly unsatisfactory for
games by some of the lowest-rated White.
players I have ever seen in any data­ 8 ...l:txf7 9 tt::lxe4 ds!
base. - 6 ... tt::l xe4 (6 ... i.. g 4 is m ore com­
bative) 7 tt::lxe4 (7 i.. xf7+ <tlxf7 8 tt::lx e4
:fs 9 dS 'it>g 8 ! 10 'iVd2 'ii'd7 ! ! 11 dxc6? !
�g4! o r 1 0 i.. g s tt::l e s with a tiny edge)
7 ... ds 8 i.. d 3 dxe4 9 i..xe4 l2Jxd4 10
tt::l xd4 'i!i'xd4 11 'it'xd4 i..xd4 12 i..x c7
i.. xb2 13 l:tbl fS 14 i..f3 i.. d4 with a
level endgame.

(41: 6 0-0

White has four knight retreats and

they're all bad - he h as reached an in­
ferior version of 5 h 3 i.. g 7 6 i..c4 tt::l x e4!
etc in line Cl (see note 'c' to White's 6th
move). The most frequent is 10 tt::l c s ? !
and now, rather than 1 0. . .1\Vd6? ! a s al­
ways played, 10 ...b6! 11 tt::lb 3 'iid 6 12
tt::l g s l:tf8 13 .:te1 es gives Black a com­
fortable advantage.

6 ... 0-0 C42: 6 'ii'e 2 i..g4!

6 ... i.. g 4 is also good; e.g. 7 i.. e 3 0-0 8
h 3 i.. xf3 9 'ti'xf3 e s 10 dxes (or 10 ds
tt::l e 7) lO . . tt::l x es 11 'i!Ve2 tt::lx c4 with
.

equality.
7 h3?!
This is no good, but it is the most
common, and other moves are either
met by ... i.. g 4 or else are covered else­
where: 7 dS tt::lb 8 is line B again, while 7
i.. e 3 tt::l g 4 is C43 below.
7 ..tt::lxe4 8 i..xf 7+
.

8 tt::l x e4? ! dS 9 i.. d 3 (9 c3 ! dxc4 is 7 i.. e 3

110

GM Robert Zelcik has twice played 7
es and won (against much weaker op­
ponents). If Black takes the bull by the
horns with 7 ... .i.xf3 8 gxf3 liJxd4 9 exf6
ttJxe2 10 fxg 7 l:tg8 11 �xe2 ! ? (or simi­
larly 11 .i.xe2 - White is 2-0 here as
well) 11 ...l:i.xg 7 we reach this mess:
Queen and two pawns is a lot for
three minor pieces, even three pieces
that include the bishop pair. Is there
really no way to use the material ? The
key to the position is to make sure the
pawn s stay mobile. lt is not bad to sac­
rifice a pawn if it helps to open files,
trade rooks, disorganize the white
pieces, or expose the enemy king. A
close look at the games shows that one
of the candidate masters was crushing
the GM, so we should n ot despair, and
instead resolve to emulate Black (up to
a point) in R.Zelcik-M. Djurkovic, Pula
2001 (see Gam e 8 1).
Anyway, if this is not to your taste,
Black can chicken out without adverse
consequences; e.g. 7 ... dxes 8 dxes .i.xf3
9 gxf3 tLlh s 10 f4 .lth6 which is equal,
though still not simple.
1 e4 ltJc6 2 tLlf3
7 es 8 dxes
...
8 d S ? ! tLld4! 9 .i.xd4 (or 9 'i¥d1 .ltxf3
10 gxf3 o-o) 9 .. exd4 10 tLlbs (10 e s ?
.
0-0! 1 1 exf6 dxc3 wins) 1 0. . .0-0 11 0-0
l::te 8 was much better for Black in
M.Strubreiter-K.Rogetzer, Austrian
Team Champion ship 2004, and after 12
liJbxd4? ! ltJxe4 13 �d3 ttJcs 14 �d2
Black should have cashed out: 14 ....txf3
15 ./tJxf3 .i.xb2, with a squeaky-clean
extra pawn to go with his positional
advantages.
8 ./tJxes 9 .itb3 o-o 10 o-o-o
...
10 h3 .i.h s 11 o-o-o as simply
transposes.
1o a s
...
This is a new move. Black has equal­
ized and should start thinking about
how to exert the most pressure. White
will n ot enjoy playing a2-a4, but it is
necessary: 11 h3 .th s ! 12 a4 (not 12
g4? ./tJfxg4 13 hxg4 .i.xg4 and
14 ... .txf3) 12 ..../tJxf3 13 gxf3 tLld7 14
1:1hg 1 ttJes 1 5 .l:r.g 3 c6 is still equal, Then
16 .ltcs .ith6+ 17 .ite3 ! .ltg7 repeats,
since 17 �b1 ? dxcs ! 18 l:txd8 l:Iaxd8
should terrify White.
111
The Dark Knight System

C43: 6 i..e 3 o-o tural advantage with ... .ixc3) 11...i.. xh6
12 tZ'lf3 c6 and White is only slightly
worse.
9 ... gs
This potentially king-weakening
move is acceptable because White has
also castled short.
10 i.. g 3 tZ'lf6

7 0-0
7 'ii'd 2 transposes to 7 .ic4 in line
C3 (see note 'c' to White's 7th m ove).
1 .. .tbg4 s .igs
In stead:
a) 8 .if4?! tbxd4! (fork tricks every­
where !) 9 tbxd4 es 10 .ie3 tbxe3 11
fxe3 exd4 12 exd4 .ie6 was H .Hughes­ Chances are equal; e.g. 11 h 3 tbxe4,
K.Richardson, British League 2004. or 11 .ib3 .ig4 12 dS tbas 13 h3 .ih s
Black had the edge and went on to win 14 'ii'd 3 .ig6.
(see Game 82), although 12 ... c6! was
simpler with a comfortable advantage. CS: 5 .igS .i g 7 6 11i'd 2 h61
b) I don't know who would play it,
but Houdini likes 8 .ic1, when 8 ... es
allows White a small advantage after 9
.ig s ! i..f6 10 .ixf6 tbxf6 11 ds tbb8.
Instead, 8 ...tbf6 returns to 6 ...0-o in line
C41 and offers (or bluffs) a repetition.
s ... h6 9 .ih4
On 9 .if4?! tbxd4! is best, as in the
previous note, even though White has
the extra possibility 10 tbxd4 es 11
.ixh 6 ! (after 1 1 tbe6 ! ? fxe6 12 'ii'x g4
exf4 13 i..x e6+ �h 7 14 .ixc8 'ir'xc8 This bishop hunt does work.
Black's active bishop allows him a small 7 .if4
plus - which he m ay convert to a struc- Others:

112

a) 7 i.h4 g 5 8 �g 3 ltJh 5 9 d5 t2Jb8 (it
is good form to delay ... t2Jxg 3 until the
last moment, even without any specific
idea in mind) 10 l2Jd4 c5 1 1 �b5+ 'ifi>f8
was equal in A.Grilc-G.Mohr, Slovenian
Team Champion ship 2008, though the
better player (Black) went on to win
because White could not control the
dark squares (see Game 83).
b) 7 �e3 l2Jg4 (8 �f4 t2Jxd4! 9 t2Jxd4
e5 equalizes; e.g. 10 h3 exd4 11 l2Jb5
l2Je5 12 t2Jxd4 'Wie7 13 �b5+ �d7 14
�xd7+ t2Jxd7 1 5 0-0-0 0-0-0) 8 0-0-0
t2Jxe3 9 'i!Vxe3 0-0 10 h3 a6 11 g4 b5 was
equal in D.Janowski-F.Yates, Marienbad
1925 (!), a wild game that ended in a
draw (see Game 84).
1 ds1
...
Abruptly, Black aborts the bishop
chase in order to seize this opportunity
for an unexpected central thrust. Thi s
is a new move.
B exd s
Or 8 �b5 t2Jxe4 9 t2Jxe4 dxe4 10 l2Je5
�d7 11 �xc6 �xc6 12 t2Jxc6 bxc6 13
0-0-0 'i'd5 14 'ifi>b1 0-0-0 1 5 C3 g 5 16
1 e4 lL\c6 2 ltJf3
�e3 f5 17 g 3 e5 18 'i'e2 �5 19 'iVxb5
cxb5 20 dxe5 i.xe5 2 1 �xa7 'it>b7 with
some chances in the endgame for Black
due to his extra space, superior bishop,
and "queenside" (i.e. away from the
kings) pawn majority.
s ttJxd s 9 ttJxd s 'Y.Ii'xd s 10 c3
...
10 �xc7 is too greedy; e.g. 10 ...�g4
11 c3 (11 i.e2 .l:.c8 12 i.g 3 �xf3 13
i.xf3 'i!Ve6+ 14 ii'e3 t2Jxd4 1 5 �xe6
t2Jxf3+ 16 gxf3 fxe6 is equal) 11...�xf3
12 gxf3 :l.c8 13 �f4 'i!Vxf3 14 .l:tg 1 e5 1 5
dxe5 �xe5 16 �xe5 t2Jxe5 17 i.b5+
'ifi>e7 and Black h as slightly fewer king
problem s than White.
10 �g4 11 �e2 o-o-o
...
Black h as reached a very satisfactory
cross between a Modern and a Scandi­
navian, and has excellent activity to
compensate for White's extra pawn
presence in the centre, with ... g6-g 5
and/or ... e7-e5 to follow. Now 12 o-o
e5?! 13 t2Jxe5 doesn't quite work be­
cause the g4-bishop is h anging, but
12 ... �h 5 ! threatens both 13 ... e5 and
13 ... g 5 14 �g 3(?!) f5, with the initiative.
113
 C h a pte r S ix
1 e4 l2Jc6 2 l2Jc3
To play the Dark Knight against
both 1 e4 and 1 d4 (as opposed to just
against 1 d4) requires some plan for
dealing with this relatively rare move.
As with 2 lLlf3, White h opes to re­
turn to familiar territory with 2 ... e5,
and while this is fine theoretically,
learning the Spanish Four Knights, Bel­
grade Gambit, Vienna Gam e, Vienna
Gambit, and Scotch Four Knights is not
a practical answer to an "uncommon"
variation.
Another possibility is 2 ... e6 3 d4 d5,
transposing to a French sideline cov­
ered in Wisnewski's Play 1 . tLlc6!.
. .
114
Ideally we should find something
that resembles - and is likely to trans­
pose into - positions we already know,
but we need to be careful of our move
order because the Dark Knight Pirc is
not well suited for generating counter­
play against the Argentinean Attack.
(For instance, 2 ... g6 3 d4 i.g7 4 ..te3 d6
5 'ii'd 2 tt:Jf6 6 f3 e5 7 tt:Jge2 is tricky for
Black at best, though 7 ... a6 8 0-0-0 b5 is
a decent try.)
2 tt:Jf6 3 d4
..•
Others:
a) 3 f4 d5 4 e5 d4 5 exf6 (5 tt:Jce2?!
d3 ! 6 cxd3 tLld5 7 a3 i.g4 8 'iib 3 liJb6 is
a mess for White) 5 ... dxc3 6 fxg7 cxd2+
7 'ilt'xd2 'ii'x d2+ (or 7 ... i.xg 7 8 'ifxd8+
tt:Jxd8) 8 i.. x d2 i.. xg7 is just equal.
b) 3 lLlf3 d6 followed by ... g 7-g6 will
be a Dark Knight Pirc.
3 d6
...
We intend to fianchetto but we are
not yet committed, so it is impossible
for White to start an Argentinean At­
tack. The most common reply by far is 4
tLlf3, transposing to Chapter Five.

Here, we will take a close look at:
A: 4 f4 115
8: 4 d 5 11 7
But first:
a) 4 i.e3 makes no sense because of
4....�:Jg4!. If 5 .ig 5 h6 6 .ih4 g 5 7 .ig3
i.g7, we have gotten our fianchetto
with gain of time, reaching positions
similar to Chapter Five. While this does
not actually give easy equality, it is cer­
tainly satisfactory - 4 .ie3 is almost
never played.
b) 4 f3 e5 5 lt:Jge2 exd4 6 lt:Jxd4 .ie7
is a kind of Philidor where White's f­
pawn does not belong on f3. White
normally continues in "Argentinean"
style with 7 .ie3 o-o 8 'ifd2, when Black
has 8 ...lt:Jxd4 (8 ... d5 ! ? 9 o-o-o dxe4 10
lt:Jxc6 'ii'x d2+ 11 .l:.xd2 bxc6 12 lt:Jxe4
lt:Jxe4 13 fxe4 offers a tiny endgame
advantage for White) 9 .ixd4 c6 10
o-o-o and n ow 10 ...b5 with an equal
game (though perhaps we should call
such a position unclear), or 10 ... .ie6,
which transposes to A.Mista-M.Szelag,
1 e4 lt:Jc6 2 lt:Jc3
Koszalin 1999 (see Game 85).
c) 4 i.g 5 h 6 ! (Mestrovic's move,
which he has had a chance to use four
times, with two wins and two draws
against very strong opposition) 5 .ih4
(this doesn't seem con sistent, but it is
the most common; also, if 5 i.xf6 exf6 !
6 d5 lt:Je7, White has little to combat
the incredibly slow but effective plan of
7 ... a6, 8 ... g6, 9 ... .ig7, 10 ... 0-0 and 11 .. .f5,
with approximate equality) 5 ... g 5 6
i.g 3 .ig 7 (6 ... e 6 ! ?) is also equal. White
tried 7 h4 g4 8 h 5 in I .Jakic-Z.Mestrovic,
Zadar 2001 (see Game 86), where 8 ... e5
9 d5 lt:Jd4 was quite sufficient; while
R.Zelcic-Z.Mestrovic, Nova Gorica 2003,
saw 7 f3 o-o 8 .if2, but Black used the
same idea, taking advantage of White's
slow play: 8 ... e5 9 d5 lt:Jd4! 10 lt:Jge2 c 5 !
11 dxc6 bxc6 12 lt:Jxd4 exd4 13 .ixd4
and although Mestrovic was successful
with 13 ... c5?! (see Game 87), 13 ....l:.b8 ! is
stronger, with excellent compensation .
A : 4 f4 e5
5 dxes
other moves don't offer much :
115
Th e Dark Kn ig h t System
a) 5 lt:Jf3 exd4 6 lt:Jxd4 is some sort
of Philidor where White's 4 f4 is prema­
ture.
Black must be alert in order to prove
this - and it is worth noting that, until
n ow, he has not been up to the task:
6 ... �e7 7 .i.e2 o-o 8 .i.e3 lt:Jxd4 9 .i.xd4
and here the new move 9 ... d5 ! 10 .i.xf6
.i.xf6 11 'ii'x d5 .i.xc3+ 12 bxC3 1i'f6 13
'ii'd 2 �d8 14 .i.d3 (14 'ii'e 3 is met by
14 ... 'ifh4+ 15 'ii'f2 1i'f6 16 e5 \i'c6, or 1 5
g 3 'ifh 3 and White h a s some trouble
getting his king safe and activating his
h 1-rook; e.g. 16 .i.f3 b6 17 e 5 11b8 18
'ili'e2 .i.b7 19 .i.xb7 .l::t xb7 20 :d1 l:tbb8
with compensation) 14...'ii'b 6 15 1Wf2
'ii'b 2 16 0-0 1Wxc3 17 e5 1i'd4 18 1Wxd4
�xd4 19 �ad1 �f8 20 c3 �d7 2 1 .i.e4
�b8 and Black unravels safely.
b) 5 d5 lt:Je7 6 lt:Jf3 (6 fxe5 dxe5 7 .i.e3
lt:Jg6 8 a3 .i.d6 is fine for Black; or 6 f5
c6 ! 7 dxc6 bxc6 8 lt:Jf3 1i'c7 with usually
... d6-d5 coming soon, and sometimes
...�b8 and/or ... g7-g6) 6 ...exf4 7 .i.xf4
lt:Jg6 8 .i.g 3 .i.e7 9 'ii'd 2 0-0 10 0-0-0 lt:Jd7
intending ... .i.f6, ... l:te8, ... a7-a6, ... b7-b5,
...lt:Jc5, ....i.b7. White's h 2-h4 must be
116
met by ...h7-h6 so that the enemy pawn
does not reach the h6-square. The posi­
tion is equal.
c) 5 fxe5 dxe 5 6 d5 lt:Jd4! (6 ...lt:Je7
transposes to 6 fxe5 in n ote 'b') 7 lt:Jf3
.i.c5 8 lt:Ja4 ! ? (or 8 .i.g 5 o-o 9 'ii'd 3 1Wd6
and since 10 o-o-o? ! lt:Jg4! is strong,
White is clearly worse) 8 ...lt:Jxe4 9 lt:Jxc5
lt:Jxc5 10 lt:Jxd4 'ii'h 4+ 11 g3 1Wxd4 12
1Wxd4 exd4 and although White's bish­
ops are nice, it is not easy for him to
recover the pawn with a good position
- Black is a little better.
d) 5 .i.b5 ? ! exd4 6 'ii'xd4 .i.e7 7 lt:Jf3
0-0 8 .i.xc6 bxc6 9 o-o l:te8 10 .:td1 .i.b7
with two bishops and a big bull's-eye
on the e4-pawn - more than enough to
make up for our space disadvantage.
s ... dxes
Black often throws in 5 ... .i.g4 6 lt:Jf3,
but this does not help the situation.
6 1Wxd8 �xd8 7 lt:Jf3 �b4 8 .i.d3 .:te8
This position has been reached only
once: in Ma.Tseitlin-C.Barlocco, World
Seniors Championship, Port Erin 2004,
which continued 9 o-o exf4 10 lt:Jd5
lt:Jxd5 11 exd5 lt:Je7, and White went on

to win . However, Black can equalize
with 9 ... h6! 10 fxes ltJg4 11 h3 (or 11
l:td1 ..tcs+ 12 �f1 ..td7) 1 1 ... ltJgxes 12
..tf4 (or 12 ..tbs ..td7 13 ttJds ltJxf3+ 14
gxf3 ..td6) 12 ... ltJxd3 13 cxd3 �e6.
B: 4 d S liJbB
Don't get any funny ideas. 4...ttJes ? !
5 f4 ltJed7 6 liJf3 c 6 7 dxc6 bxc6 8 es i s
no good for Black.
After 4 ... ltJb8, White will usually
play liJf3 soon (and Black ... g 7-g6),
tran sposing to line B in Chapter Five.
Here we investigate some independent
set-ups by White.
5 f4
The most often played, but it is not
a strong move. The problem is that,
after s ... c6, White will be unable to
maintain the e4/f4 pawn duo, unless
he plays the otherwise undesirable 6
dxc6. In stead:
a) 5 �g 5 g6 6 �xf6 exf6 7 .td3 (to
hinder .. .f6-f5) 7 ... .tg 7 8 �d2 o-o 9
o-o-o ltJd7 10 h4 ltJc5 11 h 5 f5 with full
counterplay.
b) 5 .te3 (this logical move has h ad
1 e4 ltJc6 2 ltJc3
only one obscure trial) s ... c6 6 "i¥d2
ltJbd7 7 o-o-o cxds 8 exd5 a6 should
make for a lively game, with ...b7-b5
and ... ..tb7 coming, and probably ... g7-
g6 and ...�g7, with maybe ....:c8, ... "iic 7
and/or ...ltJb6. White will be unable to
organize the rapid pressure on the e­
file that would make this uncomfort­
able for Black.
s c6
...
6 dxc6
Other moves:
a) 6 liJf3 cxd5 7 exd5 g6 8 ..te3 .tg 7
is equal . Black will continue with ... 0-0,
... ltJbd7, .. ."ilie7, ... a7-a6, ...b7-b5, ... ltJc5
(or ... ltJb6), and ... �b7 (or ... �d7). Since
White's central pawns are split, the f­
pawn is mainly a weakness. Then 9
.te2 o-o 10 0-0 ltJbd7 transposes to
M.Drasko-Z.Mestrovic, Bosnian Team
Champion ship 2003 (see Game 24).
b) 6 .te2 g6 will probably transpose
to 6 liJf3, after 7 liJf3 cxds 8 exd5 �g7 9
.te3 for instance. Instead, A.Stefanova­
M.Zielinska, Dresden 2004, continued 7
.te3 'i!Va5 8 .tf3 ! ? ..tg 7 9 ltJe2 cxds 10
exd5 ..tg4, which was about equal, but
117
Th e Dark Knig h t System
Black did go on to lose. He can improve
slightly by 7 ... i.g 7 and 8 ... 0-0, rather
than committing the queen so early.
6 ... tZ:lxc6 7 tLlf3 g 6
8 i.c4
Alternatively:
a) 8 e 5 ! ? is a good try, though it has
never been played: 8 ... tLlh 5 (or 8 ... tZ:ld7 ! ?
9 exd6 i.g 7 ! 1 0 dxe7 'it'xe7+ 11 'i¥e2
tZ:lb4 12 1Wxe7+ <Ji;xe7 13 'itd1 l:td8 with
compensation) 9 exd6 i.g 7 10 i.c4 0-0
11 0-0 exd6 and White's holes balance
Black's holes.
b) 8 i.e3 i.g 7 9 'ifd2 o-o 10 o-o-o
tZ:lg4 11 i.g 1 'ii'a 5 12 h 3 i.xc3 13 'ii'x c3
118
'ii'x c3 14 bxc3 tZ:lf6 with slightly the
more comfortable g ame for Black.
8 ... i.g7 9 o-o o-o 10 'ith1 i. g4 11 i.e3
.l:tc8 12 i.b3
The position is equal, but Black
needs to find something more active
than 12 ... a6? ! 13 h3 i.d7 14 ii'e1 b5?! 15
e5!, after which White was much better
in N.Raghavi-K.Szczepkowska, World
Junior Championships, Istanbul 2005.
Black can choose between 12 ...tZ:ld7,
12 ...'it'd7, 12 ... tLlh 5, and 12 ... tZ:la5 (since
the complications after 13 i.xa7? ! b6
14 e5 tZ:ld7 ! 15 h 3 i.. xf3 16 'i!Vxf3 tZ:lc6 17
tLlb5 dxe5 favour Black).
 Section Th ree
Others
1 c4 - Chapter Seven
1 tLlf3 - Chapter Eight
Others - Chapter Nine
Miscellaneous Topics - Chapter Ten
This section deals with almost eve­
rything except the opening moves 1 e4
and 1 d4, including some quite un­
usual openings. Except for 1 c4, which
needs to be taken seriously, the chal­
lenge for Black is to give himself
chances to wrest the initiative early,
and/or make sure that White does not
reach the type of position with which
he is experienced and comfortable. In a
few cases, I am forced to admit that
1 . . ltJc6 is not suitable, in which case I
.
like 1 ... g6. I choose to believe that this is
not too much of a departure, since the
fianchetto is so common in the Dark
Knight System.
I will warn the reader that the cov­
erage in section 3 is less detailed than
in the other sections. This is because
these openings are far less common,
and also because you may already have
your own systems of defence, which
there is n o need to abandon.
119
 C h a pter Seve n

1 c4 ltJc6

lt:lc3 and now Black can certainly play

You could play the Dark Knight
against 1 e4 and 1 d4 and then play
5 .. .f5, transposing to line B1; but 5 ... d5!
is even better, as in Ge. Lambert­
A.Labarthe, Vichy 2000. Not without
reason, White worried about the centre
opening up and played 6 c5?, when
6 ... h6 7 lt:lh 3 �xh 3 or 6 ... b6! would have
been strong; but even after 6 a3 �xc3 7
bxc3 lt:lf6 8 cxd5 1Vxd5 9 f3 o-o 10 fxe4
lt:lxe4 1 1 lt:lxe4 'i!Vxe4 12 1t'd3 �f5 Black
maintains an advantage along with his
grip on the e4-square.
2 e5
...

any favourite system against 1 c4, but

it makes some sense to use 1 ...lt:lc6 here
too, since there is a fair chance of
reaching positions you already need to
know to meet 1 d4.
2 lt:lc3
Others:
a) 2 d4 e5 returns to Chapter Two.
b) 2 g 3 e5 3 �g 2 f5 will transpose to
line A below once White (inevitably)
plays lt:lc3.
c ) 2 lt:lf3 e5 3 d 4 e 4 4 lt:lg 5 (4 lt:lfd2 f5 White will now choose from the fol-
5 lt:lc3 lt:lf6 6 e3 is line 82) 4 ... �b4+ 5 lowing:

120

A: 3 g3 1 2 1
B: 3 'bf3 123
A: 3 g3 fs!?
A reversed Grand Prix Attack, which
is "not a good idea", according to Hi­
karu Nakamura. Nonetheless, it was
most likely the great disparity in play­
ing strength that was the difference in
our encounter - upon examination, the
positions are just fine so long as Black
does not use the typical plan of trying
to deliver checkmate right out of the
opening. Come to think of it, he may
have been saying that the opening was
a bad choice against him in particular,
which is no doubt true given that I had
nothing prepared other than the usual
recipe.
4 i..g 2 'bf6 5 d3
After 5 e3, the gambit 5 ... d5 ! ? h as
been very successful. The i dea is to
open up the game to reach the weak
light squares on White's queenside,
particularly the gaping hole on d3. If
White tries to remain a pawn ahead,
1 c4 lbc6
Black obtains dangerous compensa­
tion; e.g. 6 cxd5 'bb4 7 'ilfb3 ? ! 'bd3+ 8
�f1 e4 9 'bh 3 i.. d 6 and White h as great
difficulties with both king safety and
development. Instead, 7 d3 lbfxd5 8
lbxd5 lbxd5 9 'bf3 is the normal con­
tinuation, when Black's position is a bit
loose but White needs to create some
sort of off-board diversion to sneak his
e-pawn back to e2. L.Gofshtein ­
N.Mitkov, Lisbon 1999, continued
9 ... i.. d 6 10 o-o 'bf6 11 iih3 "fle7 12 e4
fxe4 13 dxe4 and the game was equal
after 13 ...1li'f7 14 "ifc3 'iVh 5 1 5 'bd2 o-o,
though White went on to win (see
Game 88).
s...i.. b4 6 i..d 2
otherwise White's c-pawns will be
doubled, providing long-term compen­
sation for surrendering the bishop pair.
A brutal example is J.Ramirez­
J.Schuyler, Las Vegas 2007, which saw 6
a3 ? i.. x c3+ 7 bxc3 d6 8 l:tb1 o-o 9 e3? e4
10 d4 b6, followed by ... i.. a 6 (see Game
89).
6 ...0-0
7 'bf3
121
Th e Dark Kn ight System
In stead:
a) 7 e3 worked great for White in
H.Nakamura-J.Schuyler, Las Vegas 2008:
7 ... ..txc3 8 ..txc3 d6 9 lt:le2 �e8 10 h 3,
and since 10 .. .'iVh 5 ? ! would be m et by 11
f4! I was already out of ideas and lost
quickly. Meanwhile, Black is actually
fine after 10 ...b6! ? 11 0-0 .i.b7 to oppose
White's strong bishop and work to­
wards relieving him of the bishop pair.
This is nearly equal according to Hou­
dini, and can be used just as well
against 10 'it'd2 (though 10 0-0 is a seri­
ous test).
Going back further, Black h as 8 ... d5 !
which is three for three in my database,
most significantly in O.Foisor-J.M.De­
graeve, Le Touquet 1996 (see Game 90).
And if that's not enough, in J.lruzubieta­
B.Gulko, San Sebastian 1996, Gulko
casually played 7 ... lt:le7 ! ?, and after 8
lt:lge2, 8 ... c6! to guard the d5-square and
preserve the dark-squared bishop. This
too was nearly equal (see Game 91), al­
though 8 a3 is more testing.
Last but not least, Black can try
7 .. .f4 ! ? which scored a full point in
N.Spiridonov-K.Spraggett, Cannes 1992
(see Game 92). In case it's n ot obvious,
my main recommendation is 7 ... .txc3 8
.txc3 d 5 ! .
b ) 7 a3?! is very often played, and
with very bad results. lt is h ard to jus­
tify spending a tempo to force an ex­
change which Black is interested in
making anyway: 7 ....txc3 8 .ltxc3 d6
(again 8 ... d5 ! ? 9 cxd5 lt:lxd5 10 .i.d2 h 6 !
11 lt:lf3 e4! o r 11 'i!fb3 ? ! .i.e6 12 'it'xb7
122
lt:ld4 with excellent compensation) 9
lt:lf3 'it'e8, and with White having
squandered a tempo on a2-a3, Black's
"typical plan" (described below) has
done well, while if White delays cas­
tling, he run s other risks. F.Bruno­
B. Kurajica, Lugano 1985, was a short
lesson for White (see Game 93).
7 e4!
...
This rare line is best, taking advan­
tage of White's previous move to estab­
lish a knight on the e4-square.
For the record, the "typical plan" is
7 ... d6 8 0-0 .ltxc3 9 .txc3 'ii'e 8, intend­
ing 10 ... 'ii'h 5, 11 .. .f4, 12 ... .th 3, 13 ... lt:lg4,
14 .. .fxg 3 1 5 fxg 3 .txg 2 16 'itxg 2 .l:.xf3
17 'itxf3 �f8+ and wins ! Some of the
most astute readers may have noticed
that five of White's moves were
skipped - indeed that is a little snag. To
see how Black's attack can work
against an opponent with a pulse,
check out M.Sher-K.Spraggett, Andorra
1993 (Game 94), though Spraggett did
require a large assist from his GM op­
ponent, and 7 ... d6 is not my recom­
mendation.

8 dxe4 ..ixc3!
8 ... !Dxe4 will normally transpose,
but this new move order cuts down on
White's options.
9 ..ixc3 !Dxe4 10 l:tc1
I expect White to try to preserve his
pawn structure as he has in the analo­
gous positions from the alternative
move order. However, if 10 o-o ! ?
!Dxc3 ? ! 11 bxc3, White's open lines
make it easy for him to pressure the
queenside. So Black should retain the
strong knight: 10 ... d6 11 'ir'c2 as 12
lbd4 fie7 and 13 .....id7 with just a
slight edge for White.
10 ... d6 11 0-0
The only time this position has been
reached, Black started losing the thread
with 11 ... ..ie6? ! 12 lbd4 l2Jxd4?! 13
'ir'xd4 l2Jxc3 ? ! 14 l:txc3 cs 1S "ii'd 2, when
White had a pleasant long-term advan­
tage and went on to win in the game
Har.Becker-Joa. Franz, G erman league
1997.
Instead, after 11 ...1fe7 12 l2Jd4 ..id7
13 "ifh3 l2Jcs 14 'ii'c 2 aS, White's edge is
tiny.
1 c4 lbc6
8: 3 !Df3 fS
This move has a solid reputation
and is a popular alternative to 3 ...!Df6.
4 d4 e4
And now mainly:
81: s ltJgs 124
82: s l2Jd 2 125
But also:
a) 5 ltJes - only larry Christiansen
(and Houdini) seems interested in play­
ing this move, though he has an im­
pressive 3-0 with it. After s ... l2Jf6 6 ..if4,
as in l.Christiansen-S.Conquest, Oviedo
(rapid) 1992, the new move 6 ... !Dh s ! ? 7
..id2 l2Jf6 is fine for Black if he is con­
tent with a draw, because 8 e3 d6 9
l2Jxc6 bxc6 10 ..ie2 ..ie7 1 1 o-o o-o is
nothing for White. Otherwise Black can
try 6 ... d6 7 l2Jxc6 bxc6 8 e3 ..ie7 9 ..ie2
0-0 10 0-0 ..ie6, with a slight edge for
White.
b) s ..ig s ..ie7 6 ..ixe7 l2Jgxe7 7 !Dd2
l2Jxd4 8 l2Jdxe4 !De6 9 l2Jd2 b6! 10 e3
..ib7 1 1 liJf3 0-0 with easy equality and
chances for more; e.g. 12 ..ie2 f4! is
123
good for Black, is 12 g 3 ? ! f4! 13 exf4
as
�4 14 gxf4 :!.xf4 1 5 i.g 2 {or 1 5 i.e2)
15 ... 'iff8 ! .
B1: s lLlgs i.b4
6 lLlh 3
This retreat is necessary.
a) G.McKenna-J.Schuyler, Richmond
2008, instead continued 6 f3 ? ! h 6 ! 7
lLlh 3 exf3, threatening 8 ...'ifh4+ and
9 ...'iixd4. My opponent found 8 e3!
which keeps White in the game,
though 8 ...'ifh4+ 9 g3 i.xc3+ 10 bxc3
'iih 5 11 i.g 2 'ii'f7 12 i.xf3 'ii'xc4 13
.ih 5+ �d8 would have allowed Black
an edge. {A. Beliavsky-V.Bagirov, Minsk
1983, also saw White play 6 f3 ?!, but
Black did not punish him and went on
to lose.)
b) 6 g 3 ? ! , as in J .Timman-I.Sokolov,
Dortmund 1999, is n ot a good idea ei­
ther - the bishop should not be caught
dead on the g2-square. After 6 ... lLlf6 7
d5 lLle5 8 "ifb3 'ii'e 7 9 i.g 2 ? ! h 6 10 lLlh 3
'iic 5 11 i.f4? ! lLlxc4 White had little for
the pawn.
After the text move, the position
124
again resembles a reversed Grand Prix
Attack. White's king's knight will soon
reach the excellent f4-square, but the
four moves it takes to get there is a
high price, even in a blocked position.
6 ... lLlf6 7 e3 i.xc3+ 8 bxc3 d6 g l2Jf4 o-o
White's c4-pawn is weak and his
bishops are not yet working, but he
should be able to force through the c4-
c5 break. If Black is not careful, this will
bring White a significant advantage.
However, White needs to be careful as
well, because if the position opens up
at the wrong time or in the wrong way,
he will be punished for his slow devel­
opment.
10 h4 b6 11 i.a 3 lif7!
This is a new move, and it makes us
more comfortable by stepping out of
the influence of White's a3-bishop and
avoiding lLle6 forks. lt is also useful to
guard the seventh rank and shield our
king . The similar 11 ... 1ie8 was tried
successfully in R.Koch-M.Wiedenkeller,
Reggio Emilia 1982/83, but it accom­
plishes a bit less and leaves Black vul­
nerable to the il.b5 pin . lt is worth not-

ing that Black's results have been excel­
lent in this variation even without the
text move, though a few high-level
players have successfully advocated
White.
12 C5
If White does not play this now,
12 ... i.. a 6 is coming, and the c4-c5 break
will be forever ineffective.
12 ... bxcs
12 ... d5 ! ?.
13 i..c4
Naturally, if 13 dxc5 then 13 ... d5! 14
c4? d4 and White's structure is awful.
13 ... ds 14 i.. b s lbbs 15 i..x cs c6 16 i.. e 2
i.a6
White h as relieved him self of his
pawn weakness, but the coming bishop
exchange brings equality. Black's plans
are to manoeuvre a knight to the lovely
c4-square, trade off White's f4-knight,
and fight for the b-file; e.g. 17 l:!.b1
i.xe2 18 'ifxe2 lbbd7 19 lbe6 'ifc8 20
i.d6 lbb6 21 lbc5 lbfd7 22 lbxd7 l:!.xd7
(or 22 ... 'itxd7 23 i.f4 .l::t c 8 and 24 ... c5)
23 i.f4 lbc4 24 o-o .l:.b7, which is still
equal. Black has achieved his objec-
1 c4 lbc6
tives, but White's "bad" bishop is very
strong . Black will not be better unless
he can arrange to exchange the other
pair of rooks.
B2: s lDd2
This move is not common here, but
the position is important since it can
be reached from a few different move
orders.
s ...lbf6 6 e3
6 lbdb 1 ! ? is a cute move, intending
7 i.g 5 . Houdini likes 6 ...b6 ! ? opening
paths for the bishop and sketching out
some territory on the queenside. Then
7 i.g 5 h6 8 i.xf6 i!Vxf6 9 e3 (or 9 lbd5
i.b4+ 10 lb1c3 i.xc3+ 11 bxc3 �d6 12
e3 lbe7 13 lbf4 g5 14 'ifh 5+?! 'itf8 15
lbe2 i.a6 and Black is better) 9 .. .'ii'f7 10
lbb5 'itd8 11 i.e2 g 5 leads to a murky
position which Mr. H calls equal. Unfor­
tunately, there are no games to draw
on, and the computer convincingly
shot down all my sensible ideas, so we
are stuck with this nonsense. He con­
tinues 12 a3 h 5 13 lbd2 g4 14 'it'c2 h4
15 o-o-o h 3 16 g3, still equal.
1 25
Th e Dark Kn ig h t System

novelty that prevents Black from get­

ting overrun on the queen side, reach­
ing a satisfactory reversed Classical
French.
9 cxds
Or 9 a3 i.e6 10 b4 a6 !, strongly dis­
couraging the further advance of
White's queenside pawns.
g lZ'lb4 10 'i¥b3 lZ'lbxds 11 lZ'lc4 c6 12
.•.

i.d2 filc7

6 ... i.e7l
6 ... g6 is just about always played,
but the fianchetto is slow and some­
what accommodating, releasing the
b4-square. Sokolov and Sigurjonsson
both got squashed by White's huge
queenside after l:tbl and b2-b4. The
text move keeps Black focused on
where the play will actually take place.
7 i.e2 o-o 8 o-o d S I
Previously 8 . . .'it'e8 and 8 ... d 6 have Intending 13 ... i.e6, 14 ...l:tfd8, and
been tried. The text is an unplayed Black is fine.

126
 C h a pte r E ight

1 t2Jf3 t2Jc6

4 if.. g 2 g61
An excellent recipe. White's extra
tempo amounts to little.
5 o-o i.g7 6 e4
White challenges the centre with 6
c4 less often. In Bu Xiangzhi­
V.Ivanchuk, Khanty-Man siysk 2011,
Black played 6 ... dxc4 7 dxc4 'ii'x d1 8
.!:Ixd1 e4 9 tbfd2 fS, when White needed
10 f3 to equalize. Instead, after 10 tbc3
if.. e 6 11 tbdS ? ! o-o-o lvanchuk went on
If you play 1 ... tbc6 against 1 e4 and to win (see Game 9 5).
1 d4, you may as well play it against 1 6...tbge7 7 tbbd2 o-o 8 c3 a s 9 a4 h6
tbf3, since White will usually respond
with 2 d4, 2 e4, or 2 C4, transposing to
Chapters One, Five and Seven respec­
tively. The only continuation with in­
dependent significance is:
2 g3 e5 3 d 3 d S I
A reversed Pirc is quite safe for
Black, as long as he is careful not to
choose a sharp variation. I 've played
3 .. .fs here in the past because it is ex­
tremely effective against the King's
Indian Attack, but its value is question­ Black h as outscored White here at
able against other set-ups. all levels of play. K.Arakhamia Grant-

127
Th e Dark Kn ight System

A. Raetsky, Bern 1995, showed Black A.Baburin, San Francisco 1997, saw
expanding and using his space advan­ Black's central control turning into a
tage after 10 .l:.e1 ..te6 11 exds ..txdS ! powerful tactical strike against White's
(see Game 96); while V. Frias Pablaza- king after 10 exds ll'lxds (see Game 97).

128
 C h a pte r N i n e

Oth e rs

I used to "pre-move" 1 ... lt:Jc6 in online dini and I have cooked up something
blitz games - after all, it's playable new and fun against Larsen's Attack: 1
against everything, right? Eventually I b3 g6 2 ii.b2 lt:Jf6 3 e4 .i.g 7 ! ? (3 ... d6 is
was embarrassed (several times!) by a much more common) 4 es lt:Jds 5 c4
player who sometimes opened 1 b4. lt:Jf4!
After 2 bS, I did not feel like I had found
the refutation to the Orang-utan !
2 ... lt:Jes 3 ii.b2 ? ! lt:Jc4 ! ? is not bad, but 3
e4 or 3 d4 leaves Black with no excuse
for his knight placement.
In all seriousness, 1 .. lt:Jc6 is just not
.

the best way to start challenging White

after 1 b3, 1 f4, 1 e3, 1 a3, or especially
1 b4. Play whatever you like, but in my
opinion, 1 ... g6 - the "Dark Bishop"? - is
simple and strong against all of these.
Against 1 b4 and 1 b3, 1 .. g6 may.
come as a surprise to the opponent,
which is always nice. lt also makes sense
to oppose White's bishop on the long
diagonal, especially since his will be
loose, while ours (after we castle short)
will be guarded - a tactical advantage.
Larsen's Attack
Based on the counter-fianchetto, Hou-
This move is provocative and strong .
White can keep advancing pawns, with
tempo in some cases, but he will not be
able to control the territory he imag­
ines he is conquering, and he will be­
come weak in an area of the board I like
to call the "deep centre" - d3, e3, and
the surrounding squares. After 6 g 3
lt:Je6 7 d 4 d 6 8 ds?! lt:Jcs 9 b 4 (otherwise
the knight is very powerful in conjunc-

129
tion with .ifs ) 9 l2Jcd7 10 f4 o-o 11
... ...
lDf3 c 6 ! 12 'it'b3 cxds 13 cxds dxes 14
fxes aS 1S bxas t'Llcs, White's position
is approaching the later stages of a
long and painful decline. lt is easy to
criticize White's play, but he was al­
ready worse very early; for instance, 8
t'Llf3 o-o 9 t'Llc3 dxes 10 dxes t'Lld7 11
t'Lle4 b6 12 �d3 .i.b7 13 'i!Ve3 fS ! 14 exf6
exf6 1 S 0-0-0 still leaves White strug­
gling after 1S ... .:e8 or 1 S .. .fs.
Sokolsky's Opening (aka
The Orang-utan, a ka The Polish)
1 b4 g6 2 .i.b2 t'iJf6 3 C4 .ig 7 4 t'iJf3 0-0
S e3 d6 6 d4 t'Llbd7 7 .ie2 es 8 o-o e4 9
t'Llfd2 .l:1e8
I know, there are other ways the
game can go, but this is typical. Black
h as reached a King's Indian Attack in
reverse, and in practice Black's attack
has proven stronger (+12 -2 =4 is
77.7%). White's extra tempo, .ib2, is
just about useless because the bishop
belongs on the a3-square.
In spite of his successes, Black has
yet to produce a model for correct play.
130
In the games I can find, Black keeps
allowing the enemy pawn to a6, which
accelerates White's queenside initia­
tive. The reader is better off studying
the classic King's Indian Attacks by
Fischer, Petrosian, Bronstein, Larsen,
etc. They go (in reverse - I have trans­
lated here) ... t'Llf8, ...h 7-h S, ... .ifs, ... a7-
a6 (when White's pawn reaches as),
... t'Ll8h7, ... h S-h4 (and ... h4-h 3 if possi­
ble), ... t'Llg s with scary invasions coming
on the light squares or (if White has
played h 2-h 3) scary sacrifices on the
h 3 -pawn.
Bird's Opening
Against 1 f4, I would warn against play­
ing ... d7-dS, which gives White what he
wants: the eS-square. In stead, after
1 ... g6 2 t'Llf3 .ig7 3 e3 d6 4 d4 t'Lld7,
Black is +17 -1 =3 according to my
database. Wow! Obviously White's po­
sition can't actually be that bad. What
the statistics mean is that, for those
interested in winning chess games, it is
extremely important to combat the
opponent's automatic plans. Then S

ii.d3 e5 6 c3 'flie7 was no fun at all for
White in A.Capaliku-J.Gombac, Nova
Gorica 2010 (see Game 98); while 5
3Lc4 e6 has scored 100% for Black, most
notably in A.Spichkin-D. Reinderman,
Rijeka 2010 (see Game 99).
Of course, White does not have to
play a Stonewall - just as popular is a
reversed Leningrad Dutch with 3 g3,
when one system has performed ex­
tremely well for Black: 3 ... b6 ! ? 4 .ltg 2
�b7 5 0-0 e6 6 d3 ti:Je7 7 e4 d6
with 8 ... l2Jd7, 9 ... c5, 10 ...'i!Vc7, and
11 ... 0-0 to follow. Dzindzi pithily states
that "any opening named after an
Oth ers
animal is bad", but here the Hippo­
potamus is well-suited to combat both
White's kingside expansion and his
fianchettoed bishop. Black certainly
made it look easy in P.Auchenberg­
To.Christensen, Helsingor 1997 (see
Game 100). An important idea is to
play .. .f7-f5 just when White is ready
for his f4-f5 advance (though Christen­
sen did not find this necessary).
Other others
As for 1 ti:Jc3 and 1 g3, 1 ... ti:Jc6 should
transpose to our repertoire at some
point.
As I m entioned already, 1 ... g6 is a
good an swer to 1 e3, and also to 1 a3,
when White is challenged to find any
use for his opening move. If 1 h3, 1 h4,
1 ti:Jh 3, 1 lLla3, 1 a4, or 1 f3, you're on
your own . 1 f3 is particularly troubling
because it is not clear whether White is
intending 2 ti:Jh 3 and 3 ti:Jf2, or 2 g4!, or
2 �f2 ! and 3 �g 3 ! ! . May you be con-
fronted with these problem s fre­
quently.
131
 C h a pter Te n

Misce l l a n eous Topics

Reducing the Workload - but anyone who prefers t o fianchetto

The Dark Knight for Dumm ies can study less and play his own favour­
Perhaps some readers have g otten the ite positions more.
impression that the study material Notice that when playing the Sys­
{which in my opinion is very small} can tem this way, White's f- and c-pawns
be substantially simplified. Indeed, it are both free, so White could play a
can . There is no sensible way to avoid " Four Pawns Attack" if he so chooses.
Chapter One, but Chapter Two is, in Without going into detail, I will point
some sense, unnecessary. After 1 d4 out that this is not a good way for
lt:\c6 2 c4, Black can simply play 2 ... d6, White to fight for an advantage: 1 e4
which will almost certainly transpose ctJc6 2 d4 d6 3 dS ctJb8 4 C4 g6 5 f4 .itg7
to Chapter One. Likewise, Black need 6 ctJf3 ctJf6 7 ctJC3 o-o 8 ..te2 e6!
not learn Chapter Three. After 1 d4 ctJc6
2 dS, Black can just play 2 ... lt:lb8, with a
likely transposition to Chapters One or
Five. Ch apter Four? Toss it. Just play 1
e4 lt:\c6 2 d4 d6, probably transposing
to Chapters One, Five, or Six.
In other words, don't play for an
early ... e7-e5 and plan on fianchettoing
the king's bishop whether it is neces­
sary or not. The thing is, the positions
from Chapters Two, Three and Four, are
among my favourites in the Dark 9 o-o {9 e s ! ?} 9 ... exds 10 cxds c6! 11
Knight System, so I would rather know dxc6 lt:\xc6 12 '>i;>h 1 fte8 13 .itd3 'ilfb6, or
them and play them than avoid them, 13 ... a6 and 14 ... bs, with equality.

132

The Light Knight Attack!?
If the Dark Knight System is so great,
why not play it with an extra tempo?
Well, the good news is that nobody can
stop you; e.g. 1 lL'lc3 ! ? e5 2 d3 d5 3 g 3
d4 4 lL'lb1 lL'lc6 5 i.g 2 with a reversed
Dark Knight Pirc. Furthermore, there
are many other possible move orders
White can use. The bad news is that it
will often be Black fighting for the ad­
vantage, not White. How can a system
be good for Black and not for White?
The answer is that Black and White
have different opening objectives. In
the D KS Black often loses some time
with his knight, but this is mitigated by
the fact that he has reached a position
where tempi are relatively unimpor­
tant. Therefore, playing such a position
a tempo up represents only a small
gain, sometimes not enough even to
equalize.
As for 1 lL'lc3 d5 2 e4,
I believe White can fight for an ad­
vantage after 2 ... d4 (a practical one, if
not a theoretical one), but in my ex-
M iscella n eous Topics
perience Black usually plays 2 ... e6,
2 ... c6, or 2 ... lL'lf6 in stead, transposing to
his favourite defence (French, Caro­
Kann, and Alekhine's, respectively), or
just 2 ... dxe4 3 t2Jxe4 and 3 ... �f5,
3 ... lL'lf6, or 3 ... lL'ld7, with a good version
of Black's favourite Caro-Kann varia­
tion. While these positions are all play­
able for White, there is little chance of
getting Black out of his comfort zone.
Of course, real chess is not the same
as theory, and for the right player
and/or the right opponent, the "Light
Knight Attack" could be the perfect
weapon. The grandmaster Jorg Hickl
has played 1 g 3 e5 2 �g 2 d5 3 d3 lL'lf6 4
lL'lf3 i.d6 5 lL'lc3 c6 6 e4 (or 6 o-o o-o 7
e4) m any times with good results.
Readers might choose to view the
position as a Dark Knight Pirc reversed,
and Hickl's move order is an excellent
way to reach it since Black never had a
good opportunity to play ... d5-d4. There
is not much else to learn because
Black's set-up is an extremely common
one against 1 g 3 .
133
 I l l u st rative Ga mes
The games section serves several pur­
poses:
.t lt shows typical middlegame and
endgame ideas in the Dark
Knight System - there is a whole
game to be played, after all.
.t lt serves as a place to demon­
strate some alternatives to the
main recommendation s .
.t lt shows by example why certain
popular variations were rejected.
Consider these g ames a warning.
.t lt shows how successful the Dark
Knight System can be in practice.
.t lt shows how much fun it can be
to play the Dark Knight System
in your games!
A "good" King's Indian
Cam e l
J.Paasika ngas Tella­
T.Lindqvist
Fi n n is h Tea m
C h a m p i o n s h i p 1996
1 d 4 d6 2 c 4 es 3 e3 lbc6 4 d S lbce7 5
lbc3 fS 6 .i.d3 lbf6 7 ..tc2 g6
134
W e would reach this position via
the move order 1 d4 lbc6 2 e3 es 3 dS ? !
lbce7 4 c 4 d 6 etc. Black now plays an
excellent version of a King's Indian -
not an uncommon occurrence again st
an opponent inexperienced with the
Dark Knight.
S lLlf3 .i.h61?
Why not? Now if White plays e3-e4,
Black can relieve himself of a problem
piece. If White does not play e3-e4,
Black has the powerful es/fS pawn
duo.
9 b4 o-o 10 .i.b3 as 11 bxa s lba s 12 a4
'5ith8 13 lbbs lbe4 14 o-o?l c61 15 dxc6
bxc6 16 lba3 .i.e6
White h as no compensation for
Black's huge centre.
 Illustrative Games

32 ... .i.b4! 3 3 lt:lb1 i.xc4 34 l:i.g1 ..ic5 3 5

l:!.e1 .i.f2
Black soon won.

An a nti-Stonewall -
rare light-squared play

Gam e 2
J.Vialatte-F .Giroux
Pa ris 2006

17 lt:ld2 lt:lc5 18 ..ic2 l:!.a6 19 .i.b2 .i.g7
20 e4 f4?! 21 'it>h1 'i*'d7 22 f3?!
White weaken s his dark squares
unnecessarily. Black should have to
work for this.
22 ....l:!.fa8 2 3 lt:lab1 lt:lc8 24 lt:lc3 lt:lb6 25
..ib3 "iVf7 26 ..ic2 i.f8 27 i.a3 lt:lcxa4 28
lt:lxa4 lt:lxa4 29 i.xd6 .i.xd6 30 .l:txa4
lba4 31 i.xa4 "fic7
1 d4 lt:lc6 2 f4 d5
Black has not signed a contract to
play on the dark squares, and White
has already given himself a big hole on
e4. Furthermore, with Black not having
played ... lt:lf6, White is not guaranteed
an outpost on es.
3 lt:lf3 i.g4 4 e3 f6!?
Black may or may not get ... e7-e5 in,
but he has certainly stopped lt:les !
5 ..ib5 'ii'd 6 6 o-o a 6 7 ..ixc6+?!
This is not going to help the situa­
tion on the light squares.
7 .. JWxc6 8 c3?! lt:lh6 9 lt:lbd2

Black has somehow managed to

avoid winning any pawns, but his
bishop pair provides a nearly decisive
advantage.
32 h3? 9 ... 0-0-0?!
3 2 'ifc2. Black's knight needs to get to the

135
Th e Dark Kn ight System
d6-square as soon as possible - 9 ...lZ'lfS !
10 .l:!.e1 lZ'ld6 ! with a pleasant advan­
tage. Black eventually gets the position
he should have, but only with some co­
operation from his opponent.
10 �e1 .ifs 11 'ir'e2 e6 12 .l:i.e1 i.e4 13
C4 i.b4!?
Black's plan, which he will soon exe­
cute, is to give up both bishops and
play "good knight versus bad bishop".
This fails to take into account White's
development and queenside counter­
play.
14 cxds exds 1S a3 ..itxf3 16 'ilfxf3 i.xd2
17 ..itxd2 fs 18 .l:.ac1 'ife6 19 'ife2 l:!.d7
20 b4 cJi>b8 21 bs axbs 22 'i'xbs lZ'lf7 23
i.b4 lZ'ld6
24 'ii'a 4?
An incomprehensible positional
mistake. The knight that lands on c4 is
worth far more than the "tall pawn" on
b4. However, in this roundabout way
we see the proper fruition of Black's
early advantage.
24...lZ'lc4 2 S .ics b6 26 'ii'a 6 'ii'c 6 27 l:tb1
'ifa8 28 'fibs? 'it>c8! 29 a4
29 i.b4 c6 traps the queen.
136
29 ... bxcs 3 0 �xcs 'i'a s 31 'i'c6 l::!. d 6 3 2
'ifb7+ 'it>d7 3 3 .l:!.ec1 .l:!.a8 3 4 'fibS+ �xbs
3 S .l:.xbs .l:.xa4 36 l:tb8 lt:'lxe3 3 7 l:te1
l:te6 38 .l:.g8 l:te7 o-1
Can White be made to
pay for delaying d4-d 5 - ?
Game 3
M. Tratar-M.Srebrnic
S l ove n i a n C h a m pion s h i p,
Lj u blja n a 2010
1 d4 d6 2 lZ'lf3 g6 3 c4 .ig7 4 lZ'lc3 lt:'lc6
We would reach this by 1 d4 lZ'lc6 2
lZ'lf3 d6 3 c4 g6 4 lZ'lc3 .ig7 - in this par­
ticular game White did not even have
the option of playing 4 ds.
s ds ttJes 6 lt:'lxes .ixes 7 e4 lZ'lf6
lt seems odd to cut off the retreat
for the dark-squared bishop, but the
piece is very active where it is, and the
danger is far less than it appears.
White's lame attempt to trap the
bishop on move nine goes nowhere.
Such lovely dark square control !
8 .id3 o-o 9 lZ'le2
In addition to the threat of f2-f4,
White makes sure Black cannot dam­
age his structure with ... .ixc3 - how­
ever, this retreat is still not the most
accurate.
9 ... lt:'ld7 10 h4 hs 11 lZ'lf4 lt:'lcs 12 .ic2
e6!
White has forgotten about his de­
velopment and now stands worse.
White is a GM, by the way. The rest of
the game is kind of brutal .

13 .l:::!. b 1 as 14 f3?1
Be careful on the dark squares,
White!
14 exds 15 cxd s c6 16 dxc6 bxc6 17
••.
tZ'le2 �a6 18 �gs 'il'c8?1 [18 ... 'ilfb6] 19
�f2 .l:tb8 20 b3 .l:!.e8 21 :e1 tZ'le6 22 i.. e 3
dSI 2 3 exd s cxd s 24 Itc1 �d8 2 5 ..td3
'ii'x h4+ 26 �f1 �b7 2 7 'iid 2 d4 28 �g1
lZ'lf4 29 tZ'lxf4 �xf4 30 'it'xa s �xf3 31
.!:l.c2 �g3 0-1
Making White pay
for delaying d4-d 5
Game 4
A.lpatov-R.Antoniewski
G e rm a n League 2011
1 d 4 d6 2 tZ'lf3 g 6 3 c4 �g7 4 lZ'lc3 tZ'lc6 5
ds tZ'les 6 tZ'lxes ..txes 1 e4 lZ'lf6 8 i.d3
o-o 9 lZ'le2 lZ'ld7 10 o-o tZ'lcs 11 f4 �g7
12 fs?J
No sooner has White conquered the
es-square than he immediately relin­
quishes it. White ( another grandm as­
ter) receives the proper punishment -
eventually.
Illustrative Games
12 e6 13 �c2 exfs 14 exfs lZ'ld7?1 15
..•
�f4? ..txb2 16 ..th6 1i'h4 17 �xf8 tZ'lxf8
18 lZ'lf4 i.xa1 19 'il'xa1 �d7
White has sacrificed a pawn to
trade off Black's powerful bishop, only
to find that the exchange has not
helped him. This is a typical result.
20 'iic 3 :e8 21 cs �bs 22 g3 "ike7 2 3
lib1 � a 6 24 �a4 'it'e4 2 5 i.c2 'i!Ve3+ 26
'ifxe3 l:txe3 27 cxd6 cxd6 28 �a4 l:ta3
29 .i.b3 lZ'ld7 30 l:i.e1 �f8
31 tZ'le6+??
White's position was very unpleas­
ant anyway, but he gets no compensa­
tion for the knight.
31 .fxe6 32 dxe6 tZ'les 33 l:tc1 gxfs 34
••
137
Th e Dark Kn ight System

l:tc8+ <J;; e 7 35 l:tc7+ 'it>f6 36 l:txh 7 .ic4 3 7 tt::lxe4 19 'ii'e 3 .ixd4 20 �xd4 tt::l c s 21
e7 .i b s 38 h 4 d s 39 h s tt::lf 3+ 4 0 <J;;g 2 :e1 'iif8 Yz-Yz
tt::l g s 41 :h6+ rj;;x e7 42 .l:tg6 tt::lf7 43 h6
.l:ta6 0-1 White avoids tt:Jxes

White plays an early e4-e5, Cam e 6

accomplishing nothing R.Fischer-J.Schuyler
Rich mond 2008
Gam e s
Y.Balashov-G.Kuzmin 1 tt::lf3 tt::l c 6 2 e4 d6 3 d4 tt::lf6 4 ds tt::l e s s
USSR C h a m pi o n s h i p tt::l c 3 tt::lxf3+ 6 'ii'xf3 g6 7 .igs i.. g 7 8
Vi l n i u s 1980 i.. b S+
White's bishop has no squares so he
1 tt::lf3 tt::lf6 2 c4 g6 3 tt::lc 3 .ig7 4 e4 d6 5 trades it, but I appreciate the extra
d4 o-o 6 .ie2 tt::lc 6 7 d S tt::l b8 breathing room .
We would reach this by position 8 i..d 7 9 i..xd7+ 'iix d7 [9 ... tt::l x d7 ! ?] 10
.••

with the move order 1 d4 tt::l c 6 2 tt::lf3 d6 �e2 o-o 11 o-o e6

3 c4 g6 4 dS tt::lb 8 5 tt::l c 3 .ig 7 6 e4 tt::lf6 7
.ie2 o-o.
8 h3 tt::l a 6 9 .ie3 tt::lc s 10 es tt::lfd7 11
exd6 exd6 12 .id4

12 .l:.ad1
On 12 'ii'f3, I intend 12 ... tt::l e 8 fol­
lowed by 13 ... c6.
12 exds 13 exd s?!
•••

12 tt::lf6
••• Instead 13 i.. xf6 ! .ixf6 13 tt::l x ds is
A valid choice, though Black could equal .
just as comfortably have all owed the 13 .. Jlae8 14 'iif3 tt::l g4 15 h3 tt::l e s 16
exchange with 12 ... as or 12 ....U.e8. li'g3?!
13 0-0 a s 14 kte1 .:res 15 .if1 .id7 16 16 �e2 is the lesser evil .
.l:.xe8+ 'ii'x e8 17 'iid 2 tt::lfe4 18 tt::lxe4 16 .'i!Vfs!
••

138

Black is better here, but there are
many mistakes in the coming compli­
cations - White's next move really stirs
things up.
11 lt:Jb517 h6 18 ..te3 'Yi'xc2 19 ..td471 [19
..txa7 ! ?] 19 ... a6? [19 .. .'i¥c4! 20 tt:Jxa7
"i!Vxds] 20 .l:.c1? 'Yi'e4 21 .l:.fe1 'ii'xd 5 22
Ci:Jxc7 'ii'x d4 2 3 tt:Jxe8 .l:txe8 24 �h1 l::td 8
25 :c1 'iib 6 26 �e77 �f8 0-1
White's alternate plan
with g2-g3
Came l
L.Aitounian-J.Schuyler
La s Vega s 2008
1 lt:Jf3 tt:Jc6 2 d4 d6 3 d5 tt:Je5 4 tt:Jxe5
dxe5 5 c4 e6 6 tt:Jc3 lt:Jf6 7 g3 exd 5 8
cxd 5 ..tc5 9 ..tg2 o-o 10 o-o ..tg4
White has managed to keep his
space plus without committing his e­
pawn, which leaves me in some doubt
as to the best squares for my pieces. My
last is intended to provoke White to
weaken his king position, a pl an which
I continue throughout the game.
Illustrative Games
11 �b3 ..tb6 12 h 3 ..tf5 13 �h2 h6 14
�c471
This only helps me find the right
plan. After this game, the knight tran s­
fer became the standard way for me to
combat this whole variation with 3 ds.
The alternative try 14 a4 would at least
have given me a little problem to deal
with.
14 ... tt:Je81 15 lt:Ja4 lt:Jd6 16 �b3 ..te4
White is intending to grab the
bishop pair, so I already start working
to relieve him of it. lt is awkward for
him to avoid the trade.
11 f3 i..f5 18 ..td2 i.. d 7 19 tt:Jxb6 axb6
20 il.b4 .l:Ie8 21 e4
139
Th e Dark Kn ig h t System

levon mentioned to me after the

game that he played this to stop
21 ... l2Jfs, oddly overlooking the alter­
nate route.
21 ...t2Jb5 22 .ic3 c5 23 dxc6 bxc6 24
l:tfd1

Whatever you do, don't let this hap­

By n ow I am actually better if I just
play the simple moves 24 ... cs and
2 S ... l2Jd4.
24 ... 'ii'c 7?! 25 .if1 .ie6 26 .ic4 .ixc4 2 7
'iixc4 l2Jxc3 28 'iix c3 l:ted8 29 a 3 :d6 30
.l:tac1 J:.ad8 31 l:txd6 Wxd6 3 2 l:tc2 'iid 1
3 3 l:tf2 .l:.d3 34 'iix c6 'it'e1 3 5 'ii'c 2 .l:.d1
Yz-Yz
After 36 'ii'c 8+ 'itth 7 37 'iif S+ g6 3 8
'iixf7+ the result is a draw by perpetual
check.
pen to you!
7 ... .txb5 s 'ii'h 5! .ib4+!?
Following the disaster on move 5,
this is the best chance in a terrible posi­
tion.
9 lLlc3?! 'ii'd 4?
Much better is 9 ... .txc3+ 10 bxC3
'iif6 when the continuation 11 exf7+
'ittf8 12 fxg 8+ 'ittx g8 is not so terrible for
Black
10 'ii'xf7+ 'ittd 8 11 .id2 .txc3 12 o-o-o!

Don't play s e6?? ...

Gam e 8
J.Bonin-J.Schuyler
N ew Yo rk 1988
Ow!
1 d4 l2Jc6 2 l2Jf3 d6 3 d5 l2Je5 4 l2Jxe5 12 ....txb2+ 13 'itt b 1 l2Je7 14 .ie3 .ic3 15
dxe5 5 e4 e6?? [s ... l2Jf6 !] 6 .ib5+ .id7 7 'i!i'xg7 .:gs 16 'i!i'xe5 1-0
dxe6!! The carnage is unspeakable.

140

Equal plus symmetrical equals
win - fight the good fight
Game 9
H.Keskar-J.Sch uyler
H a m pton 2011
1 d4 lZ'lc6 2 lZ'lf3 d6 3 d S lZ'les 4 lbxes
dxes s e4 lZ'lf6 6 i.d3 e6 7 dxe6?!
Obviously White can forget about
an advantage after this - the question
is, how do I beat my lower-rated oppo­
nent from this nearly symmetrical po­
sition ? The fact is, if I play at all dy­
namically and keep my eyes open,
there will be opportunities to create
imbalances. One way to start is to delay
castling.
7.....txe6 8 o-o ..tcs 9 ..tgs h6 10 ..th4
'Wie7 11 a3 l:td8?! [11 ... 0-o-o] 12 lZ'ld2 gS
13 i.g3 ..tg4 14 Wie1 lZ'lhs 1S ..te2 lbxg3
16 hxg3 ..te6 17 tLlf3 f6 18 l:td1 hs
Black is nowhere near winning, but
White certainly has a lot to worry
about with ...h S -h4 and (after g3xh4}
... g 5 -g4-g 3 coming.
19 l:.xd8+ 'ii'x d8?! [19 .. :�xd8 !] 20 'iVc3
Illustrative Games
'Wie7 2 1 ..tc4 i.c8 22 b4 i.b6 2 3 a4 a6 24
as i.. a 7 2S bs?! axbs 26 ..txbS+ c6?!
[26 .. .<;£;>f8 !] 27 a6!
I h ad overlooked this.
27 .. Ji'cs 28 'ifxcs?! [28 axb7! i.. xb7 29
..tc4] 28.....txcs 29 axb7 ..txb7 30 i.. a 4
'it>d7 31 l:td1+ �c7 32 ..tb3 ..tc8
In spite of all the trades, this renews
Black's kingside threats.
33 ..tf7?! h4 34 �f1 hxg3 3S fxg3 ..tg4
36 i.b3 �b6 37 lid2 ..tb4?!
lt is best for Black to cash out with
37 .. J:th 1+ 38 �e2 l:.g 1 39 �d3 ..txf3 40
gxf3 l:txg 3 . Instead, I kept trying to
squeeze out a clear win with no risk,
but I h ave underestim ated White's
ability to generate counterplay.
38 .l:!.d1 i.. c s 39 .l:td2 ..te3? 40 .l:!.d6!
Somehow I overlooked this. Sud­
denly Black has no advantage at all.
40...fs ! 41 exfs e4 42 .l:lxc6+! �xc6 43
lZ'les+ �cs 44 lZ'lxg4 i.. d 4 4S f6
My opponent reasonably offered a
draw here. I turned over my scoresheet
and wrote the numbers 5 1 through 100
on the back.
4S ... ..txf6 46 tLlxf6 l:i.f8 47 'it>e2 .l:lxf6 48
141
Th e Dark Kn ight System
�e3 �d6 49 �xe4 .l:tf2 so �e3 l:.f1
If 50 .. J!xg 2 51 �f3 .J::! g 1 52 �g4 and
5 3 'i.t>xg 5 draws.
51 �e2 i:tf8 52 �e3 �es 53 ii.c4 l:td8
Somehow I will need to win one of
the g-pawn s and then generate
zugzwang and/or mating threats. I
wasn't sure how, but I kept on playing.
To make the draw easier on himself,
White needs to be alert for the chance
to push the c-pawn, though he is un­
derstandably reluctant to loosen his
bishop.
54 ii.d3 l:[d4 55 �f3 l:!.a4 56 �e3 l:d4 57
'i.t>f3 �ds ss �e3 �cs 59 �f3 �b4 60
�e3 �c3 61 �e2 l:ta4 62 �e3 �b2 63
�3 'itc1 64 �e2 l:tb4 65 ii.fs .:ta4 66
ii.d3 .l:!.d4 67 �e3 l:tds 68 �3 l:td4 69
�e3 l:tb4 10 �e2 .l:f.a4 71 ii.fs .l:tas 72
ii.d3 g4!?
lt was important to try eve rything
before playing this committal move,
which presents slightly different chal­
lenges to White.
73 �e3 :a4 74 ii.e4 [74 c4!] 74 ...�d1
7 5 �f4?! �e2 76 �fs �f2 77 �f4 l:tc4
78 �es 'ite3 79 ii.d3 .l:.a4 so �5 �2 81
142
ii.e4 'iii>xg3 82 �es �f2 83 ii.d s �e3 84
c4 :as ss �fs g3 86 �g4 �f2 87 �4
.l::t c s 88 �es �e3 89 �5 .l:.c8 go �es
Unfortunately, the time situation
demanded that I stop notating here. In
addition to zugzwang and mating
threats, there is now a third winning
possibility, which is somehow to drive
the white king to d8 (or thereabouts),
sacrifice the rook on d5, and win the
resulting king and pawn endgame. This
is, in fact, how I won, somewhere
around move 120-130. Naturally,
White could have held the position
with correct play, but this is hardly
relevant from a practical standpoint.
Winning the draw
Came 1 0
D.Haessei-J.Schuyler
Pawtucket 2008
1 d4 tt::'lc6 2 tt::'lf3 d6 3 d s tt:Jes 4 e4 tt::'lf6
I am planning 5 ...tt:Jxf3+, but first I
want White to figure out how to de­
fend his e-pawn.
 Illustrative Games

s lLlxes dxes 6 i.bS+ i.d7 7 �d3 a6 8 22 �c2 hxg4 2 3 hxg4 :h2+ 24 l:ld2 i.cs
i..x d7+ 'ii'xd7 9 lLlc3 e6 10 ..tgs i.b4 11 2S lLlc3 i.d4 26 1:lfd1 i.e3
o-o-o o-o-o 12 f3 1i'e7 13 �c4 h6 14
i.xf6 gxf6 15 g4

As the game develops, I am starting

to understand better why the bishop
The position has been equal since and rooks work well together - one
move 9, and now I am faced with the reason is that the bishop disrupts the
problem of generating some winning co-ordination of the enemy rooks.
chances again st my somewh at lower­ 27 .l:te2 l:r.8h3 28 .l:txh2 .l:!.xh2+ 29 'iii> b 3
rated opponent. Somewhere in my l:!.f2 30 I:td3 i.d4 31 a4 �d7 3 2 a s �d6
chess education I learned that Q+N 33 lLla4 <i;; e 7 34 �a3 l:tc2 3 5 �b3 llf2 36
work better together than Q+B, while �a3 fS I?
R+B co-ordinate better than R+N. lt is In spite of my hard work, Black is
time to arrange the ideal exchange - still nowhere close to winning. This is
which, fortunately, is not too difficult. as good a try as any.
1S ...�cSI 16 �xcs i.xcs 17 .l:f.d3 h s 37 gxfs exfs 38 llb3 fxe4 39 fxe4 fs 40
My half-open files will not suffice l:.xb7 fxe4 41 .l:txc7+ <i;;e 8
for this position .
18 h 3 l:lh7 19 .l:.f1 l:tdh8 20 liJd1 i.d6
My idea is to block the d-file and
cross my king over to the e7-square - a
small improvement to my position. My
opponent is sufficiently worried about
this to weaken his dark squares in or­
der to stop it.
21 c4?1 i.e7
I can't explain this move, but the
tempo is not important.

143
Th e Dark Kn ight System
42 cs??
Oddly, White can hold with the
"grovelling" manoeuvre 42 .l:!.h 7 ! e3 43
l:r.h 1 .l:!.d2 (43 ... e2 44 Ite1) 44 ctJc3. For­
tunately for me, my opponent could
not bring himself to consider this de­
fence seriously.
42 ... e3 43 d6 e2 44 lt:\b6 e1'i¥ 45 d7+
r3;e7 0-1
Mestrovic tries 7 g6!? and wins!
•..
Gam e 1 1
D.Rasic-Z.Mestrovic
C roatia n Tea m
C h a m pi o n s h i p 2001
1 e4 lt:\c6 2 d4 d6 3 ctJf3 '2Jf6 4 d s lt:\es s
lt:\xes dxes 6 i.bs+ i.d7 7 'i!Ve2 g6 8
i.xd7+ �xd7 9 ctJd2?!
9 c4 and 10 lt:\c3 is simpler and
stronger.
g ....Jtg7 10 ctJC4
1o ...lt:\hs
10 ... 0-o ! ? 11 lt:\xes 'fia4! and Black
recovers the pawn one way or another.
11 g4?! ctJf4 12 i.xf4 exf4
144
White has kindly released Black's fi­
anchettoed bishop from its prison.
13 o-o-o o-o 14 es e6 15 d6 cxd6?!
Too co-operative, allowing White to
land his knight on the juicy d6-square.
Instead, 1S ... 'fia4 16 'it>b1 bS 17 lt:\a3
cxd6 and Black is better.
16 lt:\xd6 'it'a4 17 'itbl f3 18 'ii'xf3 i.xes
19 l:f.hel i.xd6 20 llxd6 .:.ac8 21 b3
'ifa s 22 :edl 1li'cs 2 3 .l:[1d2 bS 24 h4 as
2 5 .l:.6d3 'ifes 26 l:te2?! 'ifh2! 2 7 h s
'ii'g l+ 28 'it> b 2 a4 29 a3 b4?! 30 axb4
.l::!.a 8?! 31 .l:r.ed2 axb3 32 .l:!.d1
By now White has regained the ad­
vantage, but it is still not simple to play
against heavy pieces with poor king
protection .
3 2 ... "i¥h2 3 3 .l:.xb3 'ii'e s+ 34 �c3 �gs 3 5
f3 lUeS 36 'ii'd 4? [ 3 6 b S ! ] 36 ....l:.ed8 37
l:!.a3 l:r.ab8 38 'ii'e 3 .l:.xb4+ 39 <t>a2 'iff6
40 .l:!.xd8+ 1\Vxd8 41 h6 1li'dS+ 42 l:tb3
l:!.a4+ 43 <t>b2 l:Ia8 44 'ii'f4 es 45 'ii'b4
'ii'd 8 46 'i!Vcs 1li'd1 47 'fia7 .l::!.c 8 48 .l:.c3
l:tf8 49 .l:!.d3 'i¥e1 so 'ii'e 3 'ii'a s 51 :a3
.l::!. b 8+ 52 ..ti>c1 i/ic7 53 'fic3 'fie7 54 'fie3
'iVh4 55 :b3 .l:.c8 56 <t>b2 'ii'f6 57 :b6
ilie7 S8 11i'b3 e4 59 f4?

After a long defence, White cracks.
59 .. ."ii'c 7 60 .l:[b7 'tlt'xf4 61 'ili'd 5? e3 62
l:r.a7? 'i¥f6+ 0-1
Mestrovic wins again with 7 g6!? •••
Game 1 2
J.Barle-Z.Mestrovic
S l ove n i a n C h a m pion s h i p,
Krsko 1997
1 lbf3 t:Lic6 2 d 4 d6 3 e 4 t:Lif6 4 d 5 lL'le5 5
t:Lixe5 dxe5 6 .tb5+ i.d7 7 "ife2 g6 8 o-o
.ltg7 9 .txd7+ 'tlt'xd7 10 c4 o-o 11 lL'lc3
tLih 5 12 g3 c5 13 dxc6 "ifxc6 14 .te3 e6
15 b4 f5
Illustrative Games
White is still better. Black's kingside
pawn majority is not as dangerous as it
looks. Mestrovic soon drifts into a bad
position without making any obvious
mistakes, which is why I'm not crazy
about this variation .
16 b5 'ii'e 8 17 .tc5 l:i.f7 18 f3 b6 19 .td6
l:td7 20 llad1 .:tc8 21 c5 bxc5 22 1i'c4
t:Lif6 2 3 .txe5 l:lcd8 24 l:lxd7 lbxd7 2 5
.txg7 �xg7 26 f4?!
26 lbe2 ! is stronger, intending 27
t:Llf4.
26 ..."ife7 27 exf5 exf5 28 lL'ld 5 'ii'e 6 29
'ii'c 3+ �h6 30 g4 fxg4 3 1 'ii'd 2 �e4 32
f5+ g5 3 3 .I:.e1 'ii'xf5 34 lL'le7 'iWf6 35
'ir'd 5 �h5 36 lL'lf5 lbf8 37 'ii'x c5 'ir'b6 38
'iWxb6 axb6 39 lL'lg7+ �h6 40 lL'lf5+ �g6
41 tLie7+ �f6 42 t:Lic6 .l:.d2 43 :f1+ �g7
44 lL'le5 lbg6 45 .l:.f7+? �g8 46 .l:tf5??
:d1+ 0-1
Not 46 ... .:tds ?? 47 lbxg6 .l:.xfs 48
tLie7+ and White wins, but after
46 ... .:.d1+ 47 �f2 (or 47 �g 2 t:Llh4+)
47 .. -l:Lds 48 lL'lxg 6 �xfs is check.
Beating the London System
Gam e 13
P .B.Pedersen-D.Bekker Jensen
Da n i s h Tea m
C h a m pion s h i p 2008
1 d4 t:Lif6 2 lL'lf3 g6 3 .ltf4 .ltg7 4 e3 0-0 5
.i\.e2 d6 6 h 3 liJfd7 7 0-0 liJc6 8 C3 e5
We would reach this by 1 d4 t:Lic6 2
t:Lif3 d6 3 .ltf4 lL'lf6 4 e3 g6 5 .te2 .ltg 7 6
h3 o-o 7 c3 lL'ld7 8 o-o es.
9 .th2 f5 10 dxe5 dxe5 11 t:Lia3 g5 12
145
Th e Dark Kn ight System

l2Jc4 'iif6 13 b4 f4 Beating the London System again

Game 14
R.Valenti-V.Tkachiev
Corsica (ra p id) 1997

1 tL'lf3 tL'lf6 2 d 4 g 6 3 i.f4 i.g7 4 e3 d 6 5

h 3 o-o 6 i.e2 t2Jc6 7 o-o tL'ld7?! [7 ... e s ! ] 8
i.h2 e5 9 c3 "ilke7 10 a4 f5 11 t2Ja3 'i£i>h8
12 b4 e4

Houdini completely hates Black's

position, but an aggressive stance on
the king side often reaps benefits. lt is
very difficult for White to solve the
problem of liberating his entombed h2-
bishop while at the same tim e keeping
his king safe.
14 t2Jfd2 tL'lb6 15 tL'la 5 'iig 6 16 e4 �dB
By now even Mr. H understands
that White is not better. Black decides he will be able to get
17 �c2 t2Jxa 5 18 bxa 5 tL'ld7 19 l:!.ad1 .. .fS-f4 in later - this is slightly optimis­
l:!.e8 20 'ifa4 <i£i>h8 21 f3 h 5 22 tL'lb3 t2Jf6 tic, but it pays off.
2 3 a6 g4 24 hxg4? hxg4 2 5 axb7? g3!! 13 l2Jd2 g5 14 �c2 t2Jf6 15 b5 t2Jd8 16 C4
t2Je6 [16 .. .f4 ! ? ] 17 J:.fc1 f4 18 l::!. a b1 'ii'e 8
19 t2Jxe4?? fxe3 20 t2Jxf6 exf2+ 21 'it>h1
i.xf6 22 'ii'e4 t2Jxd4 0-1

Losing to the London System

Game 1 5
V.Golod-E.Sutovsky
N ata nya (ra pid) 2009

1 d4 t2Jf6 2 tL'lf3 g6 3 i.f4 i.g7 4 e3 d6 5

26 bxc8'ii' :axeS 2 7 i.xg3 fxg3 28 �xa7 i.e2 o-o 6 h 3 t2Jc6 7 o-o tt::\d 7?! 8 c3 e5 9
'ii'h 6 29 .l::tfe1 .l:.a8 30 'ii'c 5 i.fB 0-1 i.. h 2 f5 10 a4 'ii'e 7 11 b4 e4?!

146

Opening up the game for White's
dark-squared bishop. This is clearly bad
since Black has little chance of achiev­
ing .. .fS-f4 to close the diagonal again.
12 ltJfd2 ltJd8 13 a s a6 14 c4! cs?
This stops White's planned 1 5 c s ! ,
but i t weakens too many squares, most
especially dS and d6.
1S ctJc3 cxd4 16 ltJds "ikf7 17 exd4 ltJe6
18 .i.xd6 ltJxd4 19 .i.xf8 ltJxf8 20 ctJb3!?
Giving back the exchange is not
necessary, but it does help to clarify the
position.
20 ... ltJxb3 21 �xb3 Si.xa1 22 .l::t xa1 .i.e6
23 .l:td1 .:!.d8 24 �c3 .i.xd s 2 s l:txd s
2S .. Jlxd S?
Illustrative Games
This gives White's queen easy ac­
cess, though Black's position was lousy
anyway.
26 cxd s 'iig 7 27 �cs 'ii'e s 28 .i.c4 'ii'a 1+
29 �h2 'ifes+ 30 g3 Wg7 31 'ii'b 6 �h6
32 'ir'd8? c;i;g7??
My computer kindly tells me that
Black can save the game with
32 ... ctJd7 ! ! 33 'ir'xd7 'ir'd4. The text move
is hopeless.
33 d6 it'b2 34 'ife7+ �h6 3S 'i¥xf8+ �hs
36 Si.f1 e3 37 d7 e2 38 .i.g2 1-0
How to fight in the critical lines
of the 1 d4 Dark Knight System
Gam e 1 6
H.Kmoch-F.Yates
H a sti ngs 1927/28
1 d 4 ctJf6 2 ctJf3 g 6 3 C 4 .i.g7 4 g3 0-0 S
.i.g2 d6 6 o-o ltJc6 1 d s ltJb8
Here 7 ... ttJas ! ? is recommended by
theory, and is objectively best, though
Yates makes a good case for the text
move - and note that we would nor­
mally not have this option, reaching
the game position by 1 d4 ltJc6 2 ctJf3
d6 3 g 3 g6 4 ds ttJbs s .i.g 2 .i.g 7 6 o-o
ltJf6 7 c4 o-o or one of many similar
move orders.
8 ctJC3 aS! 9 ctJd4 ctJa6 10 h 3
White prevents 1 0. .ttJ g 4 but, a s I
.
am so fond of saying, a tempo is a
tempo. This also necessitates a later
�h 2.
10...ttJcs 11 Si.e3 .i.d7 12 'iii'd 2 �c8 13
�h2 eS
147
Th e Dark Kn ight System
14 tt::ld b5
White should have preferred 14
dxe6, though Black is fine after
14... tt:Jxe6, or even 14 ... .1txe6 1 5 tt:Jxe6
�xe6 16 .ltxcs dxcs 17 .ltxb7 .l:.ad8 18
�f4 'ifb6 19 .ltf3 'ii'xb2, when Black's
strong bishop and activity m ake up for
the crippled queen side.
14 ... b6 15 .l:.ac1 tt::l h 5 16 b3 f5
Yates's play has a very modern feel
to it, but his next move is too optimis­
tic.
11 f4 g5?! 18 .ltxc5?!
After 18 fxg s f4 19 .ltf2 fxg 3+ 20
.ltxg 3 tt:Jxg 3 2 1 Wxg 3 White at least has
a pawn for his trouble.
18 ... gxf4?
Yates bluffs again with this fake
(and unnecessary) zwischenzug. How­
ever, the rest of the game he conducts
masterfully.
19 gxf4? bxc5 20 e3 'iii> h 8 21 .ltf3 tt::lf6 22
�g1 'iid 8 23 :g3 .lth6 24 .l::tf1 .lte8 25
tt:Je2 .ltg6 26 .l:tgg1 "W/e7 27 tt:Jbc3 l:tg8
28 tt::lg 3 �af8 29 .ltd1 �g7 30 Itg2 .l::tfg8
31 �ff2 tt::ld 7 32 tt:Jce2 'if'h4 3 3 tt::l g 1?!
exf4 34 exf4 tt::lf6 35 Wh1 l:te7 36 .:!.f1
148
l::tg e8 37 tt::\ 1 e2?
The losing mistake, though White
was about waist-deep in it anyway.
37 ... tt:Jg4 38 l::tf3 �xh3+ 39 'iiig 1 .ltg7 40
l:td3 l:i.e3 41 lbe3 Uxe3 42 .ltc2 .l:.xg3!
43 t'Dxg3 .i.d4+ 44 <;t>f1 tt::l e 3+ 0-1
s tt:J as!? - a strong alternative
•••
to s tt:J bs
•••
Gam e 1 7
A.Gal liamova-M.Krasenkow
Kosza l i n 1997
1 d4 d6 2 tt::lf3 g 6 3 g 3 .ltg7 4 .i.g2 tt::l c 6 5
d 5 tt:Ja 5
lt is possible to consider this square
any time White h as fianchettoed,
though it is more usual with a white
pawn on c4 to harass. As it turn s out,
White does not enjoy the omission of
c2-c4, as the ds-pawn becomes a tar­
get. In any case, Black must be active
quickly on the queenside to justify the
knight's position . The typical method is
... c7-c5 to gain space and make sure the
knight doesn't get trapped - in this

gam e Black has other ideas.
6 o-o c6 7 e4 lbf6 8 'ii'e 2 cxd 5 9 exd5
.i.d7 10 lbfd2 b5 11 b4 lbc4 12 lbxc4
bxc4 13 �xc4 .l:!.c8
14 'it'e2?!
lt was already dangerous to win the
pawn, and now White chooses the
wrong retreat: 14 "ii'h 3 lbxd5 ? ! 1 5 .i.xd5
.i.xa1 16 .i.xf7+ �f8 17 c3 is good for
White.
14...lbxd5 15 .i.xd5 .i.xa1 16 c3 o-o 17
.i.g5 .i.c6?! [17 ... .i.f5 !] 18 .i.xc6 l:txc6 19
.i.xe7 l:te8 20 'iff3 �d7 21 .i.f6 .l:.a6 22
lbd2 .Uxa2 23 �f4 �f5 24 'ifxd6
24 ... .i.b2?!
Black should make space for his
Illustrative Games
king with 24 ...h 6 or 24 ... h 5 . Against the
text White i s right back in the game
after 25 .i.d4.
25 lDb3? 'i!Ve6 26 �xe6 .Uxe6 27 .i.d4 .l:!.a3
28 lba 5?
2 8 lbd2 was a better try, intending
28 ... .i.xc3? 29 lbb1.
28....i.xc3 29 lbc4 l:ta1! 30 .i.xc3 .l:txf1+
31 �xf1 l:i.c6 32 .i.d4 .:.xc4 3 3 .i.xa7
l:!.xb4 34 h4 �g7 3 5 �g2 �6 36 i.e3
�e5 37 �3 h5 38 �g2 �e4 39 i.c5 .l:tb5
40 i.e3 �d3 41 .i.f4 �e2 42 .i.e3 l:!.b3 43
i.c5 l:!.f3 44 .i.d4 f5 45 �g1 .l:.d3 o-1
Black, given plenty of rope,
tries to hang himself but fails
Game 18
R.Aghasaryan-A.Chibukhchian
Kaj a ra n 2 0 1 1
1 d4 lbc6 2 c4 e5 3 dxe5 lbxe5 4 e3
White's unambitious play has al­
ready left him with no trace of an ad­
vantage. Frankly, the rest of the game
sees Black try too hard to win, for
which he was unjustly rewarded.
149
Th e Dark Kn ight System
4... ttJf6 5 tiJc3 .tb4 6 .td2 o-o 7 ..te2 c6
8 tiJf3 d6 9 o-o .l:!.e8 10 �b3 ..tas 11
.:tad1 ..tc7 12 ttJd4 'iie 7 13 h3 ttJg6 14
.l::i.fe1 ttJh4 15 g3 ttJg6 16 ..tf1 hs 11 f4?!
17 ... h4?! [17 ... ..tb6 18 ttJa4 ttJe4!] 18 g4
.tb6 19 ttJa4 .tc7?! 20 e4 ttJxe4?! 21
.tg2?!
2 1 ttJc3 ! was stronger, with the idea
21 ... 1!Vf6 22 ttJxe4! 'ikxd4 23 .te3 ! 'ii'x e4
24 .tf2.
21 ...ds? [21 ...'ii'f6 !] 22 cxdS iif6 2 3 gS
1!Vxd4+ 24 ..te3 'ilt'xd1 2 5 'ii'xd1 cxd s 26
fs? [26 'ii'x ds] 26 ...-txfs 27 1!Vxd s ttJd6
28 l:tc1? l:txe3 29 :xc7 .l:l.d3 30 'ii'c s?!
ttJf4! 31 .tf1? .l:!.d1 3 2 ttJc3 ttJxh3+ 33
Wg2 :d2+ 34 Wh1 ttJe4 o-1
1 50
Crushed on the d-file -
don't let this happen to you!
Game 1 9
A.Hoffman-A.Ferna ndez
Ma r d e l Plata 1996
1 d 4 ttJc6 2 c 4 es 3 ds .tb4+ 4 ..td2
.txd2+ 5 1!Vxd2 ttJce7 6 d6 ttJc6?
Circum stances may change, but
right now the best square for thi s
knight is e 7 , from which i t controls d s
and h as the option t o attack d 6 with
... ttJfs . Black should only consider ... ttJc6
after White has played e2-e4, clamping
on dS but opening a hole on the d4-
square. Even after the correct 6 ... cxd6,
Black can still be crushed in this fash­
ion if he is not alert to all the methods
of counterplay, so study Position Two.
In this game, Black never got on the
board.
7 tiJc3 cxd6 8 ttJbs tiJf6 9 ttJxd6+ 'it>f8 10
tiJf3 h6 11 e3 e4 12 ttJd4 'ii'a s 13 1!Vxa s
ttJxa s 14 ttJ4b5 b6 15 .te2 We7 16 o-o-o
tiJb7 17 .l:[d4 ttJcs 18 ttJc3 ttJe6 19 l:td2
a s 20 .l!thd1 ttJcs 21 ttJds+ ttJxds 22

:Xd5 g6 23 lbxe4 lbxe4 24 .l:.e5+ �d8
25 l:!.xe4 .l:.e8 26 .l:.ed4
26 ....l:.e6 27 .i.f3 .l:Ia7 28 h4 h5 29 l:!.f4
�e7 30 ii.d5 .l:!.f6 31 .li!xf6 'lt>xf6 3 2 .l:.d4
d6 33 �d2 �e7 34 �c3 .l:.c7 35 a4 .i.e6
36 .l:.d2 l:i.c5 37 b3 'ii;>d 7 38 f3 .l:.c7 39 e4
!:i.c8 40 e5 .ixd 5 41 .l:txd 5 l:i.e8 42 .l:.xd6+
rj;; c 7 43 'ii;> d 4 .l:!.g8 44 f4 .l:!.e8 45 'it>d 5 .l:.e7
46 .l:!.c6+ Wb7 47 e6 1-0
Middlegame and endga me
ideas in the 3 .i.b4+ structure
•••
Gam e 2 0
H. Titz-C.Ba rlocco
Dresd e n 2 004
1 C4 lbc6 2 d4 e5 3 d5 i.b4+ 4 .id2
.ixd2+ 5 �xd2 tbce7 6 lbc3 d6 7 e4 f5 8
exf5 .ixf5 9 i.d3 lbf6 10 lb ge2 o-o 11
o-o i.xd3 12 iVxd3 tbh5 13 g3 �d7 14
f3 a6 15 .l:!.ad1 l:.ae8 16 lbe4 h6 17 c5
lbf6 18 tb2c3 lbf5 19 b4 lbxe4 20 lbxe4
g 5 21 'it>g2 lbd4
Thi s is the position where we left off
when analysing in the theoretical sec­
tion.
Illustrative Gam es
22 cxd6 cxd6 2 3 .l:!.c1 l::tf7 24 lbd2 'lt>g7
2 5 tbb3 lbxb3 26 axb3 �f5 2 7 'ii'xf5
.l:!.xf5 28 .l:!.c7+ .l:!.f7 29 l:Ixf7+ 'lt>xf7 30
�cl J::[ e 7 31 �f2 e4 32 f4 gxf4 3 3 gxf4
e3+ 34 'lt>e2 'it>f6 3 5 :c3 l:te4 36 l:txe3
'lt>f5 3 7 'it>d3 .:.xb4 38 �c3 l:.xf4 39 l:!.e6
�h4 40 .l:!.xd6 We5 41 .l:.d7 b5 42 d6 b4+
43 'ii;>d 3 'ii;>d 5 44 .l:!.a7 .l:!.h3+ 45 'it>c2
.l:i.xh2+ 46 Wd3 l:!.h3+ 47 'lt>d2 'it>xd6 48
.l:!.xa6+ 'it>d5
49 l:!.b6? 'it>c5
49 ... l:!.xb3 (or on the next move) so
llxh6 Itc3 reaches the winning end­
game rather sooner.
50 .l:.f6 .l:!.h4? 51 'lt>e3 'it>d5 52 .l:tf5+ �e6
53 l:tb5 Wd7 54 .l:.b6 h5 55 'lt>d3 l:!.g4 56
151
Th e Dark Kn ight System
.l:!.h6 h4 57 'it>e3 'it>e7 58 'it>f3 .l:!.g3+ 59
'it>f4 .l:txb3 60 l:txh4 'it>d6 61 l:.h5 'it>c6 62
'it>e4 llc3 63 'it>d4 'it>b6 64 l:tg 5 'it>a6 65
�h5 �c1 66 .l:!.hB 'it>bs 67 'it>d3 'it>a4 68
'it>d2 �c7 69 l:th3 b3 70 �h6 'it>a3 71 :b6
'it>b2 72 'it>d3 �d7+ 73 'it>c4 'it>c2 74 'it>b4
b2 75 l::t c 6+ 'it>d2 0-1
A warning: beware the static
king - a nice attack by White
Gam e 2 1
M.Gu revich-D.Zoler
Antwe rp 1998
1 d4 l2Jc6 2 C4 eS 3 dS .i.b4+ 4 l2Jd2
lbce7 5 a3 ..txd2+ 6 .i.xd2 d6 7 e4 fs 8
'ii'h 5+1?
8 ...'it>f8?1
8 ... g 6 9 'ii'h 4 fxe4 i s not the most
fun ever, but it does leave Black with a
pawn for his trouble. The text m ove is
less accurate and demands m ore preci­
sion from Black in the coming moves in
order to avoid a disaster like the one in
the game.
9 f3 [9 exfS ! ] g...lbf6 10 'ii'h 4 lbg6 11
1 52
'iif2 'it>f7?1
11 .. .fxe4 12 fxe4 'it>f7 will allow
Black to finish castling artificially with
approximate equality.
12 exfs .i.xf5 13 g4 .i.d7 14 h4 :es 15
h s lbf4 16 o-o-o .i.a4 11 :te1 bs 18 lbh3
lbxh3 19 :txh3 'iid 7 20 :th4 bxc4 21 gS
lbxd s 22 ..txc4 �c6 2 3 'it>b1 .i.b5 24
.i.a2 'it>fB 25 f4 e4?
Black's position has been bad for a
long time, but this ends the game.
26 'ii'd 4 .i.d3+ 2 7 'it>a1 e3 28 h6 l:.e5 29
hxg7+ <3te7 30 fxe5 exd2 31 exd6+
'it>xd6 32 l:th6+ .i.g6 3 3 'iix d2 1-0
No problems for Black
after 3 .i.b4+ 4 ..td2
•••
Game 22
E.Arlandi-M.Lanzani
Sa n Ma ri n o 1998
1 d4 l2Jc6 2 C4 e5 3 d S .i.b4+ 4 l2Jd2
lbce7 5 a 3 .i.xd2+ 6 .i.xd2 d6 7 e4 f5 8
exfs ..txf5 9 lbe2 lbf6 10 lbg3 .i.g6 11
..te2 o-o 12 o-o

12 ... t2Je4 [12 ... t2Jfs] 13 t2Jxe4 .i.xe4 14 f3
.i.fs 15 .i.e3 b6 16 'ii'd 2 �d7 17 l:.ac1
Wh8 18 ltfd1 as 19 b3 h6 20 'ifb2 .i.h7
21 Wh1 t2Jg6 2 2 .i.f1 t2'lf4 2 3 g3 t2'lh3 24
.i.g2 tL'lgs 25 I:!.f1 e4 26 .txgs exf3 Yz-Yz
Fight for the e4-square
or suffer the consequences!
Came 23
G.Grigore-P.Brochet
C reon 1999
1 d 4 t2Jc6 2 c 4 es 3 d s .i.b4+ 4 t2Jd2
t2Jce7 5 a3 .txd2+ 6 i.xd2 d6 1 e4 fs 8
exfs .i.xfs 9 t2Je2 t2Jf6 10 t2'lg3 o-ol?
Illustrative Games
Black offers his other bishop, an in­
vitation White is wise not to accept: 11
tL'lxfs ? ! tL'lxfs 12 .td3 e4 13 .tc2 'i!Ve8 14
0-0 'i!Vg 6 and 1S ...I:!.ae8 with a harmoni­
ous and menacing position . Who's
afraid of the bishop pair?
11 .i.e2 �e8 [11 ... 'ii'd 7] 12 o-o t2Je4
[12 ... .tg6] 13 .te3 t2Jxg3 [13 ... 'ii'g 6] 14
hxg3 b6
Although this was also played in
Game 2 2, I don 't think it helps Black. If
he wants to, White can break through
with b2-b4 and c4-c5 anyway, and now
Black h as to worry about penetration
on both the c7- and c6-squares. Fur­
thermore, it rules out the possibility of
Black breaking with ... c7-c6. 14 ... SLd7 is
more flexible.
15 g4 .i.d7 16 .td3
Taking advantage of tiny inaccura­
cies, White has built an advantage, tak­
ing over the b1-h 7 diagonal, along with
the critical e4-square. lt's not so easy
playing a Grandmaster. Remember to
fight for e4! As it happens, Black still
had plenty of chances.
16 ...t2Jg6 17 ..te4 t2'lf4 18 g3 'ili'e7 19 f3?1
1 53
Th e Dark Kn ig h t System
[19 'ifc2 ! ) 19 ...tt::l h 3+ 20 �g2 tt::l g s 21
..txgs ifxgs 22 'ifd3 g6 23 l:th1 as 24
l:th4 �g7?! [24...ltf7] 2 5 ltah1 �h8 26
'ii'c 3 h6 2 7 b3? [27 f4!) 27 .. Ji'f6 28 f4
gs?!
28 ... exf4 29 'ii'xf6+ �xf6 30 �xh 6
llxh6 3 1 l:txh 6 �g8 holds the balance.
29 fxgs 'it'xgs 30 .l:.hs! 'it'xg4 31 �e3
�7 3 2 llsh4 'ii'g 7 3 3 .l:!.xh6 :xh6 34
�xh6 �h8 35 �g6 'ir'f8 36 'ii'g s �e8 37
�f6 'ii'e 7
38 'ii'g 6+?!
38 .tfs maintains more pressure.
Presumably the mistakes here are due
to a tim e scramble.
38 ...�d8
1 54
39 �f7 'it'e8??
39 ... ..te8! holds.
40 'it'f6+?
40 l:r.xd7+! win s immediately.
40 .'it>c8 41 ..th7 'ifd8 42 'iWg7 l:te8 43
.•
g4 ..txg4 44 'ifxg4+ �b8 45 l:tg7 �f8 46
'Wd7 'ii'x d7 47 l:txd7 �f4 48 .l:td8+ �a7
49 a4 e4 so .:tg8 .l:.h4 51 ..tfs .:tf4 52
..td7 e3 5 3 l:.e8 1-0
3 ... d6!? an interesting
-
a lternative to 3 . . e6.
Came 24
M.Drasko-Z.Mestrovic
Bos n i a n Tea m
C h a m pio n s h i p 2003
1 d4 tt::l c6 2 d S tt::l e s 3 e4 d6
Previously Mestrovic h ad success­
fully played 3 ... e6, but m aybe he no
longer believed in it? The text m ove
winds up leading to a position consid­
ered in Chapter Six (see 6 tt::lf3 in line B).
4 f4 tt::ld 7 5 tt::l c 3 c6 6 tt::lf3 cxds 7 exds
tt::l gf6 8 ..te3 g6 9 ..te2 ..tg7 10 o-o o-o
11 'i!i'd2 a6 12 a4 'ikc7 13 ..td4

13 ...4Jb6?!
The position was equal, until this
time-waster hands White an advan­
tage. 13 ... 4Jcs i s better.
14 as 4Jbd7 15 4Ja4 4Je4 16 'iie 3 �xd4
17 4Jxd4 4Jef6 18 4Jc3 'ii'c s 19 l:tfd1 bS
20 axb6?! [20 b4] 20 ... 4Jxb6 21 b4 Yz-Yz
Perhaps nobody was in the m ood
for a fight - the position is equal again
anyway.
GM uses Diebl's novelty 4 exd s ! •.•
Game 25
V.Erdos-R.Ra pport
H u n ga ri a n Tea m
C h a m pions h i p 2012
1 d 4 4Jc6 2 d s tZJes 3 e 4 e 6 4 f4 exd s 5
fxes 'ifh4+ 6 'it>e2 'ii'h s+ 7 �d2 'ii'h 6+ 8
�c3 'ii'c 6+ 9 �d2 'iVh6+ 10 �d3 'ii'a 6+
11 �d2 'i'h6+ 12 �e1 'i'h4+ 13 �d2
I can't help but get the feeling these
guys are m essing around. Black's next
constitutes a risky attempt to play for
the win .
13 . .Ji'f4+ 1 4 � c 3 'ii'xes+ 15 'iVd4
Illustrative Games
1S ... .tb4+ [1S ... 'ii'x e4 ! ?] 16 �d3 dxe4+
17 'iixe4 'ii'xe4+ 18 'it>xe4 4Jf6+ 19 �e3
.tcs+ 20 �f3 bs 2 1 .td3?!
According to Houdini, White is bet­
ter after 21 �e3 ! .tb7+ 22 �e2 �xe3
2 3 �xe3, though it hardly looks like a
safe advantage. Then again, the text
move doesn't look safe either, and
White clearly has no advantage there.
23 ... .tb7+ 22 �g3 .td6+ 23 �f2 4Jg4+
24 �1 4Jxh2+ 25 �f2 4Jg4+ 26 �1
4Jh2+ 2 7 �f2 4Jg4+ Yz-Yz
An alternative to 4 fxe6!? -
...
the "endgame" with 4 dxe6 ...
Gam e 2 6
S.Gordon-N.Short
B riti s h C h a m pio n s h i p,
S h effie l d 2011
1 d4 4Jc6
I must say, this warms my heart.
2 dS tZJes 3 e4 e6 4 dxe6 dxe6 5 'i'xd8+
'it>xd8 6 f4 4Jc6 7 c3 .tcs 8 4Jf3 as 9 a4
4Jh6
In this position, 9 ... 4Jf6 is also good.
1 55
Th e Dark Kn ight System
However, it is nice for this knight to
have a stable home, controlling impor­
tant squares. Short's 10 .. .f6 en sures
th at he will not be squeezed on the
kingside, as does 14 ...h s.
10 i.d3 f6 11 'iii>e 2 ri;e7 12 lt:la 3 i.d7 13
lt:lc4 tiJf7 14 h4 hs! 15 i.e3 i.xe3 16
lt:lxe3 b6 17 l:thgl gS 18 g3 g4 19 lt:ld4
lt:ld6 20 l:.ac1 .l:Ihc8 21 b3 lt:le8 22 l:tgfl
lt:ld8 2 3 es fs 24 �fd1 lt:lb7 2 5 .l:.d2 lt:lcs
26 l::lc d1 lt:lg7 27 lt:lc4 lt:lxd3 28 l:.xd3
.l::!.d 8 29 lLla3 lLle8
30 'iii>e 3 .l:!.ac8 31 l:f.3d2 .l:.b8 32 l:!.d3 l:.bc8
33 .l:.1d2 .l:tb8 34 'iii>f2 l::t bc8 35 'iii>e 2 .Ua8
36 We3 l:rab8 37 �e2 .l::[a 8 38 'iii>f2 .l:tac8
39 lld1 .l::!. b 8 40 l:.3d2 .l:!.bc8 41 'iii>e 1 l:ta8
156
42 l:.td3 l:!.ab8 43 lhd2 l:ta8 44 'iii>d 1 lt:lg7
45 ri;c2 lt:le8 46 'iio> b2 lt:lg7 4 7 lt:lc4 ..te8 48
'it>a3 l:.dc8 49 lt:lbs i.c6 so l!td4 lt:le8 51
lt:lbd6 �d8 S2 lt:lxe8 .l:.xd4 53 �xd4 Yz-Yz
Wow, that was boring ! Are you sold
on 4 .. .fxe6 yet?
A better advertisement
for 4 .. dxe6
.
Game 2 7
M.Gu revich-M.Rohde
Ph i l a d e l p h i a ( b l itz) 1989
1 d4 lt:lc6 2 d s lt:les 3 e4 e6 4 dxe6 dxe6
5 'i¥xd8+ Wxd8 6 i.f4 lt:lg6
Although ... lt:lc6 is normally the best
retreat for thi s piece, the difference i s
n ot so great that Black would decline
the tempo gain. Here 6 ... i.d6 ! is also
good, threatening 7 .. lt:ld3+ or 7 . . lt:lf3+.
. .
7 i.e3 lt:lf6 8 lt:lc3?!
This careless move allows Black to
h arass White's poor dark-squared
bishop further, disrupting White's de­
velopment. 8 f3 was called for.
8 ... lt:lg4 9 i.d2 i.cs 10 lt:lh3 i.d7 11
 Illustrative Games

�e2 tt::lf6 12 0-0-0 rJ;;e 7 13 f4?! dxe6 fxe6 6 e4 ii.c5?! [6 ... ds ! ] 7 ii.d3?!
Again 13 f3. lbh6 8 'i!Ve2 o-o 9 g3 a6?! [9 . dS ! ] 10
. .

13 ii.c6
..• tt::lc 3 b5 11 e5 ii.b7 12 tt::le4 ii.b6 13 ii.d2
tt::lfs 14 0-0-0 h6?! 15 .l::t hf1 c5?!

By now Black is a little better -

White's position i s too loose. Black gets tired of sitting around
14 i.d3 tt::l h 4 15 e5? waiting for the brutality, but thi s only
Thi s i s a terrible idea, releasing the makes thing s worse due to the new
strong c6-bishop. White n eeded to let hole on d6.
go of the g-pawn and seek counterplay 16 tt::ld 6 ii.d5
against Black's king with 1 5 fS ! .
15. ..tbd5 1 6 tt::le4 ii.e3 17 c 4 ii.xd2+ 18
lbd2 tt::l b4 19 ii.b1 l2Jxa2+ 20 rJi;d1 l:.hd8
21 g3 tt::lf3 22 l:!.xd8 J::!.x d8+ 23 'it>e2 tt::ld4+
24 'iW2 tt::l b4 25 .:.c1?! lbb3 26 �c3 tt::ld 2
27 tt::lx d2 .l:.xd2+ 28 <Ji>e3 l:!.xb2 29 ii.xh7
.l:.xh2 30 l2Jg5 .Ug2 31 tt::le4 a6 32 <Ji>d4 b6
33 tt::lg 5 tt::lc 2+ 34 <Ji>d3 tt::le 1+ 3 5 <Ji>d4 as
36 Si.g8 a4 37 ii.xf7 ii.d7 38 cs 0-1

Don't play 6 -tcs ?!?!

•••

17 c4?
Game 2 8 White should play 17 ii.xfs !, main­
A.Onischuk-I.Shkuro taining the wonderful d6-knight.
U kra i n i a n Tea m 17 bxc4 18 tt::lxc4 tt::lfe7?1 [18 ... l:tb8] 19
..•

C h a m p i o n s h i p 2 009 h4 l2Jf5 20 I!g1 tt::lg e7 21 g4 l2Jd4 22

l2Jxd4 cxd4 23 l:tgf1 l::t b 8 24 rJi;b1 i..c 5 2 5
1 d4 tt::lc 6 2 ds l2Je5 3 f4 tt::lg 6 4 tt::lf3 e6 5 f5 tt::lc 6 26 g5 exf5 27 gxh6 YWxh4 2 8

157
Th e Dark Kn ight System
hxg7 .l:.f7 29 ltJd6 i.xd6 30 exd6 i.e4 31
.l::tf4 i.xd3+ 3 2 'ili'xd3 'iff6? 33 l:Ih1 l:txg7
34 .l::txf5 �e6 35 �h3 1-0
In spite of the inaccuracies, there is
feeling of inevitability about the result
of this gam e due to White's uncon­
tested superiority in the centre and on
the kingside.
4 i.e3 fs!? - take one
Came 29
A.Beliavsky-A.Miles
E u ro pea n C h a m pio n s h i p
Sa i nt Vi n ce nt 2000
1 d4 tt:lc6 2 e4 e5 3 d 5 ltJce7 4 i.e3 f5
5 f3
White resigns himself to allowing
Black a good King's Indian with an
early and easy .. .f7-fS. In stead, if 5 ltJc3
tt:lf6 6 tt:lf3 d6 7 exfs c6 (7 ... a6 ! ?),
White's centre disintegrates. (In fact,
White can still play for an advantage,
but it is easy to see why thi s was not
appealing.)
5 ...tt:lf6 6 liJh 3 d6 7 ltJf2
1 58
The great Beliavsky has not suc­
ceeded in casting any doubt on Black's
idea. Now Miles should just get on with
his comfortable King's Indian : 7 ... g6.
The c-pawn can wait.
7 ... c6 8 c4 c5 9 g3 g6 10 ltJc3 i.g7 11 g4
f4 12 i.d2 g5 13 b4 b6 14 bxc5 bxc5 15
i.d3?!
Doubtful . 1 5 h4 looks more to the
point.
15 ... h5! 16 h3 �f7 17 �e2 ltJg6 18 'i!Va4
i.f8 19 liab1 i.e7 20 i.c2
20 ... hxg4?!
I suppose it's a draw either way, but
there is no disadvantage to maintain­
ing the favourable tension . After this,
I 'm not sure who's playing for the win,
or why.
21 hxg4 .:.xh1 22 l::tx h1 i.d7 23 "ii'a 3
l:i.b8 2 4 tt:lb1 'ili'b6 25 i.c3 l:l h 8 26 .l:g1
i.c8 27 ltJd2 'ii'a 6 28 'ii'x a6 i.xa6 29
l:tb1 i.d8 30 i.a4 i.c7 31 i.b5 i.c8 3 2
ltJf1 <;t;e7 3 3 i.c6 ltJh4 34 ltJd2 ltJg2 3 5
i.b7 i.d7 36 i.a6 i.b6 37 ltJf1 ltJh4 3 8
i.e1 ltJg2 39 i.d2 ltJh4 4 0 i.b5 i.c8 41
i.e1 ltJg2 42 i.d2 ltJh4 43 l:f.b3 'it>d8 44
i.c6 Yz-Yz

4 i.e3 fS !? - take two
Came 3 0
D.Cam pora-A.Miles
Sevi l l e 1993
1 e 4 tLlc6 2 d 4 es 3 d s t2Jce7 4 .i.e3 f s 5
f3 tLlf6 6 tLlc3 d6 7 �d2 g6 8 o-o-o .i.g7
9 h3?! f4 10 .i.f2
The pawn structure on the kingside
makes it impossible for White to open
lines there without creating a complex
of weak squares and pawns. Black does
have ways to create play on the queen­
side, as we will see.
10...0-0 11 'i£?b1 a6 12 tLlge2 i.d7 13
tLlc1 bs 14 a3 'ii'b 8 15 tLlb3 :d8 16 tLla s
cs 17 dxc6 tLlxc6 18 tLlxc6 i.xc6 19 i.h4
l:Id7 20 i.xf6 i.xf6 21 tLld s .i.h4 22 tLlb4
.i.b7 2 3 C4
White tries to free his terrible fl­
bishop, but in the process open s lines
again st his own king.
23 ...'i£i>g7 24 .i.d3?! a s 2 5 tLla2?! [25
tLld5] 2 S ...b4! 26 axb4 axb4 2 7 'it'xb4?!
.i.f2?!
Black's "bad" bishop turns into a
Illustrative Gam es
mon ster. Nevertheless, Black should
insert 27 .. .'W/a7 2 8 t2Jc3 .i.c6 29 �3 and
then 29 ... .i.f2 perh aps, though Black
has other good ways to continue the
attack here.
28 t2Jc3 .i.d4 29 �c2 �c8 30 tLlbs i.cs
31 'i*'b3 i.a6 3 2 .l:ta1 .i.e3 33 :hd1
White has actually done an excel­
lent job repairing his position and
Black should no longer be able to break
down the fortress.
33 ....l:tb7 34 tLlxd6??
Oops!
34....l:.xb3 0-1
White realizes too l ate that 35 tLlxc8
will be m et by 35 ... .l:r.ab8 ! intending 3 6
ltxa6 J::!.xb2+ 3 7 W c 3 :8b3 m ate. Of
course White can defend the mate, but
then he will remain a piece down .
Fighting the eternal ttJgs
Came 31
M.Kravtsiv-K.Ta rlev
Evpatoria 2007
1 e4 tLlc6 2 d4 es 3 d s tLlce7 4 tLlf3 tLlf6?
1 59
Th e Dark Kn ight System
Perhaps this i s a database error? In
any case, the reader should know that
this is not an acceptable move order: 5
l'Llxes l2Jxe4 6 'ii'e 2 l'Lld6 (not 6 ... l'Llf6? 7
d6 ! cxd6 8 l'Llc4 "fic7 9 .i.f4 an d Black is
busted) 7 l'Llc3 already leaves Black with
no good way to complete his develop­
m ent.
5 l'Llc37 l'Llg6 6 h4 h5 1 l'Llg5!7 .i.b4 8
.i.e2 d6 9 'ii'd 3 .i.d7 10 a3
10 .....txc3+
As I m entioned in the theoretical
section, it is better to retain thi s strong
bishop with 10 ... .tcs. This is particu­
l arly true since Black will not inflict any
structural damage with the exchange.
White's queenside stays mobile and his
bishops should come into their own
sooner or l ater. Frankly, from thi s point
on, I do not care for Black's position
until the g ame is nearly over.
11 'ti'xc3 c6 12 'ii'b 37!
lt m akes more sen se for White to
open the game with 12 dxc6.
12 ...cxd5 13 exd 5 o-o?! [13 .. -'i!Vc7] 14
l2Je6! fxe6 15 dxe6 .i.xe6 16 'i¥xe6+
rJi>h8 17 .i.g5 'ii'b 6 18 o-o-o 'ii'xf2 19 i.f3
1 60
e4 20 .i.xf6 l:!.ae8
21 'it'd77! [21 .i.e7 !] 21 ... gxf6 22 .txh 5
l:te7 2 3 �xd6 'ii'f4+ 24 'iWd2 'iixd2+ 2 5
l:!.xd2 l'Llf4 26 .i.d1 e3 27 .Ud4 l2Jxg2 2 8
.te2 f5 29 rJi>dl .l:.ee8 30 il.f3 .l:.d8 31
l:!.xd8 .:!.xd8+ 3 2 'iii>c 1 l:td2 3 3 .l:Ig1 l'Llxh4
34 .txb7 .:td77 [34 ... .l:.h 2] 35 .l:th17! [35
.i.a6] 35 ....l:.xb7 36 .l:f.xh4+ rJi>g7 37 rJi>d1
'iii>f6 38 We2 rJi>g5 39 l:td4 .l:.e7 40 c477 f4
41 l:rd1 .lah7 42 Wf3 .l:!.h3+ 43 'ito>e4 e2 44
.l:le1 l:te3+ 45 rJi>d4 .l:.e8 0-1
Game 3 2
H.Meissner-A.Miles
E u ro pea n C u p, S l o u g h 1997
1 e 4 l2Jc6 2 d 4 e5 3 d5 l2Jce7 4 l'Llf3 l2Jg6
5 h4 h5 6 g3 ii.c5 1 il.g5 f6 8 i.d2 d6 9
l'Llc3 .i.d7
Or 9 ... a6 - Black must play to pre­
serve the dark-squared bishop.
10 .te2 l'Ll6e7
Miles prepares to defend the h­
pawn, which was about to drop off.
11 l'Llh2 g6 12 �cl c6 13 .te3 .txe3 14
'ifxe3 'ii'b 6 15 'i!Vxb6 axb6 16 dxc6 bxc6
 Illustrative Games

11 o-o bs 18 a3 t2Jc8 19 .l:[fd1 t2Jh6 20 Yikes! Black sure has a lot of centre
tbf1 rri;;e 7 21 t2Je3 pawn s !
21 a4. 30 l:ta1 rri;;d 6 31 l:!.e2 d4 3 2 'ri;;f2 cs 3 3 c 3
21 tbb6 22 l:tac1 t2Jg4 23 i.xg4 i.xg4
••• li.d7 34 l:!. d 2 'lii' e 6 3 5 rri;;e 2 l:.ab8 36 'lii'e 1
24 .l':!.d3 li.a4

Very slowly White is being out­ 37 .:a2?

played. Miles m akes the most out of a White grows tired of the thankless
small thing - his single, uncompro­ defensive task.
mised pawn group. White needs to or­ 37 1i.b3 38 :a1 1i.xd1 39 rri;;xd1 :xb2
•••

ganize a pawn break, but he is showing 40 .l:txb2 :xb2 41 a4 .l:i.g2 42 .l:.a3 rri;;d 6
no inclination to do so. 43 a s rri;;c 7 44 a6 '>t>b8 45 a7+ rri;;a 8 46
24 t2Jc4 25 t2Jxc4?!
••• .l:ta6 .l:i.xg3 47 !:txf6 rri;;x a7 0-1
Now Black h as b2 as a target. The
immediate 2 5 :Lb1 was better.
25 bxc4 26 l:!.e3 .l:i.hb8 27 .l:i.b1 l:.b7 28
••• Came 33
f3 i.e6 29 tbd1 d S L.Christiansen-J.Benjamin
U S C h a m pion s h i p,
Seattle 2000

1 e4 t2Jc6 2 d4 es 3 d s t2Jce7 4 t2Jf3 t2Jg6

s h4 hs 6 1i.gs t2Jf6 1 t2Jc3 1i.b4 8 a 3
Thi s seems odd. Doesn't White have
anything better to do? Indeed 8 tbd2 is
better, but one way or another White
does need to force our dark bishop off
the board or we will be happy to pre­
serve it with 8 .. a6.
.

1 61
Th e Dark Kn ight System
8 ...i.xc3+ 9 bxc3 c6!
I am always reluctant to change the
pawn structure on the queen side after
saddling White with the doubled
pawn s - after all, the structure is al­
ready favourable, right? There i s some
logic there, but better players than me
have demon strated time and time
again that Black should be willing to
play on both sides of the board in this
type of position, particularly since
White has already m ade inroads on the
king side. Therefore 9 ... c6 !, the only
convenient pawn lever for eith er col­
our. Black targets the centre and pre­
pares to break the annoying pin on f6.
10 c4 d6 11 lLld2 'it'a s 12 i.d3 lLlg4!
Benjamin reminds his opponent
that 5 h4 cuts both ways - the g4-
square is lovely thi s tim e of year. If 13
f3 ? f6 14 fxg4 �xg4 15 .i.e2 i.xe2 16
'fixe2 fxg 5 17 hxg 5 lLlf4 18 'ii'f3 cxd5 19
exd5 o-o with an obvious advantage
for Black.
13 �e2 f6 14 .ie3 lLlf4
Houdini doesn't like this, but Ben­
jamin acquires the gorgeous e 5-square
1 62
for his other knight.
15 i.xf4 exf4 16 o-o cs 17 lLlf3 ttJes 18
�fb1 iVc3 19 lLlxes �xes 20 a4 gS
White is in trouble. Even the coming
endgame gives little relief.
21 hxgs fxgs 22 l:ta3 g4 2 3 'ifd2 .l:lh7 24
�C3 �XC3 2 5 l:lxc3 Jle7 26 g3 f3 2 7 aS
rJitf7 28 'it>f1 .l::t b 8 29 rJite1?!
Crossing to the queen side with the
king loses the e-pawn, though staying
put is no bargain either. Black will
bring his king to e5 and prepare to
break with ... h 5 -h4.
29 ... i.fs 30 rJitd2 i.xe4 3 1 i.xe4 �xe4
32 l:te3 .l:txe3 33 'it>xe3 'it>f6 34 'it>f4 'it>g6
35 a6 b6 36 c3 .l::tf8+ 0-1
Game 34
W.Weisser-L.Trumpp
Germa n League 2004
1 d4 lLlc6 2 d s ttJes 3 e4 ltJg6
I don't believe in this move order
because of 4 h4!.
4 ltJf3 es s c4 .ics
5 ...lLlf6 is more accurate here, and if
 Illustrative Games

6 lt:Jc3 then 6 ...i.b4. In addition to the

structural damage we can inflict, it is
not convenient for White to defend the
e-pawn. On the other hand, White could
force the game variation using the
move order 4 c4 es s lt:Jc3 ..tcs 6 lt:Jf3.
6 ..te2 lt:Jf6 7 lt:Jc3 0-0 8 0-0 a s ! 9 lt:Je1
d6 10 lt:Jd3
This is an odd way to pursue the
bishop - and quite unsuccessful, as
White has released the d4-square.
10 .i.d4 11 .i.f3 ltJh4?!
•.. 18 lt:Jf5 'iWf6 19 lt:Je3 'iWg6 20 b3 f5 21
i.a3 lt:Je4 22 lt:Jd1 i.c2 2 3 ..tc1 f4 24
�e6+ ii'xe6 2 5 dxe6 lt:Jc5?! [2s ... .l:!.fe8]
26 e7 .l:i.fe8 27 lt:Jc3 lt:Jd3?
Black has let slip most of his advan­
tage.
28 .l:!.e2 ltJxc1 29 l:i.xc1 .iH5

Setting a trap, which White fall s

right into, but i t is best prepared by
11 ...h 6 to prevent 12 ..tg s .
12 lt:Jb5? lt:Jxf3+! 13 'i*'xf3?! .i.g4 1 4 'ii'g 3
.i.e2!
Suddenly Black's light-squared
bishop is all in White's business. 30 c5?! [30 lt:Jds] 30 J:be7 31 lt:Jd 5 .l:!f7
••

15 l:!.e1?! 32 cxd6 cxd6 3 3 .l:!.d2 .:!.dB 34 f3 i.e6 3 5

15 lt:Jxd4 i.xf1 16 lt:Jfs lt:Jh s 17 'iWf3
..txd3 18 ii'xd3 i s a better chance.
15 lt:Jxe4 16 'iWh3 .i.xd3 17 lt:Jxd4 lt:Jc5
..•
The smoke h as cleared and Black
has an extra centre pawn for no com­
pensation. 17 ...ii'f6 ! would h ave been
lt:Jc7? :c8 3 6 .l:!.dc2 i.f5 37 .l:!.c4 llffB 3 8
'iii>f2 'iii>f7 39 lt:Jb5 'iii> e 7 40 I!xc8?
Equivalent to resignation, even if
the gam e somehow lasted 27 more
moves.
40 l:Ixc8 41 I:!.xc8 i.xc8 42 lt:Jc3 'iii>e 6 43
•••

even strong er. lt:Je4 h6 44 'iii>e 2 d5 45 lt:Jc3 i.d7 46 'iii>d 2

1 63
Th e Dark Kn ig h t System
'it>d6 47 <;itd3 wcs 48 'it>d2 i.c6 49 'iii>d 3
i.bS+ 50 'it>d2 i.f1 51 g3 fxg3 52 hxg3
hs 53 tbd1 gs 54 tbe3 i.bs ss tbfs i.d7
56 lbe3 bs 57 Wd3 h4 58 gxh4 gxh4 59
a3 h3 60 tbf1 i.fS+ 61 'it>c3 b4+ 62 axb4+
axb4+ 63 Wd2 Wd4 64 tbg3 i.d3 65 tbh1
h2 66 tbg3 i.f1 67 tbfS+ 'it>cs 0-1
Gam e 35
B.Perrusset-I.Moul l ier
Pa ris 2005
1 d4 tiJf6 2 C4 tiJc6 3 tiJc3 eS 4 d S ltJe7 5
e4 tbg6 6 i.e2 i.cs [6 ... i.b4! ] 7 tbf3 o-o
8 o-o d6 9 a3 as 10 i.d2 ife7
Thi s generally useful m ove often
makes ... lbh s possible; e.g. 11 'ii'c 2
lbh s ! ? 12 tbxe s ? ? "iVxes . In thi s gam e
Black had other ideas.
11 �b1 i.d7 12 b4?! axb4 13 axb4 i.d4!
Once again the slippery dark bishop
finds a home on the lovely d4-square.
14 i.d3 l:!.a3 15 .l:Ib3 .l:!.xb3 16 Vi'xb3 l:.a8
17 'ii'c 2 i.. b 6 18 'iib 2 h6 19 I:ta1 .l:!.xa1+
20 'iix a1
White has finally dealt with the can
1 64
of worms he opened up on the a-file,
but his remaining pieces are not very
well co-ordinated. Suddenly Black
starts attacking. All the attacking
m oves are typical of the variation, and
the rooks are not needed.
tbg4 21 i.e1 tbf4 22 i.c2?
The only satisfactory defence is 2 2
i.f1, though Black will still have some
initiative.
22 .. .'�f6!
Black i s threatening the crushing
2 3 ... tbh 3+! 24 'it>f1 i.xf2 ! 25 i.xf2
lbxh 2+ 26 �e2 tbf4+ 27 �d1 tbxg 2.
2 3 h3? 'ii'g 6
23 ... lbxh 3 + ! and 23 ... tbxf2 ! also win.
24 hxg4 'i¥xg4 25 lbh4 �xh4 26 i.d1
tbd3 27 i.f3 i.xf2+ 28 i.xf2 �xf2+ 29
�h1 tbf4 30 tbd1 'ifh4+ 31 �g1 i.g4 32
"iVa 3 i.xf3 3 3 �xf3 'ilke1+ 0-1
Game 3 6
D.Bara midze-E.Griezne
Ba u n ata I 1999
1 e4 tbc6 2 d4 es 3 dS tbce7 4 c4 tbg6 5

.::Jc 3 i.c5 6 lt::lf3 lt:Jf6 7 i.e2 o-o 8 o-o d6
g 11t'c2 a6
In the theoretical section I recom­
m end 9 ... i.d7 and usually 10 ... as, but
the game move is quite reasonable. In
either case, Black uses the same attack­
i n g ideas on the king side.
10 a3 'iiie 7
In addition to preparing ... lt::lh s, this
holds up White's c4-c5 advance.
1 1 b4 i.a7 12 lt::l d 1 lt:Jh5 [12 ... lt:Jf4! ?]
13 lLle3?!
Here 13 g3 i s best, which certainly
makes it clear why White needs to omit
h2-h 3 .
13 ...lt::lgf4 14 tt:Jf5?! Vi'f6?!
At some cost, White has arranged
not to be checkmated on the kingside.
Black should slow down and play
14 ... i.xfs 15 exfs e4 or 1 5 . J:tfe8 with
. 25 .. .'ii' b 2 26 i.e2?! Vi'xa3 27 f6?
positional advantages based on space,
activity, and White's weak fs-pawn.
15 g3! i.xfS 16 exf5 e4 17 i.xf4 lt:Jxf4
18 gxf4 exf3 19 i.xf3
You don't see this every day. Black i s
still slightly better, and his position is
simpler to play, as we will see.
Illustrative Games
19 .. .'fih6
20 I;iae1?
White decides that with three f­
pawns, he won't miss one. However,
even tripled pawns control squares,
and the f4-pawn is particularly impor­
tant since it covers weak dark squares.
Far better to give up the fs-pawn which
is one m ore obstruction to White's sad
bishop (though in fact White should
not be eager to give up any of his
pawns). After this his gam e slowly de­
teriorates.
20...'ii'xf4 21 'il¥e4 'il¥g5+ 22 'il¥g4 'il¥f6 23
l:[e4 l:f.ae8 24 .l:i.fe1 l:.xe4 25 .l:i.xe4
Since the f4-pawn's disappearance
White h as had to be careful not to al­
low ... i.d4 and ... i.es, but he can't de­
fend everything.
Tim e trouble?
27 ...'ii'a 1+ 28 �g2 Vi'xf6 29 f3 'i!Vd8 30
'i!Vf4 g6 31 i.d3 �g7 32 h4 h 5 3 3 'ili'g5
.U.e8 34 'ii'x d8 .l:r.xd8 3 5 .l:!.e7 i.b8 36 b5
axb5 3 7 cxb5 �f6 38 l:te4 i.a7 39 �f4+
�g7 40 .Ue4 i.b6 41 �g3 .l:ra8 42 �4
l:.a3 43 i.e2 i.f2 44 .l:r.e7 lla4+ 45 .l:!.e4
1 65
Th e Dark Kn ig h t System
l'.txe4+ 46 'ii;lxe4 .i.xh4 4 7 f4 'ith6 48
.i.d1 .i.e1 49 .i.f3 .i.c3 50 .i.g2 .i.b2 51
.i.f1 fS+ 52 'itf3 gS 0-1
Gam e 3 7
G.Kaidanov-A.Miles
Pa l m a d e M a l l o rca 1989
1 d 4 ltJc6 2 e 4 eS 3 dS ltJce7 4 C4 ltJg6 5
.i.e3 .i.b4+ [ s ... lt'lf6] 6 ltJd2 lt'lf6 1 f3
"fie7 8 g3 o-o 9 .i.h3 c6!
The strongest players are much
more likely to do thi s than look to
blockade the queen side.
10 a3 ..tcs 11 lt'lf1?! bs!
White's slow m anoeuvring does not
take into account Black's option to
open the position up. Black was better
already, but now White is in serious
trouble.
12 b4 .i.d4!
That square again ! When will White
learn not to play c2-c4 ? -
13 .i.xd4 exd4 14 'ifxd4 bxc4 15 d6 'ii'e s
Understandably, White tries to close
things again. Miles has correctly seen
1 66
that h e will be winning the endgame if
it comes to that .
16 lt'le2 as 17 lt'le3?! [17 bxas] 17 ... axb4
18 lt'lxc4 'ti'h s 19 .i.g2 cs 20 "fie3 .i.a6
21 lt'lb6 :ab8 22 axb4 l:!.xb6 23 bxcs
.l:lc6 24 :a s ttJes 25 lt'lf4 1/i'gs 26 �d4
.l:tb8 27 .l:ta1 h6 28 .l:!.xa6 l:txa6 29 o-o
lt'lc6 30 "fic4 .l:!.a s o-1
Brutality in the i.. e 3 variation
Game 38
S.Brudno-J.Benjamin
Boston 2001
1 d4 lt'lf6 2 c4 lt'lc6 3 lt'lc3 es 4 dS lt'le7 5
e4 lt'lg6 6 .i.e3 .i.b4 7 f3 .i.xc3+ 8 bxc3
d6 9 cs o-o 10 .i.d3 lt'ld7! 11 cxd6 cxd6
12 lt'le2 'if as 13 o-o lt'lcs 14 ..tc4 .id7 15
.i.b3 l:[ac8 16 g3 fs
After con structing an ideal position
on the queen side, Benjamin opens up
the second front. The m ore I look at the
Dark Knight System, the more I'm
struck by how often it is Black who h as
the convenient pawn breaks, and how
useful that is.
 Illustrative Games

As I m entioned in the theoretical

section, this dubious move happens
like clockwork. Incidentally, I worked
out Black's plan independently during
this game.
9...0-0 10 .i.d3 lLld7 11 cxd6 cxd6 12
lZJe2 �a s 13 o-o lt::lc s 14 'ii'd 2 lLlxd3?!
Thi s is a big commitment. 14 ... b6 is
more accurate - White's bishop isn't
going anywhere.
1S 'ili'xd3 b6 16 c4 ii.a6?!
17 .itc2 fxe4 18 fxe4 �xf1+ 19 �xf1 Having taken the bishop, 16 ...fs (or
�h3+ 20 '>t>g1 .l:!.f8 21 'ii'd 2 �xa2! 1S .. .fs) is the logical follow-up. Black is
Ow! better on the queenside, but it is hard
22 .l:!.e1 to play against a single weakness.
If 22 .l:.xa2 l':!.f1 mate. 17 'i¥b3 .l:!.ac8 18 l:f.fc1 l:!.c7 19 l:!c2?!
22 ..J!i'c4 23 ..tf2 as 24 lLlc1 l:[f3 2S lt::le 2 White intends a perpetual on my
a4 26 li'gs lZJxe4 0-1 queen, but there i s a tactical snag. The
right way to implem ent this plan is 19
ii.f2 l':!.fc8 20 .te1 'fics+ 2 1 ii.f2 etc.
Came 39 19 ....l::tfc8 20 .td2 'it'cs+ 21 .te3 .txc4 22
R.Mitcheii-J.Schuyler ..txcs ..txb3 23 axb3
B l oom i n gton 1991

1 d 4 lZJc6 2 C 4 eS 3 dS lZJce7 [3 . . .ii.b4+!]

4 e4 lt::lf6?! [4...lZJg6] S lt::lc 3 lZJg6 6 ii.e3
.i.b4 7 f3 ii.xc3+ 8 bxc3 d6 9 cs?!

23 ... bxcs?
This g ame was a long time ago, but
I remember that I never even con sid­
ered the right move, which is 2 3 ... as !,
even if Black i s actually not much bet-

167
Th e Dark Kn ig h t System
ter after 24 lLlg 3 ! lLJh4 2 5 lLlf1 h6 26
lLle3. Still, White must find these
moves. After the text, Black gets all
tangled up and is lucky to extricate
himself. The rest of the gam e is full of
inaccuracies, most of which I will not
bother to mention.
24 .:!.a6 .l:td8 2 5 .l:tca2 [25 b4! c4 26 l:l.c6]
2S ....l:Ib7 26 .U2a3 .l::.d d7 27 lLJc3 [27
.l:.c6 !] 2 7 ...'it>f8 28 lLla4 lLJf4 29 lLlb2 We7
30 �f2 fs 31 g3 lLJhs 32 lLJc4 fxe4 3 3
fxe4 lLlf6 34 lLJa s lib6 3 5 lLJc6+ 'iii>f7 36
We3 lLJg4+ 3 7 �d3 lLJxh2 38 .:.xa7 .l:!.xa7
39 l:!.xa7+ Wf6 40 'ifilc4 lLlf1 41 l:td7
lLJxg3 42 ltxd6+ �gs 43 l:.d7 g6 44
.l:.xh7 lLJxe4 45 l:.e7 �f4 46 b4 cxb4 47
lLJxb4 lLlf2 48 l':!.f7+ �g3 49 WcS l:.b8 so
lLJc6??
Things have been approximately
even the whole time despite my fre­
quent pawn advantage. Here White
falls into a trap. The coming knight and
pawn endgame is winning for Black in
all variations because White's king and
knight are both so far from the passed
pawns.
so...lLJe4+ 51 Wc4 lLJd6+ 52 Wd3 .l:.b3+
1 68
53 Wc2 lLJxf7 54 'ifilxb3 e4 5 5 Wc3 Wf2
s6 lLle7 gs 57 lLlfs e3 0-1
Game 40
I.Jelen-E.Dizdarevic
Lj u b lj a n a 1992
1 d4 lLJf6 2 c4 lLJc6 3 lLJc3 es 4 dS lLJe7 5
e4 lLJg6 6 i.. e 3 i.. b4 7 f3 i..x c3+ 8 bxc3
d6 9 cs 0-0 10 h4?!
White must attack or Black will
tighten the screws one by one. How­
ever, thi s is too soon and too crude.
1o ... lLJhs 11 ..tf2 c6!
Over and over we see thi s m ove,
typically at the most surprising and
inconvenient tim e for White.
12 g4 lLJhf4 13 h s lLle7 14 cxd6?!
The complication s are dizzying,
even for my computer. it's hard for
things to work out well for White with
his king in the middle, but he needs to
try 14 h6 or 14 lLJe2.
14.. .'�xd6 15 c4 cxds! 16 cxd s i..x g4! 17
h6
17 fxg4 runs into 17 ...�4+ and

18 .. ."i!Vxe4+, winning the rook.
11 gs 18 'ii b 3 i.d7 19 li:Je2
•••
1 9 lt:Jeg6?
..•
19 ... i.a4! 20 �e3 .l:.fc8! and White
cannot last long.
20 1Ig1 f6 21 lt:Jc3 .l:i.fc8 22 a4 <it>h8 23
�d2 b6 24 .i.a6 lt:Jh3 2 S l:f.gf1 lt:Jxf2 26
i.xc8 l:i.xc8 2 7 .l:!.xf2 �cs 28 l:tff1 lt:Jf4 29
lt:Je2 'ifa s+ 30 lt:Jc3 lt:Je2 3 1 .l:i.fc1 lt:Jxc1
32 J::tx c1 Wg8 33 f4?
White isn't having any fun at all,
but thi s only m akes things worse.
33 exf4 34 es? fxes 3S d6+ Wf8 36
.••
"Yi'b1 'it>g8 3 7 1lt'b3+ 'ifi>f8 38 �b1 'ilkcs o-1
Black loses because
he's playing Karpov
Gam e 4 1
A.Karpov-D.Chevallier
Fra nce 1993
1 d4 li:Jf6 2 c4 lt:Jc6 3 lt:Jc3 es 4 ds lt:Je7 s
e4 lt:Jg6 6 i.e3 i.b4 7 f3 i.xc3+ 8 bxc3
d6 9 �d2 lt:Jd7 10 h4 h6 11 g3 b6 12
tt:lh3 tt:Jcs 13 lt:Jf2 'i!Vd7
There i s nothing wrong with this
Illustrative Games
move, but 13 .. .fs can be played, and so I
believe it should be played.
14 .l:.b1 as 1s f4?!
In his annotations, Palliser gives thi s
an ' ! ' and indicates that i t i s the begin­
ning of the end for Chevallier.
1S exf4 16 gxf4
•••
16 'i!Ve7?!
•••
After this Black i s a little worse.
Meanwhile, should Black find 16 ...h s,
he is a little better! In one stroke, Black
stabilizes his knight, clamps down on
White's weak g4-square, and immobi­
lizes White's i solated h-pawn. Of
course, with a small edge (or even a
small disadvantage) Karpov will proba­
bly win - he does outrate Chevallier by
430 points.
17 .i.xcs ! bxcs 18 hs lt:Jf8 19 .l::!.g 1 f6 20
.i.d3 li:Jd7 21 'ii'e 2 i.a6?
Black's situation will not improve
with his king in the centre. For better or
for worse, he must castle and try for
.. .f6-fs.
22 tt:lg4 lt:Jb6 2 3 lt:Je3 i.c8 24 lli'g2 l:t.g8
2S 'i!Vg6+ 'iii>d 8?! [2S ... Wf8] 26 'it>d2 .i.d7
27 �h7 'i!Vf8 28 .l:f.g2 Wc8 29 l:i.bg1 'it>b7
1 69
Th e Dark Kn ight System
30 �xg7 l:Ih8 31 'iikg 6 .i.e8 3 2 �fs l:td8
3 3 'iike 6 i..x hs 34 lt::lfs lt:Jc8 3S e s fxes 36
fxes �e8 37 l:Ib1+ 'it>a8 38 .l:rxc7 'it'xe6
39 dxe6 dxes 40 .l:!.xcs e4 41 lt:Jd4 l:.xd4
42 cxd4 exd3 43 l:r.xh s 1-0
Overall, an impressive game by Kar­
pov, but not one that casts any doubt
on Black's opening.
Gam e 4 2
W.M.Buehi-J.Benjam i n
Re no 1999
1 d4 lt::lf6 2 c4 lbc6 3 lt::l c 3 es 4 d S lt::le 7 S
e4 lt::lg 6 6 lt::lf3 .i.b4 7 i.. d 3 i..xc3+ 8 bxc3
d6 9 o-o o-o 10 h3? tt:Jhs 11 'it>h2!?
11 ...lt::l hf4
Because of 11 Wh 2, Black's n ormal
attacking plan needs to be m odified -
11 ... tt:Jgf4?! 12 .i.c2 1Wf6?? 13 g4 traps a
knight.
12 lbg1?! lbxd3!
They say bad bishops defend good
pawns - but they don't if they're dead.
13 'iikxd3 fs 14 exfs?! [14 f3] 14 ... .i.xfs
1S 'ii'g 3?! 'i!Vd7! 16 f4?! 'ii'a 4
1 70
This i s why I like doubling White's
pawn s so much - it provides a ready­
m ade target on the other front if the
king side play is running out of steam .
17 fxes 'it'xc4 18 .l:!.d1 tt:Jxes 19 l:d4 Vies
20 lbe2 .l:!.ae8 21 lbf4 lt:Jg6 22 tt:Jxg6
i..xg6 23 .i.gs .l:i.e2 24 a4?! .:f.ff2 2S l:.g1
.i.e4 0-1
Game 43
A.Hahn-J.Bonin
New Yo rk (ra pid) 2003
1 d4 lt:Jc6 2 C4 eS 3 dS lt:Jce7 4 e4 lt:Jf6 S
lt::lc 3 lbg6 6 lt::lf3 .i.cs
I prefer 6 ....i.b4 here, but we must
know thi s position anyway via 4 ... lt::l g 6
s lbc3 .i.cs 6 lDf3 lt::lf6.
7 .i.d3 o-o 8 o-o a s 9 a 3 d6 10 .:!.b1 lbhS
11 b4 axb4 12 axb4 i.. b 6 13 lt:Ja4 .i.a7
14 .i.c2 lt::l hf4
Here again, the h-knight goes first.
Thi s is to allow ... .i.g4 and ... lbh4 - an
effective plan with White's bishop al­
ready on c2. Black h as a significant ad­
vantage and a powerful initiative.

15 cs .i.g4 16 g3 'fif6?
Black is alm ost winning at once
with the alternative 16 .. .<!iJh 3+ 17 Wg 2
Yi'd7 18 .i.e3 fs !, bringing a rook into
the attack.
17 .l:1b3 .i.xcs 18 ltJxcs dxcs 19 bxcs
Ita1 20 gxf4 lLlxf4 21 'it>h1 lLlh3 22 �e2
.l:!.fa8? [22 . .J::t a2!]
23 l:tc3??
Instead, with 2 3 �e3 ! White has ex­
cellent chances to realize her extra m a­
terial . After the text, she is completely
busted.
23 ... It8a2 24 'iitg 2 .l:.xc1 25 .l:.xc1 lLlf4+
26 'it>f1 ltJxe2 2 7 '.t>xe2 'iff4 28 l::te 3
.ixf3+ 0-1
Illustrative Games
The catastrophe that is
White's plan of g2-g3 and f2-f4
Cam e 44
E.Schiendorfer­
D.Recuero Guerra
E u ro pea n J u n ior C h ' s h i ps,
H e rceg N ovi 2006
1 d4 liJf6 2 c4 liJc6 3 tt::l c 3 es 4 d S lLle7 5
e4 tt::lg 6 6 g3 .ib4 7 .ig2 .ixc3+ 8 bxc3
d6 9 liJe2 o-o 10 o-o 'ii'e 7 11 �d3 .i.d7
12 .i.a3 .l:.fe8 13 l:tae1 b6 14 liJc1 'ii'd 8
Not the most inspiring play so far.
Now comes typical hara-kiri by White.
15 f4? exf4 16 gxf4 liJhs 17 liJe2 'ii'h4 18
.i.c1 fs 19 'it>h1 '/J.e7 20 exfs?? Us 21 1i'f3
21 'ifxfs l:l.xe2 2 2 .l:!.xe2 lLlg 3+ 2 3
Wg 1 lLlxe2+ 2 4 Wh 1 lLlg3+ 2 5 'it>g 1 lLlxfs
is hilariously brutal.
21 ... i..g4 22 �f2 J!txe2 0-1
Cam e 45
C.Baluta-A.Cioara
Buch a rest 1996
1 d4 liJc6 2 c4 es 3 d s ltJce7 4 e4 liJg6 s
1 71
Th e Dark Kn ight System
a3 lLif6 6 t2Jc3 �cs 7 h 3 d6 8 lLif3 as 9
g3
Thi s is amazing . I'm wondering
when White will get around to moving
his b- and f-pawns.
9 o-o 10 �g2 c6 11 o-o cxds 12 cxds
•••
�d7 13 1ie1 'iVc8 14 h4? tL:Ig4 1S l:te2 bS
16 �d2 b4 17 tL:Ia4
17 �bs?!
..•
Black has built his position admira­
bly so far, but now he starts to lose the
thread. 17 ... �a7 18 axb4 �bS was bet­
ter.
18 tL:Ixcs 'ii'x cs 19 hS tL:Ie7?! [19 ...�xe2]
20 i.h3?1 [20 �e3 ! ) 20 �d7?
•••
After thi s lemon, Black is actually
worse for a mom ent.
21 �g2?
But only for a m oment. 21 h 6 ! g6 2 2
'iiVb 3 was correct.
21 bxa3 22 .l:.xa 3 fs?! [22 ... �bs ! ) 2 3
•••
�e3?! t2Jxe3 2 4 l:.exe3 fxe4 2 S l:txe4
l:tab8 26 'ii'd 2 a4 2 7 tL:Ih4 l:tb7 28 l:tee3
.l:.b4 29 l:.ac3?! 'ii'b 6 30 .l:.e2 h6 31 �h2?
.:d4? [3 1...�b5) 3 2 l'ld3?? �xd3 3 3
'iixd3 �bs!
Finally!
1 72
34 'ii'e 3 'ii'xe3 3S l:txe3 l:.xf2 36 �g1
lbb2 37 �h3 l:tb1+ 38 �f2 l':tb3 39
�e6+ �8 40 lle1 a3 41 l:tc1 a2 42
tL:Ig6+ t2Jxg6 0-1
A loss for Black -
can this be right?
Game 4 6
Bu Xia ngzhi-L.Christia nsen
Deizisa u 2000
1 d4 tL:If6 2 c4 tL:Ic6 3 lLic3 es 4 d s tL:Ie7 s
e4 tL:Ig6 6 a 3 �cs 7 �d3 as 8 l::t b 1 d6 9
h 3 o-o 10 b4 axb4 11 axb4 �a7 12 g3?1
12 l2Je8?!
•••
This awkward m ove prepares the
.. .f7-f5 break, but Black never gets to
execute this plan. Besides, 12 ... c6! is
just sitting there waiting to be played.
Yet again. One benefit to ... c7-c6 is th at
after ... c6xds c4xds White can no
longer shut out the monster a7-bishop .
Black would then be significantly bet­
ter.
13 lLif3 �d7
Black realizes that 13 .. .fS ? ! hurts his

position, trading off White's l ame light­
squared bishop and allowing White to
castle without dropping the h 3 -pawn .
After the text m ove, White has equal­
ized - though not for long.
14 .tg5?!
14 'itf1! 'ii'c 8 1 5 Wg 2 was correct.
14 ...iVc8 15 g4 c6 16 tt:Je2 h6 17 .td2
cxd5 18 cxd 5 'ilr'd8 19 tt:Jg3 .tb6 20 o-o
tZ:If4! 21 .txf4 exf4 22 tZ:Ih5 tZ:If6 23 tZ:Ixf4
.l:!.e8
Black's pawn sacrifice has removed
White's best minor piece from play and
opened the gates for Black's pieces.
Thi s is a thematic idea in the Dark
Knight System, and indeed in any posi-
Illustrative Games
tion where the opponent has put most
of his pawns on one colour (especially
in the vicinity of his king).
24 'itc2 .l:i.c8
Houdini points out 24 ... g 5 ! 25 tt:Jg 2
(the knight's idiot square; but if 2 5 tt:Je2
h 5 !) 25 .. Jlc8 and 26 ...tt:Jxe4. In the
game White's knight becomes a real
pain in the keister.
25 'itb2 tt:Jxe4 26 tZ:Ih5 tt:Jc3 27 l:tbc1 g6
28 "ir'd2
H ad White played 27 :fell he could
have gone for 2 8 .l:!.xc3 �xc3 29 'itxc3
gxh 5 30 gxh 5 with no fear of 30 ....txh 3.
28...tt:Je2+ 29 Wg2 .l:!.xc1 30 'i¥xh6 tZ:If4+
31 'ii'xf4 .l:!.xf1 3 2 'itxf1 gxh 5 3 3 'ti'h6 f5
34 'ti'g6+ 'ith8?!
Christiansen, who has been defend­
ing perfectly until now, slips. lt is not
good to allow 'i!Vxh 5 with check, and by
winning the pawn White defends his
own peon on h 3 .
35 gxf5 .l:.f8 36 'itxh 5+ 'itg8 3 7 1i'g6+
'ith8 38 'ili'h6+ 'itg8 39 f6
39 ...�f7??
39 ... .txh 3+ 40 "iii'xh 3 .l:.f7 or 40 ....l:.xf6
offers legitimate chances of survival.
1 73
Th e Dark Kn ight System

40 tt::l g 5! i.xh3+ 41 tt::l x h3?? White is much worse here.

Suddenly Black is back in it. 41 11 i.e2 lle8 12 h3 cxd6 13 hxg4 dxc5
Wel l ! is completely winning; e.g. 14 'ii'x d8 .l:.xd8?! [14 ... i.xd8] 15 tt::lfd2
41 .. .'ii' d 7 42 'fkg 6+ wfs 43 i.bs ! . tt::l e 7 16 tt::l a 3 tt::l d 5 11 g3 i.c7 18 i.f3
4 1 .'itxf6 4 2 i.h7+ Wh8 4 3 i.g6+ 'it>g8
• • tt::l e 3 19 g5
44 i.xf7+ <j;;xf7 45 'ifh7+ 'it>e8 46 'i:Vxb7
'ti'd4? [46 .. JWfs !] 47 'i!Vc8+ i.dS 48 'i!Ve6+
i.e7 49 b5 'ii'd 3+ 50 'it>g2 'i!Vxb5
Black has recovered one pawn, but
his queen is out of play and White
takes the opportunity to enter a win­
ning pawn endgam e.
51 tt::lf4 'ii'b 2 52 tt::l g 6 'ii'f6 5 3 'ii'x e7+
'iWxe7 54 tt::l x e7 'it>xe7 55 '>t>f3 'it>f6 56
Wf4 <j;; g 6 57 <j;;g4 <j;;f6 58 f4 1-o

Hort plays the Kevitz, 15 i.t5?!

.•.

and lets Hi.ibner off the hook 19 ... tt::lfs ! is better, forcing 20 tt::lf1
(or 20 tt::l e4, which is n ot very good ei­
Gam e 4 7 ther), because 20 .l:.g 1? loses to tt::l x g 3 !
R.H i.ibner-V.Hort 2 1 l:txg 3 l:txd2 ! 22 <iii>x d2 i.xf4+ and
Germa n League 1984 2 3 ... i.xg 3 .
20 tt::ld c4 tt::lxc4 21 tt::lxc4 i.d3
1 e4 tt::lc 6 2 d4 e5 3 dxe5 tt::lxe5 4 f4 tt::lc6 There is still some play, but Black
5 i.e3 i.b4+ [s ... ds!] 6 c3 i.a5 1 tt::lf3 has no real advantage.
tt::lf6 s e5 [8 'i¥a4 ! ?] s tt::l g4 9 i.c5? d6!
••• 22 b3 .l:.ab8 2 3 Wf2 b5 24 tt::l e 3 c4 25
10 exd6 o-o tt::l d 5 i.b6+ 26 'it>g2 wfs 2 7 1Ihe1 .l:te8 28
b4 i.f5 29 .l:lxe8+ l1xe8 30 a4 bxa4 31
.l:.xa4 h6 3 2 gxh6 gxh6 3 3 tt::lx b6 axb6
34 'it>f2 lidS
By now it is White who is trying for
a win, though it is unlikely.
35 .l:la8 .:txa8 36 i.xa8 We7 37 'it>e3 'it>d6
38 'it>d4 b5 39 i.d 5 f6 40 i.t1 i.d7 41
'it>e4 cJ;;e 7 42 i.g6 i.c6+ 43 'it>d4 'iii>d 6 44
i.h5 f5 45 i.g6 i.d7 46 i.f7 i.cs 47
i.eS i.a6 48 'it>e3 'it>e7 49 i.g6 i.c8 50
'it>d4 'iii>d 6 51 i.e8 i.a6 Yz-Yz

1 74

Game 4 8
M.Orso-G.Bordas
B u d a pest 2000
1 e4 lt:lc6 2 d4 e5 3 dxe5 lt:lxe5 4 f4 lt:lc6
5 i.c4 l'Llf6 6 l'Llc3 i.b4 1 e5 d5 8 exf6
dxc4 9 �e2+ 'iii>f8!? [9 ... i.e6 ! ] 10 .ie3
Had White been tempted by 10
fxg7+? <it>xg7, Black's l ast move would
have been amply rewarded. The rook's
rapid arrival on the e-file is a serious
problem for White.
10 .. .'i¥xf6 11 o-o-o i.e6
The inconvenience suffered by
Black's king is not enough to make up
for White's missing pawn and shortage
of light-squared bishops.
12 lt:le4 'iWe7?! [12 .. .'i¥fs] 13 l'Llf3 c3 14
b3?! [14 bxc3] 14 ... a5 15 'it>b1 a4 16
'iWb5??
White is now lost. 16 .i.f2 is essen­
tial, though Black has a clear edge.
16 ...axb3 17 axb3
17 ...i.xb3 18 cxb3 'i¥xe4+ 19 l:td3 .l:!.a 5
20 'ifc4 'iixc4 21 bxc4 rJite1 22 lt:ld4
lt:lxd4 23 i.xd4 l:r.ha8 24 :e3+ Wd7? 2 5
Illustrative Games
.l:!.he1?
25 ..ixc 3 ! i.xc3 26 lld1+ and Black
still has to show good technique.
25 .. Jla1+ 26 Wc2 .l:l1a2+ 27 rJitd3 l:!.d2+
28 'iii>e 4 f5+ 29 'iii>xf5 .l::!.x d4 30 c5 .i.xc5
31 l::tx c3 l:tf8+ 32 rJitg5 i.e7+ o-1
Gam e 49
S.Fedorchuk-A.Miles
E u ro pea n C h a m pio n s h i p,
Oh rid 2001
1 e4 l'Llc6 2 d4 e5 3 dxe5 lt:lxe5 4 l'Llc3
i.c5 5 f4 lt:lc6 6 lLlf3 d6 1 lt:la4 i.b6 8
.id3 d5?!
An interesting attempt to take ad­
vantage of the decentralized knight,
but Black i s spending a tempo to open
the position when he is behind in de­
velopment and White will soon have
the bishop pair.
9 lLlxb6 axb6 10 e5?!
Not terrible, but it makes no sense
to close the position . The fS-square will
be useful for all of Black's pieces.
10 ...lt:lge7 11 o-o i.f5 12 lt:lh4 i.xd3 13
1 75
Th e Dark Kn ight System

'ii'xd3 'ii'd 7 ..tcs 5 tLlb3 .i.b6 6 tLlc3 d6

Black h as equalized.
14 i.d2 'ii'g4 15 i.e1 o-o 16 h3 Vi'd7 17
i.d2 .l:Ia4 18 i.c3 tLlg6 19 tLlxg6 fxg6 20
g3 tLle7 21 �h2 h S 22 b3 .l:!.e4 23 l:tae1
Vi'c6 24 .i.d2 tLlfs
For our purposes, it i s important to
use the precise move order 6 ...tLlge7 7
'ii'e 2 d6.
1 'i!Ve2 tLlge7 8 .i.e3 o-o 9 o-o-o fsl

There's no tim e like the present.

2 5 'i*'c3?1
The black knight becomes strong af­
ter this. 2 5 g4 was best.
2S ... h4 26 gxh4 tLlxh4?1
26 .. .'ii'xc3 27 ..txc3 �f7 ! was better.
27 "itxc6 bxc6 28 �g3 tLlfS+ 29 �g4 �f7
30 l:.xe4 dxe4 31 .U.e1 Itd8 3 2 ..te3?1
.l:.d11 3 3 l:Ie2 �e6 34 i.f2 Itf1 3 5 a4 cs
36 as?l bxa s 3 7 �gs? .l:!.h1 38 ..txcs
�xh3 39 �xg6 .l::tg 3+ 40 �h7 e3 41 �g8
gs 42 fxgs .l:.xgs+ 43 �h7 llg7+ 44 �h8
.l:!.g3 45 .l::t h 2 �xes 46 .i.a3 l:tg4 47 i.b2+
�e4 48 ..tc3 .l:.h4+ 49 .l:!.xh4+ tLlxh4 50
..txa s e2 51 �g7 tLlf3 52 �f6 �e3 0-1
10 g3 fxe4 11 tLlxe4 tLlfs 12 .i.f4 'ii'e 8
Mr. H says 12 ... �h 8 or 12 ... h 6 with
advantage to Black.
13 ..tg2 as 14 .l::[ h e1 'i¥g6 15 c3?1 h6 16
a4 ..te6?1 17 'i*'c2? [17 tLlec S ! ] 17 ...tLlh4
18 ..th1 ..tg4 19 !Id3 l:.ae8 20 f3? tLlxf3?
[2o ... .il.e6] 2 1 ..txf3 i.xf3 22 l:.xf3 d s 2 3
tLlbd2 dxe4 24 l:tff1 tLl e s 2 5 :xe4 tLlg4
26 l:!.fe1 l:td8 27 .l:te7?1 'i!Vxc2+ 28 �xc2
gS 29 �1e6 .l:tf7

Gam e 50
Goh Wei Ming-F.Bellini
T u r i n O lym p i a d 2006

1 e4 es 2 tLlf3 tLlc6 3 d4 exd4 4 tLlxd4

1 76

30 h3??
Instead of thi s blunder the simple
30 .Ug6+! �h 7 31 ltee6 keeps White in
the game.
30....:1.xe7 31 .l:!.xe7 gxf4 3 2 hxg4 fxg3
0-1
Game 5 1
P Bontem pi O Jovanic
.
-
.
Nova Gorica 2008
1 e4 e5 2 tiJf3 tbc6 3 d4 exd4 4 ttJxd4
it.c5 5 ..te3 'iif6 6 tLlb5 ..txe3 7 fxe3
�h4+ s g3 'i!Vds 9 'i!Vg4 g5!
This Basmaniac-esque move i s a bit
funny-looking at first, but if we think
about it, it is the dark squares that
need coverage since we have a light­
squared bishop and we are about to
have a knight posted on the lovely es­
outpost. To be a little more concrete, it
is beneficial to take the f4-square from
White's queen.
10 ttJ1c3 lLle5 11 'ii'e 2 d6 12 h3 c6 13
lLld4 tLlf6 14 o-o-o 'ii'e 7 15 i¥f2 ..te6 16
..te2 o-o-o 17 tLlf5 ..txf5 18 exf5
Illustrative Games
White's pawn structure has im­
proved somewhat, but his bishop has
not, nor h as the g aping hole on es. Fur­
thermore, White no longer has the fS­
hole to use.
18 ...ttJed7 19 i.f3 tLlc5 [19 ... ds ! ?] 20
.l:[he1 h 5 21 �d4 'ii'e 5 22 h4 .:1.hg8 23
.l:!.ed1 gxh4 24 gxh4 <Jilc7 2 5 l:!f4?! .!:[deS
26 'iitd 2?
26 ...ttJcd7?!
Black is reluctant to put any pawns
on light squares for long-term posi­
tional reasons, but White won't reach
the long term after 26 ... ds ! 27 .:te1
tbce4+ 2 8 i.xe4 tLlxe4+ 29 lLlxe4 dxe4,
which exposes the ludicrous placement
of White's king ; e.g. 30 �c1 �as 3 1
�b1 :i.g 2 3 2 'iWf1 'i¥d2 3 3 l:tc1 �xe3 34
�xg 2 'ii'xf4 with a winning position.
Instead, soon comes an avalanche of
blunders that is presumably time­
induced.
You know what? You don't want to
see the rest of this. lt's ridiculous and
embarrassing.
Black eventually won the game on
move 81.
1 77
Th e Dark Kn ight System
Cam e 52
S.Vajda-S.Skembris
N a ujac 1999
1 e 4 es 2 lLlf3 ltJc6 3 d 4 exd4 4 ltJxd4
�f6!?
Apparently this is a playable move
order, though it gives White the option
of 5 tLlbs ! ?, which is good enough for
some edge.
5 .i.e3 .i.cs 6 c3 'ili'g6 7 f3 ltJge7
In the theoretical section, I recom­
mend the immediate 7 ... a6.
8 �d2 a6 9 ltJc2 .i.xe3 10 ltJxe3 d6 11
.i.e2 fS
12 exfs .i.xfs
12 ... ltJxfS is equally good.
13 0-0 0-0
13 ... 0-0-0 is at least as good too.
14 tLla3 .l:tae8 15 .l:tae1 'it>h8
it's time to take away White's op­
tion to snap off the bishop; i.e.
1S ... .i.e6.
16 ltJac2 'if'h6 17 .i.d3 .i.xd3 18 'if'xd3
ltJg6 19 ttJd s?! ltJces 20 �d4 c6 21
ltJde3?! [21 f4!] 21 ...ltJf4 22 �h1 'ii'g 6?!
1 78
[22 ... ltJed3 !] 2 3 g3? [23 'ifd2] 23 ... ltJfd3
24 l:.d1 ltJxb2 2 5 �b1 t2Jed3 26 'ifb6
ltJa4 27 'ii'x b7 ltJxc3 28 �b6?! [28 �b3]
28...tLles 29 �xa6 .l:lxf3 30 .l:!.xf3 'ili'e4 31
tLle1 lLlxf3 3 2 ltJ1g2 ltJbs! o-1
Came 53
A.Motylev-S.Giigoric
Yugos lav Tea m
C h a m pio n s h i p 2000
1 e4 es 2 lLlf3 ltJc6 3 d4 exd4 4 ltJxd4
.i.cs 5 .i.e3 'ilif6 6 c3 �g6 7 ltJd2 lLlf6 8
f4 ltJxd4 9 cxd4 .i.b4 10 fS 'ilig4 11
�xg4 ltJxg4 12 .i.f4 d S 13 h3 lLlf6 14 es
lbe4 15 g4 g6 16 fxg6 fxg6 17 a3
.i.xd2+ 18 .i.xd2 hS 19 .l:.g1 hxg4 20
hxg4 .i.e6 21 0-0-0 �d7 22 .i.g2
22 ... tLlf2?!
Black keeps an edge with 22 ... ltJxd2
23 .l:.xd2 .l:t.af8 or 23 �xd2 .i.xg4.
23 .l:ldf1 .Uaf8 24 .i.gs ltJxg4 25 l:.xf8
.l::!.xf8 26 .i.xd s .i.xds 27 .l::!.xg4 .l:t.f1+ 28
'it>d2 l:tf2+ 29 'iii> c 3 a s 30 .i.h4 l:tf3+ 31
�d2 .i.f7 3 2 d s .i.xds 3 3 l:txg6 l:.b3
Yz-Yz

Game 54
P.Hromada-L.Ostrowski
Moravia n Tea m
Cha m p ion s h i p 2003
1 e4 e5 2 ll:Jf3 ll:Jc6 3 d4 exd4 4 ll:Jxd4
'iff6 5 .i.e3 .i.c5 6 c3 'ii'g 6 7 ll:Jb5 .i.xe3 8
ll:Jxc7+ 'it>d8 9 ll:Jxa8 .i.f4 10 'it'f3 .i.b8
I recommend 10 ... .i.h 6 ! in the theory
section.
11 ll:Jd2 ll:Jf6 12 o-o-o l:te8
13 l:i.e1?!
White should be willing to sacrifice
the e-pawn, for which he will gain easy
development; e.g. 13 g 3 ll:Jxe4 14 ll:Jxe4
'ifxe4 15 'ifxe4 l:txe4 16 .i.g 2 l:te7 17 f4
and, for one thing, it is no longer clear
how Black will ever corral the wayward
knight. The text move is far too passive,
a recurring problem for White as the
gam e goes on.
13 ... b6 14 .i.b5 .i.b7 15 ll:Jxb6 axb6 16
�e3 .i.c7 17 f3 ll:Jd5 [17 ... 'it>c8] 18 'iVf2
ll:Jf4 19 g3 ll:Je6 20 .i.xc6?!
A bad idea, which weakens the light
squares around White's king and sim-
Illustrative Games
plifies Black's task of using his bishop
pair.
20 ....i.xc6 21 ll:Jb3 ll:Jc5 [21 ... 'iit e 7!] 22
ll:Jxc5 'ii'g 5+ 23 'ii'e 3 'ii'x c5 24 'ii'x c5 bxc5
25 b3 f5 26 exf5 llf8 27 .l:f.e3 :xf5 28
.l:Ihe1 .i.d6
29 :t1? [29 f4] 29 ... g5 30 '>t>b2 h5 31
J:!.d1 �C7 32 .l:.f1 �b7 33 a3 C4 34 b4
.i.c7? [34 ... h4!] 35 'it>c2?! [ 3 5 .l:i.e7 ! ]
3 5 ....i. b 6 36 lie7 'iitc 8 3 7 f4 .i.g2 38 .l:!.d1
gxf4 39 .:dxd7??
Here 39 gxf4 d5 is no bargain for
White, but the text ends the gam e im­
mediately.
39 ...f3 40 l:!.b7 .i.d8 41 l:!.f7 f2 0-1
Game 55
D.Ca m pora-V.Tkachiev
B i e l 199 5
1 e4 e5 2 ll:Jf3 ll:Jc6 3 d4 exd4 4 ll:Jxd4
.i.c5 5 .i.e3 �f6 6 c3 'i!Vg6 7 ll:Jb5 .i.xe3 8
ll:Jxc7+ ..ti>d8 9 ll:Jxa8 .i.f4 10 �f3 .i.h6 11
'ii'f5?! 'ii'xf5 12 exf5 b6 13 ll:Ja3 .i.b7 14
ll:Jxb6 axb6 15 ll:Jc4 'it>c7 16 .l:.d1 d5!
A clever thrust.
1 79
Th e Dark Kn ight System
17 lba3
If 17 l:txds?! lbf6 18 .l:.d1 .l:f.e8+ 19
�e2 �a6, White regrets his pawn grab.
11 ... lbf6 18 lDbS+ �d7 19 �e2 lieS 20
�f1 l:r.es 21 h4 lbe4 22 l1h3 lbd2+ 23
�e1 �a6 24 a4 lbc4 2 5 b3 lbd6 26
.l::r. h d3 d4! 2 7 cxd4?!
White's rooks will not enjoy lan­
guishing behind the i solated d-pawn .
However, the position i s n ot rich
enough to offer Black good winning
chances (in GM play).
27 .. J!e4 28 g4 �f4 29 �f1 lbxbs 30
axbs �xbs 31 �3d2!
Unlikely as it seems, this m ove saves
the day, since 31 ... �xd2 3 2 �xbs
1 80
(threatening both 3 3 I1xd2 and 3 3 ds)
3 2 ... �xd4 33 '>i;>e2 regain s the piece.
31 ..J:txe2 32 l:txe2 �d6 33 �e1 i.xe2
34 'iii>x e2 �d s 35 Wd3 h6
Not 3 S ... tbxd4?? 36 �c3 i.es 37 f4
i.f6 38 g s .
36 .l:.e1 �d6 3 7 l:te8 �e7 38 hS �f6 3 9
l:tc8 lbxd4 4 0 l:!.c7 lbc6?!
40 ... lbxb3 is a better try.
41 f4 lbe7 42 �a7 �c6 43 �e4 Yz-Yz
Game 5 6
I.Grynfeld-A.Bisguier
H e l s i n ki O lym p i a d 1 9 5 2
1 e4 e5
Once, when GM Bisguier was ana­
lysing one of my games, he said he'd
known all about ...lbc6 and ... e7-e5
since long before I'd been born. Thi s
m akes sense because he was coached
by Alexander Kevitz, by whose nam e
this system i s known .
2 lbf3 tbc6 3 d4 exd4 4 lbxd4 �cs 5
lbxc6 'iVf6 6 lli'f3 'iVxf3 7 gxf3 bxc6 8
i.e3 �xe3 9 fxe3

As I m entioned in the theoretical
section, White has done pretty well
here. Bisguier's plan takes care of any
potential problems, though I think it's
simpler to start with 9 ... d6.
9 .. l'Llf6 10 l'Llc3
. J .Smeets-A.Beliavsky
After 10 e s ! ? l'Llg8 11 l:!.g 1 g6 12 l'Llc3
f6 13 exf6 l'Llxf6 14 o-o-o o-o, White has
little to nothing either, but I still see no
reason to allow him this option.
10 ... 0-o 11 o-o-o [11 e s l'Llds] 11 ... Ile8
12 i.h3 d6
Also possible i s 12 ... g6 (a useful
semi-waiting move}, intending ....l:!.e s­
h s . After the text, Black slips into a pas­
sive position.
13 i.xc8 .l:.axc8 14 .l:[d4 as?! [14 ... l'Lld7]
15 .l:.c4 cs 16 .l:.a4 :as 17 l'Llbs .l:.ec8 18
l'Lla 3?! [18 e s ! ?] 18 ...l'Lld7 19 l'Llc4 l'Llb6
20 l'Llxb6 cxb6
Black h as equalized (again). Nothing
else happens.
21 c4 Wf8 22 l:ta3 �e7 2 3 l:tb3 l:!.ab8 24
a4 'it>e6 2 5 l:td1 .l:tc6 26 J::[g 1 g6 2 7 �d2
.:.fs 28 .l:.g3 fs 29 exfs+ .l:.xfs 30 e4 .l:.hs
31 h3 gs 3 2 �e3 h6 3 3 .l:i.d3 l:th4 34 1Ib3
�es 35 :d3 .:!.c7 36 .l:tb3 l:tb7 3 7 l:td3
Illustrative Gam es
'it>e6 38 l:!b3 '>td7 39 l:td3 'itc6 40 �f2
.l:tf7 41 �g2 Yz-Yz
Came 5 7
N et h e rl a n d s-Siove n ia
ra pid m atch, Ma ri bor 2004
1 e4 es 2 l'Llf3 l'Llc6 3 d4 exd4 4 l'Llxd4
i.cs 5 l'Llxc6 �f6 6 �d2 dxc6 7 l'Llc3
i.d4 8 i.d3 l'Lle7 9 o-o l'Llg6 10 l'Lle2
i.xb2!
Beliavsky courageously grabs the
pawn - the "principled" move (what­
ever that means).
11 ..txb2 �xb2 12 f4 'ii'a 3 13 fs l'Lles 14
'it'gs i.. d 7
I am recomm ending 14... 'i¥f8 !, when
White has worries about proving full
compensation, though Black h ad his
chances in this gam e as well.
15 �xg7 'ifcs+ 16 'it>h1 o-o-o 11 'i¥f6 hs
18 h 3 �b8
18 ...'i¥d6 ! ? with a structural advan­
tage in the endgame.
19 l:tf4?! i..c 8 20 l:r.h4?
1 81
Th e Dark Kn ight System

if Black wants to win ? Pokazanjev's

m ove is very risky - in fact, it's objec­
tively poor, but sometimes it's worth
taking such risks. Perhaps Black should
have deviated earlier: 10 ... 0-o is not a
bad move, and avoids the present di­
lemma; 10 ... 'i!fh4 is also reasonable ac­
cording to Houdini.

20 ...l:thg8?
Black is nearly wmnmg with
20 ... tt:Jxd3 2 1 cxd3 'il¥f2 22 lt:Jf4 .l:!.hg8,
because White's queen, knight, and
king's rook are stuck in a very un­
healthy eo-dependent relation ship.
21 �f1 l:!.d6 22 'ir'e7 'il¥a 5?! 23 llxh 5 l::[d 7
24 'ii'f6 .l:td6 2 5 'fie7 .l:i.d7 26 'ii'f6 J::r.d 6 27
'fie7 .l:!.d7 Yz-Yz 12 ... 1i.d7?! 13 lt:Ja4?!
Instead of repeating moves, Black Out of book, White m akes a poor
was much better after 27 ... tt:Jxd3 28 choice. As we will see, Black does not
cxd3 (or 2 8 'ii'xf7 lt:Jf2+!) 28 .. .'fVxa2 29 even have to allow his bishop to be
lt:Jf4 �xd3 ! (and not 30 tt:Jxd3 ?? 'ii'xg 2 forced off the strong diagonal.
m ate).

Cam e 58
B.Sultimov-N.Pokazanjev
Russia 2007

1 e4 e5 2 lt:Jf3 lt:Jc6 3 d4 exd4 4 tt:Jxd4

1i.c5 5 tt:Jxc6 'ir'f6 6 'ifd2 dxc6 7 lt:Jc3
1i.d4 8 1i.d3 lt:Je7 9 o-o lt:Jg6 10 Wh1 lt:Je5
11 f4 lt:Jg4 12 �e1
White h as certainly lost the theo- 13 ... a6 14 c3 1i.a7 15 b4 "fie7 16 .l:!b1
retical battle since 12 ...lt:Jxh 2 13 'it>xh 2 0-0-0
Vi'h6+ is a draw by perpetual. But what Black is not fooled by White's

1 82

queenside demonstration, which is
mainly bark and has little bite.
11 �g3 h 5 18 h3? i..f5!!
19 tt:Jc5
The bishop should not be taken : 19
exfs ? tLlf2+! 20 l:txf2 'ii'e 1+ 21 i..f1 (or
21 Wh 2 i.. xf2 and 22 ... 'i!Vg 1 m ate)
21 ... i.. xf2 22 i.. d 2 �xb1 23 'i!Vxf2 .l:!.xd2
and wins - though White is just about
lost anyway.
19 i..x c5 20 bxc5 i..xe4 21 i..xe4 'ii'xe4
•••
22 .l::t b4 'i!Ve2 23 it'f3
The endgame will bring White little
relief - he cannot possibly hold onto his
six isolated pawn s.
23 'ii'xf3 24 gxf3 tt:Jf6 2 5 f5 .l:f.d5 26 i..e 3
••.
Illustrative Games
.l:f.xf5 21 i..d4 tLld5 28 .l::t b2 .l:th6 29 c4 tt:Jf4
30 .l:tfb1 tt:Je6 31 i..g1 .l:txf3 32 I:txb7
lb:h3+ 33 i.. h 2 11d3 34 .l:!.a7 .l:!.f6 35 .!:!.xa6
l:tf2 36 a4 1::t h 3 37 l:ta8+ Wd7 38 �g1
.i:.hxh2 39 1Ie1 l:tfg2+ 40 'it>f1 tZ::lf4 0-1
Dealing roughly with 7 ..tf4?!
Game 59
D.Von Wa ntoch
Rekowski-J.Peric
Yugoslav J u n ior C h ' s h i ps,
Tivat 2001
1 e4 e5 2 tLlf3 tt:Jc6 3 d4 exd4 4 i..c4 tLlf6
5 e5 tt:Jg4 6 o-o i.. e 7 1 i..f4 g5! 8 i.. g 3
h5! 9 h 3 h4!
10 i.. h 2?!
10 hxg4 hxg 3 11 fxg 3 d6 is relatively
best, though still with advantage to
Black.
1o tt:Jxh2 11 'it>xh2 d6 12 i.. b 5 dxe5 13
.••
tt:Jxe5 'ii'd 6 14 f4 i.. e 6 15 c4? gxf4 16
'ili'f3 Wd8?! [16 ... �f8 ! ] 17 tt:Jxc6+ bxc6 18
i..xc6 .i:.b8 19 b3 .l:tg8 20 tt:Jd2 l:i.g3 21 c5
'i¥xc5 2 2 .:tfc1 .l:.xf3 2 3 i..x c5 l:.xh3+ 24
gxh3 i..x c5 2 5 tt:Je4 .:!.b6 26 i.. a 4 i.. d 6 27
1 83
Th e Dark Knight System
l:td1 .ie5 28 tt:Jg5 <:J;;e 7 29 lt:Jf3 f6 30
tt:Jxd4 I:td6 31 tt:Jc6+ 'it>f8 3 2 l:txd6 .ixd6
33 tt:Jd4 f3+ 34 �h1 .td5 o-1
Black did not have a high rating, but
she produced the model g ame.
Gam e 60
J.Becerra Rivero-A.Miles
Ando rra 199 5
1 e4 tt:Jc6 2 lt:Jf3
Like most high-rated players, White
tries to narrow the knowledge g ap with
thi s second move. Instead, 2 d4 es ( or
in some cases 2 ... d5 ) fights in Black's
home territory. Of course, there is noth­
ing stopping Black from also knowing 2
lt:Jf3 d6 positions well.
2 ... d6 3 d4 lt:Jf6 4 C3 g6 5 .id3 .ig7 6
lt:Ja3
Becerra Rivero i s the only one to
have tried thi s m ove h ere. Although
decentralizing, it has the merit of not
interfering with White's other pieces.
6 ...0-0 7 0-0 e5 8 dxe5?1 tt:Jxe5 9 tt:Jxe5
dxe5
1 84
White's decision to exchange in the
centre is not consistent with his previ­
ous play. The c3-pawn i s intended to
shorten the diagonal for Black's fi­
anchettoed bishop and/or hold the d4-
point, neither of which are meaningful
anymore. White's d3-bishop is also
misplaced, obstructing the open d-file.
H e no longer has any trace of an ad­
vantage. Perhaps White is trying to
make a draw, in which case he has cho­
sen the wrong opponent.
10 �e2 file7 11 .ig5 h6 12 .ixf6?1
This small inaccuracy leads to
White's future problems. We will wit­
ness dark-square torture yet again.
12 .. .'ili'xf6 13 .tc4 .td7 14 .l:.ad1 .l:.ad8
15 l::t d 3 h5 16 "ife3 a6 17 .l::f.fd1 .ig41? 18
f3 .ic8
Miles's bishop manoeuvre softens
up the dark squares for his unopposed
dark bishop. There is certainly nothing
concrete yet, but it is getting to where
White needs to be careful .
19 .tb3 :xd3 20 l:txd3 h4 21 h3 Wi/e7 22
lt:Jc4 �h71 2 3 f4?1
Thi s gam e reminds m e of the old

story of the old lady who swallowed a
fly, and a spider to catch the fly, and a
bird to catch the spider, etc. White pre­
vents ... .i.h6 for now, but the new prob­
lems are bigger than the old ones.
23 ... exf4 24 'iixf4 b5 2 5 tt::l e 3 Si.e5
26 'ti'f3 [26 tt::l d s!] 26 'ifg5 21 .kld1 'ito>g1
••.
28 .:.d 5 'ii'g 3 29 tt::l g4 'ifxf3 30 gxf3 i.f4
31 �g2 .l:le8 32 l:!.d1 f5 3 3 tt::lf2 Si.g3 34
ttJh1 Si.d6 35 tt::lf2 �f6 36 i.d 5 :th8 37
tt::ld 3 .Uh5 38 f4 fxe4 39 Si.xe4 l1h8 40
.i.f3 l:te8 41 Si.g4?
In principle, this is a favourable
trade for White, who relieves Black of
the bishop pair, but the devil is in the
details.
41 Si.xg4 42 hxg4 .l:.e3 43 'ito>f1 Si.xf4 44
•••
tt::l b4 'it>g5 45 tt::l d 5 l:tf3+ 46 'ito>e2 l:r.g3 47
l:rd4 Si.d6 0-1
Game 6 1
R.Zelcic-A.Miles
P u l a 1994
1 e4 tt::l c6 2 tt::lf3 d6 3 d4 tt::lf6 4 c3 g6 5
i.b5 Si.d7
Illustrative Games
s ... a6 i s al so fine, or even s ...Si.g 7 6
dS a6 7 Si.a4 bS 8 dxc6 bxa4 9 'ii'x a4 o-o
with compensation .
6 1i'e2 i.. g 7 7 0-0 0-0 8 l:td1 'i!ke8 9 C4 e5
10 dxe5 tt::lx e5 11 ..txd7?! tt::l xf3+ 12
'ifxf3 tt::l x d7
Black i s already better.
13 'ii'e 2? f5 14 tt::ld 2 tt::lc 5 15 .1:te1 'iff7
[1S ... 'i!Va4!] 16 exf5 .l:.ae8 17 fxg6??
hxg6?
Overlooking 17 ... �xg 6 18 �f1 l:i.xe1
19 'i!Vxe1 Si.d4, which win s immediately.
18 'iif1 .l::!.x e1 19 'ifxe1 tt::l d 3 20 'i&'f1
'iixf2+ 2 1 'it'xf2 lbf2 22 tt::lf3 .l:.c2 2 3
Si.e3 tt::lx b2 24 l:tb1 tt::l xc4 2 5 Si.xa7 b 6 2 6
..tb8 tt::l e 3 27 tt::l e 1
1 85
Th e Dark Kn ight System
27 ....l:te2?
Miles puts the gam e in jeopardy
with this mistake. The c-pawn is Black's
most valuable.
28 i..x c7 i.. c 3 29 lbf3 l:!.xg2+ 30 'it>h1
�f2 31 lbg5 i..e 5 32 i..x b6 l:!.xh2+ 33
'lii> g 1 .l:!.e2 34 lLlf3 i..f4 35 .:e1 Ibe1+ 36
lbxe1 lbc4 37 i..d 4 Wf7 38 Wf2 We6 39
lbg2 i.. d 2 40 'it>f3 g5 41 We4 d5+ 42
'lii>d 3 i.. a 5 43 lbe3 lbxe3 44 i..x e3 g4 45
i..f4 i.. b 6 46 i..g 3 �d7 47 i..f4 Wc6 48
i.g3 i.. c 5 49 a4 i..d 6 50 i.. e 1? [so i..h 4!]
50 ... g3 51 'iii>e 2 g2 o-1
Gam e 62
L.Rozman-J.Schuyler
Wa s h i ngto n 2012
1 e4 lbc6 2 lLlf3 d6 3 d4 lbf6 4 c3 g6 5
i..d 3 i.. g 7 6 o-o o-o 7 h 3 e5 8 lba3 d 5 1
White h as the slow c2-c3 and h 3-h 3
under his belt, and the decentralizing
lba3, so I felt it was time to hit back in
the centre.
9 i.. g 5?1 dxe4 10 i..xe4 exd4 11 i..x c6?!
White was hoping to damage my
1 86
structure in stead of accepting the iso­
l ated d-pawn. This would have been
fine for Black, but I saw something I
liked better.
11 ... dxc3! 12 'ilfa4?!
The critical tries involved saving the
bishop, but since Black was getting
three pawn s for it, including the dan­
gerous b2-pawn, I was willing to take
my chances (and White was not).
12 ...bxc6 13 'ifxc6?
Either 13 bxc3 or 13 .l:!.ad1 was cor­
rect.
13 ... cxb2 14 .l:tad1 i.d7 15 'it'c2 'ilic8?!
I was eager to break the pin s, but
analysis shows I could have safely
m ade it harder for White to recover the
b-pawn with 1 S ... .l:.b8 ! .
1 6 'i1Vxb2 lbe4 1 7 'ii'b4 .l:.e8
Incidentally threatening 18 ... i.. xh 3
19 gxh 3 'it'xh 3, which will win the piece
back on the kingside.
18 �h2?
Not a working defence, though nei­
ther one of us noticed the obvious refu­
tation : 18 ... i.. xh 3 ! 19 gxh 3 'iifs, again
recovering the piece.
 Illustrative Games

18 �f8! 19 'ii'a s i.d6+ 20 'it>g1 �xh3

.•. takes, most of White's pressure has
2 1 gxh3 'i!Vxh 3 22 l:td3 lLl g 3 ! 2 3 fxg3 dissipated.
'it'xg3+ 24 'it>h1 .ri.e2 0-1 29 lbbs 30 l:tccl fs 31 lba3 ll:lxa3 3 2
•••

'i!Vxa3 'ii'd 7 3 3 ..ib6 l:tbc8 34 � b 3 �h6

Do not try this at home, 35 .l:1.a1 � gs 36 l:ta4?!
or even when away from home

Game 63
A.Zapata-A.Miles
Mata nzas 199 5

1 e4 lbc6 2 ll:lf3 d6 3 d4 lbf6 4 lbbd2 g6

s c3 �g7 6 i.bs o-o 1 o-o i.d7
7 ... a6 is better, trying to pick up the
bishop pair. Black's bishop m ay prefer
to deploy on b7 or a6.
8 .l:i.e1 li'e8 9 i.f1 es 10 h3 'ii'd 8 11 a4 By now, White is actually worse.
l:te8 12 as a6 13 dS lbe7 14 c4 lU8 36 �d8 37 .l:.b4 fxe4 38 �xd8?! 'ii'x d8
•..

39 .l:i.xb7 �xa s 40 'iVd1? [40 'ifl>4]

40 'iVcs 41 'i¥d2 .l::f.f4?
.••

41 ... l::t xf2+! 42 'ii'xf2 l::tf8 wins.

42 �gl 'iVd4 43 .l:i.dl 'it'xd2 44 :xd2 as
45 .l:i.b6 .l:tcl+ 46 Wh2 I1f6 47 l:.bs l:tfl 48
�g3 hS 49 l:txa s .l:i.f4 so f3?? exf3?
H ere 50 ... e 3 ! 51 l:tc2 l:th l 52 h4 g 5 .
51 gxf3 l:l1xf3+ 52 �g2 h 4 5 3 l:.a6 llf6
54 l:la4 l:t3f4 5 5 b4 �g 7 56 .l:tb2 'lfi>h6 57
z:ta7 .l:!.f3 ss bs? [58 �h 2] ss l:lg3+ 59
•••

�h2 l:tff3 60 l:ta4 l:txh3+ 61 'lfi> g 1 .l:!.hg3+

While each of Black's m oves had its 62 l1g2 .l:i.xg 2+ 63 �xg2 .l:i.b3 0-1
purpose, the overall impression is very
poor. This is a bad King's Indian . Try this at home
15 cs lbe8 16 ll:lc4 lbc8 11 'iVb3 �bs 18
i.e3 l:tb8 19 l::ta c1 lbf6 20 lbfd2 lbd7 21 Game 64
cxd6 cxd6 22 'ii'a 3 lLlf6 2 3 ll:lb6 i.xfl 24 F .Nijboer-A.Miles
Wxfl lba7 2 5 'ii'b4 lbbs 26 lbb1 lbe8 27 Li n a res 1995
lLlc3 lbxc3 28 .l:i.xc3 lbc7 29 ll:lc4
Without making any obvious mis- 1 e4 lbc6 2 lLlf3 d6 3 d 4 ll:lf6 4 lbbd2 g6

187
Th e Dark Kn ig h t System
s �bs a6 6 �xc6+ bxc6
As we will see, thi s position is much
easier to play than ... i.d7 of the previ­
ous game. Miles's m ethods of play are
worth remembering and repeating.
7 o-o �g7 s es tt:ld s 9 c4 tt::l b 6 10 'it'e2
0-0 11 l:!.d1 a s 12 tt:lf1 �a6 [12 . .f6 ! ?] 13
. and 10 ... i.a6.
b3 �d7 14 tt:lgs �cs 1S f4 a4 16 �e3
f6! 17 tt:lf3 'ii'es 1S h 3 fxes 19 fxes h6
20 l:!.ac1 axb3 21 axb3 g5
Black has been preparin g ... �fS,
though after White's n ext the plan
changes.
22 ds cxds 2 3 cxds i.a6 24 'ii'd 2 �xf1
2 5 .J:f.xf1 'i!Vbs 26 llfd1?! 'iix b3 2 7 Itxc7?
tt:lc4 2S 'ii'd 3 'ii'x d3 29 l!xd3 l:ta1+ 30
�f2 tt:lxes 31 l:.d2 .l::tf7 32 .l:a7? .l:!.b1?
[32 ... .l:r.h 1] 33 �a4 'iii> h 7 34 'iii>g 3 llb3 3 5
I1e4 l:t a 3 36 h4?! tt:lxf3 37 gxf3 gxh4+
3S 'iii>g4 i.f6 39 f4 .l:tg7+ 40 'iii>fs h3 41
l:!.h2 I1g3 42 �cs? dxcs o-1
Gam e 65
E.Sveshnikov-T.Gelashvili
Ca p pe l le l a G ra nd e 2009
1 88
1 e4 tt:lc6 2 tt:lf3 d6 3 d4 tt:lf6 4 tt:lc3 g6 5
.i.bs a6 6 i.xc6+ bxc6 7 o-o �g7 S 'ife2
0-0 9 l:i.d1 �g4
This m akes it easier to fight for the
central dark squares and advances a
clear plan. Nonetheless, I prefer 9 ... as
10 h 3 �xf3 11 �xf3 tt:ld7 12 �e3 es 13
'ife2 .:.es 14 �c4?!
White could have bottled up Black's
bishop and rook with 14 d S ! .
1 4 exd4 1S i.xd4 i.xd4 16 'ii'x d4 'ii'g 5
•••
17 'ili'd2 'ii'h 4 1S .l:te1 tt:lcs 19 f3 l:tes 20
'ii'f2 'fle7 21 tt:le2 tt:le6 22 tt:ld4 tt::lx d4 23
..Wxd4 as 24 'ii'c 3 ltcs 2 S iVd2 .l:.es 26 a4
1i'f6 2 7 c3 .l:!.bS 2S l::te 2 .l:!.es 29 l:.ae1 gS
30 'ife3 c5 31 'ii'd 3 'ife6 3 2 'i¥a6 c4 33
'i¥a7 :cs 34 'i¥d4 lieS 3 S Wf1 f6 36 iVd2
�f7 37 �g1 <J;e7 3S Wh2 r:3;d7 39 'ii'd 4
'ilif7 40 'ii'f2 hS 41 �g1 cs 42 .l:.d2 fS 43
.l:ted1 .l:!.Se6 44 'ii'e 2 fxe4 45 'ii'xc4
Black has been pressing h ard for the
win, but without anything to work
with. Now he is in danger.
4S ...1i'f4?! 46 'ii'a 6?!
46 ifbs+ We7 47 ii'b7+ Wf8 48 fxe4
is stronger.
 Illustrative Games

46 ... exf3 47 .l::r.x d6+ �e7 48 'ii'a 7+ �f6 i.c4 li:Jc6 6 d S li:Jb8 7 h3 o-o 8 o-o c6
49 l:!.xe6+ l:.xe6 so 'ifxcs �g6 51 'ifd4 [8 ... e s ! ] 9 a4! as 10 l:!.e1 tt::lfd71
'ii'fs 52 g4 hxg4 5 3 hxg4 'iff7? 54 .l:Ul?l
And h ere 54 b4! .
5 4...l::.f6 s s 'it>f2?1 'it'b3 56 'ii'd 3+ Wg7 57
.:d1 �xb2+ 58 lld2 'fibS 59 'i!id7+ 'iii> h 6
60 �f1 'ii'b 1+ 61 l:!.d1 'ii'e 4 62 'ii'd 2??
After thi s move White i s losing. 62
:d2 held the bal ance.
62 ...f2 63 �d s 'ii'xg4 64 'ii'h 1+ 'it>g6
[64 ... �g 7 ! ] 6S l:d4

Black clamps down on the dark

squares, incidentally stopping White
from considering the e4-e5 break
11 i.e3 lt:Ja6 12 i.d4 [12 i.xa6 !]
12 ...lt:Jb4 13 i.xg7 �xg7 14 li:Je2 'fic7 15
c3 cxd s 16 i.xd s li:Jc6 17 lLled4 li:Jf6?1
Weakening the dark square control.
White now has 18 i.xc6 bxc6 19 e s
with a pleasant advantage.
6s .. JWc8? 18 'ifb3?1 lt:Jes l 19 lt:Jxes dxes 20 tt::lf3
6S .. .'�e6 (or 6s ...'ii'fs) 66 'ii'e 4+ l:ta6 21 l::ta c1 lt:Jhs 22 i.c4?1 .l:[f6 2 3
(forced) 66 ... 'i!Vxe4 67 l:.xe4 �h S win s, lt:Jd2?! 'i!Vcs 24 � e 3 l:.d8 2 5 li:Jfl?l li:Jf4
since if 68 .Ue2 then 68 ...�g4 69 I!xf2 26 'ifbs 'Wic7 2 7 'ii'b 3 h s 28 'iia 2 .l:.fd6 29
l:.xf2+ 70 �xf2 �f4 etc. 'fib3 i.d7 30 i.bs i.e6 31 'iic 2 h4 32
66 'ii'e4+ �h6 67 'iYe3 'ii'f s 68 'ti'd3? :tee1 'ii'c s 3 3 l:.e3 gs 34 f3?1 lt:Jhs
�xd3+ 69 l:!.xd3 .l:!.g6?? [34 .. .'ii'c s ! ] 35 .litce1 [35 :eel] 3 S .. :i1Vcs
69 ... �h 5 still wins. 36 'itr>h2 'iii>f8 37 .l:!.d3?1 l:!.xd3 38 i.xd3
70 C4 Yz-Yz i.xh3 ! 0-1

Gam e 6 6 Game 6 7
N.Praznik-A.Beliavsky K.Nemcova-F .Oiafsson
B l ed 1999 Mari a n s ke La zne 2008

1 e4 d6 2 d4 li:Jf6 3 li:Jc3 g6 4 li:Jf3 i.g7 5 1 e4 d6 2 d4 g6 3 li:Jc3 i.g7 4 i.e3 lt:Jc6 5

1 89
Th e Dark Kn ight System
ds lt:lb8 6 'ii'd 2 c6 7 lt:lf3 lt:lf6 8 h 3 bs?!
This is a great idea if it works out
tactically, but it doesn't. 9 dxc6 b4 10
i.bs is a problem for Black. So let us
imagine instead that the game went
8 ... 0-0 9 i.d3 bS 10 a3 a6 11 dxc6 etc.
9 a3 a6 10 dxc6 lt:Jxc6 11 i.d3 o-o 12
0-0 i.b7
Black's opening has been extremely
successful - White's gam e has a
cramped feel in spite of h er central
space advantage.
13 .:tfe1 ds?!
This "freeing" m ove i s quite unnec­
essary. Black has m any improvements
he can m ake (such as ... .l:!.c8, ... lt:ld7,
.. :ii'as, or even ... e7-e6) before he needs
to consider taking action. lt is White
who will be left without anything con­
structive to do.
14 i.h6?
I've never heard anyone say that it's
a good idea to react to central action
with play on the wings.
14 dxe4 15 i.xg7 'it>xg7 16 lt:Jxe4 lt:Jd4
.•.
17 lt:Jxd4 'ilt'xd4 18 'it'c3 'it'xc3 19 lt:Jxc3
e6
190
White i s slightly worse due t o h er
very restricted bishop, though she
solves this problem quickly.
20 lt:Je4 .l:!.fd8 21 lt:lcs i.c6 22 a4 bxa4 2 3
lt:Jxa4 ..t b s 24 lt:J c s :ac8 2 5 b 4 Ad4 2 6
..txbs axbs 27 c3 .:.c4 28 .l:!.a7 'it>g8 2 9
.l:tb7?! [ 2 9 l:l e 3 lt:lds 30 l1f3] 29. . .l:txc3
30 .l:!.xbs lt:ld s 31 .l:!.b7 lt:Jxb4 32 l:txb4
l:t8xcs
A theoretical draw, but Black has
some practical ch ances.
33 l:.e2 .:.sc4 34 .:bb2 gS!
This stops White from constructing
the ideal defensive pawn formation for
this endgame: f2/g3/h4. The point of
the formation is that Black cannot cre­
ate a passed pawn without trading off
all the rest of the pawn s.
35 'iii> h 2 <t>g7 36 .Ubs 'it>g6 3 7 x:tbb2 hs
38 1:ta2 h4 39 .l:l.e1 .l:.cs 40 'it>g1 'itis 41
'iti1 es 42 f3 �hc4 43 .l:!.ee2 r;t>f4 44 �2
fs 45 .l:!.a3 .l:!.bs 46 .l:!.ae3?
The endgame has been getting
h arder to defend. Now, White is lost.
46 ....l:!.c1 47 .l:!.e1 .l:lb2+ 48 .l:13e2 .Ucc2 49
l:!.xc2 .:txc2+ so �1 l:lb2 51 r;t>g1 'iii> g 3
0-1

Game 68
J.Hjartarson-F.Oiafsson
Reykjavi k 199 5
1 e4 g6 2 d4 d6 3 tZ:lc3 .i.g7 4 .i.e3 tZ:lc6 5
ds tZ:lbS 6 'it'd2 c6 7 tZ:lf3 tZ:lf6 8 h 3 o-o 9
i..e 2 b51 10 a3
White has nothing better than this
l am e try. Black h as equalized already.
10 a6 11 o-o .i.b7 12 I:l.fe1 tbbd7 13
•••
dxc6 .i.xc6 14 .i.d3 .i.b7 15 .i.h6 .:!.c8 16
�ad1 1i'b6 17 ..txg7 Wxg7 18 :e3 tbes
19 tbxes dxes 20 .l::i.e 2 .:!.fd8
By now Black is a little better, but
the players soon decided to call it a day.
21 'i!Ve3 'ifd4 22 .:!.de1 tbhs Yz-Yz
Game 69
Y.Gruenfeld-I.Smirin
I s ra e l i Tea m
C h a m pion s h i p 1997
1 e 4 d6 2 d 4 tbf6 3 tbc3 g 6 4 tbf3 .i.g7 5
.i.e2 o-o 6 o-o tbc6 7 d s tbbs 8 h3 es 9
dxe6 .i.xe6 10 ..tgs h6 11 .i.e3 tbc6 12
Illustrative Gam es
�d2 Wh7 13 .l::!. a d1 .:!.e8 14 .:!.fe1 .i.d7 15
i.c4 i.e6 16 .i.f1 a6
17 'i!Vc1
Very soon Black has all the play he
needs. White can make him work
harder with 17 a4.
17 .. -'i!Vcs 18 a3 bs 19 tZ:ld4 tZ:lxd4 20
.i.xd4 't!Vb7 21 �f4 tbd7 22 .i.xg7 'it>xg7
23 �d2 'i!Yb6 24 .l:.e3 �adS 25 g3 tZ:lf6 26
.i.g2 a s 27 b3 .l::. b 8 28 a4 bxa4?1 [28 ...b4]
29 tZ:lxa4 'ii'b4 30 tl:\c3 tl:\d7 31 Wh2 tZ:lcs
32 f4 f6?1 33 'i!Ve2?l [33 fs !] 33 ....i.f7 34
tZ:lds .i.xds 35 exd s .l:be3 36 'ili'xe3 l:tf8
37 fs gs 38 �e2 tbd7 39 l:ta1 tbes 40
l:!.a4 �c3 41 l:!.e4 .:!.b8 42 �h s .l::!. b4 43
.:!.e2 a4 44 'iVeS axb3 Yz-Yz
Game 70
B.Chatalbashev-M.Popchev
Caca k 1991
1 d 4 d6 2 e 4 tbf6 3 tZ:lc3 g6 4 tZ:lf3 i..g 7 5
.i.e2 o-o 6 o-o i.. g4 7 i..e 3 tZ:lc6 8 d S
tbb8 9 tZ:ld4 .i.xe2 1 0 �xe2 cs 11 tbf3
'i'ib6 12 I:tab1 1lia6
Chemin points out that this queen
191
The Dark Knight System
m anoeuvre is both typical and strong.
13 'iVd2 tt'lg4 14 ..tg5 .:tes 15 'ii'f4 tt'le5
16 tt'lxe5 ..txe5 17 "YWh4 "iWc4
18 .l::!. b e1
White gains nothing by grabbing
the pawn : 18 ..txe7 tt'ld7 19 f4 ..id4+ 20
Wh 1 ..txc3 21 bxc3 "YWxe4 is equal .
18 ..txc3 19 bxc3 f6 20 ..ih6 tt'ld7 21
•..
'fi'g4 e6 22 f4 .Uad8 2 3 dxe6 �xe6 24
'i!Vg3 'ii'x a2 2 5 f5 'i!Vf7 26 fxg6 hxg6 27
'i!Vxd6 �h7 28 'iff4?1
lt i s tim e to extricate the bishop.
28 l:.e5 29 ..tg5 .:tfs 30 "YWh4+ 'ifilgs 31
•••
..if4 �e6 32 llf3 g5 33 'ii'g4 lUeS 34
..tg3 tt'le5 35 ..txe5 .l:.xe5 36 h4?1 'ii'e 6
37 'ii'h 5 �g7 38 .l::tef1??
Oops! Black had an advantage, but
thi s throws the game.
38 .l:th8 39 'ii'x h8+ 'ifilxh8 40 .Uxf6 'ir'c4
••.
0-1
Gam e 71
R.Ziatdinov-A.Chernin
N ew York Open 1998
1 e 4 g 6 2 d 4 ..tg7 3 tt'lf3 d6 4 ..t e 2 tt'lf6 5
192
tt'lc3 o-o 6 o-o ..ig4 7 ..te3 tt'lc6 8 d 5 tt'lb8
9 tt'lg5 ..txe2 10 'ii'xe2 c6 11 �ad1 'ii'a 5
12 f4 'ii'a 6 13 'ii'f3 tt'lbd7 14 li'h3 h61 15
tt'lf3 h51 16 ..id4?1 c5 17 ..ixf6 tt'lxf6 18
'ii' h 4?1 b5
White hasn't gotten anything done
on the kingside, but he has serious
problem s on the queen's wing.
19 tt'le2?1 'ii'a 4 20 tt'lg3 'ii'x c2 21 e5 dxe5
22 fxe5 tt'lg4 2 3 .l:l.d2 'ili'c4 24 "iix e7 I;Iae8
25 d6?1
A desperate try for complication s
which fails after Chemin's accurate
treatment .
2 5 ... ..th61 26 b3 'ifb4 2 7 e6 l:Ixe7 28
dxe7 ..txd2 29 exf7+ �xf7 30 exf8'ii'+
'iitxf8 31 tt'lxd2+ 'lt>g7 3 2 tt'lge4 'i!Vd4+ 3 3
'it> h 1 c 4 34 bxc4 bxc4 3 5 h3 tt'le3 36 .l:r.f4
c3 3 7 tt'lf3 'ii'd 1+ 38 'ifilh2 tt'lf1+ 39 'iii> g 1
tt'ld2+ 40 'it>f2 tt'lxe4+ 41 .l:f.xe4 c2 0-1
Game 72
So.Polgar-J.Fries Nielsen
R i m a vs ka Sobota 1991
1 e4 g6 2 d4 ..tg7 3 tt'lc3 tt'lc6 4 tt'lf3

If 4 i.e3, Fries Niel sen intended
4 ... ds, an interesting and surprising
m ove that I don't believe in.
4 ... d6 s h 3 tt:Jf6 6 i.gs o-o 1 'ii'd 2 ds
In the next game, Black tries the en­
terprising 7 ... a6.
8 exd s tt:Jxd s 9 i.h6 tt:Jxc3 10 i.xg 7
Wxg7 11 �xc3 �dS 12 0-0-0!
12 .ltc4? ! 'fias is ineffective.
12 ... i.e6 13 b3 i.fs 14 tt:Jh4 i.d7 1S
.J:td2?! �g8 16 �e3 'i!Vd6 17 C3 aS 18
tbf3 'fia3+
19 �d1?!
Asking for trouble. 19 .l:!.b2 was cor­
rect.
19 ... a4 20 'ii'h 6 f6 21 i.c4+ e6 22 dS
tbe7?
Since White has counterplay in ei­
ther case, it does not help to give up
the pawn. Hence 2 2 ... tt:Jd8 ! .
2 3 dxe6 i.c6 2 4 tt:Jd4 axb3 2S tt:Jxc6
bxc6 26 i.xb3 .l:tab8 27 :e1?!
The way to extricate the king i s 27
We2 ! l:txb3 2 8 axb3 'itxb3 29 l:thd1.
27 ....l:i.xb3 28 axb3 'itxb3+ 29 'ite2 'itc4+
30 l:td3?
Walking into a dangerous pin . 30
Illustrative Games
Wf3 ! i s still approximately equal.
3 o tt:Jds 31 g 3?
..•
My computer tells me that White is
okay after 3 1 e7!! lle8 3 2 Wd2 and that
all other moves lose brutally. I won't try
to explain, especially since I don't un­
derstand it myself.
31 ... 11a8 3 2 �d2 .l:ta2 3 3 .l:!.b1 tt:Jxc3+ 34
We1 �e4+ 0-1
3 S ... 'fih 1 mate follows.
Game 73
M.Yilmazyerli-D.Arutinian
I sta n b u l 2007
1 e4 tt:Jc6 2 tt:Jf3 d6 3 d4 tt:Jf6 4 tt:Jc3 g6 S
h3 i.g7 6 i.gs o-o 7 �d2 a6 8 o-o-o bs
9 a3 .l:tb8 10 i.h6 b4 11 axb4 tt:Jxb4 12
i.xg7 'iitx g7 13 es tt:Jfds 14 tt:Jxds tt:Jxds
1s h4 h s 16 tt:Jgs?!
The knight can stay here as long as
it wants, but it does not have impor­
tant targets in Black's position.
16 ... c6 17 i.c4 'ii' b 6 18 b3 i.fs 19 exd6
exd6 20 g4 i.xg4 21 .l:.dg1?! i.fs 2 2
tb h 3 'itb4 2 3 'itxb4 .l:lxb4 24 .l:tg3 as 2 S
193
Th e Dark Kn ig h t System
..ixd s cxd s 26 .l:.d1 :c8 21 .l:!.d2 a4 28
�b2 axb3 29 cxb3 .l:.cc4 30 ltJf4 .l:.xd4
31 .l:!.xd4 .l:!.xd4 32 ltJxh5+ 'iitf8 3 3 lt:lf6
.Ud1 34 h s gxhs 3 5 lt:lxh s d4 36 l:i.gs?!
.l:!.d2+ 3 7 'iii>c 1 .l:!.xf2 38 lt:lg3 ..ig6 39 .:td s
I:. g 2 40 ltJhs .l:.f2 41 ltJg 3 l:!.f3 42 lbe2 d3
43 lt:lc3 �e7 44 b4 �e6 45 l:i.d4 .U.f2 46
ltJbs .l:.c2+ 47 'it>d1 ..ih5+ 48 'iit e 1 .l:!.e2+
49 Wf1 d2 o-1
Gam e 74
N.Ryba-J.Schuyler
Wa s h i n gto n 2012
1 e4 ltJc6 2 lbf3 d6 3 d4 ltJf6 4 ltJc3 g6 5
i.e2 ..ig7 6 o-o o-o 7 h 3 a6!?
Against a lower-rated opponent
there is some concern that the natural
m ove, 7 ... es, will lead to a position with
too few winning chances, in which case
7 ... a6 is often a good alternative.
8 a4 e5
lt is nice for m e to have the bS­
square covered and his b4-square
weak. Therefore, if 9 dxes, Black will
play 9 ... dxes, keeping the extra pair of
knights on the board.
9 ..ie3?!
lt i s alm ost always a bad idea for
White to try to preserve the central
ten sion in the Dark Knight Pirc. This is
especially true when he has played h2-
h 3 . Black i s already at least equal.
9 ... exd4 10 ltJxd4 J:le8
One Pirc author recomm ends
10 ... .i.d7 first in this type of position,
but I prefer to have my c-pawns dou-
1 94
bled! The half-open b-file co-ordinates
well with the dark-squared bishop, and
it i s nice to have the ds-square securely
guarded.
11 ltJxc6 bxc6 12 ..if3 ltJd7 13 'il¥d2 .Ub8
14 .l:tab1 �f6 15 i.e2 li'es 16 l:!.fd1 'i!Va s
lt i s often deflating to computer­
check your own games. This queen
m anoeuvre, which I was proud of dur­
ing the g ame, accomplishes less than
nothing after 17 b4! l:.xb4?! 18 lt:lds !
.Uxa4 19 'i!Vxas .Uxas 2 0 lt:lxc7 l:!.xe4 2 1
l:i.xd6 with an edge for White.
17 i.d4 ..ixd4 18 'ifxd4 .l::t b4 19 'ifd2
ttJcs
Thi s will not get anything done ei­
ther. Black is not even threatening
20 ... ltJxa4 because of 2 1 1:ta1 and 2 2 b3.
20 b3 ltJxe4 21 ltJxe4 .l:!.exe4 22 .i.f3 J::.e s
2 3 ..ixc6 ..ifs 24 i.f3 'i!Vcs 2 5 c3 l:!.b8 26
b4 'il¥c4 2 7 .:!.b2 l:!.be8 28 'it'd4 'it'e6 29
�h2 .i.e4!
lt isn't much, but I will have a little
something to play for after stripping
White's king of its best defender. Who
knows? The e-file might even be worth
something.

30 .i.xe4 l:f.xe4 31 Wkd3 a s 3 2 'ilid s 'if'f6
33 1!i'xa s 1!i'xc3 34 .l:.bb1 cs?!
This is another move that seemed
very strong to me when I played it, but
White can equalize starting with 3 5
ft 6 ! (the text m ove is not bad, either).
34 .. Jie2 i s a better try. But by now we
were running low on tim e to reach
move 40.
35 �bs cxb4 36 .:txd6 l:!.8es 37 'ii' b 8+
�g7 38 .:td8 .:tfs 39 l::tg 8+ �h6 4o 'iWf8+
'lt>g s 41 �d8+ f6 42 l:te8 l:td4 43 "ille 7?
Houdini finds 43 f4+! :fxf4 44 l:teS+
with equality. White i s lost after any
other move.
43 ...l:Ixf2 44 :e1 lld3??
Illustrative Games
Threatening 4S ....l:!.xh 3+ 46 �g l
.l:!.xg 2+ and wins ! Somehow I over­
looked both of White's defences.
45 'ii'x h7
Or 45 h4+! 'it>xh4 46 'i!Vxh 7+ 'i.t>g s 47
.lhe4, when Black has to find
47 .. Jixg 2+! 48 �xg 2 l:td2+ 49 l::t e 2 lid4!
and draws (apparently).
4S ...l:!.dd2
Thi s i s the m ove that would have
won l ast turn.
46 .l:.g1??
Snatching defeat from the jaws of a
draw. My opponent saw the correct
continuation but misevaluated it: 46
h4+ �fs 47 l:t8es+ "il/xes (47 .. .fxes 48
"ilif7+ �g4 49 .l:!e4+ .l:!f4 so 1Wxg 6+ 'i.t>xh4
51 'ii'h 6+ draws) 48 I!.xes fxe s 49 'ii'f7+
We4 so 1!i'b7+ is no worse for White.
46 ...Wkc6 47 h4+ 'it>fs o-1
Not the most beautiful game, but
the opening was a success.
Gam e 75
N.J h u njhnuwala-S.Giigoric
Luce r n e O lym piad 1982
1 tt:lf3 g6 2 e 4 d6 3 d 4 .i.g7 4 tt:lc3 tt:lf6 5
.i.e2 0-0 6 0-0 tt:lc6 7 h3 es 8 dS tt:le7 9
.i.e3 c6!
White is unable to keep his centre
intact, nor is he fast enough on the d­
file to inconvenience Black.
10 dxc6 bxc6 11 'ii'd 2 Wkc7 12 .l:rad1 d s
13 exds tbexds 1 4 tt:lxds?!
Helpfully completing the opposing
centre. Black starts building his edge.
195
Th e Dark Kn ig h t System
14 ... cxds 15 c3 i.. b 7 16 i.. h 6 lL'le4 17
'iWc1 lUeS 18 i..xg 7 �xg 7 19 ltJd2 lL'lcs
20 lL'lb3 lL'le6 21 .l:tfe1 as
22 i.. b s .l:.e7?!
Black does not h ave a good reason
to disconnect his rooks. However,
White's next makes it easy to fix the
problem.
2 3 i..f1? [23 a4] 2 3 ... a4 24 lL'la1 l::t d 8 2 5
lL'lc2 f6 26 lL'lb4 1Wb6 2 7 1Wc2 ttJ c s 28 a3
l:.ed7 29 .l:te3 lL'lb3 30 .l:.ee1 :Ld6 31 1We2
d4 3 2 'iib s 'tWxbs 3 3 i..x bs dxc3 34
.l:.xd6 .:txd6 35 bxc3 ttJcs 36 .1:te2 .l:f.d1+
37 �h2 e4 38 �g3 fs 39 h4 'iW6 40
i.. e s? h6? [40 ...l:.h 1!) 41 i.. b s? [41 �h 2 ! )
4 1. .Jih1 42 f4 gs 43 hxgS+ hxgs 44
fxgS+?! 'ii?xg s 45 �f2 f4 46 i.. c 6? i..xc6
0-1
Since after 47 lt:'Jxc6 the reply
47 ... lL'ld3 is mate.
Gam e 76
G.Bastrikov-E.Gel ler
Ta s h kent 1958
1 e 4 es 2 lL'lf3 lL'lc6 3 lL'lc3 d6 4 h 3 g6 s
196
d4 exd4 6 ttJxd4 i.. g7 7 i.e3 lL'lf6 8 'ii'd 2
o-o 9 o-o-o .:tes 10 f3 ttJxd4 11 i.xd4
i.. e 6 12 i..f2 a6 13 �b1 bS 14 h4 cs 15
g4?
White can play the aggressive 15
h S ! ? or the defensive 15 a3 - the text
move is too slow.
15 ... b4 16 lL'le2?!
16 ...1Was?
Black g ets a big advantage with
16 ... ttJxe4! 17 fxe4 'ii'f6, threatening
m ate and White's bishop .
17 lL'lc1 i.xg4?! 18 i..g 2?
Black's sacrifice is thematic, but not
quite sound. 18 fxg4 lt:'Jxe4 19 'iWf4
lL'lc3+?! 20 bxc3 bxc3 2 1 i.. c4 is defence
and counterattack.
18 ... i.. e 6 19 'ii'x d6 .:!.ac8 20 lL'lb3 'Yi'a4 21
i.. h 3? i..f8?
Unnecessarily removing the bishop
from the m ain diagonal and misevalu­
ating the m ost direct continuation :
2 1 ... i.. xh 3 2 2 l:txh 3 c4 2 3 lL'lc5 'ifh5
threatens 24 .. J:tc6.
22 'ii'h 2?! i..x h3 23 'i!Vxh 3 c4 24 lL'lc1 c3
25 i.. d 4 cxb2 26 lL'lb3 i.. g 7 27 i.xb2 'tWc6
28 .l:.h2 J:.ed8 29 .l:thd�? .l:!.xd2 30 .l:.xd2

30 ... lDxe4! 3 1 fxe4 ..ixb2 32 h s [32
'iitxb2 ? ! 'tWc3+] 3 2 'tWc3 3 3 'ii'x c3 ..ixc3
•..
34 �d1 'it>g7 3 5 .l:!.d6 .l:!.c4 o-1
Game 7 7
D.Saduakassova­
Art.Minasian
D u ba i 2011
1 e4 lbc6 2 lDf3 d 6 3 lDc3 lDf6 4 d 4 g6 5 h 3
..ig7 6 ..ie3 o-o 7 'tWd2 e s 8 ds lDe7 9 ..ih6
White trades her best minor piece
for Black's obstructed bishop in the
hope of weakening his king position .
9 i.. d 7!
•••
Illustrative Games
Preparing ... c7-c6 and/or ...b7-bS.
10 i.. d 3 c6
10 ...bs 11 a3 as and 12 ...b4 is also
good.
11 dxc6 ..ixc6 12 .ixg 7 'it>xg 7 13 o-o
'iic 7 [13 ... b s ! ?] 14 �e2 l':[fe8 15 ..ibs a6
[1S ... d S ! ] 16 i..xc6 'i!Vxc6 17 l:tadl :ac8
18 lbd2?! ds 19 exd s lbexd s 20 lbxd s
lbxds 2 1 c4 lDf4 22 'ii'e4 fs 2 3 'ii'xc6
l:!.xc6 24 b3 tLle2+ 25 'iii> h 2 lDc3 26 Ua1
l:i.d6 2 7 lDb1 lDxb1 28 liaxb1 �d2
Rook endgames are notoriously
drawish, but according to Jesus de l a
Villa in 1 00 Endgames You Must Know,
this reputation i s unearned. With his
better rooks and king, Black has his
winning chances.
29 a4 'it>f6 30 'it>gl e4 31 l:!.fel lied8 32
'it>f1 'it>es 33 cs .l:.c2 34 liedl .l:f.d3 ! 35 b4
lla3 36 l:tdc1 l:f.xc1+ 37 Itxc1 .l::[x a4 38 c6
bxc6 39 lixc6 'it>ds 40 .l:!.c7 .l::tx b4 41
l:f.xh7 l:r.b6 42 h4 'it>c4 43 g3 as 44 .l:!.c7+
'it>bs 45 hs gxhs 46 .l:.f7 'it>c4 47 lixfs
l:r.a6 48 llxhs a4 49 :hs a 3 so .l::tc 8+
'ito>d4 51 .l:.d8+ 'iite s 52 !:!.d1 a2 53 .l:!.a1
'it>fs 54 'it>e2 l:ta3 55 'it>f1 'it>es 56 �e2
'it>d4 57 g4 'it>es ss 'it>d2 'iii>f4 59 'ito>c1
197
Th e Dark Kn ight System
'iii>xg4 60 'iii> b 2 :as o-1
Gam e 78
A.Sakha rov-A.Adorjan
Soch i 1976
1 e 4 d6 2 d4 ltJf6 3 lbc3 g6 4 lbf3 .i. g7 5
.i.e2 o-o 6 o-o ltJc6 7 .i.gs h6 8 .i.f4 ltJg4
9 h3 es 10 dxes ltJ gxes 11 ltJxes dxes
12 .i.e3 ltJd4 13 .i.c4 'i¥h4 14 lbd s?!
Not appreciating the danger. By re­
linquishing control over e4, White sub­
jects him self to a powerful attack. 14 f3
or 14 .i.d3 is still approximately equal,
but nobody likes to play such m oves.
14 ... c6! 1S lbc7 .i.xh3 !
Now White must defend accurately
to survive, and even then he will be
worse - there is no good way for him to
retain a material advantage.
16 lbxa8 'i¥xe4
The queen's arrival on this square is
a nightmare for White, who h as far too
m any bishops lying around.
17 gxh3 lbf3+ 18 'iii> h 1 lbh4+ 19 f3 'i!Vxe3
20 lbc7 'iif4 21 'ii'd 3 'ii'g 3 22 'ii'e 2 e4?!
198
[22 ... ltJfs !] 23 'ti'h2 'i¥xh2+ 24 <;i;>xh2
lbxf3+ 2 5 �h1 lbd2 26 lbe8?? [26 lbe6 ! ]
26 ...ltJxc4 27 ltJxg 7 Wxg7 28 l::tfe1 f s 2 9
.l:.ad1 1:1f7 30 b3 ttJes 31 lld6 lbf3 3 2
J:1ed1 .l:.e7 3 3 l:!. d 7 <3;;f7 34 <3;; g 2 f4 3 5
l:tld6 lbh4+ 36 �f1 e 3 3 7 c 4 e2+ 0-1
Game 79
Z.Brata nov-B.Chatal bashev
B u lga ri a n C h a m pio n s h i p,
Sofia 2004
1 e4 g 6 2 d4 .i. g 7 3 lbf3 d6 4 lbc3 ltJf6 5
.i.e2 0-0 6 0-0 ltJc6 7 .i.e3 es 8 dxes
dxes 9 'ilxd8 l::txd8 10 .i.c4
One might think that White plays
for a win with little risk. However,
White (a G M) also lost in M.Hebden­
E.Sutovsky, I sle of Man 1999. Black's
next i s to prevent the annoying 11
ttJg s.
10...h6 [1o... ltJe8 ! ?] 11 h 3 b6 12 ltJd s
ltJa s ! 13 lbxf6+ .i.xf6 14 .i.e2 .i.b7 15 b4
ltJc6 16 c3 a s 17 a3 lbe7 18 lbd2 .i.gs!
So far it's not much for Black, but
trading the worst minor for White's
 Illustrative Games

best is the first step on the road to a where 7 ... e s ! works a little better than
win. usual and 7 ... a6 works a little worse
19 .i.xgs hxgs 20 l:tad1 axb4 21 axb4 than usual.
'i.t> g 7 22 f3 l:!.a2 23 tt:'lc4 Z:.xd1 24 .i.xd1 8 dS! tt:'lb8 9 a4!
.i.a6 2 5 .i.b3 .i.xc4 26 .i.xa2 .i.xa2 27
.l:!.a1 .i.c4 28 l:i.a7 bs 29 .l:!.xc7 'it>f6
Perhaps White should hold this, but
he has a long and thankless defensive
task ahead of him. Black's task is to
penetrate with his king .
3 0 .l:!.a7 'it>e6 31 .l::t a 6+ W d 7 3 2 1:1a7+ Wd6
33 .l:!.a6+ tt:'lc6 34 <;t>f2 <;t>d7 35 Wg3 tt:'ld8
36 h4 gxh4+ 37 �xh4 �e7 38 �g3 tt:'le6
39 �f2 tt:'lf4 40 g3 tt:'le2 41 lla3 �f6 42
'lt>g 2 �gs
And now White is lost. How did that Since a2-a4 is normally m et by ... a7-
happen ? as, it stands to reason that Black's a­
43 'it>f2 fs 44 exfs gxfs 45 'iio> g 2 f4 46 pawn is misplaced.
gxf4+ �xf4 47 '>W2 tt:'lc1 48 .l::ta 8 tt:'ld3+ 9 c6
•••

49 <;t>g2 �e3 so <;t>g3 Wd2 51 l:!.a3 'iii>c 2 I n ow believe that 9 ... b6 is the best
52 <;t>g4 <;t>b2 53 .l:!.a7 �xc3 54 'it>fs Wxb4 move in the position, intending to fol­
0-1 low with ... c7-c6, ... .i.b7, ... tt:Jbd7. At the
time I was reluctant to try thi s since
White h ad not spent a tempo on h 2-h 3
Game Bo to prevent ... .i.g4.
K. Wang-J.Schuyler
Wa s h i ngton 2012

1 e 4 tt:'lc6 2 tt:'lf3 d6 3 d 4 tt:'lf6 4 tt:'lc3 g6 5

.i.e3 .i.g7 6 'ili'd2 o-o 7 .i.e2
White tries to play without h 2-h 3 -
an uncomm on plan. I was aware that
7 ... e s was the main m ove, but as in
Game 74, I wanted to spice things up. I
couldn't remember for sure, but 7 ... a6
is usually a reasonable option.
7 ... a6?! 10 as! tt:'lbd7 11 0-0 'ilic7 12 .l:tfe1 tt:Jcs
As it turns out, thi s i s one position Thi s picks up the bishop pair, but I

1 99
The Dark Kn ight System

will not be enjoying my pawn struc- 'it>xe2 1:.xg7 12 i.h6 .l:!.g8

ture.
13 i.xcs dxcs 14 dxc6 bxc6
Ugly as thi s is, it is worse to allow
White access to the dS-square.
15 i.c4 l:tb8?!
Instead, 1S .. .tLig4! frees the bishop
and control s White's e-pawn.
16 es tt:J g4 17 'ii'e 2 l:txb2
Losing a bit of m aterial, but it is no
worse than anything el se.

" Everybody knows" three pieces are

18 h3 tt:Jh6 19 i.b3 c4?! [19 ... tt:Jfs] 20
tt:Ja4 cxb3 21 tt:Jxb2 bxc2 22 't!Vxc2 i.e6
23 lLid3 i.d s 24 lt:Jd4 .l:r.d8? [24 ... e6] 25
'tics tt:Jfs 26 tt:Jxts gxfs 27 'i!Yb6 'iix b6 28
axb6 .l:!b8 29 .l:!xa6 i.c4 30 .l:!.a3?1 e6? 31
tt:Jcs i.f8 3 2 tt:Jd7 i.xa3 33 tt:Jxb8 i.d s
34 b7 cs 3 5 lt:Jc6 1-0
much better than a queen, but is this
true regardless of how m any pawns
the queen has? Regardless of structural
problem s? Regardless of king position ?
A s much fun a s these positions can be
to play, White is asking too much from
his minors.
13 h4
This is played mainly to keep the
bishop from getting cut off by an even­
tual ... g6-g S .

Came 8 1
R.Zelcic-M.Djurkovic
P u l a 2001

1 e 4 g6 2 d4 i.g7 3 tt:Jc3 d6 4 i.c4 tt:Jc6 5 13 ... c6!

lLif3 tt:Jf6 6 'ii'e 2 i.g4 7 es i..xf3 8 gxf3 An important move to control
tt:Jxd4 9 exf6 tt:Jxe2 10 fxg7 l::t g 8 11 White's minors, and uses those extra

200

pawns well.
14 J:.he1 'ii'd 7! 15 .U.ad1 'ii'fs
Activating the queen quickly is cer­
tainly the right idea, but 1S .. JWh 3 ! (16
.S.h 1 'iVfs) is m ore accurate.
16 i.gs o-o-o
17 �f1?
The queen is supposed to be h arder
to play, but not when the pieces h ave
an exposed king. lt is the GM who
makes the first big mistake.
11 ... ds 18 i.d3?! 'it'h3+?!
This check would have been m ore
useful to Black after grabbin g the f­
pawn .
19 �g1 f6 20 i.f1 'ii'xf3 21 l:i.d3 'ii'g4+
22 l:!.g 3 Vi'b4!
Black's queen is not done making
trouble.
23 a3?! �a s?
White's bluff works. Every white
pawn that disappears de-stabilizes
White's pieces m ore. They are running
short on support points, and they can't
be left lying around loose. H ence
23 ...�xb2 ! .
2 4 i.d2 es?! 2 5 tt:le4 Vi'b6? [2S . . .�c7] 2 6
Illustrative Games
tt:lxf6 l:Igf8 27 ltb3?
Just as White is getting back into
the game, he plays this awful interme­
diate m ove, forcing Black to improve
his queen position. White is not so
badly off after 27 l:.f3, threatening 2 8
i.h 3+ and 29 tt:ld7, while if 27 .. .'!J.f7
then 28 c4! �b8 (28 ....l:.df8 29 i.h 3+
�c7 30 cs!) 29 tt:lxds .l:Ixf3 30 tt:lxb6
l:1xd2 3 1 tt:la4 l:tdxf2 and Black is "only"
better.
27 ...�c7 28 .l:.f3?
White i s in a bad way after 2 8 tt:lg4
e4, but at least he doesn't lose more
material.
28 .. JU7??
28 .. .'ii'g 1 29 i.h 3+ �c7 3o i.g s l:.d6
wins.
29 i.h3+ �b8 30 .l:txes �xes 31 i.f4
'i*'xf4 3 2 .l:lxf4 .l:tdf8 3 3 tt:ld7+ �c7??
Yikes! When it rains, it pours.
33 ...�a8 ! i s still approximately even.
The text walks into 34 tt:lxf8 .l:.xf4 3 5
tbe6+.
34 tt:lxf8 1-0
Game 82
H.H ughes-K.Richardson
British League 2004
1 e4 g6 2 d4 i.g7 3 tbc3 d6 4 i.c4 tt:lc6 5
tt:lf3 tt:lf6 6 o-o o-o 7 i.e3 tt:lg4 8 i.f4?!
tt:lxd4! 9 tt:lxd4 es 10 i.e3 tt:lxe3 11 fxe3
exd4 12 exd4 i.e6 [12 ... c6!] 13 'ii'd 3
i.xc4 [13 ... cs] 14 �xc4 c6 15 �h1?!
�b6 16 l:.ad1 l:tad8?! [16 ...'i!Vxb2] 17 b3
'ii'a s 18 :f3 ds
201
The Dark Kn ight System

ti:Jd4 c5 11 .i.b5+ 'it>f8!

The king is fine here. Black keeps the
sen sitive fS-square under control.
12 tt:Jf5 .i.xf5 13 exf5 'ii'a 5 14 .i.e2 tt:Jxg3
15 hxg3 tbd7 16 o-o c4 17 .:!.fe1 .:!.c8

19 exd5 cxd 5 20 'iWc5 'i!Vxc5 2 1 dxc5 d4

22 tt:Je2 1Ife8 2 3 tt:Jc1 :c8 24 b4 b6 25
cxb6?!
lt is no good to open the c-file for
Black and allow him to press on the
weak c2-pawn. 2 5 tt:Jb3 bxcs 26 bxcs Mohr has no shortage of ways to
offers White better chances of a suc­ improve his position and has gained
cessful defence. the advantage.
2 5 ... axb6 26 .l:tf2 l:tc4 2 7 4:Jd3 It.ec8 28 18 .I::i. a b1 .l::t c 7 19 'it'e3 .i.xc3 20 bxc3 tt:Jf6
'it>g1?! [ 2 8 a4] 28 ...1:i.xc2 29 .litdd2? [ 29
a4] 29 .. Jixd2 30 l:txd2 .i.h6 3 1 l:tb2 l:tc3
32 tt:Je5 d3 33 tt:Jxd3 llxd3 34 'it>f1 .l:.d2
0-1

Gam e 83
A.Grilc-G.Moh r
Slove n ia n Tea m
C h a m pion s h i p 2008

1 e4 tt:Jc6 2 4:Jf3 d6 3 d4 tt:Jf6 4 4:Jc3 g6 5 21 .i.xc4?

.tg5 .i.g7 6 �d2
Thanks to thi s move, and the fact
that White i s giving Black ... h 7-h 6 for
free, the bishop n o longer has enough
squares to escape.
6 ... h6 1 .i.f4 g5 8 .i.g3 tt:Jh5 9 d5 tt:Jb8 10
White is understandably unhappy
with his position, but things are not yet
as desperate as this. There will be no
compensation for the piece.
21 ....l::!.xc4 22 �xe7+ Wg7 2 3 1i'e3 Wf8
24 �e7+ Wg7 2 5 l:!.xb7 .l:.f8 26 Wxd6

2 02

�xd s 27 �a6 l:f.xc3 28 g4 :xc2 29 llxa7
�d4 0-1
Game 84
D.Janowski-F. Yates
Marien bad 1 9 2 5
1 d 4 l'Lif6 2 l'Lif3 g6 3 l'Lic3 i.g7 4 e4 d 6 5
i.gs ltJc6 6 'ii'd 2 h6 7 i.e3 l'Lig4 8 o-o-o
If the white bishop tries to slip away
with 8 i.f4, Black equalizes immedi­
ately with 8 ... ll:ixd4!.
8...ltJxe3 9 �xe3 0-0 10 h3 a6
With the two players castled on op­
posite wings, the race is on.
11 g4 bs 12 I:!.g1 ll:ia s 13 i.d3 c6 14
ll:ie2?!
White i s "racing" a little slowly. 14
e5 or 14 h4 is better.
14 ...lLJc4 15 Ji.xc4 bxc4 16 �C3 aS 17 eS
'it'b6 18 exd6 exd6 19 l'Li g 3
White is much worse, with no bish­
ops and a weaker attack.
19 ... cs?! [19 ... l:.b8] 20 ll:ifs ! cxd4 21
l'Li3xd4 i.xfs 22 gxfs d s 2 3 k!.g2?! l:taes
24 f3 .l:.e7?!
Illustrative Games
lt i s important to prevent 2 5 'i!Vd2,
which allows White to escape the un­
pleasant pin and reorganize his de­
fence. H ence 24 ... a4! 2 5 a3 l:i.b8, threat­
ening 26 ....l:!.fe8 and 27 .. J::t e 3.
2 5 'it'd2 a4 26 a3 :bs 27 c3 lleb7 28
fxg6 fxg 6 29 �c2 gS 30 'ii'x a4 'ii'c s 31
l:Idd2 �f8 32 �a s �h8 3 3 �xd s?
'iixa3 ! 34 'fie4 �al+ 3 5 �bl
What's the best move in thi s posi­
tion ?
3 s .. JIVa4?
Yates overlooks the problem-like
35 .. .'ii'a 8 ! !, which sets White the unen­
viable task of defending against ....l:.a7-
a1. (In fact Yates didn't play the second­
best move either - 36 ... �a5.)
36 l:!.ge2 �a s 37 l:.e3 i.f8 38 �d1 �c7
39 �fs .l:r.a7 40 ttJbs l::i. a l+ 41 �c2 �h7
42 iVxh7+ �xh7 43 ll:id6 i.xd6 44 l:txd6
l:ta2 45 l:.e7+ �g8 46 <lt>d2 l:taxb2+ 47
�e3 J::i.c 2 48 l:tdd7 J::i.x c3+ 49 �e4 'it>f8 so
llf7+ �g8 51 l:!.g7+ �f8 52 f4 gxf4 53
.l:tdf7+ �e8 54 l:tc7 l:rg3 55 .l:.h7 �f8 56
l:.h8+ l:tg8 57 l:txh6 lieS+ ss 'it>fs l:tg7
59 llh8+ l:!.g8 60 .l:.hh7 .l:.gl 61 .l:.h8+
Yz -Yz
203
The Dark Kn ig h t System
Game 85
A.Mista-M.Szelag
Kosza li n 1999
1 e4 d6 2 d4 lt:lf6 3 lt:lc3 e5 4 lt:lge2 i..e 7
5 f3 o-o 6 i.. e 3 exd4 7 lt:lxd4 lt:lc6 8 'ilfd2
lt:lxd4 9 i..x d4 i.. e 6 10 o-o-o c6 11 g4 bs
12 b3 lt:ld7 13 .l:.g1 b4 14 lt:le2 cs 15
i.. b 2
1s ...lt:les?!
Initiating complication s that should
not work out well for Black. 1 S ...l:!.c8 ! ?
was better, with ideas of ... cS-c4.
16 it'e3 it'a s 17 f4 c4 18 fxes c3 19
lt:lxc3 bxc3 20 i..x c3 'ii'xa2 21 exd6?!
2 1 'iit d 2 ! it'a3 (or 2 1 ... i..xb3 2 2 i.. d 3
i..x c2 23 .U.a1) 22 :tal it'cs 23 'ii'x cs
dxcs 23 'it>e3 is good for White.
21 ...i..x d6 22 i..x g7 i.. a 3+
2 2 ... .l:r.fc8! was stronger.
23 'iitd 2 i.. b4+ 24 'it>c1??
Better is 24 i.. c 3 .l:.fd8+ 25 Wel
l:txdl+ 26 'iitx dl .l:Id8+ 27 'iit e 1 when
White can still defend.
24...f6 25 i..xf6 l:txf6 26 'ii'd 4 �f2 2 7
�b2 i.. a 3 0-1
204
Game 8 6
I.Jakic-Z.Mestrovic
Zad a r 2001
1 e4 lt:lc6 2 d4 d6 3 lt:lc3 lt:lf6 4 i.. g 5 h6 5
i.h4 g5 6 i.g3 i.. g 7 7 h4 g4 8 hS e5 9 dS
lt:ld4 10 f4?! exf4 11 i.. h 4?! cs 12 'ii'd 2?!
White's gambit was a poor choice,
as in thi s position he h as no real com­
pensation for the pawn he has sacri­
ficed.
12 ..Ji'b6 13 o-o-o o-o 14 'ii'xf4 lt:lxh s 15
ii'd2 i..d 7?
lS .. .fs ! is just good for Black,
whereas now White gains time for his
attack.
16 i..f2 lt:lf6 17 i.. d 3 .l:tfe8 18 i..e 3 hs 19
i.h6 lies 20 l:tf1 l:IaeS 21 'ii'f4 'i!Vd8 22
lt:lge2 lt:lxe2+ 2 3 lt:lxe2 c4 24 lt:lg3 cxd3
2 5 i..x g7 'iitx g7 26 lLlxh S+??
Missing the much stronger 26
.l:txh S ! �xh s 2 7 iVxf6+! 'ii'xf6 28 lt:lxh s+
�g6 29 lt:\xf6, when White comes out
on top.
26 .. Jixhs 27 .l:t.xh 5 lt:lxh 5 28 'ifxf7+ 'it>h6
29 cxd3 'ii'e 7 30 'ii'f2 i.. b s 0-1

Game 8 7
R.Zelcic-Z.Mestrovic
Nova Gorica 2003
1 e4 tbc6 2 d4 d6 3 tbc3 tDf6 4 ..tg s h6 5
i.h4 gS 6 i.g 3 i.g7 7 f3 o-o 8 i.f2 eS 9
dS tbd4
White has been getting very little
done, so it is easy to justify Black's ag­
gression.
10 tbge2 cs 11 dxc6 bxc6 12 tbxd4 exd4
13 i.xd4 cs?! [13 ... .:.b8 ! ] 14 .txf6?!
'i¥xf6
When will they ever learn ? From
now on White has a severe weak colour
complex on the dark squares, and
Mestrovic's play from thi s point on is
impeccable.
15 .l:tb1 .ie6 16 i.e2 :labS 17 o-o 'iif4
18 �h1 i.xc3 19 bxc3 .l:.xb1 20 'ii'x b1
'ii'e 3 21 .id3 c4 22 .:te1 1i'd2 23 .tf1
'it'xc3 24 h3 .:tds 25 'it'c1 .:tbs 26 .:te3
'ii'e s 27 l:a3 .l:i.b2 28 'it'd1? as 29 'ii'c 1 c3
30 .l:.a4 'ii'f4 31 'ii'xf4 gxf4 3 2 i.c4 .l:txc2
33 .td s i.xds 34 exds .l:.b2 3 5 .l:.c4 c2
36 �h2 :xa2 0-1
Illustrative Games
Game BB
L.Gofshtein-N.Mitkov
Lis bo n 1999
1 c4 es 2 tDc3 tbc6 3 g3 fs 4 i.g2 tbf6 5
e3 d S 6 tbxds tbxds 1 cxds tbb4 8 d 3
tbxd s 9 tDf3 i.d6 1 0 o-o tDf6 11 'ii'b 3
'iie 7 12 e4 fxe4 13 dxe4 'iif7 14 'it'c3
'iih s 15 tbd2 o-o 16 f3 ..th3 17 tDb3 .l:.f7
18 i.e3 l:'taf8 19 'it'c4
19 ... h6?!
After this, White has an edge. lt is
simpler and better to keep the knight
out with 19 ... b6.
20 tDa s i.xg 2 21 �xg2 gs?l [2 1...b6] 22
h3 1i'g6? 23 tbxb7 hs 24 .:tad1 g4 2 5
tt'lxd6 cxd6 2 6 .l:.xd6 1-0
Black h as no real counterplay.
Punishing the doubled c-pawns
Game 89
J.Ramirez-J.Sch uyler
La s Vega s 2007
1 c4 tt'lc6 2 tt'lc3 es 3 g 3 fs 4 i.g2 tt'lf6 5
2 05
Th e Dark Kn ig h t System
d3 i.b4 6 a3?
White greatly overestimates the
value of the bishop pair and h alf-open
b-file, spending a tempo to reach a po­
sition Black would happily play with a
tempo less.
6...i.xc3+ 1 bxc3 d6 8 l:b1 o-o 9 e3?
White's previous sin s are minor
relative to thi s positional catastrophe
(which is nonetheless an extremely
common mistake among amateurs).
White's pawn structure will self­
destruct in 5 .. .4 ... 3 ... 2 ... 1...
9 ...e4! 10 d4 b6! 11 tt:Je2 i.a6 12 lt:Jf4
i.xc4
13 d5?
206
A hopeless try at complications.
13 tt:Jes 14 tt:Je6 'fie7 15 tt:Jxf8 lt:Jd3+ 16
•..
'itd2 'i!Vxf8 17 h 3 lt:Jxf2 18 'fia4 b5 19
'ii'a 5 tt:Jxh1 20 i.xh1 'fif7 21 .:tb2 'fih5
0-1
White's play was poor, but not un­
common for a club player. That was
how to win.
Game 90
O.Foisor-J.M.Degraeve
Le Tou q uet 1996
1 c 4 e5 2 g3 tt:Jc6 3 i.g2 f S 4 lt:Jc3 tt:Jf6 5
d3 i.b4 6 i.d2 o-o 7 e3 i.xc3 8 i.xc3
d5! 9 tt:Je2 i.e6 10 b3 'fie7 11 a3 l:tad8
Simple chess has brought Black a
small advantage, but thi s will dissipate
unless he takes action now.
12 'iii'c 2 i.f7?!
12 ... dxc4 13 dxc4 lt:Je4! was better.
13 o-o i.h5?1
Thi s bishop transfer, typical of the
Stonewall Dutch, makes far less sense
without a closed centre.
14 l:tfe1 'fif7 15 i.b2 'ith8 16 :ac1 l1d6
 Illustrative Games

17 b4 dxc4 18 dxc4 l::tfd8 19 i.xc6?

Black's pieces are already poised to
jump into White's holes after thi s ill­
conceived trade.
19 .. Jbc6?
After 19 ... bxc6 ! White doesn't even
get the e-pawn for his trouble (20
.txes ? l::td 2).
20 .txes lt:Jg4 21 l:led1 .l:i.e8 22 .tf4
lt:Jxh2!? 2 3 Wxh2 i.f3 24 g4 fxg4 2 5
lt:Jg3 'i¥f6
Black i s already better.
15 dxc6 i.xc6 16 e4 l:f.c8 17 i.e3 i.xe4
18 i.xe4?!

26 'ilt'd3??
26 'it>g 1 ! �4 27 Wfl holds on . The
text just loses.
26 ... g5 27 'ii'd 4 gxf4 0-1 18 ...lt:Jxe4?
After 18 .. .fxe4, Black keeps his
pawn, and with it a large advantage.
Came 9 1 19 lt:Jxe4 fxe4 20 lt:Jc3 i.b8 21 ctJxe4 iVc6
J.lruzubieta 22 lt:Jd2 ct:Jfs 23 1We2 i.c7 24 bs 'ii'd 7 25
Villal uenga-B.Gu l ko .txa7 b6? 26 as bxas 27 b6 ctJd4 28 �e4
Sa n Sebasti a n 1996 lt:Jc6 29 bxc7?! lt:Jxa7 30 l:;Ixas lt:Jc6 Yz-Yz

1 c 4 es 2 lt:Jc3 lt:Jc6 3 g3 f s 4 ..tg2 lt:Jf6 5 Cam e 92

d3 i.b4 6 ..td2 o-o 7 e3 ctJe7 8 lt:Jge2?! N.Spiridonov-K.Spraggett
8 a3 or 8 ctJf3 is preferable. Ca n nes 1992
8 ... c6 9 o-o d6 10 d4 �h8 11 a 3 .tas 12
b4 i.c7 13 d s i.d7 14 a4 'i!Ve8 1 c4 es 2 g3 lt:Jc6 3 .tg 2 f s 4 d3 lt:Jf6 s

207
The Dark Kn ight System

lt:'lc3 i.. b4 6 i.. d 2 o-o 7 e3 f4 dxes 1 4 'ii'g s, and there i s no attack.

13 fxe5 dxe5 14 d4 i.. h 3

Thi s i s not the best, but it i s aggres­

sive and sound. If 8 exf4 exf4 9 i..xf4,
Black gets good play with 9 ... ds ! .
8 lt:'lf3 �e8 9 o-o i..x c3 1 0 i..x c3 d6 11
exf4 exf4 12 d4 'ii'h 5 13 d5 lt:'le7 14 l2Jd4
i.. g4 15 f3 i.. d 7 16 l:.e1 lt:'lg6 17 lt:'le6
i..xe6 18 .I:.xe6 fxg3 19 hxg3 'ili'g 5 20
'ili'e1 h5 21 Wh2 h4 22 gxh4 lt:'lxh4 2 3
'iii'g 3 'ii'h 6 24 wg1 lt:'lf5 2 5 'ifh3 'ii'g 5 2 6
l:t.ae1 l:tf7 2 7 .U1e2 l:.af8 28 i.. d 2 'ii'g6 2 9
Wh2?! [29 :2e4] 29 ...lt:'ld4 30 i.. c 3? [30
l:t2e4] 30 ...lt:'lxe2 3 1 l:.xe2 'iii'd 3 3 2 .l:!.d2
'ii'h 7 3 3 i..d 4 b6 34 b4 lie7 35 c5? lt:'lh5
36 cxd6 cxd6 37 '>t>g1 �b1+ 3 8 i..f1 lt:'lf4
39 'ii'h 2 llf5 o-1
15 dxe5?!
Black's aggression usually pays
dividends, but it is not without risk.
Even at thi s late stage, White can es­
cape to a good position with 15 i.. xh 3
'ii'xh 3 16 lt:'lxes lt:'lxes 17 dxes lt:'lg4 (or
17 ... .l:tad8 18 'i¥f4 and 19 'ii'g 3) 18 �dS+
.laf7 19 'ifg 2 .
15 ....l:!.ad8 16 'ii'f4 i..xg2 17 Wxg2?! [17
exf6 !] 11 ...l2Jd 5

Gam e 93
F Bruno
.
-
B . Kurajica
Luga n o 1985

1 c 4 e5 2 lt:'lc3 lt:'lc6 3 g3 f5 4 d 3 lt:'lf6 5

.i.g2 i.. b4 6 i..d 2 o-o 7 lt:'lf3 d6 [7 ... e4!] 8 18 'ii'c 1??
a3 i..x c3 9 i..x c3 'ili'e8 10 o-o 'ii'h 5 11 White is worse now, but he survives
ii'd2 f4 12 gxf4 h6 after 18 1i'h4.
Otherwise White will play 13 fxe s 18 ... lt:'lf4+ 0-1

208

Came 94
M.Sher-K.Spraggett
Andorra 1993
1 C 4 eS 2 g3 l2Jc6 3 i.g2 f S 4 l2Jc3 l2Jf6 5
d3 i.b4 6 i.d2 o-o 7 lL'lf3 d6 8 o-o i.xc3
9 i.xc3 'ii'e 8 10 e3 i.d7
Black, who is down a tempo on the
previous game, is not ready for 10 .. .f4? ! .
Instead, 1 0. . .e 4 11 dxe4 l2Jxe4 i s fine for
Black, but this is not why a player like
Spraggett plays the reversed Grand Prix
Attack.
11 �cl 'it>h8 12 b4 lt:Jd8 13 bS .l:.b8 14 a4
lt:Je6 1S lt:Jd2 f4! 16 exf4 exf4 17 i.xf6
Removing one of Black's m ost dan­
gerous attacking pieces, but pulling his
rook into the action. Besides, that was a
very good bishop !
11 ... l:txf6 18 lt:Je4 �h6 19 d4 'ii'g 6 20
.l:!.e1 .l:!.f8 21 .l:.c3 b6 22 .l:f.d3 i.c8 2 3 as
�hs 24 i.f3 l:!.h6 2 5 axb6 axb6
According to Houdini, White h as
been at least a little better the whole
time, but that doesn't take into consid­
eration the difficulty and fatigue of
Illustrative Games
defence. White makes a move he cer­
tainly wouldn't have made if he were
fresh - but he isn't fresh, and that is no
accident.
26 't!Ve2?? fxg 3 27 fxg 3 lt:Jf4 28 'ii'd 2
lt:Jxd3 29 'ii'xd3 i.g4 30 i.g2 'ii'h s 31 h4
.l:!.e6 3 2 l:!.f1 .!:!.xf1+ 3 3 ft'xf1 fifs 34 'iVa1
.l:!.e8 3 5 'ii'a 7 i.f3 0-1
Cam e 95
Bu Xiangzhi-V.Ivanchuk
FIDE World C u p,
K h a nty-Ma n s iys k 2011
1 lt:Jf3 d s 2 g3 g6 3 i.g2 i.g7 4 o-o es s
d3 lbc6 6 c4 dxc4 7 dxc4 'ili'xd1 8 .l:[xd1
e4 9 lt:Jfd2?! [9 ttJg s] 9 ... fs 10 lt:Jc3 i.e6
11 tt:Jd s? o-o-o
Thi s is just awful for White, who
can't activate any of his pieces. He soon
pitches a pawn to free himself, but
there is no compensation for this sacri­
fice.
12 lt:Jb3 lbf6 13 i.gs i.xds 14 cxds
.l:!.xds 15 f3 exf3 16 .l:txds lt:Jxds 17 i.xf3
lt:Jdb4 18 'it>f1 l:te8 19 tt:Jcs i.d4 20 a 3
2 09
Th e Dark Kn ight System
i.xcs 21 axb4 tLlxb4 22 g4 tLlc2 2 3 .l::ta s
i.e3 24 gxfs i.xgs 2 S fxg6 lLle3+ 26
�g1 h6 2 7 l:txa7 c6 28 g7 <J;; c 7 29 .l:ta4
.l:. g8 0-1
Gam e 9 6
K.Arakhamia G rant­
A.Raetsky
Bern 1 9 9 5
1 lZJf3 d s 2 g3 g6 3 i.g2 i.g7 4 o-o e s s
d3 lLJc6 6 lZJbd2 lLJge7 7 e4 0-0 8 C3 aS 9
a4 h6 10 l:le1 i.e6 11 exds i.xd s !
Thi s prevents the active 12 tLlc4 be­
cause of 12 ... e4, exploiting the vulner­
able situation of the white knight at c4.
12 'i¥c2 fs 13 b3 'ii'd 7 14 i.a3 lUes 1S
.l:.ad1 gS 16 l2:lc4 tLlg6 17 l2:le3 i.f7 18 d4
e4 19 l2:ld2 h S 20 f3 f4 21 tLlec4 fxg 3 22
hxg3 exf3 23 lLlxf3 'ii'g4 24 lLle3 J::tx e3
2S .l:.xe3 �xg3 26 'i¥f2 Vi'xf2+ 27 <J;;xf2
g4 28 tLlgs i.h6 29 lbxf7 i.xe3+ 30
'it>xe3 'iii>xf7 31 l:.f1+ Wg7 32 i.ds J::te 8+
33 Wd 3 tLld8 34 l:lfs c6 3S ..li.e4 lLle6 36
c4?? lLlef4+ 37 �e3 tLlg2+ 38 rJi;d3 lLle1+
39 We3 lLlc2+ 40 �d3 tLlxa3 0-1
210
Game 9 7
V.Frias Pa blaza-A.Baburin
Sa n Fra ncisco 1 9 9 7
1 lLlf3 ds 2 d3 g 6 3 g3 i. g 7 4 i.g2 es s
o-o lbe7 6 e4 o-o 7 tLlbd2 tLlbc6 8 c3 a s
9 a 4 h6 1 0 exd s tLlxd s 11 lbc4 ..tfs 12
.l:.e1 :es 13 lLlh4 i.e6 14 i.d2 'ii'd 7 1S
�b3 tLlde7 16 i.f1 .l:.ad8 17 .l:.ad1 b6 18
i.c1 gs 19 lbg2 i.g4 20 i.e2 i.xe2 21
.l:.xe2 lLlfs 22 .l:.ee1??
Black is already better, but thi s al­
lows a winning shot as Black quickly
exploits the weakened White king side.
22 ...tLlfd4! 2 3 cxd4 tLlxd4 24 'ii'a 2 lLlf3+
2 S �f1 [25 �h l ? 'iVh3] 2 S ... lLlxh2+ 26
�g1 lLlf3+ 2 7 �f1 'ii'h 3 28 tLlce3 tLld4 29
'iii>g 1 .l:.e6 30 b3 .l:.ed6 31 i.b2 tZ'lf3+ 32
�f1 e4 33 dxe4 l:td2 34 l:!.xd2 :xd2 0-1
Game 98
A.Capaliku-J.Gombac
N ova Gorica 2010
1 f4 g6 2 tLlf3 i.g7 3 e3 d6 4 d4 tLld7 5
 Illustrative Games

il..d 3 es 6 c3 'W/e7 1 e4 lDgf6 8 fxes dxes ..tc4 e6

9 o-o o-o 10 il.. g s h6 11 il.. h 4 cs 12 lt i s m ost important to blunt the
ttJbd2?1 bishop.
12 ds c4 13 il.. c 2 'ii'd 6 i s best, though 6 o-o ttJe7 1 lDc3 o-o 8 i.b3 cs 9 'it'e2 d s
this is not a problem for Black.
12 ... cxd4 13 cxd4 exd4 14 ttJxd4?1 ttJes
1S il.. c 2 .Ud8 16 ltJ4b3?1 [16 lD2b3]
16 ... il..e 6?1
White is in trouble, but 16 ... as! is
stronger; e.g. 17 a4 b6 and 18 .....ta6.
17 'ife2 l:tac8 18 llac1 ..tg4 19 'iff2 gS
20 ..tg3 ttJhs

White's light-squared bishop i s now

both passive and in danger.
10 a4 b6 11 l:td1 ..tb7 12 il.. d 2 a6 13
.i.e1 ttJfs 14 il..f2 'it'c7

I sense a dark-square catastrophe

on the horizon for White.
21 'it>h1 ttJxg3+ 22 'ili'xg3 as! 23 lDf3 ..txf3
24 gxf3 ttJg6 2S ttJxas .l:!.d2 26 ..tb3 .l:!.xc1
27 .l:.xc1 ..tes 28 Vi'e1 I;lxh2+ 29 �g1 ltJh4
30 'i¥e3 il..d4 31 lieS+ �h7 32 .:.hS+ �g6
33 .l:.gB+ �hs 34 'ifxd4 lDxf3+ o-1
1S g3?1 tiJd6 16 h4 C4 17 ..ta2 bS 18 h S
b 4 1 9 ltJ b 1 ltJb6?1 [ 1 9. . .a s ! ] 20 hxg6
Game 99 fxg6 21 aS ltJa4 22 C3 bxc3?J 23 ltJxc3
A.Spichkin-D.Reinderman ttJxc3 24 bxc3 l:tab8 2 S .i.b1
E u ro pea n C h a m pion s h i p, White, who was practically lost, i s
R ij e ka 2010 now back in the game.
2 s .....tcs 26 ttJgs?l ..tf6 27 lDf3 l:lbs 28
1 f4 g6 2 lDf3 il.. g 7 3 e3 d6 4 d4 ltJd7 S 'i¥a2 il.. d S 29 ..tc2 llxa s 30 Vi'b1 l:Ixa1 3 1

211
The Dark Kn ight System

'ii'xa1 ttJbs 3 2 i..a 4 i..e 7 33 i.. e 1 i.. d 7 34
ttJes i..e 8 35 l:.c1 i..d 6 36 tiJf3 h6 37
i..c 2 tiJa3 0-1
Did White's flag fall ? Black hasn't
made any progress on the board since
winning the a-pawn.
Black needs t o start using the c-file
as soon as possible. His next few m oves
do not work towards that, and he starts
drifting.
14....i.a6?! 15 i.. e 3 ttJcs 16 .U.c1 'ii'd 7 17
tiJe1 tiJc6 18 'iVd2 i..d 4?! 19 .txd4 ttJxd4

Gam e 1 00
P .Auchenberg­
To.Christensen
H e l s i ngor 1997

1 f4 g 6 2 tiJf3 i..g 7 3 g3 b 6 4 i.. g 2 .i.b7 5

o-o e6 6 d 3 d6 7 e4 tDe7 8 �e2 tiJd7 9
tiJbd2 0-0 10 g4 cs
Black h as actually gained the ad­
vantage with his simple development
scheme. White's plan to attack on the
king side is slow.
11 �f2 'ilic7 12 tiJf1 c4 13 tiJg3 cxd3 14
cxd3
20 'ii' b4? [2o fS ! ] 20 ... es 21 i..f1? 'iixg4
22 fxes dxes 23 .l:r.g2 'ii'f4 24 'ii'd 2 'ifxd2
2 5 .l:!.xd2 llac8 26 llcd1 ttJce6 2 7 i.. h 3 fS
28 exfs gxfs 29 'itf2 ttJf4 30 .i.f1 h s 31
h4 'ith8 3 2 tiJg2? tiJde6? [32 ... tiJh 3 + ! ] 33
tiJe3 tiJg7 34 i..e 2 l:tcd8 3 5 i..f1 :d4 36
ttJc2 .:.d6 3 7 tiJb4 i.b7 38 tiJc2 .:tfd8 39
tiJb4 .:td4 40 tiJc2 Il4d6 41 tiJb4 a s 42
tiJc2 tiJg6 0-1

Adieu
With this, I bid my readers adieu. I hope
this book was more fun for you to read
than it was for me to write. Har! I wish
you many successes with 1...tiJc6 and
the Dark Knight System .

212
 I nd ex of Va riatio n s

Chapter One
1 d4 t2Jc6 2 t2Jf3 (others - 15) 2 ... d6

A: 3 c4 g6 4 d5 t2Jb8 5 t2Jc3 i.g 7 6 e4 t2Jf6 7 i.e2 o-o - 19

A1: 8 h 3 - 20
A2: 8 0-0 - 2 1
8 : 3 d5 t2Je5 4 t2Jxe5 dxe 5 5 e 4 t2Jf6 - 2 2
81: 6 t2Jc3 - 2 3
8 2 : 6 i.b5+ - 2 4
C: 3 i.f4 t2Jf6 4 e3 g6 5 i.e2 i.g7 - 2 6

213
Th e Dark Kn ight System

Cl: 6 0-0 - 27
(2: 6 h 3 - 2 8
D: 3 g 3 g 6

D l : 4 ds - 3o
D2: 4 .i.g 2 - 3 1
E : 3 ..tg s - 3 3

Chapter Two
1 d4 ltJc6 2 c4 es 3 ds ..tb4+

214
 Index of Variations

A: 4 i.d2 i.xd2+ 5 'ii'x d2 ti'Jce7 - 3 6

A l : 6 d6 - 3 6
A2: 6 ti'J c 3 - 3 8
B: 4 ti'Jd2 - 40

Chapter Three
1 d4 ti'Jc6 2 ds ti'Jes

A: 3 e4 - 43
B: 3 f4 - 46

215
Th e Dark Kn ight System

Chapter Four
1 e4 l2Jc6 2 d4 (others - 5 5) 2 ... e5

A: 3 d5 l2Jce7 - 57
Al: 4 l2Jf3 l2Jg 6 - 5 8

All: 5 h 4 ! - 59
A12: 5 .te3 - 6 1
A 2 : 4 C4? ! l2Jg6 - 62

A21: 5 lLlc3 - 62

216
 Index of Variations

A22: 5 �e3 - 64
A2 3 : 5 tLlf3 - 67
A24: 5 �d3 - 68
A2 5 : 5 g 3 - 70
A26: 5 a3 - 71
B: 3 dxe5 t2Jxe5 - 73

B1: 4 tLlf3 tL::l xf3+ 5 'ii'xf3 tL::lf6 - 73

B11: 6 tL::l c 3 - 74
812: 6 �c4 - 74
B13: 6 e 5 - 75
814: 6 �d3 - 75
B15: 6 �e2 - 75
B16: 6 �9 5 - 76
B2: 4 f4 tL::l c 6 - 76
821: 5 �C4 - 77
B22: 5 tLlf3 - 78

21 7
Th e Dark Knig h t System

C: 3 tLlf3 exd4 - 79

Cl: 4 tt:Jxd4 ..tc5 - 80

C11: 5 ..te3 'ikf6 6 c3 'it'g 6 - 81

(111: 7 t2Jd2 - 82
c112: 7 tLlb5 - 84
(12: 5 tt:Jxc6 - 8 5
C 2 : 4 ..tc4 tt:Jf6 - 8 8

(21: 5 0-0 - 8 8

218
 Index of Variations

C22: 5 e 5 tt:\g4 - 89
C221: 6 'ii'e 2 - 89
(222: 6 0-0 - 90

Chapter Five
1 e4 tt:\c6 2 tt:\f3 d6 3 d4 tt:\f6 4 tt:\c3 g 6

A : 5 i.b5 - 94
B: 5 d5 tt:\b8 6 i.e2 i.g 7 7 o-o o-o - 96

Bl: 8 h3 e5 9 dxe6 .i.xe6 - 9 7

B l l : 10 .i.g 5 - 98
B12: 10 tt:\d4 - 99
B2: Others - 99
C: White doesn't play d4-d5 - 102
Cl: 5 h3 - 102
C2: 5 i.e2 - 106
C 3 : 5 i.e3 - 108
C4: 5 .i.c4 .i.g7 - 109
(41: 6 0-0 - 110

219
The Dark Kn ig h t System

C42: 6 'it'e2 - 110

C43 : 6 .te3 - 112
C S : 5 .tg s - 112

Chapter Six
1 e4 t'Llc6 2 t'Llc3 t'Llf6 3 d4 d6

A: 4 f4 - 115
8: 4 d5 - 117

Chapter Seven
1 C4 t2Jc6 2 t2Jc3 e S

A : 3 g 3 - 121
8: 3 t'Llf3 fs 4 d4 e4 - 123
81: s t'Llg s - 124
82: 5 t'Lld2 - 1 2 5

Chapter Eight
1 t'Llf3 t'Llc6 - 127

220
 In dex of Varia tions

Chapter Nine
Oth ers - 1 3 1
1 b 3 - 129
1 b4 - 1 3 0
1 f4 - 1 3 0

221
 I n dex of Ga mes

Ag hasaryan.R-Chibukhchian.A, Kaj a ra n 2011 149

.................................................................

Altounian.L-Schuyler.J, La s Vega s 2008 139

.............................................................................

Arakhamia Grant.K-Raetsky.A, Bern 1995 210

........................................................................

Arlandi.E-Lanzani.M, Sa n Marino 1998 152

..............................................................................

Auchenberg.P-Christensen.To, Helsi ngor 1997 212

...............................................................

Balashov.Y-Kuzmin.G, USSR C h a m pions h i p, Vi l n i u s 1980 138

...........................................

Baluta.C-Cioara.A, Bucha rest 1996 1 71

.....................................................................................

Baramidze.D-Griezne.E, B a u n ata 1 1999 1 64

.............................................................................

Barle.J-Mestrovic.Z, Slove n i a n C h a m pion s h i p, Krs ko 1997 145

..........................................

Bastrikov.G-Geller.E, Tas h ke nt 1958 196

..................................................................................

Becerra Rivero.J-Miles.A, Andorra 1995 1 84

.............................................................................

Beliavsky.A-Miles.A, E u ropea n C h a m pionship, Sa i nt Vincent 2000 158

..........................

Bonin.J-Schuyler.J, New York 1988 140

......................................................................................

Bontempi.P-Jovanic.O, Nova Gorica 2008 1 77

.........................................................................

Bratanov.Z-Chatalbashev.B, B u lga ria n C h a m pio n s h i p, Sofia 2004 198

...........................

Brudno.S-Benjamin.J, Boston 2001 166

.....................................................................................

Bruno.F-Kurajica.B, Luga no 1985 208

.........................................................................................

Bu Xiangzhi-Christiansen.L, Deizisa u 2000 1 72

.......................................................................

Bu Xiangzhi-lvanchuk.V, F I D E World C u p, Kha nty-Ma nsiys k 2011.. .......................... 209

Buehi.W.M-Benjamin.J, Reno 1999 1 70
.....................................................................................

Campora.D-Miles.A, Sevi l l e 1993 159

.........................................................................................

Campora.D-Tkachiev.V, Biel 1995 ....................................................................................... 1 79

Capaliku.A-Gombac.J, Nova G orica 2010 210
..........................................................................

Chatalbashev.B-Popchev.M, Caca k 1991 ........................................................................... 191

Christiansen.L-Benjamin.J, US C h a m pion s h i p, Seattle 2000 1 61
........................................

Drasko.M-Mestrovic.Z, Bos n i a n Tea m C h a m pion s h i p 2003 154

........................................

Erdos.V-Rapport.R, H u nga ria n Tea m Cha m pionsh i p 2012 155

...........................................

Fedorchuk.S-Miles.A, E u ropea n C h a m pion s h i p, Ohrid 2001 1 75

.......................................

Fischer.R-Schuyler.J, Richmond 2008 138

..................................................................................

222
 Index of Games

Foisor.O-Degraeve.J.M, Le Touq uet 1996 .

............................ . . 206
....................... ..... ...............

Frias Pablaza.V-Baburin.A, San Fra ncisco 1997 . . .. ............. . .... 210

.. .................. .... .................

Galliamova.A-Krasenkow.M, Kosza l i n 1997 . . .. . . . ... .. .... ..... ... 148

... ..........................................

Gofshtein.L-Mitkov.N, Lisbon 1999 . . . .

.............. .
.. ....... . .
.... ....... 205
....................... ..... ................

Goh Wei Ming-Bellini.F, Turin Olym piad 2006 . .

...................... 1 76 .......................... ..............

Golod.V-Sutovsky.E, Nata nya (ra pid) 2009 ... .. . ..... .. . . . ............... 146 ...... ..................................

Gordon.S-Short.N, British C h a m pion s h i p, Sheffield 2011 . . . . . 155..... ..... ................. ........ . ...

Grigore.G-Brochet.P, C reon 1999 . ..

............ .. .
........... .
........... . .. . 1 53
.... ..................... ........ ... ... ....

Grilc.A-Mohr.G, Slove n i a n Tea m C h a m pionsh i p 2008 . . . ......... 202 .................. ...... ...............

Gruenfeld.V-Smirin.l, Isra e l i Tea m C h a m pionsh i p 1997 . 191

........................................... ...

Grynfeld.I-Bisguier.A, Helsi n ki Olym piad 1952 . . 1 80

........................................ ............ .........

Gurevich.M-Rohde.M, Ph i ladel p h ia (bl itz) 1989 .. . . .. .... .... . . . .. . 156

.................... ................. . ....

Gurevich.M-Zoler.D, Antwerp 1998 . .

...................................... 152
........... .................................

Haessei.D-Schuyler.J, Pawtucket 2008 . . . .. .. .. .. . .. ... . ................. . 142

............ ....................... ........

Hahn.A-Bonin.J, New York (ra pid) 2003 . . ....... .

...... . .
............... 1 70
....................... ..... ................

Hjartarson.J-Oiafsson.F, Reykjavi k 1995 . .. . .

....................... ... 191
.............................................

Hoffman.A-Fernandez.A, Ma r del Plata 1996 . ..................... 150

.............................................

Hromada.P-Ostrowski.L, Moravi a n Tea m C h a m p i o n s h i p 2003 . 1 79 ........... ......................

Hubner.R-Hort.V, Germ a n League 1984 .

........................... . 1 74
......................... ......................

Hu g hes.H-Richardson.K, B ritis h League 2004 ................................................................. .201

lpatov.A-Antoniewski.R, Germ a n League 2011 ............................................................... 1 3 7
lruzubieta Villaluenga.J-Gulko.B, Sa n Sebastian 1996 . . ... . .... .207
...................... ... . .............

Jakic.I-Mestrovic.Z, Zada r 2001 204

............................................................................................

Janowski.D-Vates.F, Marienbad 192 5 ................................................................................ 203

Jelen.I-Dizdarevic.E, Lj u b lj a n a 1992 . .. ........... . . .. .
.. ....... .
........ . ... .... 1 68
. ............ ........... . ..............

Jhunjhnuwala.N-Giigoric.S, Lucerne O lym piad 1982 ..................................................... 195

Kaidanov.G-Miles.A, Pa l m a de Ma l lorca 1989 ................................................................. 1 66
Karpov.A-Chevallier.D, Fra n ce 1993 ................................................................................... 1 69
Keskar.H-Schuyler.J, H a m pton 2011 .................................................................................. 141
Kmoch.H-Vates.F, Hastings 1927/28 .. .. .
............... . . . .. . .
. . .......... . . .. . 147
...... .. .... .... ........ . .... .........

Kravtsiv.M-Tarlev.K, Evpatoria 2007 .................................................................................. 159

Meissner.H-Miles.A, E u rop'n C u p, Slough 1997 .. .. .. ............. 1 60
............................................

Mista.A-Szelag.M, Kosza l i n 1999 ......................................................................................... 204

Mitcheii.R-Schuyler.J, B l oo m i n gton 1991 .. . . . . . ...... .. . .. .. . .
..... . . . . 167
............ .. ......... ..... ........ .. . ...

Motylev.A-Giigoric.S, Yugos lav Tea m Cha m pion s h i p 2000 ......................................... 1 78

Nemcova.K-Oiafsson.F, Ma ria n ske Lazne 2008 ............................................................... 1 89
Nijboer.F-Miles.A, Li n a res 1995 ............................................................................................ 1 8 7
Onischuk.A-Shkuro.l, U k ra i n i a n Tea m Cha m pions h i p 2009 ........................................ 1 5 7

223
Th e Dark Kn ig h t System

Orso.M-Bordas.G, Buda pest 2000 . ........ . . .. . 1 75

................................................... ...... ........ ........ .

Paasikan gas Tella.J-Lindqvist.T, F i n n ish Tea m Cham pio n s h i p 1996 . 134 ...................... ..

Pedersen.P.B-Bekker Jensen.D, Da nish Tea m C h a m pions h i p 2008 .. ... . ... 145 ......... ...... . . .

Perrusset.B-Moullier.l, Pa ris 2005 .

................................... . 1 64
....................................... ............

Pol gar.So-Fries Nielsen.J, Rimavska Sobota 1991 192

...........................................................

Praznik.N-Beliavsky.A, B led 1999 .. ..

....................... ................. .. . 1 89
................................ . ........

Ramirez.J-Schuyler.J, La s Vega s 2007 .. .

.................... .. ...... . .. .... . 205
................. .............. . ...... . .. .

Rasic.D-Mestrovic.Z, C roatia n Tea m Cha m pions h i p 2001.. .. . . ...... . . . . . 144 ... .... . ..... ... . . ........ .

Rozman.L-Schuyler.J, Was h i ngton 2012 . . .

.................. ...... ....... . 186.........................................

Ryba.N-Schuyler.J, Was h i n gton 2012 . .

....................... ........................194 ................................

Saduakassova.D-Minasian.Art, D u ba i 2011 . .. ....

...................... ..
. 197
........ . . . ................... .......

Sakharov.A-Adorjan.A, Soc h i 1976 . . . . ... ..

............... ... .. .
.............. .. . 198
. .... . ......... ........................

Schiendorfer.E-Recuero Guerra.D, E u ro J u n ior Ch'sh i ps, Herceg Novi 2006 . . . 1 71 .. . .. ....

Sher.M-Spraggett.K, Andorra 1993 .

........................... . . 209
...................................... ... ..............

Smeets.J-Beliavsky.A, Netherla nd s-Siove n ia ra pid m atch, Maribor 2004 1 81 ...............

Spichkin.A-Reinderman.D, E u ro pea n C h a m pionsh i p, Rijeka 2010 . 211 ... .........................

Spiridonov.N-Spraggett.K, Ca n nes 1992 .. . .

............. . . ..
.............. 207
..................... ... ................

Sultimov.B-Pokazanjev.N, Russia 2007 . 1 82

................................................................... ..........

Sveshnikov.E-Gelashvili.T, Ca ppelle la G ra nde 2009 . .. 1 88

................................... ... ............

Titz.H-Barlocco.C, D resden 2004 .

................................ . 151
............................................ ............

Tratar.M-Srebrnic.M, Slove n i a n Cham pions h i p, Lj ublj a n a 2010 136 .................................

Vajda.S-Skembris.S, N a ujac 1999 .. ....... 1 78

...............................................................................

Valenti.R-Tkachiev.V, Corsica (ra pid) 1997 . .. . ..

...................................... 146 ............ . .. ............

Vialatte.J-Giroux.F, Paris 2006 ............. . . 135

................................................................... ...... .....

Von Wantoch Rekowski.D-Peric.P,Yugoslav J u n ior Ch'shi ps, Tivat 2001 . . .. 1 83 .. . ..........

Wang.K-Schuyler.J, Was h i ngton 2012 .

......................................... 199 .....................................

Weisser.W-Trumpp.L, Germa n League 2004 .................................................................... 1 62

Vilmazyerli.M-Arutinian.D, I sta n bu l 2007 . . 193
......................................................... .. ............

Zapata.A-Miles.A, Mata nzas 1995 . ... .

..................................... . .. 187
. ....... .............................. ...

Zelcic.R-Djurkovic.M, P u l a 2001 . ....... . . 200

............................................................. ........ ............

Zelcic.R-Mestrovic.Z, Nova Gorica 2003 205

.............................................................................

Zelcic.R-Miles.A, P u l a 1994 1 85
....................................................................................................

Ziatdinov.R-Chernin.A, New York Open 1998 .

.............................. 192 ...................................

224