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 demonstrated by the variation 9 e4 lbc6 10 lbf3 'ii'a5 ! 11 o-o-o (11 i..d3 :es 12 i..c2 b6 13 a3 'ii'c5 or 12 o-o J::r.e6 13 'fi'a3 'ti'xa3 14 bxa3 and White can stop bragging about his superior structure) 11 ... a6 12 'it>b1 .:te8 13 a3 b5 14 cxb5 axb5 15 i..xb5 i..a6.
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,
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'a5+ 15 t2Jc3 t2Jxc3 16 bxc3 t2Jd6 17 J:txd7 .l::tab8 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'xd6 'ii'xf2 15 t2Jxe5 and 15 ... i.f5, 15 ... i.g4!?, 15 ... i.e6! ?, or 15 ... t2Jc6! ? with equal chances i n all cases) 12 ... b6 13 e4 l2Je7 14 o-o .ib7 15 'ii'd3.
By playing for ... d6-d5 (with the queenside fianchetto) Black has man aged to provoke e3-e4 again, which is
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 ... t2Jg6 16 g3 t2Jf8! 17 l:tad1 l2Je6 ! 18 iVxd6 l2Jd4 (or 18 .. ."ifxd6 19 l:txd6 l2Jc5) 19 'ii'xd8 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 b5! 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 'iic2 d5! with advantage) 11 ... e4 12 l2Jd4 t2Jxd4 13 .:txd4 e3 and although Black's d-pawn remains, White now has a matching one on the lovely e2-square, and has at least as much to worry about.
7 e4
7 lL'lf3 fs 8 g3 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 downsides; 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 g3 'ii'd7 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).
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 against the straight-laced, booked-up senior master. We pick up after 1 d4 li:Jc6 2 c4 e5 3 d5 i..b4+ 4 i..d2 i..xd2+ 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 ... g7-g 5 possible. Here Black prepares to play 10 ... fxe4.
10 exf5 i..xf5 11 i..g2 o-o 12 o-o
I was justifiably happy with my po sition. Houdini prefers Black and sug gests ... �d7, ... i..h7, .. 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:Jeg8 18 g4 'ike7 19 h5 i..e8 20 'iVd3+ 'it>g7 21 li:Je4?!
I have 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..d7 23 'ii'e3 b6 24 g5?!
hxg5 25 'ii'xg5+ 'it>h8 2 6 li:Jf5 i.xf5 2 7 .l:txf5 .l:iaf8 28 h 6 'ii'e8
29 'ii'g6??
White, short on time and frustrated about being unable to break through, commits a horrible blunder, allowing his queen and rook to be forked. In stead, after 29 l:txf6 ! ? li:Jxf6 30 l:i.xf6 "fle7 31 ki.xf7 'ii'xg 5 32 l:lxf8+ 'it>h7 a crazy endgame 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 have 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 h6! 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 ... h7-h6 is generally very use- ful as well. Houdini claims 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, Israeli Team Champi onship 1999 - an annoying move be cause it pins the d-pawn, complicating the defence of es (if Black has to play ... lbg6 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 11 .i.h 3 d6 12 lt:\d4 .i.d7 13 �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 h3 h 5 or 1 4 ... lLlh4 gives Black considerable pressure for the pawn) 10 ... tLlf6 11 tLlxe5 d6 12 lLlf3 (or 12 tLld3 .l:r.e8 13 .i.d2 .i.fs 14 e3 c6 15 dxc6 tLlxc6 and 16 ... d5) 12 ... c6! 13 dxc6 .i.f5 14 �a4 bxc6 with full compensa tion; e.g. 15 .l:.d1 .l:.e8 16 .i.d4 lt:\e4 17 e3
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 h6 with just a big mess.
For the cowardly it is not 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 15 ... lt:'Jd6) 10 ... d6 11 g3 c6 12 dxc6 ltJxc6 13 lt:'Jxc6 bxc6 14 .i.g2 d5 15 cxd5 cxd5 16 .i.c3.
c) 5 'ii'c2! ? 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 'iWh5 12 g4?! 'ir'h4+ 13 'ito>d1 h5 14 g5 'ifxg5 15 lt:'Je4 'ifxf5 16 lt:'Jg 5 e4!, or 11 h3 'ifh5 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 ... h6! 8 'i!Vxg7?! lt:'Jg6 9 lt:'Jf3 (White should probably prefer 9 h4! lt:'Jf6 10 h5 �h7 11 �xh7 lt:'Jxh7 12 hxg6 fxg6 13 l:txh6 lt:'Jf8 14 lt:'Jf3, when 14 ... c6
or 14 .. JWd7 gives Black slightly the bet ter chances) 9 ... lt:'Jf6 10 lt:'Jg 5 "ille7 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, Black spoiled some fine work with 15 ... .i.xg4?! 16 lt:'Je4 .i.f5 17 lt:'Jf6 which was approximately equal, whereas 15 ... lldg8! 16 .i.xd7+ �xd7 17 lt:'Jf5 Uxg7 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 kingside storm (see Game 21). Instead, 8 ... g6 9 'ii'h4 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 ... 'iite8! ?, 15 ... 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 game Black contrived to lose.