MICHIGAN CHESS ASSOCIATION

Analysis by Vasik Rajlich


Gershon,A.(2435 FIDE)-Rajlich,V. (New York Open, Round 9)
King's Indian: Classical (Stein) [E92/25 (2nd ed.)]
1998.03.22. USA New York, NY [TC:40/120; G/30]

1.d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 O-O 6.Be2 e5 7.d5 a5 8.Bg5 h6 9.Bh4 Qe8
Though we eventually transpose back into the normal line, 9...Qe8 is a slightly inaccurate move order.  Before I present the drawback of 9...Qe8, I will present the argument which recommends it:
In the main line (9...Na6 10.0-0 Qe8 11.Nd2 Nh7 12.a3) Black, in order to force White to continue more slowly on the queenside with 13.b3 and 14.Rb1 prior to an eventual Pb4, plays 12...Bd7, since other moves are met with the immediate 13.b4.  12...Bd7, however, could be thought to have two drawbacks:
1) The d7-square may not be the long-term destination for the c8-bishop. For example, in the above-mentioned main line, after 12...Bd7 13.b3 f5 14.exf5, Black can play 14...Bxf5.
2) The d7-square might be useful for the black queen.  For example, again in the above-mentioned main line, after 12...Bd7 13.b3 f5 14.exf5, Black will have to sacrifice an exchange after the thematic 14...gxf5, i.e. 15.Bh5 Qc8 16.Be7, etc.
Hence, the logical question is:  Why play 9...Na6?  With the knight still on b8, a2-a3 before b2-b3 can be met with ...a4 even when Black's bishop is still on c8, since Black's a8-rook still controls that square.  Should White play b2-b3 before playing a2-a3 with the idea of continuing with a2-a3 and b2-b4 without wasting time on Rb1 (as he must - or at least normally does - when Black's knight controls b4), then Black will finally play ...Na6 when White has played b2-b3 without Black having had to force it with ...Bd7.
Unfortunately, there are two caveats which I present below:

10.Nd2 Nh7 11.O-O
With this move, my opponent decides to head back for the main line, correctly ascertaining that I will be unable to take advantage of having refrained from an early ...Na6.  More confrontational would have been 11.a3.  The only reply which stops 12.b3 (and thus the only reply consistent with Black's refusal to play ...Na6) would then be 11...a4 (In the only game I have which reached this position - M. Hoffmann-Knuth, Budapest 1994 - Black continued instead with 11...b6 12.Nb5 Qd7 13.b4 ["White has saved considerable time" - Nunn], etc. [+/-]) 12.Nb5 Qd7 (12... Na6 13.Qxa4 [+/-]) when Black's position is somewhat awkward.
On the other hand, it is not completely clear that White has improved on the main line here.  He can win a pawn with 13.Nb1, intending Nb1-c3xa4, and none of Black's queenside pieces can move in the meantime, but Black can nevertheless cook up some serious play while this manoever is in progress: 13...f5 (taking advantage of White's loss of control over e4) 14.f3 fxe4 15.fxe4 h5 16.N1c3 Be3 and Black has tremendous play.  White probably should at least wait with 13.Nb1, but then Black has achieved his normal goal in this opening of stalling White's queenside play with ...a4.
I consider 9...Qe8 slightly inaccurate mainly because White has the option of playing 11.a3 for free, as after 11.0-0 Black in unable to prevent a transposition to main lines, and not because 11.a3 is overpowering.

11...h5
This heads directly for the main line, while 11...Na6 is the main line.  Black had one way to give his play independent significance: 11...f5 12.exf5 gxf5 (12...Bxf5 would be horrible: 13.g4 and White has complete control over e4.  This line should be compared to the line 9...Na6 10.0-0 Qe8 11.Nd2 Nh7 12.a3 Bd7 13.b3 f5 14.exf5 Bxf5 15.g4, when only 15...e4 gives Black chances for a playable position.), attempting to take advantage of the fact that unlike in the 9...Na6 10.0-0 Qe8 11.Nd2 Nh7 12.a3 Bd7 13.b3 f5 14.exf5 gxf5 line White cannot win an exchange with 15.Bh5.  Unfortunately, after 13.Bh5 Qd7 14.f4, I believe that White would have a sizeable advantage.  This particular pawn formation is quite difficult for Black to play even in normal cases, and here e6 is particularly sensitive due to the lack of coordination between Black's queenside pieces (i.e. 14...e4 15.Nb5 and 16.Nd4).

12.f3 Bh6 13.Kh1 Be3 14.b3 Na6 15.a3 Bd7
The point of White's play is that Black does not have any real alternatives to transposing back to the main line.  One way to look at this is that this entire variation is not particularly tempo-critical, and it is not particularly promising for either side to look for superprecise ways to avoid playing modest developing moves.

16.Rb1 Bc5 17.Qc1
Black has played his obvious moves and now must make a choice.  If left alone, White's next two moves will be 18.Na2 and 19.b4.  The full page of analysis of this position in Nunn's "The Main Line King's Indian" focuses on 17...f5, 17...Kh8 (with the idea of ...f5), and 17...f6.  17... f6 (Black's idea is to play ...g5 and ...g4 while opening a path to the kingside for his queen along the light squares, and he begins, as is normal in this variation, with the least commital part of the plan) is a clever idea which I did not consider at the board and which I very much like.  17...f5, however, I simply do not understand. 
My conclusion, both at the board and now, is that after 17...f5 18.exf5 gxf5 19.f4 e4 White is getting a very favorable version of this particular pawn structure.  Normally, in arriving at this pawn structure Black can at least count as an achievement the activation of his c5-bishop, which is already active here. Furthermore, his pawn on h5 is a weakness and if anything clogs his play, since a rook cannot be swung to the h-file while no piece can use h5.  In addition to this, I believe that this entire pawn structure is suspect in the first place.

In "The Main Line King's Indian", however, after 17...f5 18.exf5 gxf5, Nunn doesn't even consider 19.f4, mentioning only 19.Na2 Qg6 20.b4 axb4 21.axb4 Be3 and 19.Bf2 Qg6 20.Bxc5, when in both cases Black gets reasonable kingside play.
17...Qb8?!
Everything happens too slowly in this variation for me to demonstrate that after this move White gets an edge, but nevertheless I do feel that this move is inaccurate.  Black needs to attack on the kingside, and while Nunn analyses 17...f6 as leading to an edge for White after 18.Na2 g5 19.Be1 g4 20.b4 axb4 21.axb4 Be3 22.Nc3 Qg6 23.Nd1, etc., it seems to me that in this line Black is getting reasonable play and should somehow emerge in decent shape.  After 17...Qb8, Black's position is too passive.

18.Na2 Qa7 19.b4 axb4 20.axb4 Be3 21.Qc3 g5 22.Be1 h4
The idea behind this move is to play ...Nh7-f6-h5-f4 and perhaps ...h3.  After both 22...f5 23.exf5 Bxf5 24.Bd3 and 22...g4 23.Nb3 Black's attack is struggling with his queen not participating.

23.Nb3
23...f5
This leads to an edge for White but should nevertheless be played, since otherwise White will continue with 24.Rb2 and 25.c5 when he is having his way.
23...Nf6, as according to the original plan, would undermine Black's ability to respond to 24.c5 with 24...Nxc5, since 23...Nf6 24.c5 Nxc5 25.bxc5 Qxa2 26.Qxe3 Qxb1 27.Nd2 and 28.Qxg5+ would win for White.

24.c5
24.exf5 Bxf5 25.Bd3 was stronger and highlights the problems with 17...Qb8, as Black's airy kingside does not combine well with the location of his queen.  25...e4 26.fxe4 Bxe4 27.Bd2 would only make Black's problems worse.
After 24.c5, the position is roughly equal.

24...Nxc5 25.bxc5 Qxa2 26.Rb2 Qa3 27.Rb1 Qa2
After 27...Bxc5 28.Nxc5 Qxc5 29.Qxc5 dxc5 30.Rxb7 White shouldn't stand worse.

28.Bd3
Objectively speaking, White should take the draw with 28.Ra1, since 28.Bd3 leaves Black slightly better.

28...h3 29.Rb2
29.gxh3 fxe4 30.Rb2 Bxh3 would be good for Black.
After 29.Rb2, the following critical position has arisen:
29...hxg2+??
This allows White to completely turn the tables on the kingside and gives him a huge edge.  29...Qa3 was forced and appears to leave Black with the better chances: 30.exf5 (30.cxd6 cxd6 and Black threatens 31...Rfc8 while 31.Rb1 could at least be met with 31...Qa2 though there may be even better; 30.c6 bxc6 31.dxc6 Bxc6[-+]) 30...dxc5 (30...Bxf5 31.Bxf5 Rxf5 32.Ra1 [+-]) 31.gxh3 (31.Rb1 Bd4 32.Nxd4 Qxc3 33.Bxc3 exd4 34.Bd2 b6 and Black is better.) 31...Bxf5 (31...Bd4 32.Nxd4 Qxc3 33.Bxc3 exd4 34.Bd2 b6 35.Rg1 and White is better) 32.Bxf5 Rxf5 33.Rb1 Bd4 and Black is better.

30.Rxg2 Qa3 31.exf5 dxc5
31...Bf4 was necessary, though White would have a huge edge.

32.f6!
32.Qxe5 would lose to 32...Bd4 33.Nxd4 Qxd3 34.Ne2 Rfe8

32...Bf4 33.Bxh7+ Kxh7 34.Qc2+ Kh6 35.h4 Rg8 36.Rfg1 Qa6 37.hxg5+ Rxg5 38.Rxg5 Bxg5 39.Qh2+ Kg6 40.Bd2 Qxf6 41. Rxg5+ 1-0.