The Definitive Gusev - Auerbach, 1946 thread

[Paloma]

There is still a forced win after 29...cxb3

1. ... cxb3 2. axb3 Qe7 3. b4 a6 4. Kg2 Qe8 5. c4 Qe7 [(5. .. Qa8 6. Kg3+-)]
6. c5+-

Win involves a zugzwang, so SF might have difficulty at higher depths.
Thanks for this comprehensive analysis and comments by tpoppins.

2.(30.)...Qe7 is a mistake, better is 2.(30.)...a5 with draw

SF shows after 28. Bxc4 bxc4 29. b3 cxb3 30. axb3

Analysis by Stockfish 251018 64 BMI2:

1. = (0.00): 30...a5 31.c4 g5 32.c5 a4 33.bxa4 Qe7 34.c6 Qc5+ 35.Rf2 g4 36.c7 g3 37.hxg3 Qxf2+ 38.Kxf2
2. +- (15.69): 30...Qe7 31.b4 Qd8 32.Kg2 Qe7 33.Rf2 Qd6 34.Rf7 g5 35.h4 Qd8 36.Rf1 Qe7 37.Rf2 Qg7 38.Bxg7

(depth was 48)

So Sven and Peter are right.

[peter]

Yuri Gusev - E. Auerbach
Molniya Sports Society 1946
Sicilian Defense, Dragon, Classical 6.Be2 without Be3 (ECO B70)
Annotations by Y. Gusev

1.e4 c5
2.Nf3 d6
3.d4 cxd4
4.Nxd4 Nf6
5.Nc3 g6
6.Be2 Nc6
7.Nb3 Bg7
8.O-O

[d]r1bqk2r/pp2ppbp/2np1np1/8/4P3/1NN5/PPP1BPPP/R1BQ1RK1 b kq - 0 8
8...Be6?
The correct continuation in this position is 8...0-0 and on 9.0-0 there follows 9...b5, equalizing. This maneuver has been recommended by V. Rauzer.

9.f4 Rc8
10.f5 Bd7
11.g4 Ne5
12.g5 Ng8
13.Nd5 f6
14.Be3 b6
15.Nd4 Kf7

[d]2rq2nr/p2bpkbp/1p1p1pp1/3NnPP1/3NP3/4B3/PPP1B2P/R2Q1RK1 w - - 0 16
16.c3
Intending 17.Qb3.

16...Qe8 17.Ne6!
Threatening to win immediately with 18.Nd5-c7.

17...Bxe6 18.fxe6+

[d]2r1q1nr/p3pkbp/1p1pPpp1/3Nn1P1/4P3/2P1B3/PP2B2P/R2Q1RK1 b - - 0 18
18...Kf8
Not 18...Kxe6? because of 19.Qb3 Kd7 20.Qa4+ Kd8 21.Qxa7 and White wins.

19.Nxf6! Nxf6
19...exf6 is met with 20.Qxd6+, followed by 21.Qxe5

20.gxf6 Bxf6

[d]2r1qk1r/p3p2p/1p1pPbp1/4n3/4P3/2P1B3/PP2B2P/R2Q1RK1 w - - 0 21

Here White was presented with the choice between 21.Bg5 Kg7 22.Bxf6+ exf6 23.Qxd6, winning a pawn but leaving the opponent with a strong knight on e5, and the chance to continue the attack. White picks the second option.

21.Bh6+ Kg8
22.Rxf6 exf6
23.Qxd6 Rc6

[d]4q1kr/p6p/1prQPppB/4n3/4P3/2P5/PP2B2P/R5K1 w - - 0 24

Wonder if anybody ever commented, that White could have won some more easily by 23.Qd5 instead of Qxd6.

Of course, that's not as beautiful as the main line, but Black has much less answers and White can fail some times less easily.
And engines find it much more easily than the Queen- sac, there's much less Zugzwang involved.

After Backward of a corresponding line and 23.Qd5:

[d]2r1q1kr/p6p/1p1pPppB/3Qn3/4P3/2P5/PP2B2P/R5K1 b - - 0 1

Analysis by ExtendFish 64 POPCNT:

23...De7 24.Td1 Td8 25.h3 Te8 26.Dxd6 Dxd6 27.Txd6 Sf7 28.exf7+ Kxf7 29.Lc4+ Ke7 30.Te6+ Kd8 31.Txf6 Txe4 32.Lg5 Kc8 33.La6+ Kc7 34.Lb5 Kb8 35.Lf4+ Txf4 36.Txf4 Td8 37.Kg2 Td1 38.Tf1 Td2+ 39.Tf2 Td1 40.Kg3 Tg1+ 41.Kh4 a5 42.a4 h6 43.Le8 Kc7 44.Td2 Kb7 45.Td7+ Kb8 46.Td6 g5+ 47.Kh5 Tg3 48.Ld7 Tg2 49.Lb5 Txb2 50.Txb6+ Kc7 51.Tc6+ Kd8 52.Td6+ Kc7 53.Td7+ Kc8 54.Kxh6 Th2
+- (6.15) Tiefe: 39/64 00:03:23 4796MN

Still the eval is even a little bit higher having kept the main line in hash by Backward, but the difference isn't big enough, I'd say, to go such a complicated way over the board, having a much easier one too.
If it was a study, I'd call solution a major dual at the position before 23rd move.

[d]2r1q1kr/p6p/1p1pPppB/4n3/4P3/2P5/PP2B2P/R2Q2K1 w - - 0 1

Analysis by ExtendFish 64 POPCNT:

23.Dxd6 Tc6 24.Dxe5 fxe5 25.Tf1 Dd8 26.Ld1 Dxd1 27.Txd1 Txe6 28.Td8+ Kf7 29.Txh8 g5 30.Lxg5 Kg7 31.Ta8 Tg6 32.Txa7+ Kg8 33.Kf2 Txg5 34.Tb7 Tg6 35.a4 Tc6 36.Ke3 h6 37.Kd3 Kf8 38.h4 Ke8 39.b4 Kd8 40.h5 Kc8 41.Te7 Kd8 42.Txe5 Kc7 43.Te7+ Kd6 44.Tg7 Tc8 45.Tg6+ Kc7 46.Txh6 Ta8 47.Kc4 Txa4 48.Kb5 Ta2 49.e5
+- (7.03 --) Tiefe: 41/71 00:04:48 7200MN

Took ExtendFish this time only because of its nullmove off- parameter again.

Edit: Not to mention 17.Qb3 instead of Ne6 already.

[MikeGL]

I have no doubts about the position after 27...b5 being won for White, so the correctness of the queen sac remains untouched. However, the game did not proceed by 28.a4, which was demonstrated to be clearly winning for White in the post you referred to, but by 28.Bxc4 bxc4 29.b3, and my question was exactly about that latter position and the possible reply 29...cxb3. Mike replied there were a forced win after 30.axb3 Qe7 31.b4 but the stronger continuation for Black seems to be 30...a5, as written by TalJechin in the post you referred to. I tried hard to find a zugzwang or another way to win for White but I failed so far. Any idea?

[d]4q1kr/7p/4P1pB/p3p3/4P3/1PP5/7P/5RK1 w - - 0 31

EDIT: From http://www.chessgames.com/perl/chessgam ... 6#reply142:
"NeverAgain: Now let's turn to AK's claim that the game continuation <25...Rc8> draws. He cites <29...cxb3> as a critical improvement and I won't argue there, because White, in his turn, can improve earlier with <28.a4!>, as Vass of ChessPub discovered in 2011."

So it seems Gusev just showed he is just human too, and didn't play precisely with 28.a4! +- instead of 28.Bxc4= where Auerbach could've escaped with a draw.
You are indeed correct about 28.Bxc4 bxc4 29.b3= being drawn as pointed out by Peters analysis too. But sorry for my variation, I picked up the 30...Qe7 defence from your SF PV in your original post. Seems the point of tpoppins is valid too, we can't just blindly trust SF. Since the second move (30...Qe7 +-) of SF's PV at depth 75 is still wrong by not finding 30...a5= . Had SF at depth 75 just displayed 30...a5 on its PV above, I could have replied easily that it was indeed a draw, but no, I heavily analysed the wrong 30...Qe7 of SF and incorrectly claimed +-