7r/pp3k1r/2bp1pR1/4pP2/2PnP3/1PN3RP/Pq4BK/6Q1 b - - 0 32

[Site "?"] [Date "2014.10.22"] [Round "?"] [Result "*"] [TimeControl "60/120"] [TimeControl "60/120"] 1. Nf3 {I am playing Black, with 90 minutes for 36 moves. My opponent is a very experienced club player with much the same grade as me, who had beaten me in our previous two games.} 1... c5 {I am in the process of switching my repertoire from the King's Indian to the Benko, but it's not so easy to play the Benko here.} 2. c4 Nf6 ({The Accelerated Benko} 2... b5 {might be worth a punt.}) 3. Nc3 ({I was hoping to get a gambit in my new repertoire with} 3. d4 cxd4 4. Nxd4 e5 5. Nb5 d5 6. cxd5 Bc5) 3... Nc6 (3... d5 {gives some interesting variations but I thought my opponent might know them better.}) 4. d4 cxd4 5. Nxd4 g6 6. e4 {Giving a Maroczy Bind structure that also arises in the Accelerated Dragon Sicilian. In the symmetrical English version, White often delays this move and fianchettoes the king's bishop to impede Black's queenside development, but there's nothing wrong with my opponent's approach.} 6... d6 7. Be2 Bg7 8. Be3 ({If} 8. O-O {I was looking at} 8... Nxe4 9. Nxc6 Nxc3 10. Nxd8 Nxd1 11. Nxf7 Kxf7 12. Rxd1) 8... O-O 9. O-O Ne5 {With considerable difficulty, restraining my penchant for incorrect forcing lines in the opening.} ({It took me quarter of an hour to discover that} 9... Ng4 {loses to} 10. Bxg4 ({not} 10. Nxc6 Nxe3 11. Nxd8 Nxd1 12. Raxd1 ({or} 12. Nxf7 Nxc3) 12... Rxd8) 10... Bxg4 11. Nxc6 ({not} 11. Qxg4 Nxd4) 11... Bxd1 ({or} 11... bxc6 12. Qxg4) 12. Nxd8 {and White is up a piece. Not surprisingly, the machine confirms that I can't win it back:} 12... Bg4 13. Nxb7 a5 14. Nd5 ({or} 14. c5) 14... Rfb8 15. Nc7 {is one picturesque line.}) 10. h3 Qc7 {My opponent commented afterwards that White needs to strike fast in this variation and that this move made his e5 break difficult; but it does risk enhancing the Nd5 jump. I have little experience with this structure, and found myself playing it like a Scheveningen Sicilian, where White's c pawn is on c2. In the Maroczy Bind it is normal for Black to play a6 and b5, but I was worried about weakening b6 at this stage.} 11. b3 Bd7 12. f4 Nc6 {I was disgruntled at the loss of two moves with the knight and thought I was in trouble. But I manage to make some retrospective sense of provoking f4.} 13. Rc1 Nxd4 14. Bxd4 Bc6 {Targeting e4 in Scheveningen style.} 15. Qd3 Nd7 16. Bxg7 {This exchange leaves White's dark squares weaker than Black's.} (16. Be3 Nc5 17. Qb1 {is possible, holding e4.} ({not} 17. Nd5 Nxd3 18. Nxc7 Nxc1 19. Nxa8 Nxe2+)) 16... Kxg7 17. Kh1 Qa5 {Here I began to feel I might be clambering out of the pit.} 18. Bg4 {My opponent thought for twenty-three minutes here and caught up with me on the clock. He is angling for a break.} 18... Nc5 19. Qd4+ f6 (19... Kg8 20. Nd5 e6 21. Ne7# {is worth noticing.}) 20. Rc2 ({I was afraid of} 20. Nd5 Bxd5 ({my first idea} 20... e6 {is hit by} 21. b4 {and the complications favour White} (21. Ne7 e5 {is fine for Black})) 21. exd5 {though the machine likes} 21... Qxa2) 20... e5 {Now Black is well back in the game.} 21. Qe3 (21. Qxd6 Rad8 22. b4 ({or} 22. Qe7+ Rf7) 22... Rxd6 23. bxa5 {is fine for Black.}) 21... h5 (21... f5 {is a fine mess but it seems positionally dubious to reinvigorate White's bad bishop.}) 22. Bf3 Ne6 23. g3 (23. f5 {is possible, e.g.} 23... Nd4 24. Rd2 gxf5 25. Bxh5 f4) 23... h4 {Here my optimism kicks in. It feels positionally right to weaken White's dark squares further, but I underestimated his play on the g file.} (23... f5 {is again interesting.}) (23... Qc5 {is another way of fighting for the dark squares.}) 24. f5 Nd4 ({Here I thought} 24... Qc5 {would be safer, but decided to go with the flow. Actually Black is in danger either way.}) 25. Rg2 hxg3 (25... gxf5 {may be better.}) 26. Rxg3 Rh8 ({Amazingly I had planned} 26... g5 27. Rxg5+ fxg5 28. Qxg5+ Kf7 {but now I spotted} 29. Bh5#) 27. Rxg6+ Kf7 28. Kh2 ({A teammate came up with the beautiful idea} 28. Bh5 {??!!} 28... Rxh5 29. Rfg1 Rh7 (29... Rah8 {is winning for Black}) 30. Qh6 {! with the idea} 30... Rxh6 ({the machine scrambles to a lost ending with} 30... Bxe4+ 31. Nxe4 Rxh6 32. Rg7+ Ke8 33. Nxd6+ Kd8 34. Nxb7+) 31. Rg7+ Ke8 32. Rg8+ Kd7 33. R1g7#) 28... Rh7 29. Rfg1 Rah8 30. R1g3 {Basically White is a pawn up with chances of consolidating. But now it's my opponent's turn to underestimate my play.} 30... Qa3 {Resorting, with nine minutes for seven moves in this sharp position, to the principle of improving the worst-placed piece. It's a bit more relevant than it looks.} 31. Qg1 {A fatal oversight. Alekhine's gun won't get to fire here.} ({My opponent and I looked at many variations here after the game.} 31. Qf2 Qc1 {and the machine still prefers White.}) 31... Qb2+ {!} 32. Bg2 ({If} 32. Ne2 {then} Nxf3+ ({or} 32... Rxh3+)) {Here my pulse was racing, but fortunately there's not much to calculate. Diagram at the top of the page.} 32... Rxh3+ {!} 33. Rxh3 Nf3+ {The file and the rank converge rather nicely here.} 34. Kh1 Rxh3+ 35. Bxh3 Nxg1 {and White resigned.} * 0-1