|
|
18. gxh3 Qh4
|
'Forking' the bishop on h6 and the castled King position. |

|
|
|
19. Bxg7
|
[!] Absolutely correct. For the moment two pieces behind, Black has a strong incentive to take the draw. |

|
|
|
19... Qg3+ 20. Kh1 Qxh3+ 21. Kg1 Kxg7
|
[!] Embedded in this recapture is Black's decision to eschew the bird-in-the-hand draw, and go all out for attack. Although objectively White might even have the whisker of an edge here, and there was a distinct chance that all I would have taken is a scenic route to a draw, yet I just felt that this was going to be ... fun. It has to be said, though: it was touch and go whether I would get much of anything. |

|
|
|
22. Bg4
|
[!?] This was the move I was expecting, though White might have done better with 22.Qd3, forcing Black back. |

|
|
|
22... Qg3+ 23. Kh1 Qh4+ 24. Kg2
|
[!] Very good! The draw is now off the table; White is also playing for the win. By playing the K to g2, the possibility of a White rook sliding over to h1 has to enter Black's thinking. |

|
|
|
24... h5
|
[!] I had really intended ...f5 as the beginning of the infantry assault, but after some thought didn't much like prospects after 24...f5 25.exf5 gxf5. Who would achieve final control over the g-file? Apart from that, I also wanted the g-pawn as a lever. That was not on the cards after 24...f5 etc. This move at least forced upon White the decision which direction the bishop would retreat. |

|
|
|
25. Bh3
|
[!] Best, I think. Keeps watch over f5, for what it's worth... |

|
|
|
25... f5
|
[!] I was going to play 25...g5?? then realised that 26.Bf5 would just about have left Black with nil attack. That was no use: I had to open lines. White's King position does look porous, doesn't it? Yet Black still has to find the way to crack it open and bring in his underemployed rooks. |

|
|
|
26. exf5
|
[!] Probably best, again. Instead 26.Qd3 f4 27.Qf3 g5 looks distinctly unhealthy for White. |

|
|
|
26... g5
|
[!] This was the plot! The f5-square being occupied, White's bishop can not escape that way. |

|
|
|
27. f3
|
[!] To hold back ...g4. |

|
|
|
27... Be5
|
[?] Wrong. Much better was to bring up the reserves: 27...Rae8. But I was lured by the possibility of the mate on h2... (...Qh4-g3-h2#). |

|
|
|
28. Ne4
|
[?] This was the move I anticipated, but 28.f4! offers much better chance of saving the game, according to the GK engine, which gives 28.f4 Bxc3 29.bxc3 g4 30.Bxg4 Qxg4ch 31.Qxg4 hxg4 32.Kg3 Kf6 33.Kxg4 Rf7 (High time something was done about the attack on b7!) 34.Rh1 Rg8ch. Although GK gives White a +0.67 at the end of this line, the thing looks very drawish to me. Incidentally, 28...g4 seems to lead only to a transposition of movesm after 29.Bxg4; and if White tries 32.Rxb7ch Kf6 33.Rb6 Rfd8 34.Kg3 Kxf5 this rook ending looks drawish as well. |
2 comments
|
|
|
28... Qf4
|
Threatening ...Qh2#. |

|
|
|
29. Rh1
|
Instead, 29.Kf2 g4 looked good for Black. However White did have an interesting alternative - a counter-attack: 29.f6ch Rxf6!? 30.Rh1 Rg6 31.Be6 g4. I feel that Black's prospects look pretty good in that line as well. |

|
|
|
29... g4
|
Closer, and ever closer, Black batters his way towards the enemy King. At last it looks as though Black at least will get his piece back! |

|
|
|
30. fxg4 Qxe4+
|
A definite gain of ground, and material is level again, but not quite a breakthough - not yet. |

|
|
|
31. Qf3 Qc2+ 32. Qf2 Qxa4
|
Rather unexpected fruits of the success Black has so far obtained...
|

|
|
|
33. gxh5
|
[?] Keeping the material level, but by opening up the g-file in this way, White's propects take a sharp turn for the worse. I was looking at the possibility of 33.b4!? a pawn sac to cut the Black Queen off from the K-side. But even better would have been 33.g5! The GK gives 33...Qe5ch 34.Qf3 Qxf3 with fair prospects for White to stay in the game, but I'm fairly certain I would not have exchanged Queens - not yet at any rate. |

|
|
|
33... Kh7
|
Here come the rooks - or one of them at any rate. White can do nothing to challenge possession of the g-file: 35.Kf1 Qe4 or 35.Kf3 Qf4ch. |

|
|
|