BOOKS aimed at the great mass of chess players, especially those books with a high endgame content, tend to emphasise how drawish endings of opposite-coloured bishops can be.
Examples usually include ones where a two-pawn advantage, let alone one pawn, is not enough for a win.
Novices and "improving" players, quite naturally, tend to be impressed by this - so much so that they often get offended if an opponent a pawn up in such an ending refuses a draw offer.
But experience in the hurly-burly of practical chess reveals there are many exceptions, and anyway care is required by the defender.
In this series I will look at examples, in chronological order from my own praxis, where the draw was not achieved despite the loser being only one pawn down (and in some cases with level material).
Spanton (147) - C Wright (179)
Barbican Rapid 1991
 |
| Black has just captured on d6 - what should the result be? |
*****
*****
*****
*****
*****
*****
*****
*****
*****
Black's extra pawn, his well-placed king and the difficulty of White effecting a blockade mean Black is winning, according to the analysis engines Stockfish12 and Komodo11.01 (but see below).
40.Kg1
White's first move is uncontroversial, but how should Black reply?
*****
*****
*****
*****
*****
40...f4?!
The engines do no like this natural-looking advance. I believe the point is Black should get his pieces into as good positions as possible before advancing pawns. That calls for something like 40...Bc5+ 41.Kf1 Ke5, although the engines' assessment has changed from Black is winning to Black has the upper hand. One continuation runs 42.Ke2 g5 43.Bf7 f4 44.b4 Bb6 45.Be8, when it is far from clear how Black makes progress, so it may well be that the diagrammed position is drawn (from past experience I suspect it will become clear in this series that engines have trouble evaluating this type of ending, tending to put too much emphasis on material).
41.Bd5 Ke5 42.Bb7 Bc5+ 43.Kf1 a5 44.Bc6 b4 45.axb4 axb4 46.Ke2 g5 47.cxb4 Bxb4 48.b3 h5
Black has a two-pawn advantage on the kingside, but is it enough?
*****
*****
*****
*****
*****
Komodo11.01 still gives Black the upper hand, but Stockfish12 is probably right in reckoning Black only has a slight edge. In other words, the game should be drawn.
But what should White play now?
*****
*****
*****
*****
*****
49.Be8?
This natural-looking move (oh how often "natural-looking" moves are mistakes!) loses. Correct is 49.Bd7 to stop ...g4.
49...h4?
Another natural-looking move, and giving it a question mark might seem harsh, but Black has a win by returning the extra pawn with 49...g4!, eg 50.Bxh5 gxh3 51.gxh3, giving a position (see diagram below) where the connected black passers win as long as Black avoids 51...e3?, which allows White to blockade with the bishop. No better for White is 50.hxg4 hxg4 51.Bh5 f3+ 52.gxf3 exf3+ (52...gxf3+? 53.Bxf3) 53.Kf1 Kf4 etc.
 |
| Analysis position after 51.gxh3 |
50.Bc6 Kd4 51.Bb7
51.Bd7 is the clearer way to a draw.
51...Bc5
Interesting is 51...g4!? 52.hxg4 h3!? 53.gxh3 f3+ 54.Kf2 Be7 when 55.b4 seems to save White, eg 55...Ke5 56.b5 Kf4 57.Bxe4! Kxe4 58.Kf1 Bc5 59.h4 Kf4 60.g5 Kf5 61.Ke1. Komodo11.01 still gives Black the upper hand, but there is no way to make progress.
52.Bc6 g4!? 53.hxg4 h3!? 54.gxh3 f3+ 55.Kf1 Ke3
Komodo11.01 reckons 55...Be7!? wins, but Stockfish12 disagrees and is almost certainly right. Komodo11.01's main line goes 56.h4 Bxh4 57.b4 Ke3 58.b5 Bd8 59.g5! Bxg5 60.b6 Bd8 61.b7 Bc7, when it still insists Black is winning, but the engine is clearly wrong, eg 62.Bb5 Kd2 63.Kf4 (63.Kd2 Bc6) Bc6 64.e3 Bb5 - Black cannot make progress.
56.g5 Kf4 57.g6 Bd4 58.b4 e3 59.Bb5 Kg5 60.Bd3 Kh4 61.b5 Kxh3?
Any reasonable move, including 61...Kh5 and 61...e2+, ie 62.Bxe2 fxe2+ 63.Kxe2 Kg5, draws.
62.b6 1-0
No comments:
Post a Comment