But I have never seen an attempt to quantify exactly how drawish they are, so I thought I would have a go using a database of my games.
Fortunately ChessBase has a tool that makes searching for such endings easy, and using it I found 72 of my games with such an ending.
Most of the games (42, or 58.3%) were drawn, but that leaves 30 that ended decisively - a higher percentage (41.7%) than I had instinctively thought.
Many of them I have already posted on this blog in a series entitled Opposite Exceptions, but here is an interesting one I omitted.
 |
| Black has just captured on c7 in L Demmel (163 BCF) - Spanton (140 BCF), Doncaster & District League 1981 |
*****
*****
*****
*****
White's extra two pawns on the queenside are, at least for the moment, firmly restrained by the black bishop. However the white king is much the better-placed monarch for operations on the other flank.
39.h4?!
But this is probably not how to go about kingside play. Stockfish15.1 and Komodo14.1 give best-play as 39.g4 Kc6 40.f4 Kd5 41.fxg5 hxg5 42.Bxg5 Kxe5, after which Black has a winning advantage thanks to having passed pawns on both flanks.
39...Kc6?
Correct is 39...gxh4 40.Bxh6 Kc6 41.Bg5 Kd5 42.Bf6 Kc4 43.Bxh4 Kxc3 44.Be7, after which White is still two pawns up but, crucially, only has one passed pawn. This significantly cuts White's winning chances. The engines reckon White has the upper hand, but that the advantage is worth less than a pawn, Indeed they reckon Black's two best continuations, 44...Bd3 and 44...Kb3, are exactly equal evaluation-wise, often a sign an ending is drawn.
40.hxg5 hxg5 41.Bxg5 Kd5
 |
| Black's active king means this is still a tricky ending despite White being three pawns up |
*****
*****
*****
*****
42.Be3?
Best seems to be 42.Bf4, eg 42...Kc4 43.Bd2 Kd5 44.f4, but also better than the text, according to the engines, is 42.f4?!, although then the c pawn will probably fall.
42...Kxe5 43.g4 Kd5 44.Bg5 Ke5 45.Be7 Kf4?!
It is not clear how White makes progress after 45...Kd5.
46.Bf8?
Winning is 46.Bd6+, eg 46...Kg5 47.Kg3 Kg6 48.f4 f6 49.Kf3 Bb7+ 50.Kf2 Ba6 51.Ke3 Kf7 52.Kd4, and White continues to make progress.
The engines reckon a better try is 46...e5!?, but then 47.Bc5 Kg5 48.Be7+ f6 (48...Kf4 49.Bh4) 49.Ke3 looks strong, eg 49...Kg6 50.Bd6 Kg5 51.Bc7 Bb5 52.f4+!? exf4+ 53.Bxf4+ Kxg4 is a win after 54.Bh2!, as the Syzygy endgame tablebase shows.
46...f5?
The black king should return to e5, with a probable draw.
47.Bh6+ Ke5 48.g5
White has passers on both flanks, which almost always is a winning advantage.
48...f4
 |
| White to play and win |
*****
*****
*****
*****
49.Bf8
The winning plan runs 49.g6! Kf6 50.g7 Kf7 51.Bxf4! Kxg7 52.Be5+ Kg6 53.Ke3, eg 53...Kf5 54.f4 Bf1 55.Kd4, after which White wins on the queenside. Black can try 49...Bd3, but 50.g7 Bh7 51.c4 also wins on the queenside.
49...Kf5 50.Be7 Bc4 51.Ke1 e5!?
The engines prefer marking time with the bishop, eg 51...Bb5 52.Kd2 Bf1 53.Kc2 Bc4 54.Kb2 Be2 55.Ka3 Bxf3 with complete equality, according to the engines.
52.Kf2 e4?
This lets White swop off his weak pawn (weak because it is fixed on the same colour complex as the black bishop) and bring the black king out of position. Instead, marking time with the bishop still seems to give good drawing chances.
53.fxe4+ Kxe4 54.g6 f3
There is no defence, eg 54...Kd5 55.g7 Kc6 56.Kf3 etc.
55.g7 Kd3 56.b5 1-0
I found 23 decisive out of 77 in my game history.
ReplyDeleteIn general the drawing method is to find good squares for the Bishop and King and then shuffle one or the other. Usually you need to be two pawns ahead to win.
RdC
I would say: two pawns ahead OR the ability to create a passed pawn on each flank, even if that means allowing the opponent a passer too
ReplyDelete