Author: Bruce Moreland
Date: 02:48:16 10/23/98
Go up one level in this thread
On October 22, 1998 at 13:16:10, Robert Hyatt wrote: >looks hopeless to me, at present. It ends up with the pawn on h7, bishop >holding it on g8 (also holding e6 pawn as well) with the black king on g7 >or h8 and the black bishop staying on the diagonal to cover e7. If I remove >the pawn on e7 I get instant draw scores as I recognize the wrong bishop + >RP ending. But with the passer on e6 I don't. This one has to be solved >by evaluation, as the repetition or 50-move rule will take a while to kick >in since white thinks he is winning and will avoid repeating until it is >no longer possible. I think it is a fascinating problem because of the complex wrong rook pawn situation. The reason I became interested in this position is that I had something similar in a real game the other day. My program was playing black and had a pawn on c2, the bishops were opposite, and my opponent had an h-pawn and an a-pawn. I had a negative score on this, even though it was very obvous that the game could be drawn by forgetting about the c2 pawn, and running for the "wrong rook pawn" corner. My program didn't do this, but it still managed to draw. I can imagine losing such a game if the knowledge is wrong. If you run to the wrong corner with the king, the enemy bishop can protect that pawn, and the king can win the other bishop for the wrong rook pawn, leading to a trivial won case. I was surprised that DarkThought doesn't get this (according to their web page), since it supposedly has a lot of endgame recognizers. If you are reading this, Erst, could you describe some of the ones that you do have? I think bishop and wrong rook pawn knowledge is crucial. It is one of the primary ways of escaping from lost positions. bruce
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.