Author: Uri Blass
Date: 21:54:14 09/26/03
Go up one level in this thread
On September 27, 2003 at 00:15:33, macaroni wrote: >My program (macaronibot on icc) can play a fairly reasonable middlegame at the >moment, but the second it gets to the ending, it turns into a real fiasco. This >ridiculous position for instance. > >[D]8/5r2/P7/3K1k2/8/8/7p/1R6 w - - > >Naturally this is nothing but drawn, but my computer managed to lose by playing >Rf1+ Kg4 Rh1 Kg3 Rb1 (no doubt white could sac the pawn, and then get his pawn >on the 7th, I think it's a draw?) Rh7 and now the horror, a7??? just throws any >chance away Rxa7 and my computer lost some moves later. > >But what can be done about this. For instance, to see the fact that it can sac >the rook for the pawn on the 7th would take a very deep search, as black can >check check check, and still be a rook for pawn up. > >Is the only way to have endgame tablebases, or is there some way of programming >that sort of stuff. You do not need tablebases for this problem. First it is trivial for my program even without them that Rf1+ lose a pawn when Rh1 does not lose a pawn. Not playing a7 is solved by evaluation and Kc6 evaluated as sligtly better than a7 thanks to some evaluation of pawns relative to the kings. Uri
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.