Author: Ed Schröder
Date: 20:45:05 04/16/00
Go up one level in this thread
On April 16, 2000 at 13:35:44, James Swinney wrote: >Here is a brilliant little problem that a friend of mine came up over 10 years >ago that we have tried on countless chess progams and none have ever figured >this out. I tried it on Fritz 5.32 with 64 meg hash tables overnight and it blew >up at the 41st ply !! >Here is the FEN >6bn/2k2prq/5prp/5pbn/5p1p/1Kp2PpP/2P3P1/8 w - - 0 1 > >The problem is white only sees that he is down tons of material and is quite >content to find a draw by perpetual check after queening his pawn. While a human >would see that all of blacks pieces are useless and play it out as a won king >and pawn ending. >Please try this and see what results you get. >James Don't blame chess programmers for not understanding these kind of positions by evaluation. These cases do not occur in normal games so a chess programmer (in general) don't care. *IF* these types of positions would happen frequently in a game the problem would have been solved 15-20 years ago. Ed
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.