Author: Heiner Marxen
Date: 10:43:03 12/10/02
Go up one level in this thread
On December 10, 2002 at 12:28:06, Daniel Clausen wrote: >On December 09, 2002 at 11:15:35, Dieter Buerssner wrote: > >>On December 09, 2002 at 09:55:50, Robert Hyatt wrote: >> >>>Some have tried. Monty Newborn had a pawn-endgame-only program called >>>"peasant." It >>>couldn't do this one either... >> >>The program Wilhelm by Rafael B. Andrist will solve it at ply 1 (score +1.8). It >>did not switch away in few minutes. It has, according to its author, >>sophisticated knowledge about corresponding squares. >> >>When I do Kb2 on the board, Wilhelm shows 0.0 immediately. >> >>Regards, >>Dieter > >If I remember right, Rafael once mentioned in CCC, that his program knows what >'Gegenfelder' are. Not sure what the English term for that is though. But it's >the theory, where you know that if whites king is on square X, blacks king has >to be on field Y, in order to hold the draw. Its called "corresponding squares" IIRC. Yes, I've seen his code (on my request he sent it per email to me). He recognizes potential candiadates, and does a backwards analysis (like that in EGTB generation), and caches some recent results. Backward analysis of this type can be much faster than the equivalent search. I'll experiment with that in Chest, also. >A very good example in my opinion, which shows what one can do in an engine with >chess knowledge. In Fine70s position, it's not just 'pure luck' that Kb1 wins, >whereas Kb2 only draws... > >Sargon I agree. Cheers, Heiner
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.