Author: James T. Walker
Date: 14:28:18 05/29/02
Go up one level in this thread
On May 29, 2002 at 16:07:29, James T. Walker wrote: >On May 29, 2002 at 10:22:42, Jeroen van Dorp wrote: > >>addition (sorry) .... and also black won't benefit from using tablebases. >>Tablebases don't make the program win - they just tell the outcome with optimal >>- any- play. A decent program playing white *without* tablebases won't screw up, >>and play as optimal as it gets. >> >>J. > >Well my question was not hypothetical. And any program without tablebases must >find Kd6 on it's own to qualify. And I'm not sure I agree that tablebases will >not help in the defense. And of course it must win without violating the 50 >move rule. >Jim In an attempt to answer my own question I used Century 4 as the program without tablebases vs several opponents. Century 4 plays Kd6 quickly. I found that Crafty seems to offer the most resistance but still gets mated quickly after Kd6. Some programs that do not quit searching immediately upon finding the mate in 66 may find the right move. Crafty found Kd8 and mate in 48 in less than 1 minute. Gandalf 5.1 in Chesspartner GUI plays Kd6 and announces mate in (17) a little over 3 minutes on my 1.4G AMD. I also discovered that Gandalf 5.1 running under the Chessbase GUI is really screwed up. Will not run in analysis mode in Junior 7. Moves immediately in 40/2 level and plays Kd6 with a score of 0.00. In the Fritz 7 GUI Gandalf 5.1 does not seem to be using tablebases (except when 5 men are actually reached) and finds Kd6 quickly. Jim
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.