Author: Mike Hood
Date: 07:27:42 11/21/05
Go up one level in this thread
On November 21, 2005 at 03:25:07, Uri Blass wrote: >On November 21, 2005 at 02:38:09, Mike Hood wrote: > >>The following position is a better example of the usefulness of endgame >>tablebases. The only winning move for White is Rg6. I allowed Fruit 2.1 (without >>tablebases) 10 minutes for a search. > >This example show the negative value of EGTB > >How much time does your program need to find Kxb3 that is the only practical >chance to save the game > >[D]8/8/R7/8/1k6/1B1P4/6b1/K7 b - - 0 1 > >Fruit2.2.1 with all the 4 piece tablebases and 3 vs 2 tablebases cannot find it >in few minutes > Uri, thanks for your one-halfply backup from my position. I think the "problem" here is that we're dealing with extreme positions. In cases where the forced mate is very long (>50) and there is only one correct move, an engine might not find it if it is a move not obvious to typical computer algorithms. My position was an extreme position that shows the usefulness of EGTB's. Your backed up position was an extreme position that shows the potential harm of EGTB's. But let's be honest... how often do positions like these occur in practical play? On the other hand, extreme positions are what computer chess (and programming in general) is all about. If an obscure position can be manufactured where a usually successful algorithm fails it might be possible to deduce a common position where the same algorithm has a not easily apparent weakness.
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.