Author: Uri Blass
Date: 11:32:10 01/19/01
Go up one level in this thread
On January 19, 2001 at 13:57:56, Dieter Buerssner wrote: > >>Crafty17.14 cannot see a forced mate for white after Rb6 in a few minutes and >>after more than 10 minutes on PIII450(11:19) it can only see 12.76 pawns for >>white at depth 20. >> >>Deep Fritz and Crafty need some minutes to avoid Rb6 in the original position >>and are slightly faster than Yace in this task but they do not see the forced >>mate. >> >>I see often in analysis that Crafty can see that white wins a pawn in the main >>line but it cannot see a forced mate or a draw inspite having the KRP vs KR >>tablebases. >> >>I guess that something like this cannot happen to yace. > >I accidently have changed the "TB-strategy" in Yace yesterday night. Therefore >I also was interested in this position. I believe the reason for different >behaviour of the programs here is the decision of the programmer, when to >probe TBs. Earlier versions of Yace did not probe TBs, when very deep in >search (at the last ply, or in lines, where many extensions have been executed). > >But this yields in inconsistences, and in the behaviour you noted above. >The pawn capture will be delayed, until a depth, where no TB access is done >anymore. But probing throuhg the whole search has the disadvantage, of >possibly much less speed. Actually, with another strategy, in the position >above Yace could avoid Rb6 after 1 minute. In other postitions I tried, >probing throughout the whole search (I did not try probing in quiescence search) >showed faster solving times. > >I do not know yet, which is better. > >-- Dieter I think that it suggests that a better strategy in games may be to decide about the strategy based on the root position. I think that probing throughout the whole search may give better results when the number of tablebase hits is small so you can decide to change this strategy only after you see a lot of tablebases hits. 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.