Author: Dieter Buerssner
Date: 11:17:34 05/12/05
Go up one level in this thread
On May 12, 2005 at 09:23:13, Uri Blass wrote: >On May 12, 2005 at 09:00:31, Ross Boyd wrote: > >>On May 12, 2005 at 07:55:03, James T. Walker wrote: >> >>>As soon as either white pawn is captured, 6 man tablebases will tell you it's a >>>draw. >> >>Amazing. A Draw?????!!!!!! >> >>Thanks for pointing this out. It doesn't look possible, but I believe you. >>I wonder how many engines see the draw without 6 men EGTBs? >> >>Cheers, >> >>Ross > >Yace cannot even after capturing the pawn >I delete most of the analysis and a question to Dieter is how is it possible >that the selective depth 110 is smaller than the depth 126. This is easy to answer. All paths in the search tree end, before the depth is reached. Reasons are: stalemate, mate, repetition, 50 moves rule, cutoff from TT. >Note that Yace can see the draw easily one ply later without learning at depth >54/109. > >It suggest that there is some bug in yace, Possible. Perhaps I am going to investigate this by a tree dump. I guess, it has to do with the fact, that TTs handle repetion and 50 moves rule wrong. Often, this is "self healing" with higher depths. But not in this case. I thought I had some code, to not proble HTs when remaining depth + half move counter without capture/pawn move > 100. Perhaps I have taken out this code again ... Even with this code, the result will not be reliable. But it may reinforce the self healing. Regards, Dieter
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.