Author: Sune Fischer
Date: 13:40:04 09/21/02
Go up one level in this thread
On September 21, 2002 at 16:26:24, Daniel Clausen wrote: >On September 21, 2002 at 13:41:51, Sune Fischer wrote: > >>On September 21, 2002 at 13:17:15, Sune Fischer wrote: >>> >>>Type of move / % of these moves in search / % cutoffs by these moves: >>>hash.......: 0.44% 93.35% >>>null.......: 27.91% 76.20% >>>wcap.......: 3.39% 67.11% >>>ecap.......: 0.45% 60.75% >>>lcap.......: 3.87% 22.38% >>>kill.......: 3.10% 22.34% >>>ncap.......: 60.84% 0.50% >> >>The above was done with a long analysis of a position, >>running a game changes things, the hash is worse but the capture >>sorting works better: >> >>hash.......: 1.09% 88.80% >>null.......: 21.35% 67.28% >>wcap.......: 4.97% 68.31% >>ecap.......: 0.43% 43.53% >>lcap.......: 4.53% 5.87% >>kill.......: 2.94% 15.29% >>ncap.......: 64.67% 0.43% >> >>Looks like killers should be tried before losing captures. >> >>-S. > >Very interesting statustics, Sune! I'll have to do something similar in my >engine too. > >Sargon Thanks, but I realized there is vital piece information missing, there should be another column with the move ordering statistics for each type of move. Perhaps that is why my percentage of non-captures cutoffs is that low (I use a history table, but could be I reversed a sign or something). I wonder if others search 60% non-captures (not incl. the killers)? In any case it would be nice to "hot spot" which type of moves needs better sorting. -S.
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.