Author: Robert Hyatt
Date: 10:16:52 02/06/02
Go up one level in this thread
On February 06, 2002 at 12:11:52, Dan Andersson wrote: >But how efficient will it be? Consider how effective the normal hash is. The >pawn hash will be even more so because there are fever unique pawn positions. >Hashing in real applicatins are full of non obvious results. And the different >pawn positions taper off much more quickly than full positions. Another >distinction is that most normal hash hits are on the same ply (something in the >order of 99%) while pawn hash hits are more evenly distributed at different >depths of search. So a pawn hash is bound to be 'very' effeftive ir real >performance. > >MvH Dan Andersson Another thought is to consider how many really unique pawn positions there are in a given position, since some are missing and will never come back, and once a pawn is pushed, it can never back up. The number of unique positions really shrinks when you think about how static parts of the pawn structure end up during a real game...
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.