Author: Brian Richardson
Date: 08:11:56 01/23/01
Go up one level in this thread
On January 22, 2001 at 12:52:04, Peter Kasinski wrote: > >I am trying to understand the impact of _not_ having some of the tablebases on >programs like Crafty or Deep Fritz. Loading K+Q+Q+R vs K on my hard disk seems >almost frivolous. Would it weaken the program if some of these _clearly always >won_ cases were not present? > >thanks, >PK This has turned out to be a fairly significant problem for Tinker on ICC. I don't have room currently for all of the 5 piece files, but do have most of the pawn combinations. What happens in Tinker's case is that the EGTB score indicating a mate is higher than the normal eval function, which still shows a clearly winning score, but not a hard "mate in n" result. What has happened is that Tinker sees the EGBT hit for one position, but not the result from a promotion a ply deeper, which then has a lower score than the egtb position, so it looks for another position and ends up drawing. I have yet to fix this. Tinker uses a signature to only probe egtb files when that exact combination of pieces is found at startup. I am thinking of checking prior plys when the piece count is <= the egtb count and if a tablebase file is not found for this ply, but was at an earlier ply, just to use the earlier score, adjusted by depth or something. Any comments welcomed (aside from just buying a larger hard disk and a CD with tablebases)...it is not the money for me, but the time to do the upgrade... Brian
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.