Author: Carmelo Calzerano
Date: 10:09:41 02/02/01
Go up one level in this thread
On February 02, 2001 at 12:41:09, Andreas Herrmann wrote: >On February 02, 2001 at 09:51:17, Carmelo Calzerano wrote: > >>On February 02, 2001 at 06:44:47, Andreas Herrmann wrote: >> >>>If i work with exact- and upper bounds, sometimes my program Holmes spends the >>>opponent a figure. >> >>Do you mean "gives pieces away", I guess? >> >>>If i only use lower bounds, this never happens. >>>What could be the reason for this? >> >>A bug, of course... :) >> >>Are you sure you store/retrieve the depth (draft) correctly? Maybe you store >>wrong drafts in the hash table, or retrieve them incorrectly... This could >>happen in case of search extensions or razoring, for example. >>Just an idea... >>:-) >> > >Thanks for your tip, but that's not the problem, because such things i debugged >very often in the past. >My program checks the hash entry. I generate all possible moves and then i look >if the move in the hash is a legal move, if not i don't use the hash entry, and >i have a hash kollision. Normally i have about 0,1 to 0,3 percent of those hash >kollisions. I'm not sure you got my point: I was referring to bugs in HT _draft_ storing/retrieving, not to hash collisions. Ie, if you sometimes incorrectly take into account an HT hit, when the current search depth is bigger than the stored position depth. It was just a really simple tip, anyway :) Bye, Carmelo
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.