Author: Tord Romstad
Date: 06:20:53 02/11/06
Go up one level in this thread
On February 11, 2006 at 09:12:59, Steve Maughan wrote: >Here's another position where Glaurung 1.0.2 announces a mate that is shorter >than the optimal. This is a Mate in 21 yet Glaurung announces Mate in 19. > >The problem is almost certainly a hashing problem. How do you store mate values >in the hash table? Do you store "distance from root to mate" or "distance from >current position to mate"? I store "distance from root to mate". Because I don't allow transposition table cutoffs at PV nodes, this rarely causes incorrect mate announcements. The exception is endgames with lots of transpositions, like your position. This problem would be easy to fix, but I have been too lazy so far. It doesn't have any influence on playing strength, and therefore has low priority. Tord
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.