Author: Robert Hyatt
Date: 20:11:55 08/10/01
Go up one level in this thread
On August 10, 2001 at 18:04:19, Artem Pyatakov wrote: >Thank you for responding, Dr. Hyatt. > >Yes, I do implement the aging process in my hash table, so at least I am not >making an error there. > >I think the source of the problem MIGHT be that I do not "reduce" the history >values enough during my search, so when I turn on ponder, there is more >searching, so there is more history values that are too high in my search. > >I am going to do some more testing on this ponder issue. > >Artem I believe that I simply shift all the history counts right 8 bits after each search, whether I am pondering or doing a real search. The history counts are not worthless from a ponder search if the opponent makes an unexpected move, but as the game progresses, you want old history counts to shrink so new history moves can overtake them quicker when needed. >>8 is not bad, as after 4 searches a non-used history move will have a count of zero... no matter what it started out as.
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.