Author: Miguel A. Ballicora
Date: 06:42:36 12/30/01
Go up one level in this thread
On December 30, 2001 at 08:43:58, Gian-Carlo Pascutto wrote: >On December 30, 2001 at 08:28:35, Thomas Mayer wrote: > >>When I get a tablebasehit, I update the age-counter of that entry to current >>age-counter to prevent that to much is overwritten in the primary table. > >I presume you mean transposition table hit? > >Basically you prevent the entries you have used from being >replaced. Sounds like a good idea, but I wonder about >the effect of not having any move ordering info for the >rest of the tree. I guess that it should not be a problem because the table carries from the previous search not only the information for the PV but also the information for the rest of the tree. I just have an idea (perhaps it is more of a rethoric question). What would happen if you erase, before the search, the scores from the PV entries (ONLY for these handful of few entries) to make them unusable "as a transposition hit". In other words, make them for instance (-infinite, LOWER). In that way you keep the information for move ordering but disallow a quick hit and force the engine to search the rest of the tree. If the rest of the tree has all the info it will return quick, otherwise it will search the holes. I think I will try it at one point, let me know if you do. Regards, Miguel > >-- >GCP
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.