Author: Martin Giepmans
Date: 03:54:20 08/09/01
Go up one level in this thread
On August 08, 2001 at 20:31:49, TEERAPONG TOVIRAT wrote: > > >>Well yes, but is the tree really smaller? In the endgame the program with "old >>cutoffs" is sometimes faster, sometimes slower. Overall the gain is not clear. >>In the middlegame there is usually no significant difference, but in some of my >>testpositions the old stuff slows the program down and the output shows >>instability. Too much mustard on the hotdogs perhaps ... >>Of course a lot depends on the environment (eval etc.). Maybe program A benifits >>from old cutoffs and program B doesn't benifit. That's always the problem. >> >>Martin > > >Hi, > >Let me ask you some questions here. > >1.What does your "old cutoffs" mean? >Do you mean any old score from hashtable or only >draw score? > >2.How you define internal,external repetition? > >Teerapong "old cutoffs": my shorthand for "cutoffs based on old scores". I call scores "old" if they are stored in the hashtable during a previous search (before the opponent moved). Internal repetition (I don't know if this is the "official" name): a position in the searchtree occurs deeper in the tree (on the same branch). External repetition: a position outside the tree (that has occurred during actual play) occurs in the tree. And thanks for your earlier post! It's interesting stuff. Martin
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.