Author: Vincent Diepeveen
Date: 07:53:43 01/25/01
Go up one level in this thread
On January 25, 2001 at 10:30:37, Gian-Carlo Pascutto wrote: >On January 25, 2001 at 09:34:05, Robert Hyatt wrote: > >>This is what null-move search does, in essence... > >Mnmnm...yes...but... > >Nullmove does more than just reducing depth, it also >does the nullmove! This makes it a lot unsafer than >just reducing the depth. > >If you had said Fail High Reductions I would have >agreed, but I don't think you can really call >nullmove a negative extension. Sorry let's discriminate between nullmove and FHR. What FHR is doing is very incorrect, what nullmove does is correct, though for some manoeuvres you might need a bit bigger depth. Nullmove assumes the worst thing you can do in a position is doing nothing. That's a far stronger assumption as forward pruning based upon evaluation and a few static threat detectors only, even weaker (and incorrect if you use hashtables) is what FHR is doing, just reducing based upon eval versus beta comparision and some other noise. Vincent >-- >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.