Author: Peter Fendrich
Date: 12:22:10 01/18/00
Go up one level in this thread
On January 18, 2000 at 06:23:34, Vincent Diepeveen wrote: >On January 17, 2000 at 16:47:24, Dann Corbit wrote: > >>I have a theory that NULL move hurts at very, very long calculation times. >> >>The reason I suspect that it is true is that if one position in ten thousand is >>a zugzwang type and you consider a billion positions, a lot of them will be NULL >>moves. Most of them will be the kind that really should be avoided (maybe 99% >>or even 99.9%) but the 0.1-1% that really should be responded to get ignored. >>When only a few million nodes are searched, I doubt if there is a problem except >>in rare circumstances. >> >>Opinions? Am I all wet? > >You're not correct. double nullmove detects zugzwang. >In rook endings typical sometimes a single zugzwang is the problem. >Therefore double nullmove is a great invention. Crafty fixes it by >turning nullmove off when there is 1 piece left. That is a hard fix. >double nullmove sees zugzwang also when there are n pieces left. > >Vincent. How much do you estimate that double nullmoves cost you in increased number of nodes? I'm not ready to pay much in order to catch "all" zugswang positions. //Peter
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.