Author: Uri Blass
Date: 03:00:05 05/20/02
Go up one level in this thread
On May 20, 2002 at 05:49:02, Matthias Gemuh wrote: >On May 20, 2002 at 04:54:20, Uri Blass wrote: > >>On May 20, 2002 at 04:46:27, Pham Minh Tri wrote: >> >>>On May 20, 2002 at 04:04:54, Matthias Gemuh wrote: >>> >>>> >>>>Hi Experts, >>>> >>>>how can I prove that my nullmove implementation is correct? Or that nullmove >>>>helps? >>> >>>It is correct if your new engine: >>>- searchs more depths. Your engine should search deeper 1-3 depths with same >>>time as before. >>>- has same results with test sets (or may be very slighly different). >> >>I think the result in test suites should be better. >> >>At least my experience says that null move pruning improved the results in test >>suites and not only in games. >> >>Uri > > > >Aren't there some "magic" positions which demonstrate the strength of nullmoves >:-)? > >Matthias. You can try the following position: r1b1r1k1/p1p3pp/2p2n2/2bp4/5P2/3BBQPq/PPPK3P/R4N1R b - - 0 1 My program solved it at the same depth with and without null move pruning but null move helped it to find the move faster. It was not a very big improvement. There were also positions when my program needed more time but the total result is better. Uri
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.