Author: Bruce Moreland
Date: 20:36:09 02/11/99
Go up one level in this thread
On February 11, 1999 at 16:21:06, Will Singleton wrote: > >On February 11, 1999 at 12:52:27, Bruce Moreland wrote: > >> >>On February 11, 1999 at 08:23:10, Peter Fendrich wrote: >> >>>Can't check that from here, but it "must be": >>>"Null Move and Deep Search: Selective-Search Heuristics for Obtuse Chess >>>Programs." Donninger, C. (1993). >>>ICCA Journal, Vol. 16, No. 3, pp. 137-143. (A) >>> >>>I tried it but it was no hit. >>>That doesn't mean that it couldn't work for other programmers in other >>>programs... >> >>I expect that Ernst will respond, because he refined it and I think still uses >>it. >> >>bruce > > >I tried it (only allowing null-move when material imbalance exists) and found >very little change. > >Regarding your thought about threat-detection and null-move, a reading of >Anantharaman's article (ICCAJ 1991) indicates that he found little or no benefit >to this. But I remember something about you detecting mate threats with null >move. Is that right? That seems to help on some tactical tests. Sometimes I return mate scores that are outside the window, and when I find one of these I extend all of the candidates. So, in this case, if you let your opponent move when it is your move, and as a result you are instantly mated, extend everything a little, since you might end up pushing the mate over the threshold. The null-move killer position is WAC 141. It's an obvious mate that humans can see easily, and some programs can see really quickly, but a generic null-move program can't see it very well. 4r1k1/p1qr1p2/2pb1Bp1/1p5p/3P1n1R/1B3P2/PP3PK1/2Q4R w - - 0 1 The key is Qxf4, and it's a forced mate. bruce
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.