Author: Omid David
Date: 11:15:29 07/08/02
Go up one level in this thread
On July 08, 2002 at 13:37:18, Jon Dart wrote: >On July 08, 2002 at 13:07:06, Omid David wrote: > >>Although the risks of using null-move pruning in the recursive way (several >>times along a variation), are not so high, I didn't get great savings (reduced >>search effort) from it. To the best of my knowledge in the eraly 1990s no >>program used recursive null-move search. What's the standing now? > >Recursive null pruning was described in an ICCA article by Donninger (author of >the program Nimzo) in 1993. He got the idea from a hint dropped by a commercial >programmer (Frans Morsch?). Since it became public it has been widely used by >amateur programs. The commercial guys still mostly don't say what they are >doing. > >--Jon My question was, how much does that recursive application of null-move contribute to the search? My experiments with fixed R=2 showed that the saving is not so high, and it's better to use R=2~3 or even R=3 instead of recursive R=2. BTW, Donninger already used "adaptive null-move pruning" in 1993 on his program Nimzo. Heinz misunderstood Donninger's 1993 article, and in fact Donninger suggested R=3 in upper parts of tree, and R=2 in lower parts, exactly as Heinz described in his 1999 article.
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.