Author: Peter Fendrich
Date: 05:23:10 02/11/99
Go up one level in this thread
On February 11, 1999 at 02:32:14, Bruce Moreland wrote: > >On February 10, 1999 at 21:55:32, Will Singleton wrote: > >>On February 10, 1999 at 19:32:18, Bruce Moreland wrote: >> >>> >>>On February 10, 1999 at 19:00:03, Dann Corbit wrote: >>> >>>>On February 10, 1999 at 17:20:34, Dan Homan wrote: >>>>[snip] >>>>>This is a very interesting observation. Maybe it would be right to >>>>>say that null move might be a "waste" of time if the material score >>>>>is not close to beta? Taken a step further... some kind of a "windowed" >>>>>null-move might be a good idea. >>>>> >>>>>I tried this in my program. I changed my null move code so that >>>>>a null move would only be tried *only* if the material score for the side >>>>>to move was greater than (beta-1.5*pawn_value). This reduced my >>>>>searched nodes from 5-10% in most of the positions I have tried so >>>>>far. (I have only tried a handful of WAC positions so far.) >>>>> >>>>>If this pans out, It is a perfectly safe enhancement to >>>>>null-move. The 1.5 pawn window that I tried was arbitrary; >>>>>another value might do better. >>>>What an incredibly creative idea! >>> >>>Not a new idea ;-) >>> >>>bruce >>> >> >>Implying that you either use something like this or have found it to be not >>worthwhile. If you use it, what window value do you use, and does it change >>depending on the situation? > >I think the idea is a good idea. I couldn't make it work in mine, meaning that >it didn't make anything go faster. I know I've tried it. I don't think I >invented it. > >The idea is obviously that if you are obviously hosed based upon some quick >eval, that you probably aren't going to come back with a fail high if you give >the opponent a free move. > >Another idea for you guys, which I didn't invent, tried, and don't use, is >widening the null move window in the downward direction in some circumstances. > >You not only check for fail high, and cut off if you get it, you check for fail >low. A fail low indicates that the opponent has a massive threat. In this case >you can extend. > >This idea was mentioned in the original Donninger paper on null-move, ICCAJ, >don't know what issue, but it was between 1990 and 1995. > >bruce 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... //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.