Author: Pallav Nawani
Date: 12:36:50 08/04/04
Go up one level in this thread
On August 04, 2004 at 15:07:46, Tord Romstad wrote: >On August 04, 2004 at 14:53:03, Pallav Nawani wrote: > >>Where pruning really shines is in endgame. I get 2-5 extra plies because of the >>(heavy?) pruning I do, some times even more. >> >>Pruning hurts tactical strength a bit, especially for a simple program like >>mine. > >Funny. This is almost exactly the opposite of my experience. > >In the endgame, I haven't been able to find any kind of pruning which works >(except plain recursive null-move pruning with a small reduction factor). All >the many pruning techniques I use in the middle game are ineffective in the >endgame. I might get a few plies extra if I switch on various kinds of >pruning, but it introduces too many mistakes. Heavy pruning in the endgame >hurts in test suites as well as in games. > >I haven't found tactics to be a problem, though. Heavy pruning greatly >improves the tactical strength of my engine. > >Do you use the trick of avoiding most kinds of pruning when there has >been an extension at one of the last few plies in the current line? > >Tord I avoid all pruning when the current move/node has been extended, but I don't worry about previous extensions. I also allow one or two moves on the current node before I start pruning. I implemented pruning _before_ I implemented NULL move pruning, and my observations and based upon that. Should have mentioned this before :) I prune everywhere, and indiscriminately at that :) probably that's what hurts my engine's tactical strength. Only 240/300 on Wac in 5sec/position on a celeron 333 :( But there would be other things, I am sure. What kind of pruning that you do improves tactical strength of your engine? Regards, Pallav
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.