Author: Robert Hyatt
Date: 15:29:48 03/16/03
Go up one level in this thread
On March 16, 2003 at 17:59:50, Martin Bauer wrote: >On March 16, 2003 at 13:32:36, Robert Hyatt wrote: > >>Search the first move with the window alpha,beta. >> >>Search the remainder of the moves with alpha-X, beta-X. >[...] > >Hello Mr. Hyatt, > >thank you very much for the explanation. I think I have to try. But may be you >can help me with your experience: > >Is it worth to do all this tricks that blow up and slow down my code only to do >singular extension? That's not easy to answer. In Cray Blitz, it seemed to work well. But when I tried it in Crafty, multiple times even, it has always seemed to be worse, or at least no better. That's why you see comments about SE in Crafty, but no SE code is present. I'm not sure whether the null-move stuff (R=2/3) causes SE to perform worse or not. But, in the case of Crafty, it hasn't worked well yet. >Is singular extension so powerful, that this helps me improving my engine? >Single response extension is more easy to implement and not so expensive. I >think single response you can call a weaker or more spezial form of singular >extension. How much better is singular extension compared to single response? > This is very hard to answer also. SE will make your program spot tactics quicker, at the cost of at least a ply of search depth, which means your _positional_ skills will be a bit weaker due to that lost ply. If you believe you are losing more games by missing tactical shots that your opponent finds against you, SE will help prevent many of those. But if you believe that most of your losses are due to things other than tactical shots, then SE won't help, and will likely hurt due to the loss of at least a ply, maybe even two. >Any experience? > >Thanks > >Martin
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.