Author: Tord Romstad
Date: 05:05:47 10/04/05
Go up one level in this thread
On October 04, 2005 at 07:17:37, Ross Boyd wrote: >I like the idea of making a sparring partner that rewards a certain skepticism. >If there's anything that bugs me when playing computers its that horrible >inevitable feeling that the computer never misses anything (tactically). > >Nowadays when I play a computer I like playing the older weaker dedicated units >because they allow me to play a more natural game. Against tough computer >opponents I tend to play anti-computer stuff which ruins the fun completely. I agree. What I want to try is to prune moves semi-randomly. Any move can theoretically be pruned, but the probability of pruning will be bigger for moves which tend to be difficult to see for humans. For instance, I want the probability to be higher deep in the tree than close to the root, and to be big for moves with negative SEE values. Long moves should be pruned more often than short moves. The computer will still be tactically strong, but occasionally play incorrect combinations (because it overlooks some tricky defensive move) or allow some tactical shot for the opponent. Experimenting with different types of "blunder styles" could also be interesting. You could make an optimistic style by pruning more moves for the opponent than for the program, or a pessimistic style by pruning more moves for the program and few for the opponent. Tord
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.