Author: Walter Faxon
Date: 19:18:21 12/02/02
Go up one level in this thread
On December 02, 2002 at 04:41:25, Ingo Lindam wrote: >Hello, > >is there any engine that is forced to motivate (or showing to be reasonable) a >big winning score (e.g. +- 10.0) by giving atleast a single variation that leads >to a mate? > >This variation doesn't have to be PV with oponent defending best. I rather think >of a way to win the game against a no defending oponent just making null moves. >This should be first step into direction of creating plans.... >(ofcourse just first step)... and making things like in the following example >impossible: > <snip...> > >Internette Gruesse, >Ingo Hello, Ingo. I don't know if this is responsive to your question, but long ago I had an idea about planning vs. "minimal cooperation", where the opponent makes "nothing" moves (generic "developing" moves or, today, null moves), except when faced with a direct threat. (The greater the depth of the threat being responded to, the less the "cooperation".) The idea is that if your plan doesn't work against minimal cooperation then that's a good sign that it won't work at all. Of course, real chess is a lot more complicated. You can plan against your opponent's plan. Individual moves are often the basis of several potential plans. Remember Vukovic's maxim from "The Art of Attack in Chess": "The maximum of preparation with the minimum of commitment." And you have what we might call "meta-plans" which involve the coordination of multiple simple plans. The whole area of planning applied to computer chess is still in its infancy. -- Walter
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.