Author: Sune Fischer
Date: 03:31:58 03/16/04
Go up one level in this thread
On March 15, 2004 at 18:44:39, Steven Edwards wrote: >On March 15, 2004 at 05:48:05, Sune Fischer wrote: >>On March 15, 2004 at 05:06:09, Steven Edwards wrote: > >>Wouldn't it be possible to construct a simple testbed to quickly try out a few >>of the ideas and see if they have merrit? > >No. There is too much foundation work needed to get even a simple plann >exercise to work. It should be possible to pick a position, analyze the kind of features that is needed to find the best move and try it out. Surely if it can't be made to work on a single position there is not much hope of getting it to work on all positions. >>How does this work if winning a piece requires a 4 move combination? > >Inital pattern matching sugeests the possibility of a specific successful attack >further pattern matching refines the idea and fills in details; a plan is >constructed to persue the idea; You violate my assumption :) Suppose this is pure tactics, no logic behind it, the kind of move no human would think of. For a static planning detector this is not going to be easy to solve I think. The biggest problem is going to be in the area of unclear lines, they look neither good or bad but the way I understand it they do not trigger by default. I think if it was me, I would try the whole thing in reverse. Ie. instead of only searching down prospective lines and having to code an infinite number of static attack plans, one could search down all lines by default as a regular chess program and only prune those lines which the static detector indicates is losing strategies. So the points of doing it in reverse is: 1) rules can be added one by one and tested seperatly 2) by default you have a 'sound' program with limited coding 3) the end result should be the same -S.
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.