Author: Vincent Diepeveen
Date: 08:43:26 12/10/99
Go up one level in this thread
On December 10, 1999 at 11:00:17, Bert Andersson wrote: >Is there anyone interested in trying to describe our own thinking when playing >chess? In a language for humans, far away from computers :-) > >I have done some work on this and it staggeringly difficult at first. I had to >define words etc. to get a clear picture of it. And it was hierarchical levels >involved etc. But all in all I got some very good startingpoints for programming >a knowledgebased approach in a small slice of the chesspace! > >Currently I'm working on programming, which is slow, also because I want to put >as much knowledge as possible from the beginning. But it is interesting as the >work gets very broad. I can use Words, graphics and all kinds of sortingprograms >as well as CASE's, which I haven't got into much because of their difficulty. >Any ideas? > >I just remember that it exist a book on this topic, from -50ies ?!? >Well, I don't have it in memory. Anyone who has? Note that when getting this position diep didn't search very long yet, so i analyzed by hand. It took some time to realize, but after a while i realized that white wins when he can get a rook to f1 and a rook to the 7th rank. black is in zugzwang then and has to allow doubling rooks on f1-f7 after which rf8 mates. the main problem is of course to get the rooks position like that without black being able to just give away its queen or give a perpetual check. obviously a program solving this position with egtb will simply show mate in xxx. DIEP solves it fullwidth searching not very quickly, but with +20.0 at 15 ply. with nullmove turned on diep is currently searching in the states somewhere with a lot of hashtables. Will get that log after the weekend i guess. Vincent
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.