Author: Sune Fischer
Date: 02:13:11 06/11/02
Go up one level in this thread
On June 10, 2002 at 21:14:05, Russell Reagan wrote: >On June 10, 2002 at 09:54:24, James T. Walker wrote: > >>They would have to see 100 plies to see the draw (50 move rule). > >I don't think so. I thought of a way to detect this in a matter of minutes. Of >course I have many more important things to worry about than draw detection at >this point, so that will be implemented WAY down the road sometime. > >Russell We should collect a testset of these draw positions, and then in time design a mechanism for detecting them. Not an easy task, but has anyone ever attempted? It shouldn't be something that takes minutes to detect, you'll have to see it at the leaf or close to, detecting it when its the root position is not worth a lot. One way to start would be to see how many blocked pawns there are, then only evaluate the pieces that can penetrate thru the open files. Detecting _all_ fortesses would be impossible, but a minor subset of them would be very nice. -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.