Author: Steve Glanzfeld
Date: 09:05:52 04/25/05
Go up one level in this thread
The observations from these games are important. It is obvious, that programs should use a effective "no progress" rule. For example, when they evaluate positively form themselves over a LONG sequence of moves, but the eval doesn't get significantly better, they must consider "unusual" moves which change the situation dramatically and let expect progress within small depths (as long as the eval is still positive for the program, like RxNf1 in your example). Progs really look stupid in these time losses... Steve
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.