Author: Fernando Villegas
Date: 15:03:33 05/07/98
Air mail Special for Thosrten: In relation with what has been said here in order to strengthen CSTAL in such a manner as giving him the capability to win won games that sometimes he loses, let me call your attention to page 8 of the old book “More Chess and Computers” by David Levy and Monroe Newborn. There you an read: "...Cheops would be used by an improved version of the Greenblat program -Machack- in the following way: whenever the program reached a position it considered strategically satisfactory, Cheops would take a look at the further tactical possibilities. This enables the program to avoid numerous traps...” Levy played two games against this comb in 1978 and won both, but this Janus-kind of monster did pretty well considering the level of chess computer in those days. So, this is a somewhat old approach that probably was abandoned somewhere in the road when stronger computers and programs appeared and the hope to get it all with one engine simply vanished the idea of a combined entity. But now we know better. I wonder if Chris, in the meantime he develops a new CSTAL engine capable of transforming himself according to situation, could not produce a comb as a transient experiment. I am sure he can put in code in half a week at least a reasonably strong conventional chess engine to do the work Cheops did for Machack. I believe that not even a top notch engine is necessary for this testing of tactical lines in complex positions. I intend to do myself the experiment giving to a old program like Fritz 1 the winning positions CSTAL did not win. Ciao Fernando
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.