Author: Dan Andersson
Date: 15:30:24 01/10/02
Go up one level in this thread
My main point is that the search and evaluator must work together. In mid level chess programs there are often jarring jumps in evaluation between moves. That does not happen often in games between top programs. Only once or twice in the positions when the tactical break of the game occurs. With Gambit Tiger it also happens in the transition between attack and endgame. If you wish to attribute that to inferior evaluators, so be it. But it is damn hard to retrofit a reasonably mature program with substantial amounts of new knowledge. It more often than not make the program slower and weaker. But adding higher order smoothing functions might be acceptable. Working on reducing a program search graph also imparts new knowledge into the program. MvH Dan Andersson
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.