Author: leonid
Date: 10:45:12 07/21/99
Go up one level in this thread
On July 20, 1999 at 12:12:14, Dann Corbit wrote: >I think it would be interesting to benchmark chess algorithms: From my point of view basic power of chess engin should be done in two ways: 1) Find the efficiency of moves generation of two engins to compare. 2) Compared time consumed for solving by "brute force" same position in equal number of plys. First, ideal way for finding the efficency of the generation of moves should be done when all the extra methods like alpha-beta, hash tables and & are disconnected. Two game must find the move for given number of plys. In reality you can almost never do this since you can't reach the inner part of somebodys game. But exactly this comparaison is the most outspoken for the potential force of every game in sight. Second way is less vociferous but more paractical and simple to execute. You must ask two games to solve some positions by "brute force" in given number of plys. You must only be sure (as much as possible) that parameters, like hash tables and so like, are the same or disconnected for both games. Leonid.
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.