Author: Arturo Ochoa
Date: 18:35:56 10/04/00
Go up one level in this thread
On October 04, 2000 at 19:54:27, Joshua Lee wrote: >But can you guess a rating from it? if not is maybe an old test like BS-2450 >better or is that too difficult for programs that don't make the LCT2 cut? Hello: I prefer use the tactical exercises for human players that I can get in several Chess Magazines or the Chess Informator. By this way, I can learn more about tactics and the same time, you can test the chess programs by tactical theme. Generally, the Chess Magazines classify the tactical positions by theme. Apart of the symbolic calculated Elo of the Test Suites, the most of these don't classify the tactical themes. I cannot conclude in advance if a Chess Program is bad or good tactically using the most known Test Suites. Why? Because you don't have any secure way to establish wihout objections in which tactical themes the Chess Program is failing. You need about forty positions (my own experience) by tactical theme, so you can conclude with a good grade of certainty if the tested Chess Program is really failing in a specific tactical theme. Another consideration: you should only choose the tactical positions whose solution variation has an unique tactical theme. Bes Regards, Arturo.
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.