Author: Peter McKenzie
Date: 10:39:58 12/15/98
Go up one level in this thread
I do not regard using the Nunn positions as a good way to test a chess program, but then it depends on what you are trying to test. When I test my chess program, I'm trying to find out how well it plays the game of chess (!). The game of chess starts from the initial position, hence I do most alot of my testing from there :) Note that strength in playing the game of chess is a different thing to strength in analysing a given set of positions: the two are related but NOT the same. The fact that program A is better than program B at playing the Nunn positions may or may not mean that program A is likely to win a match against program B playing the game of chess (from the START position). Perhaps you are interested in which program is better at analysing positions selected at random from GM games in which case maybe the Nunn positions may be a representative set. Personally, I'm not so interested in this. The book is an integral part of a chess program, the engine programmer makes design decisions with the book in mind. Playing test games without the book and/or from the Nunn positions is like making 2 soccer teams play test games and being forced to play in 4-4-2 formation, or like making NBA teams play test games using a triangle offense. These teams may or may not have been picked/designed to play that way - they may do better using their own formation so its better just to let them play normally. Peter
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.