Author: martin fierz
Date: 06:13:42 08/12/03
Go up one level in this thread
On August 12, 2003 at 08:00:42, Lars Bremer wrote: > >>the chinook team has completed the 10pc endgame database early this year, and >>with that, chinook will be the strongest program there is. but it's not public >>of course. > >Hi Martin, > >this would mean a bigger endgame database is more important than a deep opening >book. Or, in other words, the computed books of nemesis, cake and kingsrow are >not perfect. Are there games with endgame mistakes to prove that? i don't know. i doubt that any of the 3 programs above would lose a game against a 10pc chinook if they were allowed to use their opening books and it was regular 3-move-checkers. but that doesn't mean that a 10pc program isn't better: see for example my page www.fierz.ch/cake.htm and take a look at test position 1. it is a nice example for how the 8pc database makes a program understand something in a second which it doesn't understand for hours (or maybe ever) without that database. it will be the same with the 10pc db: a program with that db will be able to say that a position is a win or a draw in a second where the 8pc programs won't have a clue. perhaps they will play all the right moves, but they won't understand. it's a bit like with connect 4 programs: given enough time, any good connect 4 program will see that the player to start wins. but a program that sees that in a few minutes is better than one that sees it in a few hours (of course, you should take an average over many test positions...). cheers martin
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.