Author: Will Singleton
Date: 22:41:01 04/19/00
Go up one level in this thread
On April 20, 2000 at 00:41:22, Jeremiah Penery wrote: >On April 19, 2000 at 18:44:39, Mogens Larsen wrote: > >>On April 19, 2000 at 16:51:34, Jeremiah Penery wrote: >> >>>According to that page, computers will be competing in several other games, >>>including Backgammon, Othello, and Awari. To my knowledge these three games >>>(and possibly some of the other ones) have been completely solved by computers. >>>So what's the point of having a competition for them? >> >>I'm pretty sure Backgammon hasn't been solved. Difficult with the random element >>of dice I think. I have a backgammon program I can beat, unless it cheats with >>the dice off coure ;o). > >I decided to look it up, and of course you're right. None of these games has >been solved. I'm not sure where I think I read that before. :/ Would be quite a trick to solve backgammon, considering that dice are involved. :) I believe, having some limited experience, that you can't really minimax, due to the dice. Most progs use TD learning, which involves playing huge numbers of games over a long time to generate the probabilities in a given situation. So in a game, the pre-computed probabilities are used, with either no search or perhaps a two-ply deal (except for bear-off situations, in which some special strategy is employed). So matches between computers can usually be played in seconds. Will
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.