[ExI] Computer Plays "Perfect Checkers"?
russell.wallace at gmail.com
Sun Jul 22 23:18:21 UTC 2007
On 7/23/07, Lee Corbin <lcorbin at rawbw.com> wrote:
> Perhaps they've simply proved somehow that the program
> can and will reduce *any* position (starting from the initial position
> itself playing black or white) to one of the ten-piece winning/drawing
I believe they have. I looked at the interactive search tree on the web site
(which allows you to expand the tree node by node and shows the minimax
value of each node where known), and although many nodes are marked unknown
(presumably as a result of optimizations such as alpha-beta search), it
seems that right from the start the program always has access to at least
one move whose minimax value is known to be no worse than a draw.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the extropy-chat