Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: High branching factor games

Author: Bob Durrett

Date: 06:00:48 11/27/02

Go up one level in this thread


On November 27, 2002 at 00:34:45, Russell Reagan wrote:

>While I was at work tonight sitting at a computer, having nothing to do, I
>opened up the calculator program and started playing around with the branching
>factor for the game of go. I quickly realized that with such a large branching
>factor at each level (361 from the starting position), it's no wonder computer
>go programs play well below master level.
>
>Are there any games that have a high branching factor that computers are good
>at? The only games that I'm somewhat familiar with are chess, checkers, and go.
>Of those, it seems that the branching factor of the game largely determines how
>well computers play them. It seems that checkers with the lowest branching
>factor gives humans the hardest time, then chess, and then go with a much larger
>branching factor is not even in the same ballpark with the top humans.
>
>Does this relationship between branching factor and computer strength hold for
>other games? Or do some games break the pattern? If some games do break the
>pattern, why? Simplicity of the game? Magical shortcuts?
>
>Russell

My gut feel is that the use of search algorithms is inappropriate for such
games.  There must be some other algorithms, which are not search algorithms,
which would do a reasonable job.  It only remains for someone to identify them
and write a program using them.

This is a golden opportunity for some programmer to become famous!  He/she would
be known as "the programmer who cracked open the game of GO."  : )

Bob D.



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.