Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Board games and mathematical complexity: a poll

Author: William H Rogers

Date: 10:12:59 12/11/02

Go up one level in this thread


On December 11, 2002 at 12:00:12, Edward Seid wrote:

>Western Chess
>Shogi
>Go
>Xiangqi (Chinese Chess)
>Othello

Based upon the strength of the programs written for the above games I can only
comment about three of them. Go, in my opinion and the opinion of many others is
the most complecated, followed by chess. Othello, on the other hand depends on
the roll of the dice which includes 'chance' and there for should not be compare
to the others. There are both Othello and Checkers programs that have defeated
the worlds champions many different times, so I would conclude that they have
been solved.
Chess on the other hand with the way the programmers are working today might not
really be solved, but for all pratical purposes will be using the technics that
are in vogue today, ie. unlimited opening books, unlimited end game books, might
be consider to be solved in just a few more years.
I still recall a tournement held in 2000 where one program defeated another
without ever leaving its opening book. I do not consider that a chess thinking
machine or program, just a look-up program that could not solve the problem
without is extensive books.
Many people have argued about this topic for years now but I have to ask "what
is considered an opening play, the first 5, 10, or 15 moves", "Where does
mid-game start", and "when does end game play start"? Table bases for the end
games where there are only 5 or less pieces left on the board sounds reasonable,
but I think that if anyone ever programs to find the best endgame without table
bases then it will be a better program.
I would like to see a world wide tournement held which limits the pre programmed
openings to say 10 moves, period and then let the programs proceed based upon
their own 'real' strenght.
The people who oppose my operating theory the most at this time are the ones who
have already incorporated the above mentioned ideas into their programs.
Only time will tell if and when the games are really solved depending upon what
king of programming is allowed.
Bill



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.