Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Why use opening books in machine-machine competitions?

Author: Pete Rihaczek

Date: 11:49:52 11/25/03

Go up one level in this thread


For machine-machine games, if you wish to eliminate the influence of different
opening books, you can either eliminate them as you suggest, or else everyone
agrees to use the same book. If you eliminate them entirely I submit that the
games will look more computerish, going into lines known to be bad, or else
going into the same lines over and over. Pair two programs without an opening
book, and they will likely play the same game over and over. So it seems keeping
the book is the better choice. But then you may have complaints that the book
leads to positions more favorable to one program or another, etc., and if that
happens you have to throw up your hands and let people make their own books.
This is a competition, after all. If the goal is 'may the best machine win',
then having a better opening book is no different than having better hardware,
or more efficient algorithms. It seems to be a 'may the best machine win'
competition, since Brutus for example is a different animal than the others,
Diep is running on 500 processors, etc. People can do their engine vs. engine
testing at home, a world championship should be more interesting, and allow
whatever the competitors can put together to beat the rest.

The same idea carries over into man-machine competitions. We want the strongest
human vs. the strongest machine, regardless of the form the machine takes. The
whole interest of the event is in the differences both entities bring to the
table. The man is intelligent, the machine is no more intelligent than the desk
it sits on. Because he is an intelligent being, the man can use his knowledge of
what the machine is to his advantage. Kasparov made Fritz look worthless in a
closed position, and he well knew the reasons the machine would have difficulty.
An intelligent being has a whole dimension of advantages over an inanimate
object. If that is fair, then anything the computer does to win is fair as well.
Endgame tablebases allow superhuman knowledge of the endgame, opening books
allow superhuman knowledge of the opening. Neither one should be handicapped,
nor should the human be allowed access to either of those, because the machines
have not proven themselves dominant yet. Computers have not shown themselves to
be the equal of humans in chess understanding, Kasparov made that very clear in
the match. Until computers are acknowledged to have superior insight into the
game, it is premature to start offering the humans a leg up.




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.