Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site utcsri.UUCP
Path: utzoo!utcsri!coatta
From: coatta@utcsri.UUCP (Terry Coatta)
Newsgroups: net.startrek
Subject: Re: Chess, Kirk and Spock
Message-ID: <2436@utcsri.UUCP>
Date: Mon, 31-Mar-86 13:07:33 EST
Article-I.D.: utcsri.2436
Posted: Mon Mar 31 13:07:33 1986
Date-Received: Mon, 31-Mar-86 15:41:25 EST
References: <240@copper.UUCP>
Reply-To: coatta@utcsri.UUCP (Terry Coatta)
Organization: CSRI, University of Toronto
Lines: 25
Keywords: Irrational? Moves
Summary: 

In article <240@copper.UUCP> toddv@copper.UUCP (Todd Vierheller) writes:
>  +
>  +The subset is based on move strength, not on "rationality index".
>  +
> 
>Move strength is certainly isn't based on irrationality.  One expects the
>opposing player to do what is best in the short term as well as the long term.
>(Ie. He won't trade both his bishops to take my rook.)
>
>  +						+Pete Williamson
 
 
The subset is based on move strength as calculated at that point in the
game.  For the ranking to work well, the evaluation function must perform
well -- that is it must look at a board poistion and make a good "guess"
as to whether that board position can lead to a win.  It is quite possible
for the evaluation function to give a high ranking to a bad board position
if the the reason the board position is bad is that a fairly complicated
set of moves is required for the opposing player to take advantage of it.
The only way to play chess perfectly is to maintain the entire game tree, if
you don't there is always the possibility that way far down one of the 
branches you have pruned off there is a win lurking for your opponent.