Newsgroups: comp.ai
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!wuarchive!m.cs.uiuc.edu!ibma0.cs.uiuc.edu!sunc4.cs.uiuc.edu!epstein
From: epstein@sunc4.cs.uiuc.edu (Milt Epstein)
Subject: Re: Another chess question
Sender: news@ibma0.cs.uiuc.edu
Message-ID: <27F3C8DE.37CB@ibma0.cs.uiuc.edu>
Date: Fri, 29 Mar 1991 23:08:14 GMT
References: <1991Mar26.162003.7849@swift.cs.tcd.ie>
Organization: University of Illinois at Urbana-Champaign
Lines: 38

In <1991Mar26.162003.7849@swift.cs.tcd.ie> jeclarke@swift.cs.tcd.ie writes:

>	I remember reading somewhere (possibly years ago) that chess
>	Grandmasters did not gain their advantage over lesser players
>	by being able to look more moves ahead than them (I think they
>	used only look 3 or 4 moves ahead), but by the fact that they
>	could eliminate all the "useless" moves from any position, and
>	so have more time to concentrate on possibly useful moves.

This is interesting, but it raises some other questions.  If the
experts weren't looking further ahead than novices, in what sense were
they "concentrating" more on these possibly useful moves?  For
example, let's try to characterize the process as a search (which
seems a safe assumption).  Let's say D is the depth of lookahead and B
is the breadth of moves considered at each level.  Were experts and
novices searching the same total number of nodes (B*D)?  If so, does
this mean that the experts wouldn't miss moves that lesser players
would (i.e. they are better at choosing the B moves to consider)?


>	I seem to remember that this process was so unconcious that when
>	these useless moves were pointed out to them they didn't reply
>	that they had seen them and realised that they were no good, but
>	that those moves had never even occured to them.

If this is true (i.e. that experts say these moves did not occur to
them), then how is it possible to conclude that they eliminated these
useless moves from consideration, or that they ever considered them at
all? 

Maybe if you track down this study, it will shed some light on these
questions. 

-- 
Milt Epstein
Department of Computer Science
University of Illinois
epstein@cs.uiuc.edu