Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What's Fritz's IQ?

Author: Paul Massie

Date: 10:43:38 12/27/01

Go up one level in this thread


It's my belief that humans operate largely by a stored database of positions,
with associated good and bad plans and moves.  This belief is based both on my
experience at becoming somewhat stronger, and also by the rather limited
research that has been done in this area.  That suggests that weaker players
spend a lot of time looking for the right move, whereas stronger players seem to
use memory a lot more.  The result is that every position is new to a weak
player, so he/she has to grope for the right plan and move with little
background.  The strong player (IM/GM) appears to have a large database of
stored positions that allows him/her to almost instantly find something similar
and thus be able to look at only a few alternatives.  This dramatically reduces
the search tree and thus the NPS requirement.  Humans seem to have an
extraordinary ability to remember a huge number of positions, along with what
moves and plans did and did not work in each.  This is most useful because
humans can also extrapolate from a position to a "similar" position, and thus
attempt to apply the known plan.

This explains why a very strong GM may suddenly play quite weakly in a
particular game or position, if that position is something they're not familiar
with.  As players get stronger they seem to develop a larger database of
positions, as well as understanding each position in the database better.

I find this concept a little depressing, since it seems to reduce the game of
chess to little more than a massive memory exercise.  Also, I have no idea how
one would program something like this, or even if it is a good approach.

Paul



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.