Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Are computers stronger at 40/2hours compared to 3min/move?

Author: andrew tanner

Date: 14:51:27 05/14/03

Go up one level in this thread


On May 14, 2003 at 14:53:58, J. C. Boco wrote:

>I've been musing about different computers, both dedicated and software for
>PC's.  I've recently had a new thought which I'm sure is old-stuff to all of you
>veterans out there.
>
>I usually play with the level setting which states the average response time for
>the computer to make one move.  For example let's say the level is "3 minutes on
>average for each move".
>
>But I could achieve the same time control by setting the computer up for 40
>moves in 2 hours, and let the computer budget its time.
>
>While I can see that the time control of 40/2hours may make the machine a little
>stronger since it can search deeper (longer) in hairy positions, just how much
>stronger will it get?
>
>My lay guess is the 40/2hours time control is not much stronger than the 1 move
>every (on average) 3 minutes.  Does anyone have any insight?
   Here is my non-programmer, winboard entusiast point-of-view. Against
grandmasters I would prefer to set it at 3 minutes per move, thereby reducing
the possibility that it would overlook a deep positional plan. Some engines have
more efficient time management schemes, where greater amounts of time are spent
on opening preperation I guess with the idea that once the engine wins a
material advantage it will need only to make several moves before consulting the
endgame tablebases which can't lose. The freeware engines in competitions don't
seem to manage their time in the opening very well (understandably). They do a
good job for the most part on time management overall, but have no way of
knowing when an extremely critical position is reached out of the opening or any
phase of the game for that matter. It is said that the great positional players
have no trouble beating computers (probably for this reason). If there was a
program that could accurately calculate the long-term consequences of it's own
positional weaknesses then maybe we would have an unbeatable player.



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.