Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Proposal for a 3to1 time advantage

Author: blass uri

Date: 03:33:43 08/16/98

Go up one level in this thread



On August 15, 1998 at 13:34:31, Komputer Korner wrote:

>On August 11, 1998 at 04:13:47, David Paulowich wrote:
>
>>
>>Give the computer one hour for the first 40 moves and one hour for the
>>remainder of the game.  Triple the time limits for the human player.
>>Fairness is a nonissue here.  The goal is to allow human players to
>>perform at the highest level.
>>
>>I believe that the present FIDE policy of faster and faster games can
>>only lower the quality of chess.  We tend to forget the hard work and
>>detailed calculation behind a Frank Marshall brilliancy.  Incidentally, can
>>anyone tell me the time limits used in the Lasker-Marshall match?
>
>With a 3 to 1 time advantage and Crafty's guess rate of 50% on pondering, AND
>EVEN IF THE HUMAN GUESSES Crafty's moves 50% of the time, the time advantage
>reduces to 1.4 to 1. Even with a 1,000,000 to 1 time advantage the upper limit
>is only 1.999999 to 1. Just goes to show the power of pondering (thinking on the
>opponent's time).

I do not agree because
if I suppose 1,000,000 to 1 time advantage
if the computer does not expect the human's move the result may be
a stupid losing mistake of the computer.

The best idea in this case is to use the permanent brain for every
possibility of the human to be ready for a surprise.
if the human has 30 legal moves then I suppose the computer can use only 1/10 of
the time for the best move(it should not use the same time for every move and it
is better to use more time for good moves
but it must be prepared for every legal move).
in tactical position when the opponent has only 2 logical choices or 1 logical
choice it should be more than 1/10 and in quiet positions less than 1/10

Uri




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.