Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What rating has a compterprogram 2010?

Author: KarinsDad

Date: 11:49:24 04/12/00

Go up one level in this thread


On April 12, 2000 at 10:55:41, Graham Laight wrote:

[snip]
>
>I suspect that you'd eventually get to a point where nearly all games are draws.
>
>Could somebody with Chessbase please create a table of draw ratios against elo
>rating?
>
>For example, based PURELY on guesswork, it might look as follows:
>
>Elo Rating  Percentage Of Draws Against Players Of Similar Rating
>==========  =====================================================
>
>1000        5
>1500        15
>2000        25
>2500        50
>2800        60
>
>-g

Actually, I think these percentages are high, but based on your chart:
2900        64
3000        66
3100        67
3200        67.4

etc.

The only way to get to a 100% draw capability is:

1) IF the game is drawn with best play.
2) IF a pool of computers could play best play.

Otherwise, you will not get there.

Also, I have the impression (but have no data to back it up) that computers draw
a LOT less against each other than human GMs draw against each other.

Does anybody have any statistics on this?

If this is the case, there are probably a lot of "human psychology" factors
which result in more human draws (e.g. lack of desire, inability to see the
winning path, sufficient results for this portion of a tournament, etc.).

KarinsDad :)



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.