Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: NO MORE human vs computer game again

Author: Dieter Buerssner

Date: 16:33:03 08/04/03

Go up one level in this thread


On August 04, 2003 at 19:13:35, Dann Corbit wrote:

>This is a better approximation, but the tail will never be zero.

Sure. I wanted to emphasize more, that we have to trust the Elo-statistics. It
is heuristics, not based on a theoretical understanding, of how game results are
distributed. I doubt, that the model will be true in the tails of the
distribution (neither do I think, that this is very important. But one should
keep the point in mind, when evaluating the statistics for such extreme cases).

>However, if someone played 1e1000 games, against someone with a +500 ELO rating,
>I suspect that somewhere in the millions of points that they did capture, there
>would be a win.

For a human, it will not be possible to play millions of games. If we play 5
games every day, for 50 years, it will still be less than 100000 games. The only
cause of me winning vs Kramnik (if we played 1 game each day) would be, some
severe health problem of Kramnik. Otherwise, the chance will be zero. If the
chance is something like 0.01 % during my life - well, there we are at the
limits of statistics.

Regards,
Dieter

PS. I chose Kramnik as an example, because he is younger than me. So, probably,
I will suffer from Alzheimer earlier than him ...





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.