Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Normal Distribution & Computer Chess

Author: Marc Boulé

Date: 07:17:44 06/29/02

Go up one level in this thread


On June 28, 2002 at 19:29:38, Robert Henry Durrett wrote:

>DEFINITION:  Apologetic SPAM:
>
>All discussions of spam  [a canned version of ham?] and a variety of associated
>apologies.
>
>_ _ _ _ _ _
>
>These fall into a normal distribution where the independent variable is time.
>
>Or is it a different statistical distribution?  An "abnormal" distribution?
>
>Are there any spamish computer chess statisticians here?
>
>:) :(     :) :(     :) :(     :) :(     :) :(     :) :(
>
>_ _ _ _ _ _
>
>
>Bob D.

Elo's assumption was that a player's strength distribution follows the normal
distribution (the independant variable is the player's rating). However,
Glickman reports that using an extreme value distribution might be more
realistic. The extreme value distribution is biased to one side. Anyways, take a
look at his papers at:
http://math.bu.edu/INDIVIDUAL/mg/research.html

Marc



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.