Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Can someone make a program to ultimate perfection at 20 hrs. per move?

Author: Uri Blass

Date: 07:09:59 03/20/02

Go up one level in this thread


On March 20, 2002 at 09:42:47, Daniel Clausen wrote:

>On March 20, 2002 at 09:38:22, Uri Blass wrote:
>
>[snip]
>
>>I assume rating system when there is no maximal difference between the best and
>>the second best player
>
>Exactly my point!!!! And I think that the ELO system is NOT such a system! At
>least not in my country. Maybe someone can comment on the official FIDE system?
>
>Sargon

I know that the elo system in my country is similiar to your country but I also
know that this is only a linear approximation.


Here is the real system based on my memory:

Assume that player A has a rating of x and player B has a rating of y.

1)The player who get better result than the expected result earns a number that
is proprtional to the difference between the result that s(he) got and the
expected result.

The second player is losing the number that the first player earned.


2)The expected result of the game is assumed to be the same as the expected
result of the following game:

player A choose a number that is distibuted normally with expected value of x
and standard deviation of 200

player B choose a number that is distributed normally with expected value of y
and standard deviation of 200.

The winner is the player who choose a bigger number.

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.