Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Performance rating calculation

Author: Peter Fendrich

Date: 16:09:28 06/02/04

Go up one level in this thread


On June 02, 2004 at 19:04:15, Peter Fendrich wrote:

>On June 02, 2004 at 17:16:15, Dan Wulff wrote:
>
>>Hi!
>>
>>I've been thinking about using performance rating to select moves for Gandalf's
>>opening book, but I cannot seem to find the correct way to calculate the
>>performance rating od a given move.
>>
>>Say a 2600 player wins a game against an 2500 opponent, starting with 1. e4,
>>what would the performance rating be after that ??
>>
>>I've been looking at the ratings in the CB GUI, but sometimees, when only one
>>game has been played, the Performance is over 3200, which I cannot seem to fit
>>into the ELO rating formula, where the highest possible gain/loss for a single
>>game is 16 ELO points.
>>
>>Am I going about this in a completely wrong manner, or does performance rating
>>have nothing whatsoever to do with ELO rating ??
>>
>>Greetings
>>
>>Dan Wulff
>>(The Gandalf Team)
>
>Hi Dan,
>forget about performance rating (and your statement about 16 points is wrong...)
>
>You can absolutely use the rating difference as a way to set weights to opening
>moves. The rating difference is equal to an expected wining % and that can be
>translated to what one win/draw/loss is worth in ELO points in the long run.
>As you are Danish you will probably understand this table:
>http://hem.passagen.se/wirsbo/lask.html
>The rating diff translates to ELO points that you can add into your book.
>/Peter

just to clarify:
In your example the rating diff is 100 ELO and in the table you will find
win=12, draw=-4, loss=-20 for the winner and 20, 4, -12 for the loser.
/Peter



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.