Author: James B. Shearer
Date: 14:26:26 03/18/99
Go up one level in this thread
On March 18, 1999 at 16:32:25, Tim Mirabile wrote:
>On March 17, 1999 at 16:44:41, James B. Shearer wrote:
>
>> Actually as someone else pointed out it is wrong. By the same logic if
>>I started with gnuchess at 2300, moved to Hiarcs 7 at 2800 and then returned to
>>to gnuchess at 2300 I would be causing ratings deflation. This is equally
>>wrong. If I repeatedly move between 2300 and 2800 programs I am slightly
>>perturbing the rating pool but there is no net long term effect.
>> James B. Shearer
>
>Except that it annoys people who lose points because of this, and skews some
>ratings up and down in the short term. If allowed to continue constantly it
>would have the long term effect of making the whole system less accurate
>(although neither inflated nor deflated). Think of it as adding a random number
>from +x and -x to each player's rating, where x depends upon the severity of the
>problem.
True but a big K factor (as on ICC) means this noise damps out rapidly
and is probably insignificant compared to the random noise continually
introduced by the big K factor. Interesting tradeoffs.
James B. Shearer
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.