Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Chess Tiger - Is It Really 2696 ELO?

Author: Dave Gomboc

Date: 09:17:12 12/24/99

Go up one level in this thread


On December 23, 1999 at 21:01:12, Robert Hyatt wrote:

>On December 23, 1999 at 19:30:48, Amir Ban wrote:
>
>>On December 23, 1999 at 15:05:57, Robert Hyatt wrote:
>>
>>>On December 23, 1999 at 06:32:32, Graham Laight wrote:
>>>
>>
>>>>In the case of the SSDF computer pool, much of it has been there for a long
>>>>time, and is known to be broadly correct.
>>>
>>>
>>>No it isn't, but I also don't have enough time to turn this into a "statistics
>>>101 course" and explain why it doesn't work like that.  If you look at how the
>>>Elo formula works, the _last_ game you played influences your rating _far_ more
>>>than the game you played 40 games ago.  That is simply how the statistics work
>>>here.  So even if the original SSDF programs were _perfectly_ calibrated to some
>>>human rating scale, the fact that 10 years has elapsed means that the effect of
>>>that calibration is _long gone_...
>>>
>>
>>The way SSDF calculate ratings, the order of games and when they were played has
>>no consequence at all. They calculate ratings using the entire SSDF game history
>>every time. It doesn't matter which game was played last. They would get the
>>same ratings in whatever order the matches were played.
>>
>>When the SSDF calibrated their list, they fixed some reference. Since all their
>>ratings are calculated in relative, not absolute terms, this reference, once
>>fixed, is relevant forever and can't possibly grow out of date. Had they fixed
>>that reference 10 points higher, for example, all ratings today would have been
>>exactly 10 points higher, and this will still be true in 10 or 100 years.
>>
>
>
>That is _absolutely_ a flawed way of computing ratings of course, and if that
>is true, the term "Elo" should _never_ be used in the same breath with "SSDF
>ratings".  The Elo formula is quite well-defined...  and the 'order' is most
>definitely imortant in the ratings...

It made a lot of sense before programs adapted their play based on experience.
It doesn't work so well now, though.

Dave



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.