Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Voting unfair !!!

Author: Tim Mirabile

Date: 18:55:03 04/16/98

Go up one level in this thread


On April 16, 1998 at 08:02:15, Andreas Stabel wrote:
[...]
>The question to be answered at the poll was:
>1) Fritz 5 had an unfair advantage
>2) The Fritz 5 rating is accurate with respect to the other programs
>3) Abstain
>
>You had to select one of the above.
>
>Now everybody can see that alternative 2 is not the opposite of
>alternative 1, so even if you think that Fritz 5 did NOT have an
>unfair advantage, you wouldn't agree with alternative 2. As almost
>everybody here has commented the SSDF results are NOT accurate, so
>most persons, myself included, would hesitate to select alternative 2
>even though we clearly disagree with alternative 1.
[...]

I think the key words you are missing here is "with respect to the other
programs".  No rating, human or computer, can have total accuracy.

My personal opinion is that the SSDF list represents a closed pool and
that these ratings cannot be compared to ratings from other pools which
the programs do not play in, such as USCF or FIDE Elo ratings.  But I do
feel they are somewhat useful for comparison between programs if you
keep in mind the margin of error, and the fact that in the Elo system it
takes a 200 point difference to really outclass.

Certainly a 50 or maybe even a 100 point difference on the SSDF list
should not be enough to sway a purchasing decision if you much prefer
the features of the lower rated program. (But I have seen quite a few
purchasing decisions made based upon a one point difference on the SSDF.
:( )



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.