From: "Jeroen J.-W. Tiggelman" <CRMBJTI <at> HDETUD2 <dot> TUDELFT <dot> NL>
Date: 8 jul 1996
Subject: Re: Theoretical Elo ratings

> I have the feeling there is something wrong with these "theoretical" elo =
> ratings for grades: possibly the fact that they ARE so theoretical. If =
> you compare the ratings with the actual average rating of each grade =
> from the latest rating list, you get to see a picture of systematically =
> overrated theoretical grades:

Indeed the theoretical ratings are systematically higher. Whether this
implies that they are wrong or that the actual ratings are wrong, or
that the actual grades are wrong cannot a priori be determined.

However, I would like to point out that any system based on comparisons
like this needs an adequate number of reference material, i.e., there
should be a lot of players, and I believe that this is where the system
at present must fail.

Since there is no good reason to systematically alter grades, it will
not be done (hopefully). Anyway, in case we were reaching any sort of
dynamic equilibrium, the influx of relatively overrated new players
should cause the average ratings to rise. Unfortunately, both the
influx of new players and activity of the old ones seem to be inadequate
to establish an equilibrium.

I believe that the system works insofar that a good result on a
tournament will result in an increasing ELO and v.v., but I feel that
the tournament results themselves do not possess a great degree of
freedom because of few players with similar strenghts present.

The same effect could also be seen at the shogi club where I play:
we are used to a ladder system, but establishing a promotion rule
does not work, since you often play the same opponents, and an off
night of one player could immediately result in the promotion of
another. So, we are using a system with degradations too (to prevent
everyone becoming 9+ dan at the club ;-) -- for you also do not want
to set the rule so high that no one ever promotes). THis mainly results
in wild fluctuations, and there is also a problem because a few players
come often, and a few players only now and then, which has a bad effect
on the balance. (Since the balancing rules are based on relative
strength, with few players this means the entire field may move one
or two degress up or down (as it happens) within a few months; a
player that hardly shows up in that interval, however, would not
move along and find himself with unbalanced handicaps.)

Anyway, I believe the problem is with the lack of players and playing
rather than with any defect of the system. With a statistically unstable
number of players, you always get some sort of error, be it a
systematical one or an erratic one.

> I believe these theoretical grades are used for acquiring a higher =
> grade. If they are overrated this would imply that one has to perform =
> significantly above the average rating of a grade during 6 months to get =
> one, which seems to be unfair to me.

I believe acquiring a higher grade is based on relative results between
players (w/ strengths). I am fairly sure I never reached the theoretical
3 kyu mark. :) Which eliminates the unfairness (if there) if correct. ;)

Of course, as long as everyone suffers under the same rule, it is
debatable whether this could be called an unfairness in the first
place. ;-)

Jeroen Tiggelman
crmbjti <at> hdetud2 <dot> tudelft <dot> nl