14 July 2013

Ratings Deflation Confirmed?

FIDE's Qualification Commission just published an analysis of the ratings system using a 'Retroactive Ratings Calculator' devised by Jeff Sonas, the renowned statistician. Here.


Of most interest to me is the note on ratings inflation at the bottom of the report. The analysis seems to substantiate the perceived phenomenon of ratings deflation at lower levels of the ratings scale. What was most surprising is that the analysis puts what I call the 'inflection point', at around 2700. This inflection point is where ratings deflation ends and ratings inflation starts. I had argued (in a debate with Ronnie last year) that my superficial analysis had discerned ratings deflation at just below 2500, but I certainly would not have thought that the inflection point would be so high as 2700.

The deflation gets more pronounced the lower you go down the rating scale. The part of the analysis looked at data in the 2000-2700 ratings range and from only as far back as 2008 (5 years ago), and the significance was already seen. This would be more pronounced if data from even earlier periods were used. Thus a 2200-rated player nowadays would be playing stronger than a 2200-rated player 10 years ago if for no other reason than just ratings deflation.

And in the final paragraph of the report, a note was made on the increasing ratings gap between the strongest players and the rest.  It sounds to me like the 'expanding universe' theory. It's a natural phenomenon, man. You just cannot fight it (or even argue with it).

No comments:

Post a Comment