English Chess Forum

There's an article on the ChessBase site which makes a claim that in adult v junior games, the International Elo system fails to predict the aggregate result by around 50 Elo points. That's about the same as the ad hoc adjustment the theorists on the BCF grading system (Sir Richard Clarke etc.) came up with in the late 1960s to prevent the BCF grading numbers spiralling downwards. Their ad hoc adjustment was 5 BCF points which equates to 37.5 to 40 Elo ones.

http://en.chessbase.com/post/elo-rating ... s-part-two

I don't know to what extent, if at all, the author filtered for the effect of the recent increases in K factors, particularly the K=40 for most juniors.

Roger de Coverly wrote: I don't know to what extent, if at all, the author filtered for the effect of the recent increases in K factors, particularly the K=40 for most juniors.

Not at all, so far as I can see. The author's findings look very similar to FIDE's own, in the investigation on which the big k factor review was based.