Time span of ratings


Time span of ratings

Back to forum

Don Groves    (2009-01-10)
Time span of ratings

Someone else brought up this subject recently and I feel it deserves discussion:

Suppose one player starts at FICGS at ELO 1200 and gradually improves. Another player starts at 1800 and also gradually improves. Even if they are equally good players, the first player will always be at a ratings disadvantage to the second because the first player's opponents' average ELO will always contain those lower rated games.

Would it make sense to compute ratings based on, for example, only the previous one or two years of play at FICGS, or possibly the previous 100 or 200 games, rather than a player's entire history at FICGS?

I'm wondering if this may yield a truer current rating for everyone.

Thibault de Vassal    (2009-01-10 18:06:40)
Time span of ratings

Actually ratings (correspondence chess ones) are calculated based upon the previous ratings and games played the last 2 months only. In your example it may take 8 or 16 months "only" for the lowest rated to catch the other player in a tournament. Moreover, ratings at FICGS move faster than in other organizations in order to find the good category quickly. So I think that's not a real problem here.

As an example, it took not so long (one year) for Wayne to reach 2113 from 1400 ! That's quite short in correspondence chess.

More details about rating calculation :


Don Groves    (2009-01-11 02:04:43)

Thank you for enlightening me on this! I should have known you were smarter than that ;-)