ICCF Elo rating calculator
This calculator is part of my C++ class library of chess rating systems. Please go to introduction page for further information.
ICCF is the official world correspondence chess federation with a pretty simple set of rating rules.
If you have less than 30 games, you must include all your games, not only the ones from the last period. Old rating is your initial rating.
your rating calculator is a very good idea, but you must take in account the concept of “maximum resistance”, which is fixed by ICCF in 350 points, if you want to give a proper answer in games between player with more than 350 rating difference.
I’m a bit puzzled about your comment, as I have implemented such a cut-off. It takes effect even in the calculation of the opponents’ mean where it is easy to notice. Can you provide me with a proper data set on which the calculator misses your expectation?
Is this formula still active and conforms to the rating calculations used by ICCF? I compared a four game win streak between ratings calculators on USCF and your. USCF score new rating 160 points higher than your formula.
Apparently there is only a USCF rating calculator on the USCF website. While all these ratings are based on the same mathematical principles, the concrete formulae differ from one another.
Nice program but the average rating calculation has a bug.
If I enter one result, my rating =1800 and a win aganst someone with a rating of 1336 than the average is shown correctly as 1450. If I now add a second result, a draw against someone with 1800 then the average opponent rating is shown as 1568 (1336+1800)/2 although it should be 1625 (1800+1450)/2. I hope you are still there.
I think it is not really bug as the rating average is an informal value only. The new rating is calculated as sum of individual rating changes per game. The average does not occur there. And that is the reason why my test series did not reveal this inconsistency. But you are right, it is confusing. Will fix it within the next few days.