Chess Rating Comparison

Comparison of ratings between lichess.org, chess.com, USCF, and FIDE. This is an updated to a survey I ran last year comparing chess.com to OTB ratings.

www.chess.com/blog/smarterchess/chess-rating-comparison-2016

Last year's results were interesting - seems like chess.com's ratings are in fact lower than OTB ratings, after all. That seems to contradict much of the popular opinion.

One question though - why was a linear model used for the data in 2015?

I like the 2014 one from chess.com the most.

I never played enough games in cfc (very similiar to uscf) to actually get an accurate rating and its outdated...

So cant really fill out the survey...

have 1900 here (lichess blitz)
1800 chess.com blitz
estimate around 1725 FIDE
estimate around 1625 CFC (should be within 50 pts of USCF)

Mateus6969 we can definitely explore different models. What are your thoughts?

@legend. All of the questions are optional so you can fill in what you have listed there.

@1

Nice survey and data.

I actually created a program converting Lichess blitz to standard fide elo, that exe program is somewhere in this forum.

BTW I cannot see the comparison, I only see this.

I haven't posted any results yet. Right now you can take the survey with the link and view the results in a Google summary. Once we get enough responders 'll post a Google sheet.

Hmmn that is how i understand from your post #1?

Did you mean to say that i cannot view the summary if i will not take part in the survey?

My passionate median chess rating: I add all my chess ratings from any type of chess cadence or variant, and then I do a median value (average calculation). I then give it a ±10 % accuracy to that rating. With the ±10 % it fits pretty close to any of my other chess ratings.

I perform usually well in a tournament category that fits my median level.

Comparing chess ratings is like trying to compare different types of fruits. You have to mention the common points and their differences.

To me a performance rating is more important that an actual rating. It's like knowing the ± RD % (rating deviation) factor of that rating. To be able to truly compare ratings, we need to include their RD and volatility ratings.

duropo full results will be posted later.