Can Chess.com Ratings Represent OTB?

I ran some numbers comparing Chess.com ratings to both USCF and FIDE. The Chess.com ratings are within 100 points usually of OTB. Lichess is consistently 200-300 points above the OTB ratings.

Comparison tables

Explanation of the stats

Since most players are playing at home for now, does it seem like we can use Chess.com ratings to measure where we are at in OTB chess?

Oh yeah i have seen this before! Thank you for that! 

I have a question. Is this an estimate or correct for all players?

This is an estimate, with a very wide range of possibilities. The +/- by the estimates helps give an idea of how wide of a range 68% of the players fall into. For USCF especially, this range is very wide.

Thanks! I'm going to message you

It's not an error. You would be included if you filled out our survey in the past. Great job with your progress! I have a suspicion that the online ratings are increasing faster than OTB ratings for players who have both due to the pandemic. I only counted players with Chess.com RD < 150 and FIDE players in the November supplement. So if someone hasn't played OTB in a long time they wouldn't be included in the comparison.

I believe chess.com rating does represent OTB rating strength. 

I wish

Rankings are a much better indicator.  I have three ratings, and there is a significant difference between the lowest and the highest.  However, my ranking in each pool is about the same.

The charts are averages/ median ranges and indicative of the average players. To make claim ratings fall within 100 points works well for 50%. The other half can easily fall between +/- 200 points. Just too many variables/ differences between the two very different disciplines. Online skills and OTB experience are entirely different animals until experience is gained in both- by playing 100,s of games. Another prob with the stats - they are based on a persons “stated” OTB rating without verifications unless checked thru USCF/ FIDE. A lot of good statistical work shown - but I’m not sure of it’s relevance. For established players the stats work fine half the time. Myself - I fall way out of the norms due to circumstances as do many other players. Not critical of the accuracy of the stats in the least - just that quick interpretations should be avoided. There remain hidden variables not taken into consideration - as in all such studies that lack precise controls. 

Yes - percentile stats prove better indicators.

I did use the wrong term.  Percentile is what I meant.

How much stock can be placed on a single online survey - that is partly dependent on voluntary answers, by anonymous persons reporting their OTB rating in a profile? Perhaps some ... but with a grain of salt.

A good survey depends on having a representative sample.  It doesn't always happen.   Look at the polls for the last two US presidential elections.

in the chess academy server my chess.com blitz makes me a 2300 lmao

I really do hope that @smarterchess's graph is accurate... it means I might be able to become an NM.

Dream High, You can become an IM too

but if I set my expectations low, I won't be disappointed lmao

A major factor is the rating pool. Players of different sections travel to tournaments in another area - and find to their great surprise players there, with a similar rating, are very much stronger or weaker in strength. Many players online pick and choose who to play - clubs or friends. For representative samples - games 100,s of games need playing vs random players over the course with equal playing conditions.

not if covid

All I gotta do is clap a bunch of 2300s lmao