It seems pointless because there's no correlation. I think you could get similar results by comparing a non-chess rating with any chess rating. 

Well, statistically-speaking there is a correlation. But one must always remember that correlation may have NOTHING to do with declaring such a chart meaningful. 

However, no one is his right mind would disagree with the notion that the top players - AS A GROUP - in chess.com are stronger than the bottom players, or that - ON AVERAGE - they would probably fair better in OTB play than the bottom-rated players. Thus, there's always going to be a correlation between two reasonably constructed ratings sytems both with large pools of players.

In any case, YMMV (your mileage may vary).

Peace ... and good chess to you, regardless of rating, rating system, or rating pool. 

I have a chess.com blitz rating of 2786, an online chess rating of 2918, a USCF rating of 2871, and a FIDE rating of 233. Would you include me in the statistics, please? 

I've seen even more absurd claims than that, actually, such as people claiming a two-digit USCF rating, which doesn't exist. 

USCF ratings don't even correlate that well with each other. Anyone who has been to a national tournament knows that 1500 in Idaho is different than 1500 in NYC is different from 1500 in Texas. I assume this closes at the Master level somewhat as the pool of players is smaller and the players travel and play each other. This would certainly happen even more at the GM level. 

This seems like a really stupid question.  Your more making a statement about you've taken a statistics class than asking a question.

I am finding this to be the case.  Three friends from the Pittsburgh Chess Club, all active players, have a USCF classic rating of 2100 and blitz ratings of about 1950-2000, and their chess.com blitz ratings are all in the 1550-1750 range.  They peak higher, but that's where they are, typically, in that range.

Very true.  Additionally, folks who play big tourns with lots of sandbaggers tend to be lower.  I've seen quite a few players at the World Open, Chicago Open, etc., burst on the scene after two years of tournament inactivity and jump up a ton of points.  The fact that many people lose to this kind og person on the big tourn scene, ratings tend to be lower.  I once visited a small town for a tourn, and felt the ratings were much higher than what I saw in their playing ability.

As Sunshiny mentions, visually speaking there does not appear to be a correlation between the variables.  You mention that there is a correlation of 0.6, but I'm pretty confident that the values represented in your plot do not indicate a correlation of 0.6.  There is so much variability in the chess.com scores that if you know that someone's USCF rating is, say, 1665 then you cannot even remotely predict what their chess.com scores is.  Judging by the plot it is as plausible that their chess.com score is 1500 as it is that their score is 1700.

You might be on to something.  I've been away from USCF OTB chess (this is my first week back), so I can't say with respect to my rating.  However, three of my close friends, all between 2050 and 2150 USCF in classic rating, are 1700's in chess.com blitz.

http://www.chess.com/blog/smarterchess/chesscom-rating-comparison-2015 appears to contain updated data:

Chess.com Rating Comparison 2015

• NM smarterchess
• | Nov 11, 2015 at 9:00 PM
• | Posted in: Smarter Chess Blog
• | 3476 reads
• | 21 comments