What is the rating difference between chess.com's correspondence chess and OTB play.(USCF or FIDE) I find mid 1800ers say say their USCF rating is in the 1350 range. Or a 2350-2400 chess.comer is a 1750-1800 FIDE.

However we have a few titled players(GMs even) on chess.com who aren't 400ish points above their OTB rating. And you will see 2600-2700s drawing and even beating them. So if the top echelon of chess.com is really that good, but the 2000-2400 are just good club players where is the line drawn? 

How severe is the elo inflation on chess.com?

If anybody would like to share their OTB rating and then their CC rating that would be awesome!

There is no direct correlation between the two distinct rating pools. Two different populations playing in two extremely different conditions. Some of the titled players on chess.com might just come on and go through their games making their moves after a few minutes thought, while some experts pour over a single move for an hour or more. And of course, there are some patzers using chess engines.

Incidentally I am about 1900 USCF and 2250 chess.com

Matt my rating on chess.com is around 2000 but my OTB rating is approx 1600. I play a lot of rapid tournaments, 10, 15, 20 or 25 minutes (7 to 13 games in a tournamant during one day) and if I have a very good day my TPR is 1700-1800 but if I have a bad day it's 1400... so I think the rating on chess.com is 300-500 points higher than in OTB for most of the players.

Right Loomis, it also depends on how much time you spend on making your moves on chess.com.

USCF 1787 Chess.com 1834

One thing I find odd is the huge difference between Chess.com turn based ratings and Chess.com live ratings (long, blitz, etc).  Because they're ratings on the same site, maybe they could try to standardize them a little.

Or are there such an overwhelming number of new players on the turn-based side that the ratings become super inflated?

1200 is a very generous rating to begin players with considering this site attracts people who are new to the game.  If I had learned chess a month ago and went to an OTB tournament I'm sure I'd be well under 1000 and stay there for many tournaments to come.

If you played five games against average USCF players (~1400) in your first event and lost every game, your rating could be close to 1000. If you continue to lose, or lose to weaker players it will decline further. It takes some effort (or a kid's tournament) to get a rating below 1000.

A 1200 initial rating on most chess sites tends to produce an average rating close to 1500. Arpad Elo's rating system is designed around 1500 as the average rating. I do not believe that Mark Glickman (author of the Glicko rating system used by this site) altered this aspect of Elo's system in designing his own. The turn-based average here is 1384, substantially below Elo's expectation. The live ratings average close to 1200.

Excluding scholastic players, the median rating in the USCF is just under 1400. With scholastic players, the median drops several hundred below that. The USCf no longer maintains the rating distribution on their website and I'm citing information that is several years old. Despiste all you may have heard concerning rating inflation, the median and average ratings in the USCF are falling due to the large influx of scholastic players. (One more reason to praise my state where scholastic events are rated through a local system, rather than the USCF. Those that stick with chess beyond elementary school or that are serious competitors often join the USCF and get USCF ratings too.)

My USCF and chess.com ratings are available on my profile if you care to gather data. When I hit a new USCF peak, I update that info on my profile. When it drops, as it assuredly will, I may not update as rapidly. If it drops significantly, however, I will update it. I expect it to drop slightly before rising to another new peak. I'm shooting for 2000 over the next few years.

I don't show the whole quote but I agree with about everything you said -- I've also read an article about the large influx of underrated scholastic kids deflating USCF ratings -- it's wonderful your state helps keep the two separate.

About the two averages on this site though, I think the majority of lower rated players play exclusively on the turn-based side.  I think if every player played both you would see the live rating average drop a few hundred points.

If ELO intended the average to be near 1500 and this site's average is 1300 then perhaps in the meat of the bell curve, let's say with 70% of the users here, the ratings are inflated around 200 points and so it wouldn't be unfair to start each player at 1000. 

Now that this site is large it wouldn't make much sense to either decrease everyone's rating or start new players at a different rating (although that might not be too unreasonable) -- but like the old saying goes, if it ain't broke... I just found it odd / hadn't noticed it before.

I just recently started getting into chess.  I'm not in a club or anything.  But against Shredder (over a more than a hundred games) I'm 1435.  Chess.com I'm 1100 or so. (only about 10 games)