is chess.com rating roughly equal to an elo/fide rating?

How do I find  my elo rating 

Is Chess.com's ELO rating equivalent to FIDE's?
I would be tempted to think that ELO Chess.com is overrated...
Your opinion ?

Thanks.

hm

Does this mean that Chess.com players aren't fairly valued and their ratings aren't real ?
For example, a player rated 1500 Chess.com is worth around 1200 FIDE ?...

Hmm... How should I interpret this table ?

Thank you very much for this very precise demonstration, which explains a certain number of things and in particular which explains why weak FIDE rated players seem to play at a superior and very surprising level of strength... So much so that it raises questions as to the use of external help... If I understand correctly, the ELO system has been diverted from its initial use...
Thank you again for these clarifications 👍

no comparison because its a completely different pool of players. very difficult to calculate a true rating. Imagine playing against GMs every single game of your career. Your "elo" would be very low but if you somehow managed to win or draw against that level your rating should be much higher. Same goes if you play in a very low class of players, the opposite would be true. 

I am a fan and an unconditional Chess.com but it is unfortunately the FIDE rating (OTB) and the FIDE Online Arena rating (FOA) which are the only world references... There is no harmonization and that is really very regrettable.

i got 1000 elo because defeat my friends 3 time and lost 1 time

That was stunningly accurate in my case.

Assume player A plays player B and they both have FIDE ratings and chess.com ratings.
The basis of the elo and Glicko system is that the expected result is a function of rating difference.
Expected result of A playing B should be the same.
Rating difference should be the same.
Rating A FIDE - Rating B FIDE = Rating A chess.com - Rating B chess.com
Rating A chess.com - rating A FIDE = rating B chess.com - rating B FIDE
The chess.com rating is about 100 higher than the FIDE rating.

Shouldn't it vary by the drawishness of the format? Draws tend to neutralize the ELO and there are more draws at very very high and also very very low ELO. At high level because not even the weaker blunders enough to let the other win, and at low level because nobody knows how to checkmate effectively without stalemating or taking 50 moves. (well, if you get low enough, certainly this is true of random play engines). The quality of play for high level players increases dramatically with think time, so the drawishness should increase with time control and therefore hurt the peak ELO at high time controls. While allowing it to go well past 3000 at the very lowest controls.

E.G. Blitz seems to have noticeably less restricted rating range in both directions than the standard USCF ratings.

Also shoutout to the one IM level player who plays at 1800 on chess.com Blitz. That must take some doing.

As to low time controls having higher ratings, the current top 3 bullet players here have ratings of 3300+. Blitz players 3200+, Rapid players 2900+, and Daily is 2600+ with a single 2700. Were correspondence leaderboards a thing I would expect to see even lower peak ratings.

While chess.com ratings of games are in line with FIDE and USCF ratings, the puzzle ratings are way off.  My puzzle rating is about 1,000 points higher than my FIDE, USCF, and chess.com game ratings.  Is there a reason for why the puzzle ratings are so far off?

I think the better question is if there's any reason to expect puzzle ratings to correlate with game ratings in the first place.

@39

"Shouldn't it vary by the drawishness of the format?"
++ No. Two draws is the same as a win and a loss.

"Draws tend to neutralize the ELO" ++ No. Elo or Glicko only counts results and adjusts rating accordingly. Two draws is the same as a win and a loss.

"The quality of play for high level players increases dramatically with think time"
++ Yes. Quality goes up and error rate goes down with slower time control.
That does not change win expectancy and thus not rating either.

"Blitz seems to have noticeably less restricted rating range"
++ Bullet, blitz, rapid, classical , correspondence require different skills.
Almost like ping pong and tennis. Or like 100 m, 400 m, 800 m, 10000 m, marathon running.
Being good in one does not necessarily imply being good in another.

"Were correspondence leaderboards a thing I would expect to see even lower peak ratings."
++ Correspondence is unstable. One may analyse for hours and another for seconds.

@40

"Is there a reason for why the puzzle ratings are so far off?"
++ Because nobody has taken steps to calibrate puzzle rating so as to approximately coincide with play rating.

A win and a loss between players of unequal rating tends to neutralize high ratings though. Like, if a 2800 plays a 2600 and draws twice, you now have a 2792 and a 2608. And if that 2608 then plays a 2408 and draws twice, you have a 2600 and a 2416. Draws between players of unequal rating cause rating points to trickle down and away from the highest rated players.

Considering that games between 2600 and 2400s do end in draws 1/3rd of the time in classical format, that means that a 2600 needs to only lose 8% of the time and win 59% of the time against a 2400. Their win/loss ratio needs to be not 3:1 but 7:1.

If they played a theoretical format that with no draws, we would expect their ELO difference to be 325 or so, not 200. Thus, we can say that there seems to be significant ELO compression happening from draws.

Here is what the draw rate seems to look like for mid to high level players in symmetric and assymetric games. Clearly we can see that lowering the quality of play overall would decrease draws if players are less than a few hundred ELO apart. It's unfortunate they don't have data on very low rated players because I suspect some funny stuff is probably happening at 200 ELO as well.

And here is what stockfish centipawn loss does in response to format. Note how GMs are playing like 2000s in bullet games and the slope is much shallower the less drawish the format. (also note the one 1250 asserting dominance over stockfish by blundering a piece per move. Absolute legend whoever they are).

Lichess draw rates for 2500+ rated games:

ultrabullet: 4% of 35000 games draw.

bullet: 6% of 12,000,000 games draw.

blitz: 9% of 6,800,000 games draw.

rapid: 13% of 180,000 games draw.

classical: 24% of 920 games draw.

correspondence: 43% of 1200 games draw.

This heavily suggests that time control increases drawishness and therefore suppresses maximum rating.

Please remember, playing online sometimes limits us to ping, latency, connection drops, which may decrease your performance.