Chess.com rating conversion to OTB

Online ratings will never get much respect. It's way to easy to cheat.

Let me clarify how the conversion works between online, OTB, FIDE etc. for rapid games:

OTB = online ± 200 (sometimes it deviates even more).

Online = OTB ± 200. (sometimes it deviates even more).

Online = another online ± 200. (sometimes it deviate even more).

Chess engine calibrated rating (your true approximate rating) = online ± 200.

Chess engine calibrated rating (your true approximate rating) = OTB ± 200.

Chess engine calibrated rating = another chess engine calibrated rating ± 60-80.

Now then we also have bullet and blitz games. They are usually from -300 to +100 of your normal rating.

So, things are very simple, but they are not the way a lot of folks think they are. OTB ratings are neither better, nor worse than online ones. Simply put, ELO is a floating thing that peaks and plunges.

To sum it up, there will always be people with a high OTB rating and a low online rating and vice versa. This is especially true at amateur levels. At the higher end (for GMs) it should be much more precise with less leeway.

Disclaimer. These were all ballpark estimates and very rough formulas.

This is very logical indeed. Thank you. 

You're welcome

One more thing while bullet and blitz games are usually from -300 to +100 of amateur rating;  it's not that way higher up the ladder. At the very top it is the other way round: from -100 to +300. Just take a look at the ratings of Calsen or Nakamura. The explanation should be very simple. The blitz and especially bullet Elo range curve is more elongated. It starts from the very down below and goes up higher than 3100!

So in your opinion, if someone says his FIDE is 2125, yet after playing many games, has a blitz of 1600, a rapid of 1400s, and pretty poor daily as well (missing a mate in 1 against an 1100), would you believe that's impersonation / false rating?

1400 rapid is definitely too low for a 2125 rated player. Bad blitz and bullet ratings can be attributed to the quick pace of the games, but 30 mins or 15/10 should provide enough time for any 2100 otb rated player to think and play at higher than the chess.com 1400 level.

OK. Some guy claiming to be someone with a 2125 FIDE rating (which was obtained many years ago, to be fair to the guy) has those ratings, and when I posted on his thread (which he created basically talking about how he thought chess.com ratings were horribly deflated lol) that if he was ever 2125, he was nowhere near that now (I said he'd be closer to 1500 FIDE based on game quality... , and that was me being generous - he missed a mate in 1 in Daily chess)... and he blocked me! lol

Cherub_Enjel,

I also found it very weird. I believe his FIDE 2100 was a bad approximation. A chess engine would have probably given him something like  ELO 1950 when he was active in tournaments. My guess is that the rating system here is deflated by 100 points. So his ELO should be around 1850 here but if he lost some of his strength by not playing chess for many years he could have gone down to 1800. His blitz Elo was, If I'm not mistaken, around 1550. That's 250 difference between a more standard ELO of 1800. So, it doesn't look that weird after analyzing his situation. Of course, the first thing that springs to mind is he could not have had Elo 2125 and that it's some kind of bs. And yet I think it's still possible.

Vlad creates arguably the best threads for entertainment. i can't believe ppl take his stuff seriously. some of his stuff is hilarious :) thanks Vlad

Cherub_Enjel,

 My best guess would be his true engine-estimated ELO was around 1900 when he was active (but he got to 2125 through some luck) then he stopped playing and lost some strength going to 1800. Since I think ELO ratings are a bit deflated here, it corresponds to 1700 in rapid games which is totally in line with his 1550 in blitz. Of course he might have lied about 2125 but I don't want to perceive it that way or say bad things behind someone's back. It ain't right.

(1) No one takes his stuff seriously - but we play along because he is probably the best troll / comedian here.

(2) Yeah - I would say that he's probably just really really rusty and possibly beyond repair. I said to give it a few weeks to see if he'd jump back, but he's gotten even worse arguably. And I'll be fine with believing he had a 2125 FIDE 20 years ago, but right now he plays like a 1500.

There is one thing that doesn't really add up. It is that his rapid Elo was 1600, if I'm not mistaken. That's way too low to make a connection to FIDE 2125. Way too low for a semipro of 2100. So, I totally agree here.

I agree that he knows theory, and doesn't play like one of those guys who plays only online chess. But I'm guessing you didn't see the 11 move loss in the king's gambit, nor the missed mate in 1 vs. an 1100 daily player. 

Maybe he just needs even more time...

That just goes to show that even FIDE Elo jumps up and down too much. The most probable explanation is of course a nasty one of something murky there. I'm talking about his gap between the rapid 1600 and  2100 in long games. Blitz games are of course not indicative enough.

In blitz here, I run into a lot of players in the 1800s with FIDE ratings well above 3000, according to their profile pages. I beat those who don't cheat.

Ziryab, FIDE ratings higher than 3000? Really?

Well, there are no players with FIDE ELO above 3000. Carlsen is also below that mark with his blitz FIDE Elo. An online bullet Elo can be higher than 3000; that's for sure. I think it needs to be dealt with/reported. There are real GMs here, including Nakamura. Such things may disrupt them, I mean, best players. It ain't right.

What? You mean players on this site lie about their OTB ratings? Who woulda thunk?!

Ziryab, I mean there are no official FIDE players with FIDE Elo higher than 3000 even in blitz games. It's that simple. You can check it.