I am a bit confused about rating differences on chess sites. On Chess.com, my rapid rating is in the low to mid 1900's. On Lichess it is near 2200 and on Chess24 it has been as high as 2300. They all use the ELO system, so what gives?
Lichess is not ELO. See
https://lichess.org/faq#ratings
The various rating systems are configurable (i.e. parameterized). Also, the pool of players is different.
Different sites, different player pools and different starting ratings. According to my experience and collected evidence (by https://chessgoals.com/rating-comparison/) it seems that players on chess.com have about 100-200 points lower ratings than they do on lichess on average. I don't know about chess24 but considering that there are people at 3200+ when watching Magnus stream banter blitz I assume it is the most inflated rating pool out of the three. Your personal evidence would suggest similar.
Statistically chess.com ratings have the highest correlation to different OTB ratings (USCF, FIDE) so they might be the most accurate estimation (out of the three sites) of your playing strength in comparison to OTB, if that's what you're after. And before anyone argues about the differences between online and OTB, I'm talking purely about statistics and correlation, nothing more.
Personally I have about 100 points lower rating on chess.com than I do at lichess blitz, and especially the rapid pool in chess.com is really tough.
The most " really tough pool " is on https://chess-samara.ru/play/index.html (or they most cheat..)
2000 here is 1500 there I suspect (blitz)
#3 Yes rapid pool is chess.com is amazingly tough. It can be -300 points or more compared to lichess rapid.
@pointlesswindows bullet and blitz on chess.com are tough not rapid.@Sleaver i play on both sites so you can compare...My stats Chess.com: blitz 1709 Bullet: 1200(i dont playnit there offten) rapid:1778. Lichess: blitz:1888 bullet: 1802 rapid:2000 classical:2040.
I use chess.com to test random and weird stuff, because I don't trust the site, so I don't know how my rating would compare here "for real". When I watch titled players "How to Climb" series, though, it seems like chess.com is the most absurd of all sites, where 1300 players can somehow give IMs a run for their money - if you don't think that's suspicious , you're a better human being than me.
@texia you have funyny definition of correlation , Chess.com valuea are numerically closer to USCF values but that does not if more correleated following two sets of number have correleatin factor of 1.0
1 2 3 4 5 6 7 8 9
3 6 9 12 15 18 21 24 27
There is absolutely no reason to assume lichess ís less correlated. and correlation is weak at its best
https://imgur.com/E7oTGPA
And for for @SLeaver whether lichess used elo of glicko makes no difference. all measurement is based only on winning and losiing. Hence in two separated pools of players numbers are gonna mean different thing ans it is not flaw but feature. Lichess rating are for lichess and for nothing else
@Venividivici321 What do you mean?
blitz 1888-1709 = 179 difference
rapid 2000-1778 = 222
And in my case the difference goes to 300 points.
Bullet ratings are irrelevant.
This topic has been archived and can no longer be replied to.