You can only compare your rating to other Lichess players
I'll pile on and give a +1 to the "you can't really compare them" point.
It's a different pool of players, a different rating system, different common time controls, etc.
A lot of people have a tough time coming to grips with the fact that all these rating systems only give you relative differences between players in the pool of players being rated in that system.
There's no such thing as a "one true" rating relative to which other ratings are inflated or deflated.
All these rating systems (USCF, FIDE, ICC, chess.com, lichess, etc.) describe the relative differences in strength between the players rated by that system, under those rules and conditions (including things like time control, auto-pairing pool, etc.).
The absolute numbers in these sorts of rating systems mean nothing. It's only the differences between players that matter.
You can (and on some engine rating lists people have, to avoid meaningless comparisons with human ratings) just give the highest rated player a rating of 0 and everyone else negative something, and as long as you preserve the differences between players, it's just as accurate as with the numbers we're used to.
People just get fooled and sucked in by the fact that the absolute numbers look somewhat similar across rating systems, and think they can start comparing them.
No one would take seriously a claim that ECF ratings are "deflated" because they're so much lower than FIDE/USCF/ICC/chess.com/lichess; it's obviously just a different system.
With things like the common implementations of Elo and Glicko, the similarity in absolute numbers tempts people to make these silly comparisons.
Ok, off my soapbox now :)
It's a different pool of players, a different rating system, different common time controls, etc.
A lot of people have a tough time coming to grips with the fact that all these rating systems only give you relative differences between players in the pool of players being rated in that system.
There's no such thing as a "one true" rating relative to which other ratings are inflated or deflated.
All these rating systems (USCF, FIDE, ICC, chess.com, lichess, etc.) describe the relative differences in strength between the players rated by that system, under those rules and conditions (including things like time control, auto-pairing pool, etc.).
The absolute numbers in these sorts of rating systems mean nothing. It's only the differences between players that matter.
You can (and on some engine rating lists people have, to avoid meaningless comparisons with human ratings) just give the highest rated player a rating of 0 and everyone else negative something, and as long as you preserve the differences between players, it's just as accurate as with the numbers we're used to.
People just get fooled and sucked in by the fact that the absolute numbers look somewhat similar across rating systems, and think they can start comparing them.
No one would take seriously a claim that ECF ratings are "deflated" because they're so much lower than FIDE/USCF/ICC/chess.com/lichess; it's obviously just a different system.
With things like the common implementations of Elo and Glicko, the similarity in absolute numbers tempts people to make these silly comparisons.
Ok, off my soapbox now :)
Actually they discuss it all night long on different servers...
The numbers have some similarity although there are numerous deviations in calculating them, entrance level, player distribution, different spreading, "high frequency trade" (thousands of games and rating calculations each day).
According to my experience chess24 and lichess are about +200 compared to RL, playchess meanwhile (over 15 years) under-rated (blitz).
The numbers have some similarity although there are numerous deviations in calculating them, entrance level, player distribution, different spreading, "high frequency trade" (thousands of games and rating calculations each day).
According to my experience chess24 and lichess are about +200 compared to RL, playchess meanwhile (over 15 years) under-rated (blitz).
I think maybe the points I was making got lost somewhere.
I'll reformulate each of them.
1) The rough similarity in the absolute numbers returned from the major rating systems tempts people to directly compare them.
This usually results in claims that one system overrates or underrates players, missing the fact that there's not a "one true" value relative to which a particular system's rating can be high or low.
Eventually, this results in people trying to find some correlations to "correct" the system they think gives incorrect numbers.
Some people know that no one of the ratings is the "true" one, and that they're just trying to find an interesting correlation, but many don't.
2) With things like ECF grades, the numbers are so different that no one is reasonably tempted to directly compare the numbers and say "Oh my! ECF drastically underrates everybody (because the numbers I see on ECF lists are so much lower than the numbers on FIDE lists)! Let's fix that!".
Sure, people argue about the correlations and proposed conversion formulas for ECF->FIDE or whatever, but they're not tempted to directly compare the absolute values like they are for other systems.
In short, while the lichess system is every bit as separate a system from say, chess.com as ECF is from chess.com, people are much more likely to ignore the "different system" point when comparing lichess to chess.com, just because the numbers look somewhat similar.
With ECF and chess.com, for example, there's no missing that it's a completely different system.
3) The search for correlations, however interesting, is flawed, because of the vast differences in players, typical conditions, and even rules between various systems.
Obviously this isn't to say that there will be zero correlation between ratings from different systems; there will be some, of course, but the correlations are fairly weak.
I'll reformulate each of them.
1) The rough similarity in the absolute numbers returned from the major rating systems tempts people to directly compare them.
This usually results in claims that one system overrates or underrates players, missing the fact that there's not a "one true" value relative to which a particular system's rating can be high or low.
Eventually, this results in people trying to find some correlations to "correct" the system they think gives incorrect numbers.
Some people know that no one of the ratings is the "true" one, and that they're just trying to find an interesting correlation, but many don't.
2) With things like ECF grades, the numbers are so different that no one is reasonably tempted to directly compare the numbers and say "Oh my! ECF drastically underrates everybody (because the numbers I see on ECF lists are so much lower than the numbers on FIDE lists)! Let's fix that!".
Sure, people argue about the correlations and proposed conversion formulas for ECF->FIDE or whatever, but they're not tempted to directly compare the absolute values like they are for other systems.
In short, while the lichess system is every bit as separate a system from say, chess.com as ECF is from chess.com, people are much more likely to ignore the "different system" point when comparing lichess to chess.com, just because the numbers look somewhat similar.
With ECF and chess.com, for example, there's no missing that it's a completely different system.
3) The search for correlations, however interesting, is flawed, because of the vast differences in players, typical conditions, and even rules between various systems.
Obviously this isn't to say that there will be zero correlation between ratings from different systems; there will be some, of course, but the correlations are fairly weak.
yes it s, a1800 players play like 1500 here , chess.com is more accurate,gg.
Many good points have been made. Most pointing to the Fact... you just can't know until you play OTB. Remember also, the 1st few tournaments are a provisional rating. The established rating later on, in general, will be higher, in some cases quite a bit higher as often the new experience is distracting.
Some players acquire excellent on line skills, others do not perform as well in fast time controls. As stated by several posts, there are just way to many variables. Even comparing blitz ratings at different sites, unless you have played 1000's of games at each within the same time period, is fool-hearty.
There is a thread here with pages of posts about a professional statistician who downloaded every game played here, graphed every active FIDE player, looked at every profile, used weeks of comp time to create a formula he claims can predict a new players OTB rating within a very narrow +/- range. Believe that stuff if it makes you happy. From experience, I have seen 1400 online players after their 1st OTB tournament start anywhere from 1000 to 1800. The pool of players, how much previous "casual" OTB experience, how much preparation, playing for prizes or just at your Friday night club, how many games in the day, how established are the opponents ratings, what kind of pizza you ate for dinner the previous night...the list is endless.
For some players their online rating and FIDE rating are similar. For many others there is no correlation. Online blitz and OTB are entirely two different animals. Another factor not mentioned is age.
Some players acquire excellent on line skills, others do not perform as well in fast time controls. As stated by several posts, there are just way to many variables. Even comparing blitz ratings at different sites, unless you have played 1000's of games at each within the same time period, is fool-hearty.
There is a thread here with pages of posts about a professional statistician who downloaded every game played here, graphed every active FIDE player, looked at every profile, used weeks of comp time to create a formula he claims can predict a new players OTB rating within a very narrow +/- range. Believe that stuff if it makes you happy. From experience, I have seen 1400 online players after their 1st OTB tournament start anywhere from 1000 to 1800. The pool of players, how much previous "casual" OTB experience, how much preparation, playing for prizes or just at your Friday night club, how many games in the day, how established are the opponents ratings, what kind of pizza you ate for dinner the previous night...the list is endless.
For some players their online rating and FIDE rating are similar. For many others there is no correlation. Online blitz and OTB are entirely two different animals. Another factor not mentioned is age.
chess.com rated lower than otb
lichess rated over otb
Uscf rated unequal to Fide
chess24.Icc etc all different then otb
But all good for their sites only
lichess rated over otb
Uscf rated unequal to Fide
chess24.Icc etc all different then otb
But all good for their sites only
Lichess Blitz = FIDE Elo
As of today, valid for Sarg0n.
(My other account has 2263 though. Do I have multiple personalities?)
As of today, valid for Sarg0n.
(My other account has 2263 though. Do I have multiple personalities?)
I left other chess websites to focus only on lichess........reason is very simple.......this website induce the interest to play chess.when a person play cluelessly with unknown person without knowing age or name or other things that of the person,lichess way of RATING is better than other chess website rating because this helps psychologically a normal chess player to progress to the potential,........in my view , in other websites, only great player will become greater.......in lichess, normal player will become great player.........great player may become greater player......
Totally agree
This topic has been archived and can no longer be replied to.