- Blind mode tutorial
lichess.org
Donate

Any particular reason why lichess rating is sooomuch higher/differnt from FIDE??

#11

ICC has always had the closest ratings compared to FIDE...a 1600 on ICC will beat a 1600 from lichess like a red headed step child.

ICC has always had the closest ratings compared to FIDE...a 1600 on ICC will beat a 1600 from lichess like a red headed step child.
#12

well 1600 lichess is more or less 1200/1300 Fide while a 1600 fide would be more like 2100-2200...

well 1600 lichess is more or less 1200/1300 Fide while a 1600 fide would be more like 2100-2200...
#13

@gothica31 ..it seem that people on lichess care more about FIDE rating than lichess rating... the moment that they reach a milestone in ratings they then want to compare it to the FIDE rating

@gothica31 ..it seem that people on lichess care more about FIDE rating than lichess rating... the moment that they reach a milestone in ratings they then want to compare it to the FIDE rating
#14

FIDE is like a Formula 1 racing. Online is like a moped racing. That's why I am sitting online.

FIDE is like a Formula 1 racing. Online is like a moped racing. That's why I am sitting online.
#15

#13 I agree with you and that's why I made this table but FIDE rating is obsolete and doesn't represent the true level of players (we've seen it with lockdown, some 1200F can now defeat masters). But it's still the official chess rating and many people have a fide rating or desire to have one and wish to see where they stand on a rating ladder which has been completely changed with lockdown...

#13 I agree with you and that's why I made this table but FIDE rating is obsolete and doesn't represent the true level of players (we've seen it with lockdown, some 1200F can now defeat masters). But it's still the official chess rating and many people have a fide rating or desire to have one and wish to see where they stand on a rating ladder which has been completely changed with lockdown...
#16

@Kingscoffee Simplest answer: Ratings are relative i.e absolute value is meaningless. Lichess ratings exist for purpose of finding opponent of about same strength.

@Kingscoffee Simplest answer: Ratings are relative i.e absolute value is meaningless. Lichess ratings exist for purpose of finding opponent of about same strength.
#17

#16 or more generally to have some idea of how difficult to win a given pairing within given pool would be.

about same strength being particular case. And not mandatory, is it? (i lost track of how paring evolved over past 2 years, playing correspondence and have reduced my random pairing to almost nothing.

Also, as a longitudinal measure of my evolution, it could give me some idea of whether i have improved or not. hopefully. Not sure though. But really game struggle or challenge effort expectation is most likely the main usage.

Variation on the "about same" though. is +_200 within that?

I think i would like sometimes how bad i could be beaten.... or how much time i would have to put in my daily move (not game) to maintain some resistance and for how long in a game. So depending on the time-control (correspondence does exist on lichess), it is not just about the same but more about the difficulty.

But I do agree it is not meant to be external referential between different pools of potential pairs.... Although with large population the input control into such statistics meant to obtain a certain average and deviation, and knowing the emergent distribution behaviour for each pool at lest in standard deviation of the population in the pool, could allow some transposition. I would not expect it to be linear and with certainty. The prior individual rating probabilities (belief) (initial condition in prob. space, from some other type of mathematics point of view), is an input for those computing them within a community. So knowing those, and following the theory as to the stationary distribution characteristics, one could do the above. or try.

Not having stipulated such population distribution as a parameterized family does not prevent from making statistics on it, and probably establish piece-wise relations... but that would need work.

one would have to go into each rating foundation models, study how pairing probability constraints are, and the actual demographics of their population where each rating is used, to really be sure. lots of research, if it has not been done anywhere (has it?).

#16 or more generally to have some idea of how difficult to win a given pairing within given pool would be. about same strength being particular case. And not mandatory, is it? (i lost track of how paring evolved over past 2 years, playing correspondence and have reduced my random pairing to almost nothing. Also, as a longitudinal measure of my evolution, it could give me some idea of whether i have improved or not. hopefully. Not sure though. But really game struggle or challenge effort expectation is most likely the main usage. Variation on the "about same" though. is +_200 within that? I think i would like sometimes how bad i could be beaten.... or how much time i would have to put in my daily move (not game) to maintain some resistance and for how long in a game. So depending on the time-control (correspondence does exist on lichess), it is not just about the same but more about the difficulty. But I do agree it is not meant to be external referential between different pools of potential pairs.... Although with large population the input control into such statistics meant to obtain a certain average and deviation, and knowing the emergent distribution behaviour for each pool at lest in standard deviation of the population in the pool, could allow some transposition. I would not expect it to be linear and with certainty. The prior individual rating probabilities (belief) (initial condition in prob. space, from some other type of mathematics point of view), is an input for those computing them within a community. So knowing those, and following the theory as to the stationary distribution characteristics, one could do the above. or try. Not having stipulated such population distribution as a parameterized family does not prevent from making statistics on it, and probably establish piece-wise relations... but that would need work. one would have to go into each rating foundation models, study how pairing probability constraints are, and the actual demographics of their population where each rating is used, to really be sure. lots of research, if it has not been done anywhere (has it?).
#18

As said above, rating is only to compare strength within the same system.
Many new players equate the word elo with a ranking "What is your elo on Lichess? ". Of course it does not mean anything.
Better saying " what is your rating on Lichess ?"
Each rating are calculated with particular conditions.
You can't compare OTB 30+10 with internet 30+10.

As said above, rating is only to compare strength within the same system. Many new players equate the word elo with a ranking "What is your elo on Lichess? ". Of course it does not mean anything. Better saying " what is your rating on Lichess ?" Each rating are calculated with particular conditions. You can't compare OTB 30+10 with internet 30+10.
#19

@dboing yuo use so many words that I rarely make it to end. But anyway mapping from lichess to FIDE is not meaningfull mostly because it rather random

I like posting image below. It is scattergram mapping lichess rating to verified FIDE ratings. There is some correlation but apparently it is very weak. reason are multitude. Like how seriouly people play online varies more than that of OTB player. Also totally different time limits play part

https://i.imgur.com/Txd4qBH.png

also link to source including linear regression models
https://www.reddit.com/r/lichess/comments/mdvbkp/yet_another_lichess_vs_fide_rating_comparison/

@dboing yuo use so many words that I rarely make it to end. But anyway mapping from lichess to FIDE is not meaningfull mostly because it rather random I like posting image below. It is scattergram mapping lichess rating to verified FIDE ratings. There is some correlation but apparently it is very weak. reason are multitude. Like how seriouly people play online varies more than that of OTB player. Also totally different time limits play part https://i.imgur.com/Txd4qBH.png also link to source including linear regression models https://www.reddit.com/r/lichess/comments/mdvbkp/yet_another_lichess_vs_fide_rating_comparison/

This topic has been archived and can no longer be replied to.