- Blind mode tutorial
lichess.org
Donate

What is the highest possible rating on Lichess?

Skimming through a few stats of our titled players I see that Magnus (DrNykterstein) had a top bullet rating of 3355 on this site. A blitz rating of 3015 was achieved by Konevlad. There may be room for small improvements on these figures but to maintain such scores is difficult, as can be seen by the lower current ratings of these players.

Skimming through a few stats of our titled players I see that Magnus (DrNykterstein) had a top bullet rating of 3355 on this site. A blitz rating of 3015 was achieved by Konevlad. There may be room for small improvements on these figures but to maintain such scores is difficult, as can be seen by the lower current ratings of these players.

@sheckley666: There is a small difference. 1 minute ago I finished a game against a player rated almost 1000 points below me. Out of curiousity I checked the rating before and after the game. Of course it said "+0", but actually I gained 0.09 rating points. You can see the two decimal digits here: https://lichess.org/@/sheckley666/perf/rapid

@sheckley666: There is a small difference. 1 minute ago I finished a game against a player rated almost 1000 points below me. Out of curiousity I checked the rating before and after the game. Of course it said "+0", but actually I gained 0.09 rating points. You can see the two decimal digits here: https://lichess.org/@/sheckley666/perf/rapid

Thx @oberschlumpf I knew that rating is given with decimal digits. But is it internally calculated with even higher precision? I really wonder where @tpr 's 735 points come from.

Thx @oberschlumpf I knew that rating is given with decimal digits. But is it internally calculated with even higher precision? I really wonder where @tpr 's 735 points come from.

that is the point when adjustment drops below one. https://en.wikipedia.org/wiki/Glicko_rating_system from there the expected score has in r0-r1 as paremaeter as get smaller then gain gets smaller if wanna work it out
with that rate gain it takes more than 100 games to get one full point :)

that is the point when adjustment drops below one. https://en.wikipedia.org/wiki/Glicko_rating_system from there the expected score has in r0-r1 as paremaeter as get smaller then gain gets smaller if wanna work it out with that rate gain it takes more than 100 games to get one full point :)

Another point is the rating deviation. If a player won every game, wouldn't his rd not stay at a high level, and thus making him gain many points per win, even against much weaker opponents?

Another point is the rating deviation. If a player won every game, wouldn't his rd not stay at a high level, and thus making him gain many points per win, even against much weaker opponents?

no, it goes down for every game and only very little up after a "expected result" so it would hit the minimun. Glicko on seees last game (or patch of games sent for calculation). Just estimated from formulas but I am about 76.2% sure of it

no, it goes down for every game and only very little up after a "expected result" so it would hit the minimun. Glicko on seees last game (or patch of games sent for calculation). Just estimated from formulas but I am about 76.2% sure of it

I think its unlimited, or if it stops at a point and you dont get rating, it doesn't say: "My dear player of lichess, you reached the highest possible rating, and you cant have further rating, because now you are too good!" XD

I think its unlimited, or if it stops at a point and you dont get rating, it doesn't say: "My dear player of lichess, you reached the highest possible rating, and you cant have further rating, because now you are too good!" XD

@tpr I defer to your experience and knowledge of the game - however, the logic doesn't seem to follow.

There is also the tendency for mutually perfect games to result in a draw - which would affect rating growth. I can't find the reference but iirc, it was theorized that at around 5000 or 6000 elo, players (AI systems) will begin drawing almost every game which will naturally limit increases in ratings. So regardless of how many players participate in the future, "unlimited" ratings don't reflect the natural limits of the game itself.

@tpr I defer to your experience and knowledge of the game - however, the logic doesn't seem to follow. There is also the tendency for mutually perfect games to result in a draw - which would affect rating growth. I can't find the reference but iirc, it was theorized that at around 5000 or 6000 elo, players (AI systems) will begin drawing almost every game which will naturally limit increases in ratings. So regardless of how many players participate in the future, "unlimited" ratings don't reflect the natural limits of the game itself.

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