Free online Chess server. Play Chess now in a clean interface. No registration, no ads, no plugin required. Play Chess with the computer, friends or random opponents.
Create a game Tournament Simultaneous exhibitions
Chess basics Puzzles Practice Coordinates Study Coaches
Lichess TV Current games Streamers Broadcasts (beta) Video library
Players Teams Forum Questions & Answers
Analysis board Opening explorer Board editor Import game Advanced search
Sign in
  1. Forum
  2. Lichess Feedback
  3. Rating Inflation & Brand New Accounts. This is not a Cry thread.

I am NOT Complaining, or making a worthless thread.
I am simply pointing out my observations on this issue; as I have not seen people say these things.
If its been talked about before; I apologize.
If not; well listen up :)

A) You should not be able to receive a rating over 2K with a brand new account. This is the #1 reason people multi account.
Make a new account; try to win all 15-20 in a row. Hit 2300-2500.

A similarly skilled player will have a MUCH more difficult time trying to get a 1700 or less account higher than 2K.

Look at the top boards. 80% of them are made up of people who started with 2K or more. Never varied.

And, once you reach that rating its nearly impossible to drop under 2K. Play tournaments; win 50% of your games and your rating will still rise at least 10 points if not 30-50.

On the variants its even worse; almost 100%. Crazyhouse IS 100%. Not one single person on the top list who reached their level from a 'normal' rating. All 2K and up, some as high as starting with 2.5K and never going lower.

People say ratings dont matter to a game; they do.
People play differently against a low rated person vs someone over 2k. A sense of fear; or doubting yourself; or simply thinking you dont have a chance leads to many losses that shouldnt happen.

B) Tournaments also heavily inflate ratings.
One should not be able to win 1 game draw 1 game and lose 4 games in a go UP 6 rating points.
Like I did there :l

Similarly played games outside of a tournament have a much greater effect on your rating. I have no idea why this is.

C) And playing a game against a higher; or similarly; rated person should not have such a little effect on your rating.
If I play someone with 2.2K plus and lose I will only 1 or 2 RP. Sometimes its even zero. This is a non-loss scenario. You can go up 30-50 points if you manage a win but lose 0-2 if you lose?

Heres a good example of that:
I've played this guy 13 times.
I've lost 10 of the games and won 3 of the games.

I went down in rating right? No. I went up 5 or 6 points after you add them all up. (I came up with 6 but Im tired and added em in my head lol but def positive)

So 3 out of 13 and its a positive rating boost? Lol.

Hope ya'll are understanding; I'm saying I should be a LOWER rating not higher. Not crying about MY rating; or really anyone elses. I'm talkin bout the system and its appearing to me to be very broken.

Not sure how or if it could be fixed; but been noticing this stuff so thought I'd share.


I do not understand. People who are „good“ have good ratings and they should reach it soon. Everything‘s consistent, isn‘t it?

You have to admit - It does look a little bit like a cry thread.

I think you have a misunderstanding of how the rating system works. You can read about it in detail on the wikipedia page and its links (, but it's pretty technical and doesn't really go into the "whys", so I'll explain the best I can.

First, a rating system is not a reward for winning. It's not a score or even an absolute measure of strength. It's nothing more or less than a predictor of how you will perform in future games. This isn't exact, but if you're X points higher rated, you're expected to win something like Y% of the games:
X=100 Y=64%
X=200 Y=76%
X=300 Y=85%
X=400 Y=92%
And so on.

When you have a new account, you start with a high RD (rating deviation). A high RD means that your rating is very uncertain, so you win gain or lose more rating points each game as it tries to quickly get you to closer to your "true" rating. As you play more games and your rating goes up and down, your rating becomes more certain, your RD goes down, and your rating adjusts more slowly.

When you play someone with a high RD, your own rating will change less, because playing someone with an uncertain rating gives less information about your own performance. That's probably why you observed games outside tournaments affecting your rating more, because you happened to play people with a lower RD.

You seem to think that if you lose more than half your games, you should lose rating points. But really, if you think about ratings as a predictor of results, that's not always true. Your rating will only go down if you're performing worse than your current rating predicts. Some examples:
- I score 30% against an opponent I'm expected to score 30% against (according to our rating difference). My rating stays the same, with the wins and losses evening out.
- I score 50% against that same opponent, I'm performing like a player higher than my rating, so my rating will go up.
- I score 90% against an opponent I'm expected to score 95% against. My rating will go down.
And so on.

If that seems weird to you, maybe think about this - if I scored even 1/10 against Magnus Carlsen, I'd probably be a very strong player! And my rating would probably go up (unless I was already rated very highly).

Anyway, that's the explanation. There's a lot of math and real-world data that says the Glicko-2 system (which lichess uses) is pretty good at this.


Did you even read what I wrote?
How on earth could you call it a cry thread if you did. I am...talking about MY rating being inflated; and what the causes appear to be.

Go back to trollin dinnerspace. Leave me be timid lol


Thats just not what I'm seein on the leaderboards; as well as a large % of the 2.2K+ people I run into. First day on the account they rate over 2K. Never drops under.
Entire ZH leaderboard is this way; as well as around 90 of the top 100.

The good stay good. The bad stay bad :o

People who 'get better with time' do not seem to be the norm.
People who are good were always good; at their game (whatever variant it may be)
Which, is very improbable; that there is no 'learning curve'

The logical explanation for this, to me, is that when someone decides they are good enough for a good rating; they create a new account.

Which, is what A) was referring to.


Thank you for the well thought out response. <3
Unfortunately, it is 3:58 AM Where I am, and I am out of it.
Your post warrants a fresh mind; so I can actually comprehend it ; as it currently looks as difficult for me to understand as the quantum field theory lol. I will read it tomorrow and re-edit this post or somethin

I think the opening % of win/loss points for new players is a bit too wild. IE if you win or lose your first few games you will fly up or down 500+ points.

New players should have to fight their way up the ladder. If they are genuinely good players, they will in time get there.

I think it causes a bit of inflation on other player playing them and grabbing points off new players that are way over rated.

I also think Lichess in general is too inflated on average by about 100 - 150 , compared to other sites and also compared to OTB.

My rating in real translates to 1750 ish, but on here I can comfortably play in the 1800s or even higher if it's blitz (I always move very fast).

Not complaining, I love this site and the ratings make you feel better but they are definitely higher than general chess elsewhere.

@rickrenegade lichess uses Glicko2, not ELO. Thus, the ratings aren't exactly "inflated," since lichess uses one system and (for example) uses another.

@breakreign if you aren't 2000 quality, it will be very darn hard to stay above 2000 if you get there. As far as the people who jump right up to 2000 and stay there, that is because they are 2000 material, not because they got off to a good start. As far as your (to paraphrase) "more losses than wins but my rating goes up" example(s), it's because you played opponents better than you. For example, if a 2500 beats me, he'll get no more than one or two rating points and I'll lose no more than one or two. If I beat him (which probably won't ever happen) he loses (and I gain) well over 20 rating points, maybe even 30. The reason for this is that it should be hard to beat him and easy for him to beat me. Whoever accomplishes the more impressive feat gets more points. In this example, one win from me would need close to 30 wins from him in order for him or me to have a net points gain of zero. It's a bit of an extreme example, but I think that you get the idea.

EDIT: @cyanfish has a better explanation of this.

A vivid example: look at my curves :D

I reached my „true“ strength (rating) here rather quickly and will stay there for ages as long as the server is drifting slowly or I am becoming really old. So Elo +50 or + 100 should be the cap here which fits nicely in the one or other approximation. ;)

Just meaning: a skilled or educated player reaches his cap quickly. If there‘s some progress over the years it is often not visible because the rating pool of the players changes overall.

You will rarely see curves with constant rise (d rating / d time positive) over very long periods except for beginners or youngsters.

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