Lichess vs Chess.com ratings, explained
The same player is usually rated higher on Lichess than on Chess.com. Neither site is wrong — they are two separate populations with two separate starting points.
The two online scales do not agree, and never did
A Lichess rating and a Chess.com rating are different numbers describing the same player, produced by different systems from different opponent pools. Neither is wrong. They are simply not the same scale, and the difference between them is large enough to change what people believe about themselves.
The anchors below were measured here rather than repeated from the internet. 852 titled players hold an account on both sites and could be matched with confidence — 453 pairs in blitz, 375 in bullet — and the table is what those pairs actually show.
Below roughly 1700 the anchors are the previously published community estimates, and this page marks them as unverified because they are. The measurement cannot reach down there: a sample of titled players is a sample of strong players, and no amount of arithmetic turns it into evidence about 1200-rated ones.
| Lichess | Chess.com |
|---|---|
| 800 | ≈ 400 |
| 1,200 | ≈ 800 |
| 1,600 | ≈ 1,250 |
| 2,300 | ≈ 2,201 |
| 2,500 | ≈ 2,385 |
| 2,700 | ≈ 2,647 |
| 2,900 | ≈ 2,794 |
| 3,100 | ≈ 2,954 |
| 3,300 | ≈ 3,089 |
| Lichess | Chess.com |
|---|---|
| 800 | ≈ 450 |
| 1,200 | ≈ 900 |
| 1,600 | ≈ 1,350 |
| 2,100 | ≈ 2,271 |
| 2,300 | ≈ 2,408 |
| 2,500 | ≈ 2,624 |
| 2,700 | ≈ 2,826 |
| 2,900 | ≈ 3,068 |
| Lichess | Chess.com |
|---|---|
| 800 | ≈ 500 |
| 1,200 | ≈ 950 |
| 1,600 | ≈ 1,400 |
| 2,000 | ≈ 1,850 |
| 2,400 | ≈ 2,300 |
| 2,800 | ≈ 2,750 |
| Lichess | Chess.com |
|---|---|
| 800 | ≈ 500 |
| 1,200 | ≈ 950 |
| 1,600 | ≈ 1,400 |
| 2,000 | ≈ 1,850 |
| 2,400 | ≈ 2,300 |
| 2,800 | ≈ 2,750 |
A conversion is an approximation of a population, not a promise about one player, and the measurement says that far more loudly than the old wording did. Within a single 200-point Lichess band the paired Chess.com ratings spread 700 to 1,400 points. The anchors are medians; an individual can sit hundreds of points either side. Where this site knows both of a player’s accounts it shows the real number instead and drops the ≈ entirely.
The gap is not constant, and it changes sign
The received wisdom is that Chess.com ratings run a few hundred points below Lichess ratings, full stop. That is true through the middle of the range and false at the top, where it reverses: above 2100 the measured Chess.com blitz medians run between 104 and 171 points HIGHER than the Lichess ones for the same players.
This site shipped that backwards. An earlier version of the conversion table applied the middle-of-the-range gap everywhere, which meant a genuine 3,001 blitz rating was being converted into something absurd. Measuring it is what caught it, and the corrected anchors are what the table shows now.
The lesson generalises beyond this table. A single conversion constant between two rating pools cannot be right, because the pools do not overlap evenly at every level. Any tool that offers you one number to add or subtract is describing one part of the range and quietly extrapolating the rest.
What an online rating says about a FIDE rating
Less than almost anyone wants it to. The table below takes titled players who hold both a Chess.com rating and a FIDE one — 8,256 of them in blitz, 4,254 in rapid and 1,894 in classical — and reports the median FIDE rating in each online band, with the spread of that band beside it.
Look at the spread column before the median column. It runs to over a thousand points in several bands. Two players with the same Chess.com rating routinely hold FIDE ratings hundreds of points apart, which means the online number predicts the official one about as well as a coin predicts the weather.
And every player in it is titled, which is the part that decides what the table may be used for. A titled player with a low online rating in some control is not a weak player; they are a strong player who does not play that control online. So the medians run far higher than anyone expects, and the shape at the bottom of the table is an artefact of who is in the sample rather than a finding about anybody at that rating.
Read it as a warning rather than a converter. If you take one thing from it, take the spread; if you take two, take the fact that these are not people like you unless you happen to hold a FIDE title.
| Online rating band | Titled players in it | Their median FIDE | FIDE spread in the band |
|---|---|---|---|
| 800–999 | 38 | 2,123 | 1,171 |
| 1,000–1,199 | 93 | 2,143 | 949 |
| 1,200–1,399 | 151 | 2,185 | 944 |
| 1,400–1,599 | 292 | 2,200 | 1,094 |
| 1,600–1,799 | 378 | 2,223 | 1,105 |
| 1,800–1,999 | 416 | 2,268 | 920 |
| 2,000–2,199 | 317 | 2,255 | 868 |
| 2,200–2,399 | 124 | 2,317 | 909 |
| 2,400–2,599 | 37 | 2,306 | 711 |
| Online rating band | Titled players in it | Their median FIDE | FIDE spread in the band |
|---|---|---|---|
| 1,200–1,399 | 34 | 2,119 | 1,093 |
| 1,400–1,599 | 69 | 1,970 | 1,150 |
| 1,600–1,799 | 152 | 1,948 | 1,021 |
| 1,800–1,999 | 362 | 2,002 | 975 |
| 2,000–2,199 | 1,067 | 2,120 | 1,159 |
| 2,200–2,399 | 1,433 | 2,216 | 1,036 |
| 2,400–2,599 | 869 | 2,351 | 910 |
| 2,600–2,799 | 205 | 2,501 | 876 |
| Online rating band | Titled players in it | Their median FIDE | FIDE spread in the band |
|---|---|---|---|
| 1,200–1,399 | 24 | 1,714 | 854 |
| 1,400–1,599 | 57 | 1,794 | 946 |
| 1,600–1,799 | 121 | 1,830 | 1,080 |
| 1,800–1,999 | 255 | 1,914 | 878 |
| 2,000–2,199 | 663 | 2,010 | 1,162 |
| 2,200–2,399 | 1,457 | 2,093 | 1,115 |
| 2,400–2,599 | 2,366 | 2,190 | 939 |
| 2,600–2,799 | 1,949 | 2,269 | 953 |
| 2,800–2,999 | 1,007 | 2,370 | 791 |
| 3,000–3,199 | 295 | 2,489 | 875 |
| 3,200–3,399 | 34 | 2,647 | 438 |
Measured across 10,379 titled players holding both a published FIDE rating and an online account. Each row describes a group of players — it is not a lookup for your own rating.
What none of this can tell you
It cannot tell you what your FIDE rating would be. Nothing can, short of playing the games — the whole point of an over-the-board rating is that it measures performance in conditions online play does not reproduce.
It cannot compare a bullet rating on one site with a rapid rating on another and produce anything meaningful. Different time controls are different skills wearing the same units.
And it cannot tell you which site is harder, because that question does not have an answer. A pool is not difficult or easy; it is a different set of people, and your rating is a statement about where you sit among them and nowhere else.
The same study covers Lichess and this page does not show it: 7 titled players hold both a Lichess classical rating and a FIDE one, which is a row that would look like evidence and is not. Any control with fewer than a hundred pairs is dropped rather than printed small.
Common questions
Why is my Lichess rating higher?
Different starting points and different populations, not different skill. Your position within each site is the part that compares.
Which rating is the real one?
Both are real inside their own pool. If you want one number that travels, use your percentile rather than your rating.
How accurate is the conversion?
It is an approximation of a population, not a promise about one player. Two people with the same Lichess rating can sit a hundred points apart on Chess.com.
Data updated 2026-08-02
Chess.com publishes no rating distribution. Those figures are derived from the Lichess distribution mapped through the conversion anchors, and are rounded accordingly.