5

I have been playing on chess.com and lichess for more than 2 years. My rating in blitz on lichess is around 2100, and on chesscom is around 1800. Yet I highly believe that the players on chesscom are much more accurate. Like sometimes I look at the position and they don't do any blunders, 1 mistake, and a few inaccuracies in a position that's really tough to play. Like I would beat them with good moves (not the best moves) and then I would be surprised at the end that overall they played better almost every time! This is very strange. I don't know if there's more cheating on chesscom or what? I would really appreciate it if you can give me your feedback on this topic because it has been driving me crazy. Like I find it easier to beat a 1900 player on lichess than a 1800 on chesscom which is insane!

  • 16
    Have you considered that chesscom and lichess calculate ELO in a different way? – crrlos Jun 19 at 2:55
  • 4
    @Guess601 you could make a new rating system where everybody starts at 500 and you'd see that only GMs ever reach 1700 – David Jun 19 at 6:48
  • 6
    Lichess doesn't even use Elo-, but Glicko-Rating (similar, but not the same!). You really are comparing apples and oranges here. – Annatar Jun 19 at 7:00
  • 8
    Or put differently: It's much easier to become a millionaire in Japan than in the US - without being actually richer. – Annatar Jun 19 at 7:04
  • 3
    You say: "My rating in blitz on lichess is around 2100, and on chesscom is around 1800." and "Like I find it easier to beat a 1900 player on lichess than a 1800 on chesscom which is insane!", 1900 rated player on lichess is 200 points below you and 1800 rated player on chess.com is on the same level as you; it should not be surprising that it is easier to beat 1900 rated player on lichess. – Akavall Jun 19 at 16:06
5

Let's look at this from another perspective. Instead of assuming that the problem is just that the ratings at different sites mean different things, let's take that out of the equation. Let's look at the reasons that you could observe that opponents at your rating on one site play fewer mistakes than opponents at your rating on another site.

Your opponents might only appear to be stronger on one site:

  • If you are using the computer analysis from the different sites to find mistakes, the analysis on one site might be different from the analysis on another. The threshold for "mistake" might be different, or the analysis might be stronger or weaker, which affects how many mistakes are found.

  • If players at a particular site are making larger mistakes, they could make fewer of them. (Or, conversely, they could be making many very small "mistakes" that are below the computer's inaccuracy threshold.) Relatedly, if players at one site tend to play more strategical rather than tactical games, this might let the computer think they are stronger, since they aren't playing as many obviously bad moves.

  • If players at a particular site are more prone to running out of time, they could lose some of their games without playing any weak moves. (This doesn't really make them a stronger player; time management is a big part of blitz.)

Your opponents on one site might be stronger in some situations:

  • If, on one site, you tend to play more Swiss tournament or team match games, it could be that your opponents are taking those games more seriously than their other games, causing them to play stronger than their rating would ordinarily indicate. Arena style tournaments, on the other hand, might get some of your opponents looking for a quick knockout rather than playing the objectively best move, so they can get more games in before time runs out.

  • Time of day and the geographical distribution of players might even matter. If one site has a larger number of players from a particular time zone, and if the time of day you tend to play is such that players from that time zone are on but would be tired, that could make them play weaker than their rating would ordinarily suggest.

Your average opponent on one site might actually be stronger:

  • If a site has a large number of players who only play a few games and then quit or make a new account, this could have an impact. A player might be 2200 strength on a site but only have a rating of 1800 because they simply haven't played enough games for their rating to catch up. (As a sub-category of this, this could happen if players are caught cheating and have their accounts deleted and make new ones.) Under the Glicko and Glicko-2 systems that these sites use, new players don't affect the rating of established players as much, so you wouldn't lose many rating points by playing them, either.

  • The issue could be with you! If, on one site, something causes you to play stronger (maybe you like the visual layout better?) then the rating you equalize at will also be stronger (even if it's a lower number than the other site) and your opponents will thus be stronger than at the other site.

| improve this answer | |
  • "If, on one site, something causes you to play stronger (maybe you like the visual layout better?)" I really thought about this reason a lot. But do you know anybody who has the same experience/reason? I dismissed it because I thought it's just an interface, but if many people are having this problem then it's a thing! – Guess601 Jun 20 at 15:56
  • I also thought about the time of the day. I think it plays a big factor too. It's good to find other people who also think so. I thought I was subjective with this reason too. – Guess601 Jun 20 at 15:59
  • 1
    The time zone thing is hypothetical; I don't have any evidence of where people play from at particular times of day. The interface thing may be something concrete you can point to (like being able to see the clock easier) or it could just be personal preference. It's very possible that some people would play better on one interface and other people would play better on another. – D M Jun 20 at 16:27
  • 2
    @Guess601, if you are using free account on chess.com, the ads can be very distracting; it would affect my play. – Akavall Jun 20 at 17:12
23

Like I find it easier to beat a 1900 player on lichess than a 1800 on chesscom which is insane!

Why is that insane? According to one site, an 1800 blitz rating on chess.com is about equivalent to a 1915 rating on lichess (based on people who answered their survey and have ratings on both sites.) So, your experience of the 1900 lichess player being easier to beat than the 1800 chess.com player conforms to reality.

I think the issue here is simply that an 1800 rating on one site does not correspond to an 1800 rating on another site. Even if they used the same rating rules (which they don't), the ratings would likely diverge due to differences in the player pools.

In the end, rating is just a number, and that number means different things in different places. Just because 1800 rated players are a particular strength on one site does not mean 1800 rated players will be about that strength on another site.

| improve this answer | |
  • I think that you're missing the point. How can an 1800 player on chesscom have very good stats almost in every game, regardless of the position that he/she is playing. Like you cannot have only 0/1 mistake, and 3/4 inaccuracies in every blitz game that you play. That was my point. Like do you really believe that an 1800 player on chesscom can do that consistently in blitz games? I don't believe so. – Guess601 Jun 19 at 14:00
  • 3
    @Guess601 I think D M's point was that "1800" means different things on lichess, chess.com chess24 and in real life... FWIW, my lichess rating is also considerably higher than the one on chess.com. – user1583209 Jun 19 at 15:39
  • my point was even an 1800 (or even 2000) player on chesscom shouldn't be able to find such moves in like 5 seconds! – Guess601 Jun 19 at 21:16
  • 17
    @Guess601 People are answering the question you actually asked, and you're following up in the comments with "There's no way a player could find good plays this quickly!" which (combined with a mention of cheating in your question) suggests that your real question is "Do lots of chess.com players cheat?" It would be more productive if you just asked this, instead of couching it as a rating comparison, which answers have pointed out is not surprising. – amalloy Jun 19 at 23:00
  • @Guess601 I understand that you think this isn't the real issue. I'm going to leave this answer here (because I think it's the most likely explanation, and because it has lots of votes), but I'm going to make another one where I try to think of other possible reasons why it could happen. – D M Jun 20 at 13:34
16

Elo does not signify how strong of a player someone is.

Elo signifies how strong a player is in comparison to some community.

Assigning a numeric score to rank how strong someone is in chess would be very difficult to do. The way Elo works (and even the way it is designed) is simply to keep track of players in some group and make predictions about which player in that group is more likely to win against another player in that group. Having a score of 2100 in lichess means that the system has predicted you are more likely to win if you play against someone on lichess who has a ranking <2100, and more likely to lose if you play someone higher.

Having a ranking of 1800 on chesscom means that the system has predicted that you are more likely to win if you play against someone on chesscom with a ranking of <1800, and more likely to lose if you play someone higher.

Since presumably you play at the same level on both sites, that would imply that these numbers are more or less comparable. There are a lot of reasons that a player would have a different Elo rating in a different community

  • One community could simply be more skilled, or targeted at different types of players/

  • If the different communities have different "starting ratings" like you mentioned.

  • If the communities have a different size. In a smaller community, it's harder to get a very high Elo score (Imagine this taken to the extreme. If you track your Elo against one other player, you will get diminishing returns from beating them. Even if you are the best player in this "community", you will never hit Magnus Carlson level Elo scores)

  • It is possible for Elo ratings to differ between communities because of different standards of preventing cheating. But I would not expect that to be the case here, given that both sites have (Lichess, Chess.com) have written extensively about how they try to prevent cheating. If you feel like playing on one of these sites is unfair, reading through those articles can help you judge whether it's a fair place to play.

I know that there are many differences such as the starting rating of a new player on chesscom is 1200, but on lichess is 1500

That makes a huge difference. That means a score of 2100 on lichess is "1.4 times better than average" (which is an oversimplification of how Elo is really calculated and what it represents) whereas a score of 1800 of chesscom is "1.5 times better than average", which is a very minor difference.

| improve this answer | |
  • This should be the top answer. – lolololol ol Jun 20 at 4:33
  • 2
    Chess.com also uses Glicko and not Elo. – D M Jun 20 at 12:11

Not the answer you're looking for? Browse other questions tagged or ask your own question.