Difference between chess.com/lichess rapid rating

Hi 

I was wondering why my rating of 1275 on chess.com  rapidis about the 90th percentile of  players and myrapid rating of 1600 is about 60th percentile on lichess ? Is there different type of players on the 2 sites thanks Roger 

Lichess has a different rating system I believe

hi

There is a significant difference besides the ratings. Someone stated, "On lichess everyone is higher because it’s easier there until a certain point". This is also true of the lower ratings. You can get the same dynamic if you drive your account rating down to the floor (~300) and then play. Some will look at this account and say, "But you just joined." like that proves something. No, I just make a new account. When you drive, sometimes you go left, sometimes you go right, and after that you go straight until you have to turn again.

So, what else is there? On lichess, they jack up your rating. I got 272 on my previous account there with my first rapid game. It was a fluke win. I was playing someone 1600 for my first rapid game. So, my next opponent then was around 1872. If you assess your level between 1200-1500, why play someone at 1600?

Next thing, this false narrative that you are "tilting". No, you are not tilting. This is like seeing infomercials in the middle of the night trying to sell you sleeping pills. What is happening, and these sites won't admit it, is you aren't playing humans. The "players" are playing based on computer accuracy against you. As long as the final outcome is within a certain range, it looks like a legit game. You blunder a queen, and then they blunder a queen later on. You do a knight fork, they do a knight fork. The patterns are endless.

The typical one is where they give up a knight. Now, this could be a real person higher rated playing with a handicap. Could be a rook also. The astute ones will drive home early opening forced lines. But, if you defend well and reach a certain move number between 10-15, they fall apart and go from 1500-2000 play to 800-1000 play. This is NOT normal.

One thing to consider is group play. On lichess, I find it a lot easier to join tournaments. On this site, home page lists "League Division" with the cup icon. I guess you could challenge then individually. Does anyone do that though? Then there is "Play 10 min" (rapid), and this randomly pairs you up. After that, we have "New Game" which links to the same thing, and play computer as well as play friend. No tournament link seen.

The way it is, you have to click within layers to get to tournament play. It exists, it is just not as easily clickable as lichess. What this does is focus more on you playing random individual games. If you play tournaments, the players by rating tend to perform worse than if you were randomly paired.

I hazard to guess this is going to drive rating points up too. The random pairing system is highly suspect. You can't talk to these "opponents". You can chat it up before a tournament though.

Something to consider. Have a nice cup of matcha and don't fret. Assess your own chess. Don't let these chess sites do it for you.

Indeed

The difference in rating is caused by a different overall rating distribution on both sites.

The average on chess.com is approximately 800, while on lichess, it's closer to 1500 (the starting rating). Thus ratings on lichess could be considered to be "inflated", but in reality its simply a different distribution.

For example, take the FIDE rating system.

Imagine FIDE increased everybody's rating by 1000 points, and now magnus was 3850 elo.

It wouldn't make the new system any easier, but more that the average rating is now 1000 points higher, and thus the old rating of 2800 would be of similar difficulty to reach as 1800. Sure, it's easier to reach a certain rating, but only because that rating doesn't carry the same value as it did in the previous system. For example, hitting 2500 in the new system would be equivalent to 1500, but the new GM threshold would be 3500.

TLDR: The distribution of player elo on the sites is wildly different, with lichess having a higher average elo.

"Imagine FIDE increased everybody's rating by 1000 points, and now magnus was 3850 elo."

But if you increased Magnus's rating to 3850, it would give room for finer evaluation. So everyone wouldn't get 1000 points added. A formula could be set up with a 2600 who has a history of getting 50 ELO points each year for example. In this case, you could go back 2 years from when they became a GM. If that is not enough, then 5 years. This means 250 would be evaluated differently with the subsequent 50 points being weighted less each time.

Just throwing out numbers here to illustrate:

50 x %100, 50 x %80, 50 x %60, 50 x %40, 50 x %20 = 50 + 40 + 30 + 20 + 10 = 150

So, now we have 2750. A second variable could be used. How many tournaments? If they played in more tournaments in a couple of the years and say those years had 60 or 70 point increase, then we could go beyond 2750. This would reward those who are active, not losing, and getting higher rating points.

You could set a minimum, say more than 3 tournaments a year. If they don't play more than 3 you knock off 5-10 ELO points.

You get the gist, Magnus going to 3850 stretches the bar, but it doesn't bring everyone along. You would have to be actively playing and getting points based on something as I outlined.

Skill issue