same here 
But i don't play much there so idk if it's accurate
Sorry another comparitive rating question
Sort:
Yes. The puzzle rating on chess.com makes no real attempt at being comparable to any real rating system.
At Lichess - I'm not 100% sure of this but I believe it to be true - the puzzle is seeded at a rating that is based on the players who played the actual game. The average, the winner, I don't know, but it's based on the actual game. Then, that puzzle's rating is active. If a player with a 1300 puzzle rating solves it, the puzzle's rating goes down. If a 2000 player gets it wrong, the rating goes up considerably. But the puzzle's rating is, forever, active.
With chess.com - and this I'm 100% about - the puzzle starts with some random rating, then once it has been played 100 times, the rating is recalculated based on the results and the puzzle rating of the players who played it. At that point, the puzzle rating becomes fixed.
100 games isn't a lot, so sometimes you get relatively simple puzzles with high ratings, and absolutely impossible puzzles with lower ratings. But for the most part, the ratings seem accurate as in a 1900 puzzle is harder than an 1800 one, and not as hard as a 2000 one.
It seems to be the general wisdom that player ratings on LIchess will be about 300-400 higher than on this site. But my puzzle ratings are actually lower on Lichess .......anybody else found this?