lichess.org
Donate

Chess here is tight!

Whew! I'm pretty mediocre but play here is very accurate for the most part. I managed to stay around 1900 uscf but can't get much above 1500 here! made a few daft blunders as one does but 1400 here seems to be equal 1800 uscf. Are chess players just getting better but the ratings don't reflect it? A few yrs ago 1400 would be making silly mistakes all over the place, not so much the case today
1400 on Lichess is about 1700 USCF / 1600 FIDE, I think :)
@peelhead
Apart from the answers you already got :

otb chess can be giving very different results for some players than online chess.

My online chess is better because in otb chess I can be easily distracted by noise, by (non chess) remarks of fellow chess club members, the look of the bartender etc.
Also, otb chess club games and tournaments are often something like 2 hours for 40 moves, and then sudden death with Fischer clock.
That is much slower than 10 0 classical here.
Which might be good for some, and bad for others.
Impatient blitz players like me can get bored when the opponents is thinking for 30 minutes, and that can sometimes badly effect the game position and result.
For me it also happened that I got good positions and then started to feel really tired, and lost my head in the game, with draw or loss as result.
In team matches I am often very motivated to win, or to make the team win.
In internal chess club competition I'm not motivated, it is more a social thing and a weekly mental exercise.
So ... my results go up and down, and your results might do the same, otb and online.

Having said that :
On chess.com my corr. chess rating dropped below 1900, while on lichess it is still above 2200.
The pool of players is for sure different as well.

Well, my FIDE rating is about 2283 and my Lichess' rating was changing between 2360-2450. In fact, I get use to lose rating against cheaters and another strong players around here...; also, you can compare the OTB with the internet chess, and, standard time with blitz. :)
With a standings list, a fictious rating could be established to say the top player is say 2500 rated and the bottom in the standings list a 1200 rated player. So ratings is not as accurate as we might think.
To get some enjoyment, where you stand amongst a category of players is usually more important than the rating.
If you want to be first in a list of players, play in tournaments that have lower rated players. You rating will tend to be over inflated.
If you want to be last on a list of players, play in tournaments that only have high rated players. Your rating will tend to be underinflated.
If there was a way to apply centipawn rating averages into the present rating system our ratings would be recalibrated.
If a player did not try to play accurately, the centipawn would increase, so the rating results would not increase or decrease the rating as much compared to when a game had a lower centipawn rating.
That would be the end of the shark player ratings, because a shark could not truely reduce their rating by playing poorly.
Conclusion:
I believe ratings should be given after a game is analysed.
The K factor should change depending on the centipawn result.

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