provisional players with higher level than their rating.

this happens a lot in classical, as most people don't have time for it and play maybe once every 3 months, or have just recently started playing classical. now it is hard to tell what would be their classical rating if they play blitz/ fast rapid a lot more often, as they again don't have time for classical or slow rapid. I encountered such a player today, who was ~1500 blitz, ~2000? rapid and 1770 classical. after I got completely crushed I checked his/her classical, rapid and blitz games. the player's rapid games were extremely old, but the most recent one (8 months ago) features the player crushing a 2300, which is 500 points above my rating, blitz games showed the player was prone to blunders if not enough thought is given, and in classical, the player DOES have time to give a lot of thought. now finally the classical games showed the player losing to provisional players who are currently 1900 or players who are ~1900 already most of the time, yet I lost 7 points to the player, who even after the game had 88 RD, so the player was just not provisional.

#1 seems a description of provisional ratings.

I'm not sure you're reporting an issue or just venting? Do you have any suggestion?

I have encountered the problem @Akarsh_2010 mentions frequently; particularly in the weekly classical arenas. After some statistical analysis on a sample of players with at least 30 games each in both rapid & classical I find ratings were strongly correlated*. A suggestion for a solution would be to set the initial provisional rating for a classical player = their established rapid rating rather than 1500, as this is more likely to be reflective of their true classical strength.

* roughly 0.88; for others who wish to reproduce: pastebin.com/DpHcA7ti

@bughouse26 said in #3:
> A suggestion for a solution would be [...]

But why a solution? What is the problem you're trying to solve?

A provisional rating is there for a reason, it's not something to fix.

@pepellou said in #4:
> But why a solution? What is the problem you're trying to solve?

When Player X is underrated and defeats Player Y, Player Y will lose more rating points than he/she otherwise should had Player X been more accurately rated. In practice this seems to only matter in the classical time control when strong rapid players play their first set of classical games.

> A provisional rating is there for a reason, it's not something to fix.

I'm only suggesting a change in how the _initial_ provisional rating is set for classical players. Rather than start at 1500, for players who have already established ratings in rapid time control then use their rapid rating as the initial classical rating given there is an 0.88 correlation in ratings between both time controls.

Thanks for considering.