I am intersted in one interesting thing.
Suppose we sum all ratings of all users in one given time and divide this sum with number of players.
Is this average exact 1500, more than 1500 or less than 1500?
This question interested for me too.
Some time ago, a 1500-rated player was around 10.000-th place out of 200.000 registrated players, which is top 5%.
So 5% of all players were above 1500, and the other 95% were below.
One could assume that average is way less than 1500. But we don't know distribution of data.
Don't know much, but one could also expect, that this could be some kind of normal distribution with expacted value around 1500 and deviation around +/- 125.
I guess we should consider only players with, let say minimum 100 games played.
What would mean if we took average, weighted with numbers of games played?
Maybe it wold be interesting to se average rating of players online...like number of players online and number of games played.
I've noticed one interesting thing also.
If there are N players online, there is N/3 games in play, which means (chess is a game for two), there are 2/3 players playing and 1/3 doing something else.
bump
I've wondered about this too.
That's a feature
Chess.com offers, but isn't anywhere on Lichess.
Thibault, step up your game, brah! :p
Actually, he ran the numbers just under 2 weeks ago. You can see the results of that query here:
gist.github.com/ornicar/2a402e46d05074f880d6Well that was awkward ...
Thanks hellball!
I must say, you've been on quite the roll.