The curious case of 1200: The Expert's rating

The word "expert " antagonizes me it's more of a intermediate

In OTB play using classical (slow) time controls, a 1200 rating is the bottom of Class D, which is more than a standard deviation below the average club or tournament player.

people shouldn’t get upset by players starting at different ratings. The rating system is self-correcting.

suppose two players have the same „true” skill level of 1200. Now suppose they start at different ratings, say 400 and 1200.

After they play 100 or so games, they should both be rated 1200. the only difference is that one started at that level and the other took some time to get there.

I started at 400 just for fun.

This rating system shouldn't let people just pick any ol starting rating, that's stupid. But mine did average out quickly. The problem with starting at any rating is that it impacts the opponents rating more early because it's assumed they're playing a 2000 when they're really playing a 800 for example. It tinkers with the ratings unfairly.

I don’t know how chess.com assigns initial ratings. 

in the USCF, ratings for the first 24 games are provisional and reflect one’s performance rating in those games. That provisional rating is then used in determining one’s opponents change in rating. This helps stabilize the ratings. So if a new player plays at the 1700 level, that is the rating used to calculate how much one’s opponents’ ratings will change.

allowing players to identify if they are beginners, advanced, or whatever is probably a sales tool yes, the ratings are inaccurate for many, but so what? The ratings are meaningless. And if one plays enough. The ratings move to the true level.

the big problem that can result from setting one’s own initial rating is inflation/deflation. If more people underestimate their starting level than overestimate it, then there are more points taken from other players than are given to other players as the newbies move toward their actual skill level this deflates the ratings and of course the opposite would inflate them.

but in the end, who cares? Ratings online are meaningless.

I started at 1200 in 2012, then progressed up to about 1450. A few years ago, the average rapid rating was 1100. Now it is 620. This is a huge swing in the distribution. Chess.com moved the goal posts. My rating now is 1250 or so, but my percentile is about 91, which is higher than it was when my rating was much higher. They changed the definition of an average vs a good player. Has anyone else lost 200 or so points?

LOL.

But that wasn't recent. It was years ago.

Yeah leaderboard is cursed. First, numbers don't add up, they don't match the declared player count. Second, the fact that millions of players are stuck around 100-200 Elo (with 100 Elo as the floor) is telling that something is wrong. And randomness of skill among those 100-200 rated players is just crazy. People keep fantasizing about how it should stabilize after 100 games or so. But those 100-200 rated players who are either weak or strong aren't newcomers, they play thousands of games for years sometimes.

Not to mention that no human who has some thoughts in head should be rated 100 EVER. Random-move bot has higher rating.