I kind of expected something similar to that. If in the old system someone starts at like 25kyu and played other 25kyus then eventually got better, playing people roughly equally matched along the way, does the new system put those new players at like 11kyu (was it?) and then because the new players are playing other new players on the old system (assume they’re 11kyu in the new calculation) doesn’t the rating just settle at around that point and likely go up from there?
Then only players that were dropping a good bit below 25kyu in the old system, I would imagine would also drop down to lower ranks in the recalculated new system?
Not necessarily, as if you started as 25k in the old system, even if you only played other 25ks, some of those would be playing 24ks who played 23ks who played 22ks, etc until some number of degrees of separation would cover the entire rating population. Both systems should theoretically converge at the same point (relative to the population as a whole) if A) you played enough games (whatever this means) and B) there was sufficient play in the general population between ranks
I don’t think the ratings in the chat have been thrown off in every game though.
I have an old game where there’s a comment about how the players are 18kyu and they appear as 18kyu in chat.
I think there was a change in formatting in the API of the games at some stages over the years, so maybe it’s pulling the chat ranks from somewhere different in old games vs newer ones? (compare early 2017 to 2019 say)
I think someone forgot a ! somewhere >.< When I de-select “Golden supporter name” it makes me gold in group chat, and when I select it, I am black! Edit: actually it seems there’s just a lag before that setting is respected in the system. I thought it was opposites because I kept switching back and forth.
The other possible “bug” is that information about showing supporter status is persisted in the chat. If someone wanted to hide their status, wouldn’t they want it to be hidden for past chat actions?
There are many arguments involved, but if you accept a superficial answer look at post 420 above.
Look for points with x = 26 (4k), then look at their y.
Looks like we had both situations: someone became dan, someone else nearly DDK.