Yep same I went from 10k to 14k and then up to 7k
Well you still need to win your games to get a nice ranking
Kay, match-making is broken. Better get some sleep.
Welp, somehow reloading doesnāt work for some reasons, I need to re-login for it.
Also, Does blitz rank gets reset?
Thatās great!
Well, this is cool. Now that Iām SDK Iām looking forward to trying out that scary ā4-4, low approach, low extension, contactā joseki! Anyone else got big plans?
Thanks everybody for the hard work, the time spent, the effort and everything.
Thanks especially @flovo.
Wellā¦ dust hasnāt settled yet, but Iām quite astonished by the results: seems like I am 3.1k now (no way!) and like Iāve NEVER been SDK!!!
Thatās quite weird, isnāt it?
I already spent (in the past) too much time complaining about possible flaws in ratings algorithm and strange mystery about it.
Looks like mysteries are gone, thanks @anoek very much.
I actually donāt care anymore.
I hide ranks by default and Iām fine with it.
Nonetheless it looks weird to be overranked and being SDK since the start.
Having just been beaten by katrain 15k (on windows) despite my newly alleged 8k Iād kinda prefer if we had expanded below 25k instead and maybe used slower averaging at that level to deal with the volatility issue instead of shifting the kyus up to fewer and wider levels.
Thatās just me moaning. I accept the devs have a bigger picture view than I have.
Something does seem a bit overdone about the ranking change.
katrain-18k[8k]
katrain-14k[5k]
katrain-10k[3k]
katrain-6k[1k]
katrain-2k[2d]
katrain-2d[4d]
Particularly the weaker bots seem super over rated, knowing how they play.
And stronger bots are underatedā¦
Cannot accept custom games ever since the ranking update. It is displayed with a green button in the list but when I click on it it says " Your rank is too low to accept this challenge."
Can you please fix this bug?
Is there a new link for API type stuff? I wanted to look at my rating history to see exactly how much things had moved, so I loaded up my saved link at https://online-go.com/termination-api/player/165439/glicko2-history, but while the games are up-to-date (well, I played one since the update) the ratings ballpark seem to be pre-update.
Is this just the update has not been completed yet? Or is there a new link I should be looking for?
EDIT: looked through GitHub for information and found https://online-go.com/termination-api/player/165439/v5-rating-history, nevermind
Both are reality. You can expect a stronger effect of you played only a few games per rating period. Rating uncertainty stayed at ~60 for players with high play frequency.
I was also surprised that the flattening effect was barely noticeable. For comparison my graphs:
EDIT: My mistake, I didnāt pay attention to the fact that the ratings are more widely spread in the newer graph.
So the update has hit for me, I went from a 15 kyu to a 10 kyu, but on the app I dropped to an 18 kyu for some strange reason. I dont know if others have pointed this out yet but the app seems to have taking everyoneās ranking in the opposite direction for some reason
The app is third party and hasnāt been updated yet to accommodate the new ranks
On my graph it appears rating changes are half of what theyāre used to be. Shape of the graph is the same though. In some sense, that sound great, because we get the same-ish result with fewer changes to the rank.