Are codechef ratings now officially a meme?
New rating changes resulting in 4* accounts becoming 6* in a Division 2 contest. Reference: https://discuss.codechef.com/t/codechef-rating/102223/4
Ps.: Go check your Codechef ratings you might have 1 or 2 extra star now.
| № | Пользователь | Рейтинг |
|---|---|---|
| 1 | tourist | 3690 |
| 2 | jiangly | 3647 |
| 3 | Benq | 3581 |
| 4 | orzdevinwang | 3570 |
| 5 | Geothermal | 3569 |
| 5 | cnnfls_csy | 3569 |
| 7 | Radewoosh | 3509 |
| 8 | ecnerwala | 3486 |
| 9 | jqdai0815 | 3474 |
| 10 | gyh20 | 3447 |
| Страны | Города | Организации | Всё → |
| № | Пользователь | Вклад |
|---|---|---|
| 1 | maomao90 | 174 |
| 2 | awoo | 164 |
| 3 | adamant | 163 |
| 4 | TheScrasse | 159 |
| 5 | nor | 157 |
| 6 | maroonrk | 155 |
| 7 | -is-this-fft- | 152 |
| 8 | Petr | 146 |
| 8 | orz | 146 |
| 10 | BledDest | 145 |
All users’ ratings have been recalculated for all contests from SNCKPE21 (Nov 2021). This has been done after all cheaters in these contests have been penalized and moved to the bottom of the respective contest ranklists. This is done to make the ratings fairer for the non-cheaters.
So now you don’t have cheaters at the top of the ranklists, leaching away rating points from the non-cheaters.
The large increase in ratings are for users who had been wrongly penalized in plagiarism (false positives), and hence now been reversed.
After removing the cheaters, and the false positives:
Amongst the users with an increase, the maximum increase is 225, and 99% of them have an increase of less than 125.
Amongst the users with a decrease, the maximum decrease is 96, and 99% of them have a reduction of less than 60. The rating reduction is mostly in cases where the users have had a string of bad performances, and the lower penalized ratings of the cheaters in their subsequent contests adversely affects these non-cheaters.
-- https://discuss.codechef.com/t/rating-recalculation-after-penalizing-cheaters/102228
So essentially you're saying these changes are permanent? Here I reached 6* from a Div 2 contest, and Div 2's aren't supposed to be rated for 5* (Div 1) accounts.
Yes, they are permanent. And yes, this is an unfortunate side-effect. You are among the < 1% users whose rating increase is pretty high.
But such contradictions are inevitable in any rating recalculation which happens after another contest has started. But it is not as bad as you might think — your ratings are calculated only against the users whom you actually participated against (so in those two contests — it’s against the Div2 users), taking into account your abnormally high starting rating for those two Div-2 contest (hence you would have to perform even better in the Div2 contest to get a rating increase). The main unfair advantage that you got was that you got to participate in two more Div-2 rated contests.
The number of such users will drop much below the current 1% when we start doing this activity more frequently starting from now.
According to me this huge rating changes should not occur in just one go. Like in codeforces also when the cheaters are removed. hardly rating increases by 1 or 2. but increase by 100 on average that isn't a joke. some people literally went from 4 star to 6 star. So normally such big jump is officially not possible so the better way of increasing the rating would be contest by contest. and if in some contest, a participant's rating is more than rated range, he should be treated unrated for that. or else it should be left as it is. Well that's just my opinions.
Our plagiarism checks and the appeals phase take a significant amount of time, and it is not feasible to do it before the next contest starts. The appeals phase alone is a window of 7 days, and we have to manually go over hundreds of appeal emails for every contest. So our process is never going to be fast enough to be finished before the next contest starts, which makes some kind of bulk rating recalculation an inevitability.
But as mentioned, we will be doing this recalculation a lot more frequently going forward (at least once a month), so that the changes are smaller.
I haven't cheated in any of the contests and my rating has decreased, but some have turned 6* from 4* in a div 2 contests. My humble suggestion is to roll back the changes and start and continue plag check from upcoming contests. CodeChef_admin
No. Your rating changed from 2038 to 2043.
My bad! I'm sorry!
May I please know what is appealing and how does the process go?
But considering that the ratings of more than a hundred thousand users have increased significantly (especially in Div. 1), but only those of 6000 users have decreased, this is going to lead to inflation, right?
Inordinate rating fluctuations were always a problem in Div. 1, and you might have just worsened it.
Imho, you'd have been better off keeping the ratings as they were this time, and should instead run plag. checks and weed out violating accounts a lot more regularly.
Exactly!
Great decision! but why isn't the cheater's graph showing the plagiarised rating change as they use to before. Cheaters are simply given the last rank now. The previous way to show cheating was great to expose these stupid cheaters.
They still have the black triangle denoting cheating in the rating graph. Not sure what you mean. Link?
May I please know what you mean my appealing phase and what is the process?
What is your frequency of doing this rating re-calculation? I mean, why not do it on the exact same day, or if possible, after 1 month or so? I mean, you can also do it every month, and can keep people from getting demotivated by cheaters?
https://codeforces.com/blog/entry/104932?#comment-932911
Thank you so much.
We are continuously trying to improve our plagiarism processes. Only since Jan 2022 have we been able to penalize cheaters properly. There were other logistic and system issues which prevented us from recalculating for a while. Most of them are fixed now, and as mentioned above, we hope to be doing this recalculation process more frequently going forward.
Thank you so much!
Misa Misa turns 6 today <3
I was in Division 1 at some point. I had no idea xD