I saw that the rankings changed dramatically since the competition ended (not before to after deadline, but changes have happened after the deadline), and people who were super close to the top now have really bad scores (>0.9 down to ~0.17 for me). This looks really odd to me, especially since submitting the same code in practice again gives me closer to 0.9.
What’s going on? Has something gone wrong with the grading of the challenge? A lot of scores seems to be pretty much spot on 80% off, which would make a lot of sense if some submissions were never rejudged with all test cases (20% are graded before deadline, the rest after).
Edit: just tried running the current “winner” in practice and my submission was better than it.
We mistakenly rejudged the problems before the submission queue could have cleared. Also, we did quick rejudge twice. Due to this, there is an inconsistency in the scores. We need to rejudge all the submissions again to make the data is consistent. It will take few days to finish. We will revert the current rating and update the new rating afterwards.
How do you mean? Max score is 100. The best user before the deadline @ms4 now suspiciously has almost exactly 20 points. Which very much looks like it wasn’t regraded while other solutions were.
Thanks for the quick response! I’ve been quite confused about this for the last few days. Great to get confirmation that something is actually going on and that you’re working on it. Cheers!
Edit: maybe make that an answer and I could accept it?
haha. its clear they mistake in calculating points. who can get more rns4 :))). if you see last page of submit for bias you can see final testing just do on submits of this page.
@algmyr In deed the work is going on in backgroung because my very un optimised soln was graded 12pts during contest. And after contest it suddenly jumped to 20pts, which was unexpected. And after 48 hrs it has 2 pts.
So, work is going on. And i expect it to be completed soon.
We mistakenly rejudged the problems before the submission queue could have cleared. Also, we did quick rejudge twice. Due to this, there is an inconsistency in the scores. We need to rejudge all the submissions again to make the data is consistent. It will take few days to finish. We will revert the current rating and update the new rating afterward.
XD