In the recent Codeforces Round 1098 (Div. 2), I started solving the problems in a different order. I started with C1, then went on to B and A respectively, and then came back to C2, which was really just a small change compared to C1.
When submitting C2, I accidentally selected C1, which, of course, being the same problem just easier, got accepted.
After the contest, I saw my first accepted submission of C1 (at around 1:00 into the contest) got skipped, while the second (and last) accepted submission of C1 (at around 1:35) got accepted. After a web search, I found out that system testing takes into consideration only the last accepted submission of a problem. This led to me losing about 500 points and dropping (probably, I haven’t counted) a few hundred places.
Why? Just why? Why does it matter if I had 2 submissions of the same problem? It doesn’t make any sense from my perspective, and I want to hear your opinions on this. Thank you.
Sometimes it may be useful to resubmit a problem in the case of the first submission being very close to the time limit/hackable. Because of this, Codeforces only counts the last submission that passes pretests.
System testing already takes very long. It would take much longer if several submissions to one problem from one participant could be judged. Because of this, it seems most logical to simply judge only the latest submission.
This contest didn't have any open hacking phase. In these cases, why doesn't system testing judge the first accepted submission?
It doesn’t matter that it didn’t have an open hacking phase. Your solution could be hackable with a test from another successful hack for this problem, also during the round itself it could be hacked by someone else in your room. That’s why only the last submission is accepted.
I didn't know that, thanks!
Just be careful in the future :)