Codeforces's virtual contests are probably one of the best feature in CP, since it allows people who can't do the contests because of various reasons to have a chance to simulate the contest. However, the virtual contest format doesn't consider the fact that the real contestants had pretests and system tests, while in a virtual contest, the source gets evaluated to all the tests in the moment of submission.
So, I suggest that the virtual contestant should be able to choose whether he wants to get his solution tested only on pretests(just like in the real contest) or to have full feedback(the current format of VC as of now). I don't think this isn't that hard to implement, CF can probably still evaluate all the tests and show "Pretests passed" if the source passed a certain test number(the number of pretests which this contest had).
I think this is a problem which concerns most of the people who are using the virtual contest feature, so fixing this problem would improve this feature significantly.
MikeMirzayanov KAN What do you think about this? Can it realistically be implemented in CF?
Also, if there was some bot trying to hack your solution from the hack tests used in the corresponding contest, that'd be fun!
Thanks for this post,I was about to write for the same request and one more, which probably is too much to ask for but if possible, can we have the feature of hacks in virtual contests too, so that we can lock a problem and can't submit it again and try to hack someone else's solution (by getting a random room, at the start of the contest)
...And another difference between virtual contests and actual contests is that the virtual ones are only supported in the ACM-ICPC mode whereas the actual ones use different scores for different problems. So, say if someone starts by solving D first and others have solved A and B, then the first person might have a better rank in actual contests but not in a virtual contest. So, VCs do not directly reflect the scoring used in actual contest.
wow, listening twice boost ideas to improve cf. Mike should also listen twice!
I think the stuff suggested about auto-hackers and the like is a bit pie-in-the-sky, but especially for Educational rounds the basic functionality to virtually compete with system test evaluation OR pretest evaluation while in the virtual contest would be so valuable.
Has anything been done about this idea?