Nowadays cf usually use multiple testcases instead of multiple samples to decrease competitor's sample-test-time or due to other reason.
But some problems can't use multiple testcases(such as 1743D - Problem with Random Tests) and a friend of mine get penalties because of Wrong answer on test(sample) 2/3.
Can MikeMirzayanov fix it, or it is just cf feature?
It's the cf rule.
But I think sample #1 and sample #2/#3 are equal so what does this cf rule target at...?
I like to believe that it targets accidental submissions. Also you can try agnocpeed to test multiple test cases at once :)
But I have been accustomed to CCH and CP-Editor and I don't want to try agnocpeed though it seems powerful XD
My opinion is that, the rule is to pretend wrong submissions across problems, for example I want to submit a problem C solution but I missubmitted it to problem B.
But if you can pass the first sample, you are not submitting the wrong problem, are you?
You do have a point.