10290 Huge time limit, but why?

[Robert Gerbicz]

Ok, I know that all time limits are 3/10 on the new judge as it was on the old judge. But I'm sure that for example on this problem it's too large.
Now it is 33.333 sec., however today I've solved it in 0.02 sec. Looking the old stats 29 people solved it in 10 seconds (on the new judge this would be 3 sec.) the other 373 solved it in the terrible 10-100 seconds interval. I wouldn't call this too bad (there are some problems on acm uva, which has been solved by <10 peoples). Or are we following the spoj's (probably new) strategy? They posted a problem where the time limit is 300 seconds... I think we'll also reach this, or even 3600 seconds, and then everbody for every attempted problem get AC.

A note: how large time was on the old judge the 10 second:
looking my old stat, when I solved 610 problems: among them there are only 15 problems solved by more than 1 second.

Ps: Now on p197 the limit is 200 seconds. And look that stat, there are 9 peoples who solved it in less than 1 second on the old judge.....

[Robert Gerbicz]

They posted a problem where the time limit is 300 seconds...

Hm. And after seeing my 0.02 sec. solution on spoj for the problem they reduced the time limit from 300 seconds to 10 seconds.

[Carlos]

Time limits will be reviewed. Some of them were taken from real contests, or calculated with a non-optimal solution. I agree with you that problem 10290's time limit should be smaller.

Btw, no new strategy in our side, we always think only good submissions should be AC. If 3s time limit is too big, we won't reduce it: it means there aren't enought test cases.

[tobby]

I believe such a huge limit is to allow a "slow" algorithm that factorizes each input number to pass. IMHO such an algorithm, though not optimal, is by no means easy to implement.

-- EDIT --
I just resubmitted my old slow PASCAL code and it ran in less than 10 seconds.

By the way, Robert said that such "generous" time limit would let every attempt get AC. This is far from truth (843 users attempted this problem and only 421 have got AC).

[Carlos]

You are right, this problem is too easy to need 33s to solve it. Time limit has been set to the standard 3s and we're rejudging all submissions.