MathCrusader's blog

By MathCrusader, history, 7 years ago, In English

The ICPC South Pacific Regional Finals were held this past weekend in Sydney, Australia. The top 12 teams from the region (Australia, New Zealand, Fiji) competed for an invitation to World Finals.

Two World Finals spots were awarded:

The full scoreboard can be found here. I will put up the contest into the Gym in the next few days once I relax a little bit from all of the organization last week. =)

Congratulations to the teams and good luck at World Finals!

UPD: I have added the contest to the Gym. I hope everyone enjoys the contest!

UPD 2: The editorial for the contest can be found here.

Full text and comments »

  • Vote: I like it
  • +64
  • Vote: I do not like it

By MathCrusader, history, 8 years ago, In English

I am a coach of my programming team, and I'm just starting to use Polygon and have a couple questions.

  • In ICPC-style events, when your solution is wrong, you will receive the result of "Wrong Answer", "Time Limit Exceeded", etc. However, on CodeForces (even on ICPC-style contests), it tells you "Wrong answer on test 5". I want to hide the test case from the contestants to more closely represent an ICPC contest. Is there any way to prevent this information from appearing? (for example, in the Gym or in contests that I make for our Group?).

If you wonder why this is important, consider these two cases: you submit a problem and get "Wrong answer on test 5", then you find a bug and submit again and now get "Wrong answer on test 6". This means that the team knows that they fixed something in their submission, but they would not know in an ICPC contest. This can dramatically alter the way that teams debug their code. They may also get "Wrong answer on test 84", which may tell them that their solution is mostly right, but they are missing a corner case (which is, in my opinion, giving them way too much information).

  • In Polygon, is there a way to generate your output files from a program that is not marked as "Main correct solution"? This normally isn't desirable, but for me (being as picky as I am), I would like to avoid the situations where my judging program outputs 0.4000000012 instead of 0.4000000000. You may wonder, "Does this really matter?" Probably not, but it does cause an issue where the following results will give Wrong Answer instead of Accepted: 0.399999000 with an epsilon of 10 - 6. You also get an issue where 0.4000010010 gets Accepted but should get Wrong Answer. So if we used a program which computed the solution with lots of extra digits of precision (e.g., uses BigDecimal), I would be more confident that the answer will be correct (but this program could easily give Time Limit Exceeded, so it shouldn't be the "Main correct solution").

Thank you in advance for the help!

Full text and comments »

  • Vote: I like it
  • +19
  • Vote: I do not like it