Блог пользователя bboy_drain

Автор bboy_drain, 10 лет назад, По-английски

Hello everyone,

I was just notified by a friend that the USACO contests provide judge feedback on your submissions, so it is possible to know your score before they are released!

Good luck on the first contest!

  • Проголосовать: нравится
  • +23
  • Проголосовать: не нравится

»
10 лет назад, # |
Rev. 4   Проголосовать: нравится +22 Проголосовать: не нравится

UPD: I'm sorry, I didn't read the contest information!

If your program correctly solves the sample input case (included in the text of each problem), you will receive feedback on how it performs on each of our judging inputs. For example, you will be told if your program produced the right answer, timed out, experience a run-time error, etc. Note that although you receive this feedback, the remaining judge inputs beyond the sample case are not disclosed during the contest. Problems are intended to be challenging; it is rarely the case that a large number of competitors receive near-perfect scores!
  • »
    »
    10 лет назад, # ^ |
      Проголосовать: нравится +1 Проголосовать: не нравится

    I think it doesn't mean that. It just means that we can't see the input and output values for test cases except the sample one.

    • »
      »
      »
      10 лет назад, # ^ |
      Rev. 2   Проголосовать: нравится +5 Проголосовать: не нравится

      OK, you're right. I was confused because the contest rules haven't change yet. I'm sorry for causing confusion.

»
10 лет назад, # |
Rev. 2   Проголосовать: нравится 0 Проголосовать: не нравится

Full feedback means things like tle, mle, runtime error or ac on ALL of the test cases. Am I right?

»
10 лет назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

Does anyone know, that if you get a 1000 on the bronze contest, whether or not you get to take the silver contest on the same day?

»
10 лет назад, # |
  Проголосовать: нравится 0 Проголосовать: не нравится

It states that the Judge may add tests after the contest. So it might be possible that you are getting at least the feedback score in the final result but not full marks.

  • »
    »
    10 лет назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится

    Thanks for pointing this out. However, I think this would be a rare situation. At least at the IOI, this should never happen.

    • »
      »
      »
      10 лет назад, # ^ |
        Проголосовать: нравится 0 Проголосовать: не нравится

      It's not as easy as it looks to make good test sets for problems. It may be case that one corner case is not present, or that incorrect greedy solutions are accepted, or that one particular bruteforce strategy fails to get time limit.

      But I agree that it's going to be rare, I think they'll use the opportunity to change test cases only if it's really disturbing the results (lots of wrong codes get accepted).

      Upon reading the instruction page, this rule looks a lot more bizarre: "it is NOT guaranteed that all possible legal datasets will be solvable within the time limit". What?

      • »
        »
        »
        »
        10 лет назад, # ^ |
          Проголосовать: нравится 0 Проголосовать: не нравится

        I agree that it's often difficult to design good tests. But adding new tests after the contest is a very bad idea when full feedback is used.

  • »
    »
    10 лет назад, # ^ |
      Проголосовать: нравится 0 Проголосовать: не нравится

    No, for this contest, these are the final tests.