Rating changes for last rounds are temporarily rolled back. They will be returned soon. ×

Candidate_Master_2021's blog

Why do contest setter do this to us?
By Candidate_Master_2021, 3 years ago, In English

Why do you make pretest as weak as Round 700's Div 2 C... I mean seriously? All pretest were under n <= 100. This is not even a mistake.this almost feels like u guys were trying to do some kind of a prank on us. why,why....please give some kind of explanation? To the people coordinating and testing the round. did u try to submit a brute force on it?

this is crazy...what a waste

  • Vote: I like it
  • -53
  • Vote: I do not like it

»
3 years ago, # |
  Vote: I like it +8 Vote: I do not like it

Maybe I'm crazy and I am not familiar with the official position, but I think problemsetters should be allowed to create weak pretests, why not? You may not like weak pretests but I think creating weak pretests should not be considered as something that is fundamentally wrong.

Back in the day this was normal. Pretests guaranteed that your submission had some semblance to a real solution, nothing more. The idea that pretests must be just like systests is recent.

→ Reply
  • »
    »
    3 years ago, # ^ |
      Vote: I like it -15 Vote: I do not like it

    If there must be some semblance, it should be strong shouldn't it? Otherwise the sample test should be enough to provide some semblance and pretests would be redundant.

    → Reply
    • »
      »
      »
      3 years ago, # ^ |
        Vote: I like it +15 Vote: I do not like it

      What's your point exactly? I think it is possible for pretests to be stronger than samples but weaker than full tests.

      → Reply
      • »
        »
        »
        »
        3 years ago, # ^ |
          Vote: I like it 0 Vote: I do not like it

        I agree, but in my opinion the pretests should at least cover the basic cases(e.g the full breadth of constraints) otherwise I don't see the difference between sample test cases and pretests.

        → Reply
        • »
          »
          »
          »
          »
          3 years ago, # ^ |
            Vote: I like it 0 Vote: I do not like it

          To be honest, I like that too. In my ideal CP site, pretests would be as strong as systests. Furthermore, I would write statements in a way that minimizes dumb edge cases (for example, disallowing $$$n = 1$$$ completely in Div1A).

          The intention of my original comment was less about what I prefer and more about the current reality — I don't think there is any reason for us to be mad if pretests are not strong (although I did not see the claim in the announcement about pretests; yeah, that's just evil).

          → Reply
  • »
    »
    3 years ago, # ^ |
      Vote: I like it +2 Vote: I do not like it

    If as the blog author said that n<=100 was in pre test is true then pretest are terrible. Because pretest with n = 100 can't guarantee any resemblance with correct solution since actual n could be upto 1e5. So yeah that's one thing.

    Another thing just because pretest used to be not so strong, there's no reason to not make them strong now.

    And yeah it should be up to problem setters how they want to have pretest. Some might want participants to have proved solutions while others might believe strong pretest is a better idea.

    → Reply
    • »
      »
      »
      3 years ago, # ^ |
        Vote: I like it +8 Vote: I do not like it

      Because pretest with n = 100 can't guarantee any resemblance with correct solution since actual n could be upto 1e5.

      Sure it can. It can't check everything, but it can filter out a lot of nonsense.

      Another thing just because pretest used to be not so strong, there's no reason to not make them strong now.

      Well, sure. I just wanted to give some context to show that it is not a given that pretests are strong.

      → Reply
  • »
    »
    3 years ago, # ^ |
      Vote: I like it +13 Vote: I do not like it

    I believe you're right in saying that weak pretests are not fundamentally wrong. It might be great fun to even have only a handful of cases / samples and let people compete on their accuracy as well.

    However, that should be transparent from the start. Again, it's not 'fundamentally wrong' to do this, but add to that lines like these from the announcement post

    I have tried my best to [...] make strong pretests.

    and all weak pretests do then is create a false assurance and a sense of security.

    → Reply
    • »
      »
      »
      3 years ago, # ^ |
        Vote: I like it -10 Vote: I do not like it

      Well "tried my best" doesn't mean succeeded. People should always view pretests as it means their sol is likely to succeed but should make sure their code is actually right.

      → Reply
      • »
        »
        »
        »
        3 years ago, # ^ |
          Vote: I like it +1 Vote: I do not like it

        Agreed. And I'm completely against bashing problemsetters for hiccups. But the initial comment seems to, to some extent, invalidate the idea that strong prestests are important.

        → Reply
»
3 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Why do you downvote? Did you gain rating because of others failing system tests?

→ Reply