tem_shett's blog

Setting problem difficulties
By tem_shett, history, 4 years ago, In English

About 1.5 years ago small innovation on Codeforces appeared, it was connected with problem difficulties. This innovation is really useful for all Codeforces' members, of course. But ther is one problem, that can prevent you from improving your skills. It's the method with which the difficulties are set, that problem is connected with problems from div3 rounds in most cases. Let's parse one example: in the div3 round the last problem has difficulty 2600. If we follow words of MikeMirzayanov from his article, we will know that participant with rating 2600 would solve this problem on typical round with a probability 0.5. But in case of this div3 participants(not official) with rating 2600 have solved all problems for about an hour. Maybe it caused by other problems were to easy, but this means that difficulty of problem directly connected with difficulties of other problems from a round. In any case, Codeforces need to do something with that. Also, as I heard, problems get their difficulties not using some algorithm, but by headquarters. If that is true, why don't to come up with some algorithm, which can set difficulty using information about rating of participants, who have solved this problem, and who haven't solved this problem on the contest. It seems like there are enough data to do that. For more examples of bad difficulties you can check last problems of div3 rounds. What do you think about that?

  • Vote: I like it
  • +136
  • Vote: I do not like it

»
4 years ago, # |
  Vote: I like it -43 Vote: I do not like it

As for me, I don't care, because I can't solve last problems from div3, unfortunately.

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

Auto comment: topic has been updated by tem_shett (previous revision, new revision, compare).

→ Reply
»
4 years ago, # |
  Vote: I like it +18 Vote: I do not like it

Really good idea but there is something the algorithm will not catch it

Recently a lot of blogs have created saying that many people cheating in the rounds, so if someone have a rate < 1200 solved Div2-E(By cheating it) in the round the rate of the problem will be not realistic

I think the best option until now is updating by headquarters until catching all cheaters

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it +40 Vote: I do not like it

    I think there are not so many cheaters on CF, so it is not very tragic.

    → Reply
  • »
    »
    4 years ago, # ^ |
    ← Rev. 2   Vote: I like it 0 Vote: I do not like it

    But headquarters aren't able to catch cheaters, so they will set difficulty using number of right solutions, but in this case difficulty of the problem won't be more fair

    → Reply
    • »
      »
      »
      4 years ago, # ^ |
      ← Rev. 2   Vote: I like it 0 Vote: I do not like it

      yeah that's the point.

      while headquarters can't catch most of cheaters I think its better they estimate the difficulty of the problem on their own instead of writing algorithm based on rating of people who solved it.

      sevlll777 may be right, but its my opinion

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

        But it there won't be any changes. And it would be better for headquerters.

        → Reply
»
4 years ago, # |
← Rev. 2   Vote: I like it +11 Vote: I do not like it

1141F2 - Блоки равной суммы (усложненная редакция)

Another good example of strange difficulty

→ Reply
  • »
    »
    4 years ago, # ^ |
      Vote: I like it -49 Vote: I do not like it

    Oh, that's bad problem! I couldn't solve it on contest!!!!!

    → Reply
    • »
      »
      »
      4 years ago, # ^ |
        Vote: I like it -19 Vote: I do not like it

      Oh I'm so sorry!!! I haven't participated in this contest(I have lied). I thought it was different problem

      → Reply