JAPL's blog

Rating discrepancy in lower divisions
By JAPL, history, 5 years ago, In English

No hate intended

Lately I have been noticing something peculiar , or at least I find quite unfair !

Lower divisions such as Div.3 facing problems with difficulties reaching ranges varying between 2000-2500 .

From what I understand problems are supposed to be solved by users with ratings similar to the difficulty of the problem

example : newbie should be able to solve problems in the range of 600-1200 , pupil from 1200-1400 etc... correct me if i'm wrong

Following the logic I described above what's the need to put problems with such degree of dificulty knowing that only users rated experts or below can participate ?

Feel free to give me your opinion.

  • Vote: I like it
  • -11
  • Vote: I do not like it

»
5 years ago, # |
  Vote: I like it 0 Vote: I do not like it

Ratings for div 3 aren't accurate. https://codeforces.net/blog/entry/65383?#comment-494119

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

    i am also referring Div.2 , who is going to solve such a problem there

    → Reply
    • »
      »
      »
      5 years ago, # ^ |
        Vote: I like it +16 Vote: I do not like it

      Harder problems from div2-only rounds don't have an accurate rating too. This measure works mainly for div1 problems.

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

      A similar idea applies, just to a lesser degree. The upper bound on ratings in Div 2 is either 1900 or 2100 depending on the contest, and those ratings are doable. Of course, not many people will solve high rated problems, but that is the point of having those problems- to distiguish people based on their skill. If a large number of people were finishing every problem, then this would be hard.

      I believe for the problem ratings the approximation is something like if a problem is 400 points above your rating you have about a 10% chance of solving it, so problem ratings o 2000-2500 with the number of contestants is reasonable.

      → Reply