Please read the new rule regarding the restriction on the use of AI tools. ×

macaquedev's blog

By macaquedev, history, 117 minutes ago, In English

Is it just me being unable to do the good codeforces rounds, or have they recently just gone down in quality...? Like honestly, what even was today?

B was a standard well-known problem, and its only difficulty was sqrt precision errors. Is that really what we want codeforces to be? Or are we trying to build a community that (dare I say) focuses on algorithmic thinking and problem solving??

C was a "guess the most ridiculous solution" type constructive problem, which can be fun, but again, this "proof by AC" that has become so common lately in codeforces just seems antithetical to the purpose of the website.

Most ridiculous jambloated nonsense

D was a nice problem, I must admit.... but then there's the last two problems. E was simply a constant optimisation problem, which again is absolutely hideous. Is there really nothing in this world that is more interesting than having an O(n log^2 1024) solution and having to optimise the constant factor, and watching other people pass O(1024n) solutions using some sort of pragma witchcraft? And, it turns out that F is explainable in two lines using an obscure technique.

Rounds like this just make me a bit sad...

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

»
113 minutes ago, # |
  Vote: I like it 0 Vote: I do not like it

Please do not name blog such. It can be confusing to non English speakers. I really thought this was the actual blog for the contest for a while of time.

»
112 minutes ago, # |
  Vote: I like it 0 Vote: I do not like it

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

»
105 minutes ago, # |
Rev. 2   Vote: I like it +6 Vote: I do not like it

$$$O(1024n)$$$ passes without any single pragma under $$$1$$$ second. The hardest part of this solution is to believe that it will pass the time limit :)

»
100 minutes ago, # |
  Vote: I like it +1 Vote: I do not like it

How is C a proof by AC problem? A bit-by-bit construction is obvious, since the result of one bit cannot possibly overflow to another; from there, the a = c^d observation just requires a bit of extra thinking, and is certainly not necessary. Contrast this with a previous C, https://codeforces.me/contest/2003/problem/C, where proving the solution validity is much harder.

I agree E was a bad problem in that it allowed O(1024 * n) solutions to pass, but the intended solution does not require any kind of constant optimization, and runs in O(n log^2) rather than O(n log^2 1024) as you stated.

And how can you say F is a bad problem when you don't even understand the "obscure technique" it uses?

  • »
    »
    90 minutes ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    my E solution ran in O(100n) (since 100 = log^2(1024)) and it was TLE until I optimised it so that the vector isn't recopied 100 times (which surely shouldn't make such a big difference). Also, strictly speaking, O(100n) didn't pass, but O(65n) did, where I multiplied the pairs by 2 where i was not equal to j. See my previous submissions