I solved ABFHJK in CS3233 final contest. Here are my thoughts:

  • A: shit constant time TLE, also oolimry got trolled, I had to solve for him L
  • B: literally copy $O(n^\frac 1 4)$ factorization from kactl and first blood
  • F: not that interesting, large $n$ force me to use KMP, which I never used in like 3 years
  • H: I got trolled by the function and started expanding it immediately cause it seemed like convex hull. I was lucky to draw points on paper, then I realized that this cost function was literally euclidean distance. Funny troll problem.
  • J: It is quite obvious that $ans \leq 1$. Checking that $ans=0$ literally just involves simulating the path of Alice. Quite boring.
  • K: Standard suffix array/tree problem

The other problems:

  • C: I told pavement it was $V-E+F = 2$​​ thing and he bashed the math L
  • D: we tried finding example graphs for small $n$, idk. For odd $n$, obviously only good graph is complete graph
  • E: idk bro, oolimry: “my only contribution was to theory solve E”
  • G: i only knew it was DAG+flow
  • I: conjecture: bob only wins if the xorsum of terminal nodes is 0 and the rest of the nodes have 0 value. Rama says it is false L. Counterexample: the graph of 4 vertices $1 \to 0 \quad 1 \to 0$.
  • L: observation: each row of lasers left/right shifts a subsequence. idk how to proceed after that
  • M: didnt read

Generally, I find the contest atmosphere of this year to be much less enjoyable than previous years. I think the main reason for this was that there wernt many high school teams. It is quite well accepted that generally, the high school teams participating in CS3233 are much better than the actual students participating in the CS3233 final contest. So I didn’t feel that there was much competition other than Cossacks and Samurai.

Also, while I can’t comment on the difficulty for DIL, the rest of the problems seem to be easy while DIL are completely unapproachable. So the difficulty gradient probably wasn’t the best… The contest experience is easily mindsolve solving everything minus DIL, fighting over 1 computer to code, and then just staring at blank space for DIL…

But I don’t blame contest authors for mostly standard problems since I am not the intended audience of this contest. Perhaps, the only reason I find CS3233 contest being worse than in 2021/2022 was because I became too good for the doable parts of the problemset that my team was able to quickly mindsolve them.

No idea if I want to participate in CS3233 final contest next year.


<
Previous Post
Sandwich Editorial
>
Next Post
Magic Integral by Parts