i'm still not sure if it's actually constructive algorithm

Heya there, I just want to ask a couple of things to anyone reading this blog regarding contests strategy (mainly revolving around normal CF round).

So, most Codeforces round are hold in a very interesting way of scoring, where each problem is initially given a fixed score from the score distribution, and as time goes by, each problem's score will drop until it reaches the minimum (e.g. if the score for a problem is $$$500$$$, then at some point it will not decrease below $$$250$$$ or something, correct me if i'm wrong pls .w.). Now, here are the questions for you guys, as i'm little bit a fan for data and statistics :p.

1. What is your strategy for these type of Codeforces contests? Is it the classic strategy (from first to last), modified rainboy strategy (choose the hardest problem you think you can solve (e.g. problem E maybe), then solve from that problem to first, before returning back to classic strategy), or other strategy you crafted yourself?
2. Does your strategy involves skipping problems? If so, how does it work? If not, have you ever skipped a problem in CF contest before?

After reviewing leaderboards from IOI 2020 to 2023, I realised that in IOI 2021, the problem "Mutating DNA" (codename "dna") has a lot of AC's in contest that around 85% of the contestants manage to get score 100. And now, I have two questions regarding this (and IOI problems in general lol).

• If this problem were in a CF round, what would be the difficulty (or the problem rating) and the placement on a problemset?

• What do you think is the proper difficulty (or CF problem rating) for the "easiest" problem in IOI (in general)?

Do you personally prefer tree problem with $$$N-1$$$ pair of integers input where $$$i^{th}$$$ input ($$$u_i$$$ and $$$v_i$$$) denotes the nodes that $$$i^{th}$$$ edge connects and all the edges guarantee a tree, or $$$N-1$$$ integers input where $$$i^{th}$$$ input ($$$p_i$$$) indicates that there is an edge that connects node $$$i+1$$$ to node $$$p_i$$$ and all the edges also guarantee a tree (usually this comes with the $$$1 \leq p_i \leq i$$$ constraint to do that)? I personally prefer the latter.