Hello Codeforces!
I guess all of you know that the last Div. 2 had "a bit too many" cheaters — and I created a new entertainment material out of them!
Introducing, Div2FCheaterGuessr!
Yes, the name was indeed inspired by GeoGuessr of course, and the gist of it is pretty similar!
The rule is as follows.
- You are given a screenshot of random submission from a rated participant on 2103F - Maximize Nor, from the status page.
- Then, you must guess which data structure it contains — there are three options: a segtree,
std::map, or astd::priority_queue(if it is neither of the three, just pick another submission). - You gain a score if you guessed it correct.
Simple rule, right? And you will enjoy it the more you play it!
Let's try it out!
What do you think? A segtree, std::map, or a std::priority_queue?
The answer is...
How about this one?
The answer is...
Another one!
The answer is...
It's pretty fun, I hope many people can relieve their anxiety gained from GPT cheaters by enjoying this game!
Please try it yourselves (Maybe you can use comment threads to play the game on?) and tell me what you think!
And here's what I think:
What I think about this (major spoilers if you want to enjoy the game!)
Considering how much GPT usage was used in F, I was guessing that all of them would be seg-tree considering how much GPT likes to abuse it from the problems I've tested it on
Same. Most of the cheater (and legitimate!) solutions I saw on F used segment tree.
I like this game a lot except this:
aww, I couldn't make it into an actual website because that'd make it not an account that's "used to crawl", but a real crawler actively crawling instead ;-;
As someone who is not at the level to solve Div2 harder questions, nor did I participate in this contest, I want to kindly ask, if the tell that someone is a cheater is that they up-solve problems that are deemed too hard for their current rating? Or is the code they write too GPT-like?
Yeah, solving too hard problems for their rating and comparing their last performances is enough to detect cheaters. Only dumbs dont change gpt, so it's hard to tell is it look like gpt or not, but we can also compare their last submissions and current. No one change their codestyle significantly in short of time, so this is also sign of cheating.