jaiveer_017 is clearly cheating.
I was looking at the standings for today’s contest and found something absolutely ridiculous. A user currently rated 1491 (Specialist/Newbie level) somehow pulled off a Grandmaster-level performance, solving A, B, C1, C2, D, and E1 in less than two hours with barely any penalties.
Solving D and E1 that fast with almost zero wrong answers is completely unnatural for a 1400-level player. But the real proof is in the source code. If you look closely, it's incredibly obvious that these solutions are either copy-pasted from a Telegram group or generated by AI mid-contest.
Suspicious Evidence #1: The Shifting Template
If you look at his submissions for A, B, C1, and C2, he uses a standard copy-paste template. But suddenly, in submission D, he randomly adds #include <cstdint> and using ull = unsigned long long; to his standard boilerplate. Why would you change your base template in the middle of a contest just to solve D?
Suspicious Evidence #2: Completely Unnecessary Type Casting
His template literally has #define int ll to handle large numbers globally. Despite this, in his C2 submission, the code suddenly switches to using explicit uint64_t arrays and bizarre casts like uint64_t ua = a; alongside stoull(s). This is a classic sign of someone pasting an AI-generated solution (which often defaults to strict uint64_t types for large ranges) into their existing #define int ll template without knowing how it actually works.
Suspicious Evidence #3: The Compilation Error
On C2, he got a Compilation Error (submission 374783367) at 17:15, and then magically got an Accepted verdict just 4 minutes later (374787736). This usually happens when a cheater messes up their copy-paste job from a Telegram channel or ChatGPT window, hits submit without testing locally, and then quickly fixes the missing bracket or mismatched variable.
It's really sad seeing people artificially boost their ratings like this. You don't jump from 1491 to solving E1 flawlessly without external help.
It would be great if user himself accepts that he indeed cheated and apologise in front of community by commeting down this blog. I am one of his well wisher and want him to grow but not by wrong means. He seems to be student at iitd and is setting wrong example for his juniors, its really sad to see that.
Dear Codeforces team and contest writers (Sanae), please look into this and take action.
(Feel free to check the codes yourself: 374741006, 374752298, 374766811, 374787736, 374811210, 374840800)
Sanae
Auto comment: topic has been updated by MrTripathi (previous revision, new revision, compare).
Sanae also please take action against these four toppers the_chill e4rendel nguyenphong233 oleinikowc. just take a look at their graph or code. They cheated in the last contest for sure.
You should have solid proof, you shouldn't say that the cheated just by looking at their graph.
Note: I'm not saying that you are wrong when you report cheaters, I'm just thinking that you need more evidences.
Well just look at their clean codes and AC on one go. Even many high rated coders have struggled to solve these problems in one go. Not every time you need evidence to prove that someone have cheated. But yes i get your point. I can't say with 100 percent accuracy that the first two (the_chill e4rendel) have cheated (but with 99.99 percent yes for obvious reasons).
Secondly the third and fourth I mentioned (oleinikowc nguyenphong233) have cheated for sure. I mean their curve is not realistic at all. Just go to their past contests where you can see that they struggled to solve even problem C in div 2 and suddenly somehow they got capable enough to solve the entire set? i mean comeon. How much realistic is that? These two cheated in the contest with 100 percent certainty.
I completely understand your point. I am just saying that any one here can tell that they have cheated. Sometimes you don't have to prove everything. You just know it. Their sudden performance rise in one night is just not realistic and i have mentioned only them because they are at the top of the standings with zero brain.
something to consider when looking at their C2 CE submission is how they forgot to include the
cstdintheader foruint64_t(which is basicallyunsigned long long).so it might just be a mistake after all, it doesn't take too much time to insert
#include <cstdint>in that code/template.furthermore, they might have added
using ull = unsigned long longafter that error, and also I personally do not think AI would be "smart" enough to writestoull(string to ull function) instead of either making an entire function for it or using some other built-in function that is very straightforward. something like that.still, it doesn't hurt to prepare for the worst (especially since they solved E1), i'm just sharing my opinion here.
While the person could be cheating, note that the contest in question was their first rated contest in half a year. Also, they've switched languguage from Python to C++. People do improve over time, and weird C++ code could just be because the person has yet to get used to using C++ in a competitive programming setting.
There are however some things that are strange, such as submitting a compile error, and then 4 min later submitting an AC solution with changes that fixes more things than typos. This is definitely a bit sus, but it's not at all a smoking gun. Maybe they simply reallized there was a bug while fixing the compile error?
I just feel like one should be careful calling out others for cheating. Especially when it is not a clear cut case.
One last remark.
That is clearly not what happened. Those lines were added when trying to fix the compile error on C2.