I used AI to fix my grammar, so this blog might look like it was written by AI.

Ladies and Gentlemen, it's my honor to introduce to you: Cheater Got Talent (Yeah, those colors were picked from the CF logo). This is a series where I, CheaterGotTalent, along with some other users, find evidence and write blogs about the cheaters we find. The series is planned to be published twice a week.

And now, welcome to Cheater Got Talent Episode 1!

Our guest today is (drumroll...) BobB1707!

At first glance, this user looks kinda clean:

Let's dive into his submissions and look for proof:

1. In Educational Round 190, he spammed WA1 on A. Still, this is legit—maybe because he's new to CF, so he didn't know there are sample tests he wanted to save some time.

2. Moving on to Round 1102, where he solved the most problems out of all his contests. A and B seem to be fine. But on C, he first submitted a CE, and just 2 minutes later, he got an AC. He must have copied the wrong thing from his AI, I thought. It turns out I was wrong. He had simply submitted C++ code as Python.

3. In the next contest, Educational Round 191, these were his submissions:

As you can see, he spammed WA1 quite a lot, which is a sign of LLM usage. However, it may have taken a lot of time to check the sample tests for problems like B. Or maybe he thought there were no sample tests.

Although this account is kinda suspicious, if I were him, I could easily defend myself against this evidence. In fact, whenever I investigate cheaters, I always try to think of a defense for every piece of evidence I present. This helps me determine which evidence is strong and which is useless. In conclusion, I think that this account is totally legit.

WRONG

One thing I forgot is that skipped contests don't appear in the rated contests section. If you change the "Only rated" filter to "Only unrated", you'll find this:

Of course, the April Fools Contest is unrated. But what about Round 1103? Let's have a look:

Bruh.

But maybe those skipped submissions are misleading? No, they're not, because:

(1) he was struggling with 1400-rated problems (2233C), and then suddenly he solved both 1700- and 2200-rated problems; and

(2) his F1 and F2 code is extremely clean, while his other submissions (including those outside this contest) are painful to look at. My theory is that he used AI during Round 1103, then tried to make his code for A–E look like something he had written himself. However, when he got the F1 and F2 solutions, there wasn't much time left, so he panicked and submitted them without trying to make the code look like his own.

For the reasons above, I'm convinced that BobB1707 used AI and therefore violated the Codeforces rules.

Well, that's the end of Cheater Got Talent Episode 1! Thank you for reading, and have a nice day.