Idea for anticheat system

I thought of a possibly good idea for an anticheat method. If people are reported enough, you could have a program (a bot) that has a human name, human profile, human stats etc challenge the person and/or vs them in their next matchup and it plays REALLY well, as in super high ranked (way above the ELO it would have for on the profile) and it if the person ends up beating it then the case could be sent to an actual admin to look over the game. The bot would play as if it was stockfish level 7 or something, so not nearly perfect stockfish moves but close... That way if the person is actually using stockfish full-force it would be likely they were cheating if they beat stockfish with a minor handicap??? (at least for lower elo players).

Let me know what you think, maybe its a bad idea... Feel free to explain.

Luckily there's no need for measures such as that, lichess has a very good anticheat system in place, and a great team of volunteer moderators. But it's not a good idea regardless. It would basically mean that lichess would be cheating against its users. Pretty wasteful and pointless.

Although I don't think this particular idea is worth discussing, I appreciate the initiative. It's good to have a culture of innovation in a community.

I heard that once people get marked for cheating, they're paired with other cheaters, which is pretty funny. That's what this sorta reminds me of.

Also, someone cheating doesn't mean they play perfect moves all the time. The harder to spot cheaters combine their own moves with engine moves, or play second to third best moves purposely to avoid detection.

Detection tools have become sufficiently advanced that despite being an engine developer, I get rating refunds from opponents I thought were legitimate.

@Toadofsky, what you are saying reminds me of a friend, who I saw personally get banned on lichess due to him cheating. I was next to him when it happened and had been watching him for the last let's say 2 hours, and he was banned for cheating after beating several very high rated players and being low rated himself. I can testify that he had not cheated since I had seen him play on prior days too but regardless he was banned. I understand though, overall the system works very well but some mistakes are impossible to stop from happening.

#5 Unless you've been sitting next to him during EVERY game he played, you cannot testify for it. Him getting banned while you were watching was probably just a coincidence that he was caught at that time, for games that you weren't watching.

@Synky why would they accept the challenge in the first place ? (especially if they are cheaters, especially if they know such an anti-cheat system is in place).

Then as #2 said, that idea would have _lichess_ playing bots impersonating humans (i.e. cheating), that's not great.

And last, what if the cheater smells something fishy (this 1600 player who challenged me out of the blue is playing almost perfectly vs my 2600 engine !) and decide to lose that one game ?

The system in place is pretty good already ;)

@lecw good points, I'm convinced this was a bad idea after all, haha.

@Risk_It_4Biscuit I totally feel you... I can honestly see it happening maybe? I don't know how accurate the AI or w.e is that is in place. For example, I'm UTTER garbage, but I've had some really good games randomly that no doubt had to of shocked my opponents (and possibly even thinking I was cheating?). Take for example this game I played against like a 1890 ELO player... My proudest game yet as of starting chess just 2 months ago: lichess.org/A3FZ0or1

Like when I analyzed that I was so proud of myself for how little mistakes and only 16 centipawn loss, but like if I was my opponent I would maybe think I was cheating?

#8 And you're not banned, so the system is better than you might think. Winning against a stronger opponent won't get you banned, if you didn't cheat to do it. False positives exist, but they're the exception, not the rule.

github.com/clarkerubber/irwin