Do Lichess use bot’s to play against a player suspected for cheating, to check if he’s cheating?

Hi. Do Lichess use bot’s to play against a player suspected for cheating, to check if he’s cheating?

I’m asking this because sometimes my opponents feels a bit «inhuman» in their play behavier. So I’m just curious.

And if so, how many games do Lichess bot’s play against the suspected player before the check/test is done?

They have some kinda A.I detector afaik

"irwin
irwin is the AI that learns cheating patterns, marks cheaters, and assists moderators in assessing potential cheaters."

https://github.com/clarkerubber/irwin

I cannot imagine how this would help them in finding cheaters.

No, they don't.

Thats a very intersting question...
To view if a person is a cheater, it will be recomended to analyse first his game play (with a bot like stockfish), his profit, and his elo points.
The game analyse it could be seen as the time has been move the pieces. The more dificult the moves are, the more effort the players need to think... So, if the player moves very fast on a very hard position and the player has like 800 elos, it could have probaility of cheating. But, quite the person are relative, it imposible to determine by few evidence of a certain person whos is cheating, the counter is, it could be a GM or a Pro player playing as a low elo account, or, a player who on propuse got lowed down then try to step up on a high level, and finally, thinking if the player is a Cheater by proving it.
And the best way, to attack a cheater, is using a bot. If the player lost, probably it wasnt using any engines, but if won, considering the points, times and others, then its a cheater...

@RoadFires said in #6:

And the best way, to attack a cheater, is using a bot. If the player lost, probably it wasnt using any engines, but if won, considering the points, times and others, then its a cheater...

This is a dysfunctional idea.

It is easy to find cheaters who are using an engine in every game for every move. Just compare their moves with engine's suggestions. No need to give them extra chances against bots.

The problem are the intelligent cheaters, who are using an engine only for complicated middle game positions, or only as a blunder check. Or those, who are cheating only every other game, or only against titled players, or whatever.

These players would probably lose against a bot, and therefore would never be caught by this idea.

@sheckley666 said in #7:

It is easy to find cheaters who are using an engine in every game for every move. Just compare their moves with engine's suggestions.

This is a very crude and unreliable way to catch cheaters. Methods used by modern chess servers are much, much more sophisticated than that. "Intelligent cheating" isn't a real problem either, at least not the method you describe. That type of cheater will be caught quite quickly.

Yes, but the point is: For players using an engine for every move of every game, it would be sufficient.

@sheckley666 said in #9:

Yes, but the point is: For players using an engine for every move of every game, it would be sufficient.

Not really, or only in some specific cases. I often play games with very high accuracy and 0 inaccuracies, 0 mistakes and 0 blunders.

By your method, those games would be marked as 'cheated'.

It may be one ingredient in a large set of tools and indicators, but by itself it's not reliable.