How to Detect Cheaters

When chessdotcom released there 'dirty dossier' on Hans Neimann they accidently revealed - via some simple arithmetic - that 'titled' players were at least ten times more likely to cheat than Ordinary Joes. Bit of a shocker but there you go.

chessdotcom have also stated they ban 300K to 400K accounts each month - no all necessarily cheaters. However the statistic is not all that useful as they have no way of knowing whether it's the same people or spam-bots each month. They quote this statistic to provide some reassurance to the player base, personally I think it does the opposite.

I played blitz for three years on chessdotcom and only twice received rating refunds. This concurs with my subjective game assessment - cheating is very rare in the 1300 to 1600 level. Only once have I been suspicious following a game on LiChess - player instantly played the first 10 to 15 moves - quite possible of course but quite different from the norm.

Personally I wouldn't worry about cheating unless you're are a 'titled player'.

When chessdotcom released there 'dirty dossier' on Hans Neimann they accidently revealed - via some simple arithmetic - that 'titled' players were at least ten times more likely to cheat than Ordinary Joes. Bit of a shocker but there you go. chessdotcom have also stated they ban 300K to 400K accounts each month - no all necessarily cheaters. However the statistic is not all that useful as they have no way of knowing whether it's the same people or spam-bots each month. They quote this statistic to provide some reassurance to the player base, personally I think it does the opposite. I played blitz for three years on chessdotcom and only twice received rating refunds. This concurs with my subjective game assessment - cheating is very rare in the 1300 to 1600 level. Only once have I been suspicious following a game on LiChess - player instantly played the first 10 to 15 moves - quite possible of course but quite different from the norm. Personally I wouldn't worry about cheating unless you're are a 'titled player'.

boilingFrog

There's one giveaway I can mention here, and that is when your opponent's 'presence indicator', that little green light, goes off when it's their turn and then turns back on some several seconds later ... and then they play one of the top 3 or 4 moves ...

Once again, not entirely conclusive since they may have been away chatting with someone, but if you're chatting with a 3rd party during a game that's pretty sus too, right ?

And finally, there's practically limitless variations on all the above ...

There's one giveaway I can mention here, and that is when your opponent's 'presence indicator', that little green light, goes off when it's their turn and then turns back on some several seconds later ... and then they play one of the top 3 or 4 moves ... Once again, not entirely conclusive since they may have been away chatting with someone, but if you're chatting with a 3rd party during a game that's pretty sus too, right ? And finally, there's practically limitless variations on all the above ...

emaN-drawkcaB

@boilingFrog said in #22:

when your opponent's 'presence indicator', that little green light, goes off when it's their turn and then turns back on some several seconds later ... and then they play one of the top 3 or 4 moves ...
I do that too, however my moves are usually top 11-15...

@boilingFrog said in #22: > when your opponent's 'presence indicator', that little green light, goes off when it's their turn and then turns back on some several seconds later ... and then they play one of the top 3 or 4 moves ... I do that too, however my moves are usually top 11-15...

boilingFrog

edited

@emaN-drawkcaB said in #24:

I do that too, however my moves are usually top 11-15...

I'm going to ask Dr. Fauci to make a nice vaccine for folks like you ...

@emaN-drawkcaB said in #24: > I do that too, however my moves are usually top 11-15... I'm going to ask Dr. Fauci to make a nice vaccine for folks like you ...

Kinzhali

Gut feelings should not be ignored. The most prominent indicator of cheating, in my opinion, is extreme variance in strength of play between different phases of the game (e.g, blundering a piece in the opening and then subsequently organizing an impenetrable fortress swiftly and with time to spare, thereby forcing a draw or winning on time).

In my opinion, the most common form of cheating is from players who consult engines only after making near fatal errors. I suppose the reason that doing so merely "evens the odds" since "that is not typically a mistake that I would make." Over time, the threshold for errors that trigger corrective engine consultations grows smaller and smaller; at that point, the cheating starts at every feeling of uncertainty (i.e. to prevent errors rather than attempting to correct them)

Gut feelings should not be ignored. The most prominent indicator of cheating, in my opinion, is extreme variance in strength of play between different phases of the game (e.g, blundering a piece in the opening and then subsequently organizing an impenetrable fortress swiftly and with time to spare, thereby forcing a draw or winning on time). In my opinion, the most common form of cheating is from players who consult engines only after making near fatal errors. I suppose the reason that doing so merely "evens the odds" since "that is not typically a mistake that I would make." Over time, the threshold for errors that trigger corrective engine consultations grows smaller and smaller; at that point, the cheating starts at every feeling of uncertainty (i.e. to prevent errors rather than attempting to correct them)

petri999

what ever advice you get is just wrong most of the time. catching cheater no where as simple as looking at centipawn loss /move or accuracy. Or time between the moves. All these statistical measures that have some meaning only as part large sample. Best way to increase

And number of cheaters "has grown lately" as long I have been here. Lets assume it is true then absolutely everyone would be cheater by now. So problem would be similar to this https://www.youtube.com/watch?v=6DpLYHKzqbE

Yes there are cheaters but it is not rampant. And most people when they think opponent cheated they just played badly.

what ever advice you get is just wrong most of the time. catching cheater no where as simple as looking at centipawn loss /move or accuracy. Or time between the moves. All these statistical measures that have some meaning only as part large sample. Best way to increase And number of cheaters "has grown lately" as long I have been here. Lets assume it is true then absolutely everyone would be cheater by now. So problem would be similar to this https://www.youtube.com/watch?v=6DpLYHKzqbE Yes there are cheaters but it is not rampant. And most people when they think opponent cheated they just played badly.

petri999

@boilingFrog said in #22:

There's one giveaway I can mention here, and that is when your opponent's 'presence indicator', that little green light, goes off when it's their turn and then turns back on some several seconds later ... and then they play one of the top 3 or 4 moves ...

I can thing on any reason why cheater would need be " not present" to cheat. Only thing that indicates is a bad connection.

@boilingFrog said in #22: > There's one giveaway I can mention here, and that is when your opponent's 'presence indicator', that little green light, goes off when it's their turn and then turns back on some several seconds later ... and then they play one of the top 3 or 4 moves ... > I can thing on any reason why cheater would need be " not present" to cheat. Only thing that indicates is a bad connection.

oneiro_naut

@AlexiHarvey said in #21:

chessdotcom have also stated they ban 300K to 400K accounts each month - no all necessarily cheaters. However the statistic is not all that useful as they have no way of knowing whether it's the same people or spam-bots each month. They quote this statistic to provide some reassurance to the player base, personally I think it does the opposite

Exactly. "Our restaurant kills 300-400 cockroaches every month!"

@AlexiHarvey said in #21: > chessdotcom have also stated they ban 300K to 400K accounts each month - no all necessarily cheaters. However the statistic is not all that useful as they have no way of knowing whether it's the same people or spam-bots each month. They quote this statistic to provide some reassurance to the player base, personally I think it does the opposite Exactly. "Our restaurant kills 300-400 cockroaches every month!"

Brian-E

To answer the title question "How to Detect Cheaters", the answer has to be: it's very difficult indeed and you need specialist software plus a great deal of expertise in how to use it.

Much of the pioneering work was done by Ken Regan, and here is a rather dated article about him and his work:
https://cse.buffalo.edu/~regan/personal/JuneCLarticleKWR.pdf

How much has moved on in the field since then, I don't know. But I know enough to leave it to the experts!

Report any suspicions you have about your opponents by all means. But don't expect to know anything for sure.

And above all, don't worry about it. Just enjoy the game.

To answer the title question "How to Detect Cheaters", the answer has to be: it's very difficult indeed and you need specialist software plus a great deal of expertise in how to use it. Much of the pioneering work was done by Ken Regan, and here is a rather dated article about him and his work: https://cse.buffalo.edu/~regan/personal/JuneCLarticleKWR.pdf How much has moved on in the field since then, I don't know. But I know enough to leave it to the experts! Report any suspicions you have about your opponents by all means. But don't expect to know anything for sure. And above all, don't worry about it. Just enjoy the game.

Your network blocks the Lichess assets!