Cheating in chess

[Matt Bridgeman]

What's the difference between acting dishonestly and using a ploy to gain an unfair advantage? For example, someone employing a parapsychologist to sit in the crowd and stare you down the thrown you off your game, as someone once did whilst attempting to become world champion?

The will be Karpov - Korchnoi 1978. ('...whilst attempting to be a world champion?')
Team Karpov, then Karpov was the world champion, had Dr. Zhukhar in the audience to put Korchnoi off and Korchnoi really believed that he was being hypnotised and it was affecting his play. Team Korchnoi put anti Zukhar measures in place by employing a couple of yoga experts to restore Korchnoi's confidence.

The match was 3-3 (6 draws) but in game 7 Karpov suddenly refused to shake hands before the start of the game. This upset Korchnoi so much so he lost. (cheating, gamesmanship, rude, or unsportsmanlike?)

Fast forwards a few decades then we had toiletgate!

It’s funny in a way how that ironic image ruined Rausis. Whereas now you have these faceless individuals in America kind of making a business model out of milking FIDE points for American junior players using Hybrid Chess. Their crimes are arguably much bigger, but their punishment is more along the lines of a firm, ‘you can stop that now’ from FIDE. But none of the punishment and humiliation that Rausis got. Unless FIDE is following up and looking to sanction people? I may be jumping the gun here.

[Simon Brown]

Toiletgate was 2006, not Rausis

[Ian Thompson]

It’s funny in a way how that ironic image ruined Rausis. Whereas now you have these faceless individuals in America kind of making a business model out of milking FIDE points for American junior players using Hybrid Chess. Their crimes are arguably much bigger, but their punishment is more along the lines of a firm, ‘you can stop that now’ from FIDE. But none of the punishment and humiliation that Rausis got. Unless FIDE is following up and looking to sanction people? I may be jumping the gun here.

I don't see any significant difference between paying an opponent to lose/only draw a game in a hybrid event and doing the same thing in a normal OTB event. We all know the latter's been going on for decades to get norms. How many players have been sanctioned for doing it? Hardly any, as far as I'm aware.

Perhaps the tricker situation is the norm tournament where there's no explicit agreement to fix the result of games, but an understanding that the titled players won't try too hard against the norm-seekers if they want to get invited to future tournaments.

[NickFaulks]

Perhaps the tricker situation is the norm tournament where there's no explicit agreement to fix the result of games, but an understanding that the titled players won't try too hard against the norm-seekers if they want to get invited to future tournaments.

That was the background to the recent rule change that one norm must come from a "big Swiss", in which the likelihood of anything of that sort is presumed to be minimised. Unfortunately, the usual vested interests got that watered down in the details to the point that it has almost no teeth.

[Ian Thompson]

Anyone who thinks they're any good at spotting computer use in chess might like to try do this study being run by David Smerdon.

[Roger de Coverly]

Anyone who thinks they're any good at spotting computer use in chess might like to try do this study being run by David Smerdon.

I've long thought that something along those lines to detect the reliability or otherwise of methods purported to detect in-game engine usage was desirable.

[Christopher Kreuzer]

Anyone who thinks they're any good at spotting computer use in chess might like to try do this study being run by David Smerdon.

I've long thought that something along those lines to detect the reliability or otherwise of methods purported to detect in-game engine usage was desirable.

You would measure the reliability, not detect it. I don't thinkk it is possible (from game moves alone) to reliably detect engine use if such use is sporadic and low-level. (Continuous engine use is a different matter.) For occasional use, you simply don't have enough data. You have to have random 'testing' and changing of playing conditions to disrupt any methods being used. Plus you need those who cheat to become 'white knights' or 'white hats' (admit to their methods and help construct defences) so that the methods being used can be guarded against.

[Ian Thompson]

I don't thinkk it is possible (from game moves alone) to reliably detect engine use if such use is sporadic and low-level. (Continuous engine use is a different matter.) For occasional use, you simply don't have enough data.

That's what this study is trying to determine. You're told some players were provided with engine moves at critical moments in the game and asked to decide which players based on player rating, moves played, thinking time per move and game evaluation each move.

[Christopher Kreuzer]

I don't thinkk it is possible (from game moves alone) to reliably detect engine use if such use is sporadic and low-level. (Continuous engine use is a different matter.) For occasional use, you simply don't have enough data.

That's what this study is trying to determine. You're told some players were provided with engine moves at critical moments in the game and asked to decide which players based on player rating, moves played, thinking time per move and game evaluation each move.

I had picked up on that. It is an interesting study. I am pondering whether to take part, because I am not sure if the time limit of 1 hour to review 11 games is realistic (that is only about 5 minutes per game) - I understand they had to give a time limit, and they do provide game evaluation to help with the assessment, but if you are unsure or run out of time, do you guess or is there an option to indicate "more evaluation needed"?

All of the games were played by real players (Elo 1700-2300) in a special "Cheating Challenge" rapid tournament (20min + 20sec) we held earlier this year.

With 11 games to review, is that 6 players in an APA and 15 games. Or are the 11 games you review part of a much larger tournament (e.g. 45 games from a 10-player APA). Or maybe it was a Swiss format?

It is also interesting that the ELO range chosen is so low. Will this be applicable to players with ELO in the 2300-2800+ range? Maybe if the results are meaningful, the experiment will be extended?

I do wish the researchers all the best with their efforts, and I hope that they get reproducible results. (I would guess that players with ELO 2400+ taking part will be able to detect the cheats with a high rate of accuracy.)

[Ian Thompson]

I am pondering whether to take part, because I am not sure if the time limit of 1 hour to review 11 games is realistic (that is only about 5 minutes per game) - I understand they had to give a time limit, and they do provide game evaluation to help with the assessment, but if you are unsure or run out of time, do you guess or is there an option to indicate "more evaluation needed"?

There are only 7 games. It says 11 games in some places and 7 games in other places, but 7 is right.

They ask you to say how confident you are in your answer, on a scale of 1 - 10, so you could make that quite low if you needed more time.

All of the games were played by real players (Elo 1700-2300) in a special "Cheating Challenge" rapid tournament (20min + 20sec) we held earlier this year.

With 11 games to review, is that 6 players in an APA and 15 games. Or are the 11 games you review part of a much larger tournament (e.g. 45 games from a 10-player APA). Or maybe it was a Swiss format?

They don't say.

I would guess that players with ELO 2400+ taking part will be able to detect the cheats with a high rate of accuracy.

I wouldn't. My guess is that most people will get 3 or 4 right, meaning that 'expert' opinion is no better than guessing.

[LawrenceCooper]

Kramnik's protest at being paired with someone he has previous reported for cheating: https://www.chess.com/game/live/8524762 ... mirkramnik

[Geoff Chandler]

Kramnik's protest at being paired with someone he has previous reported for cheating: https://www.chess.com/game/live/8524762 ... mirkramnik

Hi Loz,

Same day, same two move mate v Kramnik but different opponent.

https://www.chess.com/events/2023-title ... v_Yaroslav

[Tim Spanton]

Fool's Mate may have to be renamed Kramnik's Gambit

[Roger de Coverly]

Same day, same two move mate v Kramnik but different opponent.

https://www.chess.com/events/2023-title ... v_Yaroslav

I think it's actually the same opponent, as Liminov is a handle.

Someone famous or infamous in Russia used that name as well it would seem.
https://en.wikipedia.org/wiki/Eduard_Limonov

[Geoff Chandler]

I'm reading a post from a few days ago that when playing Kramnik his opponent MVL. threw a game.

"Vlad dropped a queen from a mouse slip. Playing from his phone due to computer problems. MVL graciously threw the next game to even the score."

"threw the next game" I've not been bothered to look for the MVL v Vlad thrown game, (anybody? I find that site a bit of nightmare to navigate.) but could it be that fools mate is standard at that level to show a deliberate loss for sporting reasons. (just a thought)