Public Accusations Of Cheating Are Bad

The story about a Russian trainer accusing a 13-year-old girl of cheating led to a huge debate on our site. Currently, the report is close to reaching 700 comments. The case is another example of how dangerous it can be to accuse someone without having clear proof.

Nobody would disagree that cheating is a problem in chess that has to be dealt with. An engine on a smartphone is stronger than Deep Blue in 1997, which means that in a tournament without anti-cheating measures it's too easy to cheat. Over the past decade (and beyond), there have been many examples of people getting caught cheating in chess.

Recent versions of the FIDE Laws of Chess include measures that prevent players from having electronic devices on them during a game. Arbiters are also allowed to inspect a player's clothes, bag or other items.

Furthermore, in 2014 FIDE together with the Association of Chess Professionals (ACP) established an Anti-Cheating Commission (ACC), which took off with a set of guidelines published in November of that year (here in PDF).

The document mostly deals with actual cheating cases during tournaments. Albeit briefly, unfounded accusations are also discussed. Currently, the ACC's guidelines suggest a maximum suspension from FIDE rated events of three years for a first-time violation of the anti-cheating regulations.

The problem with the recent case was not that the Russian trainer, grandmaster Evgeniy Solozhenkin, had no reason to suspect the girl of cheating. It's debatable whether her sharp rating increase in previous months was suspicious, or that she found some very strong moves in certain positions, or that she alternated very strong performances with weaker ones. But, combine this with his daughter hearing her ask "what's the evaluation?" in a toilet cubicle, and you have enough material for a justified request to the arbiter to keep an eye on the girl for the remainder of the tournament.

And that's what the trainer did.

This was perfectly rational behavior, and not the reason why Solozhenkin was sanctioned by the FIDE ethics commission. (It was the ethics commission where the girl's mother had filed a complaint.) The commission reasoned that the trainer should have stopped there; he should have refrained from publishing his accusations in several articles online, in which, you might say, he performed character assassination on the girl, without awaiting the results of the investigation.

The trainer's penalty doesn't seem utterly unfair, and the public letter signed by grandmasters might have been the result of a misunderstanding. With a maximum term of three years, the commission went for half, and in practice probably a quarter of that: 18 months, half of which is suspended. Besides, he is allowed to participate as a player in FIDE events inside Russia and to pursue activities as a trainer, coach or manager there, meaning he can retain an income for himself and his family.

We have seen multiple examples of unfounded accusations in recent years, the most famous being "Toiletgate." At the Elista world championship in 2006, Vladimir Kramnik visited the toilet numerous times during his games with Veselin Topalov. The latter, and his manager Silvio Danailov, accused Kramnik of cheating, but their allegations backfired: evidence of computer assistance was never found, and many chess fans turned against the Bulgarians. By the way, in the 2005 world championship tournament in San Luis, it was Topalov himself who had been accused of cheating by a small group of grandmasters.

During the 2009 Aeroflot Open in Moscow, Shakhriyar Mamedyarov, now the world's number-three in the live ratings, accused Igor Kurnosov of cheating. Mamedyarov's main argument was that his opponent's moves matched those of an engine. Kurnosov, who tragically died in a traffic accident four years later, was never caught cheating.

And then there was the "Sandu case," only three years ago. WGM Mihaela Sandu was not accused of cheating, but her performance at the European Individual in Chakvi, Georgia led to a group of participants writing two letters to the organizers: one asking for broadcasting the games with a 15-minute delay (which makes it harder to get external assistance), and another to not include Sandu's games in the transmission altogether. The organizers then made the mistake of putting these letters on a public clipboard, after which they were published by chess media. Sandu eventually won a case with the FIDE ethics commission and Natalia Zhukova, who had led the protest, received a three-month ban which was wholly suspended for a period of one year.

These quite different cases, all have something in common: there was no clear proof, but the accused players were nevertheless suspected— by the accuser obviously, but also by many chess followers. They tend to reason "when there's smoke, there's fire," and argue that this trumps the assumption of "innocent until proven guilty."

Another similarity is that in each case, there were only losers. While history says Kurnosov and Sandu were not guilty of cheating, their names will always be linked to cheating in Google, and this also counts for their accusers. Chess itself loses the most, because such stories only make the game less popular, and less interesting for sponsors.

The most important similarity, however, is that the cheating allegations became public before an investigation had started, or before the results of an ongoing investigation were published. Mamedyarov wrote an open letter, the letters by the women in Chakvi were made public by the organizers, and Solozhenkin put the information online himself.

Freedom of speech is an important value in a modern society, but it comes with consequences, e.g. in cases when expressions are harmful or offensive. Public accusations of cheating have the potential to be extremely harmful for the accused player, and thus should be avoided—especially because an actual procedure for dealing with cheating suspicions exists.

The correct procedure, initially followed by Solozhenkin, is to file a complaint to the tournament arbiter, who can then take necessary measures such as searching the player. He is also entitled, in contact with the anti-cheating commission, to request an analysis by Kenneth Regan, Associate Professor at Buffalo University and FIDE's authority where it comes to discovering whether a player is playing like a computer.

There are two ways to prove cheating: discovering material evidence such as an electronic device on the body of a player, or a statistically significant collection of data.

Some chess fans believe that the statistics cannot prove cheating, but they can—in the same way that fingerprinting does not provide actual physical evidence of a crime, but with a 1/64,000,000 chance of a false positive. Chess.com’s own system of cheating detection has some similarities to other published statistical methods but is arguably much more sophisticated and based on more data. At the core of this highly complex tool is a statistical model that evaluates the probability of a human player matching an engine’s choices, and a comparison to the possibility of their actual strength of play across many games.

Statistical systems cannot prove a 100 percent proof of cheating in the same way that video footage or an in-person witness could, but, they have become so strong that numbers like "99.99999 percent certainty" are not uncommon. These kinds of results can direct arbiters to do even stronger checks of players at events. Local circumstances combined with analytical research can, and should, be enough to determine if someone’s been cheating or not.

A technical, analytical instrument like what is used on Chess.com should be used in future over-the-board events as well, if only because always detecting physical cheating devices or cooperative systems may be impossible.

Sure, FIDE should probably invest more in measures such as metal detectors at its official events (if it's true that at the World Youth they were only available from round five, this is an obvious mistake and organizers/arbiters should be penalized). 

But there will always be events, such as open tournaments, that cannot afford, or simply don't want to have such strict measures. In many cases, the cure might be worse than the actual disease, as Arne Moll argued back in 2012.

Cheating cases and unfounded cheating allegations will surely happen again in the future. Let's hope the parties involved will keep their cool, and follow the procedures at hand. And if these procedures, or the commissions behind them, are not sufficient, let's try to improve them. Any other way of dealing with it, might cause more harm than necessary to players and the sport itself.

This is a personal blog which doesn't necessarily reflect the opinion of the whole editorial team of Chess.com.