The chess world is currently consumed by a drama as lurid and compelling, in its way, as the "Don't Worry Darling" fracas. Involving implications of cheating at the highest levels of play, the feud between the world champion and an upstart challenger has prompted speculation on the existential threat to chess posed by an AI engine tiny enough to be concealed somewhere on — or in — the body.
The idea — unsupported by any evidence, it must be emphasized — that a player could surreptitiously consult an unbeatable chess engine even when playing over the board has been batted around for years. But recent events have made people think seriously about the possibility and what it might mean for the future of the game.
The saga began two weeks ago, when current chess world champion and one of the strongest players in history, Magnus Carlsen (pictured above), began a match at the Sinquefield Cup with Hans Niemann, a 19-year-old grandmaster who has ascended from respectable to downright dangerous over a remarkably short period of time.
Carlsen was playing with the white pieces, and therefore going first — an advantage he is particularly adept at using, having not lost a game in years with white and seldom even taking a draw. Yet soon he had not just forfeited the game (which you can watch here), but he had also withdrawn from the tournament, cryptically tweeting what seemed to many to be a veiled accusation of cheating by Niemann. He has not elaborated on these actions despite officials, fans, colleagues and even the likes of former world champion Garry Kasparov asking him to speak out. (After this article was published, Carlsen told an interviewer he will give a statement after the Julius Baer tournament currently underway, in which both are playing.)
Niemann, for his part, has naturally and emphatically denied any cheating and said that an apparently miraculous preparation for the unusual line of attack Carlsen took was one he happened to include after seeing it in a game from years earlier. The tournament organizer has stated that there is no indication of any suspicious behavior or wrongdoing. Others have examined the record and found no indication of cheating.
The event and its fallout (only given in outline here) have sharply divided the chess world, as even the conservative approach of "let's wait and see" tacitly sustains the idea of Niemann cheating, so there is precious little neutral ground to occupy. FIDE, the official international chess organization, is expected to issue a statement soon (Update: here) that may shed light on things, but it won't change what's already happened.
(Update 9/26: Carlsen has published a statement confirming that he suspects cheating but stopping short of making direct accusations or factual claims.)
A checkered history
Promotional image from a recent matchup between Carlsen (left) and Niemann (right). Image Credits: Meltwater Champions Chess Tour
To be clear here, there is no question that Niemann is an extremely high-level player — he has played hundreds of games against extremely strong players in situations where cheating is all but impossible and won decisively.
It was noted by other GMs that Carlsen had played poorly (for him), and Niemann had simply gotten lucky with his prep, played well and perhaps rattled the champion, leading to an advantageous position. But Carlsen is not easily rattled, nor is he one to storm off after making a blunder — still less to cast unfounded aspersions on an opponent. They have faced off before — one fan even caught them playing a friendly barefoot match on the beach in Miami just weeks earlier.
Gölgede ve güneşte satranç :P Oynayanlar Magnus Carlsen ve son dönemin yükselen genç yıldızlarından Hans Niemann yer de Miami. Buradan muhtemelen Niemann'ın Ankara'da oynanancak Türkiye İş Bankası Süper Ligine gelmediğini anlıyoruz.:) pic.twitter.com/dFoAx7x5oB
— Ozgur Akman (@ozgurakmann) August 14, 2022
But Niemann, like anyone, has a past. It came out that Carlsen was likely aware that Niemann has been caught cheating online before — at least twice on Chess.com, when he was 12 and 16, the latter time seemingly leading to a six-month suspension from prize games. He admitted this in an interview, calling it the foolishness of a young, ambitious player — though he is still young and ambitious — and that it involved asking someone else in the room to furnish him with moves from a chess engine.
I was just a child. I have never ever in my life cheated in an over-the-board game. I wanted to gain some rating so I could play stronger players, so I cheated in random games in chess.com. I was confronted, and I confessed. It was the single biggest mistake of my life and I am completely ashamed.
While there are measures against this type of thing online, they are hardly foolproof. Chess.com has implied that there was more to Niemann's behavior than this, but the details are not public.
Of course it is behavior unbecoming of a grandmaster, but whatever you make of his character, the fact is that cheating online is relatively easy — but cheating over the board is practically "Mission: Impossible." Cheating versus the world champion — at a disadvantage — at a major tournament? The very idea is ludicrous.
Or is it?
The chess community at large, a diverse group of players and commentators of all ages and skill levels, could not help but think about how, if one were insane enough to try to cheat in an over-the-board game with Magnus holding the white pieces, how would you do it? Purely theoretically, for argument's sake, devil's advocate and all that?
It turns out to be not quite so ludicrous as one might think — and there is just enough (admittedly highly circumstantial) evidence to admit a shadow of doubt.
How tech crept in
DeepMind's MuZero is an example of a generalized game-playing neural network strong enough to dominate human players. Image Credits: DeepMind
The funny thing about cheating in high-level chess is the idea that the player would need help in the first place.
In a match between grandmasters, who exactly is one of them going to ask for advice? They can't find a confederate to wink at them from the galleries: Nearly everyone in the world is worse at the game. Chess is more popular than ever, but nevertheless the number of people playing at Carlsen and Niemann's level is in the dozens. Collusion is unthinkable.
A chess engine, however, plays at an even higher level. You may hear the phrase and think of Deep Blue and Kasparov, the man versus the supercomputer, but nowadays engines infinitely superior to Deep Blue are available on any smartphone. Indeed, one could conceivably run one on a tiny computer like a Raspberry Pi Zero. Something you could slip into a pocket, or a shoe, or perhaps somewhere no one would think of looking.
Players undergo strict security measures, of course, and you can see Niemann himself being examined here. The possibility is taken seriously, but advances in technology always move faster than countermeasures.
The difficulty presented by such a theoretical device is twofold.
First, how would it even know the state of the game? After all, they're playing with real pieces over a real board. We can dispense with the idea of wiggling toes or tapping feet to select moves — it becomes impractical (and easily noticeable) very quickly. But as it turns out, most OTB games at this level are streamed online and notated in real time — very shortly after a player places their piece, a virtual board is updated and the move is registered online for others to discuss, play along with, and so on. It would be trivial to pull this information from online, passed to the device wirelessly.
It happens that, according to at least one analysis, Niemann has performed better on OTB games streaming live in this way, and more poorly on ones that aren't (some dispute the analysis, or offer other reasons for this). And in the case of the game against Carlsen, shortly after his forfeit, the stream was placed on a 15-minute delay, which eliminates the possibility of cheating in this way. Odd, but far from conclusive — hardly even suggestive to anyone not already suspicious.
The second difficulty is how the device would communicate its suggestions to the player. One can hardly see a screen with the engine proposing various lines, but you don't need to. Chess is efficiently notated: Qh5, for instance, means Queen to white's far right column for white, fifth row up. People pointed out that a handful of short signals, in Morse code or the like, could provide complete information.
Let us admit that seems a little far-fetched — imagine a grandmaster attempting to look like they're focusing on the game while the engine in their shoe stutters out a handful of promising defenses. In fact, it's been tried and detected. But the truth is it's much simpler than that: As champions of the game have said for decades, anyone good enough doesn't need to be told what play to make — only that there is a play to make.
"All I would need in order to be almost invincible"
Reviewing their matches, even the greatest players spot moments where, had they seen a given line of attack or defense, they could have crushed their opponent or snatched a draw out of the jaws of certain defeat. It's the chess version of "l'esprit d'escalier," when you think of the perfect comeback to some jibe hours later as you climb the stairs to bed.
If a player at the grandmaster level could rely on being told, even once in a game, that this move was potentially crucial, they would be practically unbeatable. There's no need for Morse code — the simplest of signals would suffice to inform the player that there is a play to be made, trusting to their skill to find it.
Carlsen expressed this himself in a (translated from Norwegian) 2021 interview:
The people who get caught are those who cheat in a really obvious and stupid manner. The problem was that he [i.e., a player caught in 2016] was not good enough to see what would've made sense.
Had I started cheating in a clever manner, I am convinced no one would notice. I would've just needed to cheat one or two times during the match, and I would not even need to be given moves, just the answer on which move was way better. Or, here there is a possibility of winning, and here you need to be more careful. That is all I would need in order to be almost invincible, which does frighten me.
At the end of the day, the game doesn't work if you do not trust your opponents. I'm not going to sit here and spread rumors, but it would not surprise me at all if we've had a lot of cheaters, even in big tournaments, that have won and not been caught.
And here we find another little quirk of Niemann's: He is occasionally not great at explaining his chess. Postgame analysis is an enormous part of chess commentary, and players frequently discuss positions, moves and alternatives. In discussion with others who play at his level, Niemann occasionally appears (to others in a position to know) unable to express the reasoning behind a move, what led to it or where it would lead.
This is not that strange in and of itself. Chess is both analytical and intuitive, but flashes of insight may not be equally well remembered by all brains, especially neurodivergent ones common to the game. Not everyone has the expected clinical, comprehensive viewpoint associated with the mindset — as chess has grown, it has embraced new approaches and personalities. Niemann is one such personality, outspoken and opinionated, streaming and tweeting and generally taking part in the discourse like any talented 19-year-old might with their favored community. His way of communicating his chess doesn't have to match what is expected of him.
But in the context of the recent drama, this occasional incapacity to explain his own thought process has been counted against him by his detractors.
Chess will survive tech (again)
There are two other obvious alternative explanations for all of this: First, that Niemann simply beat Carlsen fair and square and this is all a big misunderstanding (though one that would be devastating to Carlsen's reputation for several reasons). The second — for which, again, there is no evidence — is that someone leaked Carlsen's strategy to his opponent, a much more prosaic form of cheating that requires no technology whatsoever.
Should either of these be the case, the Hans-Magnus kerfuffle has still let the genie (back) out of the bottle. High-tech cheating has been an issue for years, basically since chess engines passed a human level of play. Commentators have considered it even in some high-level games but credible accusations remain few and far between. Security measures like metal detectors, banning of all devices at venues, delays on game broadcasts, and so on have been put in place to stop the obvious methods. Yet the possibility remains.
One only has to spend a few minutes thinking of methods to do it with today's technology to conceive of something technically doable, and subtle enough that no one would suspect anything strange was happening. As Carlsen said, a clever cheater would be invincible if they were good enough to compete in the first place. It would not show up in statistical analysis or trip the intuition of strong players, both of which are finely tuned to detecting computer-type chess. (AI's style is inimitable, it seems, in chess and other games it has come to dominate.)
As I was writing this piece, Carlsen and Niemann faced off again in a live-streamed game; Carlsen forfeited after two moves, stunning the chess world and quickly prompting criticism from his peers. It's one thing to harbor suspicions, they said, but to participate and deliberately forfeit a game like that is dishonorable and unnecessary and puts his status as world champion in jeopardy. But others took it as the action of someone who can't say what he knows and would rather lose ignominiously than play in bad conscience. (There is speculation he has presented his case to FIDE and awaits their decision, and is prohibited from discussing it publicly. Indeed, neither of them has posted to Twitter in weeks.)
It is potentially a crisis of confidence in the chess world — the specter of cheating, always present but seldom mentioned, is suddenly in every headline. Such a reckoning may lead to major changes in the chess world on the order of how chess engines did two decades ago. Chess, of course, will remain — but just as players had to learn that they would never be as good as an engine, they may have to accept that undetectable cheating at the GM level is at best possible and at worst systemic.
How will that change the game and community? Many thought that, following Kasparov's defeat at the hands of Deep Blue, humanity would lose its taste for a game it couldn't win. In fact, the opposite happened and the chess scene has become even more vibrant, the level of play higher than ever. Could the same thing happen with the idea that an AI may be concealed in one's opponent's shoe, their tooth, their watch? It's only outlandish until you find out someone has been getting away with it for years.
The drama is still unfolding and it may, in fact, be less far-reaching than this. But the community can't forget and it must reckon with the possibilities it contemplated, if only in theory. Chess will survive and thrive, but it will never be the same again