Cheating in chess ain’t so black & white


On October 4, the first day of the 2022 US Chess Championship, Chess.com released a 72-page report detailing an internal investigation of grandmaster Hans Niemann. Magnus Carlsen, the reigning world chess champion, accused Niemann of cheating after he lost to him in the Sinquefield Cup last month. On his part, Niemann has confessed to cheating in online games, but not the particular game against Carlsen. The report states he probably cheated in many more games.

Cheating in chess – or being accused of cheating in chess – is not new. In 1997, world champion Garry Kasparov accused the IBM supercomputer Deep Blue team of cheating. The Russian believed humans helped the supercomputer to win. Twenty-five years on, Carlsen believes Niemann of cheating by taking artificial intelligence (AI) help.

Cheating and its accusations in chess go longer back in time. An apocryphal story tells us of a dispute that supposedly led England’s 11th century king Cnut to kill Ulf Thorgilsson, a Danish nobleman. A rare mention of chess in William Shakespeare’s plays is in The Tempest, where Miranda accuses Ferdinand of ‘playing [her] false’ (Act 5, scene 1). Then there’s the 18th-19th -century mechanical ‘Turk’ that allowed a human chess player to hide inside to operate the machine.

In the 1978 world championship, Viktor Korchnoi wore reflecting mirror sunglasses to apparently shield his eyes from Anatoly Karpov’s glare. Karpov complained that the light was reflecting off the mirrors and into his eyes. Korchnoi’s team even alleged that the colour of the blueberry yogurt served to Karpov during a match may have been an encrypted message for offering a ‘draw’.

Ever since Deep Blue’s triumph over Kasparov, AI has transformed the way humans play chess — and the way they cheat too. In the 2006 world championship, Veselin Topalov‘s manager tried to justify ‘statistically’ – with correlation figures – that Vladimir Kramnik had been cheating with the chess program Fritz 9. Australian FIDE (Federation Internationale des Echecs) master Bill Jordan not only discussed the old ‘art,’ but also a bit of ‘technology’ in his 2021 book, The Art of Cheating in Chess: The Many Faces of Cheating. Technology-based cheating has become an integral part of online – especially post-pandemic – chess.

Today, a free chess engine like Stockfish has an Elo rating – method for calculating the relative skill levels of chess players – of 3,546. The highest ever human score is Carlsen’s 2,882 in 2014. In online chess, guidance from Stockfish is just a click away. Aided by it, even I can beat Carlsen.

Chess.com has already closed more than 500,000 accounts for cheating, including 500-plus titled players. Combating online cheating is never easy, though. Millions of chess games stored in the database are used to create a statistical model that assesses the low probability that a human player will match the top choices of an engine. And computer programs don’t just play better, they play differently. As Jordan writes, ‘A high correlation between the moves of a human player and an engine could indicate cheating.’

A cheat may seek the help of an engine only in a few critical positions making it difficult to identify the correlation, yet substantially enhancing the winning chance. Such statistical evidence, however, is just an ‘indication’ of possible cheating. That’s all. False positives may very well occur.

The Chess.com report stated that Niemann had ‘likely cheated’ more than 100 times in online chess – ‘His results are statistically extraordinary.’ Once again, such statistical evidence is no definite proof. It can, at best, shed some doubt. And the report didn’t find any ‘direct evidence that proves Hans cheated at the September 4, 2022, game with Magnus’.

Allegations of online cheating create paranoia. The way Carlsen sees it, the future of the 1,500-year-old game is at stake. So is the future of Niemann. The inconclusive ‘statistical’ proof could well lead to the 19-year-old grandmaster, effectively, being cancelled.

The writer is professor of statistics, Indian Statistical Institute, Kolkata



Source link