A charity Simul online chess event on Sunday, with Viswanathan Anand playing against nine celebrities for 'Checkmate Covid', made headlines for all the wrong reasons after one participant admitted to having taken external help. The participant, young Indian billionaire entrepreneur Nikhil Kamath, later had his profile on Chess.com - where the Simul was played - closed for violation of fair play policy.
It also put the spotlight on the ethics of charity online chess events, and indeed the problems in monitoring online chess, an increasingly preferred option in the pandemic age.
Kamath, who claimed to have not played chess since his younger days, recorded a high level of play of around 99% accuracy. He later posted a strangely-worded tweet in defense of his action, almost implying that it was always open knowledge to Anand and the wider world that he'd been assisted in his play.
It is ridiculous that so many are thinking that I really beat Vishy sir in a chess game, that is almost like me waking up and winning a 100mt race with Usain Bolt. �� pic.twitter.com/UoazhNiAZV
- Nikhil Kamath (@nikhilkamathcio) June 14, 2021
Late on Monday evening, Kamath put out another tweet, admitting he was wrong and apologising, while still playing down the exercise of seeking the assistance of Stockfish -- a free, open-source Chess engine -- as being part of a 'fun game'.
The incident set social media abuzz through Monday and Anand even received a sympathetic call from actor Aamir Khan, who too had participated (and lost) in the Simul.
"I had a sensation during the game...you know in a normal game there's usually an ebb and flow," Anand told ESPN, "You get into trouble, your opponent makes a slight mistake and then you get out and then back in, especially so in a Simul. So while my positions were dangerous I still hadn't quite prepared myself for what was happening."
The Simul had Anand playing 30-minute games with White over two sessions. Kamath began his game from a pawn blunder on the first move. "He started on the first move by giving a pawn, so you get a mental picture of his level that, 'Fine, he can't be too good," Anand said. "My first reaction when the game got critical was that, far from letting up, the level seemed to rise, the moves were all accurate and in one piece, like an orchestra.
"Every move of his was reinforced by the next and the combination was perfect. You don't expect that level from someone who plays chess occasionally. At the end of the game I thought, 'Wow, I didn't get a single chance to get back in the game'. I didn't feel confident alleging anything but there was this sense I had..."
Yesterday was a celebrity simul for people to raise money It was a fun experience upholding the ethics of the game.I just played the position onthe board and expected the same from everyone . pic.twitter.com/ISJcguA8jQ
- Viswanathan Anand (@vishy64theking) June 14, 2021
Anand could have flagged Kamath and beaten him on time since the latter had just 13 seconds on his clock but instead he resigned after 34 moves. . It may well have been the winning move from the Indian GM, since his surprising loss immediately shone the spotlight on Kamath's surprising levels of accuracy for an amateur. A couple of other participants had also hit curious levels of play and it was growing clearer that Anand may have gone up against a handful of Stockfish engines in the Simul.
"It's sad that it's come to this. It's made us wary of Simuls in the future quite honestly," Anand's wife, Aruna, told ESPN.
The loopholes in online chess
Online platforms like Chess.com where the Simul was played, use algorithms to monitor how closely a player's moves resemble that of chess engines. The level of play - whether a player is titled or not is also a question - accuracy rate in past games, excessive switching between tabs and behavioural data are mined before someone is barred. Until August 2020, Chess.com claims to have closed nearly half a million accounts for cheating, averaging 500 closures per day for engine use last year itself -- 12.6% of the titled players' accounts closed for fair play were those of Grandmasters. Since the start of this year, they've had 190,000 Fair Play closures so far. The Fair Play team consists of chess experts (including multiple titled players) and engineers who specialise in algorithms used to detect "anomalies" and "patterns" of non-human behaviour.
GM Pravin Thipsay, who coached one of the celebrity guests for three days before the Simul, points out that it's not the hardest task to spot anomalies, especially among non-titled players. "Strong players make more human moves, they bring in their natural style of play as a camouflage unlike not-so experienced guys, who end up using engines for every move, which makes it easier to detect. Ideally, I would have expected the positions to not remain equal beyond first 15 moves against someone like Anand, but in some games even until Move 25 it wasn't so, which was odd," he says.
Earlier this year, YouTube chess star Levy Rozman alleged that one of his untitled opponents was resorting to cheating, citing a dramatic ranking climb and accuracy being eerily similar to that of a chess engine. It resulted in the account being blocked by Chess.com. Rozman would later discover that Dadang Subur, the opponent who goes by the online moniker of Dewa Kipas, was an Indonesian bird-feed seller, and Rozman had to endure threats and hate messages on social media. A rebuttal by the opponent's son came through claiming his father Subur was a retired tournament player and the charges were unfair. Chess.com, however, countered the narrative of fair play pointing that algorithms suggest that Subur's moves matched those of chess engines at an almost non-human rate.
Even over-the-board chess is vulnerable. Six years ago, Thipsay recalls calling out an opponent for suspicious play in a tournament in New Delhi. "The guy was carrying two mobile phones on silent mode inside his shoes, and a concealed micro earpiece which had to be eventually pulled out with a magnet," he says, "He would tap once on the table to convey to whoever was helping him remotely that I hadn't moved my king and double tap to confirm a queen move. What made me even more certain of underhanded tactics was when even before the simplest of moves, he began to take a couple of minutes. He was clearly waiting for the earlier move I'd played to be fed into the engine perhaps and was buying out time. His game was nothing like the 1500-rated player he was and I lodged an official protest. He was frisked and found to be completely wired. In junior events, cheating is especially rampant. Coaches are looking for overnight success from their young students and they are sort of driving these practices."
How chess is policed
Though anti-cheating regulations are in place, the challenge lies in making them foolproof among players competing out of home. According to chess24, for all their gallery of online events, including the qualifiers, challengers and Champions Chess Tour tournaments, they have internal engines providing anti-cheat evaluation as well as third-party anti-cheat evaluation services that prepare reports on each game in addition to multiple cameras in each players' room, screen share, audio capture and earpiece checks. Players are also not allowed to leave their desk and camera view while a game of theirs is in progress without the permission of the Arbiter; this includes bathroom breaks.
There's however, always wriggle room for those who want to cheat as they can use apps like Stockfish on their phones. At the Pro Chess League last year, Armenian Grandmaster Tigran Petrosian was banned for life from playing on the Chess.com server after their fair play panel concluded that he violated rules. The players were monitored through webcams and Petrosian was seen to be looking down often, presumably checking for computer assistance.
"I think this sort of thing (engine aid) is broadly part of the online game and the algorithms that state what percentage of your moves were computer moves, whether it's disproportionate to your level, especially in faster time controls, become relevant in such a scenario," says Anand. "Of course, there are no easy answers here."