During the pandemic, there has been an explosion of people playing games online. While that has been enjoyable for friendly games, problems can arise when it comes to serious tournaments that can influence the rankings that players have.
In the case of chess, the website chess.com is widely used. The main problem with cheating in chess arises because there are now very powerful chess engines that can rapidly analyze situations and suggest the best move. These engines are used by players to practice their game and by analysts at tournaments who report on games. In face-to-face contests, the use of such engines is not allowed while play is going on but when people are online, it becomes hard to monitor whether someone is having an engine running simultaneously to guide them. The only way to detect such cheating is statistical, to see if the quality of the moves made by a player lie outside the range that one might expect of someone with their ranking.
In chess, a controversy arose recently when Dadang Subur, an amateur player who had been playing exceedingly well online under the name @Dewa_Kipas and whose ratings had risen considerably in a short time, was accused by another player Levy Rozman of using chess engines to decide on moves. This article explains what happened.
On March 2, 2021, in one of his streams, Rozman (@GothamChess) played a 10-minute game on Chess.com against member @Dewa_Kipas. As Rozman noticed that the account had gained over 800 Elo points in just 11 days, he became suspicious. This suspicion was fueled by how his opponent regularly spent 10 seconds making moves, even when the choice was obvious. Such time usage often reveals that a player is using external assistance such as a chess engine.
When he lost the game, in front of 12,000 viewers, Rozman checked the account and saw that in almost all of his wins, Kipas had an accuracy of over 90 percent. Rozman concluded that he was dealing with a cheater and reported the account to Chess.com. A few hours later the account was closed for violating the Fair Play policy.
There was then a counter-protest, claiming that Subur had worked very hard to improve his game.
Deddy Corbuzier, an Indonesian actor and television presenter, suggested that Subur play a face-to-face game in his studio with Irene Sukandar, an International Master in chess, to see what could be learned. The live-cast of the games drew 1.25 million viewers. And the result?
Monday’s match between Sukandar and Subur involved serious money. The Indonesian technology and e-commerce company Tokopedia appeared to be the main sponsor, providing the equivalent of $10,500 that was then doubled by Indra Kesuma, an Indonesian businessman and YouTuber.
As it turned out, Subur could not repeat the high level of play as he had shown online. He lost 3-0 to Sukandar, making fairly basic mistakes.
Chess.com estimated his performance in these three match games at an Elo of 1127. His online performance rating exceeded 3000 Elo, the highest performance Chess.com’s Fair Play system calculates for human games. Despite the poor showing, he still made roughly $7,000 while Sukandar earned about $14,000.
It is of course possible that Subur was just having a bad day. But now the burden shifts to him to provide evidence that he is as good a player as his online rating suggested.
The chess problem isn’t new. The pandemic has exaggerated the problem though.
There have been other professional sports that can be done online but before the pandemic hadn’t moved there.
Like the various car sports. The drivers regularly train in a simulator anyhow. In one of those races the result was this, cheating by having a better driver at another location take your place.
Well, you can do it with cameras monitoring the players, checking that they’re not using other software (like a chess engine, tablebase, etc. including on smartphones or whatever), and taking some additional measures like that. This happens in at least some online tournaments. However, a game between two random people is another story, which isn’t to say that those couldn’t be made more secure but that they’re not.
How about a low tech solution? Play on camera with a mirror behind so everyone can see what the player is seeing.
A second camera might be an idea, but its limited field of view makes cheating possible.
Doesn’t deal with the option of some other person watching the game, feeding the software, and then passing the results back to the player using some kind of covert communication…
Back in the Dark Ages of the web I remember several friends who played Scrabble online but stopped when they realized that two of the players were using available Scrabble engines to produce seven-letter solutions several times in any one game.
This is a persistent problem with any online game that gets popular. The more popular the game the more effort people are willing to put into cheating. It’s particularly obvious with the games that have both online and live tournaments, all of which have players with super rankings online that are mediocre live.
MMO RPGs have a huge problem with this. There are entire companies that have employees that play the MMOs and collect rewards. The companies then sell those rewards to other players for real world cash. It mucks up the balance and economy of the MMO and turns them into pay-to-win games.
It may seem like a silly comparison but it’s really the same problem. It would be like a chess player hiring somebody to feed him hints during a tournament. They are buying their way up the rankings and status rather then working harder and playing better.
Even in online bridge tournaments they have detected cheating by being able to inform the partner
by SMS the cards they have
Dunc, #4:
Could you clarify that? What do you have in mind?
“It is of course possible that Subur was just having a bad day.”
As a chess amateur and math teacher I understand a few things about ratings.
“the account had gained over 800 Elo points in just 11 days”
This is possible, even likely, when a world class player starts from scratch.
“Chess.com estimated his performance in these three match games at an Elo of 1127. His online performance rating exceeded 3000 Elo.”
Chances are higher that you’ll find a lottery ticket on street and win the jackpot -- even if the world class player has a bad day.
For details: google ELO rating.
cr, @ #8: There are all sorts of possibilities, from very simple options like a screen carefully positioned so that it’s visible to the player but out of sight to the monitoring cameras, through slightly more tricky options like very small earpieces (there are hearing aids with Bluetooth capability which are practically invisible), all the way up to the sort of tactile systems sometimes used by card counting teams… You don’t need to be able to pass very much data, just a single move, so you could pretty easily come up with a system to encode that via something like a buzzer in a shoe.
Serious tournaments are played with two webcams. The player has one camera on their face for streaming, one behind their back showing their hands and their workplace, and they must share their screen the whole time too.
_________________________________
I have observed this issue closely because I am subscribed to Levy Roseman’s Youtube channel. There is no doubt in my mind that Dewa Kipas was cheating
I have too suspected one player of cheating against me, which I do not do often. I am fairly charitable to my fellow humans. I have been reported him but my report was dismissed. However, one month later he was indeed banned. He was an intermittent cheater, so it possibly took that much time to amass enough statistical evidence, but he did get the banhammer in the end.
Chess.com is fairly secretive about how they identify cheaters, but it is possible to make some educated guesses. They have probably several criteria to look at, and here is my guess at what they are:
1. The time between moves. Experienced players can play memorized opening moves within a fraction of a second. If someone consistently has a high rating and takes a long time to make beginning moves, it is an indicator of engine use.
2. Distribution of times the moves take during a game. I have not made a proper analysis, but my guess based on looking at my own games would be that they should conform to a Weibull distribution.
3. The length of winning-losing streaks. These should probably be pretty randomly long. Consistent patterns of extremely long winning streaks and no losing streaks are a bit suspicious.
4. The win/loss ratio. The site does a fairly good job at pairing people of similar strength, so it should be about 50/50. Even when your ELO is going up. I have gained 300 ELO over half a year and I do have circa 50/50 win to lose ratio.
5. Game accuracy and consistency. It is possible, even for weak players like me, to get accuracy over 90%, or even an occasional perfect game without mistakes and blunders. But a streak of twenty nearly flawless games is unlikely, even for titled players.
6. Rating growth speed. Titled players can send in their certificate and they get assigned rating accordingly, they do not need to start at the basic rating like everyone else. For an untitled player, the faster they gain rating, the more suspicious it is.
AFAIK, Dewa Kipas ticked all of these to some extent.
There are people who pay actual money for cheat mods in online shooting games even though there is no benefit (it’s not a tournament, there are no prizes, there is no fame). They just want the “win”, and they see that win as legitimate even though they cheated. Many of these cheaters throw a fit if they end up getting banned for doing this.