Chess cheating blog post

[kvasir]

One perspective is that lot of people in the chess world are depended on Magnus. He will be a big shareholder in chess.com soon when the acquisition of play magnus group (chess24, chessable, ...) is concluded. These same people may be giving Magnus a lot more leeway than is reasonable.

javaness, Did you mean US Virgin Islands, not British Virgin Islands? I think there was a problem with two chess organization and all kinds of irregularities in the US Virgin Islands. Hopefully not all the virgin islands have a chess problem.

[RobertJasiek]

As a chess beginner, I find this explanation of Carlsen's loss convincing: https://m.youtube.com/watch?v=ceccUxOKZpk

[Javaness2]

Yes, my mistake, I got the wrong island group there!

[Elom0]

Chess is like soap opera now.

If Hikaru no Chess has a suspicion of cheating plot

https://www.youtube.com/watch?v=Rel6kLRZ-2w

[kvasir]

Magnus finally came clean with a statement that he doesn't want to play with Hans. I'd say he is also accusing Hans of cheating in the Sinquefield cup game specifically, or is he saying that he was unhappy with the loss and this caused him to revaluate the situation? Maybe I am unfair with Magnus to interpret his words this way but I think the cheating concerns could have been directed to FIDE's fair play commission and then he could just say that he complained, but when it comes to the wrong energy this is nothing in chess history.

Dear Chess World,

At the 2022 Sinquefield Cup, I made the unprecedented professional decision to withdraw from the tournament after my round three game against Hans Niemann. A week later during the Champions Chess Tour, I resigned against Hans Niemann after playing only one move.

I know that my actions have frustrated many in the chess community. I’m frustrated. I want to play chess. I want to continue to play chess at the highest level in the best events.

I believe that cheating in chess is a big deal and an existential threat to the game. I also believe that chess organizers and all those who care about the sanctity of the game we love should seriously consider increasing security measures and methods of cheat detection for over the board chess. When Niemann was invited last minute to the 2022 Sinquefield Cup, I strongly considered withdrawing prior to the event. I ultimately chose to play.

I believe that Niemann has cheated more — and more recently — than he has publicly admitted. His over the board progress has been unusual, and throughout our game in the Sinquefield Cup I had the impression that he wasn’t tense or even fully concentrating on the game in critical positions, while outplaying me as black in a way I think only a handful of players can do. This game contributed to changing my perspective.

We must do something about cheating, and for my part going forward, I don’t want to play against people that have cheated repeatedly in the past, because I don’t know what they are capable of doing in the future.

There is more that I would like to say. Unfortunately, at this time I am limited in what I can say without explicit permission from Niemann to speak openly. So far I have only been able to speak with my actions, and those actions have stated clearly that I am not willing to play chess with Niemann. I hope that the truth on this matter comes out, whatever it may be.

Sincerely,
Magnus Carlsen – World Chess Champion

[hyperpape]

There are claims Niemann’s online games show evidence of cheating. I haven’t read that much, and there was some confusion in the thread I read. Obviously important if true.

[kvasir]

There is usually a lot of really bad analysis floating around for cheating cases but there are exceptions. If you have links to the actual reports then I'd actually be interested, especially if it is not bad arguments and bad statistics.

[RobertJasiek]

In "top" go tournaments, there tends to be a duty to play all one's games, or all those for the time for which one has registered, especially because tournament result tiebreakers can depend on the opponents' tournament results or otherwise influence the fair competition among participants. E.g., this is so for EGF tournaments. Basically, a player has to play against anybody against whom he is paired. Cheating suspicion cannot make an exception because it is not the player who decides whether an opponent might have cheated but the arbitration bodies decide. Resigning prematurely for reasons unrelated to the board position and not having allowed justification (such as sudden illness) also tends to be prohibited as unsportsmanlike behaviour.

There might have been an exception in European go tournaments when quite a few players declared their intention to boycott playing against (Bela-)Russians or tournament participation not due to cheating suspicions but due to ethical considerations related to war; this exception seems to have been mostly dissolved by prohibiting participation of (Bela-)Russians in EGF tournaments. As to cheating suspicion, however, I am not aware of any such exception.

Apparently, some top chess tournaments do not have tournament rules preventing (all) unjustified dropping of rounds or premature resignation. Cheating allegations aside, maybe chess should learn from go and establish similar tournament rules?

Refusing to play particular opponents in tournaments is said to damage their reputation. However, this does not per se involve defamation. Statements besides tournaments of suspecting that certain players might cheat (again) are not exactly defamation, either. However, many perceive it as such nevertheless. Players making (too many) unproved allegations can damage their own reputation.

There is a theory that Magnus, as he has been becoming more involved in Chess.com, might recently have learned about more (real or alleged) cheating instances of Niemann, became more concerned about cheating possibility due to his hard loss against Niemann and therefore now refuses to play against him in tournaments. (Niemann's alleged or misinterpreted sandbagging or weak play in friendly beach fast games against Magnus might have added to the suspicisions.) However, this theory might be replaced by the simpler theory that Magnus simply does not want to play against Niemann in tournaments any more because of suspecting him to be a notorious cheater.

Improved anti-cheating measures can be considered but, eventually, what can one do against body implantations if a player has medically approved metal implantations or heart pacemakers? Passive modern body scan machines like those at airports? Tournament organisers digging rest rooms? Hopelessly ambiguous, over- or misinterpreted AI anti-cheating tools? Promotion of ethical behaviour among a game's world-wide players community?

[RobertJasiek]

On https://www.fide.com/news/1999 the FIDE President mentions "the anti-cheating protocols" of FIDE. What are those?

EDIT:

In https://www.youtube.com/watch?v=jfPzUgzrOcQ&t=702s Yosha Iglésias uses some tool to talk about accuracy / correlation between a game's moves and AI moves and the press mentions some 100% occasions for Niemann. What are those accuracy / correlation numbers? The percentage of identical moves themselves? Some derived alleged likelood of cheating?

[kvasir]

I think what is meant by "the anti-cheating protocols" is the whole body of rules and regulations. This includes

• Anti-cheating regulations
• Anti-cheating protection measures
• Complaint forms

There are also other relevant documents:

• The ethics and disciplinary commission procedural rules
• Ethics and disciplinary code
• and an interpretation guide for the above two but that might be out of date.

Fide has and ethics and disciplinary commission (EDC) and a fair play commission (FPL). Cheating and improper accusations of cheating are the domain of the FPL, I am not sure about other similar transgressions but I think the EDC is the default and that the FPL is about anti-cheating in a restricted sense. For example hiding your opponents queen is, I assume, not the kind of cheating the FPL is concerned with.

Apparently, some top chess tournaments do not have tournament rules preventing (all) unjustified dropping of rounds or premature resignation. Cheating allegations aside, maybe chess should learn from go and establish similar tournament rules?

I think the Sinquefield cup had a rule that it was not allowed to drop out after round 4. I heard that in the official live commentary but I have not idea if it is an actual rule or simply a convenient explanation. Magnus dropped out in round 4 and might not have strictly violate any rule. The argument that this violates the spirit of the competition is I think valid but players have a lot of leeway before commissions start sanctioning them. Possibly, Magnus is heading down a road where he will risk sanctions but it seems likely that this will blow over.

I'll have to watch that vide with the "most incriminating evidence" later, maybe I'll buy popcorn

[kvasir]

In https://www.youtube.com/watch?v=jfPzUgzrOcQ&t=702s Yosha Iglésias uses some tool to talk about accuracy / correlation between a game's moves and AI moves and the press mentions some 100% occasions for Niemann. What are those accuracy / correlation numbers? The percentage of identical moves themselves? Some derived alleged likelood of cheating?

There is some chessbase documentation online but it is rudimentary. As far as I can tell the "correlation" being talked about is a ratio of coincidence between the game moves and any move from a large committee of computer programs. That is, it is basically recall but measuring if the game moves match any of the computer moves.

For example, if someone plays a 20 move game every move could easily coincide with moves from any of 10 computer programs if we don't require that it is the same program every time or that the computers are not simply predicting 10 different moves in the more difficult positions.

Besides that the statistical argument doesn't seem to lead anywhere there is a problem with what is being argued from the beginning. Namely that Hans would be cheating in about one-in-ten games, having little regard to winning prize money or the fact that he would be a strong favorite in many of the games anyway. I say "from the beginning" because the argument is presented without looking at the games when he played terribly poorly in light of this statistic.

[ez4u]

I didn't bother to read any of the material so apologies for that. However, did they do the logical next step and check the correlation for the rest of the top 20 players (including Magnus) to see how they would fare?

[RobertJasiek]

Not "they" but some individuals have been digging in some games and found something like the following correlation numbers (which I try to recall from memory, I recall partly wrongly and show you only to give you a very rough idea of how these numbers are):

100% Hans Niemann in a few(?) of his best-correlated games. (I do not say that this would indicate cheating. I am just reporting numbers.)

Various (including high) numbers for Hans Niemann checked by a supposed expert statistician on chess cheating: no evidence for cheating found. (Magnus Carlsen has (subjectively?) interpreted for himself similar but unpublished (internal Chess.com?) numbers for Hans Niemann as clear cheating evidence. Maybe he thinks that each 100% value would be evidence beyond any doubt and regardless of statistical verification? Most of the press jumps on 100% and copies that without knowing what such numbers mean. Most Youtubers and most of their viewers throw around such numbers and have no idea whatsoever on statistics.)

100% in one game and 98% in one game of a proved cheater in these two games.

82% in one game a (Super-(?))GM checks for himself as his subjectively best game ever.

72% Bobby Fischer in a famous 20 games winning series.

70% Magnus Carlsen in some of his top games

69% Garry Kasparov in some of his top games

67~68%(?) Super-GM

65~66%(?) GM

[RobertJasiek]

Such numbers vary. At different times, different engines might be used. Chess.com and Chessbase numbers differ greatly.

[RobertJasiek]

My opinion on a player having relatively many 100% games: most likely, this means a) he has studied with AI a lot and applies his acquired knowledge in AI playing style or b) he cheats. Statistics cannot distinguish these two possibilities if correlations are derived from several engines.