Niemann - Carlsen

[Brad Thomson]

This thread is staying fairly respectful of others, this is good. :)

I confess that the "evidence" tends toward the possibility that Niemann may have cheated over the board, but to me it falls short of any "proof", and we must be considered as innocent until proven guilty. I really do think that chess needs to clean up its act here in a very serious way, as there are a lot of would-be chess parents who will say no to their children over this scandal, and the game will suffer as a result. Sid has suggested some kind of actual implant that is state sponsored, and I do not consider this to be outrageous at all, though I am not convinced that this is the most likely explanation. The only thing we know about all of this is that Carlsen has disgraced himself and tossed at least one game.

[Peter McKillop]

.... The only thing we know about all of this is that Carlsen has disgraced himself ......

I disagree (surprise!!). I would say, Carlsen's methods may not have been perfect but he deserves a lot of credit and respect for standing up for the integrity of the game in such a public way! And let's not forget that Carlsen is putting his money where his mouth is. He would have had a good shot at six-figure prize money in St. Louis if he hadn't decided to make a big statement on cheating.

[Brad Thomson]

Peter, Magnus tossed a game after one move, this does little for the integrity of the game, and is cheating. Two wrongs do not make a right. All of this could have/should have been done quietly, but by airing the dirty laundry in public all of chess has disgraced itself. Yes, chess needs to do all it can to avoid cheating...

[Sid Belzberg]

His name is Ken Regan, not Ken Rogan, which you typed in twice.

As someone who has spent considerable time (using my Game Performance Rating [GPR] methods) on analyzing top-level games and compared moves to moves chosen at high search depth (anywhere from 32 to 50 ply depth) by the latest Stockfish engine, I can say a few things about all this.

100% correlation may not mean as much as you think. For example, the situation in which neither player is playing for a win. At the top levels of chess, many games fit into this category. The players might not admit it, and the game may have gone 40 or more moves, giving the appearance both players were trying for an edge, but the reality is they were not.

In this situation, it can be easy to achieve very high correlation because a game can have many plies where there is only 1 or 2 or 3 solid moves, and the other moves that are not outright bad are at least risky and would only be played by someone trying to create complexities (and thus trying to win by forcing a mistake out of their opponent, at the risk of the opponent not making a mistake and achieving an advantageous position).

So in not playing for a win, both players simply choose these solid moves, ply after ply, and the game results in a colorless draw. Very high correlation to an engine like Stockfish is then likely, as I have seen firsthand. I have had to disqualify some games from GPR rating altogether.

This to me is as big a problem at the top levels of chess as cheating. And if cheating is not really happening (over the board), which i still consider possible with all the top 100 or so players, then this problem is bigger than cheating and threatens the long-term viability of top-level chess.

What is possible is that Neimann could be cheating on only specific moves, key moves that can dramatically change a game's direction. So let's say for 30 moves, each player is playing the solid moves and each player knows this and can sense that the game is going to be an agreed draw. Then suddenly Hans, IF he is cheating, decides (at a moment when a solid move would rank 2nd to a really strong move) to engage his cheating method and find the really strong move, and plays it.

I am not claiming this is happening. But I will say that in a court of law, DNA matching is not absolutely 100% certain, but is strong enough to justify a verdict based on that matching. Therefore there must be a number of correlations at which Hans COULD be considered guilty of cheating, and similarly a number at which he could be considered innocent of cheating.

I think 6 consecutive tournaments is a large enough sample size... BUT if there was indeed 100% correlation in ALL the games of the 6 tournaments, we would have to disqualify the games which could be considered a colorless draw, i.e. an entirely risk-free game from BOTH players.

I don't know if Ken Regan considers this in his methods, but I would bet he does

Here is another way of analyzing. The difference between the best move on stockfish and what a player plays is called "centipawn" loss. A 2300 player will typically show an average centipawn loss of 40. A 2300 player that improvers his game and ascends to 2500 will drive down his average centipawn loss to around 22.

Here is a program that you can get on GitHub that allows you to scrape all games of any player has ever played and run the centipawn analysis on all their games.
https://github.com/rafaelvleite/fide_crawler

Every titled player that has ascended from master to Grandmaster has the same decreasing average centipawn score as described above. Now, when we take a look at Hans Niemann's game and analyze all the games he played from when he was 2300 up to his ascension of 2700 the average centipawn score never changes. He has a unique pattern compared to all other ascending Gm rated players in the world.
I hate to say it, but I personally have no doubt he is cheating on the basis of this analysis.

I sent this post to Ken Regan, who I happen to be friends with on facebook, and asked for his opinion.

[Fred Harvey]

This thread is staying fairly respectful of others, this is good. :)

I confess that the "evidence" tends toward the possibility that Niemann may have cheated over the board, but to me it falls short of any "proof", and we must be considered as innocent until proven guilty. I really do think that chess needs to clean up its act here in a very serious way, as there are a lot of would-be chess parents who will say no to their children over this scandal, and the game will suffer as a result. Sid has suggested some kind of actual implant that is state sponsored, and I do not consider this to be outrageous at all, though I am not convinced that this is the most likely explanation. The only thing we know about all of this is that Carlsen has disgraced himself and tossed at least one game.

I still do not understand your attitude to Carlson. To me, he has reacted quite predictably for a guy who feels he is dealing with a cheater, whether right or wrong, and your attempt to introduce his mental state as the evidence mounted was a tad offensive...yes?

[Sid Belzberg]

Here is another way of analyzing. The difference between the best move on stockfish and what a player plays is called "centipawn" loss. A 2300 player will typically show an average centipawn loss of 40. A 2300 player that improvers his game and ascends to 2500 will drive down his average centipawn loss to around 22.

Here a program that you can get on github that allows you to scrape all games of any player has ever played and run the centipawn analysis on all their games.
https://github.com/rafaelvleite/fide_crawler

Every titled player that has ascended from master to Grandmaster has the same decreasing average centipawn score as described above. Now, when we take a look at Hans Niemann's game and analyze all the games he played from when he was 2300 up to his ascension of 2700 the average centipawn score never changes. He has a unique pattern compared to all other ascending Gm rated players in the world.
I hate to say it, but I personally have no doubt he is cheating on the basis of this analysis.

I sent this post to Ken Regan, who I happen to be friends with on facebook, and asked for his opinion.

Interesting correspondence going on between myself and Ken Regan, will update later.

[David Ottosen]

I confess that the "evidence" tends toward the possibility that Niemann may have cheated over the board, but to me it falls short of any "proof", and we must be considered as innocent until proven guilty.

We only require to be considered innocent until proven guilty in the eyes of the law, which Niemann is. He hasn't been expelled from any tournaments, had any prize money seized, or any other form of action taken against him.

Carlsen is free to believe what he wants and take whatever decisions he wants, even if his justification is nothing more than "I don't like that guy's face." and he's not required to justify it to any of us. Just like you're not required to justify your opinion that Carlsen is mentally ill, no matter how unfair it is.

[Brad Thomson]

I still do not understand your attitude to Carlson. To me, he has reacted quite predictably for a guy who feels he is dealing with a cheater, whether right or wrong, and your attempt to introduce his mental state as the evidence mounted was a tad offensive...yes?

Thank you, Fred, for your comments. My opinion is that Carlsen KNEW (or damn well should have) what was bound to happen to Niemann (he would be tossed under the bus and a big stink would be raised with respect his possible cheating) when he, Carlsen, decided to withdraw from this prestigious event. Carlsen has expressed having felt reservations about this event before the tournament even started. Magnus goes into the tournament with an unusual pre-game mind set, and promptly loses to the dude?! (I say he tossed this game too, to have a pretext for his coming forth with allegations). Then the man withdraws altogether and proceeds next to tossing an online game and now all hell done broke lose. And Magnus absolutely knew what was going to happen. What good can this or he possibly do for chess? It is a brutal public spectacle and a severe black eye on the game. In my opinion, Carlsen is behaving very erratically. What should he have done? I do not know, I cannot say. But his moves are not/were not forced. To repeat, Magnus "vaguely" intimated cheating without any evidence and then quit a big event and then tossed a game online. He knows what he is doing. I think it is very bad for them game.

Now, because I am no expert I am wrong to opine on the man's state of mental health, agreed. And I apologize if I was offensive. But simply as a fellow, brother human I cannot help but feel sympathy and concern for this person, also a rather deep sense of pity. I am worried about a man who as world chess champion announces his intention to step down first from being world champion, then withdraws from a big event in Saint Louis in such circumstances as he did, then tosses a game and precipitates all of this mess without any reasonable proof/evidence.
If they demonstrate the method of cheating, then guilt shall have to be pronounced. This means detecting the technology being used. No theories about his game-scores in comparison with computer analysis is evidentiary. This trial in public is very sad, chess is being tarnished, laughed at, it is losing its respect/status due to all of this. Our game is losing power.

Finally, thank you again Fred, and I am sorry about calling into question Magnus' thought processes.

[Fred Harvey]

Thank you, Fred, for your comments. My opinion is that Carlsen KNEW (or damn well should have) what was bound to happen to Niemann (he would be tossed under the bus and a big stink would be raised with respect his possible cheating) when he, Carlsen, decided to withdraw from this prestigious event. Carlsen has expressed having felt reservations about this event before the tournament even started. Magnus goes into the tournament with an unusual pre-game mind set, and promptly loses to the dude?! (I say he tossed this game too, to have a pretext for his coming forth with allegations). Then the man withdraws altogether and proceeds next to tossing an online game and now all hell done broke lose. And Magnus absolutely knew what was going to happen. What good can this or he possibly do for chess? It is a brutal public spectacle and a severe black eye on the game. In my opinion, Carlsen is behaving very erratically. What should he have done? I do not know, I cannot say. But his moves are not/were not forced. To repeat, Magnus "vaguely" intimated cheating without any evidence and then quit a big event and then tossed a game online. He knows what he is doing. I think it is very bad for them game.

Now, because I am no expert I am wrong to opine on the man's state of mental health, agreed. And I apologize if I was offensive. But simply as a fellow, brother human I cannot help but feel sympathy and concern for this person, also a rather deep sense of pity. I am worried about a man who as world chess champion announces his intention to step down first from being world champion, then withdraws from a big event in Saint Louis in such circumstances as he did, then tosses a game and precipitates all of this mess without any reasonable proof/evidence.
If they demonstrate the method of cheating, then guilt shall have to be pronounced. This means detecting the technology being used. No theories about his game-scores in comparison with computer analysis is evidentiary. This trial in public is very sad, chess is being tarnished, laughed at, it is losing its respect/status due to all of this. Our game is losing power.

Finally, thank you again Fred, and I am sorry about calling into question Magnus' thought processes.

Well thank you for demonstrating that it is possible to have a civilized discussion on here without the shrill nonsense so often encountered! I still believe that Carlson felt strongly that he was dealing with a cheater, and faced a very difficult decision. Perhaps the way he handled it was not your preferred approach, but it was the action of a normal person, and I suspect he will be totally vindicated in time.

[Brad Thomson]

We only require to be considered innocent until proven guilty in the eyes of the law, which Niemann is. He hasn't been expelled from any tournaments, had any prize money seized, or any other form of action taken against him.

Carlsen is free to believe what he wants and take whatever decisions he wants, even if his justification is nothing more than "I don't like that guy's face." and he's not required to justify it to any of us. Just like you're not required to justify your opinion that Carlsen is mentally ill, no matter how unfair it is.

Everything you say here is self-evidently true, there is no room for rational disagreement.

I do think that social norms should probably generally require a bigger gripe than the look of another person's face before launching into such a public display of allegations of cheating at this highest level of the game. Magnus has the freedom/right/power to do as he is, yes of course. But what the hell good is he doing, doing it like this? We have a tantrum taking place before our eyes. I think Magnus' judgment is very bad and his pattern of conduct questionable. It is not at all good for chess, we are getting very bad publicity, it will affect prospective chess parents and deflect them and their children from the game.

[Neil Frarey]

It's going to lead to AI determinations ... solely AI's 'judgement ... whether or not a player has cheated.

So for instance if I were to enter a OTB tournament in the U1800 section and sweep it (naturally, 'cause I'm still raked in the mid 1400's) I most likely would be accused of cheating especially if my average centipawn loss was around 40, ha!

Banned for life, ha!

[Sid Belzberg]

This means detecting the technology being used. No theories about his game-scores in comparison with computer analysis is evidentiary.

According to that logic, we may as well throw out entire scientific branches that rely on indirect analysis and probabilistic theory. This would include particle physics, Quantum mechanics, Molecular Genetics, etc.

In a court of law, "beyond a reasonable doubt" is the standard of evidence. If the weapon used to murder someone is not found, but the alleged assailant's blood has a DNA match at the scene of the crime, it is enough to convict or, in many cases has, exonerated the accused when other blood is found at the scene of the crime. Either way, DNA matches are not 100 percent. However, the number is so closer to 100 percent that it is valid evidence to convict someone.

The same is true here. I doubt if we will ever know what technology was used or how it works, but the evidence, in my opinion, is now overwhelming that Hans Nieman has routinely cheated since 2018, when his ascent from 2300-2700 began. Every other GM has a steadily decreasing average centipawn score as they progress from Master to Grandmaster. But not Hans Nieman, who simply has a flat line all the way through. This indicates that he used a computer aide for a long time.

[Roger Patterson]

.....
In a court of law, "beyond a reasonable doubt" is the standard of evidence. If the weapon used to murder someone is not found, but the alleged assailant's blood has a DNA match at the scene of the crime, it is enough to convict or, in many cases has, exonerated the accused when other blood is found at the scene of the crime. Either way, DNA matches are not 100 percent. However, the number is so closer to 100 percent that it is valid evidence to convict someone.

The same is true here. I doubt if we will ever know what technology was used or how it works, but the evidence, in my opinion, is now overwhelming that ....

US and Canadian law is littered with examples of innocent people being convicted on the basis of faulty correlation analysis and forensics. I remember one case particularly, the suspect was convicted on the basis of testimony that the chemical analysis of the crime bullet matched bullets from a box in his possession so it was his bullet that klled the victim. The US government had a whole department devoted to chemical analysis of bullets and they testified as to the match and that it must be his. That person spent many years in jail before being exonerated and the forensics exposed as being bogus and untrustworthy. There is an ongoing problem in law with suspect forensic pattern matching that is accepted by the courts but cannot withstand double blind testing.

Here we have a bunch of amateur analysts coming to highly motivated conclusions without any thought as to how to fairly test and confirm their analysis. These results must be considered highly suspect.

[Sid Belzberg]

US and Canadian law is littered with examples of innocent people being convicted on the basis of faulty correlation analysis and forensics. I remember one case particularly, the suspect was convicted on the basis of testimony that the chemical analysis of the crime bullet matched bullets from a box in his possession so it was his bullet that klled the victim. The US government had a whole department devoted to chemical analysis of bullets and they testified as to the match and that it must be his. That person spent many years in jail before being exonerated and the forensics exposed as being bogus and untrustworthy. There is an ongoing problem in law with suspect forensic pattern matching that is accepted by the courts but cannot withstand double blind testing.

Here we have a bunch of amateur analysts coming to highly motivated conclusions without any thought as to how to fairly test and confirm their analysis. These results must be considered highly suspect.

Roger, the correlation analysis I specifically referred to is DNA analysis. Ballistics analysis is not a valid comparison. Thanks to the advent of DNA analysis, the innocence project has exonerated over 350 people, most serving life sentences, including 18 that were on death row. https://history.innocenceproject.org/ .

Here we have a bunch of amateur analysts coming to highly motivated conclusions without any thought as to how to fairly test and confirm their analysis

The lead investigator is a professional data scientist and computer programmer. His methodologies and source code are fully disclosed on GitHub and are reproducible. In fact, I invited Ken Regan to do just that today, and he has said he will. https://github.com/rafaelvleite/fide_crawler. https://github.com/rafaelvleite/cent...oss_calculator

Here is a follow-up presentation that compares the ascent of many other GMs, including Magnus Carlsen, Fabio Caruana, and others to Hans Nieman's chart that uniquely shows no correlation showing an inverse relationship between average centipawn score and chess ratings. If anything, it is highly suspect to consider that there is not an issue with Hans Neiman.

https://www.canva.com/design/DAFOAuf...ce=sharebutton

Expanded Medium Article posted by lead investigator and Data Scientist Rafael Vliet today.

https://medium.com/@rafaelvleite82/h...l-bea9485055de

[Pargat Perrer]

Roger, the correlation analysis I specifically referred to is DNA analysis. Ballistics analysis is not a valid comparison. Thanks to the advent of DNA analysis, the innocence project has exonerated over 350 people, most serving life sentences, including 18 that were on death row. https://history.innocenceproject.org/ .



The lead investigator is a professional data scientist and computer programmer. His methodologies and source code are fully disclosed on GitHub and are reproducible. In fact, I invited Ken Regan to do just that today, and he has said he will. https://github.com/rafaelvleite/fide_crawler. https://github.com/rafaelvleite/cent...oss_calculator

Here is a follow-up presentation that compares the ascent of many other GMs, including Magnus Carlsen, Fabio Caruana, and others to Hans Nieman's chart that uniquely shows no correlation showing an inverse relationship between average centipawn score and chess ratings. If anything, it is highly suspect to consider that there is not an issue with Hans Neiman.

https://www.canva.com/design/DAFOAuf...ce=sharebutton

Expanded Medium Article posted by lead investigator and Data Scientist Rafael Vliet today.

https://medium.com/@rafaelvleite82/h...l-bea9485055de

Actually what this entire scandal is exposing is the sheer lunacy of top-level chess.

I watched the youtube video, Sid, and I was struck by one thing: every chess player who strives to be the best chess player s/he can be is ACTUALLY striving to be a mechanical machine. They are striving to reduce their "centipawn loss" compared to Stockfish to as near as zero as possible. They are striving to be ALL ALIKE.

All EQUAL TO STOCKFISH. All striving to be a machine, the machine called Stockfish.

It reduces competitive chess at the top levels to meaningless drivel.

This problem is unique to chess, although a game such as Go could also have the same problem except that Go is much more complex than Chess, so humans are much farther down on the comparison scale to the machine, which in that case would be AlphaGo I think.

I"ve completely lost interest now in my Game Performance Rating work, and in competitive top-level chess itself. As I believe Mr. Crowhurst here in this thread put it, it is all "as boring as dirt".

I do have an interest in John Leslie's Hostage Chess, a much more freewheeling version of chess. But I do fear even there that the ultimate fate is for everyone to be compared to a machine.

Give me hockey instead, with totally unpredictable play and no possible comparison to any machine!

Those talking about the "integrity of the game" might as well be talking about the "integrity of Tic Tak Toe". Absolutely meaningless.