Carlsen withdrawal after loss to Niemann

[swami]

GM Firouzja Alireja was also banned by Chess.com
Should they not also make a public statement asking Alireza to defend why he shouldn't have been banned. That's just crazy. He has better things to do, such as winning top tournaments like St.Louis, one at a time.

[acepoint_de]

Actually the whole discussion here and elsewhere on the internet is more and more strong evidence for putting a higher, independent authority (an expert panel) in place, which develops «state of the art» rules and regulations for investigating, detecting and treating cheating cases on- and offline.

Ciao

acepoint

[DrCliche]

OK, let’s do the same for a pool of players of about the same age who play about as frequently as Hans Niemann. I’m willing to bet we’ll see more “damning evidence.” So far you only have an incomplete, Mr. Cliche. No Dr. title for you.

I would welcome this sort of analysis, and wholeheartedly encourage you to do it. As I previously mentioned, I did perform a cursory check to see if I could identify any of Niemann's peers in those tournaments that seemed to display a similar pattern of performance, but due to time constraints I didn't do a formal analysis of any other player. I also hypothesized that identification of other players with such a huge correlation between broadcast status and performance would most likely mean that those players were also suspicious, not that Niemann wasn't.

Obviously, implicit in this view is the assumption that almost all players should and do have little to no correlation between broadcast status and performance. I believe this is a reasonable default position to take, since it's difficult to imagine a mechanism—other than cheating—that could account for a large correlation. But, as I already mentioned, I only performed a cursory examination of other players.

I hope there's a person or team that has the time and resources to commit to gathering much more extensive (and unquestionably accurate) data, and uses it to perform a more thorough analysis. It won't be me, though. Gathering and curating data is incredibly time consuming and tedious, and I spent well longer than I wanted to duplicating and verifying the original tweet's Niemann dataset. (And, as I already mentioned, I wasn't able to definitively confirm the broadcast status of all the tournaments.)

That's surely why, despite a fair number of disputatious souls on these forums—all of whom are happy to try to poke conceptual holes in the analysis of others—none of them can be bothered to do the slightest bit of leg work in furtherance of their arguments. Well, I did as much work as I had time for, and in good faith. From the outset I was entirely open about the limitations of both my analysis and the dataset I was using for it. If you believe my data was incorrect, or my analysis was flawed, or that the pattern of performance exhibited by Niemann is typical of his peers, I invite you to show it.

Believe it or not, I'm not out to get Niemann. (I didn't know who he was prior to this, and my initial reaction was to think that Carlsen was being a sore loser.) I follow chess from time to time, and the drama was interesting. I was curious about some things that I read and dug a little deeper. That led me to various conclusions, but I'd be happy to see them supported or refuted by more thorough research. You and others blithely claiming, "I'm pretty sure if you gathered more data this is what it would show," however, doesn't constitute more thorough research.

I’m sure everyone is familiar with the classic mistake of confusing correlation and causality when it comes to school boy shoe size and IQ. All the small sample size issues aside, Mr. Cliche has lots of work to do to demonstrate causality.

Though not the end all be all, a high degree of statistical significance (e.g. a regression coefficient with a very low p-value) can often be taken as an indication that your sample size is sufficient. As for causality, if broadcast status isn't correlated with the error term in the regression, and nobody can even suggest an alternate mechanism to explain a large correlation between broadcast status and performance, I'm not really sure what else you could want, especially given the limited scope of my data and analysis. (Bringing arcane causal inference methods to bear also seems premature given the questions that remain about the data itself.)

[M ANSARI]

I think if you are an online serial cheater then you are responsible for people doubting you have reformed and are not cheating. The idea that somehow cheating online is OK while cheating OTB is not OK is beyond ridiculous. A lot of the invites to money OTB tournaments come from how well a player does on online tournaments that are followed all over the world. Nieman has cheated in Titled Tuesday ... a tournament that has tens of thousands of viewers live and then probably millions more who go through the games after broadcasting. Dismissing online cheating is just not right. Personally I would never invite any online cheater to a prestigious OTB tournament ... a cheater just doesn't deserve that.

As for Firouzja and even MC being accused of cheating on Chess.com, they banned them temporarily until they could check the games in more detail and after checking they received apologies and were reinstated. That did not happen with Hans and with many other GM's and IM's that were accused of cheating online. The software can sometimes flag a game as suspicious and they get a temporary ban ... I personally think that the software is not to blame but rather an over zealous human operator imposes the temporary ban when a very strong GM complains about a player who just ouplayed him. That happened to Magnus Carlsen in earlier years and also happened to Fiouzja, but that was quickly rectified.

[DrCliche]

Shubham Kumthekar (look him up on Twitter) states that the key YES and NO column, on which Cliche and even Larry Kaufman depend for their accusations against Hans, are incorrectly labelled and the table looks very different if the labelling wasn’t false.

From the start I stated plainly that I wasn't able to definitively confirm the broadcast status of all of the tournaments in the table, and that questions about this information are the primary weakness of my exploratory analysis. If the original tweeter systematically erred or lied in just the right way, it would have been possible for him to motivate any conclusion. (I saw no evidence of this.)

But I'm also not really sure how to go about gathering more conclusive data. I checked archives of the official tournament websites for mentions of live broadcasts, but I suppose it's possible that lots of live broadcast tournaments habitually fail to mention the fact that they're being broadcast live. I don't have any sense for how likely that may be. Perhaps surveys of the tournament directors, organizers, participants, or trusted observers may be able to shed some light on this, but that's well beyond my limited resources.

Other people have trawled tournament archives on chess.com, lichess, chess24, followchess, and the like, and used that as an indicator of broadcast status. I have no idea whether that's valid. Is it always true that if a tournament is thus recorded, that it was broadcast in real time? I suspect not, but again I have no real sense of it either way. Perhaps logs from those websites could be requested and analyzed, revealing when move data was received, but again, that's well beyond my limited resources.

I certainly don't give credence to tweeted claims without meaningful evidence. (And while the original tweet that prompted my analysis was just a tweeted claim without meaningful evidence, I made a serious attempt to verify its data, and wasn't able to find any verifiably contradictory information.) It's compelling that Shubham Kumthekar, in his capacity as FollowChess Broadcast Operations Manager, "remembers broadcasting many of the events," but even if he's right, does that mean they were broadcast live? There remain lots of irregularities that I don't know how to interpret.

If he broadcast the World Open, why are some of its rounds missing? Maybe the World Open wasn't broadcast after all, and some low-level employee (or volunteer) was just supposed to uploaded a dump of the day's games to the usual websites each night. In such a scenario it's then easy to imagine how entire rounds could be missing from a "broadcast"; the nightly upload was simply neglected a few times. Maybe there are equally reasonable explanations for missing entire rounds from actual live broadcast tournaments. I truly don't know.

Anyway, I did perform a robustness check by changing the broadcast status of the most significant tournaments in the dataset. The suspicious results still hold (though to lesser degrees of statistical significance) even after flipping the two most significant tournaments.

Given the abundance of other, independent (and perhaps less controversial) evidence that points to possible instances of OTB cheating (and unadmitted instances of online cheating), it seems to me that this issue is well worth exploring more fully. Even if the data from the original tweet was erroneous, and my analysis of it meaningless, it remains a strong possibility that analysis of the other suspicious evidence will prove conclusive.

There’s also an allegation (which appears correct) that several tournaments during the period covered were omitted.

This has already been addressed, and is incorrect. The original tweeter correctly excluded rapid and blitz events from his table. Other people, not realizing that the USCF has "dual rated" tournaments for intermediate time controls, have erroneously concluded that any tournament that resulted in an adjustment of Niemann's classical rating must have been a classical tournament. That's simply not true.

USCF rapid tournaments aren't FIDE rated, don't allow norms to be earned, and invariably have small prizes, if they have prizes at all. I suspect that most people agree the incentives to cheat in such tournaments are significantly reduced. Thus, it's well motivated and probably correct to exclude them from the analysis.

[swami]

It's a known fact that many stronger Indian players, beyond certain elo, never play tournaments in India because their rating gets drastically affected. They would play only in the tournaments abroad, mostly European and International invitational events. However, there are few who tried playing events in India and their performance rating is usually 200-300 points lower than their actual rating! Many have gone from 2400 to 2200! And oh such tournaments here used to have NO LIVE broadcasting.

When they play in European events, their performance rating exceeds their own rating and they also score more elo points and further their rating to go beyond peak. Very easy to go from 2300 to 2600. And these tournaments almost always had LIVE broadcasting.

Such players retain inflated rating for so long and remain in limelight, until one day, they had to play an important event that has more youngsters and they will see that their actual performance is miserably shocking.

If average rating of the event is lower, there wouldn't be any LIVE broadcast for obvious reasons, lack of funds, and prospects of doing better and gaining elo in such events is dim.

Peer strength is LOW, so you have more RISK.

If the average rating of the event is higher, there would obviously be LIVE broadcast and chances of doing better and gaining elo is higher.

Peer strength is HIGH, so you have more OPPORTUNITY.

[lkaufman]

It's a known fact that many stronger Indian players, beyond certain elo, never play tournaments in India because their rating gets drastically affected. They would play only in the tournaments abroad, mostly European and International invitational events. However, there are few who tried playing events in India and their performance rating is usually 200-300 points lower than their actual rating! Many have gone from 2400 to 2200! And oh such tournaments here used to have NO LIVE broadcasting.

When they play in European events, their performance rating exceeds their own rating and they also score more elo points and further their rating to go beyond peak. Very easy to go from 2300 to 2600. And these tournaments almost always had LIVE broadcasting.

Such players retain inflated rating for so long and remain in limelight, until one day, they had to play an important event that has more youngsters and they will see that their actual performance is miserably shocking.

If average rating of the event is lower, there wouldn't be any LIVE broadcast for obvious reasons, lack of funds, and prospects of doing better and gaining elo in such events is dim.

Peer strength is LOW, so you have more RISK.

If the average rating of the event is higher, there would obviously be LIVE broadcast and chances of doing better and gaining elo is higher.

Peer strength is HIGH, so you have more OPPORTUNITY.

This is a valid point. However I don't think it's a major factor in the U.S. the way it apparently is in India. I've played in many local events with lots of young players here (and hence lost lots of Elo points!) that were broadcast. Most events in the U.S., except for a few round robins like the U.S. Championships, have a lot of young players nowadays. This has indeed wreaked havoc with the ratings.

[Chessqueen]

It's a known fact that many stronger Indian players, beyond certain elo, never play tournaments in India because their rating gets drastically affected. They would play only in the tournaments abroad, mostly European and International invitational events. However, there are few who tried playing events in India and their performance rating is usually 200-300 points lower than their actual rating! Many have gone from 2400 to 2200! And oh such tournaments here used to have NO LIVE broadcasting.

When they play in European events, their performance rating exceeds their own rating and they also score more elo points and further their rating to go beyond peak. Very easy to go from 2300 to 2600. And these tournaments almost always had LIVE broadcasting.

Such players retain inflated rating for so long and remain in limelight, until one day, they had to play an important event that has more youngsters and they will see that their actual performance is miserably shocking.

If average rating of the event is lower, there wouldn't be any LIVE broadcast for obvious reasons, lack of funds, and prospects of doing better and gaining elo in such events is dim.

Peer strength is LOW, so you have more RISK.

If the average rating of the event is higher, there would obviously be LIVE broadcast and chances of doing better and gaining elo is higher.

Peer strength is HIGH, so you have more OPPORTUNITY.

This is a valid point. However I don't think it's a major factor in the U.S. the way it apparently is in India. I've played in many local events with lots of young players here (and hence lost lots of Elo points!) that were broadcast. Most events in the U.S., except for a few round robins like the U.S. Championships, have a lot of young players nowadays. This has indeed wreaked havoc with the ratings.

Is this a good solution to cheating or there is a better solution ? https://en.chessbase.com/post/revisited ... cast-delay

[Damir]

Actually the whole discussion here and elsewhere on the internet is more and more strong evidence for putting a higher, independent authority (an expert panel) in place, which develops «state of the art» rules and regulations for investigating, detecting and treating cheating cases on- and offline.

Ciao

acepoint

These are facts

Hans Niemann 2019-2020 Tournament Results:

USCF performance rating with NO LIVE GAMES:2404

USCF performance rating with LIVE GAMES:2610

USCF Rating Change with NO LIVE GAMES: -112

USCF Rating Change with LIVE GAMES: +131

[Mike Sherwin]

When I was young I played in a simul against an IM (I forget his name). He let us play white. The day before I was studying a game from Tal's 100 best games. It was a Najdorf with quick expansion on the queenside. Tal sacked a knight on b5 for two pawns but failed to win, iirc. Anyway, I analyzed that instead of Tal's Re1 (or was it Rd1). Tal should have played Rf1 and his sac would have worked. I was in total shock the next day when the IM played into that line. I ended up sacking two pieces just like I analyzed the day before and got a quick checkmate. I was the hero that day. Everyone was amazed! I never told anyone that I had just studied that very line the day before. I just let them think that I would be a future world champion! The point is that what Han's said about just happening to study that very line is totally believable. Even if he really did not know why.