Hikaru & Fabiano on Hans, Magnus, and the Cheating Scandal
VloĆŸit
- Äas pĆidĂĄn 12. 09. 2024
- On September 21, 2022 GMs Cristian Chrilia and Fabiano Caruana published a podcast where Fabi drops some truths about what we know about Magnus Carlsen and Sinquefield, the cheating scandal, Ken Regan and more. Original material at:
đ MERCH âș streamlabs.com...
âïž LEARN CHESS & PLAY WITH ME âș go.chess.com/h...
đ GIVE đ CHESS âș www.chess.com/...
đŹ CLIPS CHANNEL âș www.youtube.co...
đïž MORE GMHIKARU âș www.youtube.co...
đ TWITCH âș / gmhikaru
đ INSTAGRAM âș / gmhikaru
đŠ TWITTER âș / gmhikaru
âš TIKTOK âș / hikarugm
đ DISCORD âș / discord
đ FACEBOOK âș / gmhikaru
đ SUPPORT âș streamlabs.com...
đ€ SUPPORT MY ORG Misfits Gaming âș us-shop.misfit...
đ€Ł REDDIT âș / hikarunakamura
âââââââââââââ
đš Thumbnail âș JojoChessNoob / jojochessnoob
đChannel Management âș Team Hikaru
đ§ Business inquiries âș TeamHikaru@WMEAgency.com OR GMHikaruBiz@yahoo.com
#gmhikaru #chess #chessdrama
Hearing Fabiano talking is like listening to a very good professor give a speech. Didn't realize that before.
Yup. He's very organized in his thoughts and it's very easy to understand him.
Yeah he was really good for some of the small commentary he gave during the world championship
@@ghosthunter0950 Exactly I thought
Terrible to watch tho, why even put a camera in this guy's face? he's absolutely terrified to look into it.
I agree such a great player
He didn't cheat two times. He got caught two times.
which means that he cheated two times. No one knows, but I get what you are trying to say.
nice try Hans!
@@polokucoch8112 come with real account hans
@@polokucoch8112 I see your point (he can't conclude he cheated more than twice) but it's not accurate for you to say he cheated two times- from a logical perspective (philosophical argumentation) that does not follow. It would be accurate to say "He cheated AT LEAST" two times, but you cannot conclude how many times he cheated by how much he got caught- it has to remain open-ended.
@@thingsmac innocent until proven guilty. "He cheated AT LEAST" two times is not accurate as "WE KNOW IT"
However, Hans cheated two times is FACT. You will not be incorrect for saying that, but you could be incorrect for saying "AT LEAST" twice. The rest is just speculation.
Fabi has a lot of respect from his peers and the chess community.
Man, Iâm glad Fabiano is having a podcast. To my surprise, he is a great commentator.
Also in the recaps after the match he is good
Why is it a surprise?
@@Jeffrey_Tyler he looks very shy, speaks too calmly, he has not have a lot of dynamism in his voice, but still he is great.
@@MagnusAnand he's more of a reserved than shy I think
@@MagnusAnand are you from the US? this is not meant offensive just wondering xD
So Fabiano was absolutely sure this person was cheating and yet Regan didn't detect it... This is the exact situation Magnus is in.
Reagan system won't detect a 2600+ GM cheating, because as these GM pointed out at that level you only need 2 or 3 moves to cheat on and that it, his system can't detect those 2 or 3 moves as unnatural statistically but a GM can clearly see something is not right.
Yes. The difference is that Magnus had the courage to stand by his belief, even if he has handled it in a kinda clumsy way
Regan claims that he can spot the anomaly if there are 3 sketchy moves.
@@NotQuiteFirst yes a lot of courage while not reporting Niemann or showing any shred of evidence. It takes a lot of courage to resign and make a lot of non statements.
@@max1311 Get back to licking those beads
When Fabianos playing career is over, he will be the next Yasser/Leko as a commentator. How he speaks, how fast his mind works and how well he keeps everything together is very rare. I think he demonstrated perfectly in Magnus-Nepo WC match how he can be analytical when need be and switch to more humorous and not-so-serious talk without any problems.
He was ranked second in the world and did a great job against Magnus in the World Championship Match. He's gonna be a lot more than "the next Yasser/Leko". He'll be able to provide insights they never could
@@sCraNDoMxD tbf the commentator role is far less demanding than actually playing the games. Yasser and Svidler are perfectly fine for instance.
I'm pretty sure Yasser could have been world champion at one point. I'm not 100% sure bit I think he was ranking like #2 in the world in tournaments consistently at his peak. He's not a push over at all although alot of people view him that way.
@@sCraNDoMxD (Not trying to downplay Fabi at all here) but remember Peter was also a WCC challenger and got really close to beat Big Vlad at his peak. So no small feat there. I think we are extremely lucky to get commentary from guys who were/are at the very top of the game - It ain't something you usually get at other sports.
You could even add Judit Polgar to this list of commentators. At the candidates this year on Chess24 she even analized Anish Giri out! đ
These things would never come to light if Magnus had not done things the way he did. Might forever hurt his reputation, but may help us go toward fixing a serious issue in chess.
For real. Personally I have realised I did not properly appreciate Magnus' actions. He has put his whole repuation/legacy on the line for the good of chess - whether he's right or wrong.
Nah it does not hurt his reputation in any way
@@williamrobert9898 dont talk simply without knowing consequences
Haha, hurt Magnus's reputation? I doubt it
@@renishisrael9520 you should take your advise quite seriously
When Fabi speaks, I listen. He comes across so wise and intelligent. One of my favorite people in chess. I am always rooting for him.
Agreed. Unlike Hikaru who comes across like an a..hole.
Fabi is definitely a great standout in the chess world. But I think actually all the top GMs are like this to one degree or another. It's so refreshing to follow a sport/game where all the top players are so articulate and analytical in their comments. As opposed to like, soccer, where (and not to denigrate any soccer players: it's fundamentally a different game) the top players are all basically like "yeah we did well we scored more goals" or "yeah we didn't do well we scored less goals" in every interview.
When Fabiano speaks, I listen
When Fabiano thinks, I applaude
When Fabiano posts, I agree
when Hans speaks, I ignore him
When Hans thinks, he doesn't
When Hans posts, I disagree
We are so lucky to have a world champion who is completely clean and can therefore protect the integrity of the game, otherwise chess would end up like cycling. We owe Magnus a whole lot of gratitude.
Cycling? I have not heard of a cheating cyclist in like⊠3 hours đ
Cycling is much cleaner because of Floyd Landis.
Wrong. Time will prove that magnus is the lance armstrong of chess.
Could be (not likely but could be) that Magnus know he himself is cheating, so Hans has to be cheating to beat him.
The one time I was sure someone was smurfing on their account (pretending to be lower rated) was when I was smurfing and lost to a 600 rated player playing insanely good moves
@@wsemenske different circumstances, the most likely kind of cheating a GM would indulge in on open-board would be someone leaking them plans.
Amazing video. I cannot believe Fabi just torched Regan like that. Fabi is always so chill, he must feel REALLY strongly about this.
He is right to do so. People look to Regan as an answer to a problem he hasn''t solved completely and as Fabi said, if a strong player is motivated to cheat it will be very difficult to catch them. Especially if they know how the anti-cheat systems work or have someone instructing them who does.
It's warranted. Regan might be a great statistician, but he doesn't know the nuances of high-level chess. At best, his model can predict general trends and catch clear outliers, but it will never have granular detail. It's like when physicists try to model stock prices - it's fuzzy at best.
i mean its not black magic, as far as i learn from these hikaru videos, regan calculates conformity with engine decisions at critical positions but if chess players only need info on 1 position that model can never find statistically significant deviation.
@@GBGB000 Regan is an IM. He understands high level chess. However, that said, he probably also has blind spots in his statistical models.
@@juanquntos7123 the difference between a 2400 and a 2750 player's understanding is astronomical
This video raises a very serious problem: FIDE is the only organization that could ban a player for life. And itâs decisions are based on the work of a single statistician, whose methods may well be faulty
@Silent Man Thats simply not true
@Silent Man "oh, objects fall to the ground. It makes it hard to make a building taller than a hut because I can't extrapolate the reasons why objects fall to the ground." I wonder how the pyramids were made.
@@FA-ul4pr haha the burn. Guy thinks buildings were not build before Newton lol. đ
@Silent Man The difference is that all of Newton's theory is and has been public for centuries and countless physicists have tested his theory. Regan's algorithm isn't public and can't be checked by other experts. Obviously there's a very good reason why he can't just make his algorithm public (It gives the information to cheaters on exactly how to dodge it), but it *is* just the work of one man.
@Silent Man yes of course :D there are really high buildings from before gravity was âdiscoveredâ
it seems that Reagan's test focuses on making sure there are no false positives. The drawback from this type of test is that negative results are less reliable. This means that if there is a positive result (the algorithm concludes the player cheated) you can be 99% sure. But if it's negative, it's ambiguous. So, a negative results shouldn't be used as evidence that someone is not cheating
As something close to a statistitian myself, Regan's method will mainly work if the cheating is quite constant. However, if there is a player cheating very rarely in key games, it is tricky to be detected.
Realistically it is probably not only in key games but specifically in key situations of a game.
The problem is that the only evidence people have at the moment is saying âhe is giving bad analysis in interviews, this is less than GM analysisâ. If that were the evidence then it would mean he has to be cheating in every game, otherwise how can he maintain his rating. It necessarily means he is cheating in all his games, if that is something you want to say is suspicious.
@@Gaddo_ to be fair, Hans is clearly a gm+ level player. The question is whether is like 2600ish or 2700+. If you play particularly well against great competition and play 2600 the rest of the time you will improve (probably not to 2800 butâŠ). Also over time his rating would rise and that would increase the amount of moves/games he can cheat in without it being suspicious.
If magnus cheated no one would question it.
@@augustuscaeser5895 tbh the games against magnus, hans played well but not particularly well...it was magnus basically play bad, not at his level at all.
And that could be because of psychology at work, people tend to play way worse against an opponent who they think cheats.
That said, hans game against firouzja (2800) was indeed particularly (but strangely) brilliant
Isn't it more like blatant fluctuations in the level of his play?
44:12 Fabiano: Regan's methods if applied on Hans' games during his cheating years will come back as "not cheating".
Fabiano threw a shade so big that it would reverse global warming.
Fabi 2024?
Hikaru getting leveled by the closed captioning when he said "that are people I know and respect immensely" (talking about the Super GM circle) the caption said "that are people I don't respect immensely".
lol
At an open tournament in Agoura Hills, CA I actually caught my opponent in the bathroom looking at his cellphone with a chess position on his phone using some chess software. I made a complaint to the organizers and they blew me off!! They said "how can you be sure it was your position if you're looking through a bathroom stall, etc.". So even when you can prove it you can't prove it.
love that podcast, respect for supporting you're fellow chess players
Your not you're
@@stevedeall425 true
Hikaru and the other super GMs should play a bunch of games against each other where they agree to consult the engine a couple of times in critical positions and then send them to Ken Regan to see if his analysis can detect the cheating. If it can't detect this sort of cheating then his analysis is largely useless when it comes to cheat detection at the highest level. I don't have any confidence that engine use will be detectable merely from the moves played for super GMs. Even if a super GM doesn't receive any moves, simply knowing the evaluation of a position is massive and how in the world could it ever be possible to detect that a player knows the evaluation just from the moves played?
Heck I agree with Fabi where his methods probably wont even be able to detect Hanâs 12 / 16 yo cheating
Super solid point. This should be highlighted.
Definitely agreed this is a good idea
That one video that just came out I think proved his research to be useless. Hans has a LOT of games at 100% accuracy over 20 moves. One is 48 moves. I think it was Hilary who said that computers donât understand human intuition and top GMs know what a weird move is even if itâs not detected. I believe this is because if you make the strongest move at a clear point in the game it doesnât flag it just because the general accuracy is around normal even if it shifts the tide of the game
I have no doubt that Ken Regan is a world class statistician, but I think the problem is, as he's admitted, he doesn't play chess. And so I think his model, which I'm not smart enough to understand, has huge blindsides as a result of that.
He even admitted, that based on his algorithm, the games that Feller got caught cheating in, his algorithm thinks it was cheating if you look at his best 4 games, but not if you look at all 9. And that alone should be a giant flag for using Regan's system
That it doesn't always detect known cheaters is a giant issue
wtfff....
Never seen fabi like this...
this is savage bro
Fabi is so balanced and objective while being firm and credible. I really appreciate CZcams giving a forum to get a peak into these greats and their personalities.
My takeaway from hearing Regan explain how his model works was basically exactly what Fabiano said: someone has to be very blatantly cheating for it to work. It's not exactly useless. It's great for catching people who don't know how to play chess and just copy engine moves all the time, it just isn't that useful for when a master only needs to for cheat one or two moves in some of his games.
well Danya did specify that this one or two moves needed to win apply for top level chess not all GMs you know he said that a regular GM would need maybe 4 to 7 moves and he still can not see how this can be detected and honestly I agree the statistics will show nothing suspicous
@@williamrobert9898 Would you consider using some punctuation? I really can't tell what you're trying to say.
29:24 Hikaru: What?
(Voiceover: It was at this moment that Grandmaster Nakamura realized that unbeknownst to him, he had been excluded from the secret superGM whatsapp group.)
18:02 "Fabiano will never get flagged for saying the same exact thing as you did"
I think the difference is that Fabiano does care.
From a stats point of view, if a player cheats through the game the pattern can be uncovered but if it is limited to a couple of moves where a perfect move replaces an imperfect one in the context of a longer game then it would likely show as being statistically insignificant.
Even harder to measure, and arguably more effective in super gm chess would be imperfect but not losing moves which are rare and difficult to respond to. A 7th best move that drops from .1 to -.1 but makes the opponent find a difficult only move is probably the strongest move possible
Especially among the strongest players in the world
@@augustuscaeser5895 exactly. What many people don't understand is that chess can be a very paradoxical game. Some winning positions are very often an all-in situation. This means you need to find the only winning move and at the same time the only not losing move. But finding that move is so hard that you'd rather have a slight winning position instead.
@@augustuscaeser5895 like hans' g4 in the levon game?
@John Mullane these people clearly don't know stats, sure, but I'm not convinced his model is particularly accurate at the highest level. If a 1200 plays two brilliancies in a game, clearly cheating. But novelties are being played daily, and computer-aided study (not cheating) leads games in inhuman directions sometimes... so unless he did a full experiment involving super GMs to determine, say, preferred structures, transpositions, etc where they themselves do and don't cheat, the science isn't as secure. Tl;dr, I'm sure his work is very accurate, but super GMs are inherently fringe cases
Itâs kinda funny that everyoneâs searching for a way to detect cheating when they already can. Like Fabiâs 100% sure that someone was cheating after looking at their game, and lots of people suspected Hans was cheating. Itâs been detected. Now get a panel of 3-5 people to analyze the games and if they all come back with the same answer, thatâs plenty enough proof. Weâre never going to get physical evidence like we would with athletic doping, and people need to realize that.
FACTS
I think Fabia is saying the statistical anti-cheat system doesn't work for someone who's cheating 1-3 moves a game (i.e. someone playing at a very high level). This makes perfect sense because the signal-to-noise is very low.
I think the same, but on the other hand, just 2 or 3 moves a game would be not enough. If the engine says a crazy move but then you can't develop it correctly you ll find yourself in a much worse situation
@@victornavarrete1572 nah for 2700s like them they will see the idea 90% of the time when they know it's the best move, since they don't need to calculate anything else because they know it's the correct approach. Also 1-3 moves is actually more than what's needed, even one computer move in a pivotal moment of the game basically closes the gap between Super GM and World Champion easily, forget 2 or 3.
@Victor Navarrete the thing is at that level, as the super GMs will confirm, any clue can be enough for them to find the killer blow
@@victornavarrete1572 like when a "2700" makes a crazy move that "must be winning" that evaluates to -3 and cannot explain the lines or analysis?
@@victornavarrete1572 Not for super GM. They don't even need to be told what's the move, sometimes just being told that there is something at certain point is enough.
It sounds like Fabi has almost as much of an issue with Ken Regan as Magnus does with Hans
Hikaru is so loud and Fabi is so quiet. Itâs hard to find a good volume that suits them both!
Yeah let's dislike this video
@@josemanuico5613 đ
true though
I never said I dislike it. Only that the volume is different
What are you trying to say then??
Keep it coming , love this stuff .
when there's moke there's fire
When there's speaking there is chess
There was no smoke
VERY intelligent commentary by Fabiano. He seems a bit boring in interviews sometimes but you realise listening to this that he just likes to think carefully before he speaks.
Hard not to love Fabi.. such a smart individual.
Actually I wouldn't have listened to Fabi for one hour. But having Hikaru curate the interview and react to it makes Fabiano more bearable.
@@da96103 nah i feel disturbed when hikaru speaks in between.
But yeah it's his channel so who will speak if not hikaru.
Such a rare trait and so under valued
Wish they could lower Hikarus' mic while he watches videos. Maybe it's just my phone but I can't here the clips then Hikaru is like, "HEY EVERYBODY!! OKAY!! LET'S CONTINUE!!!"
It's because he turned up their video all the way, but their audio sucks and was too quiet.
In the beginning of the video Hikaru said he had their volume up to 100%.
Yeah I know and being live Hikaru can't do anything about it. I was just wondering or wanting to know if the editors could do it in post/editing because many of the videos lately have the same issue, for me.
â@@MichaelDavis-zf6ntGood suggestion, they should be able to boost it a bit at editing-time, albeit there is a limit, which when crossed will make the audio distorted. However, they should be able to give it a bump before that happens, and then it should be more clear. I have the same issue as you. edit: it's not too bad though, just thought I'd chime in, in light of your comment.
@@Whateverworksism easier to compress louder sources and boost the whole track if it's already close to clipping
@@DuBLD33 Ah interesting! Shows how little I know, but if I understand you right (and please allow me to use layman terms), we take the louder source (Hikaru) and compress it down to the less-loud source (podcast) and then just boost the whole shebang.. That's pretty smart actually..
As time goes on and considering the cheating problem in chess, the figures who will be at a premium will be those that explain their play in great detail publicly and logically demonstrate the thought process of the highest level of chess play.
This is basically proof that Niemann is a cheater. Because I have never see a GM that's anywhere as bad as he is when analyzing their own game. It's very very sus.
This qualitative aspect of cheat detection is greatly overlooked. If I could highlight replies this would be one.
@@ncs9753 Agreed when GMs do not know how to explain a move of their opponent they simply say it and when it comes to their own moves they explain it perfectly and most of the time the engine agrees
They say that being caught cheating doesn't impact the careers of the offender's enough, but then they won't name any names. If the governing bodies actions are insufficient, then the community should start by actually naming the offender's
Apparently they would all get sued then
Am i the only one that imagines the super GMs just chatting on their whats app group and sending memes
Fabi, Dominguez, Levon with some St.Louis guys might have one group
Put a stick to one end of Regan and call him a marshmallow coz fabi is roasting him đđđ
This is my first time hearing Fabiano speak for an extended time. He seems very intelligent
I definitely want to start a magnesium business with Dlugy.
Does anybody remember during the Sinqfield Cup when Hans was interviewed and he gave this long speech he was asked who his current coach was and he didnât want to answer?? I think it was because it was Dlugy and he knew it would raise even more suspicion if he told people that.
Of course man, I thought the same as soon as I heard Magnus
You may also add that journalists asked him that question for this reason
But to be fair, chess players don't give their coach name because it is a very important information cuz we will directly look for your coach games too if we re ever playing you since he will obviously share with you his prefered positions
Wow good catch.
Exactly. That is exactly why he was acting like such a weirdo when he was asked. Stumbling and fumbling for words.
His coach is Garry Kasparov.
You can tell Fabi has a lot of respect for Hikaru's game.
One amusing detail is that given all the suggestions re Max Dlugy, HE was the person interviewed in the chessbase article about famous cheater Borislav Ivanov and he claims Ivanov had the device in his shoe. Wikipedia also has this.
He reportedly also said that if you give 2600 rated player something similar kid of device, it would be near impossible to catch them when they do nuanced cheating.
Shoe is an interesting theory!!!!
32:53 . True to what Fabiano was saying. The statistical method that Ken Regan uses requires a very large numbers of moves that correspond to engine's lines (relatively to the 1-2 moves needed by top GMs to have an advantage) to become detectable (ie., to be 2 standard deviations away from norms).
Statistical method really is useless if the person only needs 1-2 moves from the engine.
or even 3 or 4
Fabi is clearly not a fan of the anti cheat detection
Wrong. But this specific anti cheating detection Fabi has his doubts (which might be true). That's a big difference. Fabi is well aware that it is really difficult to detect cheating in chess on every occasion for sure. That's what chess needed but isn't working right now.
Cheating is the doping of chess. đ±
@@MusikPiratCH I was referring to that specific anti cheat detection.
should have shown the part where fabi says he would be shook to his core if he found out Hikaru had EVER cheated
Wow! Time stamps of the podcast?
Fabi being SAVAGE about the 4 year cheating challenge
Facts but trust me even if Reagen does it and finds out that his algorithm does not see anything suspicious he will not admit it đ€Ł
Fabi is such a good analyst, such a high-level, academic and rational approach to every subject
Nobody can disagree with fabi
Even if they do
In reference to Fabiano stating that he is aware of a player who certainly cheated and was not caught by a statistical model, it is important for people to understand that no hypothesis test in statistics yields certain results.
If Kenneth Reagan's Null hypothesis was that this person did not cheat and his test led him to reject the null hypothesis whole the person was indeed cheating, this would be a type I error.
Please look further into statistical hypothesis testing and type 1 and type 2 errors for more information.
Tell us further about type 1 and type 2 error
Well Mr Solomon, Your statement doesn't mean anything. "Certain" results don't have to be binary yes/no one or zero answers.
Statistical models can be completely bogus, not have enough working data or be measuring and comparing irrelevant data that outweighs the relevant data.
It is very annoying when people like you say a whole lot of wordy nothing.
Well Mr. Dumbo
I'm sorry you didn't understand. Certainty is in fact a binary state of having no possible doubt. As far as the discipline of statistics is concerned, type I and type II errors happen on occasion. If you're concern is related to "bogus" statistical models, that would be an issue with the particular party responsible for a specific study.
Hope this helps!
@@CapeBone Certainty isn't a binary state unless you are a rigid autistic idiot lol.
DNA tests never give 100% matches even with the same individual but they are still certain results when within a certain percentage range.
Nobody would ever be able to use statistics for anything if they weren't capable of giving certain results in certain situations.
probabilistic mapping techniques such as electron microscopes are incapable of giving 100% identical images due to the random nature of electron and atom movements. Yet we still obtain certain results.
Everybody understands you Solomon, you just aren't saying anything intelligent.
Anyone can regurgitate crap they read from a textbook.
From the regan interview theres a couple things he mentioned about his model.
1) weighted distribution. We are testing for cheating in 1 move therefore using a weighted model is wrong ( even if there is to be cheating in approx 30% of the moves, his model will not pick this up) if the data input is erroneous the result would be so as well.
2) he mentioned 2 unit points. I heard in ur other video this refers to sigma. 1.96 sigma is the 5% confidence interval for a 2 tail test. Cheating should be using 1 tailed test, 1.64 sigma. To be caught using these normal distribution tests would be almost a guarantee of cheating beyond reasonable doubt (95% confidence as described in the test name). This is a really high bar, when accompanied with point 1) i am certain fabiano is correct that this model will not catch hans 2016.
3) his model uses correlation. Once again if there is no cheating in the majority of the moves the correlation is bound to be extremely low and any anomaly of 3/4/5 datapoints with perfect correlation to the engine will not raise any significance to the tests.
4) using correlation and problem with sampling. Im not sure where did regan gets his data for testing. But if he is basing off hans past data this will be a problem as the data would have included his "cheating" datapoint.
5) in summary, the problem is too complex to resolve without additional inputs from supergms. We could modify this model by having supergms to sieve through hans database to pull out the "interesting" moves and feed it to regans model. Of course the supergms has to be unbiased and cannot check these moves against engine beforehand and should rate "interest level" on instinct and not on calculations. Of course this still isnt foolproof as we can still bamboozle the system by playing interesting moves that dont actually work(esp in events that are less consequential). This might give a high sigma but can be logically explained that he "yolo" his moves.
the thing is Fabi already said that when they asked him to give his input about that specific game he mentioned he told them that this player is definitely a cheater but then they ignored his input just because Reagan exonerated the dude
You should have a link to the c2 podcast in the description lol
4:04 Funny how Fabi's evaluation of Dlugy is the same as Magnus' evaluation of Hikaru
@Melon Husk Like your life even matters. đđđ
@Melon Husk Not as much as you đ„ș
Lol, yeah, I noticed that too. I think that's why Hikaru laughed when he heard it.
People keep saying his moves against magnus weren't that special at a grandmaster lvl, you are all asuming Hans is even truly at a grandmaster lvl. lolol
Confirmed. There is a super-GM discord, invite-only of course where these things are "known" and discussed.đ
Oh yes. You would poopoo in your panties if you saw the logs.
10:48 Pulling things out of my you-know-what đ
Bro watch Nepoâs new podcast. He had a problem as well and actually wanted extra security measures when he heard Hans was added.
I'm pretty sure those in this super GMs circle already know that Hans is a habitual cheater lol.
Can you link or name the original source of this video?
The moment I listened to Ken's explanation of his algorithim, in previous Hikaru video, I explained exactly why his analysis is ridicilous and means nothing. There you have it, Fabiano's statement clearly lights it out again.
Nepo said on his own channel that he talked with the organisers at St Louis - before the event started - about increasing security in light of Hans joining.
Fabi seems like a genuine gentleman
Fabiano should do more stuff. Honestly really good. He skirts the line of informative and engaging very well.
Dlugy was hiding in the ceiling, if you look closely at the board you can see a laser pointer hit the correct square for a split second.
đ€Łđ€Ł
Big if true đ
No, it's not Hans its AI humanoid Robot with Hans face
please insert link to original material in the infobox
Fabi is pure class .intelligent well spoken , world champion material .
ok seriusly, Why don't some of you super GMs get together and play with one player having access to stockfish and see if you can cheat in a way that would be undetectable. Send the games for analysis to Ken and see if his algorithms will spot the cheating. Should be an easy way to test how effective his methods are.
I'm surprised this never has been done, the algorith should have been tested with the best players
Hikaru is so funny - loves this drama so much he has turned into a reaction channel
he loves money
Is it true that GMâs only need to cheat on a few moves to have enough advantage to win? If so that would be pretty difficult to detract.
Yes that is true. Magnus did an interview in Norwegian about it around a year ago and itâs available on CZcams
Ken Regan bears a huge responsibility for whatâs going on
It's like, he positions himself as an expert...and then proceeds to do the most unsophisticated simplistic cheat detection possible.
To be fair to him, I doubt he ever said his system is bulletproof. It is possible that a truly bulletproof system is unachievable but who knows...
I think if the Ken's methods are proven to be not working or not working as much as he claimed, FIDE should be the one to take most of the blame here. Who trusts one guy with the security of an international game, in his off time. However he convinces you with his talk and analysis and position as academic. There is a reason why scientific method uses peer review. You cant just trust someone.
@@sada0101 exactly
To be fair to Regan, it sounds like he developed his method 20 years ago, and hasn't adapted it much since then. So while it detected some high-profile cases, and probably works well at lower ratings with naive and blatant cheating, I assume that it can be duped with just a small amount of sophistication
He is probably not even aware that it's obsolete, because it still works in some cases
Love Fabiano. His take on Reaganâs algo is 100% correct. Algorithm is NOT designed nor sensitive enough to catch cheating on only a few moves. From the ground up it needs to be completely re-thought.
He's not saying Ken Regan's anti-cheat system doesn't work. He's saying it is tuned to avoid "false positives" because accusing someone who didn't cheat based on this system would make it completely broken. Instead it might be missing some instances of cheating (ie. false negatives) which isn't good either, but much better than making false accusations. The question is then how often are there false negatives with this system, and it would be up to the organizers to determine if the frequency of missing a cheater is an acceptable margin or not.
Do you have any knowledge of how often games get analyzed by Ken? I ask because it seems like most of the time GMs have an intuitive feel for BS rumors of cheating and only in the case where people are taking those rumors seriously do you need an analysis done. If that is the case, wouldn't it be better to tune the system towards finding false positives rather than the way it is. If multiple reputable players and organizers have good reason to believe someone is cheating then it doesn't make sense not to analyze their games to the strictest possible standard. Even if doing so is just a initial measure to gather all possible scenarios.
@@vincenzo6412 the problem is also lack of data
@@vincenzo6412 I have no knowledge of how many games Ken has analyzed. I'm also not questioning the validity of his system. The only thing I am discussing is how a security system, in general, is measured. He could analyze 3*10^19 games, thanks Matt Parker, and it wouldn't matter unless he knew what the False Acceptance Rate (FAR) and False Rejection Rate (FRR) at least to some measurable degree. Hope that helps clear up any confusion, thanks for the comment!
Agree 100%. I canât believe Hikaru jumped immediately to thinking that Fabi was saying Reaganâs system is trash.
I don't recall Fabi actually saying that, but to be honest, you'd have to be a moron not to come to that conclusion on your own. If he detects one false positive, how would anyone even know?
You canât necessarily trust âstatisticsâ from Regan because the method doesnât occur frequently enough to pick it up. If a cheater gets âhelpâ TWICE (just 2 moves) during each game, he could not possibly pick that up. There are not enough data points for a meaningful statistical analysis. Magnus, and other Super GMs have spoken about this. It only requires a couple key moments in the game for a GM to dramatically increase his chance of winning.
Magnus said he's gonna spill the beads after the tournament
Haha
The issue is that cheaters can pick 2nd or 3rd best moves if they look more human. The cheaters themselves can use engines and remain undetected.
Ken Regan out here taking hits left and right!
6 months and 2 years for cheating? That's crazy. Someone who's willing to cheat to win shouldn't ever be on equal footing with people who want to compete to win in any competitive game.
Make a cheaters bracket they can play in.
Zero tolerance for cheating IMO.
I don't care how young you are, you need to learn a life lesson early.
30:00 any got any clue who Fabi is referring to?
Hans' silence speaks for itself.
What does it say?
@@jordanneal8979 I lied.
People don't realize how good Fabiano is
There was a chance he could have beat the shit out of Magnus and took the crown. He's the perfect guy to have as a commentator. And he's very down to Earth, approachable, light-hearted and playful, you can interact with him.
Nobody is talking about the real drama here. Fabi breaking his promise to Eric Rosen.
it was so disrespectful smh
Yeah canât believe Hikaru is just letting Fabi get this publicity after that despicable betrayal. Fabi is a monster for what he did
@@huckthatdish what happened
@@carsonteuscher2519 Fabi promised Eric Rosen he would play the Stafford in titled Tuesday and then didnât. Breaking a promise to Eric Rosen is the worst crime I can think of
@@huckthatdish Eric Rosen is precious and must be protected at all costs
Good he says that Hans has shown a side of his character by cheating in the past, and that thera are players who would never cheat on principle. Some people keep making the argument that everyone's done worse things, even to the moronic degree of claiming that any other top GM has certainly cheated in the past. And it's definitely not the case, I don't even know where those people come from.
17:48 From the clips I've seen, it seems that players often discuss the game once finished. So my guess is the talks might raise suspicion if player explains moves like Hans did in the interview.
Well that is a fair way of raising suspicion in my opinion because titled players in general not just the GMs easily explain their moves when you ask them
Reganâs method is probably outdated. He was well known for his algorithm in the early 2000âs. Computers are far more advanced today.
I'm willing to bet a million dollar that this Niemann is a cheater, lol. Anyone who understand chess can see how suspicious it is, this guy can't analyze games like a super GM yet plays like one. That is super fucking sus.
Who was Hans coach that Magnus referred to?
I heard his coach caught a cheater 10 yrs ago, with a device in his opponents shoe. And his coach stated. "If I were to cheat, I guarantee you I wouldn't get caught like this guy."
Dlugy
If you listen to Reagan descibing his work on cheating in chess and have a background in math or statistics and actually understand what he is describing you hear a lot of holes... I Aggree with Fabiano
Just to be clear, It's Reagans algorithm that's imperfect. Someone less full of himself could surely construct one that does work.
Actually heard the Ken Regan podcast before the CSquare Podcast, and at that time I already believed his methodologies are flawed. As someone who have worked with machine learn WITH a data analysis guy with multiple degrees, we can only concluded that his methods could not or highly unlikely to detect one to two moves cheating in the highest level. Reason being his methods excels in avoiding false positive, so it means that if his methods detect cheating they are dead lost, but him saying someone that has not cheats means very little.
I thought his techniques were very sound but pretty obviously not absolution of "1 or 2 move cheating"
The problem is that it's simply not possible to detect "1 move cheating" using statistics and if we're honest with ourselves it's not possible using any kind of game analysis.
Given that there's no physical evidence of OTB cheating anyone accusing or suspecting Niemann of this type of "smart cheating" may as well admit that there is simply nothing provable or actionable.
I don't think there was anything wrong with Reagans statistics in and of themselves.
I especially thought it was very interesting that when he looked at Niemanns best tournament perfomance and checked the opponents T% live they got an abnormally low score.
@@JKenny44 I understand and agree that in a statistics standpoint that one move cheating is basically impossible to detect. The reason why I think the methodology is flawed is because it seems like to FIDE if Kent Reganâs method claims someone is clean, they will be considered as clean. While thats just not always 100% true.
Like I said if Kent Regan say you are a cheater, you are dead lost but claiming that someone is not a cheater definitively is just flawed. đ€
On a separate note, I think psychologically its just hard to play a cheater, even if hes clean afterwards. Speaking from personal experience, not in chess but in another sport the second time playing them knowing they done sth cheeky is just hard
Fabi's hairline looks like it was drawn on his forehead with a sharpie
Regan doesn't understand chess well enough to apply a statistical analysis.
34:28 Skull & Pawns. Theyâll fork you over.
Hikaru: I don't know who he's referring to and what tournament
Chat: Hikaru is sweating đ
I like the respect Hikaru shows for fabi
Fabi's point about the person cheating but not getting caught in an "important event" is an important reminder that the statistical analysis and even the algorithms aren't 100% accurate. It is difficult to say how they aren't because then you would theoretically be able to make them more robust. However, it is possible that loopholes exist and cheaters are exploiting those loopholes to a certain degree.
The algorithm is likely tuned to give high false negatives (cheaters get away) in order to reduce false positives (innocent person branded a cheater). This is an unfortunate reality of statistical modeling where reducing one type of error will increase the likelihood of the other.
The permanent problem to combat cheating is that the people trying to cheat are not dummies, and when they learn how cheating is being detected, they specifically can avoid the statistical mistakes. "Don't play tooooo well", etc.
Fabi looks like he could be Levyâs brother from another mother
15:59 "we played butt house" -- of course you did đ€Ł
@GMHikaru I'd love to hear your take on the Punin analysis. It seems rather damning for Niemann, but I'm a novice. It would be super interesting to hear a GM's thoughts of his findings.
They seemed to give him credits
@@ADliver4430 Yeah, but i'd love an in debt analysis by a GM about his findings and if his methodology holds water;-)
Have you got a link to that
I really enjoyed ur commentary on this case
Fabi has gone up in my books...
Levon after playing Hans âHe played weird moves.â đ€
I'd like to point out for a lot of people who are less certain about what cheating means in chess: cheating is not limited to using an engine. Any outside assistance is also cheating. I'm a judge for the trading card game Flesh and Blood. Within our card game, there are very specific guidelines on what is considered outside assistance and it is not allowed.
The same goes for chess. Using an engine is a way to cheat, but even your friend telling you over your shoulder to move a piece is still cheating. It may not matter as much if you're just learning, or it's just a casual game with no stakes.
@GMHikaru you have to watch Yosha - Echecs video on the statistical analysis of Hans' games. He has had a bunch of games OTB that had an engine correlation of 100.0%. Including a game where as black he buries his knight on H2 so that 15 moves later the white king is trapped for awhile at the end game, a move that it seems unlikely anyone would think of at his level or any level really.
His games vs. Magnus did not have 100% correlation, but Yosha makes the point that doesn't mean anything.