I think one thing is not said enough is that those 48 games are casual games. Imaging being Hikaku, who qualified candidates, biggest chess streamer and so on, and you cheat by putting your reputation on the line so that you can win games against 2 IMs and 1 FM.
hikaru literally doesn't care about casual tournaments like the candidates or the grand swiss; that's just his side hustle anyways. His entire focus as a professional twitch streamer is on the high stakes twitch blitz matches against people 500 elo weaker than him.
If I was Hikaru, I would be more offended about someone insulting my intelligence than insulting my character. Dudes got it made, imagine risking his whole career to cheat in casual games 😂
I think the illustrative point about Kramnik is how much he simultaneously insists that he’s not accusing Hikaru of cheating, and yet baselessly insists that Hikaru’s results are impossible. He insists on investigation, yet rejects evidence presented. There is fundamental dishonesty (or delusion) to that. While it’s worthwhile to refute Kramnik’s inaccuracy on the mathematics and the particulars of online chess (both rating mechanics and actual play) for the broader community, it doesn’t seem productive to respond directly to someone who clearly is not honest, who is not participating in good faith. Whatever his past accomplishments, Kramnik’s current actions are simply not respectable.
11:21 " I dont think he (Naroditsky) ever won a Titled Tuesday." Daniel Naroditsky has won the Titled Tuesday tournament at least three times. He emerged victorious on December 30, 2014, September 1, 2015 and February 22, 2022
Yes it does. There is a significant difference between 0 and 3. Christian needs to correct the record. There's probably only a dozen people or so that have won titled tuesday, and Danya is one of them. He picked a bad example for his point. He also belittled Danya. He absolutely should say something to acknowledge the error.
Fabiano got this point right. The streak that Kramnik identifies as 47.5 out of 48 could be equally viewed as 47.5 out of 50 if someone else includes the games immediately before and after the long streak of wins.
it doesn't matter at all, what matters is that he Hikaru does it regularly 5-10 times a month. If something happens once, it's an extraordinary case, if an «extraordinary» i.e. statistically improbable case happens regularly, it's a reason to check for cheating.
@@kazimirmalevich6712and the statisticians already showed their work explaining to kramnik its not supposed to be that rare for Hikaru and it should happen as often as it already does with no shock factor.
@@kazimirmalevich6712 So, you want to say if Hikaru cheats 5-10 times a month thats normal lol But if he cheats once a year thats suspicious. haha dude hikaru is extremly suspicious.
@@Viewer13128 I'm sorry, but I can't accept "someone out there explained..." It's very simple, if 47.5 out of 48 is a rare extroordinary result with probability P=0.01, then Hikaru's five times in November have a probability P^5=0.0000000001 that's ten zeros! This does not fit into the framework of an extraordinary result. Let Hikaru explain it or let it pass the cheating test.
@@kazimirmalevich6712did you just divide one number by the other, completely disregarding these are individual probability events with a complex model explaining their likelihood? And then pretend it's stats by giving it a p-value?
@@patrickpet7905 Two things can be true. He might wanna defend Hikaru. But he also might not wanna have to be addressing these silly accusations/callouts at 10pm on a Saturday night lol
Fabi really doesnt have that much time for it, being an active top Grandmaster. Mr. Kramnik has all the time in the world now, but I would appreciate seeing his actual evidence. Recently in Kramnik's recent posts it has been simply denials of others that have opposing beliefs and statistics.
Год назад+38
In my local chess club there was an otb blitz match between a 2200 and 2100 fide classical and 2200 won 20-0 the idea that these things dont happen is crazy
I have a question for both Fabi and Christian - when you attend these parties where strong super GMs perform these “party tricks” and play with time odds against other players, are people like Kramnik, Nepo, etc. usually present? And do they also play with time odds against others?
The problem with Kramnik accusing Hikaru is that there is so much real cheating going on, and by accusing Hikaru instead who is obviously not cheating he bastardizes any future claims he may make of people cheating. Its hard to take him serious now, even if he claims to have a team of mathematicians. Also it proves that a statistical analysis alone is not enough, you need to use common sense also
Well, as far as statistics go, it demonstrates a lot better how important it is to look at the whole picture (all of the tens of thousands of games hikaru has played) instead of one small subset
@@Kayrim_Borlan Indeed, if the statistics were done properly, that should have been taken into consideration in the first place. I guess the team of mathematicians he has is quite bad at math
Statistics is not mathematics, first lesson in statistics is to be careful with interpreting numbers, on top of that there are outliers in statistics. You cannot statistically prove anything. Of anything it is drawing approximate conclusions from observations. Economists have been consistenly been mistaken because there are too many variables.
@@kimgysen10 I mean, Statistics are a branch of mathematics. So I don't know what you mean by saying they aren't mathematics. That is just objectively false. Then you jump both feet in with more fallacies. You cannot statistically prove anything. Ok?? Who said that statistics alone prove something? If statistics alone are used to accuse, they should be good enough to defend against the claims. In this scenario, the statistics help to satisfy probability. When applied correctly, they do not indicate any type of abnormalities with his win to loss ratio and don't support the idea that they point to cheating. Then you hit us with that false equivalence, talking about Economists making mistakes due to the variables in the economy. And? What does that have to do with the length of ADs in a RUclips video? Hikaru doesn't have to prove he isn't cheating. There isn't any credible evidence that he is. This is how logic works. Without evidence of him cheating, you can't use statistics to prove he is, unless you are willing to also allow it to prove he isn't. So if we can't clear his name using the proper application of statistics in regard to the probability of cheating. You damn sure can't use the improper application of statistics to make the accusation that he is, and then deny him the ability to clear his name using the method you chose to aim allegations against him with. This is what I hate about the internet, and people who just want to argue and take a position without having any foundation to stand upon. It's usually based on biases, either against Hikaru or for Kramnik(in this specific situation). You wrongly say statistics aren't mathematics, then say that statistics couldn't even prove his innocence anyway. Yet you are completely okay with using them to prove his guilt? That is top tier hypocrisy. The proof to condemn someone should be greater than the proof needed to prove one's innocence, shouldn't it? So we avoid the concept of falsely convicting or condemning those innocent of their allegations? SMH. So disappointing. You should strive to think more critically about things and to be better in life than someone who is wiling to throw out logic to appease their own opinions. Other than that, the statistics in economics have NOTHING to do with the probability of cheating in chess. So please avoid using a lie and fallacies to try to argue against logic, pwetty pwease? You are better than this. I have faith.
I think there's a really useful analogy to use that really reframes the situation and makes it easier to understand why an insane winning streak might not be as strange as it seems on first glance. Consider this : if you roll a set of six dice and you get all 1s, that's incredibly unlikely and if you can do it on demand, you're almost certainly manipulating the dice somehow. However, if you are spending all your time rolling six dice and you do it enough times (let's say a million times, to give an extreme example), the way the probabilities work, the situation flips: it actually becomes extremely unlikely that you DON'T ever get all 1s. If you're standing around rolling your six dice as usual and a random bystander notices that you rolled all 1s, from his perspective he sees this as a remarkably unlikely event. But from your perspective, you know that this is just the product of you spending all your time rolling your dice. To come back to Hikaru and his apparently insane winning streaks, basically the equivalent is that Hikaru spends enormous amounts of time playing chess, since streaming himself playing chess is his main career now. Most of the time his results are (for his ability and his opponents' ability) probably more or less average. But occasionally he gets apparently insanely good results, simply as a product of playing so much (and probably some terrible results as well). And to an observer who doesn't consider all the games he plays (and looks only at the insane winning streaks), it comes off as remarkably unlikely. To be sure, I don't know how the actual probabilities work out when it comes to how much Hikaru plays compared to how unlikely his winning streaks are, but I just think the above is good to keep in mind.
Your example is very good to try and shape a sensible intuition with regards to probabilities. Hopefully it'll reach some people who can learn from it, but I am certain it will not reach some others. Kramnik mentioned the fact many people criticized his interpretation by pointing at cherry picking, and his response in part was verbatim "of course it has to be cherry-picking!" 😂
You could also use coin flips as an example. Do enough and you'll get long streaks of either heads or tails. One way to analyze it is to imagine a million people flipping a coin, then everyone who gets heads flips again, then repeating that until you have a last toss with a head. I'm not sure exactly what the likeliest number of heads in a row is, but it would be up to 19.
Kramnik's "interesting" statistics regarding Hikaru were poorly thought out from the start. He obviously didn't get any of the statistics checked by a Statitician, he clearly didn't do any research into who Hikaru was playing, and he didn't look at any extraneous factors like accuracy or move time. Lastly, the knock out blow for me was this - Hikaru's streak(s) were statistically LIKELY. Kramnik should've owned up to his mistake then, trying to pursue it just made him look silly. Edit: RUclips will not let me mention the chess website who said Hikaru's winning streaks were statistically likely. Everytime I do it removes the comment.
@@saikat93ify The "mathematicians" and "statisticians" online did not show their work in arriving at the beginning parameters. If you use each game's Elo, arrive at probabilities of wins (and draws) and determine the probability of such a streak within thousands of games (or the absurd 30K games) then we can see why there is a question. I wish he would show his work because the "mathematicians" certainly did not. They all skipped the part where draws were likely at about 20% of the time.
@@scottekoontz I have the same problem. I can't use any URLs in comments, ironically not even links to RUclips videos. It's worse than that, though. Just now my reply to you was deleted because I used an example of how even if I leave out the dot from the URL of the site you're talking about, I can't post successfully. I used somedomainname___, where ___ was com. It seems any string of characters that ends with com isn't allowed. What really galls me is I had to report the same spammer posting links in chess video comments that redirected to a porn site multiple times before it stopped.
@@thebcwonder4850 You have a 90.2% chance of win per game, and with a string of 50 games you can expect to win about 45 of them, and at these Elo ranges the expected draws at 3.4% is lower than the expected losses at 6.4%. For higher Elos with a difference of 400, e.g., 2000 vs 1600, the WLD is about 88 / 4 / 8 which indicated that draws become more common. In 50 games your expected longest string of wins is 20. If you wanted a string of 50 wins in a row you would want to play over 1,300 games to have equal probability of achieving such a streak as not.
Chris/Fabi. As a mostly retired 'old' 2400 USCF player, It is very clear, that while cheating does take place (being a reality) at various levels, what Kramnik (and Nepo) really object to is the "Farming"...as they and other sporadic high level players that only play key online events, do not do this. Farming has occurred at all levels for many many years, whether at the very top (3000 plus players farming 2700-2800s), 2800 players farming 2500, 2500 farming 2200, and so on. Players do this to increase their ratings for a variety of reason, and I am speaking now specifically for online platforms. It doesn't matter to Kramnik/Nepo that some GM's do it for content or not, and of course it can equate to $$$, which again, Kramnik and Nepo do not enjoy having, or player Farm for other reasons such as 'status' and respect..I guess. While Cheating is an underlying issue that certainly needs to be constantly addressed and I believe is, to some regard, the Kramnik's and Nepo's (etc) think farming, if not cheating, is unethical. I appreciate and really enjoy your channel, content and boadcasts.
I've been fooled into thinking these guys are on two different locations having split screen call but then realized that's just something on the wall :D
Wtf, I was one hundred per cent sure they were at different locations and had aligned their cameras to make it seem like they were having a conversation side by side.
Now I wanna see Danya adopt Fabi in 1 + 0 online. That would be a hoot! On a more serious note, Fabi should consider hiring Danya as a coach to improve his mouse/ultra-high speed skills.
@@PeteQuadIt is quite contra intuitive and it's not really a thing that comes natural to many. This is not to slam Kramnik but maybe he has some wierd form of math dyslexia that makes this hard to get while being very intelligent in most other aspects?
Maybe because Kramnik has been so good for so long he's forgotten how intimidating it is to play a much stronger opponent. I very much doubt the IMs who play Hikaru are playing at their very best when they are expecting to lose nearly every game.
Kramnik simply does not understand that there is more to chess skills when it comes to online chess - mouse skills, board vision, art of playing second best moves quickly (which Hikaru is the best in the world in my opinion). As such, streaks can be very common for these online specialist players like Hikaru, Alireza, Danya, Nihal etc
@@schizoframia4874 I think it's a factor when you're on your tenth game in a row of getting absolutely obliterated by someone like Hikaru. People go on tilt
There is one obvious explanation for the farmed rating being inflated that you (and Kramnik!) missed: the farmed *also* farm, but against lower rated players.
It should be mentioned in all in these conversations that Hikaru recorded himself doing this run and made a youtube video of it. He literally is explaining his moves in the video. So...... Everyone talking about this has seen the video, and know the whole thing is stupid. Ages later we are still talking about it when a tiny percentage of people think anything actually happened. I guess it's fun to talk about, but it is just content. I doubt there is much left to say on it now.
Yeah hikaru is like the most bullet proof player in the world with this stuff. He has countless otb results backing up his online play, streams most of his success, and vocalizes his thought process through the majority of it
@@tvrskkngdm2333 The thing is, in a lot of these games he simply does not have the time for engine advice. We all know, stockfish and other engines need "time" to calculate. In bullet without increment, there is basically no shot the engine can help him in the end, where he tries to flag the opponent. You can't even argue "oh yeah, he used engine earlier, thats why he has the better position!", because in most games of the streak he was dead lost. He basically just played moves to survive. There is basically no shot he cheated in these games. And like Fabi said, he can utilize his skills in meaningless OTB matches as well, without preparation, without some engine vibing his ass. Fabi encoutered it in real life and backs Hikaru.
@JackHarb89 Besides his speed, there's also the fact that Hikaru has absolutely insane defensive skills, even compared to other super gms. Even from a dead lost position, there are very few people who can turn the disadvantage into a mate, and combined with his speed they'd need to be able to play the best/second best move almost instantly. Someone like Danya, magnus, alireza could most likely do it, but 2 IMs and an Fm? Probably not happening
Fabiano's explanations are so concise yet clear. I listened to Kramnik for 2 hours the other day (couldnt handle the whole 3h interview) and he rambled on for a long time without reaching any real conclusions. IF Kramnik truly believes in this he should be clear and take it further (if he can) or apologise and put the matter to rest once and for all. Thank you Fabiano for sharing your insights with us.
I used to give a demonstration in one of my classes in which I had a student write down a "random" sequence, where they could select triangle, square, or circle. So their sequence would be something like CSTTSTTCCSCT... etc. I would guess the first symbol, they would tell me if I'm right or wrong, and then I'd guess the next and so on. The point of the exercise is I'd always do better than 33% correct - usually pretty close to 50%. The trick is that I get feedback on whether I'm right or wrong, and use that to make my next guess, knowing that a human will rarely repeat when trying to generate a "random" sequence. A true random sequence will have the appearance of, say, 4 circles in a row, but that looks "unrandom" to the eye. We want to invest meaning in any sequence, but sometimes it's just luck. Or rather, luck coupled with the fact that Hikaru should win against these players 95% of the time. I also find his streak quite plausible given the factors Fabiano identified.
Isn’t that the beautiful thing about playing online, you get to play multiple games against really good players. Those IMs and FMs got a lesson and multiple games to analyze. I’d love for Hikaru or Fabi to whip me for an hour, because they’re taking their time to play ME. Great, and FREE training.
@@declupNo, playing a very strong human is much, much better than playing an engine. Humans, obviously, aren't computers, and no matter how much training they do will never be able to calculate as much as one. Because of that, it's extremely difficult for even top level players to understand most of the moves, and next to impossible to replicate their performance. Even Magnus would lose 100% of the time playing against stockfish, and only maybe be able to pick up anything useful. At the top level engines are studied because there isn't much else TO study, but anyone below that is much better off studying human play which is what they're going to be playing against anyway
I wonder what the statistically equivalent losing streak for Hikaru is, and how many times in the 30,000+ game record this has occurred. I suspect the odds of hikaru having a winning streak such as this is maybe equal to him losing 4 games in a row, which I assume he's done a few times in that 30,000 game record.
It has been discussed how the online ratings aren't reliable and also how time controls matter. But even if you get the compensations for that pretty much correct, there are still 2 statistical traps that are easy to fall into. The first and more obvious one is that you can't just look at the streak in isolation. Instead, you have to interpret every game as the start of a new streak attempt. The second, less obvious one is that it's invalid to look at a streak and ask "what were the chances?". You have to define exactly what you are looking for before you look at the data. So what streak length will be looked for and how many draws will be allowed in that streak. If you start trying to estimate chances after the fact you get absurd results, like asking "what where the chances of that leaf falling in exactly that way from the tree(wow, much worse than a trillion to one).
It’s so ridiculous that they have to explain this AGAIN, and in more detail this time for Kramnik to understand! Kramnik looks like a fool at this point
From my experience (and this just adds to it) it's just very difficult to make someone understand something they so wholly don't understand and can’t grasp. Looking at it from the side, I could recognise each time Kramnik failed to adapt to the changes the electronic age brought to chess but putting them together and trying to argue these are the reasons for his lack of... graspness, is difficult. He is obviously a methodical person(ality) who finds it difficult to adjust concepts once those are formed. There seems to be a battle, in his mind, between, shall we call it - "the Botvinnik school of chess" and this computersied era where his view seems to be that studying "move per move" as he puts it, is inferior and so coming from a place of "higher ground" he is trying to preach his truth. Yes, there's some elitism there but that's not the main issue. His refusal to accept practical chess and let go of old ethics are, in part, what's debilitating him… from adjusting. He views flagging as disrespect, gaming-the-system (i.e. not really chess), he feels sorrow for the loss of hierarchical... manners... towards former (older) champions, etc. To a modernist it seems like he is looking for favouratism but from his standpoint dirty-chess (in particular) is simply not chess at all since it’s too practical and fundamentally wrong. Unfortunately, there is a lot that he doesn't understand. In fact, it is so much and involves so many topics it'll be impossible to write it all here (it will take ten pages) and even then it will involve too many adjustments for him. Just to give one example - it is very much possible to deliberately play games that are 92%-97% (which he claims is impossible seeing as his average is 88% so how come some "nobody" has higher percentages) it's just that it will mean you will play far more draws. This is in part what he doesn’t understand. The computer’s evaluation isn’t at all the computer’s way of telling you you played a good game. A higher percentage means less imbalances were created and less changes caused too many differentials. In fact, you can play pretty bad chess that is very un-ambitious, in many ways wrong, leads to no initiative and no winning chances but since you were never in any problems and kept the structure sound and everything solid the overall score will be at least 94%. He doesn’t seem to understand that at all.
Flagging is disgusting. He is right about that. For every one game I get flagged I play against around 5 to 10 people trying to flag me in a lost position and it's a pointless waste of time. For both me and the flagger.
I just went on chess. I won a game 2 pieces up. He was trying to flag me. I had 22 seconds with a 1 second increment when he flagged. Completely pointless. It's braindead.
@@peterhardie4151 I don't disagree. This is the irony. Most of my life I didn't play blitz and I didn't view flagging as chess but over time two things changed: 1, I learned to respect the skills involved. 2, understanding that it's not about what I think. It's about accepting that this is what most people do and then it's about taking responsibility and either saying it's not for me and quitting chess, or only playing rapid with increment, or get better at flagging and become part of the movement. Look, I used to be extremely solid and technical which results only took me so far. Ironically it's only when I started playing practical chess, which is wrong, that I started winning tournaments. I still don't like it and I still feel ambivalent about it but I can't argue with the results. This is what chess is these days. Because of it and other reasons, I decided not to play professional OTB chess anymore. But that's my decision. I don't view it as something I understand and other who feel differently don't. To each his own.
@@c2c001 that's a very good description. I just think there is a limit. Trying to flag people in rook Vs rook or some nonsense is just embarrassing to me. You made a good point and you made it very well.
Agreed. It’s a shame, though, because cheating in general is a big problem, and a valid stement from a former World Champion could be very helpful. But Kramnik is just making a mess of things by sticking to his silly argument.
It is soo much deeper than this. All I am hearing on this subject are trivial probabilities of binary results (yes or no, heads or tails, black or red, even or odd). And it's not just 1111111 being an anomalous result, but things like 10101010101, or 110011001100, 110110011000110000, or any other pattern. And that's still just binary results with patterns added. On any number of games, all coin-flip combinations are equally probabilistic. Flipping 5 coins in the air has equal distribution of results because the system doesn’t care about patterns. Each flip of 5 has the same probability. The likelihood that you get 10101 is the same as getting 11111 or 00000. It's more like balls and strikes. Within balls and strikes you have pitch type, pitch count, batting average, leftie/rightie, men in scoring position, inning, number of outs, speed of pitch, strength of batter, the weight of the bat, the current pitch count, .... "The hard is what makes it great," to quote a great movie. Now, with that frame of mind, now, you can imagine a chess match with similar numbers of variables but in chess-y terms. Things like black or white, openings, end games, time to process moves, armageddon, clock resourcing (aka flagging), game speed (1+0, 3+0, etc.), rules of the match, historical game memory, instincts, and on and on. It's still just probability, but it is probability with dozens of variables, if not hundreds, and possibilities that permutate into the billions and trillions. Now. Now, we're in the world of mathematics, statistics, computers, algorithms, and yeah, multiple STEM baccalaureates to solve this problem. And calculations that require huge computing resources. With historical chess games effectively in databases throughout the world, it is possible to apply powerful machine learning algorithms to this problem and do so with many more orders of magnitude in precision than a mere human brain can comprehend. What I'm getting at is that we are complete fools if we think we can whip out a spreadsheet and a calculator to solve this problem. Instead, leave it to the experts that have no skin in the game. Instead of calling out players with anomalous performances calculated from a spreadsheet, then audit the auditors. It is up to the auditors to provide the transparency in methods and results they're using to keep cheating in check. And, up to the community to peer review and GMs to weigh in on the analysis and interpretations of the anomalies.
Not that hard to determine likelihood of such events assuming we know and agree with the Elo values, know how to arrive at win/lose/draw percents, know how to calculate or arrive at likelihood of such streaks. We have to be in agreement with the starting parameters, and in this case I can see why the streaks are interesting. Ans no, you do not need a lot of math background to "solve" the likelihood. You need to know how to arrive at the parameters to plug into statistics calculators, and then all of us can check the results because we will all arrive at the same results. Prior to performing the stats math: Should we should use his ending Elo and assume that represents his blitz strength, or use the average Elo of the streaks, or take each game individually? But after that the calculations are easy if we know what Elo is, know how to calculate win/loss/draw percentages, and then calculate likelihood. For the 45.5/46 streak, the draw percentage is what is most important and missed by other "mathematicians". They used a 90% chance of winning each game and if we trust Elo stats to hold for such games, then 90% is clearly wrong. The odds of such a streak assuming all known parameters is very low even if we use 10,000s of trials/games.
Also, as Fabi said, if the rating is based on his 3:2 games in titled tuesday AND 3:0 games vs lower rated players - the rating difference should be even bigger when he is playing 3:0. So it is even a smaller anomaly than it seems, since his performance in 3:0 is probably up at a level of 3500?
Online blitz chess rating measures the same skill that in person blitz rating measures. Additionally, there is another set of skills that is specific to online chess. It shouldn't be surprising that Hikaru is where he is. He is amongst the very best at both. And it is also explains Naroditsky is so high rated in online blitz. A rating that measures a broader set of skills ought to have a higher peak than one that measures a subset of them, since those that excel at all of the skills have that many more ways to separate themselves from the rest.
As another commenter mentioned happening, I once had a coin land on edge. Fell out of my pocket from waist high, and hit the ground standing straight up. This was very interesting. I'm not a Hikaru fan, but in the case of trying to smoke out engine abusers, I would urge Kramnik to read the story "The Boy Who Cried Wolf". His campaign is damaging rather than helpful. Plot Twist: Hikaru's streaks were actually ended by a cheater, and we will never know
I appreciate Fabi taking the time to discuss his thoughts on this matter so thoroughly. Knowing them both personally and being a world class player himself (Fabi), I am sure it is not easy navigating these discussions in such a public fashion. Vladimir obviously has taken some offense to Fabi's comments, which I'm guessing Fabi would prefer were otherwise given his respect for Vladimir's history in the game. This situation feels...unfortunate, maybe in a similar way that Bobby's opinions about certain groups of people tainted his career in part. Fortunately, Bobby's legacy in the game remains intact. Hopefully this ends in a way that Vladimir's legacy in chess remains intact.
I mean if Magnus is able to destroy Fabi, second best player of last decade, in SCC 2022 (which is with increment) 22-4, it's definitely possible for Hikaru in non-increment (which is his favourite) defeat nonGMs 47.5/48. If you'd ask me: Is more probable that Hikaru will win 50 games in a row in 3+0 against nonGMs or Magnus scoring 18W-8D-0L against Fabi in 5/3/1+1, I would definitely choose Hikaru streak.
@@styleisaweapon that is just wrong. it is a fallacy to believe that you have to respond to everyone equally. Kramnik represents basically an internet troll right now, no matter his past achievments, because they dont objectively change the point he is trying to make. Not responding to every internet trolll doesnt create one sided narratives. His claims are as valid as claims about hans using some vibration toll up his ass. There is no evidence to back up anything about it. Kramniks claims habe similar validity.
@@LunnarisLPkramnik definitely doesn't represent an internet troll. He's one of the greatest in the world, speaking publicly, and as far as I can tell, in good faith. He's not one of a million anonymous twats talking bs to get strangers upset for amusement.
20:33 Hey Fabi, I'm the grandmaster you destroyed (Benjamin Bok). My butt is still a little sore from the poundings but at least I got some good practice in against a real grandmaster. Also, forgot about all the lunch money you took from me! Could have used that to upgrade my microwave!
As a youngster in school we got in groups and flipped a 2p coin 100 times to explain probability to us. Not only were there several long streaks, but one time our coin settled on the side. So it was a good lesson.
I think that difference between ratings can be explained by level of people. For example I have 1800-1900-1950 rating. I have it because 20+ years ago I used to play chess and used to be a chess player. Today I’m no so good at chess as many years ago, so my rating is lower and I play no so well because I forgot all the openings, theory, I don’t have practice and so on. In comparison with me other guy 1800-1950 is a current young chess player. He plays every day, has classes, grows up etc. We both have the same rating right now, but it’s obvious that this guy is stronger than me and he will quickly grow up, and for me it’s my maximum rating ever.
as a 1600 FIDE rating player that have played guys from 1400 to 2200 fide rating range, playing someone who is 200-300 points better than you is a big psychological pressure.. just this thing could influence the whole streak so much..
@@youtubecensoringcomments7427 my guy you havent been in the competitive chess environment for a long time have you.. those cases were completely different, with one there was evidence of cheating and alot of top players knew that and for the other there is litterally nothing to even hint to the fact
@@goldenpaperyt9763 I think you were not able to process what I wrote, pause and think. Magnus and Hikaru did accuse hans niemann of cheating at the sinquefield cup without proof. Still to this day there is no proof... The rest is your opinion
Very few Super Grandmaster are usually playing Online as frequently as those IM Players. So, Hikaru does not have another choice but to pick what he can from the pool. Exposing the opening preparation is a risk for Fide tournament players; so, they are avoiding each other online even if they are playing in the same platform. If all top 100 players were playing online against each other, with the same level of dedication as Hikaru, the Online rating would be more in parallel with the FIDE rating.
Those weaker opponents are paying Hikaru for his instructions with rating points. That's why you don't see any of those players supporting Kramnik's (non-)claim of fraud.
I wrote a simple python script to take in the ratings of 2 players and simulate 1000 games between them. Player A was 3200 and B was 2800 in my test. The highest win streak for player A was 41 and this was only from a sample of 1000 games. Of course this ignores a lot of factors but statistically winning 40+ games in a row is well within the bounds of reality.
@@pianochess1882 regardless, if I sampled 40 games enough times I’d get a period just like the one he had. I don’t know what the rating difference was but there is a 2% of winning 40 consecutive games with a +400 rating. That ignores how triggered a player would get losing so many games too.
I completely agree faster time control is a different skill. My rating difference against my friend in blitz is +500 but in lightning he is +200 Increment and time allowed to think can hugely make a difference between performannce and overall results..
Fabi does a great job of breaking down the flaws in kramnik's analysis...the whole idea behind using ELO-based statistics is not a good approach. It seems the way GMs sniff out cheaters is that they play difficult moves too quickly and spend too much time on easy moves, and in general have bizarre clock usage. Kramnik is just barking up the wrong tree
Dear Vladimir... take a break, spend time with loved ones, enjoy time with friends and family... look after yourself. Life is short... enjoy it. Relax.
I think everyone including Hikaru has great respect for Kramnik and his accomplishments. However , his crusade against Hikaru is obviously motivated by his personal dislike for him and he is doing nothing more than losing all credibility.
I don't see how that being his motivation is obvious. There doesn't appear to be any kind of consensus opinion that that is the case, which there should be, if it was obvious. As far as I can tell, he's on a crusade against cheating, and thinks he found something. Right or wrong in his accusations, his motivation seems... obvious. Hehe:)
@@drunkrtard Perhaps but his insistence that Hikaru’s “record” needs to be checked thoroughly because there are statistical anomalies tells a different story.
I don't know, Kramnik has been on an anticheater crusade for a while now, and sometimes I wonder if he is dumb enough that he tought people wouldn't realize that one accusation was against Hikaru.
@@dan1e1473 Yeah I don’t know. He keeps saying that according to his “mathematician friends “ the statistics are bad but he not once names them or posts any of their actual proof. Meanwhile, serious statisticians have come out regularly in support of Hikaru and yet Kramnik keeps beating the same drum. If he’s not out to get Hikaru then why is he doing this? Everyone knows there is cheating going on in online chess but to call out Hikaru like that for no good reason and no proof is “very interesting” to say the least. I’d like to believe his motivations are pure but at what point will he admit he was wrong about Hikaru? I don’t believe he ever will.
Fabi, can you speak to the difference between paranoid suspicions of cheating resulting from delusions of your own strength, to the confidence you have when playing an actual game where you feel strongly a person cheated.
I remember several years ago when TT started becoming very popular and I thought to myself (as Kramnik would say) - its very interesting how there are so many online GMs in these events who are posing challenges to the top GMs in the world, such as Magnus, Hikaru, Fabiano, Ding, Nepomniachtchi, Wesley, etc. Players who would get absolutely smoked OTB, but online they are suddenly playing at a 3000+ level... And we dont even need statistics to realize what is going on, just look at the top OTB tournaments in the world. Its always the same top GMs (as it should be). This includes the rapid and blitz world championships where all of these online "comets" are nowhere to be found. Im glad Kramnik has brought a lot of needed attention to this topic.
As any and all of these top players, who are still in touch with reality would say, any of these strong players has a chance in a one off game against these guys. Some of them don't have the chance to appear more often OTB because of lack of invitations or means of travel etc. There are problems with the way top chess has been, where it's always the same players who then are in a rating bubble just exchanging rating with the same guys. Online chess removes this bubble, where now these weaker but still strong players can take a game off of them once in a while
To compare OTB games and online games is a mistake. I mean premove can’t be done OTB and Mr Kramnik looks like he doesn’t understand that, he doesn’t understand esports or online when he comes to chess. Games that Mr Nakamura won are not always winning positions but more about how fast he plays and how fast he premoves and put pressure. So you can flag someone with a losing position. Mr Kramnik uses fallacious arguments against Mr Nakamura by intentionally saying that Mr Nakamura won 46+ games with good positions.
I would investigate ALL of Nakamura's latest results, say, from the last two years, not only at online, but at over the board (classical) chess too. There's something odd about this latest "comeback" of his in classical, especially when you think of his age, and that it was never really his forte...
As far as the “farming,” I assume Hikaru is also trying to make good content, because he takes his role as a streamer at least as seriously, if not more seriously, than that of a chess grandmaster. He will be inclined to pick games that make for good content. There is a reason why he is so successful at streaming.
It was crazy to me that Kramnik tried to imply that farming was unethical. These are casual games that were agreed upon by all parties, no prize money on the line, under an online rating system that isn't particularly meaningful, at least not to the extent FIDE ratings are. Playing for fun/content is fine.
3:40 Kramnik said it's not about streaks, it's about anomalous increases in rating. he doesn't understand why his critics focus on streaks that much, Hikaru's streaks are not his point
Some specific key words that Fabiano uses shows that he has studied some math and statistics and he is not blindly commenting on this matter....weighted average,small subset,etc.
I feel like Vlad is forgetting / underestimating that chess is a psychological game, the impact your mind state can have when playing chess is massive. If you only play 1 game at a time against different players then the effect this would have on statistical variance would be fairly minimal. However when you play a large set of games in a short space of time with fast time controls and are repeatedly playing the same player / players, the effect of psychology on your performance is enormous.
I think one thing is not said enough is that those 48 games are casual games. Imaging being Hikaku, who qualified candidates, biggest chess streamer and so on, and you cheat by putting your reputation on the line so that you can win games against 2 IMs and 1 FM.
Agreed… what do the detectives say motive, then means and opportunity.
Hikaru didn’t have a motive to cheat.
hikaru literally doesn't care about casual tournaments like the candidates or the grand swiss; that's just his side hustle anyways.
His entire focus as a professional twitch streamer is on the high stakes twitch blitz matches against people 500 elo weaker than him.
If I was Hikaru, I would be more offended about someone insulting my intelligence than insulting my character. Dudes got it made, imagine risking his whole career to cheat in casual games 😂
Yah it's like a street basketball game compared to an NBA basketball game
You do realize that's not an argument?
I think the illustrative point about Kramnik is how much he simultaneously insists that he’s not accusing Hikaru of cheating, and yet baselessly insists that Hikaru’s results are impossible. He insists on investigation, yet rejects evidence presented. There is fundamental dishonesty (or delusion) to that. While it’s worthwhile to refute Kramnik’s inaccuracy on the mathematics and the particulars of online chess (both rating mechanics and actual play) for the broader community, it doesn’t seem productive to respond directly to someone who clearly is not honest, who is not participating in good faith.
Whatever his past accomplishments, Kramnik’s current actions are simply not respectable.
11:21 " I dont think he (Naroditsky) ever won a Titled Tuesday."
Daniel Naroditsky has won the Titled Tuesday tournament at least three times.
He emerged victorious on December 30, 2014, September 1, 2015 and February 22, 2022
Okay? That doesn't add anything to the video. The point he made still stands.
Yes it does. There is a significant difference between 0 and 3. Christian needs to correct the record. There's probably only a dozen people or so that have won titled tuesday, and Danya is one of them. He picked a bad example for his point. He also belittled Danya. He absolutely should say something to acknowledge the error.
He won on 2-22-22? That's a lot of consecutive 2's... Very low probability 🤔
There's like 50 different winners since 2022.
@@aajasinteresting 🤔
Kramnik does not accuse people of cheating. He just embarks on a Special Statistics Operation.
daaaaymn!
SS Operations you say?
he provides interesting and unsettling stats
OMFG :D
Interesting..
Fabiano got this point right. The streak that Kramnik identifies as 47.5 out of 48 could be equally viewed as 47.5 out of 50 if someone else includes the games immediately before and after the long streak of wins.
it doesn't matter at all, what matters is that he Hikaru does it regularly 5-10 times a month. If something happens once, it's an extraordinary case, if an «extraordinary» i.e. statistically improbable case happens regularly, it's a reason to check for cheating.
@@kazimirmalevich6712and the statisticians already showed their work explaining to kramnik its not supposed to be that rare for Hikaru and it should happen as often as it already does with no shock factor.
@@kazimirmalevich6712 So, you want to say if Hikaru cheats 5-10 times a month thats normal lol But if he cheats once a year thats suspicious. haha dude hikaru is extremly suspicious.
@@Viewer13128 I'm sorry, but I can't accept "someone out there explained..." It's very simple, if 47.5 out of 48 is a rare extroordinary result with probability P=0.01, then Hikaru's five times in November have a probability P^5=0.0000000001 that's ten zeros! This does not fit into the framework of an extraordinary result. Let Hikaru explain it or let it pass the cheating test.
@@kazimirmalevich6712did you just divide one number by the other, completely disregarding these are individual probability events with a complex model explaining their likelihood? And then pretend it's stats by giving it a p-value?
I love how, in the beginning of this video, Fabi gives off a real strong _"C'mon. I don't even have time for this shit!"_ vibe 🤣
Lmao
Which is a good sign of a healthy mental state
You don't know Fabi then. He cares strongly about defending Hikaru. All the top American players have each other's backs.
@@patrickpet7905 Two things can be true. He might wanna defend Hikaru. But he also might not wanna have to be addressing these silly accusations/callouts at 10pm on a Saturday night lol
Fabi really doesnt have that much time for it, being an active top Grandmaster. Mr. Kramnik has all the time in the world now, but I would appreciate seeing his actual evidence. Recently in Kramnik's recent posts it has been simply denials of others that have opposing beliefs and statistics.
In my local chess club there was an otb blitz match between a 2200 and 2100 fide classical and 2200 won 20-0 the idea that these things dont happen is crazy
Interesting.
2100 was probably drunk or dead... or he wasn't 2100.
Fabi is very objective about his thoughts and chess.
I have a question for both Fabi and Christian - when you attend these parties where strong super GMs perform these “party tricks” and play with time odds against other players, are people like Kramnik, Nepo, etc. usually present? And do they also play with time odds against others?
Very true, so articulate in his expression
I don't know about objective (and I'm not saying he's not), but he's very measured, precise and clear in expressing his thoughts.
he is so emotionally stable, he seems robotic
The problem with Kramnik accusing Hikaru is that there is so much real cheating going on, and by accusing Hikaru instead who is obviously not cheating he bastardizes any future claims he may make of people cheating.
Its hard to take him serious now, even if he claims to have a team of mathematicians.
Also it proves that a statistical analysis alone is not enough, you need to use common sense also
Well, as far as statistics go, it demonstrates a lot better how important it is to look at the whole picture (all of the tens of thousands of games hikaru has played) instead of one small subset
@@Kayrim_Borlan Indeed, if the statistics were done properly, that should have been taken into consideration in the first place.
I guess the team of mathematicians he has is quite bad at math
Statistics is not mathematics, first lesson in statistics is to be careful with interpreting numbers, on top of that there are outliers in statistics. You cannot statistically prove anything. Of anything it is drawing approximate conclusions from observations. Economists have been consistenly been mistaken because there are too many variables.
@@kimgysen10 I mean, Statistics are a branch of mathematics. So I don't know what you mean by saying they aren't mathematics. That is just objectively false. Then you jump both feet in with more fallacies. You cannot statistically prove anything. Ok?? Who said that statistics alone prove something? If statistics alone are used to accuse, they should be good enough to defend against the claims. In this scenario, the statistics help to satisfy probability. When applied correctly, they do not indicate any type of abnormalities with his win to loss ratio and don't support the idea that they point to cheating. Then you hit us with that false equivalence, talking about Economists making mistakes due to the variables in the economy. And? What does that have to do with the length of ADs in a RUclips video? Hikaru doesn't have to prove he isn't cheating. There isn't any credible evidence that he is. This is how logic works. Without evidence of him cheating, you can't use statistics to prove he is, unless you are willing to also allow it to prove he isn't. So if we can't clear his name using the proper application of statistics in regard to the probability of cheating. You damn sure can't use the improper application of statistics to make the accusation that he is, and then deny him the ability to clear his name using the method you chose to aim allegations against him with.
This is what I hate about the internet, and people who just want to argue and take a position without having any foundation to stand upon. It's usually based on biases, either against Hikaru or for Kramnik(in this specific situation). You wrongly say statistics aren't mathematics, then say that statistics couldn't even prove his innocence anyway. Yet you are completely okay with using them to prove his guilt? That is top tier hypocrisy. The proof to condemn someone should be greater than the proof needed to prove one's innocence, shouldn't it? So we avoid the concept of falsely convicting or condemning those innocent of their allegations? SMH. So disappointing. You should strive to think more critically about things and to be better in life than someone who is wiling to throw out logic to appease their own opinions. Other than that, the statistics in economics have NOTHING to do with the probability of cheating in chess. So please avoid using a lie and fallacies to try to argue against logic, pwetty pwease? You are better than this. I have faith.
@@nerdjournal I’m not reading junk from someone with clearly no experience in the field.
I think there's a really useful analogy to use that really reframes the situation and makes it easier to understand why an insane winning streak might not be as strange as it seems on first glance. Consider this : if you roll a set of six dice and you get all 1s, that's incredibly unlikely and if you can do it on demand, you're almost certainly manipulating the dice somehow.
However, if you are spending all your time rolling six dice and you do it enough times (let's say a million times, to give an extreme example), the way the probabilities work, the situation flips: it actually becomes extremely unlikely that you DON'T ever get all 1s.
If you're standing around rolling your six dice as usual and a random bystander notices that you rolled all 1s, from his perspective he sees this as a remarkably unlikely event. But from your perspective, you know that this is just the product of you spending all your time rolling your dice.
To come back to Hikaru and his apparently insane winning streaks, basically the equivalent is that Hikaru spends enormous amounts of time playing chess, since streaming himself playing chess is his main career now. Most of the time his results are (for his ability and his opponents' ability) probably more or less average. But occasionally he gets apparently insanely good results, simply as a product of playing so much (and probably some terrible results as well). And to an observer who doesn't consider all the games he plays (and looks only at the insane winning streaks), it comes off as remarkably unlikely.
To be sure, I don't know how the actual probabilities work out when it comes to how much Hikaru plays compared to how unlikely his winning streaks are, but I just think the above is good to keep in mind.
Your example is very good to try and shape a sensible intuition with regards to probabilities. Hopefully it'll reach some people who can learn from it, but I am certain it will not reach some others. Kramnik mentioned the fact many people criticized his interpretation by pointing at cherry picking, and his response in part was verbatim "of course it has to be cherry-picking!" 😂
You could also use coin flips as an example. Do enough and you'll get long streaks of either heads or tails. One way to analyze it is to imagine a million people flipping a coin, then everyone who gets heads flips again, then repeating that until you have a last toss with a head. I'm not sure exactly what the likeliest number of heads in a row is, but it would be up to 19.
Kramnik's "interesting" statistics regarding Hikaru were poorly thought out from the start. He obviously didn't get any of the statistics checked by a Statitician, he clearly didn't do any research into who Hikaru was playing, and he didn't look at any extraneous factors like accuracy or move time. Lastly, the knock out blow for me was this - Hikaru's streak(s) were statistically LIKELY. Kramnik should've owned up to his mistake then, trying to pursue it just made him look silly.
Edit: RUclips will not let me mention the chess website who said Hikaru's winning streaks were statistically likely. Everytime I do it removes the comment.
He didn't even bother to show any statistics. He just listed out a streak. He didn't bother to do any statistical analysis and show some work.
@@saikat93ify The "mathematicians" and "statisticians" online did not show their work in arriving at the beginning parameters. If you use each game's Elo, arrive at probabilities of wins (and draws) and determine the probability of such a streak within thousands of games (or the absurd 30K games) then we can see why there is a question.
I wish he would show his work because the "mathematicians" certainly did not. They all skipped the part where draws were likely at about 20% of the time.
@@scottekoontz I have the same problem. I can't use any URLs in comments, ironically not even links to RUclips videos. It's worse than that, though. Just now my reply to you was deleted because I used an example of how even if I leave out the dot from the URL of the site you're talking about, I can't post successfully. I used somedomainname___, where ___ was com. It seems any string of characters that ends with com isn't allowed.
What really galls me is I had to report the same spammer posting links in chess video comments that redirected to a porn site multiple times before it stopped.
@@scottekoontz if I play my 600 elo brother as a 1000 50 times, how many games do you think I’ll win?
@@thebcwonder4850 You have a 90.2% chance of win per game, and with a string of 50 games you can expect to win about 45 of them, and at these Elo ranges the expected draws at 3.4% is lower than the expected losses at 6.4%. For higher Elos with a difference of 400, e.g., 2000 vs 1600, the WLD is about 88 / 4 / 8 which indicated that draws become more common.
In 50 games your expected longest string of wins is 20. If you wanted a string of 50 wins in a row you would want to play over 1,300 games to have equal probability of achieving such a streak as not.
I love hearing fabi speak about chess I learn so much every time he explains something.
Appreciate you guys and your thoughts. Thanks for being willing to share and make this video!
Chris/Fabi. As a mostly retired 'old' 2400 USCF player, It is very clear, that while cheating does take place (being a reality) at various levels, what Kramnik (and Nepo) really object to is the "Farming"...as they and other sporadic high level players that only play key online events, do not do this. Farming has occurred at all levels for many many years, whether at the very top (3000 plus players farming 2700-2800s), 2800 players farming 2500, 2500 farming 2200, and so on. Players do this to increase their ratings for a variety of reason, and I am speaking now specifically for online platforms. It doesn't matter to Kramnik/Nepo that some GM's do it for content or not, and of course it can equate to $$$, which again, Kramnik and Nepo do not enjoy having, or player Farm for other reasons such as 'status' and respect..I guess. While Cheating is an underlying issue that certainly needs to be constantly addressed and I believe is, to some regard, the Kramnik's and Nepo's (etc) think farming, if not cheating, is unethical. I appreciate and really enjoy your channel, content and boadcasts.
Kramnik’s next post
“A certain American GM has been having the best results of his career in almost 5 years. I find this very interesting.”
The way the background aligns, I thought you guys were in different locations and stitched your cameras together 😂
dont do drugs :)
I noticed they weren't when I read your comment
They did it on purpose..
I've been fooled into thinking these guys are on two different locations having split screen call but then realized that's just something on the wall :D
i scrolled down to see if anyone else saw this cuz it tripped me out
HOLY SHIT!
Wtf, I was one hundred per cent sure they were at different locations and had aligned their cameras to make it seem like they were having a conversation side by side.
Fabi is such a nice dude. Refreshing compared to many on top.
Now I wanna see Danya adopt Fabi in 1 + 0 online. That would be a hoot!
On a more serious note, Fabi should consider hiring Danya as a coach to improve his mouse/ultra-high speed skills.
I would be far more suspicious if there were NO such streaks.
Absolutely. There has to be streaks like that for there not to be something shady. Kramnik doesn't understand statistics.
Interesting
What's shady about losing all the time? I do it all the time.
Actually, true! I mean if I wanted to cheat, I'd try to not make it so obvious. So, I'd avoid streaks that draw a lot of attention to me.
@@PeteQuadIt is quite contra intuitive and it's not really a thing that comes natural to many. This is not to slam Kramnik but maybe he has some wierd form of math dyslexia that makes this hard to get while being very intelligent in most other aspects?
Fabi making the camera shake and Christian tried telling him but to no avail. 😀
Maybe because Kramnik has been so good for so long he's forgotten how intimidating it is to play a much stronger opponent. I very much doubt the IMs who play Hikaru are playing at their very best when they are expecting to lose nearly every game.
Correct , Kramnik was a beast in his Prime and that can be the reason why he is so deluded.
Good point
Kramnik simply does not understand that there is more to chess skills when it comes to online chess - mouse skills, board vision, art of playing second best moves quickly (which Hikaru is the best in the world in my opinion). As such, streaks can be very common for these online specialist players like Hikaru, Alireza, Danya, Nihal etc
It’s a good point but I think chess players in particular are really good at dealing with stress and I don’t think it is much of a factor
@@schizoframia4874 I think it's a factor when you're on your tenth game in a row of getting absolutely obliterated by someone like Hikaru. People go on tilt
Fabi is a class act to the bone. I always enjoy listening to him speak.
There is one obvious explanation for the farmed rating being inflated that you (and Kramnik!) missed: the farmed *also* farm, but against lower rated players.
“…the farmed also farm..”
Amen.
Yep, and I'm soooo sure no money is involved in this whole scheme.
Those players didnt farm, checked
As Qui Gon used to say, there's always a bigger fish
It should be mentioned in all in these conversations that Hikaru recorded himself doing this run and made a youtube video of it. He literally is explaining his moves in the video. So...... Everyone talking about this has seen the video, and know the whole thing is stupid. Ages later we are still talking about it when a tiny percentage of people think anything actually happened. I guess it's fun to talk about, but it is just content. I doubt there is much left to say on it now.
Yeah hikaru is like the most bullet proof player in the world with this stuff. He has countless otb results backing up his online play, streams most of his success, and vocalizes his thought process through the majority of it
There's literally not a one letter in yr argument backing the idea he couldn't do it WITH engine advices...
Imagine telling a world champion how to play chess, I doubt there is much left to say on it now.
@@tvrskkngdm2333 The thing is, in a lot of these games he simply does not have the time for engine advice. We all know, stockfish and other engines need "time" to calculate. In bullet without increment, there is basically no shot the engine can help him in the end, where he tries to flag the opponent. You can't even argue "oh yeah, he used engine earlier, thats why he has the better position!", because in most games of the streak he was dead lost. He basically just played moves to survive. There is basically no shot he cheated in these games.
And like Fabi said, he can utilize his skills in meaningless OTB matches as well, without preparation, without some engine vibing his ass. Fabi encoutered it in real life and backs Hikaru.
@JackHarb89 Besides his speed, there's also the fact that Hikaru has absolutely insane defensive skills, even compared to other super gms. Even from a dead lost position, there are very few people who can turn the disadvantage into a mate, and combined with his speed they'd need to be able to play the best/second best move almost instantly. Someone like Danya, magnus, alireza could most likely do it, but 2 IMs and an Fm? Probably not happening
Love this show. Hope you keep it up after you become the world chess champ.
So basically you're asking them to stop keeping it up?
Fabiano's explanations are so concise yet clear. I listened to Kramnik for 2 hours the other day (couldnt handle the whole 3h interview) and he rambled on for a long time without reaching any real conclusions. IF Kramnik truly believes in this he should be clear and take it further (if he can) or apologise and put the matter to rest once and for all.
Thank you Fabiano for sharing your insights with us.
I used to give a demonstration in one of my classes in which I had a student write down a "random" sequence, where they could select triangle, square, or circle. So their sequence would be something like CSTTSTTCCSCT... etc. I would guess the first symbol, they would tell me if I'm right or wrong, and then I'd guess the next and so on. The point of the exercise is I'd always do better than 33% correct - usually pretty close to 50%. The trick is that I get feedback on whether I'm right or wrong, and use that to make my next guess, knowing that a human will rarely repeat when trying to generate a "random" sequence. A true random sequence will have the appearance of, say, 4 circles in a row, but that looks "unrandom" to the eye. We want to invest meaning in any sequence, but sometimes it's just luck. Or rather, luck coupled with the fact that Hikaru should win against these players 95% of the time. I also find his streak quite plausible given the factors Fabiano identified.
Isn’t that the beautiful thing about playing online, you get to play multiple games against really good players. Those IMs and FMs got a lesson and multiple games to analyze. I’d love for Hikaru or Fabi to whip me for an hour, because they’re taking their time to play ME. Great, and FREE training.
Totally agreed!
Right? it's one of the only games where you have the opportunity to play one on one with the best players in the world. That's just cool
Huh🤨
Wouldn't a chess engine be more useful? Playing a few games against a 2700 online is more a party anecdote really than training.
@@declupNo, playing a very strong human is much, much better than playing an engine. Humans, obviously, aren't computers, and no matter how much training they do will never be able to calculate as much as one. Because of that, it's extremely difficult for even top level players to understand most of the moves, and next to impossible to replicate their performance. Even Magnus would lose 100% of the time playing against stockfish, and only maybe be able to pick up anything useful. At the top level engines are studied because there isn't much else TO study, but anyone below that is much better off studying human play which is what they're going to be playing against anyway
I wonder what the statistically equivalent losing streak for Hikaru is, and how many times in the 30,000+ game record this has occurred. I suspect the odds of hikaru having a winning streak such as this is maybe equal to him losing 4 games in a row, which I assume he's done a few times in that 30,000 game record.
Fabi is a true gentleman and a great embassador of the game. Carry on gents!
It has been discussed how the online ratings aren't reliable and also how time controls matter. But even if you get the compensations for that pretty much correct, there are still 2 statistical traps that are easy to fall into. The first and more obvious one is that you can't just look at the streak in isolation. Instead, you have to interpret every game as the start of a new streak attempt. The second, less obvious one is that it's invalid to look at a streak and ask "what were the chances?". You have to define exactly what you are looking for before you look at the data. So what streak length will be looked for and how many draws will be allowed in that streak. If you start trying to estimate chances after the fact you get absurd results, like asking "what where the chances of that leaf falling in exactly that way from the tree(wow, much worse than a trillion to one).
It’s so ridiculous that they have to explain this AGAIN, and in more detail this time for Kramnik to understand! Kramnik looks like a fool at this point
From my experience (and this just adds to it) it's just very difficult to make someone understand something they so wholly don't understand and can’t grasp.
Looking at it from the side, I could recognise each time Kramnik failed to adapt to the changes the electronic age brought to chess but putting them together and trying to argue these are the reasons for his lack of... graspness, is difficult.
He is obviously a methodical person(ality) who finds it difficult to adjust concepts once those are formed. There seems to be a battle, in his mind, between, shall we call it - "the Botvinnik school of chess" and this computersied era where his view seems to be that studying "move per move" as he puts it, is inferior and so coming from a place of "higher ground" he is trying to preach his truth.
Yes, there's some elitism there but that's not the main issue. His refusal to accept practical chess and let go of old ethics are, in part, what's debilitating him… from adjusting.
He views flagging as disrespect, gaming-the-system (i.e. not really chess), he feels sorrow for the loss of hierarchical... manners... towards former (older) champions, etc.
To a modernist it seems like he is looking for favouratism but from his standpoint dirty-chess (in particular) is simply not chess at all since it’s too practical and fundamentally wrong.
Unfortunately, there is a lot that he doesn't understand. In fact, it is so much and involves so many topics it'll be impossible to write it all here (it will take ten pages) and even then it will involve too many adjustments for him.
Just to give one example - it is very much possible to deliberately play games that are 92%-97% (which he claims is impossible seeing as his average is 88% so how come some "nobody" has higher percentages) it's just that it will mean you will play far more draws. This is in part what he doesn’t understand. The computer’s evaluation isn’t at all the computer’s way of telling you you played a good game. A higher percentage means less imbalances were created and less changes caused too many differentials. In fact, you can play pretty bad chess that is very un-ambitious, in many ways wrong, leads to no initiative and no winning chances but since you were never in any problems and kept the structure sound and everything solid the overall score will be at least 94%. He doesn’t seem to understand that at all.
@@xXG3TPWNEDXx You certainly think you know things you don't know.
Flagging is disgusting. He is right about that. For every one game I get flagged I play against around 5 to 10 people trying to flag me in a lost position and it's a pointless waste of time. For both me and the flagger.
I just went on chess. I won a game 2 pieces up. He was trying to flag me. I had 22 seconds with a 1 second increment when he flagged. Completely pointless. It's braindead.
@@peterhardie4151 I don't disagree. This is the irony. Most of my life I didn't play blitz and I didn't view flagging as chess but over time two things changed: 1, I learned to respect the skills involved. 2, understanding that it's not about what I think. It's about accepting that this is what most people do and then it's about taking responsibility and either saying it's not for me and quitting chess, or only playing rapid with increment, or get better at flagging and become part of the movement.
Look, I used to be extremely solid and technical which results only took me so far. Ironically it's only when I started playing practical chess, which is wrong, that I started winning tournaments. I still don't like it and I still feel ambivalent about it but I can't argue with the results.
This is what chess is these days. Because of it and other reasons, I decided not to play professional OTB chess anymore. But that's my decision. I don't view it as something I understand and other who feel differently don't. To each his own.
@@c2c001 that's a very good description. I just think there is a limit. Trying to flag people in rook Vs rook or some nonsense is just embarrassing to me. You made a good point and you made it very well.
6:48 (Read with Kramiks voice) "Ok, Christian looking clearly second screen, yes?"
Fabi is a true gentleman
Imagine kramy saying "okay yeah I was wrong, sorry guys"
What's the probability?
It’d be suspicious. I’d expect it to be written by an AI.
@@Fergus316 very suspicious indeed haha
Less than for Hikaru's streaks.
Low enough probability that it would be "interesting"
Fabi is confirming points that Hikaru has also made. From the beginning, Kramnik was making a poor argument, and it has only gotten worse.
Agreed. It’s a shame, though, because cheating in general is a big problem, and a valid stement from a former World Champion could be very helpful. But Kramnik is just making a mess of things by sticking to his silly argument.
@@macleadg Yeah, he's making it look silly for anyone to assert that there's cheating.
Fabi being so level-headed and rational about this puts Kramnik to shame lol
It is soo much deeper than this. All I am hearing on this subject are trivial probabilities of binary results (yes or no, heads or tails, black or red, even or odd). And it's not just 1111111 being an anomalous result, but things like 10101010101, or 110011001100, 110110011000110000, or any other pattern. And that's still just binary results with patterns added. On any number of games, all coin-flip combinations are equally probabilistic. Flipping 5 coins in the air has equal distribution of results because the system doesn’t care about patterns. Each flip of 5 has the same probability. The likelihood that you get 10101 is the same as getting 11111 or 00000.
It's more like balls and strikes. Within balls and strikes you have pitch type, pitch count, batting average, leftie/rightie, men in scoring position, inning, number of outs, speed of pitch, strength of batter, the weight of the bat, the current pitch count, .... "The hard is what makes it great," to quote a great movie. Now, with that frame of mind, now, you can imagine a chess match with similar numbers of variables but in chess-y terms. Things like black or white, openings, end games, time to process moves, armageddon, clock resourcing (aka flagging), game speed (1+0, 3+0, etc.), rules of the match, historical game memory, instincts, and on and on. It's still just probability, but it is probability with dozens of variables, if not hundreds, and possibilities that permutate into the billions and trillions.
Now. Now, we're in the world of mathematics, statistics, computers, algorithms, and yeah, multiple STEM baccalaureates to solve this problem. And calculations that require huge computing resources. With historical chess games effectively in databases throughout the world, it is possible to apply powerful machine learning algorithms to this problem and do so with many more orders of magnitude in precision than a mere human brain can comprehend.
What I'm getting at is that we are complete fools if we think we can whip out a spreadsheet and a calculator to solve this problem. Instead, leave it to the experts that have no skin in the game. Instead of calling out players with anomalous performances calculated from a spreadsheet, then audit the auditors. It is up to the auditors to provide the transparency in methods and results they're using to keep cheating in check. And, up to the community to peer review and GMs to weigh in on the analysis and interpretations of the anomalies.
Not that hard to determine likelihood of such events assuming we know and agree with the Elo values, know how to arrive at win/lose/draw percents, know how to calculate or arrive at likelihood of such streaks. We have to be in agreement with the starting parameters, and in this case I can see why the streaks are interesting.
Ans no, you do not need a lot of math background to "solve" the likelihood. You need to know how to arrive at the parameters to plug into statistics calculators, and then all of us can check the results because we will all arrive at the same results.
Prior to performing the stats math: Should we should use his ending Elo and assume that represents his blitz strength, or use the average Elo of the streaks, or take each game individually? But after that the calculations are easy if we know what Elo is, know how to calculate win/loss/draw percentages, and then calculate likelihood.
For the 45.5/46 streak, the draw percentage is what is most important and missed by other "mathematicians". They used a 90% chance of winning each game and if we trust Elo stats to hold for such games, then 90% is clearly wrong. The odds of such a streak assuming all known parameters is very low even if we use 10,000s of trials/games.
Also, as Fabi said, if the rating is based on his 3:2 games in titled tuesday AND 3:0 games vs lower rated players - the rating difference should be even bigger when he is playing 3:0. So it is even a smaller anomaly than it seems, since his performance in 3:0 is probably up at a level of 3500?
His peak bullet rating is a little over 3500 online but that's a bit inflated.
Online blitz chess rating measures the same skill that in person blitz rating measures.
Additionally, there is another set of skills that is specific to online chess.
It shouldn't be surprising that Hikaru is where he is. He is amongst the very best at both. And it is also explains Naroditsky is so high rated in online blitz.
A rating that measures a broader set of skills ought to have a higher peak than one that measures a subset of them, since those that excel at all of the skills have that many more ways to separate themselves from the rest.
As another commenter mentioned happening, I once had a coin land on edge. Fell out of my pocket from waist high, and hit the ground standing straight up. This was very interesting.
I'm not a Hikaru fan, but in the case of trying to smoke out engine abusers, I would urge Kramnik to read the story "The Boy Who Cried Wolf". His campaign is damaging rather than helpful.
Plot Twist: Hikaru's streaks were actually ended by a cheater, and we will never know
I appreciate Fabi taking the time to discuss his thoughts on this matter so thoroughly. Knowing them both personally and being a world class player himself (Fabi), I am sure it is not easy navigating these discussions in such a public fashion. Vladimir obviously has taken some offense to Fabi's comments, which I'm guessing Fabi would prefer were otherwise given his respect for Vladimir's history in the game.
This situation feels...unfortunate, maybe in a similar way that Bobby's opinions about certain groups of people tainted his career in part. Fortunately, Bobby's legacy in the game remains intact. Hopefully this ends in a way that Vladimir's legacy in chess remains intact.
I mean if Magnus is able to destroy Fabi, second best player of last decade, in SCC 2022 (which is with increment) 22-4, it's definitely possible for Hikaru in non-increment (which is his favourite) defeat nonGMs 47.5/48.
If you'd ask me: Is more probable that Hikaru will win 50 games in a row in 3+0 against nonGMs or Magnus scoring 18W-8D-0L against Fabi in 5/3/1+1, I would definitely choose Hikaru streak.
Perhaps Kramnik has reached the early stages of chess madness.
Early?
Lol like magnus and Hikaru with the niemann case??😂😂
@@youtubecensoringcomments7427 Hans had cheated before
@@youtubecensoringcomments7427Not comparable. Hans is a known cheater
@@thebcwonder4850Yes, but they still were completely out of line
Dear Kramnik, thank you for shuffling the deck, he’s got the guts
Not responding to people not worth it responding does not mean lack of balls.
but it does mean a one-sided narrative
@@styleisaweapon that is just wrong. it is a fallacy to believe that you have to respond to everyone equally. Kramnik represents basically an internet troll right now, no matter his past achievments, because they dont objectively change the point he is trying to make. Not responding to every internet trolll doesnt create one sided narratives. His claims are as valid as claims about hans using some vibration toll up his ass. There is no evidence to back up anything about it. Kramniks claims habe similar validity.
@@LunnarisLPkramnik definitely doesn't represent an internet troll. He's one of the greatest in the world, speaking publicly, and as far as I can tell, in good faith. He's not one of a million anonymous twats talking bs to get strangers upset for amusement.
Why did Fabi Left the last titled Tuesday?
@@drunkrtardHey, don't insult my hobbies!
Surely you don't need a huge zoom call. Only their opponent needs to be on a call to check for cheating.
20:33 Hey Fabi, I'm the grandmaster you destroyed (Benjamin Bok). My butt is still a little sore from the poundings but at least I got some good practice in against a real grandmaster. Also, forgot about all the lunch money you took from me! Could have used that to upgrade my microwave!
Kramnik is octupling down by now 😂
Kramnik: "People will find this interesting."
Fabi: "No, I don't think I will."
I’m pretty sure that without increment there are players around 2900 that never could beat Hikaru.
Vladimir isn’t yesterday’s news, he isn’t even yesteryear, time he disappeared and opened an Aloo Tikki store in the back of beyond
Nak is almost ALWAYS going to be up on time, combined with his skill set he can get wins and flag them.
As a youngster in school we got in groups and flipped a 2p coin 100 times to explain probability to us. Not only were there several long streaks, but one time our coin settled on the side. So it was a good lesson.
I think that difference between ratings can be explained by level of people. For example I have 1800-1900-1950 rating. I have it because 20+ years ago I used to play chess and used to be a chess player. Today I’m no so good at chess as many years ago, so my rating is lower and I play no so well because I forgot all the openings, theory, I don’t have practice and so on. In comparison with me other guy 1800-1950 is a current young chess player. He plays every day, has classes, grows up etc. We both have the same rating right now, but it’s obvious that this guy is stronger than me and he will quickly grow up, and for me it’s my maximum rating ever.
as a 1600 FIDE rating player that have played guys from 1400 to 2200 fide rating range, playing someone who is 200-300 points better than you is a big psychological pressure.. just this thing could influence the whole streak so much..
At this point Kramnik is far beyond being "wrong", he is operating in bad faith with malice.
Like magnus and Hikaru did
@@youtubecensoringcomments7427 my guy you havent been in the competitive chess environment for a long time have you.. those cases were completely different, with one there was evidence of cheating and alot of top players knew that and for the other there is litterally nothing to even hint to the fact
@@goldenpaperyt9763 There was no evidence for cheating, just Magnus being butthurt.
@@VirideSoryuLangley just look it up..
@@goldenpaperyt9763 I think you were not able to process what I wrote, pause and think. Magnus and Hikaru did accuse hans niemann of cheating at the sinquefield cup without proof. Still to this day there is no proof... The rest is your opinion
Very few Super Grandmaster are usually playing Online as frequently as those IM Players. So, Hikaru does not have another choice but to pick what he can from the pool. Exposing the opening preparation is a risk for Fide tournament players; so, they are avoiding each other online even if they are playing in the same platform. If all top 100 players were playing online against each other, with the same level of dedication as Hikaru, the Online rating would be more in parallel with the FIDE rating.
Happy that Fabi changed his jacket😊
Those weaker opponents are paying Hikaru for his instructions with rating points. That's why you don't see any of those players supporting Kramnik's (non-)claim of fraud.
I wrote a simple python script to take in the ratings of 2 players and simulate 1000 games between them. Player A was 3200 and B was 2800 in my test. The highest win streak for player A was 41 and this was only from a sample of 1000 games. Of course this ignores a lot of factors but statistically winning 40+ games in a row is well within the bounds of reality.
He also didn’t win 40+ games in a row. There was a draw in the middle
@@pianochess1882 regardless, if I sampled 40 games enough times I’d get a period just like the one he had. I don’t know what the rating difference was but there is a 2% of winning 40 consecutive games with a +400 rating. That ignores how triggered a player would get losing so many games too.
I completely agree faster time control is a different skill. My rating difference against my friend in blitz is +500 but in lightning he is +200 Increment and time allowed to think can hugely make a difference between performannce and overall results..
Fabi does a great job of breaking down the flaws in kramnik's analysis...the whole idea behind using ELO-based statistics is not a good approach.
It seems the way GMs sniff out cheaters is that they play difficult moves too quickly and spend too much time on easy moves, and in general have bizarre clock usage. Kramnik is just barking up the wrong tree
Makes total sense. Increase volatility in the game then you also increase volin the results.
We need auto pairing to prove who is best like ICC used to do
Dear Vladimir... take a break, spend time with loved ones, enjoy time with friends and family... look after yourself. Life is short... enjoy it. Relax.
Good recap but the voice quality was low. There was an echo, as if you guys speak from the bathroom :)
kramnik going insane and paranoid the same way fischer did
Well, he hasn't gone for anti-Semitic conspiracy theories...yet.
I think everyone including Hikaru has great respect for Kramnik and his accomplishments. However , his crusade against Hikaru is obviously motivated by his personal dislike for him and he is doing nothing more than losing all credibility.
I don't see how that being his motivation is obvious. There doesn't appear to be any kind of consensus opinion that that is the case, which there should be, if it was obvious.
As far as I can tell, he's on a crusade against cheating, and thinks he found something. Right or wrong in his accusations, his motivation seems... obvious. Hehe:)
@@drunkrtard Perhaps but his insistence that Hikaru’s “record” needs to be checked thoroughly because there are statistical anomalies tells a different story.
@@drunkrtard I’ll amend my statement that it seems obvious to me anyway.
I don't know, Kramnik has been on an anticheater crusade for a while now, and sometimes I wonder if he is dumb enough that he tought people wouldn't realize that one accusation was against Hikaru.
@@dan1e1473 Yeah I don’t know. He keeps saying that according to his “mathematician friends “ the statistics are bad but he not once names them or posts any of their actual proof. Meanwhile, serious statisticians have come out regularly in support of Hikaru and yet Kramnik keeps beating the same drum. If he’s not out to get Hikaru then why is he doing this? Everyone knows there is cheating going on in online chess but to call out Hikaru like that for no good reason and no proof is “very interesting” to say the least. I’d like to believe his motivations are pure but at what point will he admit he was wrong about Hikaru? I don’t believe he ever will.
imagine how many of hikaru's streaks were broken by cheaters.
The statistical term you are looking for in this video is "selection bias".
7:55
It's always wise to seek refuge in a basement during an earthquake. 😂
Fabi, can you speak to the difference between paranoid suspicions of cheating resulting from delusions of your own strength, to the confidence you have when playing an actual game where you feel strongly a person cheated.
I,v seen that too in person 15 years ago Hikaru 30 sec against 5min GM even match
Unless FIDE add anti-farming to chess rating calculation to mitigate playing a weaker opponent repeatedly and encourage playing stronger players
I remember several years ago when TT started becoming very popular and I thought to myself (as Kramnik would say) - its very interesting how there are so many online GMs in these events who are posing challenges to the top GMs in the world, such as Magnus, Hikaru, Fabiano, Ding, Nepomniachtchi, Wesley, etc. Players who would get absolutely smoked OTB, but online they are suddenly playing at a 3000+ level... And we dont even need statistics to realize what is going on, just look at the top OTB tournaments in the world. Its always the same top GMs (as it should be). This includes the rapid and blitz world championships where all of these online "comets" are nowhere to be found. Im glad Kramnik has brought a lot of needed attention to this topic.
As any and all of these top players, who are still in touch with reality would say, any of these strong players has a chance in a one off game against these guys. Some of them don't have the chance to appear more often OTB because of lack of invitations or means of travel etc. There are problems with the way top chess has been, where it's always the same players who then are in a rating bubble just exchanging rating with the same guys. Online chess removes this bubble, where now these weaker but still strong players can take a game off of them once in a while
To compare OTB games and online games is a mistake. I mean premove can’t be done OTB and Mr Kramnik looks like he doesn’t understand that, he doesn’t understand esports or online when he comes to chess. Games that Mr Nakamura won are not always winning positions but more about how fast he plays and how fast he premoves and put pressure. So you can flag someone with a losing position. Mr Kramnik uses fallacious arguments against Mr Nakamura by intentionally saying that Mr Nakamura won 46+ games with good positions.
I don't know if I'm the only one, but after I discovered your channel, I just binged everything 😂
I would investigate ALL of Nakamura's latest results, say, from the last two years, not only at online, but at over the board (classical) chess too.
There's something odd about this latest "comeback" of his in classical, especially when you think of his age, and that it was never really his forte...
We haven’t heard anything from Hikaru’s opponents in this saga! Somebody should interview them 😁
Danya has won Titled Tuesday but I know the point they're making is his mouse skills on 3:0 boosts his rating
As far as the “farming,” I assume Hikaru is also trying to make good content, because he takes his role as a streamer at least as seriously, if not more seriously, than that of a chess grandmaster. He will be inclined to pick games that make for good content. There is a reason why he is so successful at streaming.
It was crazy to me that Kramnik tried to imply that farming was unethical. These are casual games that were agreed upon by all parties, no prize money on the line, under an online rating system that isn't particularly meaningful, at least not to the extent FIDE ratings are. Playing for fun/content is fine.
It almost sounds like context and nuance are important when interpreting statistical significance. Who knew.
3:40 Kramnik said it's not about streaks, it's about anomalous increases in rating. he doesn't understand why his critics focus on streaks that much, Hikaru's streaks are not his point
Love to hear Fabi talking and I'm not even fluent in enfkish
At the end Christian had enough of fabis yapping and wrapped it up 😂
Some specific key words that Fabiano uses shows that he has studied some math and statistics and he is not blindly commenting on this matter....weighted average,small subset,etc.
I feel like Vlad is forgetting / underestimating that chess is a psychological game, the impact your mind state can have when playing chess is massive. If you only play 1 game at a time against different players then the effect this would have on statistical variance would be fairly minimal. However when you play a large set of games in a short space of time with fast time controls and are repeatedly playing the same player / players, the effect of psychology on your performance is enormous.
Boys… Let’s get the mic situation under control
"Resistance decreases as you continue to get beat up" I call that bullying Christian. LOL.
I really dont think theres much point debating with him at this point tbh, hes not going to concede
lol Kramnik goes after Fabi of all people, probably the most respected and classiest person in chess.
Fabi has nothing to defend from that episode of the podcast.
He was correct.
Dania probably also use some assistent lol
Tip: put your lav-mics as close to your mouths as possible. Sound is quite bad in this video.
What Kramnik is proving is that being good at chess doesn't correspond to being good at math.
Very good response; thoughtful & sincere.
Dudes fabi is such a good guy.
thanks for being the level heads chess needs!