It's obviously fair to check every suspicious player apart from their communal status / name. That being said, I personally do not think Hikarus performance is suspicous to an extend that it needs to be checked seperately. I feel like Krammnik's analyses miss the point of flagging - which you mentioned. Performance level 3300+ sounds insane, but surely is achievable by the world's best, if they can simply win a sloppy game on time thus increasing their performance level.
An Elo performance much higher than your current Elo for a 10-game tournament happens, and is expected to happen often enough. A 3600 performance for 55 games is statistically not going to happen. To calculate a Performance rating you determine the Elo that results in win/loss/draw percentages that would indicate a 50-50 chance of such an outcome. He would have to win a sloppy game on time many times, and absolutely none of his opponents (for any game) can win or tie on time, opponents cannot have a great day or even a single game, Hikaru cannot have a mouse slip... If absolutely every and all planets are aligned in Hikaru's favor, then maybe. Not saying he cheated. But it is like having the top professional baseball win streak go from 26 to 36, then that team follows up with a 40 soon afterwards. Sure, Hikaru is farming, can have a good day, and any other excuse we can add to push all parameters in his favor, but statistically these events should be absurdly rare and certainly not one after the other. My explanation would include the possibility that his bliss Elo is higher than reported (and maybe we see his rating continue to climb) or that blitz draws are less common and we need new win/loss/draw percentage probabilities.
Kramnik's chart does not correlate with his words that he won those games in just November. Further upon checking the 55 game streak was made mid September 😂. That's almost like 3 months apart from today. He said November, gave the chart that claimed the result upto October, and used the games from September and before
It's also irrelevant, and a red herring to even discuss. There is nothing about Hikaru's gameplay (moves) to suggest cheating. This is entirely a meta accusation and it should not be entertained.
@@paulm6737 kramnik has kind of a mind where he thinks every world champion is on the same level. He can't accept that today's IMs and FMs can defeat previous world champions
Hi, I'm a professional mathematician (prof. of mathematical physics at University of Camerino, Italy), and a big fan of Fabi and of your content 😊 I wanted to stress what seems to me like the main point, here. As far as I know, Kramnik didn't hire a statistician or a mathematician to work on (and most importantly, TAKE RESPONSIBILITY OF) his "analysis". This is simply inconceivable. You cannot say that "mathematics" (he rather means statistics, btw) is accusing someone without being competent in the matter. The first and sufficient reason to disregard his claims or "requests" is that they're not backed up by an expert opinion and reputation, so that he can say whatever he wants and still, in the end, claim everything was done in good faith as he's not an expert. He cites some anonymous "friends mathematicians" who said that Hikaru's streaks had a "very low probability". Possibly they didn't know how many blitz games he plays online regularly. Otherwise, I really want to know who they are and what their credentials are 🤨 As for the matter itself: suppose Hikaru has some 90%, or 9 in 10, chances of winning against an opponent (which is realistic based on rating if the opponents are in the range covered by those statistics, I think). Now take a D&D dice with 10 faces. You can think of a "1-9" result as "Hikaru" wins and of a "10" result as "Hikaru doesn't win". Toss 10000 times your dice. If you DO NOT have some very long streaks of "not 10" results, the dice is probably unfair.
Just asked chat gpt on this probability and it says probability to have 50 streak from 10000 with 90% win chance is 0.0052. Which is quite low I would say.
@@pilot3344 which question did you ask exactly? I mean the probability of having a streak of “non 10”, not of having a streak of 10s. Also, I didn’t commit to a specific length. However, sticking to your number, there’s around 0.5% , or 1 out of 200, chances of having 50 consecutive “non 10” results. Now, how many streaks of 50 tosses do you have in 10000? 😏
@@alessandrodellacorte1737 yes, I think chat gpt just calculated (0.9)^50, so this is probability to have 50 wins in only 50 games. You are right, if we take 10000, the probability will be almost 100%. I wrote a script in python to simulate the same to get the real data. with 90% win chance we get 99% probability of having at least one 50 streak from 10000 games, but with 80% probability of win in every game there is only 3% probability of having such a streak. With 75% we already get 0.17%. So the question is how the probability of winning in 1 game is calculated by Kramnik. This is the crucial question. This can answer why we have such a difference in streaks between Carlsen and Nakamura. Just slightly worse opponents increase probability to have streak a lot.
lol dont argue w that guy hes trying to say what ChatGPT says instead of, yknow, using math???
10 місяців тому+1
The thing is tho this isnt even rolling the dice its a skill game in which if you are better than someone youre just better and can beat them consistantly
@@blender_wiki -- I'd probably sign yours as well as spnhm34's. The Internet makes way too much hay out of people's mistakes and problems. Conflict sells; gossip is addictive; the Internet enables the addiction.
@@blender_wiki -- btw, I think your autocorrect shot you in the foot. You probably meant 'sick', not 'seek'. Otherwise, the secret passphrase in your comment has just triggered a CIA op deep in the Mariana Trench.
Kramnik pretty much says such a streak is mathematically impossible. I bet he would think Usain Bolt couldnt ever win 10 races in a row either. Maybe he forgot that skill is important in the game and if you are higher rated then you are likely to win. I mostly play against my dad for example, so I would say it is really unlikely that he is cheating. He had a 11 game winning streak again me and 27 games streak without losing. While we both dont really play on chesscom against other players, so our rating isnt all that relevant, we have a relatively similar rating in tactics and our stats are 17/9/66. This means he lost 17/92 games, which translates to a 81.5% chance of not losing. This would overall put the percentage of him not losing 27 games in a row at 0.4% I am sure according to Kramnik that would clearly mean my own dad is cheating against me despite being a better player anyway in completely meaningless games. This is the so called "Math" that he is reffering to. It makes 0 sense because he is simply ignoring all external factors. Even if you only consider that players can tilt or their daily performance might vary even just 50-100 rating points for whatever reason, like not drinking enough or whatnot then that alone shows that these cherrypicked datapoints are completely meaningless.
Magnus = 10 streaks of 20+ wins in 1400 games Nake = 34 steaks of 20+ wins in 3800 games Actually its pretty proportional especially considering that Magnus is playing more difficult opponents in Titled Tue and Naka is farming weak players for content
Fair enough. I wish we had a break down of the average ELO of opponents of Naka and Magnus. Also would love to know the % of each’s game that is 3+0 and 3+1
Many years ago I played an IM OTB. There was a 700 elo difference in our strength. I did not even come close to drawing one in twenty games. Kramnik is suspicious because he thinks his strength is close to Naka’s. Nah. In classical perhaps, but not 3 or 1 min flat. Naka is a god in those time controls. Naka will adopt him easily.
With his empty allegation, Kramnik owes us an explanation of how he alleges Hikaru did this. If this is another case of anal beads, Kramnik has hurt the cause of rooting out cheating. Notwithstanding that Nieman was an admitted cheater, there is absolutely no evidence, nor even reason, to believe Hikaru has cheated.
Hey guys, love the podcast and I'm always up to buy merch from my favorite content creators but I really need tall sizes (LT, XLT, 2XLT, etc.) For some reason, t-shirt makers have been making the XL sizes shorter and now the only sizes long enough for me are absurd (like 3XL.) Add tall sizes and I promise I'll buy a couple - I'm sure I'm not the only one.
Everything about "it's coming from mathematics" is probably one of the most cowardly dodges (and ineffective) I've seen in a while. Sucks that he thinks it will fool the world.
It's pretty sad to me to see a great player like Kramnik besmirching himself by behaving like a dog with a bone. As he digs in deeper with his defamatory allegations, he twists himself up in ever more disingenuous claims like "it's not me making the accusation, it's the mathematics...," etc. As someone who was falsely accused of cheating himself, he should be more sensitive to the effect of his actions.
I know some kids in the 2000-2200 FIDE range wo are around 2700 bullet online. They are just incredibly fast, but wouldn't stand a chance against Hikaru, while a slower GM who is 2700 in bullet may.
The game is "Can you checkmate Hikaru's instant autopilot when you only have 3 minutes to think?" Hikaru will always have a 99.999% ("five nines") performance against those that cant unless he willingly gambles.
Yay merch!! You guys did a fantastic job curating a very wearable collection. Can’t wait to get my order!! 🤍 I’ve seen the recent back and forth responses but _really_ didn’t want to spend any time reading/watching it myself. Thanks for taking one for team and doing this video so that now I don’t have to!! 😂
If any GM wants to score 100 out of 100, just play me! The serious point: Some of these streaks involve Hikaru beating up on the same hapless person 10 or more straight games.
Why to bring up multiple times that streamers are not using their phones to cheat, it is not a point. It is obvious that it can be don over the headphones.
Cheaters have so many ways to deceive themselves into thinking they are doing the right thing or something okay. It is amazing what the human mind is capable of in terms of self-deception and bigotry. Also blaming a ~3300 rated player for a winning streak with a 3450 performance does not quite show a deep understanding of statistics. For example, he completely ignores how many total games Hikaru plays. In tens of thousands of games you will always be able to cherry-pick an outlier. Even if it is a one in ten thousand chance to happen, if he plays 10000 games it should happen with a chance of 50%. I could take Kramnik seriously much more easily if he actually laid out the statistics instead of just making bold claims based on a few cherry-picked data points.
I do not cheat when playing chess, so I don't know how these apps work, but I am an IT guy and know that if there is an app that can analyse position from the picture then you do not need a phone to take that picture, you can use other means, eg. a screen capture running in the background and then feed the picture to the app which can parse the position and send it to the engine. From there it will be very easy to set some way the engine talks to you to advise you the best move. I believe that many players know this and use it, some on the top level too, not that I know it for sure but I guess it must be happening if its so easy and some players salaries get significantly increased by using it.
Kramnik doesn't understand what probability really means. And neither do most people. Especially when it involves thousands of games we are talking about. There's nothing reasonable about Kramnik his 'request'. Hans Niemann wasn't checked just because it was demanded. The chess platform got sued and had to defend its decisions.
These comparisons are ridiculous. Hikaru entertains his audience by beating up on weaker players by the dozens. He makes no bones about that. And the difference between increment and no increment is enormous
Anyone recall when he went 27-2 against Hou Yifan in a tournament about 5 years ago. Im not even a fan of naka, but the guy has done this for decades. Kramnik is a moron for doing this. If weve gotten to the point where a former world champion is accusing a top 3 in the world player of cheating, the game is no longer worth playing on any level.
Hikaru is being used as a sacrificial lamb 🐑 by Kramnik to bring more attention to cheating online. Niemann was low hanging fruit because of his admitted cheating in the past. But Hikaru is probably the most famous player. Hikaru could also be credited as being a big factor for chess having a resurgence of interest from the average Joe because of his streaming success. What better person to use for bringing more attention to your cause. Sad really
my issue with kramnik's suspicions is that he is not giving any information regarding the accuracy of Hikaru's games. If Hikaru is indeed using engines (which I am sure he is not), his moves should be really accurate (>90% atleast). Just accusing him based on the number of games he has won in a streak is baseless. Hikaru is making more than enough money out of streaming for him to actually start using engines to beat random lower rated FMs and IMs.
We keep record books to celebrate exceptional achievements. By Kramnik's own logic, every single record in sports is "interesting" and requires investigation. It's on the level of thought-police.
@@scottekoontz Kramnik has repeatedly dismissed career mathematicians and statisticians who present alternative interpretations of Hikaru's chess data, and he's done so without offering any evidence to counter their assertions. Perhaps that is the flaw in Kramnik's understanding of statistics that danh72 is referencing.
@@theKashConnoisseur "repeatedly dismissed career mathematicians and statisticians who present alternative interpretations of Hikaru's chess data" And I have yet to see one of them post the parameters. Agreed, Kramnik has not posted his either, which is why we are not certain he is wrong. If you have access to any of these career mathematicians' comments then please post them, and thanks in advance. The math is not that hard after the parameters are known. I have seen one post where a "mathematician" claims a 1% chance of an event, and extrapolated from there, but it's all BS if we don't have the win% and number of trials used at each point.
@@scottekoontz My understanding is that Kramnik has deleted the comments of aforementioned experts from Reddit and from his blog, claiming that they are misleading. However, there are archived screenshots of said comments, as well as writeups posted to Github and elsewhere. Unfortunately, UA-cam does not allow the posting of links offsite, and so I'll have to bid you to do your own research. Nakamura has posted video content on his channel discussing some, and Google should be helpful to you as well. Also, I believe the number of trials (being the totality of Hikaru's online chess career) is 35,000 plus. Cheers.
@@scottekoontz I think that the debunkers (against Kramnik) have failed. I think that the key issue is not simply knowing the parameters, it is knowing what to investigate. For example, is Kramnik concerned that Hikaru is cheating online all the time, or only during his streaks, or only when extending a streak, or ...? If someone gives me a dataset and tells me their thoughts about it and what they want me to help them with then I can often help them, but if they simply ask me to look for anything suspicious then I tell them that I need a clearer task specification.
I think the main reason for the discrepancy is the flagging. I wonder how many of those streaks are 3+2. The idea that this isn't an accusation is just silly.
Hypothetically, Hikaru’s oppenents were habitually cheating and thus overrated but chose not to cheat against Hikaru, resulting in Hikraru getting easier games that expected by rating. Hypothetically Hikaru’s opponents were having an off day and didn’t play up to par during these streaks. There’s lots of explanations for these streaks that aren’t cheating. But look at the game quality and it should be clear
"Only with a very clear mathematical explanation..." You have been given a very clear mathematical explanation, but you rejected it, deleted the comment from your blog, and ad hominem'd the guy who made the comment. Kramnik is an unhinged lunatic with a grudge. People tell him you can't cherry pick streaks from a large dataset -- he ignores them or calls them names. People explain tilt and farming to him -- he ignores the new knowledge and retreats back to cherry-picked streaks and expected win percentage ^ number of games in streak. The longer this goes on, the more two things get damaged: Kramnik's reputation, and the reputation of anti-cheating advocates writ large. I couldn't care less about the former given the damage that he's doing to the latter.
I love your podcast, and I understand the need to cover this subject given the attention it’s received, but trying to present Kramnik’s lunacy in a neutral fashion not only does a disservice to your audience but it’s also not a good look. This is not complicated. Kramnik not only doesn’t understand how probabilities work, but has silenced everyone who has tried to explain things to him. You should cover this especially. The guy is embarrassing himself and perhaps even tarnishing the image of what it means to be a top chess player, let alone a world champion. One certainly wouldn’t expect such an idiot to reach the pinnacle in any field.
I think you're right but for the wrong reason. Elo is based off of the win probability. I don't think it's really designed for being flagged. That's a really good observation.
Ppl have especially mentioned this in Bullet with 1+0 vs 1+1 which is a complete other dimension for top players. In 1+1 ppl like Alireza can basically chill forever while in 1+0 if youre 10-15 seconds behind youre just fucked if your opponent doesnt blunder a quick mate.
I've never liked Hikaru's stream -- easily the worst stream in chess. He is chaotic, and talks too much (without being helpful). His latest STUNT (no other word for it) is the icing on the cake. SUMMARY: He'll do anything for clicks.
Hey Christian & Fabiano, I tested one of your hypothesis in a recent blog article covered by Hikaru & Gotham. Would love to hear your thoughts. Please let me know if I can clarify any points.
Non-statisticans who wade in deep, always do seem to get in over their heads...same here. Frankly his citing Nakamura - even if sometimes he qualifies that by saying he's not accusing Nakamura, 'the stats are' - is ironically the biggest argument AGAINST usage of his stats that one can think of. This is what former world champions do these day - invent chess variants and accuse people of things? Just stick with 'no castle' chess, Vlad.
I do rarely write comments but this topic is just absurd. The main arguments still seem to be the isolated best winning streaks that are taken out of huge sample sizes of games (while ignoring the latter). The Elo differences between online and irl are mostly not accounted for, aswell as the flagging from worse positions. What's also just ignored are psychological components of playing vs Hikaru in front of a big audience in the first place. That's something that he is surely more used to than the avg opponent. Also tilt is not factored in. A player that's say 0-5 down & still continues on to play vs Hikaru will not always perform at their peak rating. With online chess there might even be stuff like time zone differences involved. An opponent that plays at their midnight vs Hikaru at an early evening will perform much worse just because of that (I know that from other online games myself).
And then there is lagging/connection problems, mouseslips....what also goes into your psychological component argument is that players see Hikaru and the opportunity to get huge rating points from him and might deliberately enter sharper lines than they would like too just because the off chance of winning justifies the additional risk (like going all in with a shitty hand in poker if theres already a huge pot, you can win much more than you can lose).
The problem is if Hikaru plays a lot more games, he will have longer streaks just by chance. Also he could have longer losing streaks or periods of underperformance that balance out the winning streaks. So you can't just look at winning streaks alone.
That's not the problem. The problem is Kramnik doesn't know even high school level math. But for some reason he thinks people should listen to him because he used to be good at chess.
Hikaru had streaks like this five times in November already. Even though Kramnik isn’t solid on his knowledge, you can’t deny there’s lots of cheating. He’s trading his authority and his good name to spotlight the problem. Maybe the trade is suboptimal, but someone has to do something. Ok, Hikaru isn’t cheating, let’s test him first, because of the numbers, says VBK
@@mikemerinoffyou can easily do that by looking up the individual games and seeing how many he had losing positions in and won either by flagging or the opponent blundering either due to time trouble or because he’s playing a top 2 speed chess player.. If the trade off isn’t the best then why not use your time in a more optimal matter looking into suspicious games and non-human moves instead?
@@EmanCipated919 As crazy as it may sound, Hikaru's accuracy per move isn't even that high. And he often plays moves outside of theory or in a different order. There are no non-human moves in his play. Not to mention how you can literally see him think about his moves on live streams. And Hikaru blunders too.
you really want to compair fischers most outstanding performance with nakamuras 2 to 3 times longer streaks which he repeated for several times? who the heck is nakamura compaired to fischer? nakamura is even not a chessplayer. he is just a streamer ;)
How many online games has Hikaru played in the last year? How many games does the average GM play? Are the numbers inflated/affected by the fact that Hikaru is a wildly popular streamer and plays ~5 days a week as his primary source of income? Naroditsky also plays a ton of blitz online...does he have any of these "unbelievable" win streaks? I'm not a statistician, but it looks like Kramnik is cherry picking small sets of data and not looking at the big picture.
Naka has a lot of winning steaks because he plays much more than other top players and also is the only one to play lower rated players. Magnus would never play 55 games against lower rated players
At the level of chess these guys are playing im not surprised if they crush lots of lower rated gms. They are super gms for a reason. A hundred to couple hundred points is leaps and bounds
Agreed. The gap betwen 3k and 2500 is night and day. Sure, they can win against a 3000, but i wouldnt be suprised if magnus or hikaru goes off and beat them 50 in a row
One of my favorite examples of why he’s not cheating is when Hikaru is down to like 10 seconds and he’s simultaneously hyperfocused and explaining his moves at a crazy rate while making errors and excellent moves to tie or win.
@@martinpaddlethere are chessbot softwares that play the game for you. All you have to do us sit in front of a camera. I'm not saying Hikary does it, but there are a lot of ways one can cheat.
@@milankotevski1663yea even if engine is playing moves for you it’s even harder to have a coherent thought on what moves he wants and planning to play as computer moves can be very random
This doesn't say anything at all. The cheating at this level would consist of just sometimes becoming aware of the sudden changes in evaluation, that's it. Like have your phone vibrate one way when it drops and another way when it increases. All the strategies and tactics are yours. Obviously Hikaru is the one making the moves and obviously he's among the top players, but cheating like this is exactly what would allow a person to play quicker and won't preclude making mistakes at all - just help win more often
Vlad pointing out Naka's performances as being an extreme outlier when Naka is perhaps the most extreme outlier in online chess history is extremely funny. Vlad: "wow 55 out of 55 that's pretty strange" Uhh yeah man we've been saying that for the last two decades lol
@@akashsinha2880 Investigation of what? There's absolutely nothing there. Anyone who has any understanding of statistics is laughing at this. It's like saying I woke up this morning. I should be investigated for being alive. My streak of being alive is unbelievable. So many days in a row.
Utterly embarrassing for kramnik. He's exposing himself to be utterly clueless about how elo ratings actually work. They aren't a pure mathematical construct, they are only rough approximations of how strong a player *may be*. In practice they are heavily flawed, and drawing statistical inferences is not accurate. Also he doesn't consider time control which is a massive factor. He literally doesn't seem to mention it at all as far as I can see.
Kramnik holds the record of classical games without a loss at something like 82 games. Seems like a super unlikely event. I bet people would find that interesting.
@@KyleMc16 i thought Ding had like over 100 games without a loss at some point. perhaps vlad was the prev record holder. both should be investigated though 😁
Kramnik is actually harming his own course by going after Hikaru like this, I think all the real cheaters are very happy with what he’s doing right now
Can anyone explain to me how it would even be possible for Hikaru to cheat, while streaming, while explaining each move, while eating? He would need a second computer and a couple of spare hands, right?
He could have a beepingsound that indicates the evalbar on his headphones or signals him when there is a tactic oportunity. It is possible but he misses so many oportunities that even if it is possible his play suggests otherwise.
@@Monika77ful but it would still be impossible since if you watch his first response video for kramnik when he show us 3 hours of that match, the only reason hikaru managed to outplayed his opponents simply because they failed to punish him hard when he blunder. If he played againts someone like fabi, magnus or alireza then hikaru would lost many of that match.
@@lordtouchme77 I came to the same conclusion that hikaru is not cheating. But it is possible to cheat. Statistics alone will not catch cheaters. A great cheating detection instinct is found in magicians. And you would have to combine that with Mathknowledge and Chessknowledge of the highest order. We will probably never catch good cheaters. Again I dont think Hikaru cheats.
@@pizzapimp8128 not nesseceraly you could have a Program doing that. But another possibility is that the opponents of Hikaru loose on purpose. Like They farm lower rated players and then sell losses to him. Again theoretical possible but I highly dought it. Because my explanaition is wildly constructed with multiple cheater (Hikaru and the people loosing to him). If you watch Hikarus latest content his explanaition is much better for not cheating. So if we use arcoms razor that means he is not cheating. Again will always be possible but in his case extremy unlikely. He is not cheating.
Does one of the following statements have to be false? 1. Hikaru Nakamura does not cheat on Titled Tuesday. 2. There is a lot of cheating on Titled Tuesday. 3. When players who cheat play Hikaru Nakamura on Titled Tuesday they carry on cheating. 4. Hikaru Nakamura frequently wins Titled Tuesday.
Some key points the high school dropout misses: 1 - randomness: if your n (number of games) is high enough the probability of streaks like this increase 2 - elo difference, too lazy to do the math but his opponents were not Magnus and Fabi but IM/FM level and therefore the probability to beat them is high 3 - tilt/psychology factor: no explanation needed, (PTSD haha) 4 - low number of unique opponents, therefore not truly 46 unique games from a math perspective, not independent games but a couple of grouped games 5 - Bayesian statistics - dependent probability (one of the more difficult subjects in stats). Everything taken into account the probability of Hikaru destroying 4/5 IM/FM level players ain’t that crazy 6 - sampling bias - selectively choosing a streak to prove a point. Imagine looking at 46 votes to determine an ellection 7 - using annoying Russian double speak, no scientific argument but my Sovjet side of the family has thought me well how to spot this bull 💩 Stay in school kids!
Not only does Kramnik misrepresent basic statistical concepts, he seems to be completely unaware of online dynamics I'd expect literally every 1500 to be aware of. Opponent selection is one of the most obvious facets in online play. All of us have faced players at or above our rating that we feel we have a significant advantage over, and conversely players at our rating that are utterly terrifying. We all do a mental calculation whether rematching is in our favor. Time controls also drastically change ability/advantage. Hikaru is a speed specialist. The less time or lack of increment massively boosts his advantage. Hikaru is undisputedly the best or second best bullet player in the world.
It takes a very experienced statistician to design a meaningful statistical test or analysis. Kramnik’s results may be interesting, but they simply are not conclusive and have no value. Please leave statistics to the real statisticians.
@@th3cup1d92 some of the best in every game or sport have cheated. So if he does it wouldn't surprise me. If it wouldn't be easy to cheat then maybe...but it's super easy to cheat so it's likely one or a few top players are cheating.
In the digits of Pi there will be a sequence of 10000 consecutive 9's, the same happens when you play a lot of games, at some point you will have a streak of 40 won games. The chance increases when you rematch a player that gets tilted.
Simple solution: Have a Nakamura - Kramnik match with the two playing 3+0 for 46 games. Conditions: Both being on a computer, in the same room, no headphones and with an arbiter next to each of them.
I noticed that this morning, the sun rised up again. That's 1825 billion sunrises in a raw : someone is clearly f...g with us ! I'll immediately start a petition for an honest investigation.
also I think all the cheating drama is inevitable at this point, because during the pandemic a lot of people got into chess, this brought along a lot of money and where the money is involved there is always going to be cheaters and scammers, so in one hand it is good that a lot of people have taken up chess but on the other hand it is obvious that getting good at chess takes a lot of time and all the digital aids will help you push to these ratings where you can play for money, so the nikita cheater explained it perfectly that he doesnt cheat in money events but he still needs rating to get to those and so to cut time he cheats, same can be said about already established players, their competition is tougher than ever and the better the player you are the less cheating you need, for GMs I think a simple evaluation bar is enough to give a slight edge, so in a sense cheating has come to stay, it will always be there online and it will evolve to such undetectable levels that I think there is a possibility where it will become a norm to use some sort of small aid or something playing online
Kramnik is upset about cheating so he decides to go after a chess player who live streams every game he plays, has millions of people watch him play live and recorded, live discusses his thought process during each game. Kramnik, on the other hand, plays games at home with no witnesses, with no audio and visual proof he's not cheating, and doesn't broadcast/publish himself playing. Very odd times we're living in.
You just need common sense and basic understanding of maths and statistics to know that Kramnik is totally wrong. It appears to me that Kramnik did not do any proper research before making these allegations and he goes around shutting down valid statisticians comments without a proper rebuttal tells you he doesn't understand what he is doing.even GMs that doesn't share good relationships with Hikaru has come out in defense of Hikaru speak volumes.
Kramnik should look at the players with overinflated online ratings whose OTB rating are widely different - like hundreds and hundreds of difference. Nakamura has played at a very high level OTB and online AND has played against the top players like Carlssen so he has been under great scrutiny and no one doubts his strength.
C'mon. This is accusing plain and simple. I appreciate your balanced stance on this but let's be frank. We can all see this for what it is. You are right in saying that Kramnik is handling this very badly. These two points are without doubt.
This reminds me of RNG discussion for video games. People just don't have an intuitive understanding of statistics. Streaks WILL happen. The more games you play, the longer the streaks that are possible. Flip a coin 4 times. What are the odds of all heads? 1/16 or 6%. If you add one more flip, how many streaks of 5+ heads can you get? Well, the middle 3 flips must still be heads, so that's 1/8. But now either the first or last flip can be heads. The odds of that are 3/4. Combine that with 1/8 (1/8 * 3/4) = 3/32 or 9%. The odds jumped dramatically. So if you look at a streak by itself, it looks incredible. But if someone plays a lot of games, it's expected. Kramnik couldn't understand statistics even if his life depended on it. He has no clue what he's talking about.
We can still make no assumptions regarding what Kramnik is referring to because he has revealed none of the parameters used. The stats math is easy once you are given the Elo (resolves win%, loss%, draw%) and the number of successive wins, and in some cases an imbedded draw. If someone plays a lot of games (2,000? 5,000? 10,000?), then some streaks are still statistical rarities depending on what you use as a win probability. Odds may jump "dramatically" from nil to nearly nil.
i have done the maths, 33400 blitz games played, average opponent rating 2800, all time no loss rate 88%(win + draw) probability of a streak of 45.5 happening is 99.7%@@scottekoontz
@@scottekoontz Kramnik can't reveal anything. He doesn't understand what he's talking about. All he's done so far is say "in this streak, Hikaru has won these games, so he's playing like an N rated player" That makes no sense. He's taking the streaks in isolation and you can't do that. I explained this in my original comment.
@@alienrenders "in this streak, Hikaru has won these games, so he's playing like an N rated player" You can make such a statement, and his results appear to be correct. A factual statistic bothers you because you want to see results across many trials (see below) but that does not negate the work shown. Too many people do not understand statistics for many trials. Hikaru's streaks would be rare or impossible even after 3,000 trials. Some after 10,000 trials. As I mentioned, how many games are you talking about? Since the streaks are all contained within a two-months span, using 5,000 is being very generous, 15,000 going overboard. Elo difference: 350 p result: 0.787 win based on Elo math Desired Outcome: 45.5/46 One trial (start now, end after 46 games): Probability ≈ 0.00024 or 1 in 4,000 (not 1%) 2,600 trials (play 10x a day, every weekday, for a year): 9.5% 10,000 trials: 33% (i.e., play 38 games every weekday for a year and maybe this would happen) ----------------------------------------------- Elo difference: 350 p result: 0.787 win based on Elo math Desired Outcome: 55/55 One trial (start now, end after 55 games): Probability ≈ 0, not going to happen. 2,600 trials (play 10x a day, every weekday, for a year): 0.1%, very highly unlikely 15,000 trials: 0.6%
Is there a chance Kramnik is trolling everybody that he KNOWS Hikaru did not cheat ... but to bring this topic/issue on everybody's radar (including dozens of super big YT/Twitch streamers) with a potential reach on 10s of millions ... Then when all is calmed down again "new online chess rules" are made? ... Thus, "mission accomplished" ?
Sorry but Kramnik just sounds like some random conspiracy theorist. "I'm just asking questions.", "I'm just stating the facts not implying anything", ... blah blah. And underlying it with superficial data which on a quick glance might look like obvious. Bad style and by far not detailed or specic enough for any serious discussion.
The most funny thing is that Kramnik talks about "Hikaru's bots" while actually HE is the one supported by a "bot", with a username "Rimeeeeee" from Russia (and you can see the flag). I'd say even more - this user's posts are no "interesting" and nonsense that I suspect it's Kramnik himself.
I view this as Magnus and Nakamura are generational type players. Similar to like Tiger Woods in golf or Novak Djokovic in Tennis. I bet even Bobby Fischer or Garry Kasparov performances are very statistically unlikely during certain periods of their career. I wouldn't have thought I would have ever seen people dominate a sport like these players before they arrived on the scene. What is the probability of the Buffalo Bills making it to 4 straight super bowls? What is the probability of Tom Brady winning 7 Super Bowls? Probabilities depend on the past patterns predicting the future. If we are seeing something new that hasn't existed in the past it doesn't make sense to use the past pattern to predict the future. The current elite of every sport are changing the game and setting new standards all the time on what is possible. When the sport catches up to these new standards it will take the next dominate player or players to change the level of what we thing is possible.
Very sad to see Kramnik embarrass himself like this, and destroy his legacy.
lmao Kramnik is historical person, nothin could destroy it.
Did you see Danny rensch went to Hikaru's house . Streaming studio tour 😁
It's obviously fair to check every suspicious player apart from their communal status / name.
That being said, I personally do not think Hikarus performance is suspicous to an extend that it needs to be checked seperately.
I feel like Krammnik's analyses miss the point of flagging - which you mentioned.
Performance level 3300+ sounds insane, but surely is achievable by the world's best, if they can simply win a sloppy game on time thus increasing their performance level.
An Elo performance much higher than your current Elo for a 10-game tournament happens, and is expected to happen often enough. A 3600 performance for 55 games is statistically not going to happen. To calculate a Performance rating you determine the Elo that results in win/loss/draw percentages that would indicate a 50-50 chance of such an outcome. He would have to win a sloppy game on time many times, and absolutely none of his opponents (for any game) can win or tie on time, opponents cannot have a great day or even a single game, Hikaru cannot have a mouse slip...
If absolutely every and all planets are aligned in Hikaru's favor, then maybe. Not saying he cheated. But it is like having the top professional baseball win streak go from 26 to 36, then that team follows up with a 40 soon afterwards. Sure, Hikaru is farming, can have a good day, and any other excuse we can add to push all parameters in his favor, but statistically these events should be absurdly rare and certainly not one after the other. My explanation would include the possibility that his bliss Elo is higher than reported (and maybe we see his rating continue to climb) or that blitz draws are less common and we need new win/loss/draw percentage probabilities.
Kramnik's chart does not correlate with his words that he won those games in just November. Further upon checking the 55 game streak was made mid September 😂.
That's almost like 3 months apart from today. He said November, gave the chart that claimed the result upto October, and used the games from September and before
It's also irrelevant, and a red herring to even discuss. There is nothing about Hikaru's gameplay (moves) to suggest cheating. This is entirely a meta accusation and it should not be entertained.
@@paulm6737 kramnik has kind of a mind where he thinks every world champion is on the same level. He can't accept that today's IMs and FMs can defeat previous world champions
Hi, I'm a professional mathematician (prof. of mathematical physics at University of Camerino, Italy), and a big fan of Fabi and of your content 😊
I wanted to stress what seems to me like the main point, here.
As far as I know, Kramnik didn't hire a statistician or a mathematician to work on (and most importantly, TAKE RESPONSIBILITY OF) his "analysis". This is simply inconceivable. You cannot say that "mathematics" (he rather means statistics, btw) is accusing someone without being competent in the matter. The first and sufficient reason to disregard his claims or "requests" is that they're not backed up by an expert opinion and reputation, so that he can say whatever he wants and still, in the end, claim everything was done in good faith as he's not an expert.
He cites some anonymous "friends mathematicians" who said that Hikaru's streaks had a "very low probability". Possibly they didn't know how many blitz games he plays online regularly. Otherwise, I really want to know who they are and what their credentials are 🤨
As for the matter itself: suppose Hikaru has some 90%, or 9 in 10, chances of winning against an opponent (which is realistic based on rating if the opponents are in the range covered by those statistics, I think). Now take a D&D dice with 10 faces. You can think of a "1-9" result as "Hikaru" wins and of a "10" result as "Hikaru doesn't win". Toss 10000 times your dice. If you DO NOT have some very long streaks of "not 10" results, the dice is probably unfair.
Just asked chat gpt on this probability and it says probability to have 50 streak from 10000 with 90% win chance is 0.0052. Which is quite low I would say.
@@pilot3344 which question did you ask exactly? I mean the probability of having a streak of “non 10”, not of having a streak of 10s. Also, I didn’t commit to a specific length.
However, sticking to your number, there’s around 0.5% , or 1 out of 200, chances of having 50 consecutive “non 10” results. Now, how many streaks of 50 tosses do you have in 10000? 😏
@@alessandrodellacorte1737 yes, I think chat gpt just calculated (0.9)^50, so this is probability to have 50 wins in only 50 games. You are right, if we take 10000, the probability will be almost 100%.
I wrote a script in python to simulate the same to get the real data. with 90% win chance we get 99% probability of having at least one 50 streak from 10000 games, but with 80% probability of win in every game there is only 3% probability of having such a streak. With 75% we already get 0.17%. So the question is how the probability of winning in 1 game is calculated by Kramnik. This is the crucial question. This can answer why we have such a difference in streaks between Carlsen and Nakamura. Just slightly worse opponents increase probability to have streak a lot.
lol dont argue w that guy hes trying to say what ChatGPT says instead of, yknow, using math???
The thing is tho this isnt even rolling the dice its a skill game in which if you are better than someone youre just better and can beat them consistantly
We need a petition to save Kramnik from himself
Who's gonna put it up. I'd sign it
Sure, I'm in. Everybody's messed somehow or other. Kramnik's seeing little green men, so let's help a fellow messed-up earthling out.
We need a petition for UA-camrs to stop speaking about a seek man
@@blender_wiki -- I'd probably sign yours as well as spnhm34's. The Internet makes way too much hay out of people's mistakes and problems. Conflict sells; gossip is addictive; the Internet enables the addiction.
@@blender_wiki -- btw, I think your autocorrect shot you in the foot. You probably meant 'sick', not 'seek'. Otherwise, the secret passphrase in your comment has just triggered a CIA op deep in the Mariana Trench.
14:45 -- So, Kramnik is stating that he isn't accusing Hikaru of cheating but it is math itself that is accusing him?
Kramnik pretty much says such a streak is mathematically impossible. I bet he would think Usain Bolt couldnt ever win 10 races in a row either. Maybe he forgot that skill is important in the game and if you are higher rated then you are likely to win. I mostly play against my dad for example, so I would say it is really unlikely that he is cheating. He had a 11 game winning streak again me and 27 games streak without losing. While we both dont really play on chesscom against other players, so our rating isnt all that relevant, we have a relatively similar rating in tactics and our stats are 17/9/66.
This means he lost 17/92 games, which translates to a 81.5% chance of not losing.
This would overall put the percentage of him not losing 27 games in a row at 0.4%
I am sure according to Kramnik that would clearly mean my own dad is cheating against me despite being a better player anyway in completely meaningless games.
This is the so called "Math" that he is reffering to. It makes 0 sense because he is simply ignoring all external factors. Even if you only consider that players can tilt or their daily performance might vary even just 50-100 rating points for whatever reason, like not drinking enough or whatnot then that alone shows that these cherrypicked datapoints are completely meaningless.
He does not want to get sued. Has to choose his words carefully
He is a coward and delusional, what do you expect?
where is fabi? i am really curious about his take on this
Magnus = 10 streaks of 20+ wins in 1400 games
Nake = 34 steaks of 20+ wins in 3800 games
Actually its pretty proportional especially considering that Magnus is playing more difficult opponents in Titled Tue and Naka is farming weak players for content
These statistics are not very representative. Because 40+ wins are much more difficult to achieve than 20+. How many Magnus had 40+?
Fair enough. I wish we had a break down of the average ELO of opponents of Naka and Magnus. Also would love to know the % of each’s game that is 3+0 and 3+1
All Kramnik is doing is making a fool of himself. Truly pathetic.
Kramnik has accused Hikaru of cheating. It's no different than a politician getting caught in a lie and saying "I misspoke."
I am guessing Fabio is not going to talking about this. Which is fair enough.
bro is off the wagon
Many years ago I played an IM OTB. There was a 700 elo difference in our strength. I did not even come close to drawing one in twenty games.
Kramnik is suspicious because he thinks his strength is close to Naka’s. Nah. In classical perhaps, but not 3 or 1 min flat. Naka is a god in those time controls. Naka will adopt him easily.
Believe me, Kramnik thinks he is MUCH better than Hikaru. But he's not, long since.
With his empty allegation, Kramnik owes us an explanation of how he alleges Hikaru did this. If this is another case of anal beads, Kramnik has hurt the cause of rooting out cheating. Notwithstanding that Nieman was an admitted cheater, there is absolutely no evidence, nor even reason, to believe Hikaru has cheated.
Hey guys, love the podcast and I'm always up to buy merch from my favorite content creators but I really need tall sizes (LT, XLT, 2XLT, etc.) For some reason, t-shirt makers have been making the XL sizes shorter and now the only sizes long enough for me are absurd (like 3XL.) Add tall sizes and I promise I'll buy a couple - I'm sure I'm not the only one.
Everything about "it's coming from mathematics" is probably one of the most cowardly dodges (and ineffective) I've seen in a while. Sucks that he thinks it will fool the world.
I think most and foremost he fooled himself.
It's pretty sad to me to see a great player like Kramnik besmirching himself by behaving like a dog with a bone. As he digs in deeper with his defamatory allegations, he twists himself up in ever more disingenuous claims like "it's not me making the accusation, it's the mathematics...," etc. As someone who was falsely accused of cheating himself, he should be more sensitive to the effect of his actions.
But it's an accusation made by poor mathematics proposed by Kramnik.
Nice video Christian, thorough, fluid and well structured
Sounds live a dude biting the fist and I love him and have so much respect for him but still standing biting the dust
I know some kids in the 2000-2200 FIDE range wo are around 2700 bullet online. They are just incredibly fast, but wouldn't stand a chance against Hikaru, while a slower GM who is 2700 in bullet may.
The game is "Can you checkmate Hikaru's instant autopilot when you only have 3 minutes to think?" Hikaru will always have a 99.999% ("five nines") performance against those that cant unless he willingly gambles.
I wonder if many of the super GMs wouldn't be in the 3200 range naturally if they were able to play 100 games a day instead of per year.
Rasian is not a nationality ,its a way of thinking
merch!!! finally!!!!!
Yay merch!! You guys did a fantastic job curating a very wearable collection. Can’t wait to get my order!! 🤍
I’ve seen the recent back and forth responses but _really_ didn’t want to spend any time reading/watching it myself. Thanks for taking one for team and doing this video so that now I don’t have to!! 😂
If any GM wants to score 100 out of 100, just play me! The serious point: Some of these streaks involve Hikaru beating up on the same hapless person 10 or more straight games.
Why to bring up multiple times that streamers are not using their phones to cheat, it is not a point. It is obvious that it can be don over the headphones.
Cheaters have so many ways to deceive themselves into thinking they are doing the right thing or something okay. It is amazing what the human mind is capable of in terms of self-deception and bigotry.
Also blaming a ~3300 rated player for a winning streak with a 3450 performance does not quite show a deep understanding of statistics. For example, he completely ignores how many total games Hikaru plays. In tens of thousands of games you will always be able to cherry-pick an outlier. Even if it is a one in ten thousand chance to happen, if he plays 10000 games it should happen with a chance of 50%. I could take Kramnik seriously much more easily if he actually laid out the statistics instead of just making bold claims based on a few cherry-picked data points.
Kramnik making youtubers and streamers rich from milking his content lmao
I do not cheat when playing chess, so I don't know how these apps work, but I am an IT guy and know that if there is an app that can analyse position from the picture then you do not need a phone to take that picture, you can use other means, eg. a screen capture running in the background and then feed the picture to the app which can parse the position and send it to the engine. From there it will be very easy to set some way the engine talks to you to advise you the best move. I believe that many players know this and use it, some on the top level too, not that I know it for sure but I guess it must be happening if its so easy and some players salaries get significantly increased by using it.
Kramnik doesn't understand what probability really means. And neither do most people. Especially when it involves thousands of games we are talking about. There's nothing reasonable about Kramnik his 'request'. Hans Niemann wasn't checked just because it was demanded. The chess platform got sued and had to defend its decisions.
These comparisons are ridiculous. Hikaru entertains his audience by beating up on weaker players by the dozens. He makes no bones about that. And the difference between increment and no increment is enormous
Anyone recall when he went 27-2 against Hou Yifan in a tournament about 5 years ago. Im not even a fan of naka, but the guy has done this for decades. Kramnik is a moron for doing this. If weve gotten to the point where a former world champion is accusing a top 3 in the world player of cheating, the game is no longer worth playing on any level.
With all due respect, who is he to say “we must have an investigation”. Go back to your cave old man…
Kramnick is the crazy Rasputin of the chess world.
Hikaru is being used as a sacrificial lamb 🐑 by Kramnik to bring more attention to cheating online. Niemann was low hanging fruit because of his admitted cheating in the past. But Hikaru is probably the most famous player. Hikaru could also be credited as being a big factor for chess having a resurgence of interest from the average Joe because of his streaming success. What better person to use for bringing more attention to your cause. Sad really
my issue with kramnik's suspicions is that he is not giving any information regarding the accuracy of Hikaru's games. If Hikaru is indeed using engines (which I am sure he is not), his moves should be really accurate (>90% atleast). Just accusing him based on the number of games he has won in a streak is baseless.
Hikaru is making more than enough money out of streaming for him to actually start using engines to beat random lower rated FMs and IMs.
We keep record books to celebrate exceptional achievements. By Kramnik's own logic, every single record in sports is "interesting" and requires investigation. It's on the level of thought-police.
It is amazing that such a brilliant chess player has the flawed understanding of statistics you would expect from a junior high school student.
Not sure what he is missing since he has not posted the starting parameters yet. What flaws can you discern at this point?
@@scottekoontz Kramnik has repeatedly dismissed career mathematicians and statisticians who present alternative interpretations of Hikaru's chess data, and he's done so without offering any evidence to counter their assertions. Perhaps that is the flaw in Kramnik's understanding of statistics that danh72 is referencing.
@@theKashConnoisseur "repeatedly dismissed career mathematicians and statisticians who present alternative interpretations of Hikaru's chess data" And I have yet to see one of them post the parameters. Agreed, Kramnik has not posted his either, which is why we are not certain he is wrong.
If you have access to any of these career mathematicians' comments then please post them, and thanks in advance. The math is not that hard after the parameters are known. I have seen one post where a "mathematician" claims a 1% chance of an event, and extrapolated from there, but it's all BS if we don't have the win% and number of trials used at each point.
@@scottekoontz My understanding is that Kramnik has deleted the comments of aforementioned experts from Reddit and from his blog, claiming that they are misleading. However, there are archived screenshots of said comments, as well as writeups posted to Github and elsewhere. Unfortunately, UA-cam does not allow the posting of links offsite, and so I'll have to bid you to do your own research. Nakamura has posted video content on his channel discussing some, and Google should be helpful to you as well. Also, I believe the number of trials (being the totality of Hikaru's online chess career) is 35,000 plus. Cheers.
@@scottekoontz I think that the debunkers (against Kramnik) have failed. I think that the key issue is not simply knowing the parameters, it is knowing what to investigate. For example, is Kramnik concerned that Hikaru is cheating online all the time, or only during his streaks, or only when extending a streak, or ...?
If someone gives me a dataset and tells me their thoughts about it and what they want me to help them with then I can often help them, but if they simply ask me to look for anything suspicious then I tell them that I need a clearer task specification.
I think the main reason for the discrepancy is the flagging. I wonder how many of those streaks are 3+2. The idea that this isn't an accusation is just silly.
None are 3+2, Hikaru has said it, all are 3+0
this is very imp how many he really won on position and how many were just flagging
Never thought of that and it makes a ton of sense. This guy can basically see the future when playing vs people.
@@singh.jayesh1866even if not there was no flagging (technically), it affects the performance
Hypothetically, Hikaru’s oppenents were habitually cheating and thus overrated but chose not to cheat against Hikaru, resulting in Hikraru getting easier games that expected by rating. Hypothetically Hikaru’s opponents were having an off day and didn’t play up to par during these streaks. There’s lots of explanations for these streaks that aren’t cheating. But look at the game quality and it should be clear
"Only with a very clear mathematical explanation..."
You have been given a very clear mathematical explanation, but you rejected it, deleted the comment from your blog, and ad hominem'd the guy who made the comment.
Kramnik is an unhinged lunatic with a grudge. People tell him you can't cherry pick streaks from a large dataset -- he ignores them or calls them names. People explain tilt and farming to him -- he ignores the new knowledge and retreats back to cherry-picked streaks and expected win percentage ^ number of games in streak.
The longer this goes on, the more two things get damaged: Kramnik's reputation, and the reputation of anti-cheating advocates writ large. I couldn't care less about the former given the damage that he's doing to the latter.
Best comment. Bump
You both need to have Hikaru on and the three of you can sign the petition!
Hikaru himself said he'd love for the website to investigate, he's not worried in the slightest lmao.
"im not accusing hikaru, the math is!" what a weasel. everyone can see that it's a cheating accusation so just own it.
I love your podcast, and I understand the need to cover this subject given the attention it’s received, but trying to present Kramnik’s lunacy in a neutral fashion not only does a disservice to your audience but it’s also not a good look. This is not complicated. Kramnik not only doesn’t understand how probabilities work, but has silenced everyone who has tried to explain things to him. You should cover this especially. The guy is embarrassing himself and perhaps even tarnishing the image of what it means to be a top chess player, let alone a world champion. One certainly wouldn’t expect such an idiot to reach the pinnacle in any field.
I agree
The problem is obviously that 3+0 and 3+x has the same rating. This is the reason that people can overperform so much in 3+0 relative to their rating.
I think you're right but for the wrong reason. Elo is based off of the win probability. I don't think it's really designed for being flagged. That's a really good observation.
Ppl have especially mentioned this in Bullet with 1+0 vs 1+1 which is a complete other dimension for top players. In 1+1 ppl like Alireza can basically chill forever while in 1+0 if youre 10-15 seconds behind youre just fucked if your opponent doesnt blunder a quick mate.
I've never liked Hikaru's stream -- easily the worst stream in chess. He is chaotic, and talks too much (without being helpful). His latest STUNT (no other word for it) is the icing on the cake. SUMMARY: He'll do anything for clicks.
Hey Christian & Fabiano, I tested one of your hypothesis in a recent blog article covered by Hikaru & Gotham. Would love to hear your thoughts. Please let me know if I can clarify any points.
That was a really nice analysis. Hats off, beautiful plots too
Non-statisticans who wade in deep, always do seem to get in over their heads...same here. Frankly his citing Nakamura - even if sometimes he qualifies that by saying he's not accusing Nakamura, 'the stats are' - is ironically the biggest argument AGAINST usage of his stats that one can think of. This is what former world champions do these day - invent chess variants and accuse people of things? Just stick with 'no castle' chess, Vlad.
I do rarely write comments but this topic is just absurd. The main arguments still seem to be the isolated best winning streaks that are taken out of huge sample sizes of games (while ignoring the latter). The Elo differences between online and irl are mostly not accounted for, aswell as the flagging from worse positions.
What's also just ignored are psychological components of playing vs Hikaru in front of a big audience in the first place. That's something that he is surely more used to than the avg opponent. Also tilt is not factored in. A player that's say 0-5 down & still continues on to play vs Hikaru will not always perform at their peak rating. With online chess there might even be stuff like time zone differences involved. An opponent that plays at their midnight vs Hikaru at an early evening will perform much worse just because of that (I know that from other online games myself).
And then there is lagging/connection problems, mouseslips....what also goes into your psychological component argument is that players see Hikaru and the opportunity to get huge rating points from him and might deliberately enter sharper lines than they would like too just because the off chance of winning justifies the additional risk (like going all in with a shitty hand in poker if theres already a huge pot, you can win much more than you can lose).
Hikaru is basically farming players 300 elo below him. Often through flagging.
The problem is if Hikaru plays a lot more games, he will have longer streaks just by chance. Also he could have longer losing streaks or periods of underperformance that balance out the winning streaks. So you can't just look at winning streaks alone.
That’s why Hikaru called Kramnik a 500 ELO statistician…
That's not the problem. The problem is Kramnik doesn't know even high school level math. But for some reason he thinks people should listen to him because he used to be good at chess.
Hikaru had streaks like this five times in November already. Even though Kramnik isn’t solid on his knowledge, you can’t deny there’s lots of cheating. He’s trading his authority and his good name to spotlight the problem. Maybe the trade is suboptimal, but someone has to do something. Ok, Hikaru isn’t cheating, let’s test him first, because of the numbers, says VBK
@@mikemerinoffyou can easily do that by looking up the individual games and seeing how many he had losing positions in and won either by flagging or the opponent blundering either due to time trouble or because he’s playing a top 2 speed chess player..
If the trade off isn’t the best then why not use your time in a more optimal matter looking into suspicious games and non-human moves instead?
@@EmanCipated919 As crazy as it may sound, Hikaru's accuracy per move isn't even that high. And he often plays moves outside of theory or in a different order. There are no non-human moves in his play. Not to mention how you can literally see him think about his moves on live streams. And Hikaru blunders too.
Fischer scored 19/19 against Interzonal and Candidates opposition....hmm, very interesting.
you really want to compair fischers most outstanding performance with nakamuras 2 to 3 times longer streaks which he repeated for several times? who the heck is nakamura compaired to fischer?
nakamura is even not a chessplayer. he is just a streamer ;)
ooooo the bots are feisty @@purrpurr6618
@@purrpurr6618 nah nakamura is that 1500 rated elo player i beat last night. Bro played blitz and bullet because think he is hikaru.
magnus went undefeated for 125 games in classical chess, he didn't lose in classical for 2 years, propelling his rating to 2882. @@purrpurr6618
How many online games has Hikaru played in the last year? How many games does the average GM play? Are the numbers inflated/affected by the fact that Hikaru is a wildly popular streamer and plays ~5 days a week as his primary source of income? Naroditsky also plays a ton of blitz online...does he have any of these "unbelievable" win streaks? I'm not a statistician, but it looks like Kramnik is cherry picking small sets of data and not looking at the big picture.
Naka has a lot of winning steaks because he plays much more than other top players and also is the only one to play lower rated players. Magnus would never play 55 games against lower rated players
To be fair, Kramnik thinks Danya is cheating too.
At the level of chess these guys are playing im not surprised if they crush lots of lower rated gms. They are super gms for a reason. A hundred to couple hundred points is leaps and bounds
Depends on your opponents if they are below 2500 heck Hikari could beat them all day
Agreed. The gap betwen 3k and 2500 is night and day. Sure, they can win against a 3000, but i wouldnt be suprised if magnus or hikaru goes off and beat them 50 in a row
One of my favorite examples of why he’s not cheating is when Hikaru is down to like 10 seconds and he’s simultaneously hyperfocused and explaining his moves at a crazy rate while making errors and excellent moves to tie or win.
Also the way in which he pulls himself out of difficult situations, which is where engine advice is completely useless
@@martinpaddlethere are chessbot softwares that play the game for you. All you have to do us sit in front of a camera. I'm not saying Hikary does it, but there are a lot of ways one can cheat.
@@milankotevski1663yea even if engine is playing moves for you it’s even harder to have a coherent thought on what moves he wants and planning to play as computer moves can be very random
This doesn't say anything at all. The cheating at this level would consist of just sometimes becoming aware of the sudden changes in evaluation, that's it. Like have your phone vibrate one way when it drops and another way when it increases. All the strategies and tactics are yours. Obviously Hikaru is the one making the moves and obviously he's among the top players, but cheating like this is exactly what would allow a person to play quicker and won't preclude making mistakes at all - just help win more often
@@NJ-wb1cz thats not going to help in a blitz game with time pressure. it would just slow someone like hikaru down.
Topalov must be having a field day
Vlad pointing out Naka's performances as being an extreme outlier when Naka is perhaps the most extreme outlier in online chess history is extremely funny. Vlad: "wow 55 out of 55 that's pretty strange" Uhh yeah man we've been saying that for the last two decades lol
Hikaru is good enough to get that streak but nobody should be given special treatment. There should be a fair and thorough investigation.
@@akashsinha2880 Investigation of what? There's absolutely nothing there. Anyone who has any understanding of statistics is laughing at this. It's like saying I woke up this morning. I should be investigated for being alive. My streak of being alive is unbelievable. So many days in a row.
@@alienrenders😭😂😂
pretty strange that one of the best chess players ever has one of the best win rates ever lol
@@alienrenders, well, it's not the undisputable record though
Utterly embarrassing for kramnik. He's exposing himself to be utterly clueless about how elo ratings actually work. They aren't a pure mathematical construct, they are only rough approximations of how strong a player *may be*. In practice they are heavily flawed, and drawing statistical inferences is not accurate. Also he doesn't consider time control which is a massive factor. He literally doesn't seem to mention it at all as far as I can see.
Switching it up now to streaks of games without losing
Naka hasnt lost a classical game this year (43games)
Kramnik holds the record of classical games without a loss at something like 82 games.
Seems like a super unlikely event. I bet people would find that interesting.
magnus holds it a 125 games with no loss.
@@KyleMc16
People might find it interesting how often he went to the toilet during this streak.
@@martinpaddle need to edit that again pal
@@KyleMc16 i thought Ding had like over 100 games without a loss at some point. perhaps vlad was the prev record holder. both should be investigated though 😁
hans nieman must be laughing his buttplugs out these days
where is the link for the petition? want to sign it just to see Kramnik apologize later on :D
Kramnik's campaign started for a very important cause but has now become fully political
Kramnik is actually harming his own course by going after Hikaru like this, I think all the real cheaters are very happy with what he’s doing right now
Can anyone explain to me how it would even be possible for Hikaru to cheat, while streaming, while explaining each move, while eating? He would need a second computer and a couple of spare hands, right?
He could have a beepingsound that indicates the evalbar on his headphones or signals him when there is a tactic oportunity. It is possible but he misses so many oportunities that even if it is possible his play suggests otherwise.
@@Monika77ful but it would still be impossible since if you watch his first response video for kramnik when he show us 3 hours of that match, the only reason hikaru managed to outplayed his opponents simply because they failed to punish him hard when he blunder. If he played againts someone like fabi, magnus or alireza then hikaru would lost many of that match.
@@lordtouchme77 I came to the same conclusion that hikaru is not cheating. But it is possible to cheat. Statistics alone will not catch cheaters. A great cheating detection instinct is found in magicians. And you would have to combine that with Mathknowledge and Chessknowledge of the highest order. We will probably never catch good cheaters. Again I dont think Hikaru cheats.
@@Monika77ful any possibilities of cheating would have to involve at least 1 other person, right?
@@pizzapimp8128 not nesseceraly you could have a Program doing that. But another possibility is that the opponents of Hikaru loose on purpose. Like They farm lower rated players and then sell losses to him. Again theoretical possible but I highly dought it. Because my explanaition is wildly constructed with multiple cheater (Hikaru and the people loosing to him). If you watch Hikarus latest content his explanaition is much better for not cheating. So if we use arcoms razor that means he is not cheating. Again will always be possible but in his case extremy unlikely. He is not cheating.
Absolute content goldmine for Hikaru while nobody besides Kramnik actually believes he's cheating
There are a few of Hikaru haters who believe he's cheating on the grounds of "I hate this guy" (and using Kramnik's nonsense as a rationalization) :)
Does one of the following statements have to be false?
1. Hikaru Nakamura does not cheat on Titled Tuesday.
2. There is a lot of cheating on Titled Tuesday.
3. When players who cheat play Hikaru Nakamura on Titled Tuesday they carry on cheating.
4. Hikaru Nakamura frequently wins Titled Tuesday.
Some key points the high school dropout misses:
1 - randomness: if your n (number of games) is high enough the probability of streaks like this increase
2 - elo difference, too lazy to do the math but his opponents were not Magnus and Fabi but IM/FM level and therefore the probability to beat them is high
3 - tilt/psychology factor: no explanation needed, (PTSD haha)
4 - low number of unique opponents, therefore not truly 46 unique games from a math perspective, not independent games but a couple of grouped games
5 - Bayesian statistics - dependent probability (one of the more difficult subjects in stats). Everything taken into account the probability of Hikaru destroying 4/5 IM/FM level players ain’t that crazy
6 - sampling bias - selectively choosing a streak to prove a point. Imagine looking at 46 votes to determine an ellection
7 - using annoying Russian double speak, no scientific argument but my Sovjet side of the family has thought me well how to spot this bull 💩
Stay in school kids!
100% agree
Naka beat MC last speed championship, with anti cheat measures.. extra cameras... etc.. Naka not cheating
What is the problem in checking Hikaru’s game?
Not only does Kramnik misrepresent basic statistical concepts, he seems to be completely unaware of online dynamics I'd expect literally every 1500 to be aware of. Opponent selection is one of the most obvious facets in online play. All of us have faced players at or above our rating that we feel we have a significant advantage over, and conversely players at our rating that are utterly terrifying. We all do a mental calculation whether rematching is in our favor.
Time controls also drastically change ability/advantage. Hikaru is a speed specialist. The less time or lack of increment massively boosts his advantage. Hikaru is undisputedly the best or second best bullet player in the world.
Vlad needs one sober day.
calling it now: Kramnik's team of mathematicians is chatgpt
😆 quite possible
not sure chatgpt deserves that insult
It takes a very experienced statistician to design a meaningful statistical test or analysis. Kramnik’s results may be interesting, but they simply are not conclusive and have no value. Please leave statistics to the real statisticians.
This whole thing is rather silly, it is obvious that Hikaru is not cheating. He analyzes his games and moves while he is playing live on his stream.
And says stuff like I am human and the human move is this one....suspicious
@@answeris4217no way you’re buying this lmao
@@th3cup1d92 some of the best in every game or sport have cheated. So if he does it wouldn't surprise me.
If it wouldn't be easy to cheat then maybe...but it's super easy to cheat so it's likely one or a few top players are cheating.
@@answeris4217 the suspicious reply would be, “and the Stockfish move is…”
@@tonyeye101 That would be suspicious...
In the digits of Pi there will be a sequence of 10000 consecutive 9's, the same happens when you play a lot of games, at some point you will have a streak of 40 won games.
The chance increases when you rematch a player that gets tilted.
You should have invited Fabi for this convo, wouldve been interesting to hear his view. But it's pretty clear that Kramnik has lost his mind.
Fabi is in competition, he'll be back after Sinquefield Cup!
Can't really expect him to take time to talk about gossip and drama when he is in second place in a super tournament.
AHEM! As so often happens in the chess world, we say GM Kramnik is "becoming eccentric." That's the corteous way to notice this behaviour.
Fabi would probably rather not roll around in sh&@ like the rest of us 😂
Fabiano is attempting to win a tournament at the moment.
Simple solution: Have a Nakamura - Kramnik match with the two playing 3+0 for 46 games.
Conditions: Both being on a computer, in the same room, no headphones and with an arbiter next to each of them.
Kramnik is stronger than those opponents Hikaru won in those streaks tho.
I noticed that this morning, the sun rised up again. That's 1825 billion sunrises in a raw : someone is clearly f...g with us ! I'll immediately start a petition for an honest investigation.
also I think all the cheating drama is inevitable at this point, because during the pandemic a lot of people got into chess, this brought along a lot of money and where the money is involved there is always going to be cheaters and scammers, so in one hand it is good that a lot of people have taken up chess but on the other hand it is obvious that getting good at chess takes a lot of time and all the digital aids will help you push to these ratings where you can play for money, so the nikita cheater explained it perfectly that he doesnt cheat in money events but he still needs rating to get to those and so to cut time he cheats, same can be said about already established players, their competition is tougher than ever and the better the player you are the less cheating you need, for GMs I think a simple evaluation bar is enough to give a slight edge, so in a sense cheating has come to stay, it will always be there online and it will evolve to such undetectable levels that I think there is a possibility where it will become a norm to use some sort of small aid or something playing online
Nikita the Cheater! Has a kind of ring to it. 😂
Kramnik is upset about cheating so he decides to go after a chess player who live streams every game he plays, has millions of people watch him play live and recorded, live discusses his thought process during each game.
Kramnik, on the other hand, plays games at home with no witnesses, with no audio and visual proof he's not cheating, and doesn't broadcast/publish himself playing.
Very odd times we're living in.
And almost always has a headset on...
he should put a camera in his toilet
Kramnik once played one of the online events on his phone as he couldn't get the computer tech to work.
The petition speaks for itself.
I like how you guys were thoughtful enough to include those who identify themselves as mugs. :)
Bad jokes aside, finally we have C² merchhh yayy
Of my last 40 games I have lost all of them. Interesting?
Easy to disprove then from Hikaru....more cameras to cover his surrounding area, one focussed on screen to show no on-screen bot help etc
You bet I signed that petition. 👀
You just need common sense and basic understanding of maths and statistics to know that Kramnik is totally wrong. It appears to me that Kramnik did not do any proper research before making these allegations and he goes around shutting down valid statisticians comments without a proper rebuttal tells you he doesn't understand what he is doing.even GMs that doesn't share good relationships with Hikaru has come out in defense of Hikaru speak volumes.
It's the combination of "I'm just asking questions" and "Nobody tell me why my questions are stupid".
Kramnik just needs to play too aggressive but also too passive sometimes
Kramnik just bringing up stats! Can you not handle that Naka has anomalies that are not comparable to other top players?
Kramnik should look at the players with overinflated online ratings whose OTB rating are widely different - like hundreds and hundreds of difference. Nakamura has played at a very high level OTB and online AND has played against the top players like Carlssen so he has been under great scrutiny and no one doubts his strength.
Yeah. Nakamura is a 100% chess professional playing people with real jobs. 2700 online is a patzer pretty much.
C'mon. This is accusing plain and simple. I appreciate your balanced stance on this but let's be frank. We can all see this for what it is.
You are right in saying that Kramnik is handling this very badly. These two points are without doubt.
The reason why hikaru wouldn't cheat is he almost beat Magnus in Ott world blitz champion. this shows he is capable of performing at highest level
The 55 out of 55 could be 55 out of 57 and then it brings the elo back down to earth.
When the 'shepherds' say some cheats they are opposed by channels like this one. We all know the reason why, but it is a great shame, buy a t-shirt.
This reminds me of RNG discussion for video games. People just don't have an intuitive understanding of statistics. Streaks WILL happen. The more games you play, the longer the streaks that are possible. Flip a coin 4 times. What are the odds of all heads? 1/16 or 6%. If you add one more flip, how many streaks of 5+ heads can you get? Well, the middle 3 flips must still be heads, so that's 1/8. But now either the first or last flip can be heads. The odds of that are 3/4. Combine that with 1/8 (1/8 * 3/4) = 3/32 or 9%. The odds jumped dramatically.
So if you look at a streak by itself, it looks incredible. But if someone plays a lot of games, it's expected.
Kramnik couldn't understand statistics even if his life depended on it. He has no clue what he's talking about.
We can still make no assumptions regarding what Kramnik is referring to because he has revealed none of the parameters used. The stats math is easy once you are given the Elo (resolves win%, loss%, draw%) and the number of successive wins, and in some cases an imbedded draw.
If someone plays a lot of games (2,000? 5,000? 10,000?), then some streaks are still statistical rarities depending on what you use as a win probability. Odds may jump "dramatically" from nil to nearly nil.
i have done the maths, 33400 blitz games played, average opponent rating 2800, all time no loss rate 88%(win + draw) probability of a streak of 45.5 happening is 99.7%@@scottekoontz
@@scottekoontz Kramnik can't reveal anything. He doesn't understand what he's talking about. All he's done so far is say "in this streak, Hikaru has won these games, so he's playing like an N rated player" That makes no sense. He's taking the streaks in isolation and you can't do that. I explained this in my original comment.
@@alienrenders "in this streak, Hikaru has won these games, so he's playing like an N rated player" You can make such a statement, and his results appear to be correct. A factual statistic bothers you because you want to see results across many trials (see below) but that does not negate the work shown.
Too many people do not understand statistics for many trials. Hikaru's streaks would be rare or impossible even after 3,000 trials. Some after 10,000 trials. As I mentioned, how many games are you talking about? Since the streaks are all contained within a two-months span, using 5,000 is being very generous, 15,000 going overboard.
Elo difference: 350
p result: 0.787 win based on Elo math
Desired Outcome: 45.5/46
One trial (start now, end after 46 games): Probability ≈ 0.00024 or 1 in 4,000 (not 1%)
2,600 trials (play 10x a day, every weekday, for a year): 9.5%
10,000 trials: 33% (i.e., play 38 games every weekday for a year and maybe this would happen)
-----------------------------------------------
Elo difference: 350
p result: 0.787 win based on Elo math
Desired Outcome: 55/55
One trial (start now, end after 55 games): Probability ≈ 0, not going to happen.
2,600 trials (play 10x a day, every weekday, for a year): 0.1%, very highly unlikely
15,000 trials: 0.6%
@@scottekoontzNothing in your comment is even remotely true.
Is there a chance Kramnik is trolling everybody that he KNOWS Hikaru did not cheat ... but to bring this topic/issue on everybody's radar (including dozens of super big YT/Twitch streamers) with a potential reach on 10s of millions ... Then when all is calmed down again "new online chess rules" are made? ... Thus, "mission accomplished" ?
Could this dude be playing 4d chess and just trying to force more checking? He knows this dude ain’t cheating.
Sorry but Kramnik just sounds like some random conspiracy theorist. "I'm just asking questions.", "I'm just stating the facts not implying anything", ... blah blah. And underlying it with superficial data which on a quick glance might look like obvious. Bad style and by far not detailed or specic enough for any serious discussion.
The most funny thing is that Kramnik talks about "Hikaru's bots" while actually HE is the one supported by a "bot", with a username "Rimeeeeee" from Russia (and you can see the flag). I'd say even more - this user's posts are no "interesting" and nonsense that I suspect it's Kramnik himself.
I view this as Magnus and Nakamura are generational type players. Similar to like Tiger Woods in golf or Novak Djokovic in Tennis. I bet even Bobby Fischer or Garry Kasparov performances are very statistically unlikely during certain periods of their career. I wouldn't have thought I would have ever seen people dominate a sport like these players before they arrived on the scene. What is the probability of the Buffalo Bills making it to 4 straight super bowls? What is the probability of Tom Brady winning 7 Super Bowls?
Probabilities depend on the past patterns predicting the future. If we are seeing something new that hasn't existed in the past it doesn't make sense to use the past pattern to predict the future. The current elite of every sport are changing the game and setting new standards all the time on what is possible. When the sport catches up to these new standards it will take the next dominate player or players to change the level of what we thing is possible.