If you cheat, you play God: GothamChess on Magnus, Hans, and a 0.00 that is not a draw

Lex Fridman opens by asking Levy Rozman, better known as GothamChess, to guess his rating from a look at his face. He even offers a move for the listeners: e4, which he has read that Stockfish prefers because it maximizes tactical options. Rozman does not play along. The number he gives is 3400, which is not a human. It is, he believes, Stockfish.

Guessing ratings is one of Rozman's YouTube staples. He watches a dozen moves and usually lands within a couple of hundred points, because different levels fail in different, regular ways. He tells Fridman he will use the chess.com scale for simplicity. Lichess numbers, he says, often run 300 points higher, sometimes 400 or 500, before the two systems catch up around 2200 to 2300.

They keep bumping into a joke that has already escaped the chess internet: anal beads talking to Stockfish over Bluetooth. Rozman laughs, then turns serious. It is scary, he says, how many people think the gag is real. He tends to believe that a hilarious theory, believed at scale, will make the thing a reality. People manifest the meme. Humor, he adds, is a lubricant for the trajectory of civilization, which is why a sex-toy rumor can outrun a rules debate.

Before they finish the rating lecture, Fridman asks what chess.com and Lichess actually are. Rozman describes two websites that both let you play and both wrap the same engine, Stockfish 15, in a different shell. Chess.com has a free tier with daily puzzles and lessons, then paid memberships whose flashiest product is a dedicated game review: a coach voice that flags every computer-labeled mistake. The feature he thinks matters more is the library of video courses recorded by international masters and grandmasters, hundreds of hours, with training positions you can practice afterward.

Lichess is the other pole. It is open source, donation-funded, and tries to be as free as possible, with no advertisements. You can play, analyze, study openings, and sit on a live server. Nobody holds your hand. Rozman's standing comparison is that chess.com is a good personal trainer at the gym, while Lichess leaves you to figure out the machines. Combine YouTube with Lichess, he says, and you can do it, but you have to explore on your own.

He is sponsored by chess.com and admits the bias, then compliments the competitor in the next breath. He also cannot resist a joke: ardent Lichess supporters, he says, are like vegans, ready to convert you. Chess24, another live server with lesson libraries, was bought by Play Magnus, the company around Magnus Carlsen, who Rozman thinks owns something like nine or ten percent. That group now looks to be merging with, or being acquired by, chess.com. The for-profit companies, he argues, are what let grandmasters make a living and what made chess an esport. Lichess is great and puts on no events and no commentary. The ecosystem needs both.

Back on the 0-to-900 band, Rozman says the crack is the ending. With seven or eight pieces left, players at that level do not know how to convert. They do not know when to bring the king into the fight, or how a bishop outplays a knight with a few pawns. Some of those positions are solved: show the board, ask for the evaluation and the win, and there is a technique. Coaches on chess.com can walk you through it in a recorded lesson, then let you practice. Endgames are still hard, he adds, which is why Carlsen is the best. Seven hours in, when everyone else has given up, he is still squeezing juice from the fruit.

Within the first fifteen moves you can usually tell more. You see how well they knew the opening they were supposed to memorize, and how they react when pieces start hitting each other. A more advanced player has a natural reply ready. A massive giveaway is moving the king in check when you did not have to. Some people think check means the king must run. Meeting check has three answers: interpose a piece, capture the attacker, or step the king away. Low-rated players jump to the third. Stronger players scan all three, and often the king never moves.

Memorized openings do not save you, because it takes two to tango. You can drill a line, then sit across from a 900 who plays some crap not in the book. Either you answer the way you think you are supposed to, which will probably be wrong, or you have to handle a fresh position. Both reactions show the real rating. Rozman's own over-the-board number at the time was 2320, peaking at 2430. He can still force himself down to 900. Before YouTube and Twitch he taught kids in New York, and he had to play at their level and explain in a way a child could hear. That, he thinks, is a lot of why the channel grew. He still remembers how beginners think.

Fridman asks what the kids who improved quickly had in common. Rozman does not hesitate: obsession, and love of the game. If you are bored, you do not want a show. You want to boot up chess.com or Lichess. The useful metric is not a pious hour a day. It is how fast you return to the board. Hours spent and hours wanted both matter. Talent exists, he says, though he is not sure it is chess-specific. Some children pick up which squares their pieces control much faster. Their memory is not just openings. It is patterns: this position from two tournaments ago, this one from yesterday.

Then those brilliant kids go to a tournament and collapse. Give them a page of tactics and they solve faster than anyone. Sit them across from another six-year-old in library quiet, where you can hear the breathing, and the anxiousness wins. One sentence of trash talk can throw off a prize student. Rozman has watched kids disintegrate, and he has watched his own students bully other children with faces and little expressions the rules forbid and the arbiter never sees. Chess training drills calculation. The points get lost in emotion.

At the highest level the silence is a different war. Both sides, Rozman says, are in a battle they both foresee 99 percent of the time. The scary part is not a move you never saw. It is a move you saw and misevaluated three or four ply deeper. Your opponent plays it, they get excited, you get nervous, then they inaccuracy and the tide shifts. You might spend the whole game on the same plane, or on totally different ones. When a move looks sub-optimal, the first question is whether they miscalculated or you did. Getting that wrong is how you lose.

Fridman asks whether Carlsen's gift is taking people away from known openings. Rozman thinks a lot of the dominant reputation comes from not letting opponents into ultra-theoretical positions. Carlsen had just won an online event and talked about an anti-young-player strategy. By move seven or eight you go to the database and find no games. The kid is on their own, swimming in strategic complexity that Carlsen already feels. Some of that, Rozman says, is an enormous private library of historical structures. Some of it is sheer genius. A younger player cannot remember a game from a Soviet bar in 1951. Carlsen can, because he read a magazine once. Rozman calls it unfair.

What makes him so good, if you have to name it, is memory plus the sense that he just gets the game better than anybody else. Age has added stamina. Put two other 2750s, world top-ten players, into some of those positions and they would draw. Put Carlsen in as the aggressor and the winning chance doubles, Rozman says, from five percent to ten. Game six of the last world championship against Ian Nepomniachtchi is the exhibit. Stockfish sat at 0.00. Sometimes that means a dead draw, theoretically impossible to win. The joke Rozman always makes is that sometimes it means Stockfish went out for a smoke: so much complexity that the engine cannot evaluate correctly.

Carlsen kept asking questions for six or seven hours. He sacrificed a pawn, then another, to damage the structure. The evaluation did not move. A machine could have stopped him. Nepomniachtchi could not. The game ruined Rozman's day and became one of his favorite recaps. He had tried to be first and best, watching every broadcast, stuffing every variation into a video, and the adrenaline dump at the end flattened him. Most of the match was draw after draw, boring for the average viewer. What he loves is the moment the grandmaster commentators say they do not know why a move was played. After game six, Carlsen said that ended it, because now Nepomniachtchi would have to take more risks. Rozman quotes the runner Steve Prefontaine: when the race is down to pure guts, that is when I win.

The pre-match consensus was that Nepomniachtchi catches fire in the first half of a tournament and slows down after the first setback. That is how the match died, in the worst way, one-move mistakes cascading. He later won the Candidates again. Rozman puts the next title match at 55-45 and will not say whose 55. Nepomniachtchi has been on that stage and knows what a training camp feels like. Ding Liren is one of the most stable, unemotional players in the game, and his curve runs the other way, from down to up. In the Candidates he was demolished by Nepomniachtchi in round one of fourteen, jet-lagged after arriving in Spain two days early. His chances of finishing top two sat around two percent. People wrote him off. He took second.

Fridman asks how Rozman feels about Carlsen stepping away from the world championship. Disappointing, he says, but understandable. He is not a romantic about the title, and only under pressure admits he sees beauty in certain types of chess, not all of it: a part-time romantic. He also notices that Carlsen still signs as world champion. The prestige, Rozman thinks, is not what it could be. He is a UFC fan. Georges St-Pierre and Khabib Nurmagomedov walked away near the zenith, but UFC champions are taken care of, with sponsors, massive paydays, and international celebrity. Chess does not have that. The world championship prize fund, he says, has not changed much in forty years.

Tally Carlsen winning every tournament he plays in a year, which Rozman does not even think is a crazy estimate, and that sum still looks smaller than a year spent on social media, brand building, and sponsor deals. He thinks the gap is about five times. Money is not everything. The test he wants is simple. If the next championship were ten million dollars and Carlsen still said no, it is something else. Take him at his word: too much stress, too little sample. He wants to play more often, in a format closer to a long series of tests. The rating list already works that way. Game after game estimates strength better than a short match.

History makes the current product look almost gentle. The first Kasparov-Karpov match was first to six wins, draws ignored, no best-of cap. It ran something like seven or eight months. Karpov went up 5-0, Kasparov came back to 5-3, and officials called it off for health. Kasparov stormed out calling it a farce. The match was on the order of fifty games. Bobby Fischer wanted something almost as extreme: first to ten, and if it was 9-9 the champion kept the title, so the challenger had to win 10-8. Rozman would love that for YouTube recaps and not for the players or the public. Imagine your favorite tennis match lasting six months. What are we doing. He thinks Carlsen is openly challenging FIDE and the rest of the industry to redesign the event, which he calls a good thing.

Asked for the greatest of all time, Rozman will hear cases for Magnus Carlsen, Garry Kasparov, and Bobby Fischer. Capablanca was brilliant. Steinitz was brilliant. He does not put them in that trio. He leans Kasparov, and thinks Carlsen has a chance to overtake him. Longevity is the lever. World championships matter the way rings matter in basketball. Chess also has years as number one, which is what Carlsen likes, because it is built from game after game rather than a small-sample title match. Carlsen had been world number one for about eleven years. Kasparov held it for twenty. Carlsen's rating is higher.

Rating, Rozman thinks, has inflated, though he cannot narrate the exact mechanism. Online it is obvious. Three or four years earlier, some of the best blitz players on chess.com were 2700 to 2800. Now they are 3200. Over the board, FIDE once changed the formula for juniors. If you were under 16 or 18, he cannot remember which, and below 2300, your rating-change factor was three to four times higher. A normal win might move you five points. Those kids moved four times as fast, up and down. One American teenager played a month of tournaments under that factor and became nearly 2640, around top fifty in the world. He became a grandmaster later, then bled about ninety points as the inflation came off. Rozman still sees kids jumping from 2100 to 2400 after one good month.

That is why he will not let a raw number settle a cross-era argument. Rank, duration at the top, and the strength of the field travel better than a rating that drifted because a K-factor got turned up. Fridman had already tried to pin him as a romantic who treats chess as a beautiful art. Rozman first said he was not qualified, a teacher who now makes YouTube videos. Pressed, he conceded beauty in certain types of chess and settled on part-time romantic. He would rather talk about who stayed number one than about the poetry of a sacrifice.

Fridman finally names Hans Niemann, the nineteen-year-old who had just beaten Carlsen, then watched Carlsen forfeit after a move or two. Rozman says there are really three games, and they will get there. First they detour through the internet. Rozman pins mean comments. Viewers named it the Pin of Shame. He went from no subscribers in May or June 2020 to a million in a year, and by September 2022 he had about 1.6 million. He still answers people who say horrible things, sometimes with a question mark. He is, he keeps insisting, a human more than a Creator. People now fake hate comments for attention. The good thing got ruined the usual way.

Fridman had tweeted for a minute that Rozman was coming on, then deleted it because 95 percent of the replies were about cheating. In that window someone called Rozman a lame guest. Rozman replied. The man said he did not know what value a person who just makes videos about chess games would bring. Rozman's bio-read of the account was family, God, and politics. One five-second joke in a thirty-minute video can exile you from a household. Fridman wishes people would search their own old emails for the stupid things they have said. Rozman thinks chess criticism hits lighter than the existential topics Fridman covers, and still hates the disappointment of people being cruel about nothing.

The Indonesia story is the one that actually shook him. On a chess.com stream he sat down with an untitled, high-rated account flying the Indonesian flag, which anyone can pick. The win rate and accuracy looked off. He said so out loud, played anyway, lost to a move that killed a position he liked, and wrote the fair-play team off-stream. The account was banned. That night, twelve hours ahead, Twitter exploded. A Facebook post claimed the account was an older man, that Gotham had lost to someone's dad, and that the community had mass-reported him. Indonesia has the fourth-largest population on earth. Rozman learned it the hard way: tens of thousands of Instagram DMs, his like-dislike ratio collapsing, threats to kill him and rape his family, people contacting accounts he followed.

He is 100 percent sure the account cheated. At some point it won 27 games in a row at 95 percent accuracy, which even Carlsen cannot do. He geo-blocked his YouTube in Indonesia through an MCN, hated watching the work burn, and learned he is a horrible crisis actor. The father and son went on major Indonesian shows. The Indonesian Chess Federation brought in a sponsor and sat the man down with Irene Sukandar, an international master around 2400. She crushed him. Rozman puts his real strength around 1300 or 1400. The man still got paid a losing prize and vanished. Rozman shut DMs and notifications. The lesson he took is that every live minute can be blown out of proportion, and that he would rather get a drink with someone than let chat laugh them into a dark place.

Rozman had tweeted that he was retiring from all competitive chess. Preparation had outmatched him, his calculation was too flawed, and most importantly his anxiety was beyond repair. He could not do it physically or emotionally. Over the board after he became a Creator, the feeling was a kid who had not studied enough for a test. Good moves never felt as good as bad moves felt bad. He is an international master who got the title at 22, which in that world is late. There are IMs who are 11. He quit more than once as a teenager. One parent told him to sit down, that this was the only thing he was good at. He had one coach around age ten, a man in his seventies who still emails, who used to tell Rozman's grandmother he was brilliant and never worked.

The best run was 2016. Teaching a scholastic program, he asked the parents, managing directors and lawyers in seven- and eight-figure households, to help him stop private lessons for four months and travel. He kept a blog, studied six or seven hours a day, and went from 2040 to 2404 with two of the three IM norms. In 2018 he came back again, hit his peak, and then the mission was done. He had never thought he would be a grandmaster. The adrenaline dump of the IM title made him stop working. The second he started falling he could not stop. Online, in messy dynamic positions, he still beats grandmasters and thinks that if he trained full-time he would get there. He does not think he will do YouTube forever. Fridman reminds him they are going to die. Rozman once cried at a funeral when that landed.

They wander through Twitch, Skyrim without monsters, and why Rozman will not stream Overwatch anymore: teammates using racial slurs in voice chat. He grew up across three households, grandmother as Switzerland, the same backpack moving between houses, and he will not pretend there is a clean theory of why he came out alright. You need a support system, he says, and you need luck. At fourteen in a New York park a stolen basketball almost turned into a gang beating. Nothing happened. He still thinks some lives turn on those minutes. Fridman talks about almost being run over, and about meeting people who make another kind of life look possible. Then they go back to the teenager who had been doing stupid things on a chessboard.

Carlsen is, in Rozman's view, arguable greatest of all time, world number one for more than a decade, and still close to the top of his game. Niemann is 19, American, not a prodigy in the usual sense, because you do not become a prodigy at 18. Rozman met him at 11 or 12 at the Marshall Chess Club, a little kid trash-talking, already famous with counselors. Rozman can relate. He was kicked out of a camp at nine by Artur Yusupov, once top five in the world, who said it was either him or this kid. Niemann drifted in and out, around 2300, then in an open tournament in Philadelphia played nine grandmasters and got an IM norm. After the pandemic he streamed, talked trash, gained rating, went back over the board, took his last GM norm, and demolished people. The juniors were a wave: Alireza Firouzja at 2800, Vincent Keymer at 2700, a crowd of Indian teenagers. It looked awesome. Then you started hearing sprinkles. Chess.com does not always put a badge on an account. Sometimes they just tell the player they know.

Niemann has admitted to cheating online at 12 and 16. Chess.com and then Carlsen said they think it was more than that. Historically, Rozman says, a teenager who cheated online was not assumed to be cheating over the board at the ultra-elite. It was not supposed to be possible. They first met in a live esports event in Miami. Day one, Niemann lost 3-0. Day two, game one of a best-of-four, he destroyed Carlsen. The interviewer said yesterday was terrible and today you start with a masterpiece. Niemann said it speaks for itself and walked off. He then lost the rest of that match and every remaining best-of-four, finishing with zero points and zero prize money. There are photos of them playing on the beach. They were still getting along. Almost nobody needed to know who Hans Niemann was. Three weeks later he was probably a top-five all-time popular chess player.

The Sinquefield Cup in St. Louis is the classical invitational, named for Rex Sinquefield, hosted at a club with a Hall of Fame. Someone in the top ten could not come, for a coronavirus-related reason Rozman does not specify. Niemann was the replacement. Weeks later it came out that Carlsen had wanted to withdraw, and that other top players were suspicious and wanted stronger anti-cheating. Rozman calls the over-the-board measures dog shit: a metal wand, no shoes off because it would be elitist, and a live argument that a micro earpiece, a buzzer in the armpit, or something in another orifice would not be caught. The extra measures were not introduced. Niemann drew Levon Aronian while pushing, crushed another top player, then beat Carlsen with Black from start to finish. Carlsen tried a rare White idea. Niemann played the first 15 or 20 moves perfectly, converted a slightly better ending, and squeezed him to death. Nobody had beaten Carlsen with Black that way in years.

In the interview Niemann said he had looked at those moves, or something very similar, that morning. Rozman has done that himself. Combined with the rest of the interview, it got odd. Elite post-game talks usually walk through variations. Niemann's were sloppy in a different way: yeah, completely winning, I just played perfect. The next day Carlsen withdrew and tweeted a Jose Mourinho clip: I cannot speak, if I speak I am in big trouble. One theory was a mole on Team Magnus. Rozman thinks that is stupid. Carlsen's inner group is tiny. Knowing the opening helps a lot if the line is on a knife edge. Round four was another brilliant game Niemann failed to win, explaining a piece sacrifice in words rather than moves. After the first three games he played like a good grandmaster, lost a couple, and drew the rest. He admitted the old online cheating, denied over-the-board cheating, said chess.com had banned him privately from the Global Championship, and complained about Hikaru Nakamura on Twitch. His last tweet on the subject, Rozman says, was September 7.

In the online Meltwater event Carlsen played one move against Niemann and resigned. Rozman thinks he did not want to play at all, and that some sites will not let you resign before a move is made. He still finished first in the preliminaries and won the event. Niemann lost in the first knockout round to Le Quang Liem. Carlsen then put out a statement as an image. He called cheating an existential threat, said organizers should raise over-the-board security, and wrote that he believes Niemann has cheated more, and more recently, than he has publicly admitted. Niemann's progress looked unusual. During their Sinquefield game Carlsen had the impression he was not tense or even fully concentrating in critical positions, while outplaying him with Black in a way only a handful of players can. Going forward he does not want to play people who have cheated repeatedly, because he does not know what they are capable of. He said he was limited without Niemann's permission to speak openly.

Rozman will not put a probability on that specific St. Louis game. Carlsen played poorly, one of his worst, which does not help the accusation. Day by day, though, he thinks the evidence is showing more, the way Carlsen said: more than Niemann admitted, and more recently. Statistics from Niemann's 2020 and 2021 over-the-board games show engine correlation higher than Carlsen and many top grandmasters combined. The counter is that he was a very strong player beating weaker opposition, possibly a 2700 who had been incubating while listed at 2400 or 2500, and the most active player of 2021. Carlsen named Maxim Dlugy, a grandmaster and former U.S. Chess president, later revealed to have been banned on chess.com. Rozman says they had worked together. You still do not have physical proof. You have smoke. He hopes, chess or not, that Niemann is mentally strong enough for whatever is coming. If you cheat in chess, he said at the top of the show, you play God. You decide when the game is over. You can fake bad moves. Cheating the right way, you will lose plenty of games to avoid detection.

Fridman wants the engineering of it. Rozman's mind stops at sophisticated and goes back to YouTube and a book. Fridman has built chess engines and a vision-reading arm, and he likes the joke of wiring Stockfish to a Bluetooth sex toy, a rumor Elon Musk amplified with Rozman's username sitting in the copypasta. Eric Hansen read the line out of Chessbrah chat. It became late-night television, Trevor Noah and Stephen Colbert, and news in China. Lex found a Rust library, Buttplug, that talks to vibrators, then flinched at how dirty the demo would look. Rozman thinks a tutorial showing that you can encode squares in vibration would be hilarious. He will not say the words on his own channel. He also knows that only the stupid cheaters play the top engine line the whole game. Grown adults cheat in lessons. It feels like a drug.

The bots beginners love are, he says, mostly the same program wearing different faces. You can make them blunder on a schedule. The problem is the mistakes are not human. A toddler bot hangs everything. A so-called 1800 hangs a rook after a reasonable opening, which is not how you get to 1800. Then the thing enters a loop and bulldozes you, because it was only allowed a few errors. Stockfish, given a tenth of a millisecond a move, already plays better than a human, which Rozman finds disgusting. AlphaZero excited him without existential dread. It destroyed openings people thought were good, with the caveat that it may have had a full tank of servers while Stockfish did not, and DeepMind stopped developing it. Self-play is what he thinks intelligence looks like, and it inspired modern Stockfish. An engine move, for him, is two things: a concept so outlandish that a hundred grandmasters would pick two or three tries and the computer throws a haymaker, and a continuation so counterintuitive that only the future the machine has already seen makes it good. Fabiano Caruana had just said some of Niemann's games were out of his league. Rozman hears the implication. Is it possible Niemann is that kind of genius. He does not know how the story ends: a pile of chess.com evidence, a confession, a lawsuit, or a denial to the death.

Rozman's favorite opening is personal. At 15, after three years away, he found a book by the Danish grandmaster Lars Schandorff on the Caro-Kann, 1.e4 c6. On move three, at amateur level, Black already has a plus score, which is not supposed to happen. Club players do not know how to meet the structural attack. Stockfish would take the offered pawn and tell you to win it back the annoying way. Ninety-nine percent of the rating ladder will not. He plays the line against grandmasters. Openings, he says, run from two moves to ten before you are on your own. Titled players know every opening and still cannot play all of them competitively. Carlsen recently chose a move order in a position reached thousands of times that had been played maybe 0.05 percent of the time, a line supposed to let Black equalize. He does not care about equalizing. He wants a position where the opponent does not know the answers to the test. In the seventh hour, Rozman says, that intimidation is not the same as sitting at a nine-handed poker table with him. Daniel Negreanu, he jokes, may be the only person who does not respect Magnus Carlsen.

Advice for anyone trying to improve, after the obligatory plug: be prepared to lose way more than you win. His mother told him at 13 or 14, when he had just discovered he liked girls, that he would get a lot more nos than yeses. Adults learn worse than kids because they are afraid to be wrong. They preface correct answers with this is probably wrong. They assume a career of hard study will transfer. It does not. You need a beginner's mind, which means sucking in every part of the game and studying the interesting ways you suck. People do a lot of fake learning, blasting speed games and clicking through tactics. He used to tell students to do ten puzzles and get 100 percent, however long it took. Thirty puzzles with eight correct is laughable. Magnus says what matters is hours having fun. Some improvers say what matters is hours of suffering. Rozman thinks it is a mix, closer to fun, with enough stubbornness that he has sat on one puzzle for an hour rather than punch in a wrong answer. He can play four blindfold games at once and does not know how to teach that. It is a party trick.

Fridman asks for the lowest point. Rozman says of course he has been depressed. In May 2012, at sixteen and a half, he lived for something like eight months in a house where he felt like a prisoner, confiding in his grandmother, grinding through junior year by staying at friends' places and listening to sad music. He will not put the family details on a podcast. More recently he tore up a happy life with his wife. A house they had just moved into was broken into with a crowbar. They went to a high-rise. Neighbor noise started at 7 a.m. A woman at the door gaslit them. Building staff did nothing. His stubbornness decided they were leaving. They rented a too-big house, got a lunatic puppy, and on moving day he realized the spiral had been in his head. He did not want to exist. It was not a suicide plan, he says, more the wish to watch himself from a distance. Anxiety turned everything to dread for two weeks. Through all of it he streamed and posted every day, because he had a job. The emotions he is learning to name are guilt and remorse, a brain going in circles about things done and not done. Lex closes with Irving Chernev: every chess master was once a beginner.