The most overlooked problem with cheating is that the players who genuinely love the game start to lose interest in the game because its not worth it any more 😢

That "Hans" Freudian slip at 16:04 was hilarious

It's still unbelievable to me that people who make it to that level of competition would have so little respect for the game as a whole

16:01 A classic Freudian slip from Hikaru.

Was playing a friend in daily games and he was consistently turning a completely losing position into a win, few weeks later he said his account got “locked” because he messed up his emails. With his new account he hasn’t played any online games unless it’s against his friends.. I have a hunch he was cheating and got caught.

Fabi's analysis of chess and the challenges it faces is superb.

This issue has been going on for quite a long time, especially in online chess, at all levels. It's good to see the top players having this discussion. Cheating in chess, especially at lower levels is rampant.

Im wondering if different national responses to the pandemic had influence. If players on countries like the ones Hikaru mentioned, where there's a limited pool of players over a certain level and smaller improvement increments, the countries with stricter travel requirements or extended limits on larger events would limit OTB participation, it would make sense that now those players would be able to get those points. I mean, Ding almost didn't qualify for candidates for similar reasons. Anyone who was not Ding would have less opportunity

Hikaru greatly adapt his playstyle online to counter cheaters: He rather play faster moves that are maybe not best or not direct to complex positions than calculating heavily to play most accurate and complex moves. That's why his games looks less spicy than some other players - viewers can even thinj that Hikaru sometimes play for a draw / early endgame than for a win. But when Hikaru creates position where engines gives opponent +0.1 +0.2 and after ten best moves it become +0.8 then cheater has to waste a lot of time to move pieces on the second board and play ten best moves to take only a little advantage over Hikaru. In that time Hikaru plays fast and take minute or more advantage on the clock that gives him a chance to win an endgame against cheater - because with only seconds on the clock cheater cannot really use engine and has to play by himself. I think that is the reason why Hikaru has so good results online even when a lot of people are cheating. I'm pretty sure than non of us amatours could defeat Hikaru in 3+0 / 3+1 even using an engine (without bot who makes moves for us).

The problem with ratings is not elo, it's exactly what Caruana said: the highest rated players just play among themselves, so it's impossible to have a big variation on their ratings. Then the only time they play someone that is as strong as them but outside of the group they will lose a lot of points because that person inevitably will be underrated.

When he kept saying Hans 💀

It seems to me that the rating declines would be caused by a large group of younger players, growing quickly. Their rating would not be able to keep up with their current skill levels and thus they would constantly be underrated, and be draining points from higher rated players. Look at the large group of youngsters coming up out of India. Those kids are going to take points from many higher ranked players.

Once I'm lost and suspect cheating, I just start 'blundering' all my pieces. If my opponents response time doesn't increase when the game is ever more easy to win then my suspicions are probably confirmed.

Once money is introduced into anything, chess, a hobby, asking someone for a favor, it changes everything.

It is also true that cheating is pervasive at lower levels. I am a lowly National Master and even I can tell (mostly) when someone is cheating. I expect cheating is far more prevalent with lower rated people.

Love this kind of content from Hikaru!

That inflation I noticed it first in lower levels. 2 years ago I started playing again after a long pause. In the past (10 years ago) players of 1600 didn't know much of openings and made blunders in the middlegame, and now that level is 1400, 1600 players now know a lot more theory and play decent middlegame and endgame, like it was for 1800 before. I think that this deflation has come to higher levels.

6:00 Also the closed pool among 2700 players who often play invitationals with other super GMs means these strong low rated players have fewer chances to beat 2700's and gain rating.

The number of times I have been playing a game against someone who was taking their time, 20-30 seconds per move, holding even with me, and then when I win a minor piece or a couple pawns, suddenly they start taking 5 seconds per move and finding absurd tactics... it makes me suspicious. It's like someone handed off the controller to their big brother in a fighting game. Watching a player who fumbles openings just like me, misses basic math on protected pieces, then ten moves later they're finding three-move-deep fork setups lightning fast? The math, it is not mathing.

Hans confessed to cheating online while streaming and no watching his stream noticed. I don’t think a camera will help detect cheaters. Maybe online tournaments need to be on a delay

The moment you said "Hans" i was dead of laughter for 5 minutes straight😂😂. You're such a cool guy.

6:34 Could it be that either cheater influx in lower brackets makes cascaded rating drop all the way to the top or a new generation of players learning with engines rather than books creating a the cascade?

I think one thing to keep in mind is visualization. For me personally, I learned chess online and have a far easier time seeing tactics and weaknesses looking at an online board compared to OTB games. I'm not saying they're wrong about cheating, but you need a lot more evidence than "they play different online" because even missing one or two tactics/weaknesses in high level games could completely change how the games go. I wouldn't be surprised if there are players that just have an easier/harder time working through their ideas with the simplified/over head view of the online boards.

Interview is top notch and Hikaru with his words makes this more understandable. Fabi is like a surgeon where as Hikaru like a pediatrician.

I once played a game and was accused of cheating, it was the greatest compliment I ever received in chess.

It’s also opening prep. An average 2400 can play like 2800+ for 20+ moves. To test them you need to play suboptimal novelties, and therefore outcomes dont match ratings difference

100% true that cheaters will try and argue that those who think there is cheating are bad losers and that there is no cheating. They will act all innocent when it is suggested and ask how it is done and that sort of thing.

Maybe this is why Hikaru has gotten so much better, he's been playing with cheaters so much it's improving his game at a much quicker rate than someone like Fabi.

Maybe having a rating system that combined in-game strength and performance against stockfish would be an idea. Then you'd be measuring pure ability against perfect chess + ability to handle the pressure of playing another human in competition.

Thank you for bringing this content Hikaru!!! Much needed with all the online cheating lately insane!!! Even in the bracket I play in 2000-2300 so much cheating!

Another point of note… for financially strict people… its not unreasonable that both spouses would offer insights into an ongoing game, effectively making it a 2v1. There are many situations that would be cheating that can be considered fair from there perspective…

Love it! Great commentary

Its actually psychologically very explainable: Imagine you start your dream of becoming the next Magnus at 7-8 years of age. You dedicate basically your entire youth and possibly life to this cause to then end up maybe not even 2500 or maybe 2600 which is also not even remotely close to the actualy big boys of chess that earn a lot of money....at some point it might corrupt you into thinking you "deserve" just more after sacrificing so much (which imo is what happened to Niemann without opening a discussion this is MY opinion). And then you have that nowadays cheating is much easier and thus seduction way higher with all the online chess going on and all the advances in technology as well.

I think the main problem with underrated players is how much chess than they play online at a high calliper versus how much OTB Fide rated games they can play. They’re improving faster than their Fide rating can keep up with.

Cheating over the board is of course more difficult but still not very tough when there is no delay on stream. There is no way to scan players for small devices that are only receivers (devices that are listening for signals but not send any signals, especially when those type devices are turned off when scanning occurs...

Adequate id checks instead of random online accounts for attendants, 2 cameras (one showing your face, another showing the area in front of you) and 2 year ban for cheating would help a lot. That said, I don't think any amount of cameras can help because you could always have other methods of communicating with people near you that are simply not visible in any camera (which is easy to do when you get to setup the cameras).

Thanks for reacting to this podcast episode, Hikaru! I found it very lucid and interesting, so I can't wait to hear your thoughts.

Yes, online cheating has become an epidemic A person that knows chess and is a smart cheater will never be caught, unfortunately It’s ruining the game

adding cams wouldn't change the statistics. Because people will just cheat on cam, with a buzzer in their pocket that goes off whenever there is a winning combination on the board. The fact Hikaru doesn't see that immediately puzzles me. He blundered that logic

For me personally i can see how beeing in your "safe desk space" instead of high stress over the board enviroment would improve your game. Not saying cheating isn't happening, ofcourse it is but I think there is a big mental factor that should be counted in when playing over the board vs playing in the comfort of your home.

I like that Fabi has been a world championship challenger and 2800 yet still admits, sometimes you just dont see the move, "If you see the move you play it, but sometimes moves just don't come into your mind." To me that feels like a universal chess experience that everyone has. Post game or move, how did I miss that moment?

There are so many cheaters online it's insane.... Don't even take it seriously anymore just for fun. OTB is the only place Chess can be taken seriously.

It is possible that very strong players emerge from online chess and classes aswell as youtube tutorials and courses etc, and OTB they have played to little, explaining the changes in the elo ratings

I just love how this channel is all chess all the time. No matter what I do throughout my day, I can bet hikaru is either eating, sleeping, pooping, or playing chess

To one of the points made. If anyone is playing, they just need to be alerted the opponent made a mistake. For cheaters that are cheating to cheat, with little understanding of the game, that would not help. For people that know how to play, knowing your opponent made a mistake you see an evaluation bar jump, obviously you will look longer. But all you need is a signal. You could have someone on camera all day. They could be using anything as the signal. They are physically making the moves, they are actively playing, but every time their opponent makes a mistake. I think many people could convince themselves it isn't cheating. Over the board, if anyone has real time access to the moves, how hard is it to hide something, that all it has to do is send a minor shock, or someone in the room smiles. Money does motivate people to cheat, but a lot of them just do it to cheat, to beat a system. There really isn't a solution. I had one of my best friends cheat in a game against me, just because he wanted a win against me.

Man when hikaru saiu Hans name twice by mistake made me burst into tears laughing 😂

A likely possibility that cannot be dismissed is that certain chess organizers who are not implementing or resisting actual anti-cheating measures are cooperating with the cheaters in some way and benefiting or profiting from such cheating. Obviously, honest tournament directors or organizers at this point in time have no excuse for their neglect of anti-cheating measures unless they themselves are corrupt, dishonest, and cooperative with the cheaters in some way.

one reason why some may perform better online then on board is nerves. having said that ofc a lot of people are cheating. but checking online performance vs board performance can hurt some that struggle with nerves when eyes are on you

I could find only 3 people

I played against darts with a person who I shall not name but he was in the top 5 of people in our province. Averaged like 115 a lot of the time but eventually it came out that he used really small darts and when the darts were close he would still take off the high score cuz nobody was looking too close. They just assumed it must have been in there. Until he celebrated and started shaking hands one day before taking his darts out of the board. The guy he was playing against rushed past and moved his dart over so you could see it wasn't in the double. They he ounched him in the face. Then after that we realized that he could get 100s most of thr time but not 140s like it seemed. So he was still in like the top 10 anyway but it wasn't good enough for him I guess.

One reason of rating deflation is the talented new generation. I remember period when these top guys were drawing each other on purpose in big tournaments to maintain ratings. But then some tournaments got similar to Tata Steel, started inviting younger, talented players and then these top players were forced to win or faced with losing rating.

42:20 that's so funny he thinks this now but not when he was analyzing the "engine correlation" conspiracy with the Hans drama

As someone who plays chess, I can confirm that chess speaks for itself.

Fabi's voice is so similar to Eric Rosen's voice.

This aged well, rationalizing and belittling the negativity of cheaters. No big deal.

They need to stop making money tournaments in online chess.. cheating is too ridiculously easy to do in your own home.. a chess bot set to play moves at a 2600 level could beat even elite players in faster time controls while still making reasonable enough amounts of mistakes to not raise any flags... OTB should be the ONLY financial tournaments in chess.

The only way to catch chess cheaters is to recruit a team of the highest rated proven cheaters. And keep replacing them with higher rated players that they in turn prove are cheating. Its like pay back the guilt debt and you get to play again

16:04, Freudian slip 😂

Maybe not 2600 but 2700 for sure can use one or two assessment hints to beat top 10 consistently.

16:02 That's a Freudian waterslide! 🤣

The online cheating issue has gotten so bad that I have declined rematches after winning. Ever get that feeling your opponent is about to start cheating against you?

Inconsistancy is also how I detect cheaters in CS. If they play badly but then occasionally pull off stuff that would require world champion level skill I know they are toggling cheats.

Im not totally sure how the math works but I think part of the rating depreciation is because of all the great young talents coming up and taking rating points from some of the older more established players. So they are on the way up and the older ones are on the way down so we are just in a "down" time rating wise but I think in the next few years they will go back up again as the young guys establish themselves as the new super gms.

My FIDE is 1920 and I'm 1600 to 1800 online depending on the time control. I can't be that unique, can I?

One aspect I rarely hear mentioned is how “it gets into your head” and takes focus away from the game. Instead of thinking about the game; I find myself thinking about whether or not I suspect my opponent is cheating. It’s bad out there. It has gotten to the point where I only do puzzles and just play the bots themselves. That way I at least know I’m playing a bot. …still stuck on Francis bot though. I have beat Noam (3 stars) so it seems possible. Once I blundered a winning position.

Fabi is spot on - one can flag someone for inconsistency in level of play between OL and OTB over enough games - but it’s still not proof.

Arpad Elo would be astonished to be schooled on his rating system. The explanation is probably the mixing of online and over-the-board results. The solution is not to mix these and give a rating to online games and a different rating for over-the-board games. If only grandmasters played chess then all ratings would gravitate toward the middle of the curve at 1800. The rating number is not some kind of absolute indication of a player's strength, but rather the mathematical probability of him winning. You should read Arpad Elo's paper on the mathematical underpinnings of the system and I think you eventually conclude that it is not absolute perfection, but it gets pretty close over the long run.

pretty sure Fabi is referring to Xiong, whom he has never beaten in a Titled Tuesday event. Placed 2nd about 10% of the time and winning around 1 in every 10 events he competes in.

Great resources - much better than before. Young players using them. That would seem to be enough to explain all the rating issues.

Your use of the Service is deemed a “continuing acceptance” of this Agreement, meaning that every time that you use the Service you accept the then-current version of the Agreement or any of the Other Policies, even if they have changed (without notice to you) since the last time that you used the Service.

16:03 Freudian slip is the psychological phenomenon that just keeps on giving 😂

We all know the most accurate rating system is average speedrun peak ELO. Clearly this is the superior method to determine someones true ELO. If anyone has a more accurate measurement, that covers the whole rating range and a players ability over a wide range of openings in different time controls, I'd be happy to hear them!

I play a lot online chess and amount of cheating in low level/anonymous chess is insane. In 90% of my games, when my opponent is on a losing position he/she suddenly pulls up 15 move stockfish combo and beats the hell out of me.

A behind the person camera with larger view coupled with a side camera is much better solution. if the camera is seeing the whole view within the max peripheral view of a person, then it takes a lot away. And side camera can see anytjing infront of the person without showing the face. Key point is eyes can only see whats in their periphery.

Camera plus that test taking software that prevents other windows as well as sharing the tasks running on the system.

I don't think this is related with cheating, at least not significantly. The access to engines during practice and preparation allows players that would otherwise have to be creative at their own strenght to be inspired by a 3700 rated machine. Add good memorization skills and you have that today's 2500s can play confortably at the 3300 level for the first 10 - 15 moves of the game at least. This is compounded by the facts that a. the general principles of defending like an engine are becoming commonplace, b. the fact that lower rated players have the ability to select openings and replies that are more likely to lead to tamer outcomes with less play, skewing heavily in the game outcome to draws, c. the fact we have diminishing returns in additional engine depth, implying that Super GMs get less from engines than GMs. What this in turn means statistically is that we should expect lower rated players to consistently take points from higher rated ones, mostly through draws. The main reason one can consider as to why this is hapenning now and not 5 years ago, is likely due to the availability of cloud computation and engines with high depth at the public in general's disposal as the marginal benefit of further layers of depth cannot be capitalized on by humans, e.g. the 50th or 60th layer of depth is of little value in the context of an OTB game - adding to the diminished returns mentioned above in SGM vis. GM. And also lastly likely because the player pool is now full of technologically literate humans that can yield these technological tools, in the first place. I believe that it is easy to forget just how impactfull small effects can be when compounded across a populatjon through multiple iterations, and even easier to forget how Databases and Engines have shaped how everyone prepares and analyses their games. Although, cheating is likely to have its influence, I would challenge the notion that it would represent more than 1% of the total elo stolen by each elo bracket to the its subsequent. I would suggest using a regression model to test the correlation between elo loss at each 100 rating interval to its preceding, with the trend in player accuracy at each level, using the time series of the appropriate data. I'm sure Chess.com's team of Data Analysts are more than apt for the challenge. Any takers?

As a programmer I can only see this becoming worse as it becomes easier to cheat and remain undetected. It may already be in existence, but using overlays to highlight the best move, or moves would be fairly easy to implement and then, even if you have a camera on the player, if you don't have one on the monitor then you're not even going to be aware that the player is cheating.

Hearing Fabi talkin that Sailor talk 🤣🤣🤣🤣

Lazavic seemed to have a safety-first approach in the games that I saw him play in Champion’s Chess Tour. Slow and accurate, not aggressive or taking any risks. If you’re in the Stockfish generation, perhaps you grew up feeling this is the best way to play, since it’s the way computers play. Like, excluding your ego, and not caring about winning. Just playing safe.

I have a lot to say but for now I'll just say that yes, everyone is cheating but Super Grand Masters are also losing their F'ing minds. A few months ago I was accused by a GM. To make a long story short, I've been playing very seriously for over twenty years. My peak was in the 2300s. In those twenty years I played roughly half a million games (obviously including a lot of blitz and bullet), out of all these games I have 5 wins against GMs and the GM STILL thought I'm a cheater. Right... 5 out of 500,000 is 0.001%. That means that for roughly every 25 thousand games I play I manage to beat one GM in a fast time control. And the GM still couldn't accept it. Sorry people... you're losing your freaking mind and with it all of our respect towards you.

only Hikaru can play so fast- Fabi also mentioned that it is astonishing to see Hikaru play in real life, finding tactics so fast, it was super-human

It's happening everywhere. It's the monetisation people are trying to attract by faking a social media presence wit bots, and then using an cheat engine to try and monetise their gameplay, be it battlebit or chess. Since lockdown people are chasing huge incentives from online entertaining that actually were an aberration and don't exist anymore.8

From the intro, I thought Hikaru was being interviewed in this podcast but never mind 💀😂😂

I don't think ive caught a Hikaru video within an hour.

Here is my recent lichess experience as a rapid 2000 Sensei Danya: Here's this Najdorff variation that Fabi lost to in a blitz out of the opening Me: Oh yay, a real chess weapon as Black [4 out of 5] 800 ELO players playing second best Stockfish move, for 20 moves, every time 2 minutes ahead of me Me: But you are lower than me in Every Area of the Game by 200 pts(Including Puzzles, 3 Check, 960 and KOTH, not to mention, Bullet ) Lichess Mods: There is no Conclusive proof that these 4 players are cheating. There is one Tiny difference between King Hikaru and us low mortals : Hikaru would destroy most of them while they are still cheating, the rest of us could not hope to, even when they(the smarter cheaters) stop after playing 25 near pefrect moves and reaching a decisive edge. On my level 2000 rapid, 1900 blitz, it's 4 out of 5 that cheat. It's a lottery where the odds for an honest game are not in your favor:)

I played a game about a month ago where I had an unstoppable mate in 1, and the engine called that move an inaccuracy so I do tend to be of the belief that the rating system is flawed

Some have a tool make snap shot from screen or table board! to gain in 3 seconds an powerfull answer in lvl 2500 , we old chessplayer work only online with knowledge persons, so i play with my old master (Junior master in west/ost Berlin 1955) each day since two years , when i lose so i now he was an real match.

I love how Caruana is sometimes unintentionally funny

Hikaru is a master of using lots of words to say very little

Or the top players are in a bubble? In the recent classic we saw lots of top players lose against lower rating players.

27:58 Following that logic Shankland must be a cheater considering he's a top 2700 SGM OTB but online he's barely 2600 when there are a ton of CM who are 2800-2900. Clearly that's not the case, hence, this logic doesn't work.

My take on the rating deflation is that it is due to 3 main factor: a) close-pool effects in countries like India or China where some of the players are severely underrated (as an example, a 2300 elo from India/China can often mean a GM level) b) the huge crop of young talents (typically teenagers < 18 year old, especially from india but also uzbekistan and many other countries) and c) the advent of easily accessible super strong engines that levels the quality difference between 2500+ and 2700+ players, especially when these lower rated players play with the white pieces. Therefore I think the rating system should be updated, taking into account all these factors, to make it so it is less "punitive" for superGMs, in order to get more superGMs back in open tournaments. Because rating-wise, it is currently not worth for superGMs to play open tournaments where they are going to draw against a bunch of 2500-2600 rated players and therefore lose rating points

It's much harder to detect cheating if they do it sparsely. Like say a person cheats one in 50 moves. Are you gonna find out if they're cheating? Probably not if they're any kind of intelligent about it.

Not only that 2500-2600 needs to win 1 or 2 super grand masters to win TT but also needs to beat 10 more other people on his lvl wich should not be possible often.

We gamers have been dealing with this problem for years in our online video games. Game devs have largely been behind the curve on dealing with it. Chess organizations seem only now to begin to realize how dire the situation is for the integrity of the game.

He doesn't have an ace of his sleeve, he has a bishop

Camera is not going to solve cheating online, even if there are multiple cameras.

Regarding the statistics provided by Kramnik, we need to see the standard deviations. What is the range/skew vs just the mean of accuracy... otherwise even if there is a 5% difference between 2 players, it could be statistically not significant.

I really fail to see how cameras are going to help. I can thing of a lot of ways to defeat cameras. Option 1 duplicate screen in the next room where someone replicates moves with an engine. After that hiding a device outside of the cameras view to return information to the player should be trivial even with multiple cameras. Hell you can even connect a second mouse in the next room and have the person with the engine do the moves.

I disagree with one point caruana makes about players being better online than over the board , yes huge gap in ability is questionable but I personally would expect players to be stronger online than at board since that's where most people play and learn now