for sure it makes it impossible to have a natural streak of games. even in sessions where you versed no cheaters, you're so worked up about the possibility of facing a cheater that you self-sabotage. not something we should have to deal with when we play Chess.
Preach brother. Losing a game to a cheater, especially when you can tell it's a cheater, ruins multiple subsequent games, mentally. It's even worse when you see the obvious cheater not get banned.
hahah yeah its pretty insane. i think only a Chess.com fairplay analyst could answer the questions we have, and they never would give insights to how their anticheat works to prevent workarounds. it does make you wonder how much people are getting away with. glad you enjoyed though bro
I also often have win streaks of 7 to 8 games, I just don't have accuracy over 90% for all the games. My elo atm is about 1550 but my peak is 1710 about two months ago.
i'm not 100% sure on this but it would make sense if paid accounts on Chess.com had priority when it comes to in-game reports which might explain the anomaly. just a guess though & you'd need someone that works for Chess.com to know for sure but i'm fairly certain this exists for other games reporting systems. there are also many exceptions to the winstreak rule. for example, someone who has 2400 Blitz but is rated only 2100 Rapid will be far more likely to have large winstreaks as they're playing at a level lower than they should be. i wouldn't be surprised if you fall into that bracket also.
@@ResonantFrequency-sn9ud winstreaks on their own have exceptions. for example if someone is 2400 Blitz, but playing at 2100 Rapid they're more likely to have higher winstreaks than someone else their level. this is simply because they're playing below their skill level. you probably fit into that bracket considering your rating swings, if you go down 200 Elo, you'll have big winstreaks on your way back up to where you belong.
While I don't play online and only play in over-the-board tournaments (because I see 3-dimensional positions better than flat 2-D ones onscreen), I'm not too concerned about cheaters in these casual games where no money is involved. Cheaters in online tournaments with prize money, however, is entirely different.
100% agree that more priority should be given to prized events - but its the covering up of a real issue that concerns me. Chess.com insisting time and time again on the latest State of Play's that there isn't a big cheating problem in regular play and that they're doing "everything they can" to do their part. at the end of the day people pay good money for memberships, and invest hours of playtime on the platform and half of them don't even know a sizeable amount of their opponents are using engine assistance. if it were more publicly known i'm sure people would just switch platforms, however i think its been swept under the rug for far too long.
No man, some of us want to do a 20, 30 or an hour of competitive and fun time per day playing chess at home, and because it's not a pay tournament it does not mean that it's ok or it's not a big problem, it's the time of people that want to learn and improve. How many people maybe stop playing chess after a brutal loss against a cheater but they never knew it?
Chess players don't understand what 3D positions mean. When you play with real chess pieces and board, it's still 2D. Chess with 3D positions has multiple layers and pieces can move between the layers. In Star Trek, they play 3D chess. Although the rules are not explained.
@@lucianorodriguez7726 A serious player is not going to stop playing because of a single or even a series of crushing losses. It happens period. You learn from the experience. Yes it is not pleasant but you still learn from the experience.
No, he doesn’t. He solely relies on statistics to form his conclusions. This video provides clearer explanations and more reasonable arguments than anything Kramnik has posted over the two years
@@yefer-gz5vu You don't know what you are talking about. Did you even see where Kramnik reviews games of these cheaters? Probably you just watch Nakamaru and that's where your opinion coming from.
It's an almost unsolvable problem. Players can just cheat and use all kinds of smart tricks, ie think long when the software tells them it's a !! move, to hide it. There is just no way to separate a smart cheater from an inconsistent player. Just accept it and play more over the board. Or really go rigorous methods that you require videostreaming while playing, that's the only somewhat feasible way to make sure your opponent isn't cheating.
playing OTB is not a solution for this. if i want to play a FIDE rated classical tournament, there's essentially 2 per year in my city. if i travel interstate to play it'll set me back over $1000 in travel expenses, accommodation, entry fees & food. even if i only play nationally rated Chess games which are held more frequently, i'm limiting myself to a pool of around 50 players that i can ever play against (that are around my level anyway). also to play in any kind of OTB Chess event i'd need to take a 45 minute train ride into the center of the city, and the same 45 minute train ride back. even still these tournaments would still be infrequent, usually only on weekends and sometimes weeks go by without one being organized. its very easy to say "just go play OTB" if you're European or American, but in Australia at least its completely different.
I actually think as the neural nets continue to learn from human play they'll be able to tell the difference. Just like engines have a unique way they play, so do people.
Like other online games you can't have open servers, you have to have your own vetted private server. Games like GTA are totally a waste when played on the open online servers.
@@howard5992 of course people cheat. the insane part is spending 20 hours making graphs of people that "didn't cheat against me but cheated against someone else so that is cheating and I should get elo refund for all of my games"
I player a cheater otb in a national tournament who was cheating. There was no checks or anything being done so he would just leave on my turn go the bathroom and come back but i still managed to draw the game because he got to 5 minutes while i still had over an hour left, I sacrificed my knight for his promoting pawn which complicated the game and i managed to draw him.
I once started playing a guy and I suddenly got diarrhea. It was a seriously painful experience! And I'm sure he got suspicious until I started making mistakes and lost the game. I was constantly going to the bathroom. I was much younger and much more embarrassed back then. It happens.
There was an account I played around 1700 elo rapid that was about a month old. He had just got to my elo from 1000 elo and had lost a few games at that low rating, but handled me easily which I found suspicious. In the next few weeks the account was 2200 elo. Almost certainly a cheater and I was never refunded elo
True the worst feeling for me is to know and almost know my opponent is cheating and report him and he still isn't banned then i have to convince myself maybe he wasn't cheating though i know he was 😢
Jack for your own rapid improvement and my gross curiosity I would love if you were to play rapid on lichess. I personally feel like I have faced way less cheaters in the lichess rapid pool vs chesscom pool. And once you settle in your fair lichess rating range I would love if you were then to play 100 games and do a similar analysis as done here. ------ I'll add some notes here: Way more higher rated and strong players play rapid regularly on lichess. Notably GrandmasterGauri on the streamer side and plenty of other strong FMs, IMs, and even GMs regularly play rapid on lichess. As you said it'd often take you at least 5 minutes to find a game in the cc rapid pool but on lichess when you are rated accurately I'd be willing to bet you'd always get paired within 30 seconds. I do believe chesscom's cheat detection and especially refund system is better than lichess. However, by design lichess entices a higher percentage of strong players (fair players) vs it's populous size.
yeah i'd definitely be interested in doing this, might try and play some lichess Rapid on stream soon and have some fun with it (bonus points for lichess not saying JACKSARK IS CURRENTLY STREAMING HERES A LINK WHERE YOU CAN SEE ALL OF HIS MOVES AND PLAN AHEAD BASED ON WHAT HES THINKING
@@jacksarkisian lmao, nice looking forward to it. Btw your analysis of these games really has shown how much stronger you've become. Congrats on the well deserved improvement ❤
@@Willeinhelm Yeah I wasn't expecting him to play on stream. As Jack practically only plays rated rapid off stream I'd assume it'd be fine for him to play lichess rapid off stream as well.
Amazing video, much of it was things ive thought but never took time to look over (closet cheaters, account time, etc) thanks for actually bringing these things to light
thankyou bro. i would love to see a GM like Danya play ~100 Rapid games at this level and do their own analysis. i can't say with 100% certainty but i'm very convinced i versed more than 29 cheaters in my last 100 games its just i'm not strong enough to know for sure if they were or not. glad you enjoyed bro
@@jacksarkisian yea i also think so. im also curious if a gm like danya does it. he sees alot of cheaters in his speederuns already so i think theres an intersting fact
@@jacksarkisian In one of his speedrun videos he talks about this and says the amount of cheaters is the reason he does not want to continue the rating climb beyond a certain level.
@@Shadow-7295 which video is this in? i heard someone mention this before but i couldn't find the video. wanted to mention it in this video but couldn't find the clip haha
the only case this happens is when a Chess master is playing on a secret account. they play unbelievable moves and because they are quite literally comparable to a low depth engine they can be banned for it. aside from this it doesn’t happen ever with the current anti cheat.
For those saying "just play otb brroo", keep in mind that security is non-existent for all lower level (think city, district etc) OTB events, FIDE or not. Of course its still not as easy to cheat as online but I expect with even more miniaturisation of electronics in the coming years that soon cheating will become a menace even at OTB events UNLESS all of them religiously put in place metal detectors and body scanning. I don't small events able to afford the expense or manpower. This is just an additional point on top of what OP already said in reply to another comment regarding travel expenses and frequency of such events being poor or non-existent in most parts of the world. And OP is 1st worlder. Imagine a Mongolian or Congolese young talent... either having to face endless cheaters or travel 100s of miles with his non-existent cash for a single OTB event...
100%. even in Australia we have ~2 FIDE rated classical events per year in my city. to travel to a different city & play it'll cost over $1000 to play (travel expenses, entry fee, accommodation, etc). you could play nationally rated tournaments, but those aren't even held fortnightly meaning that people who recommend you "just play OTB" are essentially recommending you play one Chess tournament per month. obviously i'm still privileged to be able to play OTB as often as i can, but even in a country like Australia its not an optimal solution.
Great video. I couldn't tell you the amount of times I have a winning position (2.0-2.5 advantage) going into the middle or end game only to have the opponent play out of it and beat me. Happens way too often.
In my video, I showed an account that was many years old and had thousands of games. It played bullet, and every move was like 0.7 seconds. Even that was too hard for them to detect. Oh, and the same for the guy playing crazyhouse with every move made in 0.2 seconds.
actually saw your video the other day & made me get around to making this. super disappointing that there’s been virtually zero progress for the anti cheat in over a year now, great work on your video though 🔥
@jacksarkisian Oh wow, I'm glad to see an update, in that case! I was also wondering what the cheating rate would be if I played a lot of games now, and you did just that!
@@AyrtonTwigg hahaha yeah would love to see you make a follow up video in the bullet time control to see if theres been any change there. sadly i'm incredibly slow and only ~2050 bullet so i don't verse many cheaters but would love to see a comparison. i just got an email from some Chess.com representatives about this video so hopefully will be able to provide some more clarity to the situation over the next few days.
im a 2100-2200 rapid player myself and I can only agree with you. When I play on one of my numerous secound accounts (rated anywhere between 1400-2000) I rarely, almost never) face a cheater. But once you hit the 2100 rapid barrier around 10% of the accounts I face look sus as hell and I abort alot. I imagine that at 2300 the issue is even worse :( sad stuff and that's also the reason why I stopped playing rapid chess as much (gone from 40 games a month to 15).
this might be a thing i'm not too sure. i know titled players & Chess.com partners/staff are reviewed by the anticheat and by a human but not sure about paid accounts - could have something to it.
Cheating is rampant in every online video game right now. There are AI anti cheats being developed and in a year or 2 cheating in video games will be almost impossible.
Top 3 signs of a closet cheater. #1: They make rookie opening mistakes despite having a high ELO #2: They suddenly play perfect chess after a blunder #3: They play many top moves without thinking
This can all be explained by just playing chess though; try something because you played the same opening 30 times in a row, after a blunder they get a different mindset and try harder, just trying things and by luck doing some brilliant moves. If it's constantly in every game, it might be cheating, or it's a genuine genius stuck at 900 elo (😂😂😂)
I think the 30 day limit might be based off the assumption your lost rating will revert just through play since if you are underrated your rating will self correct.
yeah this would make sense but i think due to the amount of cheaters in the Rapid pool its less effective. paired with the fact that Rapid is a longer time control and you're far less likely to gain it all back because the games are much longer. would still love to see a more dynamic solution to this. maybe if they found a game where the player cheated, games played 15 days before and after that game would be refunded? if this was applied for every game they cheated on this would seem optimal especially if there were numerous instances of cheating over long periods.
yeah super sad. the worst part is even if they're blatantly cheating Chess.com are still yet to implement a way to remove someone from your OWN arena. even if i own the club and started the arena myself i still can't remove people from a live arena. this is a feature that was requested over 2 years ago and still hasn't been added.
Why don’t you make the tc something faster for the arenas like bullet, it discourages cheaters (using bots are much easier to spot). Even 2+0 is better than what it currently is, the players winning these arenas are usually quite high rated so bullet will still get a decent quality of chess.
@@gonnabegm 3+0 is what jack's arenas use. Because he wants it to not be too long and not too quick. And also 3+0 is what most people play anyway, why complicate it with other time controls most people aren't used to? And also before giving out the prizes it needs to pass the fair play check first.
@@gonnabegm because it massively excludes more amateur players. i actually took off the minimum amount of games required to play for the tournament because it was affecting more legitimate players than it was preventing cheaters. if the time control is too fast, you'll just end up with an incredibly tryhard tournament in whats supposed to be an event that allows newer players to grab a free membership. sadly i think the way we do it now is optimal but if Chess.com ups their game things can improve massively.
i think this alongside with titled players make sense. of course i don't think they should be given more leniency, but in any case in which a streamer or titled player has suspicious play on their account they should be carefully reviewed by a human. even though i think Chess.com has a problem with letting too many cheaters off the hook, the opposite in banning players on very minimal evidence is also not the solution either.
Have to agree that after 2300 you will face loads of cheaters when playing rapid and that really affects your game. In every single game you wonder if the opponent is cheating or not. This affects your play and it is very hard to play the way you normally would. To put is shortly I very rarely play rapid games nowadays as it is almost 50/50 if you face a cheater. Ps. I would also like to add that I have 2500 blitz rating and 2600 bullet rating (records) but in almost every time it is me who gets low on time even if my opponent has blitz rating of 2100 etc. It is a really strange feeling.
Your claim that skilled players don’t blunder their queens is factually incorrect. I’ve seen blundered queens in Titled Tuesday games multiple times. Let’s be honest, your metrics are subjective, you have no data, you have no skills to process even the data that you have, and you’re saying that your excel exercise is worth more than a team of analysts and engineers who work around the clock. This cheating accusation madness should stop.
there wasn't enough context in my statement. i meant 2200 players don't ever blunder there queen LIKE this. by like, i mean straight out of the opening with literally 10 minutes to think about their moves. of course there are times where someone will be on some kind of ungodly tilt, or has come home drunk playing Chess but this is why i said the sample size was too small & they were clearly banned for other games. however this is still an important piece of evidence to put towards their account being illegitimate. queen blunders happen all the time, theres definitely games in my last 100 where i've fully blundered my queen but its much more reasonable to happen at latter stages in the game where you're running out of time - the same can be said about titled tuesday players as the time control is much shorter. and Chess.com do not have a massive team of analysts. according to their own website they have 2 and i don't think its unreasonable to think at some stage they've missed the mark with their work.
@@jacksarkisian Even tho I'm only 1950 rapid, I have blundered my queen, bc of my friends or when I was generally distracted. I understand, bc even at this low elo of mine there are still many cheaters. After the 99 precentile people become too competitive lol.
I don't cheat but definitely think i could get away with it. If you're already decent and only check the engine when in key positions how would the fair play team catch you? I can't think of a way. The signal to noise ratio just isn't robust enough so it seems you can only really reliably catch obvious cheaters. Of course I don't see the point in cheating. Is winning a game where you had to check the engine satisfying in any way? I guess there's the shallow thrill of getting away with something but I dunno how thrilling is it really when there are literally no stakes? Maybe a former or current cheater can explain their motives?
Soemone might be 2200+ and blunder their queen just by being a mixed of tired (playing several games in a row) and drunk/high or just simply losing concentration due to other people in his room talking to him etc. It's not impossible imo
definitely not impossible but highly unlikely and as i said in the video, it cannot be your sole piece of evidence to prove they were cheating (unless you have some highly advanced and reliable AI generative learning anticheat that says otherwise). but clearly even though that sample size was too small, they were cheating in other games and were banned for it. the instance in this game simply backs up the fact that they were cheating and is useful in hindsight.
@@jacksarkisian Yeah I did realise your point. I was just picking on semantics I suppose where you had said "it's impossible to happen at this rank" or something. Not a big deal, I definitely didn't mean it as a criticism. You made a good point about "closet cheaters" that I hadn't really considered so much.
Online I'm 2500. I've got a few thousand games and have never blundered my queen to a 1 mover. Had my queen trapped a few times but that's a different story
Very well put together video. However there is one critical flaw in your analysis in the result section at around 15:00. The sample size that you are taking is very small to follow a normal distribution hence inflating the statistics. If you are interested I can help you develop an extensive and robust Machine learning model which could run through your account and find those accounts which are likely to cheat.
hey! this was a thought i had, however i think the following issue would come up. i did consider increasing the sample size to something much larger, but if i go back lets say 250 games into my Chess.com history my Elo would be around ~2000. not only is that 300 Elo gap absolutely huge in comparison to lets say 1400-1700 Elo, but i would also assume the higher rated you are, the more cheaters you will verse. i feel like if i use games from my personal account this factor would deem the data much less valuable as the Elo level i played at was not a constant. i would also assume finding an account with a suitable dataset would be quite rare. not only are there only around 3000 active players in Rapid rated around 2300 Elo on Chess.com, but you would need to find someone who is completely stagnant in their Chess and has played 250+ games in the last 2 months (2 months is also a generous timeframe for various reasons (cheating program availability, real-world events, Chess.com site changes, etc)). i suppose you could take the same 100 games over a 2 month timeframe from other players similarly rated on the website, but i still don't know if there are enough people at this level to acquire a large enough sample size. i wouldn't imagine my findings are hugely reliable, but i think it at least outlines that there is a problem somewhere which is the main point - would love to see a more extensive dive into something like this though!
My personal favorite cheater story of mine is when I played someone who had an antisemitic name, was racist in the chat and was clearly cheating. I reported him for cheating and he wasn't banned until roughly 2 months after. And I didn't get refunded, presumably as he was banned for the antisemitic name and racist comments.
someone mentioned that paid members require a human review from Chess.com to be banned, and considering they have 2 fairplay analysts you could imagine the delays considering literally anyone with Discord Nitro right now has a membership.
I'm not positive the "closet cheater" was cheating. Kh1 could have been a nearsighted move to get the g rook active on that file. If it was a mouse slip, then I still don't see how they are moving the mouse and getting Stockfish best moves all within 1 second or so, sometimes less. By saying mouseslip you're implying they are manually moving the pieces and not using some advanced automove cheat, but they are still moving as fast as under 1 second? Doesn't seem possible... The move times do not feel suspicious. I'm 2000 bullet and could intuit the rook sac given the activity of the other pieces. It also just feels like a desperation move that luckily happens to maintain a draw.
people use browser extensions to cheat on Chess.com. it shows an overlay of your Chess board and draws an arrow from the piece to where it should be moved to. afaik automove isn’t really a thing anymore unless you set the times to random. if you’re automoving in 0.1s every move you’re obviously going to be caught, most people do it semi manually now. this kind of erratic decision making might be common in blitz or bullet but it’s nonexistent at high level rapid. no one plays risky moves like that in such short amounts of time, you at least think for 15-20 seconds otherwise you’ll end up throwing more games than not.
thankyou bro but i do still believe it'll be in vain until someone with a larger following covers the topic. i've built a great community of people over the years and i think my voice is heard a lot more than others but the issue is simply too big for someone like me.
How do you know some cheaters play badly in the opening to lower their overall accuracy and play perfectly in the other phases 2:18 ? The example you showed is just some dude blundering his Queen and resigning, where is the perfect play afterwards 2:35 ? Also, people can make all kinds of mistakes even blunder the Queen. On what do you base your suspicion of having win streaks 3:33 ? So you personally have a 6 game win streak and thus a person with 7 or 10 is suspicious? Why? On what do you base your description of "closet cheating" How do you know they typically cheat to 1-200 elo higher rated than they are 4:03 ? How do you know that they usually cheat in the game after they lost the previous one? So your entire argument that this guy is cheating is that he knew how to refute your opening, he didn't fall for your mate in 1 trap, he played Kh1 and not Kh2 8:50 ? And after realising the mistake went all in? And based on this they are "undoubtedly obvious cheating" 11:05 ? How do you know that chesscom in recent years made a public effort in recent years to reduce the amount of false bans they give to players 11:10 ? 11:20 what do you mean they are given full immunity and will never be banned 11:24 ? "The opponent clearly meant to play the best engine move, mouse slipped and then suddenly found the best way to force a draw 11:54 ". Such an interesting wording. Also, how do you know that he wanted to force a draw? How do you know the rook sack is not an attacking move, and you just happened to escape the attack? How do you know that are people that are cheating under our noses that will never be banned 11:23 ? How do you know that they are cheating in a way that is undetectable by chesscom 12:28? So chesscom banning someone is not the seal of approval but your cheating analysis is? 12:37? Every game should have looked at by a the anti-cheat, did you report every player you played for the last 100 games 14:48 ? How is it obvious that they were cheating against other people 16:05? So if you claim that they are all cheaters how are you beating 50% of them? If they were truly cheating you wouldn't score 50/50 with them like with average people right?
How do you know some cheaters play badly in the opening to lower their overall accuracy and play perfectly in the other phases 2:18 ? Because it's a common trend when versing cheaters. I play viewers on my livestreams all the time and you’ll constantly see players stumble out of the opening with low accuracies, then absolutely dominating the middlegame and endgames (this is also myself playing against players with Elo’s as low as 800). Those accounts are then subsequently banned and it becomes a pattern that users that play legitimately do not follow - only cheaters. The example you showed is just some dude blundering his Queen and resigning, where is the perfect play afterwards 2:35 ? The two statements were not connected for this example, if you blunder your entire queen in the opening with no compensation even an engine won’t save you against a 2300 rated player - hence the resignation. It was also against a proven cheater, there is no speculation they were cheating as their account was banned. Also, people can make all kinds of mistakes even blunder the Queen. On what do you base your suspicion of having win streaks 3:33 ? 2300s do make mistakes, but blundering the queen straight out of the opening is typically not one of them. 2300s can blunder their queen in complicated positions where they have lower times, but straight out of the opening is almost impossible. As I mentioned in the video though, this is NOT enough to determine whether or not someone is cheating as the sample size is too small, but if you quickly analyse their account you’ll see they were blatantly cheating against other users. So you personally have a 6 game win streak and thus a person with 7 or 10 is suspicious? Why? On what do you base your description of "closet cheating" How do you know they typically cheat to 1-200 elo higher rated than they are 4:03 ? My 6 game win streak should be considered suspicious. It's an extreme outlier in data and should be considered when determining whether or not someone was cheating. The opponent in question also had a staggering 14 game unbeaten streak in the pool where they scored 90%+ accuracy in every game they played. Additionally, closet cheating is typically when players, in response to losing a game decide they’re going to cheat to win the Elo back instead of fairly fighting for it. They mostly use their own moves, but if they think the position is getting away from them or the moment is critical they’ll use an engine. They’re typically higher rated than they should be because they lose Elo fairly without needing to gain it back fairly, making them inherently weaker than the people they verse. How do you know that they usually cheat in the game after they lost the previous one? So your entire argument that this guy is cheating is that he knew how to refute your opening, he didn't fall for your mate in 1 trap, he played Kh1 and not Kh2 8:50 ? Because that's the entire premise of a closet cheater, they won’t cheat unless they feel like they have to. If they’re legitimately winning games on their own they have no reason to cheat, but if they lose a few games and can’t win on their own - they use an engine to help them. I also explicitly said a player knowing deep Sicilian theory was not a sign of cheating, I also never insinuated him not blundering M1 was evidence of him cheating either. And as I stated in the video, my evidence of him cheating was him mouseslipping the top engine move, and then immediately finding an entirely new idea that was perfectly functional without even calculating it. And after realising the mistake went all in? And based on this they are "undoubtedly obvious cheating" 11:05 ? First of all he had no time to even process what he did might’ve been a mistake because he was playing so quickly. Second of all if any legitimate player thinks they made a mistake, they don’t dig a deeper hole by sacrificing their rook without calculating it - that's complete nonsense. How do you know that chesscom in recent years made a public effort in recent years to reduce the amount of false bans they give to players 11:10 ? The CEO of Chess.com IM Danny Rensch has stated it numerous times in recent Chess.com State of Plays. 11:20 what do you mean they are given full immunity and will never be banned 11:24 ? Due to the functionality of the anticheat on Chess.com, it is unable to detect anyone cheating in this fashion and therefore they cannot be banned. "The opponent clearly meant to play the best engine move, mouse slipped and then suddenly found the best way to force a draw 11:54 ". Such an interesting wording. Also, how do you know that he wanted to force a draw? How do you know the rook sack is not an attacking move, and you just happened to escape the attack? Because there is no way for either side to win in the line he chose to play. He forced me down a line where I would be down a rook if I didn’t repeat moves in the position. Even if I just “happened” to escape the attack, that's why you take more than 3 seconds to calculate a move in Chess - to make sure it actually works as you intended. How do you know that are people that are cheating under our noses that will never be banned 11:23 ? This is an assumption that a cheating software paired with common sense can be passed as legitimate play in online Chess. How do you know that they are cheating in a way that is undetectable by chesscom 12:28? So chesscom banning someone is not the seal of approval but your cheating analysis is? 12:37? Because Chess.com doesn’t ban them for what is obvious cheating. Also Chess.com is not the seal of approval because as I stated before, Chess.com largely prioritises preventing false bans which also allows others to slip through the cracks. Every game should have looked at by a the anti-cheat, did you report every player you played for the last 100 games 14:48 ? I mentioned this at 14:05. How is it obvious that they were cheating against other people 16:05? This was a general statement at players cheating across the board. I bothered to analyse their games against other players and found evidence of cheating in those. So if you claim that they are all cheaters how are you beating 50% of them? If they were truly cheating you wouldn't score 50/50 with them like with average people right? I mentioned in the video the reasons you might beat a cheater at 16:23.
@@jacksarkisian 1). You didn't show an example for that, and how playing poorly in the opening and dominating in the other phases affects the overall accuracy, and how do you know that's the goal of a cheater and if it actually works. 2). Just before you showed the example with the Queen you reminded us, that it is connected to that 2:36. And yes he is a proven cheater, but you are making it seem that the fact that he blundered his Queen is somehow connected to him being a cheater. 3). Here are the two recent Rapid games where you as a 2300 blundered your Queen with plenty time on the clock game/live/114925732439 game/live/116230227659. 4). On what basis do you believe your 6 win streak is an outlier? So in this sense how staggering 14 win streak is actually? Is this guy suspicious "MO2002ISMAIL" ? I understand now, so you say that their elo is inflated because they cheat in the next game after they lost and gain that unfair elo premium. So what do they do after they lost 2 games in a row, they cheat 2 games in a row in response? And also if they are inherently weaker doesn't it mean that won't be able to compete on that level and would need to use the engine more and more until they are solely using the engine? 5). So what did you say here, maybe I misunderstood 8:27. I don't belive that basing your accusation on move Kh1 is enough of the evidence. P.S I think your argument would be way stronger if you had mentioned the time the player in question used during all of his moves and not just during the rook sack. Because he was playing most of the moves with 0.2, 0.5, 1.0 time frame. 6). I would consider the rook sack to be the tactic with obvious threat of Rg1 Rg7 mate. Most of the Blitz players are really good at tactics and attacking, and blitz players don't really tend to calculate it all through, it's more about intuition (sensing it's dangerous enough). You think of it as he calculated 5 moves ahead that it was a draw and went for it all in 3 seconds, but it most probably is him trying to continue to attack you which has happened to be a draw. In a similar situation, his attack could have failed and you won, and you wouldn't even slightly suspect that game. 7). I don't see the reference. 8). How do you know that due to the functionality of the anticheat on Chesscom, it is unable to detect anyone cheating in this fashion and therefore they cannot be banned?. 9). Yes but you blow that assumption into a basis of your argument as if it's a factually proven situation. It's an idea that this can be possible, but for now, it's only a speculation of the community. Because all those who try to "test" it get banned. Of course, cheaters won't be banned for using an engine for 1 move, but how influential that move would be to the outcome of the game? When to know at which point to use the engine? 10). It is not clear what you claim to be obvious cheating that chesscom is refusing to pay attention to. I am quite certain that with the right game and with the right sample of data, the same argument can be made about your account, You already mentioned that your 6 game win streak is suspicious, so all that needs to be done is to find the game that would in some way illustrate that. Your account is way younger than mine "check", Your account has suspicious win streaks "check", You blundered your Queen in rapid two times recently "check", Find a suspicious game "most probably check". And now what is left is to report you to Chesscom. Oops they didn't ban you. Well, you must be one of those undetectable cheaters well. 11). Sorry I for some reason completely missed that. Thank you. 12). Would be nice to see some examples, and on what basis you actually figured out that they were cheating. 13). I understand how you can beat 1 or 2 cheaters, or maybe even beat them regularly, but how are you beating 50% of them. Maybe they aren't cheaters after all? If we assume that all of them are cheaters and you beat 50% of them it means that they didn't cheat in at least the 50%. Now if we take the other 50%, does it mean that all of the games you lost are to people who cheated in the game? Or we can use the same assumption that some of them cheated against other people and not you? Does it mean that more than 50% of the game let's say 60% or 70% of the games you played are against people who didn't cheat during that time?
When I was a kid I liked chess. Then I found out apparently some people just memorize fully what they have to play on every possible opening. Since then I realized it isn't really the game for me, as it turns from a thinking game to a pattern recognition and memorization game. Now add to this cheating with bots... I feel maybe it is just too late for chess
in my opinion Chess just hasn't evolved with the times. world championship matches should be played in shorter time controls to prevent extreme memorization and induce mistakes. i do think that memory should be a part of the game, but it shouldn't be the main characteristic of it - creativity and problem solving should be. shortened time controls prevents cheating as a whole in OTB Chess and would also be much more exciting to watch as a spectator - i think anyone that's watched the Speed Chess Championships lately would agree with that.
@@jacksarkisian Yes, time can make a big difference for sure. Watching high level fast play is way more fun than absolute stockfish perfect slow games. Or course in any game there is some memorization...and to say I wasn't tempted to just memorize a whole opening would be a lie, but unless you plan to become a GM I felt like it would just suck the fun out of the game. Well of course the 100th time you see the wayward queen you will have it memorized, lol. Cheating on the other hand... I don't understand what those people have fun with, at all. They don't have fun, they don't improve, and they can't even do it professionally. Just for an ego boost.
Online chess has its problems. While you do make some valid points that the number of cheating accounts has been growing, I do need to say that a player is innocent until proven guilty - that's the basic rule. Based on a single game, all you can do is to say this guy is "likely" a cheater- hell, even a random mover has a non zero probability to plays 100% perfect moves! Also, 100 games is not big enough as a sample.
i agree, 100 games is not a big enough sample size but if we increase it to 200 games, we go back to when i was ~2050 Elo and my Elo is not a constant anymore and the data becomes much less valuable. i'm also fine with you saying the best you can say he's "likely" a cheater - but in my opinion even if its not definitive, i don't want to be versing "likely" cheaters 33% of my games.
A pleasure to watch this video. Its goal is Fairness and it was informative. Effective communication throughout. Clear, well organized, well reasoned, right degree of detail, right tone, and concise.
thankyou bro i appreciate it. planning to have a discussion with Chess.com's head fairplay analyst in the coming week & will make an update video on this
hahah there are of course exceptions to this. for example if you're a player that likes to trade pieces off the board early, your games will consistently have higher accuracies because there's less possibilities. but if you like to keep pieces on the board, your accuracy will be lower on average because of the increased amount of possibilities. openings also come into play with this but generally one single suspicious trait usually doesn't indicate foul play
from what i've heard, Lichess' anticheat is similarly comparable to Chess.com's. their situation is slightly different though, they have a much smaller player base with a much older age demographic. its hard to say for sure but you could imagine that 25+ year olds are much less likely to cheat than children just on a maturity basis. a lot of people who do cheat use browser extensions to do so as well. from what i've seen, most browser extensions are compatible with Chess.com and not as commonly Lichess. meaning if someone does want to cheat they're more likely to do it on Chess.com. i think there are just differing variables that make Lichess a platform with less cheaters, i'm not 100% familiar with their policies on things and how they approach cheating but you would imagine they take a different approach to it considering its a non-profit charity as opposed to Chess.com being a profitable business. realistically if anyone is going to be the forefront in innovating engine assistance detection, it should be Chess.com.
I play in Lichess since I started playing chess and I didnt play against a cheater in 580 games. But if you dont with a account you will find cheaters, at least they cheat without account so you dont lose elo
No it's worse but it's better. To make sense of that sentence; because you can expect cheating on lichess practically 80% of the time you end up playing more solidly trying to prevent mistakes rather than creating an attack that is not realistically correct.
@@equinøx984 Mate they just don't review their games I ran a statistical analysis from numbers provided by them publicly and they not only made no sense but would even in effect be void in counter-acting the problem as there are far too many games to be checked.I know this 10000000%, a user cheated for months without being detected. They, in honest, conversation explained everything to me as they themselves couldn't believe how on earth it was allowed. They reached far above 2000 without ever being detected by combining their own strength of play with an engine and from what I was told they would only ever use moves given by the engine if by then seeing them they could understand them at the least loosely, Apparently they were able to stay in a realistic range of centipawn loss and by that get completely dismissed by the detection system. In their head they started doing because they hated being cheated against and saw how it worked but nothing was being done about it so they just said **** it. Now imagine how many of them are out there.
Lichess errs on the side of not banning anyone unless they are extremely sure of cheating. So you will see less people banned for cheating, which probably results in less tilt.
the % of cheaters per 100 games is definitely different at varying ELO levels, but i'd imagine its still pretty abysmal regardless. i've been playing Blitz a lot lately instead & seems like i'm not at an ELO where cheating is a massive problem. 👍
I don't understand why they don't make a 3D-secure login function. Ban this and you will have to use someone else's Personal Identification Number- which one would run out of fast. With absolute certainty, make a "cheater pool" cheater vs cheater only, unable to player normal pool.
i know Chess.com has been looking at this in the past but as i mentioned in the video, this would need to be paired with a massive promotion of Rapid Chess (possibly getting Grandmasters on board to play again). if not, you'll just have a tiny pool of players where everyones playing the same people and waiting too long for games. might be the only good idea, i could be wrong but i think this was CounterStrike's solution to the problem with Premier.
@@jacksarkisian Force everyone to use 3D secure login. The player pool would stay the same. I would even be okay with playing the same person every 5 games to keep cheaters at a all time low. People are literally using Python scripts to automate creating new email accounts to burn.
hahah i got an email from Chess.com's head fairplay analyst & will hopefully have a meeting with them over the next week or so. will probably make an update once that's happened to clear anything up.
Why do some players play a few moves, then resign or lose by timeout? I am not talking about just a few games like that. They play many games and most are like that.
i've noticed this too which is weird. i would imagine its connection issues with chess.com or just internet issues in general. i think it might've happened 2-3 times over my last 100 games? regardless if this does happen & you email Chess.com's support team they'll refund you the Elo for this.
I know this problem, and I am only 1300 elo. There was a stretch of time when I said in chat before every game: "Lets have a fair game! If you cheat in any way, I am the winner regardless of the result. Thank you." This means cheaters lose against me 100% of the time.
@@consciousmassofatoms they cheat, because they want to feel like they are good at something. They want the other to think they are amazing player. If I write that message, then they know I wont think they are an amazing player, because I know they might be cheating, which takes the joy out of it for them. Also, they will know they cheated, so obviously I won regardless of the result, so It won't feel like a win to them, their subconscious will know. Chesscom should have a message like that which is a direct hit to the ego of a cheater. Self detection is the best prevention, make it unappealing.
I haven't played in a while. But the last time I played a for a little while I ended up getting my ass kicked and over the next few days got almost 50 elo points back from them banning cheaters. That's a lot. Wish they told me which games they were. Basically all the games I lost I guess. Lol
if you go into your game history and see which players have a red cross next to their name you should be able to see! shame they don't make it easy for you & tell you who it is in the message but its better than nothing haha
@@jacksarkisian That's true but that symbol doesn't show up when you're using the app playing from your phone which is how I play it. So I'm forced to check all the games I lost recently to try and find out which game was it. And a lot of times I lose patience and just give up because the game could have been months ago
i hear you bro, i just got fed up with this completely in rapid. I'm only 1300 elo so i don't think that this has to do anything with elo, rather it does with simply recently there are a lot more cheaters and more sophisticated ones than before. Almost all my games have an accuracy above 80% and this is almost not enough for anything. I was able to finally win a game after loosing 6 in a row against 1250-1300 elo guys who played above 90%. My accuracy was 93% in this game, and the only reason i was able to win was that my opponent got low on time and finally made a blunder with 10 seconds on the clock, but he played almost perfectly until then. Another guy also found some Hikaru/Alireza kind of fork tactic in 3 seconds, with apparently 1205 elo. I just can't believe these games at all.
Idk about the streaks. I had a 9 game win streak yesterday where 6 of them were 93%+ accuracy. I play for endgames so my accuracy tends to be higher. It’s statistically probable as I played 900 games in 90 days to get a win streak like this
i agree i love playing classical and the reason i play Rapid is to practice for it. the only thing is FIDE rated classical tournaments are hard to come by in Australia and usually you'll have to spend around $1K to travel to a different state to play in one which isn't super viable.
Also I’m about 1900 rapid, and I just went and looked at the last 10-15 games and there were at least 3 accounts opened in 2024. Only one did I suspect of cheating until I went and looked closer at the accounts. None have been banned and knowing I would only be refunded points within 30 days is so concerning. Thanks for the video, I’m definitely looking every single time and aborting any game vs an account made in 2024
Maybe they can make a trust factor system like in CS where if a player is reported a lot they will have a higher chance to play vs other low trust factor player, this would also help against smurfs or cheaters just creating new accounts as the starting trust factor can be set to low and only get higher by playing many games without getting reported for cheating. This probably wouldnt affect the new players that badly as new players would start very low in rating and would need many games to get higher elo and to get a high trust factor, also most cheaters dont cheat in low elo so it probably wouldnt affect them.
The only cheating I ever do is when I play against chess engines. If the Elo is at 3300 or above, I set the game up to where it gives me queen odds, and after a challenging 80 to 90 move game, I checkmate the computer with 3 or 4 pieces left.
Sometimes when I think my opponent is cheating, and I start seeing them play weird moves, I turn on the engine myself and compare their moves to the top Stockfish moves. When I see them playing super weird looking moves that are the top Stockfish moves, I already look at them with suspicion. When I then look at their profiles and see some massive streak of 12 wins in a row or a 90 % win rate, gotta smash that report button. I know my methods are morally questionable, but hey, I'm an anti-hero, I'd do anything if it means getting even one of those disgusting cheaters banned.
A quick rule of thumb for anyone: - Winning more than 6 games in a row is certain cheating (5 wins in a row was the longest win streak that I've ever had, I'm currently 1600) - Having an accuracy of over 90 more than twice in a row is certain cheating - Taking more than 5 seconds when they only have one legal move (perhaps they're in a check) is a warning sign - If they for some weird reason resign even though they're totally winning, it's a warning sign, they could be doing this to lower their win rate and look less suspicious
I only play daily chess and since i made my account in August of this year ive only played two people who i could tell where cheating. There is something very artificial about people who cheat at my level 1000. Its like looking at an ai picture, you can just tell something is off and artificial about it. People at my level play very human moves where you can easily follow the motive. Both times i felt the game was artificial, the accounts got banned for cheating (one of the accounts I reported myself)
Same reason why I stopped making videos... Also I don't like the lichess interface. Although I enjoy bullet more because of how easy it is to catch a cheaters. but you don't really learn anything from constantly playing bullet which is also sad
I think it's important to point out that players in the 100-1800 elo rapid pool are far less likely to face cheaters than someone in the 1800-2500 elo rapid pool. Not as many cheaters hang around at lower elos, with the exception of a few. I'm 1400 rapid, and I am refunded elo about once in every 200 rapid games. I suspect some level of cheating occurs in about 1 in 40 games or so.
If ( if ) you are right, then a simple solution would be a filter where you can request to only play accounts that are say over 3 years old or 5 years or whatever you set it at.
This is one reason ive quit playing online chess except against stockfish itself. This is a political problem they are also not punishing cheats in other facets of life.
There's checkers, backgammon and lots of other games... And some people cheat or try to cheat in all of these games. And who is so badly concerned about elo?
I agree with a lot of things in the video, but not about winning a lot of games in a row. I m currently around 1650 elo and I tend to often win a lot of games in a row, and latly not loosing many in a row because I made it a strategy to quit for the day if I loose 2 in a row since I previously ended up tilting
I only play blitz, I can't handle the idea of investing all my energy for almost an hour and realize you played a cheater. I would love to play slower time controls but its just unreasonable
Last weekend I played 50 blitz games. I lost most of them (I'm not a good player anyway). When I entered the game this weekend, suddenly my rank was 100 elo higher. I checked wtf is going on, and of 50 opponents 27 (!!!) were banned (was there a mass banning happening last week?). That's OVER 50% of my opponents... Online chess currently make no sense.
do you think cheating is more common or less common at lower elo levels? i’m under 1000 elo and am wondering if i should expect a similar percentage of cheaters as you
I have no idea on which basis you’re coming up with those numbers on % of cheaters. Very confusing. Would suggest you highlight the approach more clearly.
Its really not clear whether some of the people you claim are cheating are actually cheating. One move isnt enough to convict anyone. The game you showed against that 2300 only had one suspicious move which is reasonable
I always wonder how good these 900, 1300 in the Arena blitz tournaments are. I lost there against 1100 and 1300 (and my rating is much much higher) and even I didn't, they are playing way stronger than you would assume. There has to be something wrong. Look at the standing where these type of players are, and where the 1800er are. Often they leading the tournament by far. That does not make not a lot of sense. IRL you would never see such a strong 1100er, NEVER! I am not the type of person who accuse a lot of individual players of cheating, and I am very cautious in doing that, because I am a very bad at blitz, but here is something really wrong. So, there is clear cheating problem at all levels.
Losing is never a waste of time. When you analize the game without hate, just simojy looking at why your lost happened, you learn. I play against stockfish often to test lines or find new ideas. Sometimes I learn much more than when I play people that don't cheat! Cheating is a problem because we all want to win, that kind of gives us a sense of intellectual superiority or approval. When you lose you first try to find guilt in others. Does anyone talk about cheating when solving chess problems??
I have the same feeling about the chance of running into a cheater. Alas this made me stop playing rapid. The sad thing is, that I am unfortunately too slow for blitz. Therefore I play against the bots and do tactics. Very unsatisfactory!
It's probably more the fact that they'd rather focus on open accounts that might be currently cheating as opposed to closed accounts that won't be cheating currently.
You have a race of AI vs humans. There are, or soon will be, AI that can beat you whilst playing in a human kind of way, avoiding tell-tale patterns such as suspicious winning streaks. There is no point getting good at chess, you'll just be beat by deceptive AIs trained to be human-like, but a bit better. And when players use such an AI only at critical moments, then, there's no defence. Find a new hobby. You cannot win at online games any more. AI's are just too easy to use to cheat.
The amount of elo I’ve lost to tilt started from cheaters is actually crazy
for sure it makes it impossible to have a natural streak of games. even in sessions where you versed no cheaters, you're so worked up about the possibility of facing a cheater that you self-sabotage. not something we should have to deal with when we play Chess.
I feel you, i feel that pain
I think the opening you play may increase the percentage of cheaters you face as well. I play the Scandi and I think at least 1 in 3 cheat on LiChess
Preach brother.
Losing a game to a cheater, especially when you can tell it's a cheater, ruins multiple subsequent games, mentally.
It's even worse when you see the obvious cheater not get banned.
thats a shame. dont let them impact you like that
This rook sac in 3 seconds is hilarious xD this made it 100% obvious. the account is still not banned . i wonder when this happens. great video!
hahah yeah its pretty insane. i think only a Chess.com fairplay analyst could answer the questions we have, and they never would give insights to how their anticheat works to prevent workarounds. it does make you wonder how much people are getting away with. glad you enjoyed though bro
Rxf5 is slightly beliveable, becouse the premis is to open the king and then go with Rg1, normal attacking moves
@@Wielkie-Ksiestwo-Slupskie in 3 seconds?
Yeah, I spend much more than 3 seconds just on tilting if i mouse slip...
I had a paid account and would always get refunded elo for cheaters.
I started a new account (unpaid) and haven't been refunded elo in 898 games..
I also often have win streaks of 7 to 8 games, I just don't have accuracy over 90% for all the games. My elo atm is about 1550 but my peak is 1710 about two months ago.
i'm not 100% sure on this but it would make sense if paid accounts on Chess.com had priority when it comes to in-game reports which might explain the anomaly. just a guess though & you'd need someone that works for Chess.com to know for sure but i'm fairly certain this exists for other games reporting systems.
there are also many exceptions to the winstreak rule. for example, someone who has 2400 Blitz but is rated only 2100 Rapid will be far more likely to have large winstreaks as they're playing at a level lower than they should be. i wouldn't be surprised if you fall into that bracket also.
i have 4324 games on my account and my elo was refunded 5 times, all of them in 2021
@@ResonantFrequency-sn9ud winstreaks on their own have exceptions. for example if someone is 2400 Blitz, but playing at 2100 Rapid they're more likely to have higher winstreaks than someone else their level. this is simply because they're playing below their skill level.
you probably fit into that bracket considering your rating swings, if you go down 200 Elo, you'll have big winstreaks on your way back up to where you belong.
My rapid rating is about 1650 now. Peak was 1730. I've played 3400 games in 2,5 years, and had 5 (!!!) rating compensations.
While I don't play online and only play in over-the-board tournaments (because I see 3-dimensional positions better than flat 2-D ones onscreen), I'm not too concerned about cheaters in these casual games where no money is involved. Cheaters in online tournaments with prize money, however, is entirely different.
100% agree that more priority should be given to prized events - but its the covering up of a real issue that concerns me. Chess.com insisting time and time again on the latest State of Play's that there isn't a big cheating problem in regular play and that they're doing "everything they can" to do their part.
at the end of the day people pay good money for memberships, and invest hours of playtime on the platform and half of them don't even know a sizeable amount of their opponents are using engine assistance. if it were more publicly known i'm sure people would just switch platforms, however i think its been swept under the rug for far too long.
No man, some of us want to do a 20, 30 or an hour of competitive and fun time per day playing chess at home, and because it's not a pay tournament it does not mean that it's ok or it's not a big problem, it's the time of people that want to learn and improve. How many people maybe stop playing chess after a brutal loss against a cheater but they never knew it?
Chess players don't understand what 3D positions mean. When you play with real chess pieces and board, it's still 2D. Chess with 3D positions has multiple layers and pieces can move between the layers.
In Star Trek, they play 3D chess. Although the rules are not explained.
OTB is still 2 dimensional, its not like u can lower ur pawn into the ground
@@lucianorodriguez7726 A serious player is not going to stop playing because of a single or even a series of crushing losses. It happens period. You learn from the experience. Yes it is not pleasant but you still learn from the experience.
I got accused of cheating yesterday. Took it as a compliment.😂
MEE TOO
I was accusing someone than i check our precision and it was 40 for both than i was quiet :D
Kramnik was right all along
a broken clock is right twice a day 😭😭
I wouldn’t go that far 😂
@@jacksarkisian we can all joke about Kramnik, But he knows this games better than all of us combined.
No, he doesn’t. He solely relies on statistics to form his conclusions. This video provides clearer explanations and more reasonable arguments than anything Kramnik has posted over the two years
@@yefer-gz5vu You don't know what you are talking about. Did you even see where Kramnik reviews games of these cheaters? Probably you just watch Nakamaru and that's where your opinion coming from.
It's an almost unsolvable problem. Players can just cheat and use all kinds of smart tricks, ie think long when the software tells them it's a !! move, to hide it. There is just no way to separate a smart cheater from an inconsistent player.
Just accept it and play more over the board.
Or really go rigorous methods that you require videostreaming while playing, that's the only somewhat feasible way to make sure your opponent isn't cheating.
playing OTB is not a solution for this. if i want to play a FIDE rated classical tournament, there's essentially 2 per year in my city. if i travel interstate to play it'll set me back over $1000 in travel expenses, accommodation, entry fees & food. even if i only play nationally rated Chess games which are held more frequently, i'm limiting myself to a pool of around 50 players that i can ever play against (that are around my level anyway). also to play in any kind of OTB Chess event i'd need to take a 45 minute train ride into the center of the city, and the same 45 minute train ride back. even still these tournaments would still be infrequent, usually only on weekends and sometimes weeks go by without one being organized.
its very easy to say "just go play OTB" if you're European or American, but in Australia at least its completely different.
We need better ways of explaining to cheaters why it's wrong
I actually think as the neural nets continue to learn from human play they'll be able to tell the difference. Just like engines have a unique way they play, so do people.
Like other online games you can't have open servers, you have to have your own vetted private server. Games like GTA are totally a waste when played on the open online servers.
Confirmed after you hit 2300 elo you go insane.
That's actually why I don't hit 2300 rating. Yes, yes, the ONLY reason. ( not that I can't, of course. Hum hum... of course. )
@@howard5992 of course people cheat. the insane part is spending 20 hours making graphs of people that "didn't cheat against me but cheated against someone else so that is cheating and I should get elo refund for all of my games"
i am 2900 but currently im not going insane
@@Ragehunteroh shoot, the best bullet player right? FM ragehunter?
I player a cheater otb in a national tournament who was cheating. There was no checks or anything being done so he would just leave on my turn go the bathroom and come back but i still managed to draw the game because he got to 5 minutes while i still had over an hour left, I sacrificed my knight for his promoting pawn which complicated the game and i managed to draw him.
bro played the guy with a phone in his sock 😭😭
I once started playing a guy and I suddenly got diarrhea. It was a seriously painful experience! And I'm sure he got suspicious until I started making mistakes and lost the game. I was constantly going to the bathroom. I was much younger and much more embarrassed back then. It happens.
There was an account I played around 1700 elo rapid that was about a month old. He had just got to my elo from 1000 elo and had lost a few games at that low rating, but handled me easily which I found suspicious. In the next few weeks the account was 2200 elo. Almost certainly a cheater and I was never refunded elo
OR he was not a cheater and just far far stronger than you.
You have a strong accent :) "heitch four"
Otherwise cool video. Although I have little experience in rapid.
hahahah apparently i've just been pronouncing H wrong my whole life, its not even an Australian thing 😭😭
His accent is really fine, and he expresses himself in a very clear and articulate way.
@@jacksarkisianit may be correct in certain countries, but i think it is incorrect in Australia lol, greatings from the gold coast 🇦🇺
Doesn't sound strong to me
@@reversetops im pretty sure he's the only person in the entire world saying it this way.... if you can say the word eight you can say H properly.
True the worst feeling for me is to know and almost know my opponent is cheating and report him and he still isn't banned then i have to convince myself maybe he wasn't cheating though i know he was 😢
They aren't cheating you are just bad
Jack for your own rapid improvement and my gross curiosity I would love if you were to play rapid on lichess. I personally feel like I have faced way less cheaters in the lichess rapid pool vs chesscom pool. And once you settle in your fair lichess rating range I would love if you were then to play 100 games and do a similar analysis as done here. ------ I'll add some notes here: Way more higher rated and strong players play rapid regularly on lichess. Notably GrandmasterGauri on the streamer side and plenty of other strong FMs, IMs, and even GMs regularly play rapid on lichess. As you said it'd often take you at least 5 minutes to find a game in the cc rapid pool but on lichess when you are rated accurately I'd be willing to bet you'd always get paired within 30 seconds. I do believe chesscom's cheat detection and especially refund system is better than lichess. However, by design lichess entices a higher percentage of strong players (fair players) vs it's populous size.
yeah i'd definitely be interested in doing this, might try and play some lichess Rapid on stream soon and have some fun with it (bonus points for lichess not saying JACKSARK IS CURRENTLY STREAMING HERES A LINK WHERE YOU CAN SEE ALL OF HIS MOVES AND PLAN AHEAD BASED ON WHAT HES THINKING
@@jacksarkisian lmao, nice looking forward to it. Btw your analysis of these games really has shown how much stronger you've become. Congrats on the well deserved improvement ❤
this is not that easy because we as a chesscom partner are not allowed to do content on other chess platforms usually
@@Willeinhelm Yeah I wasn't expecting him to play on stream. As Jack practically only plays rated rapid off stream I'd assume it'd be fine for him to play lichess rapid off stream as well.
@@jacksarkisian actually you can link your profiles and Lichess will say exactly that
Amazing video, much of it was things ive thought but never took time to look over (closet cheaters, account time, etc) thanks for actually bringing these things to light
A very intersting video. You were not biased. Thank you for this interesting insight!
thankyou bro. i would love to see a GM like Danya play ~100 Rapid games at this level and do their own analysis. i can't say with 100% certainty but i'm very convinced i versed more than 29 cheaters in my last 100 games its just i'm not strong enough to know for sure if they were or not. glad you enjoyed bro
@@jacksarkisian yea i also think so. im also curious if a gm like danya does it. he sees alot of cheaters in his speederuns already so i think theres an intersting fact
@@jacksarkisian In one of his speedrun videos he talks about this and says the amount of cheaters is the reason he does not want to continue the rating climb beyond a certain level.
@@Shadow-7295 which video is this in? i heard someone mention this before but i couldn't find the video. wanted to mention it in this video but couldn't find the clip haha
Vladimir Donaldson
The video cuts make me sick
Do you believe that somes players get banned for cheating even though they are not ?
the only case this happens is when a Chess master is playing on a secret account. they play unbelievable moves and because they are quite literally comparable to a low depth engine they can be banned for it. aside from this it doesn’t happen ever with the current anti cheat.
For those saying "just play otb brroo", keep in mind that security is non-existent for all lower level (think city, district etc) OTB events, FIDE or not. Of course its still not as easy to cheat as online but I expect with even more miniaturisation of electronics in the coming years that soon cheating will become a menace even at OTB events UNLESS all of them religiously put in place metal detectors and body scanning. I don't small events able to afford the expense or manpower. This is just an additional point on top of what OP already said in reply to another comment regarding travel expenses and frequency of such events being poor or non-existent in most parts of the world. And OP is 1st worlder. Imagine a Mongolian or Congolese young talent... either having to face endless cheaters or travel 100s of miles with his non-existent cash for a single OTB event...
100%. even in Australia we have ~2 FIDE rated classical events per year in my city. to travel to a different city & play it'll cost over $1000 to play (travel expenses, entry fee, accommodation, etc). you could play nationally rated tournaments, but those aren't even held fortnightly meaning that people who recommend you "just play OTB" are essentially recommending you play one Chess tournament per month.
obviously i'm still privileged to be able to play OTB as often as i can, but even in a country like Australia its not an optimal solution.
god knows the amount of elo's i lost to this cheaters
it's time to join your local chess club
Great video. I couldn't tell you the amount of times I have a winning position (2.0-2.5 advantage) going into the middle or end game only to have the opponent play out of it and beat me. Happens way too often.
In my video, I showed an account that was many years old and had thousands of games. It played bullet, and every move was like 0.7 seconds. Even that was too hard for them to detect. Oh, and the same for the guy playing crazyhouse with every move made in 0.2 seconds.
actually saw your video the other day & made me get around to making this. super disappointing that there’s been virtually zero progress for the anti cheat in over a year now, great work on your video though 🔥
@jacksarkisian Oh wow, I'm glad to see an update, in that case! I was also wondering what the cheating rate would be if I played a lot of games now, and you did just that!
@@AyrtonTwigg hahaha yeah would love to see you make a follow up video in the bullet time control to see if theres been any change there. sadly i'm incredibly slow and only ~2050 bullet so i don't verse many cheaters but would love to see a comparison.
i just got an email from some Chess.com representatives about this video so hopefully will be able to provide some more clarity to the situation over the next few days.
im a 2100-2200 rapid player myself and I can only agree with you.
When I play on one of my numerous secound accounts (rated anywhere between 1400-2000) I rarely, almost never) face a cheater.
But once you hit the 2100 rapid barrier around 10% of the accounts I face look sus as hell and I abort alot.
I imagine that at 2300 the issue is even worse :(
sad stuff and that's also the reason why I stopped playing rapid chess as much (gone from 40 games a month to 15).
Should have a built it tilt function. Where it won't let you play if it feels you're playing too fast and with little accuracy
this wouldnt work out well. You should just be responsible for your own playing. When I lose 2 games in a row, I take some time off and that helps
I believe paid members require a human review, which delays the process, since CC is understaffed
this might be a thing i'm not too sure. i know titled players & Chess.com partners/staff are reviewed by the anticheat and by a human but not sure about paid accounts - could have something to it.
Cheating is rampant in every online video game right now. There are AI anti cheats being developed and in a year or 2 cheating in video games will be almost impossible.
it isn't but all the low elo players think it is because they are bad + gapped and can;t handle it
That may be possible in video games, but cheating in chess is considerably harder to detect for numerous reasons.
Yeah but there's also going to be AI cheats to get around the AI anti cheats. It's a cycle that can't be fixed with more tech.
Top 3 signs of a closet cheater.
#1: They make rookie opening mistakes despite having a high ELO
#2: They suddenly play perfect chess after a blunder
#3: They play many top moves without thinking
Not really. A closet cheater is already good. They cheat in difficult positions. All the things you listed are signs of obvious cheating.
The only thing you know someone cheating is time usage. They plays best moves with no times usages mean cheating
This can all be explained by just playing chess though; try something because you played the same opening 30 times in a row, after a blunder they get a different mindset and try harder, just trying things and by luck doing some brilliant moves. If it's constantly in every game, it might be cheating, or it's a genuine genius stuck at 900 elo (😂😂😂)
I think the 30 day limit might be based off the assumption your lost rating will revert just through play since if you are underrated your rating will self correct.
yeah this would make sense but i think due to the amount of cheaters in the Rapid pool its less effective. paired with the fact that Rapid is a longer time control and you're far less likely to gain it all back because the games are much longer. would still love to see a more dynamic solution to this.
maybe if they found a game where the player cheated, games played 15 days before and after that game would be refunded? if this was applied for every game they cheated on this would seem optimal especially if there were numerous instances of cheating over long periods.
Fr, most of the winners of your weekly prize arenas keep getting banned for fair play lol.
yeah super sad. the worst part is even if they're blatantly cheating Chess.com are still yet to implement a way to remove someone from your OWN arena. even if i own the club and started the arena myself i still can't remove people from a live arena. this is a feature that was requested over 2 years ago and still hasn't been added.
Why don’t you make the tc something faster for the arenas like bullet, it discourages cheaters (using bots are much easier to spot). Even 2+0 is better than what it currently is, the players winning these arenas are usually quite high rated so bullet will still get a decent quality of chess.
@@gonnabegm 3+0 is what jack's arenas use. Because he wants it to not be too long and not too quick. And also 3+0 is what most people play anyway, why complicate it with other time controls most people aren't used to? And also before giving out the prizes it needs to pass the fair play check first.
@@gonnabegm fun fact people can cheat in bullet as well
@@gonnabegm because it massively excludes more amateur players. i actually took off the minimum amount of games required to play for the tournament because it was affecting more legitimate players than it was preventing cheaters.
if the time control is too fast, you'll just end up with an incredibly tryhard tournament in whats supposed to be an event that allows newer players to grab a free membership. sadly i think the way we do it now is optimal but if Chess.com ups their game things can improve massively.
23:04 That's crazy. I nearly 2000 and it only takes me a few seconds. Crazy that 300 rating points makes that big a difference.
it does
WOW! VERY INFORMATIVE VIDEO! BEAUTIFUL CHARTS, UNDER APPRECIATED IMO, DESERVED MORE SCREENTIME!
THANKYOU QUEEN SORRY ABOUT THE LACK OF CHART SCREENTIME WILL DO BETTER NEXT TIME
Paid members might not get extra anticheating protection, yet, but its undoubted that streamers do get extra protection.
Which makes sense to me.
i think this alongside with titled players make sense. of course i don't think they should be given more leniency, but in any case in which a streamer or titled player has suspicious play on their account they should be carefully reviewed by a human. even though i think Chess.com has a problem with letting too many cheaters off the hook, the opposite in banning players on very minimal evidence is also not the solution either.
thats why i left chess and starting playing checkers. best decision ever👍
Edit:Kramnik pinned your video on twitter lol
bro MIGHT be Matteo Bernini
@@jacksarkisian thanks for the membership man ^^
@@EmFanForLife no worries bro get solving those puzzles 💪
@@jacksarkisian surely man. ✌️
Have to agree that after 2300 you will face loads of cheaters when playing rapid and that really affects your game. In every single game you wonder if the opponent is cheating or not. This affects your play and it is very hard to play the way you normally would.
To put is shortly I very rarely play rapid games nowadays as it is almost 50/50 if you face a cheater.
Ps. I would also like to add that I have 2500 blitz rating and 2600 bullet rating (records) but in almost every time it is me who gets low on time even if my opponent has blitz rating of 2100 etc. It is a really strange feeling.
Stupid to refund old games, as ratings converge to their appropriate level in the next games anyway.
Agreed. And probably after about 20-30 games you are back to where you belong. No reason to credit people rating once they've surpassed this.
Your claim that skilled players don’t blunder their queens is factually incorrect. I’ve seen blundered queens in Titled Tuesday games multiple times. Let’s be honest, your metrics are subjective, you have no data, you have no skills to process even the data that you have, and you’re saying that your excel exercise is worth more than a team of analysts and engineers who work around the clock. This cheating accusation madness should stop.
there wasn't enough context in my statement. i meant 2200 players don't ever blunder there queen LIKE this. by like, i mean straight out of the opening with literally 10 minutes to think about their moves. of course there are times where someone will be on some kind of ungodly tilt, or has come home drunk playing Chess but this is why i said the sample size was too small & they were clearly banned for other games. however this is still an important piece of evidence to put towards their account being illegitimate.
queen blunders happen all the time, theres definitely games in my last 100 where i've fully blundered my queen but its much more reasonable to happen at latter stages in the game where you're running out of time - the same can be said about titled tuesday players as the time control is much shorter.
and Chess.com do not have a massive team of analysts. according to their own website they have 2 and i don't think its unreasonable to think at some stage they've missed the mark with their work.
A kramnik minion
I am 2430 blitz and i have blundered my queen like thatx@@jacksarkisian
@@LyndsieDances said Niemann minion...lol
@@jacksarkisian Even tho I'm only 1950 rapid, I have blundered my queen, bc of my friends or when I was generally distracted. I understand, bc even at this low elo of mine there are still many cheaters. After the 99 precentile people become too competitive lol.
I don't cheat but definitely think i could get away with it. If you're already decent and only check the engine when in key positions how would the fair play team catch you? I can't think of a way. The signal to noise ratio just isn't robust enough so it seems you can only really reliably catch obvious cheaters. Of course I don't see the point in cheating. Is winning a game where you had to check the engine satisfying in any way? I guess there's the shallow thrill of getting away with something but I dunno how thrilling is it really when there are literally no stakes? Maybe a former or current cheater can explain their motives?
Soemone might be 2200+ and blunder their queen just by being a mixed of tired (playing several games in a row) and drunk/high or just simply losing concentration due to other people in his room talking to him etc. It's not impossible imo
definitely not impossible but highly unlikely and as i said in the video, it cannot be your sole piece of evidence to prove they were cheating (unless you have some highly advanced and reliable AI generative learning anticheat that says otherwise). but clearly even though that sample size was too small, they were cheating in other games and were banned for it.
the instance in this game simply backs up the fact that they were cheating and is useful in hindsight.
@@jacksarkisian Yeah I did realise your point. I was just picking on semantics I suppose where you had said "it's impossible to happen at this rank" or something. Not a big deal, I definitely didn't mean it as a criticism. You made a good point about "closet cheaters" that I hadn't really considered so much.
At least twice my finger slipped on the screen and I moved a piece to the next square of the one I intended.
@@sunway1374 true that. Happened to me lots that
Online I'm 2500. I've got a few thousand games and have never blundered my queen to a 1 mover. Had my queen trapped a few times but that's a different story
Very well put together video. However there is one critical flaw in your analysis in the result section at around 15:00. The sample size that you are taking is very small to follow a normal distribution hence inflating the statistics. If you are interested I can help you develop an extensive and robust Machine learning model which could run through your account and find those accounts which are likely to cheat.
hey! this was a thought i had, however i think the following issue would come up.
i did consider increasing the sample size to something much larger, but if i go back lets say 250 games into my Chess.com history my Elo would be around ~2000. not only is that 300 Elo gap absolutely huge in comparison to lets say 1400-1700 Elo, but i would also assume the higher rated you are, the more cheaters you will verse.
i feel like if i use games from my personal account this factor would deem the data much less valuable as the Elo level i played at was not a constant. i would also assume finding an account with a suitable dataset would be quite rare. not only are there only around 3000 active players in Rapid rated around 2300 Elo on Chess.com, but you would need to find someone who is completely stagnant in their Chess and has played 250+ games in the last 2 months (2 months is also a generous timeframe for various reasons (cheating program availability, real-world events, Chess.com site changes, etc)).
i suppose you could take the same 100 games over a 2 month timeframe from other players similarly rated on the website, but i still don't know if there are enough people at this level to acquire a large enough sample size.
i wouldn't imagine my findings are hugely reliable, but i think it at least outlines that there is a problem somewhere which is the main point - would love to see a more extensive dive into something like this though!
My personal favorite cheater story of mine is when I played someone who had an antisemitic name, was racist in the chat and was clearly cheating. I reported him for cheating and he wasn't banned until roughly 2 months after. And I didn't get refunded, presumably as he was banned for the antisemitic name and racist comments.
someone mentioned that paid members require a human review from Chess.com to be banned, and considering they have 2 fairplay analysts you could imagine the delays considering literally anyone with Discord Nitro right now has a membership.
@@jacksarkisian Ah, nice to know.
I'm not positive the "closet cheater" was cheating. Kh1 could have been a nearsighted move to get the g rook active on that file. If it was a mouse slip, then I still don't see how they are moving the mouse and getting Stockfish best moves all within 1 second or so, sometimes less. By saying mouseslip you're implying they are manually moving the pieces and not using some advanced automove cheat, but they are still moving as fast as under 1 second? Doesn't seem possible... The move times do not feel suspicious. I'm 2000 bullet and could intuit the rook sac given the activity of the other pieces. It also just feels like a desperation move that luckily happens to maintain a draw.
people use browser extensions to cheat on Chess.com. it shows an overlay of your Chess board and draws an arrow from the piece to where it should be moved to.
afaik automove isn’t really a thing anymore unless you set the times to random. if you’re automoving in 0.1s every move you’re obviously going to be caught, most people do it semi manually now.
this kind of erratic decision making might be common in blitz or bullet but it’s nonexistent at high level rapid. no one plays risky moves like that in such short amounts of time, you at least think for 15-20 seconds otherwise you’ll end up throwing more games than not.
There's no point bulletting whole rook sac when you have 7 minutes on the clock!
The moves seem inhumanly quick. JMHO
😂 nah he was obviously cheating, im telling you im 2000 blitz fide he was 100% clearly cheating
This is only valid if the time on opponent clock was very low, which it was not.
Very well documented, thanks for bringing awareness to this issues
thankyou bro but i do still believe it'll be in vain until someone with a larger following covers the topic. i've built a great community of people over the years and i think my voice is heard a lot more than others but the issue is simply too big for someone like me.
How do you know some cheaters play badly in the opening to lower their overall accuracy and play perfectly in the other phases 2:18 ? The example you showed is just some dude blundering his Queen and resigning, where is the perfect play afterwards 2:35 ? Also, people can make all kinds of mistakes even blunder the Queen. On what do you base your suspicion of having win streaks 3:33 ? So you personally have a 6 game win streak and thus a person with 7 or 10 is suspicious? Why? On what do you base your description of "closet cheating" How do you know they typically cheat to 1-200 elo higher rated than they are 4:03 ? How do you know that they usually cheat in the game after they lost the previous one? So your entire argument that this guy is cheating is that he knew how to refute your opening, he didn't fall for your mate in 1 trap, he played Kh1 and not Kh2 8:50 ? And after realising the mistake went all in? And based on this they are "undoubtedly obvious cheating" 11:05 ? How do you know that chesscom in recent years made a public effort in recent years to reduce the amount of false bans they give to players 11:10 ? 11:20 what do you mean they are given full immunity and will never be banned 11:24 ? "The opponent clearly meant to play the best engine move, mouse slipped and then suddenly found the best way to force a draw 11:54 ". Such an interesting wording. Also, how do you know that he wanted to force a draw? How do you know the rook sack is not an attacking move, and you just happened to escape the attack? How do you know that are people that are cheating under our noses that will never be banned 11:23 ? How do you know that they are cheating in a way that is undetectable by chesscom 12:28? So chesscom banning someone is not the seal of approval but your cheating analysis is? 12:37? Every game should have looked at by a the anti-cheat, did you report every player you played for the last 100 games 14:48 ? How is it obvious that they were cheating against other people 16:05? So if you claim that they are all cheaters how are you beating 50% of them? If they were truly cheating you wouldn't score 50/50 with them like with average people right?
Found Danny Rensch's alt account
Bro what is this an academic peer review
How do you know some cheaters play badly in the opening to lower their overall accuracy and play perfectly in the other phases 2:18 ?
Because it's a common trend when versing cheaters. I play viewers on my livestreams all the time and you’ll constantly see players stumble out of the opening with low accuracies, then absolutely dominating the middlegame and endgames (this is also myself playing against players with Elo’s as low as 800). Those accounts are then subsequently banned and it becomes a pattern that users that play legitimately do not follow - only cheaters.
The example you showed is just some dude blundering his Queen and resigning, where is the perfect play afterwards 2:35 ?
The two statements were not connected for this example, if you blunder your entire queen in the opening with no compensation even an engine won’t save you against a 2300 rated player - hence the resignation. It was also against a proven cheater, there is no speculation they were cheating as their account was banned.
Also, people can make all kinds of mistakes even blunder the Queen. On what do you base your suspicion of having win streaks 3:33 ?
2300s do make mistakes, but blundering the queen straight out of the opening is typically not one of them. 2300s can blunder their queen in complicated positions where they have lower times, but straight out of the opening is almost impossible. As I mentioned in the video though, this is NOT enough to determine whether or not someone is cheating as the sample size is too small, but if you quickly analyse their account you’ll see they were blatantly cheating against other users.
So you personally have a 6 game win streak and thus a person with 7 or 10 is suspicious? Why? On what do you base your description of "closet cheating" How do you know they typically cheat to 1-200 elo higher rated than they are 4:03 ?
My 6 game win streak should be considered suspicious. It's an extreme outlier in data and should be considered when determining whether or not someone was cheating. The opponent in question also had a staggering 14 game unbeaten streak in the pool where they scored 90%+ accuracy in every game they played. Additionally, closet cheating is typically when players, in response to losing a game decide they’re going to cheat to win the Elo back instead of fairly fighting for it. They mostly use their own moves, but if they think the position is getting away from them or the moment is critical they’ll use an engine. They’re typically higher rated than they should be because they lose Elo fairly without needing to gain it back fairly, making them inherently weaker than the people they verse.
How do you know that they usually cheat in the game after they lost the previous one? So your entire argument that this guy is cheating is that he knew how to refute your opening, he didn't fall for your mate in 1 trap, he played Kh1 and not Kh2 8:50 ?
Because that's the entire premise of a closet cheater, they won’t cheat unless they feel like they have to. If they’re legitimately winning games on their own they have no reason to cheat, but if they lose a few games and can’t win on their own - they use an engine to help them. I also explicitly said a player knowing deep Sicilian theory was not a sign of cheating, I also never insinuated him not blundering M1 was evidence of him cheating either. And as I stated in the video, my evidence of him cheating was him mouseslipping the top engine move, and then immediately finding an entirely new idea that was perfectly functional without even calculating it.
And after realising the mistake went all in? And based on this they are "undoubtedly obvious cheating" 11:05 ?
First of all he had no time to even process what he did might’ve been a mistake because he was playing so quickly. Second of all if any legitimate player thinks they made a mistake, they don’t dig a deeper hole by sacrificing their rook without calculating it - that's complete nonsense.
How do you know that chesscom in recent years made a public effort in recent years to reduce the amount of false bans they give to players 11:10 ?
The CEO of Chess.com IM Danny Rensch has stated it numerous times in recent Chess.com State of Plays.
11:20 what do you mean they are given full immunity and will never be banned 11:24 ?
Due to the functionality of the anticheat on Chess.com, it is unable to detect anyone cheating in this fashion and therefore they cannot be banned.
"The opponent clearly meant to play the best engine move, mouse slipped and then suddenly found the best way to force a draw 11:54 ". Such an interesting wording. Also, how do you know that he wanted to force a draw? How do you know the rook sack is not an attacking move, and you just happened to escape the attack?
Because there is no way for either side to win in the line he chose to play. He forced me down a line where I would be down a rook if I didn’t repeat moves in the position. Even if I just “happened” to escape the attack, that's why you take more than 3 seconds to calculate a move in Chess - to make sure it actually works as you intended.
How do you know that are people that are cheating under our noses that will never be banned 11:23 ?
This is an assumption that a cheating software paired with common sense can be passed as legitimate play in online Chess.
How do you know that they are cheating in a way that is undetectable by chesscom 12:28? So chesscom banning someone is not the seal of approval but your cheating analysis is? 12:37?
Because Chess.com doesn’t ban them for what is obvious cheating. Also Chess.com is not the seal of approval because as I stated before, Chess.com largely prioritises preventing false bans which also allows others to slip through the cracks.
Every game should have looked at by a the anti-cheat, did you report every player you played for the last 100 games 14:48 ?
I mentioned this at 14:05.
How is it obvious that they were cheating against other people 16:05?
This was a general statement at players cheating across the board. I bothered to analyse their games against other players and found evidence of cheating in those.
So if you claim that they are all cheaters how are you beating 50% of them? If they were truly cheating you wouldn't score 50/50 with them like with average people right?
I mentioned in the video the reasons you might beat a cheater at 16:23.
@@jacksarkisian
1). You didn't show an example for that, and how playing poorly in the opening and dominating in the other phases affects the overall accuracy, and how do you know that's the goal of a cheater and if it actually works.
2). Just before you showed the example with the Queen you reminded us, that it is connected to that 2:36. And yes he is a proven cheater, but you are making it seem that the fact that he blundered his Queen is somehow connected to him being a cheater.
3). Here are the two recent Rapid games where you as a 2300 blundered your Queen with plenty time on the clock game/live/114925732439 game/live/116230227659.
4). On what basis do you believe your 6 win streak is an outlier? So in this sense how staggering 14 win streak is actually? Is this guy suspicious "MO2002ISMAIL" ? I understand now, so you say that their elo is inflated because they cheat in the next game after they lost and gain that unfair elo premium. So what do they do after they lost 2 games in a row, they cheat 2 games in a row in response? And also if they are inherently weaker doesn't it mean that won't be able to compete on that level and would need to use the engine more and more until they are solely using the engine?
5). So what did you say here, maybe I misunderstood 8:27. I don't belive that basing your accusation on move Kh1 is enough of the evidence. P.S I think your argument would be way stronger if you had mentioned the time the player in question used during all of his moves and not just during the rook sack. Because he was playing most of the moves with 0.2, 0.5, 1.0 time frame.
6). I would consider the rook sack to be the tactic with obvious threat of Rg1 Rg7 mate. Most of the Blitz players are really good at tactics and attacking, and blitz players don't really tend to calculate it all through, it's more about intuition (sensing it's dangerous enough). You think of it as he calculated 5 moves ahead that it was a draw and went for it all in 3 seconds, but it most probably is him trying to continue to attack you which has happened to be a draw. In a similar situation, his attack could have failed and you won, and you wouldn't even slightly suspect that game.
7). I don't see the reference.
8). How do you know that due to the functionality of the anticheat on Chesscom, it is unable to detect anyone cheating in this fashion and therefore they cannot be banned?.
9). Yes but you blow that assumption into a basis of your argument as if it's a factually proven situation. It's an idea that this can be possible, but for now, it's only a speculation of the community. Because all those who try to "test" it get banned. Of course, cheaters won't be banned for using an engine for 1 move, but how influential that move would be to the outcome of the game? When to know at which point to use the engine?
10). It is not clear what you claim to be obvious cheating that chesscom is refusing to pay attention to. I am quite certain that with the right game and with the right sample of data, the same argument can be made about your account, You already mentioned that your 6 game win streak is suspicious, so all that needs to be done is to find the game that would in some way illustrate that. Your account is way younger than mine "check", Your account has suspicious win streaks "check", You blundered your Queen in rapid two times recently "check", Find a suspicious game "most probably check". And now what is left is to report you to Chesscom. Oops they didn't ban you. Well, you must be one of those undetectable cheaters well.
11). Sorry I for some reason completely missed that. Thank you.
12). Would be nice to see some examples, and on what basis you actually figured out that they were cheating.
13). I understand how you can beat 1 or 2 cheaters, or maybe even beat them regularly, but how are you beating 50% of them. Maybe they aren't cheaters after all? If we assume that all of them are cheaters and you beat 50% of them it means that they didn't cheat in at least the 50%. Now if we take the other 50%, does it mean that all of the games you lost are to people who cheated in the game? Or we can use the same assumption that some of them cheated against other people and not you? Does it mean that more than 50% of the game let's say 60% or 70% of the games you played are against people who didn't cheat during that time?
@@RomanWegera nobody: this guy when presented with facts and logic: BuT hOw Do YoU kNoW?
When I was a kid I liked chess. Then I found out apparently some people just memorize fully what they have to play on every possible opening. Since then I realized it isn't really the game for me, as it turns from a thinking game to a pattern recognition and memorization game.
Now add to this cheating with bots... I feel maybe it is just too late for chess
in my opinion Chess just hasn't evolved with the times. world championship matches should be played in shorter time controls to prevent extreme memorization and induce mistakes. i do think that memory should be a part of the game, but it shouldn't be the main characteristic of it - creativity and problem solving should be. shortened time controls prevents cheating as a whole in OTB Chess and would also be much more exciting to watch as a spectator - i think anyone that's watched the Speed Chess Championships lately would agree with that.
@@jacksarkisian Yes, time can make a big difference for sure. Watching high level fast play is way more fun than absolute stockfish perfect slow games.
Or course in any game there is some memorization...and to say I wasn't tempted to just memorize a whole opening would be a lie, but unless you plan to become a GM I felt like it would just suck the fun out of the game. Well of course the 100th time you see the wayward queen you will have it memorized, lol.
Cheating on the other hand... I don't understand what those people have fun with, at all. They don't have fun, they don't improve, and they can't even do it professionally. Just for an ego boost.
Just play chess 960. No opening theory. Its a better game for beginners
How can most of us beat a GM if we don't cheat...... is a no brainer........ ( is a joke)
Bro has his name more than any others in the thumbnail
they’re my games against cheaters 😭😭
thank you! you helped me to quit chess! this video was very insightful
Online chess has its problems. While you do make some valid points that the number of cheating accounts has been growing, I do need to say that a player is innocent until proven guilty - that's the basic rule. Based on a single game, all you can do is to say this guy is "likely" a cheater- hell, even a random mover has a non zero probability to plays 100% perfect moves! Also, 100 games is not big enough as a sample.
i agree, 100 games is not a big enough sample size but if we increase it to 200 games, we go back to when i was ~2050 Elo and my Elo is not a constant anymore and the data becomes much less valuable. i'm also fine with you saying the best you can say he's "likely" a cheater - but in my opinion even if its not definitive, i don't want to be versing "likely" cheaters 33% of my games.
A pleasure to watch this video. Its goal is Fairness and it was informative. Effective communication throughout. Clear, well organized, well reasoned, right degree of detail, right tone, and concise.
thankyou bro i appreciate it. planning to have a discussion with Chess.com's head fairplay analyst in the coming week & will make an update video on this
This is why I stopped playing online chess too many cheaters and it only gets worse
casually preparing "why 10-0-0 football is a major crisis in the Premier League" after that Arsenal game vs City
"bro the Arsenal down to 10 men against Man City meme is still MASSIVE dude like MASSIVE"
1:55 😟me looking at my last 9 rapid games all having 75+
hahah there are of course exceptions to this. for example if you're a player that likes to trade pieces off the board early, your games will consistently have higher accuracies because there's less possibilities. but if you like to keep pieces on the board, your accuracy will be lower on average because of the increased amount of possibilities. openings also come into play with this but generally one single suspicious trait usually doesn't indicate foul play
@@jacksarkisian i'm just messing. I enjoyed the video and the analysis of the mouseslip cheater. I definitely wouldn't have clocked that one.
@@ChaeDoc haha no worries man glad you enjoyed!
Is LiChess any better with regards to cheat detection or a culture of cheating on a whole from it's player base?
from what i've heard, Lichess' anticheat is similarly comparable to Chess.com's. their situation is slightly different though, they have a much smaller player base with a much older age demographic. its hard to say for sure but you could imagine that 25+ year olds are much less likely to cheat than children just on a maturity basis.
a lot of people who do cheat use browser extensions to do so as well. from what i've seen, most browser extensions are compatible with Chess.com and not as commonly Lichess. meaning if someone does want to cheat they're more likely to do it on Chess.com.
i think there are just differing variables that make Lichess a platform with less cheaters, i'm not 100% familiar with their policies on things and how they approach cheating but you would imagine they take a different approach to it considering its a non-profit charity as opposed to Chess.com being a profitable business.
realistically if anyone is going to be the forefront in innovating engine assistance detection, it should be Chess.com.
I play in Lichess since I started playing chess and I didnt play against a cheater in 580 games. But if you dont with a account you will find cheaters, at least they cheat without account so you dont lose elo
No it's worse but it's better. To make sense of that sentence; because you can expect cheating on lichess practically 80% of the time you end up playing more solidly trying to prevent mistakes rather than creating an attack that is not realistically correct.
@@equinøx984 Mate they just don't review their games I ran a statistical analysis from numbers provided by them publicly and they not only made no sense but would even in effect be void in counter-acting the problem as there are far too many games to be checked.I know this 10000000%, a user cheated for months without being detected. They, in honest, conversation explained everything to me as they themselves couldn't believe how on earth it was allowed. They reached far above 2000 without ever being detected by combining their own strength of play with an engine and from what I was told they would only ever use moves given by the engine if by then seeing them they could understand them at the least loosely, Apparently they were able to stay in a realistic range of centipawn loss and by that get completely dismissed by the detection system. In their head they started doing because they hated being cheated against and saw how it worked but nothing was being done about it so they just said **** it. Now imagine how many of them are out there.
Lichess errs on the side of not banning anyone unless they are extremely sure of cheating. So you will see less people banned for cheating, which probably results in less tilt.
Lichess isn't much better
Can you say the letter H a few more times for me please? It's literally my favorite thing to listen to Aussies say the letter 'H'
😭😭😭 at this point i may as well never say it correctly to farm engagement
Now I understand why I am losing ❤
Skill issue?
Jk sorry
the % of cheaters per 100 games is definitely different at varying ELO levels, but i'd imagine its still pretty abysmal regardless. i've been playing Blitz a lot lately instead & seems like i'm not at an ELO where cheating is a massive problem. 👍
I don't understand why they don't make a 3D-secure login function. Ban this and you will have to use someone else's Personal Identification Number- which one would run out of fast.
With absolute certainty, make a "cheater pool" cheater vs cheater only, unable to player normal pool.
i know Chess.com has been looking at this in the past but as i mentioned in the video, this would need to be paired with a massive promotion of Rapid Chess (possibly getting Grandmasters on board to play again). if not, you'll just have a tiny pool of players where everyones playing the same people and waiting too long for games.
might be the only good idea, i could be wrong but i think this was CounterStrike's solution to the problem with Premier.
@@jacksarkisian Force everyone to use 3D secure login. The player pool would stay the same. I would even be okay with playing the same person every 5 games to keep cheaters at a all time low.
People are literally using Python scripts to automate creating new email accounts to burn.
Weird colour coding: red = honest player, and green = cheater? 😂
hahahah yeah i wasn't really sure which one made more sense but in hindsight it probably should've been the other way around
Did you get the 'phone call' from Danny yet?
hahah i got an email from Chess.com's head fairplay analyst & will hopefully have a meeting with them over the next week or so. will probably make an update once that's happened to clear anything up.
Why do some players play a few moves, then resign or lose by timeout? I am not talking about just a few games like that. They play many games and most are like that.
i've noticed this too which is weird. i would imagine its connection issues with chess.com or just internet issues in general. i think it might've happened 2-3 times over my last 100 games? regardless if this does happen & you email Chess.com's support team they'll refund you the Elo for this.
Sometimes if i am in Asia the wifi, the internet and the phone connections drop out mid game often.
I know this problem, and I am only 1300 elo. There was a stretch of time when I said in chat before every game: "Lets have a fair game! If you cheat in any way, I am the winner regardless of the result. Thank you."
This means cheaters lose against me 100% of the time.
And they proceed to cheat
@@consciousmassofatoms but if they do, they know I won regardless, big hit on their ego heh
@@consciousmassofatoms they cheat, because they want to feel like they are good at something.
They want the other to think they are amazing player.
If I write that message, then they know I wont think they are an amazing player, because I know they might be cheating, which takes the joy out of it for them.
Also, they will know they cheated, so obviously I won regardless of the result, so It won't feel like a win to them, their subconscious will know.
Chesscom should have a message like that which is a direct hit to the ego of a cheater. Self detection is the best prevention, make it unappealing.
I haven't played in a while. But the last time I played a for a little while I ended up getting my ass kicked and over the next few days got almost 50 elo points back from them banning cheaters. That's a lot. Wish they told me which games they were. Basically all the games I lost I guess. Lol
if you go into your game history and see which players have a red cross next to their name you should be able to see! shame they don't make it easy for you & tell you who it is in the message but its better than nothing haha
@@jacksarkisian That's true but that symbol doesn't show up when you're using the app playing from your phone which is how I play it. So I'm forced to check all the games I lost recently to try and find out which game was it. And a lot of times I lose patience and just give up because the game could have been months ago
@@AdrianStalker ah right thats weird i had no idea. lots of weird quirks with the mobile app i'm heaing hahahah
Didn't realise Luke Skywalker played chess
Luke Skywalker is a significantly better comparison than the usual "MrBeast staying awake for 72 hours challenge" 😭😭
i hear you bro, i just got fed up with this completely in rapid. I'm only 1300 elo so i don't think that this has to do anything with elo, rather it does with simply recently there are a lot more cheaters and more sophisticated ones than before. Almost all my games have an accuracy above 80% and this is almost not enough for anything. I was able to finally win a game after loosing 6 in a row against 1250-1300 elo guys who played above 90%. My accuracy was 93% in this game, and the only reason i was able to win was that my opponent got low on time and finally made a blunder with 10 seconds on the clock, but he played almost perfectly until then. Another guy also found some Hikaru/Alireza kind of fork tactic in 3 seconds, with apparently 1205 elo. I just can't believe these games at all.
Idk about the streaks. I had a 9 game win streak yesterday where 6 of them were 93%+ accuracy. I play for endgames so my accuracy tends to be higher. It’s statistically probable as I played 900 games in 90 days to get a win streak like this
Over the board remains supreme.
i agree i love playing classical and the reason i play Rapid is to practice for it. the only thing is FIDE rated classical tournaments are hard to come by in Australia and usually you'll have to spend around $1K to travel to a different state to play in one which isn't super viable.
sad
Also I’m about 1900 rapid, and I just went and looked at the last 10-15 games and there were at least 3 accounts opened in 2024. Only one did I suspect of cheating until I went and looked closer at the accounts. None have been banned and knowing I would only be refunded points within 30 days is so concerning. Thanks for the video, I’m definitely looking every single time and aborting any game vs an account made in 2024
Are you open to coaching?
i am currently taking students! here's the link to my website where you can book: www.jacksark.com/coaching
Maybe they can make a trust factor system like in CS where if a player is reported a lot they will have a higher chance to play vs other low trust factor player, this would also help against smurfs or cheaters just creating new accounts as the starting trust factor can be set to low and only get higher by playing many games without getting reported for cheating. This probably wouldnt affect the new players that badly as new players would start very low in rating and would need many games to get higher elo and to get a high trust factor, also most cheaters dont cheat in low elo so it probably wouldnt affect them.
The only cheating I ever do is when I play against chess engines. If the Elo is at 3300 or above, I set the game up to where it gives me queen odds, and after a challenging 80 to 90 move game, I checkmate the computer with 3 or 4 pieces left.
Sometimes when I think my opponent is cheating, and I start seeing them play weird moves, I turn on the engine myself and compare their moves to the top Stockfish moves. When I see them playing super weird looking moves that are the top Stockfish moves, I already look at them with suspicion. When I then look at their profiles and see some massive streak of 12 wins in a row or a 90 % win rate, gotta smash that report button.
I know my methods are morally questionable, but hey, I'm an anti-hero, I'd do anything if it means getting even one of those disgusting cheaters banned.
A quick rule of thumb for anyone:
- Winning more than 6 games in a row is certain cheating (5 wins in a row was the longest win streak that I've ever had, I'm currently 1600)
- Having an accuracy of over 90 more than twice in a row is certain cheating
- Taking more than 5 seconds when they only have one legal move (perhaps they're in a check) is a warning sign
- If they for some weird reason resign even though they're totally winning, it's a warning sign, they could be doing this to lower their win rate and look less suspicious
In Bullet it doesn't seem to be realistic numbers at all. The pairing code is crap or something.
I only play daily chess and since i made my account in August of this year ive only played two people who i could tell where cheating. There is something very artificial about people who cheat at my level 1000. Its like looking at an ai picture, you can just tell something is off and artificial about it. People at my level play very human moves where you can easily follow the motive. Both times i felt the game was artificial, the accounts got banned for cheating (one of the accounts I reported myself)
I was watching the 2nd game like "come on, he's not a cheate..." sac the rook! "Nevermind"
Same reason why I stopped making videos... Also I don't like the lichess interface.
Although I enjoy bullet more because of how easy it is to catch a cheaters. but you don't really learn anything from constantly playing bullet which is also sad
2:13 "and importantly the game phase accuracy wasn't suspicious wither". bro the endgame accuracy is 100% 💀
I think it's important to point out that players in the 100-1800 elo rapid pool are far less likely to face cheaters than someone in the 1800-2500 elo rapid pool. Not as many cheaters hang around at lower elos, with the exception of a few. I'm 1400 rapid, and I am refunded elo about once in every 200 rapid games. I suspect some level of cheating occurs in about 1 in 40 games or so.
If ( if ) you are right, then a simple solution would be a filter where you can request to only play accounts that are say over 3 years old or 5 years or whatever you set it at.
I thought for sure this was a tarkov video haha. Man you want some views? Become the cheater. Log everything until you get banned.
This is one reason ive quit playing online chess except against stockfish itself. This is a political problem they are also not punishing cheats in other facets of life.
There's checkers, backgammon and lots of other games... And some people cheat or try to cheat in all of these games. And who is so badly concerned about elo?
I agree with a lot of things in the video, but not about winning a lot of games in a row. I m currently around 1650 elo and I tend to often win a lot of games in a row, and latly not loosing many in a row because I made it a strategy to quit for the day if I loose 2 in a row since I previously ended up tilting
A thumbnail with the other_weirdos_exist accounts would have been better, he literally was cheating in my twitch streams
I only play blitz, I can't handle the idea of investing all my energy for almost an hour and realize you played a cheater. I would love to play slower time controls but its just unreasonable
Last weekend I played 50 blitz games. I lost most of them (I'm not a good player anyway). When I entered the game this weekend, suddenly my rank was 100 elo higher.
I checked wtf is going on, and of 50 opponents 27 (!!!) were banned (was there a mass banning happening last week?). That's OVER 50% of my opponents... Online chess currently make no sense.
Im 2300 rapid and I legit cant play it anymore. So terrible so I now just play blitz.
OK, I understand. You can play rapid up to 2300, then only blitz.
do you think cheating is more common or less common at lower elo levels? i’m under 1000 elo and am wondering if i should expect a similar percentage of cheaters as you
I have no idea on which basis you’re coming up with those numbers on % of cheaters. Very confusing. Would suggest you highlight the approach more clearly.
Its really not clear whether some of the people you claim are cheating are actually cheating. One move isnt enough to convict anyone. The game you showed against that 2300 only had one suspicious move which is reasonable
I always wonder how good these 900, 1300 in the Arena blitz tournaments are. I lost there against 1100 and 1300 (and my rating is much much higher) and even I didn't, they are playing way stronger than you would assume. There has to be something wrong. Look at the standing where these type of players are, and where the 1800er are. Often they leading the tournament by far. That does not make not a lot of sense. IRL you would never see such a strong 1100er, NEVER! I am not the type of person who accuse a lot of individual players of cheating, and I am very cautious in doing that, because I am a very bad at blitz, but here is something really wrong. So, there is clear cheating problem at all levels.
Lost to the bong cloud after he sacrificed a bishop for free.
Was just ridiculous how it systematically steals everything.
I stopped playing rapid after reaching 2000, I just can’t bare the idea of facing cheaters
Losing is never a waste of time. When you analize the game without hate, just simojy looking at why your lost happened, you learn. I play against stockfish often to test lines or find new ideas. Sometimes I learn much more than when I play people that don't cheat! Cheating is a problem because we all want to win, that kind of gives us a sense of intellectual superiority or approval. When you lose you first try to find guilt in others. Does anyone talk about cheating when solving chess problems??
I have the same feeling about the chance of running into a cheater. Alas this made me stop playing rapid. The sad thing is, that I am unfortunately too slow for blitz. Therefore I play against the bots and do tactics. Very unsatisfactory!
It's probably more the fact that they'd rather focus on open accounts that might be currently cheating as opposed to closed accounts that won't be cheating currently.
You have a race of AI vs humans. There are, or soon will be, AI that can beat you whilst playing in a human kind of way, avoiding tell-tale patterns such as suspicious winning streaks. There is no point getting good at chess, you'll just be beat by deceptive AIs trained to be human-like, but a bit better. And when players use such an AI only at critical moments, then, there's no defence. Find a new hobby. You cannot win at online games any more. AI's are just too easy to use to cheat.