The Chess Cheat: A Quick Method for Identifying this Parasite

By Brendan J. Norman | Internet Chess

Whether you refer to them as a “chess cheat” a “computer abuser” or whatever, we all know who they are.

This post will discuss my perspective on cheating in online chess (not the global epidemic of cheating in open events) and will provide:

  • A short rant on the topic 🙂

and then…

  • A method you can use to get a pretty good idea of whether your opponent is a chess cheat.

So let’s get started…

The Chess Cheat


This guy is a dishonest and pathetic parasite who instead of sharing his passion for chess in an honest, fair game (and MAYBE even trying to improve his chess) prefers to be a chess cheater over the internet.

Clearly headed for great things in life.

He will typically have two windows open.

One is his chess-playing website of choice, be it, Lichess, or even my new fav Guo Xiang Lian Meng.

The other window open is his chess engine.

This is how he “plays” online chess.

Method 1: Manual Transmission

  1. Opponent moves
  2. He enters the move into the engine window and engine responds.
  3. He relays this move to chess website.
  4. Rinse and repeat until poor guy on the other end is crushed senseless and depressed (especially if he like many naive players, thinks he just “played so poorly that he was crushed“)

This is just the most primitive (but also most common) way of cheating.

There are also more “sophisticated” methods like using apps like…

Method 2: Automatic Transmission

InternetChessKiller (suitable name I know)

This app makes the entire process mentioned above automated, so no window switching is required and it even has (so I’ve heard) ways to make the engine only choose like only the 2nd or 3rd best move in order to avoid detection!

Sneaky rats.

You guys starting to see why I train more and more against engines now? At least I get to tweak, weaken and make the engines more “human” before I play them.

Better than taking on Stockfish 7 with its monstrous 3300 rating, which would demolish even Magnus Carlsen.

A friend asked me recently:

“But I don’t get it… Why would they even want to win in such a way? There can be no pride in winning by relaying engine moves mindlessly onto a chess server”

Typical online cheats are low rated, lazy fucks who never had the work ethic to study chess diligently and become a half decent player.

Don’t get it twisted and be offended by my use of “low rated”

This is not a strength issue.

Even 1600 ELO is not bad if you worked hard to get there. Great effort!

I sincerely applaud you.

One the other hand…

A chess cheat is similar to a guy who tries to use “pick up lines” to get a date, rather than just improving himself to the point that a member of the opposite sex might find him interesting.

If You’re a Chess Cheat… Like All Actions in our Lives, This is a Reflection of Your Inner Workings:

If you cheat at chess online, you clearly orient yourself towards…

  • Instant gratification
  • Moral bankruptcy
  • A slothful mindset
  • Poor self-image (need to win by any means necessary to reinforce ego and feel “successful”)

Cheating Destroys Chess Culture

One of the saddest things to see is the lack of trust now in online chess.

Strong players (myself included) after suffering a crushing loss online, immediately suspect that the opponent has had engine assistance.

Its only natural since the problem is rampant.

Just tonight even the mighty Hikaru Nakamura accused an IM who beat him at blitz of using Komodo to beat him!


Here’s a screen cap.

nakamura accuses chess cheat

 How I’ll do My Part to Deal with Cheats

At the bottom of this post I have made a video showing you a quick way to get an idea if your opponent is using an engine. Don’t accuse people until you have overwhelming proof (read: don’t be Hikaru 😉 )

How You’ll do Your Part to Deal with Cheats

Run suspicious games through the test I give below and if somebody is found to be guilty of cheating, send a report to staff of that website, complete with the links to the games and screenshots from the test.

  • Make a note of the country the cheat says they are from (some cheaters returning with new accounts after being banned seem unable to remove their national pride, and still use the same country indicator.)
  • Make a note of the account name (some cheaters returning with new accounts after being banned only change one digit in the user name to gloat to site staff. So “iMaChEaT” might become “iMaChEaT2”)

Keep all of this gathered intelligence in an “evidence file” for future reference.

How to Identify a Chess Cheat:

Here’s what you’ve been waiting for…rant over I promise! 😉

Its time for the nuts n bolts of this method.

First some no brainers that often point to engine use…

  • Consistent time usage for each move, regardless of position type (cheats don’t use their own brain. So for a decent operator it takes about 5 seconds to switch windows and get a response from the engine, then switch back and make a move. This can be made very obvious when an opponent takes 5 seconds for a forced move or recapture, which a non-cheat would do within 1 second.)
  • PERFECT defence even in scary, knife’s edge, tactical positions with an unsafe king.
  • Account Age: a lot of cheaters have accounts which are less than one year old, due to having previous accounts banned.
  • Failing the “LiChess Test” (video below)

The LiChess Test

Use the test in this video to drive the final nail into the cheat’s coffin.


Chess cheats are rampant in online chess, but I hope the methods in this post can at least allow you to put your mind at ease about whether you really lost fair and square,  and most importantly…if your opponent was playing fairly.

Keep up the training guys, take care and I’ll see ya soon!


  • joe quimby says:

    Instant gratification
    Moral bankruptcy
    A slothful mindset
    Poor self-image (need to win by any means necessary to reinforce ego and feel “successful”)
    You just described an awfully big percentage of people nowadays, whether they play chess or not.
    But given that, there’s proof enough that cheating is probably everywhere in online or even postal chess.

    • “But given that, there’s proof enough that cheating is probably everywhere in online or even postal chess.”

      And this is why I advocate training against weakened/weaker engines.

      At least you get to choose a suitable opponent. 😉

      • joe quimby says:

        I see your point, but if I suspect Engine use is going on at all – regardless of Engine strength – I don’t feel compelled to try and win it like its a normal game; its anything but. And in trying to win it, I’d perpetuate the problem.
        No, I’d rather take my lumps and the loss and use the game to sharpen my tactics. I’ve even heard this great reasoning: “I don’t use an Engine to select my move. I only use it to check if my move is a blunder”. That is by far the stupidest reasoning I’ve ever heard. That is the equivalent of using the weakest Engine available – its STILL using an Engine and its STILL cheating. And I don’t care that in some circles Engine use is allowed. I stay away from those circles.
        Engine cheats would love nothing more than for me to fight Fritz with Fritz.
        I won’t do it.

        • Hey Joe,

          I’m not sure I get what you mean…

          What I mean is, since a good portion (especially at 2100+ level) of players are cheating, it makes more sense to cut down online play.

          Instead, choosing human-like engines to train with (AGAINST) avoids cheaters, and still gives you a good game that FEELS like you’re playing with a buddy from the club.

          This goes especially for long games – a HUGE portion of players of games 15 minutes or more, are using an engine these days – its terrible, I’d prefer to play against my own computer.

  • x1134x says:

    The funniest thing about these articles is the supposed reasoning behind the cheater that is surmised from the cheated. It simply MUST be a low-brow thing. The cheater MUST be lazy. The cheater MUST want to win to achieve the same feeling that the cheated feels when they win fair and square, but is unable so they MUST resort to using a computer. Allow me to enlighten you and the lucky few who happen to read my comment, as to WHY you’re being cheated, and WHY perhaps YOU should start cheating TOO, and why laziness or a desire just to win a game is the furthest thing from the “cheater’s” mind.

    When seriously studying chess it doesn’t take anyone very long before they run into this advise: “to improve, you must study the games of the grandmasters”, or more specifically as Yasser Seirawan suggests: Put yourself in the seat of the grandmaster, and try to make the move you think is best, then check against what the granmaster did, if you match say “yay”, if you don’t make the grandmaster’s move try to understand why it was better, then make the opponent’s next move and repeat. This method actually does produce very good results, however the results are “top-heavy” i.e. this helps you best prepare for and play against *grandmaster* opposition, and will send you deep down memorized classical lines. i.e. these games will increase your chess skills, but will never appear on the board when YOU play some other lesser human. The opponent the grandmaster is playing against is also a grandmaster, So really the game is “over your head” You can study games from grandmasters vs normal people from exhibitions, but they are so few and far between that collating them and drawing statistical conclusions about grandmaster vs normal person chess with your specific flavor of opening repertoire is simply not possible. However a hybrid approach to this method is possible due to internet chess: You can actually get a human to play seriously against you without them becoming discouraged or knowing in advance they’re going to lose horrifically to a strong engine. You can then tailor yourself an opening repertoire and over time catalot a LARGE database of games using only those opening moves where an unbelievable good “master” shows you how to tear apart normal human play. If, instead of filling your database of games you study from full of “human vs human” games, you fill it with “houdini/rybka/stockfish” vs “peon” games you end up quite quickly with a new edition of Bruce Pandolfini’s popular “Chess openings: Traps and Zaps!”, and with free programs like Scid vs PC, you can build a tree and sort these games by popularity, which allows you to focus your study on the lines the humans tend to try, and fill your brain full of smashes when your opponent plays poorly. Now of course this isn’t the end-all be-all of chess-training, the opposite approach where you play 1000 games yourself with your best effort, throw out all wins, and study with a computer (or master) all of your losses, is absolutely essential as well, however when it comes to obtaining good chess “thinking” you can either feel around in the dark that is the endless amount of chess moves in chess hoping to find ideas on your own, OR you can have the jedi master walk you through it and show you how good chess is played. Sure you’re not going to memorize every single line, and every opponent is not going to be weak enough to fall for a human error you’ve already studied, but WHERE TO FIND THOSE, is through being a total jerk and forcing a human to play vs your opening but against ridiculously “perfect” play, and going over 1000s of these games, the repetition helps you learn by “osmosis” simply absorbing the style of the engine you use.

    The final thing I’d say about this is that after using an engine in such a fashion another side-effect benefit is that you come to know a quick and easy way to identify those who use a computer against you. I know what Houdini vs Houdini looks like using my opening systems. I know what houdini vs rybka looks like, I know what komodo vs stockfish looks like. Its just like if someone were to play vs you one of the “classic games” of the masters like “the evergreen game” or heck even “the scholar’s mate” that would pop up in your head if they are played become quite obvious. Human players cling to “opening rules” when faced with a non-classical opening. Chess computers disregard these rules when its more positionally advantageous. These moves become a signature, and with just a few of them in a row, you recognize. “this is houdini 1.5 vs rybka 4” When playing humans the humans when playing poorly get swallowed up, and when playing well, usually choose the same lines that do well but conform to the commonly-taught opening rules of “don’t move a piece twice”, “control the center”, “pawns first”, “knights before bishops”, “castle early” etc. These human lines usually have a few higher level tournament games you can draw from and will have studied. “Karpov V Miles 1980” comes up a lot. When my opponent seems to know the better moves that violate classical principles that KARPOV didn’t see. . .I can immediately draw a conclusion that the opponent is likely using a computer whether I am or not at the time. You don’t really have to dissect it very far, it just becomes obvious that they’re using an engine.

    Finally even my obscure opening (the saint george defense, played both as white and black, as recommended by IM Michael Basman) has had enough analysis and play that the “traps” that humans fall into were already known. For every 1 player that falls for a trap I learned from Rybka or Houdini, 4 fall into a trap I learned from Basman. So no I’m not saying this is a “core” training technique, rather a very effective “supplemental” training technique, all moral implications of defrauding a person of a human level opponent and showing them their true skill of chess vs CHESS ITSELF aside. (there has to be some level of benefit enough to show a few entries in the “pros” column when analyzing having been playing against a computer. Those of us who study chess much know you learn much more from your losses than from your wins and draws.)

    I never really understood the AWE and beauty of the play of players like Fischer who could play like an engine during their time vs normal opponents until I started watching rybka tear humans apart, not using “trade down to won endgame” human strategizing but using *CHESS*, almost every game a “brilliancy” like watching Morphy play at a kiosk in public. That type of study shows you what you’ve been missing the whole time, the things the grandmaster’s opponent DID see, but you never did when studying grandmaster games. Now that I’ve seen tactics used in such brutal fashion, I’m much more aware that THEY ARE THERE TO BE FOUND almost EVERYWHERE in chess games, and I look for them much harder and spend a lot less time thinking about “the rules of the opening say ________” or “the goal of the middle game is _________” to try to guide my move choice between two or three seemingly equal moves. “which move leads to a tactic or is most likely to coordinate pieces such that a tactic appears” is much more my thought process after 10 years of studying computers tearing humans apart. Humans really are HORRIBLE at chess. That’s why there’s such vitriol against cheaters and chess engines, we don’t want to play against CHESS ITSELF, the state-space that is the game and all its moves and beauty, we want to play against someone who misses those moves like we do. {:-)

    • What’s your ELO, may I ask?

      • x1134x says:


      • x1134x says:

        So your focus simply cannot be pried off of ad-hominem? 2200 but I haven’t played rated games in over a decade. Most people who play chess do not compete in organized play, and many would do well if they did, kinda like the best guitarists mostly have other day jobs. Many of my friends play competitively and I usually defeat them easily, unless we’re playing blitz, then I lose on time but material ahead, and still feel like I won. To me chess isn’t about you vs me, its not spassky vs fischer, it’s white vs black. “What’s your highest score on Galaga?” Doesn’t really convey anything about love, dedication, or time spend in pursuit of an enjoyable alien bug zapping experience.
        As an example, I’m using an engine vs someone in a faster game say 3 or 5 mins, I’m up a “game’s” worth of material or more, and time is coming to a close. In your mind the cheater would abandon the slow process of consulting the computer, start moving fastly, even stupidly, to secure the “win” on time, because winning is everything, and the rating is the cheater’s focus. I however will continue to consult the computer, making 3300 level moves all the way until I win or I lose on time. And then retrieve the payload, the thing I’m after, the pgn of perfect play vs humanly selected moves. I’m not looking for “I beat you” or even “here’s how I beat you” I’m after “here’s HOW TO beat this guy”. “To win” is not the focua. “To witness how to win” is. Then later I will play vs your move list, myself. I will play good and bad moves, but rather than me leading the lesson down an untenabe line, we’ll course correct the whole time. I can measure my progress by scoring my chosen move against the score of the computer’s best option . Basically your lichess method in reverse. Use it to rate vs known best play with the ability to continue lines vs human selected moves rather than engine vs engine moves when you make your first mistake. Studying your own games you have to continue down your bone headed lines or pretend you and your opponent both suddenly start playing like 3300 engines. My way, you continue down the correct line, and see both how you can still blunder it away, and how your opponents human tendencies can be defeated, because theyre still present down the line you should have chosen, because you cheated earlier when collecting the data. Also whtly I don’t get my briefs in a bunch when people are obviously using an engine against me, whn im not, hopefully they learn something from it, I know I will.

        • joe quimby says:

          Excuse me for butting in but I’d like to clarify: Are you recommending the use of an Engine during a game vs. another player? – or simply the use of an Engine to train with, outside of a game?
          If the former, then I don’t understand your defense of a cheater as someone who has other goals in mind than just winning. He knows he WILL play error- free when using an Engine to select his moves. So how then is winning not his goal? because its done under the guise of amassing research?

          • x1134x says:

            because win-loss record is not important, move selection is. Certainly some people cheat only to win. Their behavior can be easily predicted, if their goals are known.

            I guess you don’t understand my example (my intuition says you’re really not trying) but in case you are I’ll try to be more detailed.

            Lets take two chess cheaters with engine assistance, playing against two human players playing without engine assistance. The first one’s goal is to WIN. The second one’s goal is to study chess and get better. Their opponents happen to try the same opening, then fall for the same trap, and have a lost game, and in the process have lost tons of time on the clock from actually thinking, and the chess cheats are also low on time due to transfer time of each move eating up much of their time in a faster game. Now let take the first cheater who’s sole goal is to WIN THIS GAME, and feel an endorphin rush of having achieved a victory, or whatnot. Realizing that they can and or will lose if they keep consulting the computer, as their opponent seeing the clocks will simply start making quick moves, even very poor moves and secure a victory on time, they will abandon consulting the computer, and also make quick moves in an attempt to stay ahead on time, win on time, or not lose on time. The 2nd player has no such cares. A loss on time or a win on time or a draw are meaningless. The only thing that he cares about is the game moves being something which can and SHOULD be studied to get better at playing chess. Playing random moves doesn’t serve this purpose. To the 2nd player it is much better to lose the game on time having always consulted the computer to get as long a game score with 3300 level play as possible to study later.(computers do NOT play “error-free”, else they’d never lose to each other.)

            It doesn’t matter to the 2nd player if their record is 100-0 or 0-100. Winning is absolutely meaningless. What is obtained from the “cheating” is something that cannot be derived in any other way and is immensely useful to improve one’s own chess play: Human selected moves in endless “brilliancy” games.

            To understand better, I guess maybe we should discuss how people study chess? How do YOU study chess? How do YOU think one gets any better at chess? Now contrast that against say a soviet block prodigy: How did HE get better at chess? He had grandmasters showing him how to destroy human play. They had seconds and servants collating and statistically analyzing the play of other players. They correctly identified that to win at chess you do not study solely chess, you study humans, and their tendencies.

            You don’t have those resources, they’re not available for the most part, and certainly nothing catered or made specifically for your opening preferences and playing style. But it is available to you for free. You just have to collect it.

            Now I guess maybe you’re thinking “ahh this guy just copy/pastes the game moves and saves them, and plays back over a few of them, to justify to himself his cheating of others.

            I have over 10,000 games collated for just my specific opening all humans vs 3000+ ELO moves. I have a tree program that collates all of them, flips the white games to mirrors, and check them against the black games for positions that transpose, and give me statistics about what the likelihood of a human trying a specific move against me are. Having collected statistically significant data, I can then use that data not just as a guide of how to play better, but also as a guide as to where to spend my time studying. Where a book about an opening may spend 50 pages on a line grandmasters play, you may find you never play that line ever in your life because you don’t play grandmasters. But you’ve been preparing for them, because that’s what the books are written like, this is what I found. The humans who aren’t good at chess tend to select different (poorer) moves in the opening and those lines are the ones that must be studied more thoroughly. The only goal I ever have of watching a human play against a computer engine, is to learn from the HUMAN, what poor moves they tend to play. The intent from the get-go is to become a better chess player, the process of cheating is simply the collection of DATA, not “feel good wins”.

          • joe quimby says:

            On the contrary, I’m trying very hard to understand how you can take credit for a win you didn’t create and in the same breath say that winning doesn’t matter to you. Then start resigning games when you reach a won position. Otherwise you could be hiding behind the research argument and as such you shouldn’t have too difficult a time understanding why you’re being accused of it.

          • x1134x says:

            So going back to your edit:

            1. If i say “I’m going to use an engine”, I find essentially zero willing participants. No study, no data, no jazz, those willing, just fire up their engine and I watch another engine vs engine game that I can already name their engine because I’ve already seen engine vs engine down my opening lines. Defeats the purpose of the research. Under research terms it qualifies under “establish inability to consent”. Seeking consent destroys the viability of the research, so ethical review is triggered to determine violation of the rights of the studied. I don’t need to hear where you come down on that I’m sure I’m already fully aware you’d advise the research council that the participant’s “rights” were violated, even though there is no “right” to know if you’re playing chess against a human or computer in any country. I’m certain it would pass ethical review.

            2. A game is not won until checkmate. There is always something to be learned from 3300+ play in every phase of the game, each and every move is valuable, more valuable than the win/loss stat on the web site by far. Its a incomplete study to stop at an assumed “won game” and furthermore in practice of studying those games you’re brought right back to the conundrum that started up the research to begin with: no human supplied moves to practice against. Yes I’m up in a game according to the computer by 5 pawns, and should be able to win, but am without the moves to study HOW the game would proceed. All I could study would be engine provided moves, which is far beyond more useless than study GM vs GM games. LIke I said in the first post about studying human vs human games: from the point at which I resign, my human opponent isn’t going to continue to play like an engine. When studying these games I find there are many MANY lessons still to be learned, I’ll be playing blindly to the score of the game, perhaps even engine-matching well, (I score my whole performance after the entirety of the game, not move-by-move) and will have progressed into the point where the engine has an obvious advantage in material and position. . . . then I’ll select a move that I later find out loses not only ALL that advantage but the entire GAME sometimes by a brilliancy available to the opponent I didn’t see. So saying “at that point I could easily just find the win from there so I’ll resign the game now” removes a ton of teachable material. However when I get to the point where the engine finds MATE, I ALWAYS, ALWAYS ALWAYS, instantly resign, and have the computer just finish filling the game score out as though my opponent had actually found the longest resistance to mate, and study that game while giving my opponent the win, because that win does NOT matter to me AT ALL, and its the opportunity to make the account not flag review for suspension. I’ve had at least a few dozen accounts frozen from statistical analysis of my games matching engine moves. It doesn’t bother me to start over again, the score, rating, win/loss record could not be more meaningless to me. My prize is my data. My prize is going from being hit by opening novelty traps, to being impervious to opening traps. My prize is essentially always knowing the landscape of the game further than my opponent. My prize is all that time on my clock making instantaneous moves as I watch my opponent play like the human he is having to work out the problems other humans tried earlier against me because I was out of his book from move 1 or 2, but he’s in my book till move 12 to 30.

            Its night and day difference in my ability to play vs other humans before and after this research project and computer program. before I studied chess, and found I never saw on the board against a human anything I’d ever studied before beyond a few moves, or I’d memorize classical lines, and only end up playing those against people who knew them better than me, humans who sucked would divert much sooner. Now I study what humans play against my system, and am much better prepared.

            Another advantage of having built this system I have is I know right away when I’m playing vs a computer. Every popular engine even back to weaker engines like chessmaster choose the same attack against my opening, and humans have a less than 1% chance of choosing that line. If you really want an “opt out” of playing vs engine users. Do what I did. I can opt out every time.

          • x1134x says:

            Ok I’m done editing the previous post, I’ll start another one to expound further if you’d like.

          • joe quimby says:

            There’s no need. I see clearly where you’re coming from. if you’re up 5 pawns and still want to see what the engine does, that completely counteracts your claim that someday you’re gonna play honestly .

          • x1134x says:

            How? There is no logic to that statement at all. “Because he wants to learn from an engine how to play top level chess even when very far ahead, he therefore will not play honestly in the future. . . . ”

            Yeah, that doesn’t compute, they are completely unrelated.

            You are under the false assumption that there’s nothing to learn from an engine during a game when the score is lopsided. You are wrong. There’s nothing to learn by playing it out yourself and winning in 100 moves, (you already know that) and plenty to learn by being shown how to win in 22.(which you don’t know). The claim is that the motivation for cheating is data collection of human play to study, rather than the “thrill” of achieving a victory without earning it. Future honest play is off-topic.

          • joe quimby says:

            Dude, if you need to see how a win is implemented when you’re up 5 pawns I don’t think any amount of research is really gonna help you.

          • x1134x says:

            Yeah this is going to be difficult, you read the words: “but am without the moves to study HOW the game would proceed.” as “but still do not know how to convert that advantage to a victory.”

            How can I explain to you the difference between what I was saying and what communication was received. . . .? hmm. Perhaps we go back to the example of “free to play” games where a “free” player is just being the “a.i.” inside of a game character. So rather than having a live player to player interaction, one could also RECORD the game actions of the “free player” and play them back again on another character, essentially time-shifting the human player’s A.I. That’s essentially what I’m doing with the chess trainer I built. I’m time-shifting their human play to when it is convenient for me or other users to sit down and play an instructional game against REALISTICALLY HUMAN opposition. Essentially building a free teacher of chess that teaches at 3300 level and providing that system with human opponent play data to work with. So I’m not without the understanding of how to win the game. When collecting data the win is meaningless. What we’re collecting is human data to make the game/trainer play like a human. I’m without the information of *how the game would proceed*.

            Often times a computer will show the score of the game as +3 or more and material on the board for the next 6 moves or so is still even. The engine sees all avenues lead to defeat, if we stop at an engine score that shows a winning advantage, we’re without knowing what path a HUMAN opponent would take through the combination.

            We could obviously stop and study from there an engine’s longest continuation, but again the purpose of studying human play is defeated the instant you stop collecting data. We’re right back to the conundrum of using an engine to study your own human vs human games: we don’t know what the human would have played against the RIGHT move, just what they played against your wrong move.

            The reason I’ll resign when the engine finds mate is that when studying those games after playing through Seirawan-style, I’ll go through and make that “mate in X” into a puzzle online, so I actually DO end up studying whatever the human would have played in addition to “most moves to mate” because I enter every possibility into the puzzle creator.

            This sentence doesn’t make any sense to me, sorry: “your unwillingness to even take over for the Engine[sic] once you have a won game you are trying to remove all trial and error from your learning process.”

            Corrected I’d restate it: “you are removing your own human error from your program’s database data, and preserving its teaching value in future study.”

            Taking over for the engine is simply saying “I’m so smart there’s nothing I could possibly learn from here”. Yes you could win in another 65 moves. . . . and the engine can win in 20 more. There is plenty you can learn and can make you a better chess player in that data that may apply to a future game where you’re even on material and position.

            I study failure plenty, I guess maybe you don’t understand the method:

            I say I want to study an “easy” level game, the game starts and the database makes the first move in a game-score. I then make the move I’d make if I were playing them without consulting any engine, just playing with my brain. If I chose the same move the engine chose during that game, the database will provide the next human move, and the game will continue. If I do not make the right move, the piece will move back, and I have another chance to make the right move, I get 3 guesses. If my guesses are wrong, then the move the engine did play is made followed by the next move in the database, and my guesses are recorded. This process continues until the end of the game. After the game is over, the engine scores my guesses against the score of the engine move and calculates the differences, and generates a score for the game, which I can then go back over and see where each mis-guess was in the game and what its “loss” was, and what varations those poor moves lead to, if they really were poor at all, often a “wrong” move is almost equivalent to the engine move in score, so the game lets you flip through your mistakes in order of severity, so you can be take directly to your failures and start to learn why you made them.

            Essentially its like “what if those ‘personalities’ in chess games really DID play like humans instead of playing like engines tweaked to kind of but not really play like humans”, combined with “and what if I had a 3300+ player there with me to teach me by the process of trial and error, how to play top-level chess.

          • joe quimby says:

            Also, it sounds like the bulk of your 10k games are 3-5 minutes long? If the human’s poor move selection is part of what interests you (in as much as the Engine response does interest you) I would think that most of the poorness of the human move stems from no real thinking time. Why not collect the human responses from games with MUCH longer time controls? 3-5 minute games are exactly the type that the adrenalin-rush-winning-is-everything types live for. So your preference for that short time control does nothing to support your claim that you don’t care about winning.

          • x1134x says:

            I see that you have edited your post, my response was to your post when it was 1 paragraph long.

            Please try very hard to read ONLY the words written on the page and not add extra words between the lines. Please help me be a clearer writer: WHERE on earth did you get the idea that “the bulk of the games” were 3-5 mins long, and that I have a “preference” for short time controls? None of that is written in my post. I gave the short game as an example because in my experience of playing thousands of games I have of course ran into other people also using engine assistance, and know how these games can end, and they serve as a great example for me to describe the difference in motivations, between cheaters you describe, and cheaters I describe. In that scenario a choice must be made: which is more important: getting a win, or collecting useful data. Its a point in which what you surmise about the motivation of chess cheaters is shown to be false, where they have a victory within easy reach yet choose to eschew it, in pursuit of the actual payload they are after.

            I have collated games from many many time controls, and yes the 3-5 min games are classified under the “easy” section, when you want something where the answers are more obvious and the human blunders earlier and more costly, and games on longer time controls vs GMs and IMs and FMs or scored well against the computer for a long time but are not engine vs engine are in the “hard” section.

            In my mind I see them as extensions of two different books: The 3-5 min games where humans fold to engines within 10 to 25 moves are my own personally made versions of Bruce Pandolfini’s popular “Chess openings: Traps and Zaps!”

            the harder games are an extension of Yasser Seirawan’s “play winning chess” and “winning chess brilliancies”, custom-made for my opening repertoire and playing style.

          • joe quimby says:

            Right right right …lets cut to the chase. Why don’t you resign the games once you are up material?

          • This is a good question.

            You’d still get the statistical data, while not robbing the poor opponent of his hard earned rating.

          • x1134x says:

            Addressed it in another post, I do resign won games all the time.

          • joe quimby says:

            I mentioned the 3-5 minute games because you mentioned as an example a 3-5 minute game where the human player gets in time trouble. Forgive me for concluding that a game YOU mentioned as a sample did not make up the majority of your games. You have all the data at your hands, so maybe you can publish the stats on what your time controls are. I would be very surprised to see the bulk of them being an hour or longer. Because that would not fit into the cheater profile.

          • x1134x says:

            You wrote “Forgive me for concluding that a game YOU mentioned as a sample did not make up the majority of your games.”

            first let me fix it for you: “Forgive me for concluding that one game YOU mentioned as one sample makes up the majority of your games.”

            You’re forgiven for the illogical concluding that one single example represented the whole 10,000.

            The majority of the games are 2 *days* per move (correspondence time controls, so rather than playing sequentially game after game I can play many simultaneously, increasing the speed of data collected while maintaining quality, rather than reducing game length to speed up data collection.

            Sorry I don’t fit your “cheater profile”, that’s what I’ve been telling you this whole time. A lot of people cheating you are simply cataloging you, they don’t care about the win/loss stats of an account that is going to be deleted by the admins of the site soon anyway. You seem to think they want a ratings boost, I guess you should cheat just to find out how long your account doesn’t last. Any chess site worth its salt has cheat checks that identify engine users. They simply make the false assumption that once they’ve banned an IP they have gotten rid of the “cheater”, when IPs are a dime a dozen. AND they only catch people using engine moves every move. People who want to CHEAT and WIN, and achieve a HIGH SCORE. . . .know full well how to bypass the cheat detection systems. They know to use the engine in multi-move mode and to choose 4th 5th and 6th best moves when the landscape is even, strike when the human makes an error, then use human “skill” to finish the game. Those “cheaters” are undetectable and litter the top-scores of every chess site.

            Which brings me back to a point you ‘d said somewhere in this convoluted replies and edits thread we have going here, that the humans I’m playing are somehow under the impression I’m playing “fair” as a human not consulting an engine. That’s hilarious. I’d imagine most every human chess player playing online takes the opposite approach. We all assume everyone is using an engine until they prove us wrong. If we’re not there already we will be, because its the logical position to take.

          • joe quimby says:

            Sorry, I disagree. Read ‘My System’, read ‘The search for chess perfection’. Play against an engine to sharpen your tactical ability. Immerse yourself in the game and be prepared to take your licks in the first 200 games you play. THAT is how you learn and THAT is how lessons stick. You are making the same mistake that many do nowadays: – that learning can somehow be optimized to the point where there is absolutely no ‘wasted’ time in the learning process. That is a pipe dream. Learning, by definition, is one step forward, two steps back. Cheating for a supposed greater goal is still cheating. The end does not justify the means, not in my book. Though I will say you know an awful lot about how to game the system, for a non-cheater that is.

          • x1134x says:

            You couldnt’ be more wrong. Bumbling around and losing doesn’t teach you anything but how NOT to play. There’s so much how “not” to play in chess, you can’t study it all, but you can focus on how TO play, and make great gains. This is Seirawan suggest studying the play of good players for this reason, watching HOW TO PLAY RIGHT, imbues a person with much more chess skill than continually finding out how to play wrong. I have read Nimzovich and Purdy. I Play both WITH and against an engine to sharpen my tactical ability.

          • joe quimby says:

            Well then imagine how ashamed the millions of chess players throughout history must feel at having done things the wrong way.

          • x1134x says:

            Yes what chess prowess did they achieve? Millions of GMs? Beside the fact that they didn’t have the ability to do what I’m doing, or many may just have done so.

            I have read Nimzovich and Purdy. I Play both WITH and against an engine to sharpen my tactical ability. I plateaued big time and made almost no gains until I started studying human play vs an undocumented unresearched system of play, allowing a 3300 level player map out the system and the humans to statistically map the human tendencies through that undiscovered territory for me. I know the equivalent of the petroff’s, ruy lopez, caro, italian, sicilian, QGD, etc all down an unavoidable line that only 2 books have ever been written about, the latest a reprint in 1990. Without aid, that would take the same amount of time it took humans the first time, (with similar inaccuracy). Without polling human responses one could amass millions of moves and games and be completely out of sync with what humans on average play against the system.

          • joe quimby says:

            Cheating is cheating. Doesn’t matter what your goal is. If you can sleep at night, then more power to you.

          • x1134x says:

            I agree with you, yes, cheating is indeed cheating. My initial comment was about you guys surmising as to the “goal” of the cheater, if it doesn’t matter, then it doesn’t matter. It seemed by those posts that the reason really did matter, but I can see clearly now that what matters is for you to be able to tell yourself, that a computer engine user just wanted some rating points on a computer screen to read a little higher number, and couldn’t have any other reason for using an engine, and when their account is frozen or deleted, they cry like babies over it, so that you can sleep at night. Someone out there getting better because they’re learning from your poor play, gives you nightmares it seems. Knowing the real motivation behind the engine user isn’t actually very important to the person asking “why does someone use an engine?” Its more of an elementary school “put someone else down to build yourself up” behavior. A hypothetical question posed to rouse the put down posts.

            If my team cheated and we won the Indy 500 unfairly, I’d be so ashamed of that victory I’d personally NARC on all of us, I go back into stores to tell them they forgot to charge us for dog food on the bottom of our cart, and apologize profusely for our forgetfulness. I think its more in how you parse ethics, by letter or by spirit. The spirit of the “Only free refills during the same visit” rule is that you pay for a drink, and for that meal if you need more drink you’re free to refill that drink as much as you need, but please pay for the next meal’s drink. The letter of that would be “if you step foot out of the main doorways, then you can’t come back in and refill your drink”. So to you someone who fills a drink, then goes to their car for something with their drink in their hand should pay for another? Letters of “rules” get broken all the time, “spirits of rules” are commonly called for. How does this apply to “cheating” using engines? The agreement between an engine user and the website offering up the chess play is “do not use an engine, OR your account may be subject to suspension or deletion”. That is a completely acceptable agreement to me, but to me it means something a little different than perhaps it does to you. You as a player with a brain should understand, you’re not “guranteed” nor “owed” a game against a player on the internet where they use only their brain and not an engine, you signed up for the same service and same terms. You logically only have the expectation that the site will take action against engine users, not that they will not be exist. Taking the position of “I was expecting a fair game, and you took that opportunity away from me”, has the false basis in the expectation. Similarly those cheated in real life games by people scheming to get engine assistance to the table of real life tournament games, they only expectation their opponents should have is that if caught the cheater will be ejected from the tournament. Expecting the opponent not to cheat is a good way to miss catching them cheating. At least Brendan is on the right track with that.

          • I think rather than debating the METHOD used and it’s moral justification (or lack thereof), let’s just see what results are being reaped by this unusual method.

            Could you share your own rating prior to cheating and then after?

            How much have you improved your playing level? This is the final goal, right? 🙂

            (I mean the rating earned from playing games alone, and outside of your ‘research’) 😉

          • x1134x says:

            Rather than debate the topic, you’d prefer to debate the qualities of the presenter? Yes, you need to google”ad hominem”, people knew this to be a logical fallacy since the time Latin was the dominant language. The saying is “The plural of anecdote is ‘anecdotes’, not ‘data’.”, so the singular of anecdotes is that I went from struggling, and losing, to winning.

          • No. I don’t need to google “ad hominem”, as I’m fully aware of what it is.

            I’m also fully aware of other logical fallacies and communicative devices used by people – especially to aid manipulation.

            You came on my site under an anonymous screen-name, with the condescending attitude of somebody who seems to think he knows more than others – even referring to honest people who improve by trial and error (rather than cheating with an engine) as “bumbling around and learning how NOT to play”.

            Then when asked for quantifiable proof as to the validity of your method (are we being scientific here or not?), you tell me to “google ad hominem”.

            That’s just a pseudo-intellectual way of saying “I don’t need to prove ANY of the claims made in my thousands of words of comments, because you’re picking on me”.

            Not at all – I’m sincerely asking for proof of the method that (i assume) is better than “bumbling around learning how not to play”.

            I’m pretty sure that neither @joe_quimby:disqus , I or anyone else cares WHY people cheat, just that they DO…and focusing on that point alone might be a case of creating a “straw man”.

            If somebody said that your gf/wife cheated, would you be focused on verifying if she DID do it (or not), or would you be pontificating about WHY she did it?

            Just as in the case of our chess cheater, nobody cares about why.

            P.S. I know that you lied about being 2200 Elo as well, because you mentioned having the book “Chess Openings Traps and Zaps” – this was a dead giveaway.

            I had this book when I was 15 and its absolute garbage. Nobody over 1800 would find any use for it and in fact GM Tony Miles famously wrote a review for a Pandolfini book once, in two words stating it as: “Utter Crap”

            Somebody who clearly has a decent head on his shoulders, might like to use some intellectual honesty and do things the right way.

            It’ll get you much further in the long run.

          • x1134x says:

            You said “That’s just a pseudo-intellectual way of saying “I don’t need to prove ANY of the claims made in my thousands of words of comments, because you’re picking on me”.”

            Nope, its recognizing and reminding you of the fact that even if it doesn’t work at all for me, that data is completely impertinent to the validity of the proposed methodology.

            I really appreciate the laugh, though with your condescending “gotcha” of a person who studies a Pandolfini book as novice with the completely erroneous attribution of Tony Miles’ review of Eric Schiller’s Unorthodox Chess Openings book as a review of a Pandolfini book. Pandolfini was one of Josh Waitzkin’s first coaches, and he achieved a rating of 2480. I agree that book is very elementary, its just the most resembling the idea of what is obtained by sorting a human sampling of an opening vs an engine by game length. “He reads Pandolfini, therefore he cannot ever have been rated at 2200.” I thought I read you saying you understood fallacious logic.

            Coincidentally my work is based on an idea Tony Miles had.

            You also called the topic sentence of an initial post a “strawman”. The entirety of the post is about the motivations behind the engine use, adding an additional reason to the list of “lazy”, “stupid”, “instant gratification”, that were read by me and were the topic of my post. People do write up a big kindergarden list of reasons that must motivate an engine user, they’re littered all over this page and every other comment board on engine cheating. That’s all my post was addressing. Sorry I was too high detail, and got that discussion diverted into the details of how the data collected is used personally by me.

    • youdiscussme says:

      Are you going to publish your findings in the form of an openings book?

      You should post some of the opening innovations that you’ve discovered through this research on your youtube channel.

  • John Diamond says:

    Hi Brendan thanks interesting article and I agree with your
    feelings . Just wanted to add a bit about internetchesskiller as someone
    recently showed me this program running (privately and not on an actual game site I hasten to add) and I wanted to make people more aware
    so they can watch out for it. You say above that it automates the process so no
    window switching is required.Thats true but in fact it’s even more devious than
    that. You can choose to have it play the moves automatically for you to a
    set time control where it actually moves the mouse for you in which case move
    timing would be consistent regardless of complexity of the position (by the way
    this is how people use it as an autoplayer in bullet time controls like 1
    minute for the entire game) , OR you can set it to manual input mode so that
    the program suggests the engine move in a little window but you then actually
    make the move manually with your mouse. This means that anyone who has any
    chess nous can actually vary the move timing so that simple moves like piece
    recaptures are quick but complex moves/positions appear to take longer.
    Also using this method, you can choose to use the engine’s offered move or do a
    move of your own choosing, so the game is a mixture of engine moves pus your
    own moves making detection more difficult. Also you are correct that you
    can set it to offer the first, second or third engine move choice to try to
    make detection harder. You can also load any engine you choose so by loading a
    weaker engine (eg. 2000 elo and not the 3300 elo super engines) that could also
    make it less obvious. Just sharing this in case people aren’t aware that
    in the hands of a clever user this method of cheating can be very hard to
    detect and prove categorically.

  • Tomasz says:

    Hi Brendan

    I hope you can think of writing the second part of the article. At least I want to show you some evidence and suggestions.

    There are some exceptions you CANNOT say (at least with absolute certainity) anyone is cheating. What are they? It is my list of doubts 😉 🙂

    1. The game is simply too short. Here are two (real) examples (from my practice).

    1. e4 e5 2. Nf3 Nf6 3. Nxe5 Nxe4 4. Qe2 Nf6 5. Nc6+ Qe7 6. Nxe7 1-0

    1. d4 d5 2. c4 e6 3. Nc3 Nf6 4. e3 c5 5. Nf3 a6 6. a4 dxc4 7. Bxc4 b5 8. axb5 axb5 9. Bxb5+ Bd7 10. Rxa8 Bxb5 11. Nxb5 1-0

    2. All the game’s moves are forced (or very easy to find/play). It is very close to the examples presented before.

    3. You played too small number of games. If some weak opponents could have played 2-3 games against you and you (being 2200 and they are about 1500-1600) would crushed them soundly – they could think of you as a cheater.

    Of course there could be more points to consider, but they are just for the reflection and some suggestion.

    What are the conclusions?

    1. Do not take into consideration the games the ends before move 15.
    2. Do not take into consideration the games that have obvious and/or forcing moves [very easy to find for most average players like: captures, checkmate finals, combinations that leads to mate, etc.].
    3. Do not take into consideration the games the ends with a known [especially forced] variation (theory) and after that your opponent blundered (and resigned or disconnected purposely).

    And below you can see one of the best games of mine – and it consists only ONE inaccuracy. It is a proof we can play “perfect games” from time to time – without being a titled player (not to mention World Champion or using an engine). However the difference between the opponents was 300-350 rating points (Let’s say: a strong B-class player against weak C-class player).

    1. e4 e5 2. Nf3 Nc6 3. Bc4 Bc5 4. O-O Nf6 5. d4 Bxd4 6. Nxd4 Nxd4 7. f4 d5 8. fxe5 dxc4 9. exf6 gxf6 10. Be3 Nc6 11. Qh5 Be6 12. Nc3 Ne5 13. Rad1 Qe7 14. Nd5 Bxd5 15. Rxd5 Ng6 16. Bd4 Qxe4 17. Bxf6 Ne7 18. Bxe7 Qe3+ 19. Kh1 Qxe7 20. Re5 1-0

    To be completely clear: I DO NOT defend cheaters, but I want to inform/warn people about false accusations. There are additional (necessary) conditions that have to be met… before they accuse anyone of using an engine.

    Let me know if you agree on me and if my examples convinced you about my conclusions.

    BTW. I am only B-class player (approx. 1700-1750 ELO) and If I could play some (close to) “perfect” games, try to imagine how strong players (2200+) can make an outstanding performance.

    • Interesting thoughts Tomasz,

      Yeah clearly if a game is too short, we cannot judge it accurately or if the game’s moves are easy to play thats also true.

      The number of games played isn’t important though, since clear indications of cheating can be found even from a single game played (provided it fits into the criteria mentioned in this post).

      Another way to avoid all online cheating is to simply play on a paid chess site like ICC, which I have began to do. People aren’t likely to cheat (and risk a ban) when they’ve paid for their account. 🙂

  • gareth says:

    Hi Brendan

    Informative post, after spending many years away from chess
    i do believe online chess is a fantastic portal to supplement over the board
    club chess. Personally unfriendly manner and engine abuse should be dealt with
    and not at all tolerated. In reading your article in relation to the 3 games
    against that same player, is amazing has chosen to rule against your suspicion.
    I am a current premier member of and dismayed they would take
    that view. I would say there is a strong argument for creating closed clubs of
    say 100 players controlled by elected members. The idea being to hold the trump cards regarding
    dealing with cheaters and plain bad manners i.e not relying on paid admins to make a decision. Another way also is for over the board chess clubs to create a national world online chess gamesmatches and tournaments open only to club members with published
    over the board ratings. The above would easily able to root out any offenders
    and create a better online chess experience.

    • Brendan J. Norman says:

      Hi Gareth, Sorry its taken a while to get to this comment and respond. I agree that the internet is amazing for participating in chess communities and improving our chess and it really annoys me that a small percentage cheat and ruin it for others. Your idea of closed clubs is also brilliant (seriously), but we’d need to have some way to police it. So your ideas of elected representatives (assuringly with shared philosophy) worldwide is quite interesting. Perhaps some sort of “Fair Online Chess Association” or something for serious players and you need to register with your local official and register with I.D etc? 🙂

  • >