![]() |
![]() |
#1 |
"Brian"
Jul 2007
The Netherlands
327810 Posts |
![]()
Cheating at chess must be as old as the game itself. But obviously the methods have evolved out of all recognition over the centuries.
As a teenager one of my favourite chess books was "How to Cheat at Chess" by William R. (Bill) Hartston, a British International Master. It was first published in 1976. Although largely humorous (sometimes bordering on facetious, though frequently genuinely funny), the book also contains a wealth of information about how players could, and did, cheat in those days, right up to the highest levels of the game too. The subtlety varied: blatant disregard for the rules (like Yugoslav Grand Master Matulovic's taking back a blunder, saying "j'adoube" and making a different move against Bilek at the 1967 Interzonal in Sousse, and getting away with it too) was one end of the scale. More subtle would be something like offering a draw in a manner which wasn't perfectly clear - perhaps exploiting language difficulties to achieve this - in order simply to see if your opponent was amenable to a draw without actually having to commit yourself to the draw if it turned out that he was: knowing if your opponent is trying to win can be important psychologically, but it is not information which you have a right to know. Of course cheating at chess had taken on a completely new dimension during the few decades since that book was written. No prizes for guessing why. And I believe that one of the biggest headaches for people in charge of online chess servers is detecting and combatting this cheating, for which the cheating player only needs the same machine that he or she is using to connect to the server plus some appropriate software. I play sometimes on the ICC server. The administrators there are a bit secretive about how they detect computer cheating - I guess because they figure that if they publish their methods that would make it easier for cheating players to circumvent them. But I'm very interested to know what methods could be used to detect computer cheating at chess if anyone knows or suspects any. Any discussion about cheating in general is equally welcome. |
![]() |
![]() |
![]() |
#2 |
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
19·232 Posts |
![]()
A long time ago, I wrote simple robots that play ten different games at a Russian game server. (The only trick was, at the time, that the protocol was not open, nor was the client code, but I was lucky to hack it. I did inform the server owners that my attack was successful and they upped the security level on the client many fold since then. But we all had good fun about that. All in good spirit. Now, five years later, they are still working on an open protocol.)
Anyway, I specifically started with implementing a chess robot as the easiest (the least event-intensive game) communication protocol (cf. backgammon: chess - only moves and a few legal propositions: 'draw', 'new game'; backgammon - throw dice or double, get the dice event, transfer a variable amount of moves, accept or drop the opponentโs double event, etc, etc). The robot was immediately successful as a novelty, but I deliberately never put any close source game programs against people (neither I nor the server owners would need legal problems for a commercial use of Rybka or Fritz or whatnot); so for chess, it was always Crafty. It plays there for close to five years now. It is clearly marked as a computer program and people playing against it know exactly what they are getting. However, an interesting side use for the robot later emerged. All the new 'smarta..' program-assisted cheaters (they always come; every week someone thinks that they โinventedโ this neat and totally unheard of idea J) would inevitably come to play it (as to a honey pot) and beat it (Rybka beats Crafty hands down, and you can easily transfer moves at a 5 minute control), and the protocols of those games are easily analyzed later to see that a high percentage of moves were not 'human' like. The 'dumbest' cheaters (those that only transfer moves; i.e. are human 'robots') are easily diagnosed with this tool. There's still a huge problem of talented (and well technically equipped) cheaters who are fair players in their own right and who only use the computer help to get an advantage and then realize it themselves. Or complicated variations on that (using the program, then not, then using again; usually it requires a good team there behind a client computer). The diagnosing and labeling cheaters on game servers is often a controversial and a not well rewarded job. Last fiddled with by Batalov on 2011-06-10 at 18:11 Reason: "event-intensive" typo corrected |
![]() |
![]() |
![]() |
#3 |
Romulan Interpreter
"name field"
Jun 2011
Thailand
240418 Posts |
![]()
if by cheating you mean letting a computer program play instead of you, this is very easy to detect. say you play against an opponent on the web, each move he sends to you, you put his move in a computer program for chess and whatever it answers, you send back this answered move to your opponent. this way he plays in fact against a (very strong) chess program, and not against you. good computer program will beat 99% of human opponents. I used to play a lot of gomoku/renju/scrabble over the net many years ago. The script that paints your board sends a "tag" to the server every time it loses the focus. If you use a program to play, your web interface will lose the focus every time when you execute a move on the chess program (clicking on the board to make the move from your opponent), that is obviously after every move you get. More clever servers (scripts) read the signatures of the running programs on your computer, like for example renjuclass.com web interface won't start at all if you are running "fiver" or "gomoku solver" or other 10 or 20 different renju programs. Other servers are "more discrete", they will run normally, but will recors your "switching from a program to the other" activity (i.e. from your web browser to some other application as a chess or gomoku program). and your stats will go down or you will be later on labeled as "cheater", wondering wtf??? :D For example this is the way kurnik.com is doing it, except on kurnik your name becomes gray every time you switch to other program causing the web browser to lose focus. Of course you could use two computers, to have the gaming program separate, but is still "detectable", specially in gomoku, where you start with a different board configuration every time (see swap rules). Good players have their own "studied" configurations (openings) that will almost every time "cheat" the gomoku program (the program is not able to find the best answer in reasonable amount of time, so you either lose by time, or you are forced to do a bad move in reply). personally, I can detect with 100% accuracy if an opponent uses a program against me in 5-6 moves, if I am the one who open the game, and in this case, beside of the fact that I will report him, the opponent will generally lose that game. games are stored and - if reported - are analized by good players that can even say what program was used. The machine plays by certain rules and in renju/gomoku you can easily "force" it in making a wrong move, if you studied its playing stile. I assume in chess is the same (I am only an average chess player)
|
![]() |
![]() |
![]() |
#4 |
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2
100111010000112 Posts |
![]()
gomoku is an excellent example.
(go is even better - except in go, using a computer is actually a handicap. Still. Even though over just a few last years, great strides were made.) By playing against a gomoku robot with fiver, users get themselves a six-in-a-row which is always very funny to watch. ('gambler.ru' gomoku rules are with a strict five, no sixes. They have renju as well.) |
![]() |
![]() |
![]() |
#5 |
Dec 2010
Monticello
5×359 Posts |
![]()
So, it looks like there are two basic methods:
1) Looking into the computer that is supposed to be accepting moves from the player, which can be effective if the player is an amateur and isn't aware that this can happen, (typically looking at what other programs are running and whether the player is giving the game full attention) and 2) Looking at the pattern and timing of moves. Classical chess has standard openings which are 10 or 20 moves long, so playing style is much less detectable. |
![]() |
![]() |
![]() |
#6 | |
Bamboozled!
"๐บ๐๐ท๐ท๐ญ"
May 2003
Down not across
265708 Posts |
![]() Quote:
Known as a "man in the middle attack" in the security literature. Paul |
|
![]() |
![]() |
![]() |
#7 |
Undefined
"The unspeakable one"
Jun 2006
My evil lair
2·32·7·53 Posts |
![]()
It doesn't sound like much fun to play chess like that. In fact it doesn't sound like the person is playing chess at all. Where is the sense in blindly copying moves from one computer to another? Might as well be counting blades of grass in the garden, at least you get to do that outside in the fresh air.
|
![]() |
![]() |
![]() |
#8 | |
"Brian"
Jul 2007
The Netherlands
2·11·149 Posts |
![]() Quote:
But in the arena where cheating happens on a grand scale, amongst amateurs on the online chess server, the motivation is usually the Rating. With a capital R. It's the measure of your performance as updated by the server after every game depending on your results and the ratings of your opponents, and it's what marks you out as a strong, average or weak player. Most players place an awful lot of importance on their Rating. If it's higher than it usually is then it looks very nice displayed there by their name for all to see. So there's plenty of incentive to win games by whatever means necessary. ![]() |
|
![]() |
![]() |
![]() |
#9 |
Undefined
"The unspeakable one"
Jun 2006
My evil lair
2×32×7×53 Posts |
![]()
So:
Rating is to Chess as post count is to forums. Sigh, some people just seem to miss the point completely. ![]() |
![]() |
![]() |
![]() |
#10 | |
"Ben"
Feb 2007
1110100100112 Posts |
![]() Quote:
If this forum had a post rating mechanism, then someone could go out and kidnap a mathematician, bring him printouts of some topic, force him to provide an intelligent response, and post it as his own, all in an effort to get a high Rating. Man in the middle again. Defeating this would be similar to defeating chess cheaters - you have to find a way to authenticate the post (or move). Authentication could come from analyzing responses in such a way as to identify the real poster - pose questions the answers to which only he would know, for instance. Last fiddled with by bsquared on 2011-06-10 at 21:13 |
|
![]() |
![]() |
![]() |
#11 | |
Undefined
"The unspeakable one"
Jun 2006
My evil lair
2×32×7×53 Posts |
![]() Quote:
My thought is if people really want to cheat like that then just let them at it. There appears to be no harm done to anyone. If you visit the site looking for a strong game then you get it, albeit against a strong computer. At least it saves you having to run the program yourself. More cycles for GIMPS if someone else is doing the grunt work for you. |
|
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Vote chess game 4: To be decided? Some chess variant will be interesting to consider with! | Raman | Chess | 6 | 2016-12-06 06:50 |
Reset the milestone counter: New SoB & PSP was found !! | Aillas | Prime Sierpinski Project | 4 | 2016-11-08 04:24 |
Visitor counter | Rodrigo | Forum Feedback | 9 | 2010-09-14 14:14 |
Counter-examples, please | Numbers | Miscellaneous Math | 4 | 2005-12-29 11:03 |
A Counter example, anyone? | devarajkandadai | Math | 27 | 2005-05-24 04:37 |