mersenneforum.org  

Go Back   mersenneforum.org > Extra Stuff > Hobbies > Chess

Reply
 
Thread Tools
Old 2011-06-10, 15:27   #1
Brian-E
 
Brian-E's Avatar
 
"Brian"
Jul 2007
The Netherlands

2×23×71 Posts
Default Cheating at chess how to counter it

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.
Brian-E is offline   Reply With Quote
Old 2011-06-10, 18:01   #2
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2·3·17·89 Posts
Default

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 opponent’s 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 ‘invented’ 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
Batalov is offline   Reply With Quote
Old 2011-06-10, 18:01   #3
LaurV
Romulan Interpreter
 
LaurV's Avatar
 
Jun 2011
Thailand

5×11×157 Posts
Default

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)
LaurV is online now   Reply With Quote
Old 2011-06-10, 18:08   #4
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

2·3·17·89 Posts
Default

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.)
Batalov is offline   Reply With Quote
Old 2011-06-10, 19:21   #5
Christenson
 
Christenson's Avatar
 
Dec 2010
Monticello

5×359 Posts
Default

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.
Christenson is offline   Reply With Quote
Old 2011-06-10, 20:00   #6
xilman
Bamboozled!
 
xilman's Avatar
 
May 2003
Down not across

100111101010002 Posts
Default

Quote:
Originally Posted by LaurV View Post
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. ...
And trivial to defeat. Run two computers to play the same game; use a human to copy moves between the two. The on-line machine never loses focus.

Known as a "man in the middle attack" in the security literature.

Paul
xilman is offline   Reply With Quote
Old 2011-06-10, 20:17   #7
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

559910 Posts
Default

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.
retina is online now   Reply With Quote
Old 2011-06-10, 20:41   #8
Brian-E
 
Brian-E's Avatar
 
"Brian"
Jul 2007
The Netherlands

2×23×71 Posts
Default

Quote:
Originally Posted by retina View Post
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.
Successful cheating in top tournaments and matches can be financially rewarding to say nothing of the prestige of earning a master title of any kind.

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.
Brian-E is offline   Reply With Quote
Old 2011-06-10, 20:54   #9
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

11·509 Posts
Default

So:

Rating is to Chess as post count is to forums. Sigh, some people just seem to miss the point completely.
retina is online now   Reply With Quote
Old 2011-06-10, 21:10   #10
bsquared
 
bsquared's Avatar
 
"Ben"
Feb 2007

3,271 Posts
Default

Quote:
Originally Posted by retina View Post
So:

Rating is to Chess as post count is to forums. Sigh, some people just seem to miss the point completely.
Well... a high Rating implies quality in the "posts" (moves). A high post count doesn't.

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
bsquared is offline   Reply With Quote
Old 2011-06-10, 21:23   #11
retina
Undefined
 
retina's Avatar
 
"The unspeakable one"
Jun 2006
My evil lair

10101110111112 Posts
Default

Quote:
Originally Posted by bsquared View Post
Well... a high Rating implies quality in the "posts" (moves). A high post count doesn't.

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.
Okay, makes sense. But you are focussing upon the method. I was more meaning the reasoning. Some people just seem to be concentrating upon getting some arbitrary number as high as they can, in whatever way they can. Perhaps my analogy was flawed since the numbers mean different things in Chess and forums, but I think the reasoning is still strongly correlated. Of course the reasoning itself is highly suspect as to the utility of it. Both a high online chess rating and a high forum post count mean absolutely zero in any other situation outside of the domain where the number is computed.

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.
retina is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
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

All times are UTC. The time now is 14:04.

Tue Aug 4 14:04:06 UTC 2020 up 18 days, 9:50, 1 user, load averages: 1.53, 1.73, 1.83

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2020, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.