Cheating at Chess

There's a fascinating story about a probable tournament chess cheat. No one knows how he does it; there's only the facts that 1) historically he's not nearly as good as his recent record, and 2) his moves correlate almost perfectly with one of best computer chess programs. The general question is how valid statistical evidence is when there is no other corroborating evidence.

It reminds me of this story of a marathon runner who arguably has figured out how to cheat undetectably.

Posted on January 16, 2013 at 6:25 AM • 63 Comments

Comments

whomeverJanuary 16, 2013 6:38 AM

if something is hard to do but has a payoff, and cheating is easier, people will cheat

Andy BarrattJanuary 16, 2013 6:50 AM

I'm not much of a chess player so this might be nonsense but...

Is it possible he's memorised a number of plays?

Otto J. MäkeläJanuary 16, 2013 7:03 AM

Chess-related frauds have a long history, starting from The Turk. To figure out how the fraud is perpetuated, I would suggest that chess organizers consult casino security – there are numerous cases of computers being used against blackjack.

RichardJanuary 16, 2013 7:12 AM

Somehow these stories remind me of the biggest cheater of them all: Lance Armstrong. I didn't watch his fall very closely, but wasn't it led by piles of circumstantial evidence?

Aaron WJanuary 16, 2013 7:13 AM

Unlike the marathon cheater, who has a long record of unlikely performances, this guy only has a single tournament, which makes it somewhat less of a mystery.

The accounts from the tournament say he took off his shirt and organizers searched his pants pockets.

He also performed badly in matches when his moves weren't being broadcast live over the internet.

Based on that, my guess would be a receiver in his shoes that used a morse code-like buzzing to recommend moves sent via text message.

It'll be much more of a mystery if he does it again, in a truly secure tournament. But I doubt he'd try it.

FigureitoutJanuary 16, 2013 7:29 AM

First the Lance Armstrong thing (I was actually kind of inspired and saw him ride in Paris..) and now anytime someone does something extraordinary they must be cheating. I remember my brother got around a 900-1000 rating as a kid and he was pretty good so a range of 3089–3258 is freakish.

FigureitoutJanuary 16, 2013 7:33 AM

@Aaron W
--It would actually be kind of cool if he really received morse; even cooler if someone keyed it with a light and a mirror/glass. I guess we should give all chess players cavity searches now before tournaments...

NameJanuary 16, 2013 7:45 AM

I guess the more fundamental issue is that the best (human) chess master hasn't been able to beat the best available chess software for about a decade now.

For me, at least, the game stopped being interesting in '97.

Mike BJanuary 16, 2013 7:46 AM

I have always been of the mind that once a game becomes solved algorithmically human society should essentially stop playing it. Sort of like Tic-Tac-Toe its an element of progress that we have solved certain games and never have to expend time or effort playing them because they have been solved in all cases. Granted Chess is not at this level yet, but computers have basically eclipsed human players so there is no use continuing to put human talent into the pipeline when alternatives such as GO are still dominated by humans.

abcJanuary 16, 2013 7:48 AM

@Andy Barratt All chess players have plenty of games memorized. It is part of their preparation.

BWJanuary 16, 2013 8:04 AM

He must transmit and receive signals.

Nothing you can not detect with a spectrum analyzer and some logic thinking.

Try playing a tournament in a shielded environment like an underground bunker and see if he still does well.

It would be funny to detect his method of communication and apply a replay attack or direct him to lose the game in the least possible sets.

Would make a nice video :)

Dr. I. Needtob AtheJanuary 16, 2013 8:41 AM

@wiredog (the first comment) - I suspect that many of these guys are not familiar with "The Eudaemonic Pie", where a hidden computer was used to win at roulette. (No WiFi was involved.)

When I first heard of it I was skeptical, but that was before I understood how the system actually worked.

Peter A.January 16, 2013 8:45 AM

The chess tournaments aren't going to be held in military-grade TEMPEST-shielded bunkers; nor the players are going to be strip- and cavity searched. It will kill the whole fun.

Barring that, it will be always possible to set up some covert channel to an accomplice running chess software.

Apparently some chess federations allow using PDAs or similar devices with "approved" software to record one's moves. I shouldn't have to explain to the readers of this blog what possibilities it creates. Even if these were banned, one could, for one example, record one's moves on paper with a rigged pen that transmits it's movements (I've read about a commercial product like that once). Often a covert input channel to the chess software is not needed at all as tournaments get transmitted live on various online chess services. For the "output" channel there are numerous possibilities for a miniature device receiving signals of any nature and transducing them somehow to player's senses - or even an accomplice among the audience giving coded signals.

The best approach would be to apply some simple countermeasures like banning electronic devices and observing players' behaviour and letting the really smart cheaters win. It's only a game, after all.

Dr. I. Needtob AtheJanuary 16, 2013 8:50 AM

They need to call in some experts to solve this one. I suggest Penn & Teller.

customaryJanuary 16, 2013 9:34 AM

@Andy Barratt: it is more than that.

I suspect that cheating is customary in chess (for real money); this guy cheated too boldly and raised suspicion.

The highest chess encounters have some level of isolation, but it is only to prevent each opponent from being disturbed by the public.

Chess need even more isolation. Faraday won't be enough (think of a microcomputer sensitive to ultrasound).

But in real world, Chess won't get that. Chess is already rigged by cheat. Like cycling. Like pretty any sport with money involved (Judo, ...). And no discipline recovered from there.

alanmJanuary 16, 2013 10:29 AM

Cheating at chess wouldn't need comms every move.

A master player (like this guy) would be fine in the book openings because he'd have his selection memorised. Computer advantage would be useful in the relatively option-dense and short-duration mid game, and even then a batch update of the next four or five moves from a given position (opponents responses could be predicted pretty accurately by the computer too) could be decisive.

If I were doing it (disclaimer: I've never played a chess tourney) I'd take a bathroom break a few moves after the book opening had played out and get a drop from someone who's done the analysis of the available (on the internet) data. Maybe even a printed page with branching options, a strong enough player should be able to memorise it fast.

Erik V. OlsonJanuary 16, 2013 11:15 AM

He must transmit and receive signals.

In many tournaments, just receive. The moves are often sent out in real time on the net. There is a proposal to insert a 15 minute delay on that, and one French tourney did so when a suspected cheat was playing.

Cheating at chess wouldn't need comms every move.

Indeed, Garry Kasparov and Vishy Anand proved this in 1996, when they were both ranked in the top five of the world by ELO. A game was in progress, at move 20, Kasparov paused for some time before making a move. He was fighting between two moves, chose one, and some four hours later, the match ended a draw

After the match, he immediately went to a commentator -- who had Fritz, the noted chess program running in front of him -- and asked "I couldn't win, could I?" To Kasparov's and Anand's shock, the answer was "Yes, you had a win with 20.g4!" This was the other move that Kasparov had been considering.

They stood stunned as the commentator started to read out the line, and two moves into that line -- they both saw Kasparov's mate.

Basically, what they realized is that even the very top players, if they can get just *one* move at the right time can turn matches.

Kasparov also used this as an argument for his "Advanced Chess", which combines a player and a computer. I'm not sure if I buy that argument, but it became very clear that if you could check one move late in the game, you could easily turn a match.

WaelJanuary 16, 2013 12:24 PM

@ Bruce Schneier,

The general question is how valid statistical evidence is when there is no other corroborating evidence.

wouldn't that qualify as *sniffle* statistically-profiling the guy based on his past performances? :)

I used to play chess a lot and wasted many years on it (books, programs - Shredder, Fritz, and Hiarcs- , on-line...) Gave it up after I could not beat computers anymore. I still play once in a while, but not as before, and I see that a huge improvement in ELO rating is suspect.

I use

derpJanuary 16, 2013 12:31 PM

They should call up James Randi and have him watch the crowd for signals, or watch the guy for slight of hand when he throws away his equipment while getting up to handshake after the game

alanmJanuary 16, 2013 12:52 PM

A game was in progress, at move 20

Spot on, right in the mid-game window. Chess programs shine at finding a way through the incredibly complex decision tree at that stage.

He could have intentionally complicated the positions as far as possible; some openings almost guarantee this. Then at the right time (maybe signaled by a friend who is monitoring the ongoing analysis), get a drop of the data and play the computer-picked line.

This guy was a strong player who didn't need hand holding, he just needed an edge.

customaryJanuary 16, 2013 2:11 PM

@derp: a move is an 10-bit information, and with half of this information you will still able to guess what the computer chose.

It is easy to send 5 (or 10) unnoticed bits for an individual in a crowd.

MarkHJanuary 16, 2013 3:20 PM

@Mike B., who wrote "I have always been of the mind that once a game becomes solved algorithmically human society should essentially stop playing it."

I sympathize with this perspective to some extent -- I can't bear to play Sudoku, or the board game Clue, because for me they are tedious exercises that cry out for computerized solution.

But sports (at the end of the day) are about pleasure, or -- if done publicly/commercially -- about entertainment. Most people know the fascination of watching someone succeed (or at least, come close) at a task that is extraordinarily hard.

There's certainly an awesome quality (for many of us) to what great chess players can do with their merely human apparatus.

Though it's as yet unproven, I think it quite likely that with a few years of development, autonomous cars could reliably defeat the best humans at (for example) Formula 1 auto racing -- even if their potential weight-distribution and aerodynamic advantages were nullified, or even if they operated under imposed delays to compensate for human reaction time. Algorithmic vehicle control has been very sophisticated for decades, and the state of the art only improves.

Does that make it less exciting to see the limits of what people can do?

russJanuary 16, 2013 4:03 PM

A while back I read a neat blog post about ways to cheat in a public chess tournament:

http://kenta.blogspot.com/2009/10/...

>>>

We consider one very minimal situation: the channel capacity or bandwidth from the computer to the player is a very small one bit per move, and the bandwidth from the player to the computer is zero (no communication possible in this direction, the player is under intense scrutiny, and any unusual behavior will be spotted). Actually, it's the covert bandwidth from the player to the computer that is zero; we assume the game is live and public, so the confederate, a member of the audience, can see the moves as they are made. Thus, the channel is such that the player cannot ask, "is this such-and-such a move a good move?" (because of the zero covert bandwidth in that direction), nor can a confederate signal "such-and-such a move is the best move in this position" (because such a message requires more than one bit per move). Nevertheless, one bit of information will be enough to provide a significant advantage to the cheater.

This "one bit of information" may be transmitted, for example, by the player seeing the confederate in the audience. It is helpful if the confederate situates him or herself where the player does not have turn his or her head and adjust his or her gaze. The confederate performs a subtle pre-agreed signal to transmit the bit: perhaps arms crossed left-over-right versus arms crossed right-over-left.

My proposal for the one bit that is transmitted is, "Did the opponent just blunder?"

MingoVJanuary 16, 2013 4:49 PM

The accused player was searched and did not have any apparent way to receive information: no eyeglasses, nothing in or behind his ears, no electronic devices on his person, etc. The most likely explanation for his improved playing (that resembles a computer program) is that he studied intensively using that program.

SomeoneJanuary 16, 2013 5:23 PM

If I'm reading this article correctly:

Chess Ratings:
2227 - Ivanov's established rating before the tournament
2400 - "international master" typical rating
2600 - "strong grandmaster" typical rating
2861 - world record rating for a human player
3089–3258 - Ivanov's performance during this tournament

I could sympathize with an "innocent until proven guilty" line, but calling intensive study the "most likely explanation" strikes me as ridiculous. All top chess players study intensively, but only one made a sudden leap from "pretty good" to "best in the world".

GGJanuary 16, 2013 5:35 PM

"undetectable" is far too generous a characterization of Kip Litton. Eyewitness evidence provided by actual top finishers is pretty compelling.

Petréa MitchellJanuary 16, 2013 5:46 PM

MingoV:

We don't have much information about when the search was performed (Was it the moment he stood up at the end of a match? A few minutes later? After a bathroom break?) and the most complete description of the search I can find says it wasn't a very thorough one. So there is hardly definitive information that he didn't have any suspect devices on his person. And as noted in other comments, if there were an audience in the room, that opens up more cheating methods that don't require him to have any special equipment.

DocTechnicalJanuary 16, 2013 5:51 PM

Has anyone checked his yogurt? He could be getting coded messages via that...

RHJanuary 16, 2013 5:54 PM

Personally, I'm fascinated that chess has advanced so greatly that we can now put a player's moves into a computer program, and out pops an estimated rating based on that play.

AC2January 16, 2013 11:05 PM

@ Mike B • January 16, 2013 7:46 AM

I have always been of the mind that once a game becomes solved algorithmically human society should essentially stop playing it

--------------------------------

Apart from the fun element pointed out by another poster above, that's as silly as saying that since we have calculators now we should stop learning how to add.

WaelJanuary 17, 2013 12:04 AM

@ russ

I will have to disagree with that "one bit" logic. Telling a player "your oponent blundered" may not be sufficient. Counter examples include chess puzzles where one is asked to find the "best move" or find a mate in a certain number of moves[1]. Would it help to tell us the opponent blundered before we solve the puzzle in this case? It's just one example. In the story at hand, revealing this one bit of information certainly will not raise one's ELO rating by a few hundred points.

[1] my favorite puzzle that no one I asked has solved is: White mates in *drum rolls* half a move :)

customaryJanuary 17, 2013 12:26 AM

@Wael:

Re-Read that from @Erik V. Olson: "Basically, what they realized is that even the very top players, if they can get just *one* move at the right time can turn matches."

At two bit per half-move, it is better to transmit information about a move starting 3 (predicted) moves in advance. That makes 12 bits of information to inform about that one precious information (minus the protocol with messages as "oops, this half-move was not the one predicted, flushing."

WaelJanuary 17, 2013 12:51 AM

@ customery

I did read it again. I am also familiar with that incident. I do agree with the quote, and it does not only apply to top players! On the other hand, do we have an overloaded meaning of "bit"? I interpreted "bit" as described by @russ to mean "binary digit"; a zero or a one.
I still maintain my opinion. How can you tell the player all this information in one bit? What am I missing?

Time for some definitions:
Half a move is actually known as a "ply". Which is one move by one player.

In the puzzle I talked about, I should have said half a ply.

WaelJanuary 17, 2013 1:02 AM

@ customery

Regarding

After the match, he immediately went to a commentator -- who had Fritz, the noted chess program running in front of him -- and asked "I couldn't win, could I?" To Kasparov's and Anand's shock, the answer was "Yes, you had a win with 20.g4!" This was the other move that Kasparov had been considering.

How would telling Kasparov "your opponent blundered" at the right moment help him choose between the two alternate moves he was considering? That is the question.

Do we need to define "blunder"?

customaryJanuary 17, 2013 5:07 AM

@russ was talking about 1 bit per move (half a bit per "ply").

I am talking about 2 bits (2^2=4 four possiblities, named 1, 2, 3 and 4) of information a each half-move (each "ply", but I don't understand the expression "half a ply" at end of your post).

In the case of Kasparov, the person with the "Fritz" programm could have predicted correctly the 6 moves before the critical point, and sent the information "g4" in six packets of 2 bit, one packet per half-move, like:

(Kasparof plays 17.Rh3) Juri Dokhoian transmits "4", which means that from the partition "a,b", "c,d", "e,f", "g,h", the information is in the fourth set "g,h".

(Anand plays 17...g6) Juri Dokhoian transmits "1", which means that from the set "g,h,reset transmission,your opponent blundered", the information is the first one, it is "g".

(Kasparof plays 18.Qd2) Juri Dokhoian transmits "2", which means that from the partition "1,2", "3,4", "5,6", "7,8", the information is in the second set "3,4".

(Anand plays 18...Nf5) Juri Dokhoian transmits "2", which means that from the set "3,4,reset transmission,your opponent blundered", the information is the second one, it is "4".

(Kasparof 19.Nxf6+) Juri Dokhoian transmits "3", which means that from the partition "King,Queen","Rooks,Knights","Bishops,pawns", "reset transmission" the information is in the third set "bishops,pawns"

(Anand plays 19...Bxf6) Juri Dokhoian transmits "2", which means that from the set "Bishops,Pawns,reset transmission,your opponent blundered" the information is the second, it is "Do g4 with your pawn now!".

In this protocol, "reset transmission" means that a move was not correclty predicted, so transmission is restarted from first step; "your opponent blundered" means "there is an unusual move of yours now to take a decisive advantage after this blunder, you may also use the sparse set of information that I just transmitted to you".

Note that with Kasparof, only giving him the column "g" would have explained to him that "20.Bd5" was not correct, which makes this much shorter transmission:

(Kasparof 19.Nxf6+) Juri Dokhoian transmits "4", which means that from the partition "a,b", "c,d", "e,f", "g,h", the information is in the fourth set "g,h".

(Anand plays 19...Bxf6) Juri Dokhoian transmits "1", which means that from the set "g,h,reset transmission,your opponent blundered", the information is the first one, it is "do something in the column g".

By the way, your link was malformed.

QJanuary 17, 2013 7:11 AM

> White mates in *drum rolls* half a move :)

You're asking how this could happen?

Something like a white pawn moving from the 7th to 8th rank uncovers the black King for checkmate, so the mate happens before the pawn gets promoted?

jbJanuary 17, 2013 10:37 AM

Wael,
Pawn has moved from 7th to 8th row. The 2nd half of the move, which hasn't happened yet, is White deciding what piece to promote the Pawn to.

Alternatively, White pawn has just made an En Passant capture, which will result in a checkmate via discovered check once the captured Black piece is removed, but has not removed the captured piece yet.

Rick AuricchioJanuary 17, 2013 10:42 AM

@ Wael:

"How can you tell the player all this information in one bit? What am I missing?"

The Morse-code-like system mentioned earlier allows one bit to express plenty of data based on timing. (Actually, we're talking about serial data communication.)

No OneJanuary 17, 2013 12:14 PM

@Wael: I looked up the solution to your puzzle and I think it's better presented in (approximately) your form than the general form.

I think the proper puzzle would be: Construct a position in which White will mate in one half ply.

POSSIBLE SPOILER FOLLOWS:

Strictly speaking, however, I don't think that the required move should be considered two half plys, but should be considered one ply entirely and that it is only by convention that we complete it as multiple actions.

SPOILER FINISHED

Bob TJanuary 17, 2013 2:36 PM

He probably statistically analyzed his competition as though they were the game rather than the game itself with a computer program. Kind of like Michael Larson who figured out how to win at Press Your Luck. No Whammy's!!

WaelJanuary 17, 2013 3:03 PM

@ No One

Think of "moves" that are composed of more than one piece moves. For example castling (hint hint, wink, wink). In castling, the king moves first, followed by the rook. The two "moves" are considered one move.

WaelJanuary 17, 2013 3:19 PM

If I were to cheat at chess, I would not depend on one accomplice. I could arrange to have a multiple of them giving me different signals each. Say you have 9 accomplices, one running a computer program and the rest act as an 8-bit communication signal to the player. each with a known one-bit signal. One touches his ear, another coughs, a third blows his nose, etc... Or all can use the same signal to reduce confusion at the price of easier detection. In this manner, you can tell the player exactly which move to make and his poor opponent will be playing essentially against a computer that obliterate him. Another camouflage tactic is to personalize the computer program. Most programs allow you to personalize them (Aggressive, Solid, Murphy, Capablanca, etc...). Their moves will be a little different than a non-modified program, but also often weaker. Would not be far fetched if a certain country wants to help it's "citizen" to win a tournament ;)

SomeoneJanuary 17, 2013 3:51 PM

Hypothetically, if the player is good enough to identify good moves but can't always tell the BEST move, and you can give him a 1-bit hint on each of his turns, I would consider "of the two best possible moves (or perhaps 'the best two pieces to move'), is the better one the one on the left or the one on the right?"


@Wael

If we're going to speak technically, I'm pretty sure there's no such thing as "half a ply". There are various moves in Chess that could be abstracted as containing multiple steps--castling, en passant, promotion, or even moving a rook several squares down a line. But the rules don't actually let you slice your ply into multiple parts and select them independently; all of the examples mentioned must be made as a single, entire move or they are illegal.

So if you want this puzzle of yours to be taken seriously, I think it's reasonable to demand that you rigorously define what YOU mean by this magical phrase that you made up.

One of my criteria for a good puzzle is that the person solving it should be able to check a potential solution on his own, without relying on the riddler to confirm or deny it. Your goal posts are currently too vague to meet that criteria. (Unless there's some commonly-accepted formal definition of "half a ply" that I'm just not aware of.)

No OneJanuary 17, 2013 4:19 PM

@Wael, yeah, like I said I looked up the solution to the puzzle, but I disagree with it. However, I think your presentation is better than the more common presentation of showing a board and challenging the viewer to find mate in X. I think forming the puzzle as create a board position where white mates in half a ply is a better presentation. However, I don't think castling counts as two actions that each make up half a ply in chess terminology, so it's a weird case anyway.

WaelJanuary 17, 2013 4:40 PM

@ Someone

I admit its a little vague without a board. At the same time, its supposed to be a bit amusing...
I did not make up any terms. When people are playing in a tournament for example, they list their moves as such:

White Black
1 E4 E6
2 D4 D5

Each line is called a move. When computers are evaluating the next step, they are looking at a "ply" at a time. for example, if I play "E4" what are the possibilities? Then If my opponent plays "E6" then what does the situation look like, what are the rest... "E4" is a ply, and "E6" is a ply in that sense.

Now as for the puzzle. If its White's turn, the question is:
Is it possible to mate black in half a ply? The puzzle itself is not normally presented as such. A board setup is shown, and the puzzle is: White mates in half a move (ply). The king has already moved to castle, but the rook is in its original place (the king move is not shown). Once the rook moves (the other half of the castling), it gives a checkmate to the opponent.

Regarding "blunder": If you signal a player that his opponent blundered (using a single bit), it could mean the following:

1- Opponent had a forced win, but "blundered" and now its a theoretical draw
2- Opponent had a theoretical drawn game, but "blundered" and now there is a forced mate, and his game is hopeless
3- Opponent had a won game but blundered and now its a lost game.

Thats why I say a single bit will not help much.

WaelJanuary 17, 2013 4:45 PM

@ No One,

Castling involves moving two pieces. You don't use both hands to move them. You use one hand to move the king, then you move the rook. Theoretically, these are two halves of one whole (whatever).

SomeoneJanuary 17, 2013 5:30 PM

I'm aware of what a "ply" is. But just because "E4" is a ply does not imply that "E" is half a ply.

"Half a ply" is not a real thing, and therefore is no more fair than a puzzle where I say "find the move that puts black into half-check" or "what can white do to half-win the game?"

You could probably use it in a play on words to create a joke, but what you are doing right now is like telling a joke and asking the listener to guess the punch-line (and even telling people they are wrong when they guess punch lines that are every bit as funny as yours but don't happen to be the exact one you thought up in advance).

WaelJanuary 17, 2013 6:28 PM

@ Someone

"E4" is a ply does not imply that "E" is half a ply...
Of course not!

But:

Suppose white castles, and is listed as:

17- Castles (o-o)

We describe the "Castles" as:
1- K-g1
2- R-F1

17 (Castles) = 1 + 2
"1" was already played (but is an illegal move, until the other half is done)
"2", finishes the castling move.
I understand if you disagree. I just wanted to share something I thought was "cute" with you, was not supposed to be a "challenge"... And I did not say other solutions are wrong, I said they were close. If you think they solve the "puzzle", then I am ok with that too.

mooJanuary 17, 2013 9:58 PM

There's no such thing as "half a ply", its in the same category as "a little bit pregnant": either she is, or she isn't.

Regarding 1-bit per ply hint, one use for it could be "the obvious move is the best / is NOT the best". That would allow him to play faster than normal on positions where it was the best, and spend the extra time on the subtle positions (and also not have to waste any time analysing the "obvious" move since the computer has already told him its not the best).

Another idea, perhaps a better one, is to assume the player can correctly pick the 2 most plausible moves (players at IM skill level should be able to pick the two best moves the same as the computer does a very high percentage of the time, maybe 98% or something) and use that bit to tell him which one is better (so in the 5% or 10% of cases where the player thinks one move A is better but the computer thinks the other move B is better, the player can go with B). With that system, the player doesn't need to spend much time analysing his potential move, he just needs to confirm that it isn't an obvious blunder (if it looks like it is, he probably chose the wrong 2 candidate moves).

mooJanuary 17, 2013 10:12 PM

@Wael: Either the castling move is an illegal move (in the current position), or it isn't. If you playing the castling move would leave you in check, its illegal. I believe also if any of the squares the king must move across is currently attacked, its illegal (not counting the square it started from; castling out of check is legal). If the move is illegal, you can't play it. Otherwise, playing it counts as a "half-move" (a ply) and is treated as an atomic operation. You're allowed to put your hand on a piece without moving it, and then change your mind. I think once you move the first piece in a castling move, you have to complete the castling move (I think technically you have to move the king first too). All of this kind of stuff is spelled out in great detail in the FIDE rules. There are procedures in there spelling out how to handle it if one player has made an illegal move, but I can't remember what they are. I don't actually play chess, so I might have got some detail wrong in this post. YMMV. :P

WaelJanuary 17, 2013 10:59 PM

@ moo,

You got the details right. Still changes nothing :)
You either get it, or you don't.

That's all I have to say about that -- Forest Gump.

JimFiveJanuary 18, 2013 9:23 AM

@russ
My proposal for the one bit question would be: Is there a clearly superior move in this position? If the opponent blundered the answer is yes, but in the Kasparov scenario the answer is also yes and gives enough information for him to keep looking until he finds clarity.

No OneJanuary 18, 2013 10:11 AM

@Wael, "Theoretically, these are two halves of one whole (whatever)."

I know how to castle and I know that it involves two pieces. However, I don't see it as two moves because of the simple fact that castling is a single move that affects two pieces, just like capturing is a single move affecting two pieces. Also, there's no rule requiring you to use one hand to castle.

Sauce: I did play tournament chess. Never very good back then, but I played.

mooJanuary 18, 2013 10:14 AM

@Scott, customary: Thanks, I knew I was probably wrong about a few details. That Touch-move rule is pretty interesting!

hughJanuary 22, 2013 9:41 AM


The marathon cheater – clones the RFID tag that he’s issued – accomplice(s) in the spectating crowd cross over the edge of the RFID timing mats at predetermined times?
He always gets the RFID tags in advance – plenty of time to clone.
The timing of his accomplices would be hard to get perfect for the split times, so they don’t bother too much – hence some of the spilt time discrepancies in the story.
He has a history of accomplices referenced in the story.
Studied engineering before dentistry, indicator of technical knowledge/aptitude – Probably uses RFID tags in workplace too .
Solution is very simple in nature – Occam.

Leave a comment

Allowed HTML: <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre>

Photo of Bruce Schneier by Per Ervland.

Schneier on Security is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc..