unforum

[ReeKorl]

OK, Here goes my first post after registering, and I think it's going to be a long one.

The first thing that struck me when looking over this puzzle was that we don't have the required KG to make the keystream. Without this, the passcode is next to useless. MC have stated we don't need any info other than on the card, so it must be solvable in it's own right without the keyed deck.

Now this got me thinking about weaknessess in stream ciphers of this kind. There's the two sets of ten letters at either end, possibly hinting at two messages of equal length? Also the chance that the middle text is two messages tacked onto each other with the same key used on each?

Even without the key, if there are two messages in there which have been encoded using the same key, then we don't need the key at all, we've got two messages, both Modulo 26'd with the same 'mask', so subtracting one from the other will result in the two plaintext streams Modulo 26'd to each other.

Here's an example in binary:

Take one plaintext and apply the XOR mask to it-
01011010101 Plaintext 1
10111110000 XOR Keystream
11100100101 ciphertext 1

Now take the other plaintext and again assign the XOR mask to it-
11100101010 Plaintext 2
10111110000 XOR Keystream
01011011010 ciphtertext 2

Now we get to our situation here, with two ciphertext streams-
11100100101 ciphertext 1
01011011010 XOR ciphtertext 2
10111111111 result of ciphertexts

If we know one plaintext stream, XOR this with the result-
10111111111 result
01011010101 XOR Plaintext 1
11100101010 plaintext 2

and we get the other plaintext stream.

Of course, in binary this requires knowing one of the plaintexts to get the other, and the same is true for regular letters for base-26 instead of base-2, but we can take guesses at what one message (or part of one message) says and if it comes out with intelligible text on the other message, there's a high chance it's right. After all, we're more likely to pick out when a message is right in english than we are in binary!

As a starter for 10, the two 10 letter streams are:

EODMF XRUTH ciphertext 1
FHURG IFULP XOR (or subtract in Mod-26's case) ciphertext 2
ASQEA KNZRH plaintext 1 and 2 Modulo 26'd

Can anyone with a bit more knowledge of the backstory of PXC (I've got here a bit late and am sadly lacking in the history department) try running some names or other details through that and see if we get anything intelligible out of the second stream? Also, try the same with the first and second halves of the central stream of letters and see if we can get anything out of that, perhaps the phrase 'djinnworm' may crop up somewhere? I may be off on completely the wrong track here, but it's something else to stick into the melting pot at least.

phew..... yep, I was right - it was a long one.

I need more coffee.....

[lillyplop]

Hello all

I have had a thought, I dont get many, and most of them mean nothing, but I thought I would come here and jot it down.

Forgive me if its mentioned elsewhere

Has anyone tried encrypting the word SHUFFLED into something like the Enigma cypher. Then used what comes out to decodes the larger set of words.

I know I am not explaining this right but its a bit like those games you get in Word Puzzle Books. They give you empty blocks and then they tell you that all 1's are N and all 4's are C's. Then you have to work your way round the rest until you get the answer.

If SHUFFLED were to come out as

S=E
H=L
U=B
F=T
F=T
L=P
E=M
D=I

We then look at the code to see that all the E's should be S's, all the L's should be H's etc. Then try and work out waht the sentance is trying to say.

Like with lyric from the Band Etymology card. Unless you had all the answers to all the questions you would have missing letters but you could figure the words out with a bit of thought.

Maybe I am talking rubbish but it kinds makes sense to me, but then I am in a little white room wearing a jacket that ties up and theback and no one here will give me a pencil.

[lillyplop]

Ooooooh I am on a role here with ideas!

What about a Tranposition Cipher? Where you SHUFFLE a key word as per this website

http://library.thinkquest.org/04oct/00451/trancipher.htm

Maybe if we shuffled SHUFFLED?

Mmmm the more I think about it the more it sounds like a hell of a lot of work for my teensy tiny brain!

[chichiri]

That would produce HSFULFDE, I really don't think it will be as simple as this but will try anything at this point.

no da

[WolverineFan]

[BrianEnigma]

[GinGenie]

I'll play devils advocate then....

As was posted upthread, the design on the back of the card is from the Bicycle deck made by the American Playing Card Company, though its been altered slightly, putting actual bicycles into the design, so I'd guess it was a deliberate choice.

This edition was re-released as a Jubilee 2000 edition, with pictures and captions instead of court cards (which gives us many options for keywords if anyone has this deck). The original edition was called the Bicycle 808 deck.

I figured that using 808 as a caesar shift pattern (first letter + 8, second letter + 0, third letter + 8 etc) would just add another layer of complexity to the solitaire cipher, but it could be used to disguise letter distribution in something like a transposition cipher.

[doublecross]

I still think that despite what MC say, the order of the deck is the order of the playing cards appearing on the puzzle cards, although it is a little surprising that this puzzle was not chosen to be one of them. That ordering is not used to solve the big letters and so is still available.

It is too much of a coincidence that there are 54 prime numbers between 1 and 256, all the prime numbered cards (and only the prime numbered cards) have playing card symbols on them, and the solitaire cipher uses a deck of 54 cards. I suppose confirmation will occur when we find two jokers on the puzzle cards, one with a star.

[oliverkeers13]

Von's hint:

[locqust]

I know earlier in the thread people were talking about Djinn as a possible (passcode/key?) to use. However another name for Djinn is a Genie, could this be used as a passcode? To be honest most of cryptography goes right over my head, should really read those books my dad has lying around about it.
So anyway any chance Genie could work?

EDIT: Dunno if this will help anyone but before reading this thread I did play around myself with the card ie heating etc (found the letters btw) I did freeze it 1st and got a yellow smudge or something going across the bottom right corner of the deck and into the last few letters of the code. The card is warm now and the yellow smudge is gone, I reckon it could be some really small writing but I didnt have my magnifying glass handy at the time.

[locqust]

Ok after a very quick lesson in cryptography off my dad and having a play around with the solitaire cypher system, I think I understand how this card works now! Now we both sat looking at this card and ignored what has been said here (he doesnt play the game so I didnt want to cloud his thinking with other ideas that have been suggested)

Basically we came up with this: (most of which has been already said on here before)

The passcode is a cryptic clue from either the entropy quote or from the joker and star. (Im thinking the joker and star and the entropy quote was for the heat ink bit, otherwise why mention it?)

The keystream needed is the white heat sensitive letters.

The cyphertext is the 100 and somat odd characters not affected by the heat ink. (remember it is a message to help us work out who made the virus, not the actual name, which makes more sense for such a long text)

Now your probably gonna ask why would the entropy quote be related to the heat ink if the guy at the poker club told Garnet? Well it isnt mentioned anywhere on the card that the BOX wasnt marked and that Garnet must of got the cyphertext himself somehow to work it out, which means that the picture we have is probably supposed to be the same box. Along with the code and heat ink. (no idea about the made in china bit!)

Ok so now we're going through what a star associated with a joker could mean and whether the design does have anything to do with it or not! Oh anybody else notice that the top left motorbike on the design is erased somewhat?

phew! you know I think this has been my longest post yet and I don't really think at the end of it this has helped at all!

[Grizy]

Hello fellow pxc players.
I'll add some of my thoughts to this lengthy thread. Here goes

Solitaire has been mentioned. If it were a solitaire cipher should all the word blocks be in groups of 5? There are 108 characters not including the ones that disappear.
Also the pack appears to be unopened. You can see the tear open strip towards the top of the pack. If it is solitaire, the cards in the pack possibly wont give Garnet (i assume it is him?) the key.

All the info for solving the puzzle is supposed to be on the card (?), so I assume they don't expect us to use brute force.

The person passing the info to Garnet must be sure that he can decrypt it so maybe investigating info on Garnet may give a clue?

There looks to be a strong link to the solitaire cipher, but could it be a red herring sent to give us hour of fun/stress?

Cheers for listening

G

[doomsdayred]

I've just stumbled on a variant of the solitaire cipher called the One-time Pad http://en.wikipedia.org/wiki/One-time_pad

It looks as though I may have a match for the code on the pack but this is supposedly an unbreakable code.

Can anyone work with this?

[Langley Moor]

Sorry, but I don't actually think it is a match - it's much closer to a Solitaire than a Vernam cipher, especially since we clearly have two separate pieces of information - the hidden text and the visible. Also, the pack of cards is way too much of a hint. However, in the spirit of investigating every possible solution, I'll decode this message using a Vernam cipher and splitting it into two rows (assuming one row is plaintext and the other the key). Back in a bit with an answer!

EDIT: Right, first attempt at a Vernam cipher was with the heat sensitive bits of text removed, and the first twenty characters it came back with was:
ztuhi qgzlu yhvzg zedwr
Which, I think you'll agree, isn't promising.
Am now going to try switching to make the top one the code and the bottom one the plaintext, see how that works, then try both methods again with the heat sensitive ink included.

Result of switching code and plaintext:
bhgts kubpg ctfbu bwxej

With heat ink included, I did the first 20 letters to make sure the first 10 weren't dummy characters to throw it off.
Results of first attempt with line one as plaintext and line two as the code:
fgwhl growa iluvk kgjzh

And with the lines switched:
tuetp ujmea rpgfq qurbt

So unfortunately no luck Anyone want to check my working? Seems easy enough but just produced gibberish, whichever way round it was tried.
Ok, redid the first bit again just to double check and got exactly the same. Also did another ten characters to check it wasn't dummy characters, but no luck - all my results are above. Sorry guys!

[thrush]

A little spec on the possible passcode.

The bicycle motif on the cards: Could this be a reference to how Lawrence Waterhouse and Alan Turing described modulo math through a broken bike chain in Cryptonomicon? Could this be a hint towards a passcode?