unforum

[mattyjh]

[James Siegesmund]

[BrianEnigma]

Okay, so I have started to revisit Shuffled. Waaaayyyyy back when (wave 1 or 2), I wrote a program that attempted to figure out a correct deck order, given the playing card symbols on the cards we knew. It attempted to find the various permutations of the remaining cards in the remaining prime numbered "holes."

After talking with a few people in IRC and revisiting the wiki, I am starting to believe that the deck started in a "standard" order. In this case, I am taking "standard" to mean not how the cards are stacked at the factory in any given deck, but the mathematical start of the Solitaire cipher (i.e. card 1 is in position 1, card 2 is in position 2, etc.) More specifically, and stepping into meta out-of-game territory for a second, it is very likely that the puzzle's designer went over to Schneier's site, grabbed his Solitaire Perl program, then ran his message through it with a password/phrase. Schneier's program does not have a particular starting deck--it starts with everything in order (as I said above, card 1 in position 1, etc.)

This got me thinking that all we need is a password/phrase, and the message is ours for the taking! A slightly inefficient shell script later, I was throwing the standard Unix dictionary at the encryption. I've attached my results, if you're interested. I scanned them for obvious crib words (djinn, virus, worm, perplex, sente, kiteway, claire, castille, etc.) but nothing intelligible came out. You're free to try yourself. What bothers me a bit is why some of the lines are missing a letter. There is a quirk of some kind in the decryptor that eats an extra letter on certain words. For example compare "./sol.pl -d abas message.txt" versus "./sol.pl -d abase message.txt" and you will see the former decode is one letter short.

Anyway, this line of reasoning (assuming it's a 1..52 deck order and sol.pl), leads to the password being not a standard word, but a non-dictionary word or a longer phrase. I tried the obvious ("if entropy...") as well as a few names (variations on Garnet's name, given that the message was written for him), etc. The mnemonic letters of "if entropy..." anagram out to "ice slowly," so I tried that, too.

At this point, I am not sure what else to try and whether I should drop this line of query for a little while.

Okay, not an attachment because it's 16M, but here's the output: http://perplexcitycardcatalog.com/notes/shuffled-wordlist-output.zip

[Mindez]

Ugh.

I just went through the whole thing and read every line with "The".

And I found nothing =(

Does your output include the outside letters, though?

Perhaps try again with/without the outside letters?

[themandotcom]

Bri, in your output, I did a control-f on sente. There is acutally one that says." Sentehi." Coincidnce? I don't know, just lettin' ya know.

[Sh1ft]

[BrianEnigma]

[PerplexHero]

[brokasaphasia]

[PerplexHero]

[brokasaphasia]

[arnezami]

@brokasaphasia: could you post the source of your program so we can look if there is a bug in it. Because solitaire should be perfectly reversable (if you have the resulting deckorder). Its also possible there is a bug in the original source of the decryption code.

Looks like a small problem though since your ending deck is very close to the original deck. Maybe when reversing you could also output the keystream/decrypted text and compare this later when you start decrypting forwards (in other words: outputting the backwards decryption).

Btw: I suspect step 5 may not have been reversed properly somewhere (because of a joker).

Thanks for all the work btw .

arnezami

PS. Just to be sure this should also be tried: use the deck order on the primary cards as end-deck and then reverse it (= do the backwards decryption). Keep in mind after reversing there could be some "logic" to the resulting (and originally starting) deck which we might have been able to guess without the primary cards. Which would mean we didn't need the primary cards...
PPS. Its also possible the reversing doesn't take 108 steps but 10 or 20 or 118 or 128. And maybe you have to decrypt the whole 128 text while only the middle 108 makes sense after decryption.
PPPS. What would finally explain the existence of the 10 characters at both ends would be this method: reverse the deck 128 steps (entropy wins because the deck is shuffled). Then decrypt only the 108 characters (outward looks should leave you cold). This way both 10-texts would effectively be additional security measures. And of course there may be other possibilities I haven't mentioned/thought of yet.

[brokasaphasia]

Actually, I believe that my code is correct.

I believe that it is IMPOSSIBLE for Solitaire to leave the deck perfectly ordered 1..54 at the end of any number of encryption steps:

If the deck ENDS UP like this:

[arnezami]

[RedZed333]

I know Von has told us 'as for the code, everything you need is there' and, according to Perplex City Customer Services, 'You do not need any other puzzle cards in order to solve card 243-shuffled.'

Do you think it's worth asking if these hints mean we can solve this card as it stands and we don't need knowledge of any 3rd party codes/ciphers ie solitaire.

If so, who could we ask?