Register

Newcomer's thoughts on the 340

Moderators: Michael Butterfield, Aud8us

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Wed Sep 12, 2012 3:00 am

And another question for you, Oranchak. When running zkdecrypto analysis on a large amount of possibilities, did you find a way to automate it, or do you just load every file manually into the program? I am all about some experiments now, and find the lack of a proper tool chain really limiting. I currently feel like I might well have to bite the bullet and get to worker on a proper command-line only version of zkdecrypto, unless you have better suggestions from your experience.

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Thu Sep 13, 2012 6:49 am

Yeah, as you can see from the Playfair experiment, I had found that nice Playfair utility and wrote my own pivot-search. The algorithm is of course simple, I was wondering if there's something not immediately obvious that would be faster, but it turned out to be quite fast anyway, and it's sufficient to scan just 3 directions.

With a command-line version of ZKD, I'm afraid I am mostly interested in getting some good speed out of it. I think I'll get hacking at ZKD this evening, removing the GUI and Windows-specific code should not be outside the realm of feasibility.

Basically my thinking now is that certain theories about the 340 are more convenient to verify by getting a few thousand variations and seeing what happens. Or by doing experiments with other systems there, like the Playfair experiment I ran. If need be, I also have access to a decent computational cluster, though currently I do not imagine anything I'd want to do that might require that sort of resources.

Oh, and on a hunch yesterday I tried decrypting the 340 by reading it in two different spiral arrangements. Got the usual exciting but useless partial match with MOTHER and SLAIN close by and NOTYOU towards the end, but overall no better than the usual ;)

I take it, Oranchak, that you've also been satisfied for a while that the 340 is NOT a standard homophonic cipher constructed similarly to the 408?

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Thu Sep 13, 2012 8:25 am

Note this - the interesting patterns in the 408 are only clear after decryption. Spelling errors in the plaintext aren't particularly interesting - all of Zodiac's writing seems to contain them. The multiple assignment is noteworthy but also rare within the 408 and might have not even been intentional. None of this, and not the filler either, are visible if you only look at the ciphertext. In fact, if you look at the ciphertext, it's a pretty boring cipher. There is nothing completely strange that stands out.

The 340, on the other hand, has lots of those strange moments in the ciphertext. I'm quite sure it has filler (the ZODAIK signature has to be that), the pivots are statistically very unlikely, and the + symbols just make no sense. In addition, I keep getting bothered by the "look here" pattern, that is, how line 10, being in the middle, is visually highlighted with - symbols AND has a + in the middle, forming a visually striking -+- pattern. So in the 340, even the ciphertext itself provides plenty of interesting points, and I'm sure there'd be something else noteworthy to see if the thing actually gets decrypted.

Your idea of making test ciphers that are similar to the 340 in so many qualities is a very good one, but that actually sounds pretty hard to do, and like it might require quite some manual work.

Still it seems that the large amount of different symbols in the cipher, and the relatively smooth frequencies (the blasted + aside) indicate homophonic substitution expressly meant to conceal frequencies - but all the other oddities indicate the presence of some other step in the process. Once again the thought creeps up that the worst possible thing (concerning changes of decryption) would be if the author, being an amateur, did something "smart" that actually makes the cipher undecryptable.

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Thu Sep 13, 2012 4:53 pm

I offer congratulations to myself - just got the command-line version of ZKD up and running, and confirmed it works still by decoding the 408. Now I can sleep, and after a couple of small tweaks for better output, it should be ready for experiments.

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Fri Sep 14, 2012 4:26 am

Will probably attach it here. Just need to add some stop conditions to it. This is the ugliest thing I have ever done with any piece of software, I just took a chainsaw to it and ripped out the GUI. It does honestly need cleanup, but will do anyway.

The way it works now is that it simply takes a file name as a command line argument and gets to solving that cipher, a nice non-interactive way of working that should be perfect for scripting.

Oh, and the other thing I still need to fix with it is compatibility, I think those Windows specific code is still left in there. Anyway, I will post as soon as I can.

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Thu Sep 20, 2012 4:26 pm

So here is one more in the line of useless observations. I tried symmetrically flipping the cipher along the diagonal of a square. That is, for a cipher like

Code: Select all
ABCD
EFGH
IJKL
MNOP


the diagonal is AFKP, so exchange M with D, C with I, etc. Of course you can do similarly with respect to the other diagonal. To get the squares, I just used 17x17 squares, dismissing combinations of the first three and last three lines.

The result is that for all combinations (consecutive lines forming the square and both diagonals), flipping produces ciphertext that is statistically less inviting than the unmodified 340, with the ZKD score being notably lower. This is, I should say, well within expectations.

Level 5
User avatar
Posts: 1346
Joined: Wed Nov 11, 2009 1:47 pm
Location: Earth, mostly.

Re: Newcomer's thoughts on the 340

Postby smithy » Fri Sep 21, 2012 1:59 am

Duman - please stick with it! When I get away from the strange meander I'm on, looking at the letters, I'm hoping you'll still be around looking at the cipher; it needed some new energy. :clap:

Level 3
Posts: 64
Joined: Mon May 21, 2012 1:25 pm

Re: Newcomer's thoughts on the 340

Postby Duman » Fri Sep 21, 2012 12:52 pm

Thank you guys. Could be that it needs some new energy... could be that it needs another 40 years of frustration :D

As interesting as it has been studying the cipher so far, I am currently no closer to knowing what it says than on the day I first read about Zodiac's ciphers. I might, though, have learned something about ways in which it was NOT encrypted ;)

Level 1
Posts: 19
Joined: Wed Jul 24, 2013 3:39 am

Re: Newcomer's thoughts on the 340

Postby dagmajor » Sat Oct 15, 2016 6:42 pm

As basic and simple as Z408 was, I would guess Z340 is not that complicated as far as ciphers are concerned. There are a number of factors that make Z340 seem impossible to solve.

1) The mixture of symbols and letters hinders determining how the cipher was enciphered.
2) The large quantity of symbols (64?) versus the quantity of plaintext (26?) hinders frequency analysis.
3) The Zodiac Killer is probably not apt to make the same errors made in Z408 (double consonants LL, for example).
4) Too many smart people are trying to decipher it, and smart people naturally overthink the obvious, complicating the cipher themselves.

On the other hand, we can deduce certain aspects that may hold true for Z340.

1) It is not a homophonic substitution cipher (he failed with Z408, and needs to redeem himself as being "crack proof.")
2) It probably utilizes more than one "alphabet," as others have opined.
3) When it is eventually solved, we may all sit back and say, "Ah ha! That was a simple cipher," much the same way everyone did with Z408.

I am currently using brute force to attack it as a Vigenere' variation, in that rather than have a 26x26 graph of letters, I would have a 26x26 graph of letters and characters. A Vigenere' seems more within the abilities of the Zodiac Killer, gauging from his Z408. I do not think he has the patience or intelligence for anything more elaborate (if I analyzed his letters correctly.) If anyone has tried the Vigenere' in the past, I am curious to see what they have accomplished or ruled out so I am not kicking the same rocks. Would I be able to post any progress here for critique? I take criticism well since I retired, and unlike some people that continue to foster theories that have long ago been debunked, I prefer to move on with the next approach when the consensus dictates it.
Living IN America does not make you an American; Living FOR America does.

Previous

Return to November 8, 1969: The 340-symbol Cipher

Who is online

Users browsing this forum: No registered users and 2 guests