BDHOLLAND-

What you have described is the very messy process the team went through to obtain the solution. Solving ANY cipher is ALWAYS a tricky problem, with dead-ends, false leads, and etc.

Oh I understand your point it would be messy. I have always held the position the Zodiac is an amateur and doing something strange with the codes after the Z408 was solved so quickly and is why I liked Bauer’s solution.

However while solving ciphers may be tricky the standard for a cipher solve in the Zodiac case wasn’t that tricky at all. The Hardens did it by using a combination of forensic psychology with cryptology to crack the 408.

It turns out is was a known method of substitution that was actually simple. That is why the FBI were able to evaluate it’s authenticity quite quickly.

This is the sort of expectation many of us were led to believe would be what is required for a cipher pass mainly because Bauer’s solution didn’t get a pass and he is Editor-in-chief of the international journal Cryptologia. He is basically a peer who does reviews for academia. Many of the arguments against his solution are present online for everyone to read through. The arguments are quite a standard set for ciphers. Known cipher methods have those standards. Bauer’s was a custom job that had to break such standards to get a result. He had to explain things in many ways that were not considered legit cipher explanations. For example, going outside of the cipher to other evidence to explain why he did this and that. Invoking Zodiac mistakes to explain why he needed to shift things that he did. However he got a key also. You just end up omitting a section of the cipher.

This section omitted BTW contained Oranchak’s own pivot finding which indicated the Zodiac had just copied meaningless patterns to fill up the cipher. It seems in this new solution that pivot finding is no longer incorporated into it. It seems we now have to consider that just a coincidental anomaly.

That is why when I saw Oranchak’s solution on YouTube (seems publishing on arXiv or plus.org isn’t the first choice here and History channel a little bit of a long way off to do it that way) followed some of these customized methods Bauer used so I was very surprised that all of a sudden some of those methods now get a pass. It seems whatever works, works. Which, BTW, is the sort of thing many of us argued would be congruent with the Zodiac but this was often argued against as a solution because of stricter validation rules that would cause problems for those types of solutions.

Consider for a moment that the 340 was a code between two people. The person on the receiving end with the key not only has to rearrange the code (fine) but fix mistakes (not so fine) and then complain to management that the person making these codes out really needs to get themselves together and not make those mistakes and to come up with something a little clearer next time without needing so many fixes and subjectively interpreting some of the results into words that can fit into the passages to make it legible (not so fine). However in the end the person on the receiving end was still able to get it to work so it wasn’t a totally failed cipher in this respect. That is my best analogy in all of this.

I look forward to reading the responses in published academia as those tend to be more robust than these ‘early days’ as you mention. However the FBI was quite quick to say it was correct. So either they have trust in Oranchak’s methodology or the Zodiac sent them the solution and they have held it back until someone made public the same solution. I don’t think this method is something subject to verification through null hypothesis testing hence the appeals to gut feelings (subjective) it is the right one in addition to the rules and massaging done to get it to work (mostly objective).

I think it would be pretty cool for cryptographers to try to falsify the solution. It would be a very good academic exercise.

The reason the FBI crypto unit verified it so quickly is they have a lot of experience with pen-and-paper ciphers. They saw this solve as being consistent with other transpositions they’ve dealt with. They didn’t need to do the math from scratch to prove it – they just knew from their experience.

In my followup video I’m going to show how we think the original message was encoded into the cipher. The process itself is actually pretty simple. I think it seems more complicated when we have to reverse the process to discover the message. But encoding the message is pretty straightforward.

I hope I never suggested that subjectivity during the analysis process is always bad. If I did then I was wrong. I’ve tried to emphasize where it is dangerous when it not accompanied by other solid evidence. In the case of the 340, we were building on solid evidence and some subjectivity helped guide the solution. But there is a big difference between using some intuition with a firm foundation, and using some intuition when there is no foundation at all. In the latter case, you can only hope that you found the right foundation completely by accident, and then hope that you can start to make it solid.

No one is saying you can’t continue on your quest to force a straight homophonic substitution into the 340. On the contrary, it’s been quite entertaining watching you try for the last seven years. There is no conspiracy here. They have solved the cipher. Steps have been pretty clearly outlined. Maybe work on the other two ciphers.

In addition to the arguments raised, it had previously been verified and widely debated that the letters corresponding to alphabetical symbols in cipher no 408 were basically adjacent to letters on typewriter keyboards.
http://www.zodiackillersite.com/viewtopic.php?f=81&t=2419&hilit=qwerty
http://www.zodiackillersite.com/viewtopic.php?f=81&t=5031&hilit=typewriter
I noticed the same phenomenon in 340.

So, how could a solution be "forced" by taking the same path ?? !!
Marcelo

I was wondering now, if this keyboard pattern was used in the two long cipher 340,408, it may have been used in the short cipher.(z13 & 32)
I have a lot of work to close the year, I am at work in home office, I will try on the holidays that are approaching or on weekends.

I wanted to do my own confimation and try to understand the process so i made simple little solver in python using the key from doranchaks’ video.

What i learned was you only have to do a single transposition operation.
And that part3 ( the last 2 lines are Un-transposed , which is a bit odd, It could be by doing that z provded a kind of homophone-key verifier.

Here is a pastebin of the program ( don’t flame my coding skills too much )
https://pastebin.com/ZpbBWfmU

Its a simple as i could make it, with no attempt to do the corrections for part2 of the cipher, so part2 output is messy but pt1 and pt3 come out nicely.
Some of my cleartext letters are different from the video..dont know why.
Excluding the ‘Life is’ part there are at least 14 errors ( some are probably mine though )

Sample output:

z340 part1
ihopeyouarehaving
lotsoffanintrying
tocatchmethatwasn
tmeonthetvshowwhi
chbringoupapointa
boutmeiamnotafrai
dofthegaschamberb
ecaaseitwillsendm
etoparadlceallthe
-----------------
z340 part 2
soohenbecauseeoow
haveensughslavert
oworvfovmewherees
eryoneelhehasnoth
ingwhentheyreachp
aradicesotreyalre
afaaidiofdetthifa
mnoeafreaidbncaui
seivyowtshatmrnew
-----------------
z340 part3
efilwillebnaeasye
noniecidarapdeath
--END--

Cool program, letega! Thanks for sharing that and providing confirmation.

Some thoughts regarding Doranchak’s (team) solution:

At least for me it was necessary to wait for some extra illustration to understand how the proclaimed solution should be extracted from Z’s gibberish. Why so difficult to write letters above symbols and draw a line onto it? I don’t know; I felt this to be a bit reckless leaving everybody else behind with only partial informaton. Still not having received a response on that one – unlucky.

But I’m not a pussy so let’s have a look at it:

SECTION 1
– errors –
FAN (FUN)
BRINGO (BRINGS)
BECAASE (BECAUSE)
PARADLCE (PARADICE)

– main words –
HOPE
HAVING
TRYING
CATCH
WASNT
SHOW
WHICH
BRING
POINT
ABOUT
AFRAID
WILL
SEND

SECTION2
– errors –
SOOHER (SOONER)
WORV (WORK)
VNOW (KNOW)

– main words –
BECAUSE
HAVE
ENOUGH
SLAVES
WHERE
EVERYONE
ELSE
NOTHING
WHEN
THEY
REACH
PARADICE
THEY
AFRAID
DEATH
AFRAID
BECAUSE
THAT
PARADICE

SECTION3
– anagrams –

– main words –
LIFE
WILL
EASY
PARADICE

SECTIION4
– 13 letter cipher –

DREA_A_O_LEDO
(negative?)

PRO’s
– cleartext makes ‘sense’, logical sentences
– same method used on section 1 and 2 delivers cleartext (‘confirmation’)
– multiple words longer than 5 letters existing
– cipher is solved completely instead of partial only (340)

CON’s
– total of 7 errors (not a big deal, one even repeating as a ‘signature’: K vs. V)
– cipher text was cut in three different sections
– cipher text interrupted (‘LIFEIS’)
– one line is shifted
– different method in section3
– negative 13-letter-cipher result

Please note that without

– cutting the cipher into three sections
– interrupting the cipher with LIFEIS
– shifting one line to the right
– using anagramming in section 3

the proposed solution method (‘one down two to the right‘) would not work. Opposite, repeating the same method on section2 – which they actually did – has lead to longer words as some kind of definite confirmation. But is this confirmation ‘definite’?

Now, if you look into section1 trying to ‘put in for example one sentence by purpose, you will find this to be pretty much easy the case (my first try worked over the first 4-5 diagonal lines, without using any computer). Therefore, the question coming up is the following:

Is it possible to get such a ‘complete’ solution by modifying both, cleartext as well as the method of decryption itself?

For example by trial and error ‘step-by-step’ one word after another and using moreless short words with e.g. rather frequent letters?

We know that the newly used program has allowed to ‘enter’ cleartext by guessing, which is fully ok. But what if the rest of the cipher has been completed this way? And exactly this had lead to the complex method of decryption? Possible – but first of all unlikely the case. The cleartext is indeed very ‘strong’, on first appearance, by being confirmed with longer words found even in section2.

But there exist many words…would it be thinkable to create such a cleartext into the cipher? This certainly would not be easy, after all. Obviously, they found that one:

IHOPEYOUAREHAVINGLOTSOFFANINTRYINGTOCATCHMETHATWASNTMEONTHETVSHOWWHICHBRINGOUPAPOINTABOUTMEIAMNOTAFRAIDOFTHEGASCHAMBERBECAASEITWILLSENDMETOPARADLCEALLTHESOOHERBECAUSEINOWHAVEENOUGHSLAVESTOWORVFORMEWHEREEVERYONEELSEHASNOTHINGWHENTHEYREACHPARADICESOTHEYAREAFRAIDOFDEATHIAMNOTAFRAIDBECAUSEIVNOWTHATMYNEWLIFEWILLBEANEASYONEINPARADICEDEATH

SO LET’S CHECK IT!

In an overall frequency analysis, this text has a pretty common letter distribution. Approximately 70% are covered by the first ten letters (EAOTHINRSL). An example text of Mark Twain comes up with 72.9%. This, on a first look, this seems to be plausible.

In an average Mark Twain text I was able to count 115 words of length five or longer out of 298 words (38.6%).

In the proposed cleartext a total of 23 of such long words exist out of 89 (25.8%). This is significantly LESS. That could have happened either accidentially or due to the lack of Z’s language style. And true: In his Z408, Z has indeed only used 24 longer words out of 100; thus this fully complies with the theory Mark Twain simply having better language skills than the Zodiac. To no surprise.

And what about the short words? 45 words of length 3 or shorter cover a total of 100 letters or roughly 30% of the cipher. Easy to place / use as fillers but could be ok.

The frequency analysis of each letter, however, deserves to have a closer look at:

In the proposed solution, the following letters are significantly less frequent than actually expected (e.g. Scott Bryce for the English language): The letters T, I, N, S, R, H and L are all (!!!) less frequent than actually expected. Instead, the vowels E, A and O are way more frequent; this making it ‘easier’ to fill up structures with random text. In fact, the letter ‘A’ even occurrs +42.6% more often than expected. This certainly is a significant deviance. At the same time, the letter ‘S’ appears 28% less than expected.

To find the letter ‘A’ more than 42% more frequent than expected, is surprising.

So why not having a look at the double letters, too?

Eliminating double letters from words connected as well as spelling errors (‘BECAASE’), I found a total of three double letters (WILL, ALL, WILL). In the Z408, Z had used the following double letter words: KILLING, KILLING, ALL, KILL, THRILLING, BETTER, OFF, WILL, ALL, KILLED, WILL, WILL, WILL, COLLECTING. Therefore 14 times or equally 11.7 times to be somehow expected in a length of 340 letters. Instead: Only three double letters. Also, no other double letter occurs in the proposed solution. This discrepancy, from almost 12 to only three double letters, is also significant (-74%) – but could still be accidential.

Finally, a look at the CONTENT of the message:

Most of it consists of very, very common words, such as TRYING, HAVING, WILL, BECAUSE etc.

The even more expressive word sections – regarding its content – would be TRYINGTOCATCHME, TVSHOW, GASCHAMBER, ENOUGHSLAVES. That’s it, moreless.

Thus, according to the proposed solution, Z had used only four (!) phreases with inner content instead but only commonly used words (e.g. BECAUSE). Thus, let’s compare this with the Z408:

KILLINGPEOPLE, WILDGAMEINTHEFOREST, DANGEROUSANIMAL, THRILLINGEXPERIENCE, ROCKSOFFWITHAGIRL, REBORNINPARADISE, BECOMEMYSLAVES, MYNAME, SLOWDOWN, COLLECTINGOFSLAVES, AFTERLIFE.

It can be seen that Z had used lots of ‘color’ in his language in the Z408 but shall not have done so in the Z340? He did so by using pretty much specific words, such as the word REBORN, and even combining those to to what I call ‘language’ instead of ‘words’ (WILDGAME, THRILLINGEXPERIENCE). The proposed solution, however, lacks of such (four instead of expected 9.75 such phrases -41%), which I’d say is significant, too.

It should be mentioned that the terms TRYINGTOCATCHME, TVSHOW and GASCHAMBER indeed exist but actually occur in the first part of the cipher. Thus, those could very well be self-constructed (not assumingly by purpose but any intent of a cryptanalyst to set up some text). In the second part of the cipher, ENOUGHSLAVES is actually the only one, with still the very common (lots of vowel-containg) word ENOUGH.

Ok, so having a look at WHICH words have been used in the proposed solution: It strikes immediately that many words have already occurred in previous Z communications. I didn’t check all of them but would go for BECAUSE, SLAVES, PARADICE, CATCHME, ABOUT, DEATH, LOT(S), THEY, SLOW, KNOW. Therefore, plenty of such previously used words show up in the solution, too. This indicating that those kight have been ‘forced’ in e.g. by gribs or computation; consciously or unconsciously. Not very encouraging.

Instead, other things are lacking: In the proposed solution, it not at all is referred to e.g. a specific crime he had committed or similar (except TVSHOW in the first part). He also did not ‘tease’ with regard to police or any other individual, e.g. PINESPLINTERSDRIVENUNDERTHEIRNAILS or ISHOTAMAN or comparable. Mostly, however, the content even lacks of aggression – it calmfully is mentioned in a sense of "It wasn’t me on the TV show and now I have collected enough slaves, so that’s it". No reference to killing, blaming, committed crimes etc, not even a threat – which would be very very untypical for Z.

After all, my thoughts regarding the proposed solution. It could be correct, and I have no need to deny that. Because method and cleartext are obviously there if you put things in the right order; even with changing the method to anagramming in the end. Thus my highest respect to Doranchak and his team to find such one.

The only thing is:
I believe the proposed solution is neither the correct one nor the only solution. Considering the polyalphabetic structure of the cipher, this is obvious. I’d like to give an example: It may be easy to replace the word ME with MY, if the second homophone occurs e.g. only twice or once; considering 7 spelling errors, too. You end up with a different cleartext which also matches the cipher. Especially if you transpose the cipher, cut it into three parts, add a LIFEIS section, move one line to the right – it might become true to find a ‘hill’ solution, matching the structure but still being far away from the cleartext. Especially in times of computation and when you have entered e.g. 40% of the future text in advance. Although this could be hard to do, I think it is still possible (‘forced solution’).

There is Pro’s and Con’s for this one and I will not be stupid to put my doubts on what has been published and celebrated as ‘the’ solution thus broadly already. However, my doubts are there and shall be expressed why they are there:

1) Significant ‘restructuring’ of the original cipher text; including 7 spelling errors, skipping sections, shifting lines, changing decryption method etc.
2) Above-average use of vowels (especially the letter ‘A’ +42.6%) indicating a constructed instead of original cleartext (which – on average – would contain +28% letters ‘S’)
3) Overwhelming percentage of common words with little meaning; most of them repeating from previous Z communications but lack of content-loaded terms and phrases
4) Most of the message’s overall content occurs in the first and last part of the cipher (there anagrammed) but in the middle of it; indicating ‘trying’ to make the text fit the cipher instead of deriving the text itself from the cipher (sorry but I think that might be the case)
5) Significant lack of combined phrases such as KILLINGPEOPLE which would indicate some validity of the previous/following word
6) Lack of double letters used in general, no other double letter word (except LL)
7) Lack of message itself, e.g. threats; referrence to crimes committed; plans for the future; hints on police or himself; no ‘storytelling’ (e.g. ‘interesting ride’) etc.

My opinion is:

Yes, its a (somewhat forced?) valid ‘solution’. IMO this was done with a lot of effort, thus it is a great success. But at the same time:

No, it is not the (original) Zodiac cleartext. We can discuss about this until ages but according to the previous analysis, this is what I believe. Mostly because of the above average ‘A’, the overall vowel displacement as well as 95% of the text being very common words with lack of content. Especially the lack of content-related phrases (e.g. KILLINGPEOPLE) as well as the lack of aggression/message in the proposed solution make me doubt

If I had to prove the proposed solution to being not the original one, it wasn’t that difficult: "Method changed during decryption process." Polyalphabetic may have multiple solutions regularly. But knowing how hard it is to find even such a proposed solution (on a basis of 340 letters), well.

In a message from Z, there might have been more – or not.

QT

https://streamable.com/5qh0tn

http://scottbryce.com/cryptograms/stats.html

Mark Twain text example:
Behind the Castle swells a great dome-shaped hill, forest-clad, and beyond that a nobler and loftier one. The Castle looks down upon the compact brown-roofed town; and from the town two picturesque old bridges span the river. Now the view broadens; through the gateway of the sentinel headlands you gaze out over the wide Rhine plain, which stretches away, softly and richly tinted, grows gradually and dreamily indistinct, and finally melts imperceptibly into the remote horizon. I have never enjoyed a view which had such a serene and satisfying charm about it as this one gives. The first night we were there, we went to bed and to sleep early; but I awoke at the end of two or three hours, and lay a comfortable while listening to the soothing patter of the rain against the balcony windows. I took it to be rain, but it turned out to be only the murmur of the restless Neckar, tumbling over her dikes and dams far below, in the gorge. I got up and went into the west balcony and saw a wonderful sight. Away down on the level under the black mass of the Castle, the town lay, stretched along the river, its intricate cobweb of streets jeweled with twinkling lights; there were rows of lights on the bridges; these flung lances of light upon the water, in the black shadows of the arches; and away at the extremity of all this fairy spectacle blinked and glowed a massed multitude of gas-jets which seemed to cover acres of ground; it was as if all the diamonds in the world had been spread out there. I did not know before, that a half-mile of sextuple railway-tracks could be made such an adornment.

I propose to you we do the following approach:

1. We compute the entropy of information that is in a text of length 340. That is easy, it is the sum of -p*ln(p) for all letters in english frequencies
2. We compute the unicity distance for a homophonic cipher ( https://en.wikipedia.org/wiki/Unicity_distance ). Let’s take the results from http://www.practicalcryptography.com/cr … tatistics/
3. You describe a way to generate the transposition sequence

0,19,38,57,76,95,114,133,152,1,20,39,58,77,96,115,134,136,2,21,40,59,78,97,116,135,137,3,22,41,60,79,98,117,119,138,4,23,42,61,80,99,118,120,139,5,24,43,62,81,100,102,121,140,6,25,44,63,82,101,103,122,141,7,26,45,64,83,85,104,123,142,8,27,46,65,84,86,105,124,143,9,28,47,66,68,87,106,125,144,10,29,48,67,69,88,107,126,145,11,30,49,51,70,89,108,127,146,12,31,50,52,71,90,109,128,147,13,32,34,53,72,91,110,129,148,14,33,35,54,73,92,111,130,149,15,17,36,55,74,93,112,131,150,16,18,37,56,75,94,113,132,151,153,172,191,210,229,247,267,286,305,154,173,192,211,230,248,268,287,289,155,174,193,212,231,249,269,288,290,156,175,194,213,232,250,270,272,291,157,176,195,214,233,251,271,273,292,158,177,196,215,234,252,255,274,293,159,178,197,216,235,253,256,275,294,160,179,198,217,236,238,257,276,295,161,180,199,218,237,239,258,277,296,162,181,200,219,221,240,259,278,297,163,182,201,220,222,254,260,279,298,183,202,204,223,241,261,280,299,184,203,205,224,242,262,281,300,185,187,206,225,243,263,282,301,186,188,207,226,244,264,283,302,170,189,208,227,245,265,284,303,171,190,209,228,246,266,285,304,164,165,166,167,168,169,309,308,307,306,310,311,312,313,315,314,317,316,318,319,320,321,324,323,322,326,325,334,333,332,331,330,329,328,327,335,336,337,338,339 

and we count how much information you need for that. I expect agreeing on this number, is the most challenging part.
4. We convert the results of #1 and #2 and #3 into bits worth of information. This does not matter, we can use natural logarithms instead of base 2. It is just simpler to imagine, and accept, when represented as bits, people are used to them.
[/list:u:3uhaoghr]
If the sum of the unicity distance plus the amount of information needed for describing the transposition from #3, remains below the amount of information given from #1, then mathematically there is no chance to escape from accepting that the solution to the transposed cipher is the one and only one solution, with the trust level of the likelihood that is left from the remaining bits of information.

Just by gut feeling: The result will be over-whelming, in the range of sand on all beaches on earth. But I see no-one has so far done the computation, so formally I would say, you are right in requesting, and the question deserved an answer.

After that, the discussion is finished, with a result that leaves no choice to accept or not: It will prove to be right, or wrong.

 

Hi QT,

Just a few questions:

Isn’t 1) kind of the point of a cipher? Also, this cipher follows the methods of an already existing cipher. 5) When you say significant what do you measure it against? I think you are trying to compare it to Z408 too much; which, statistically speaking, isn’t very interesting seeing as that is just one other cipher. It would be different if there had been, say, 150 cracked Zodiac ciphers; patterns in his writing could then be established with more certainty.

Hi QT,

Just a few questions:

Isn’t 1) kind of the point of a cipher? Also, this cipher follows the methods of an already existing cipher. 5) When you say significant what do you measure it against? I think you are trying to compare it to Z408 too much; which, statistically speaking, isn’t very interesting seeing as that is just one other cipher. It would be different if there had been, say, 150 cracked Zodiac ciphers; patterns in his writing could then be established with more certainty.

ad 1)
True, but usually the decryption happens in one or two steps over the complete cipher consistently. Not so here,

step #1 is the "down one, two to the right" rule,
step #2 cutting the cipher into three – even differently sized – parts
step #3 LIFEIS section inserted
step #4 shift of one line to the right
step #5 anagramming as a completely different method (but still possible)

A bit too many steps, for my taste. But I fully accept those as a method. However this results in the indication that the text has been created from the first third (first attempt), then the last two lines (second attempt), finally to be completed with what text ever in the middle part. Just a thought.

ad 5)
The Z408 was only taken to compare, e.g. to show the length of words in such text compared to Mark Twain. Accordingly, the word length did not become a significant criteria.

But if a text has one main letter occurring +42% [corr.], this is pretty much unusual. When the vowels E, O and A are pulling lots of statistic from the other letters, it becomes suspicious but not evident that the proposed text might have been rather constructed instead of linguistically been put down (without saying so but of course it might be way easier with more vowels).

Of course the Z340 stands for itself, even if it contains +46% of the letter ‘A’. It’s the sum of the ingredients which make the chef’s dinner complete. And in this case, I won’t do the Gordon Ramsey but should everybody else decide for him/herself.

Thanks for asking.

QT