Unigram distance curiosity

[doranchak]

It was only through the image of Jarlve that I realized how astonishing this observation is. Nevertheless, I wonder how P15/P19 fits in. So far we assume that z340 was first transposed in clear text and then substituted. If you combine this with the current assumption that lines 1-6 and 15-20 were encoded with key "A" and the middle (i. e. lines 7-14) with key "B", we are facing a problem: You cannot solve the individual parts separately, but must know both the transposition and the key. If that is the case, I currently have no idea how to solve such a cipher automatically.

This problem reminds me of George Lasry's research about solving double transposition ciphers using hillclimbers:

https://www.evernote.com/l/AAHGCBP9mqdP ... jgfTxSMLfE

He came up with a way to split the problem into two steps rather than searching the massive combined space of all keys. I wonder if a similar approach could be applicable to the type of cipher you're suggesting.

[doranchak]

It's just a hunch but the process I described should be supported by some observations made in this thread, such as
- the unigram distance suggests the message was cut and rearranged.
- the symbols not used in the middle / top-bottom suggests the use of two ciphers.
also
- cutting the message horizontally was done in Z408 (3 eight-row blocks).

It also seems to relate to what the FBI's Dan Olson believes about the message being split:
http://zodiackillerciphers.com/wiki/ind ... _Dan_Olson

[borkky]

borkky:
Hello and welcome to the forum. It's nice that so many people are working on the z340 solution. Your widget is really great and I have experimented a lot with it. However, I made a small change and changed the highlight color from grey to red.
You're from Sweden, right? I was there once at Långsön Lake near Rimbo.I enjoyed kayaking around the lake every day. Such a beautiful landscape and many kind people!

Yea, I'm a Swede alright. Countryside outside of Rimbo is really wonderful, at least in the summertime. Used to celebrate summer there when I was younger.

I took the liberty of changing the hover color to red, too. Thanks

[borkky]

It's just a hunch but the process I described should be supported by some observations made in this thread, such as
- the unigram distance suggests the message was cut and rearranged.
- the symbols not used in the middle / top-bottom suggests the use of two ciphers.
also
- cutting the message horizontally was done in Z408 (3 eight-row blocks).

It also seems to relate to what the FBI's Dan Olson believes about the message being split:
http://zodiackillerciphers.com/wiki/ind ... _Dan_Olson

Oh, wow!
I am quite new to this topic and have much existing material yet to read! This was a very interesting link.

[doranchak]

Oh, wow!
I am quite new to this topic and have much existing material yet to read! This was a very interesting link.

Yes, it can be challenging to find all the good information surrounding the ciphers. There is a lot of noise to sort through as well. My initial frustration with that challenge led me to make my page of summaries of factual details about the ciphers: http://zodiackillerciphers.com/wiki/ind ... servations It probably is not complete but I try to keep it up to date. Let me know if you notice anything missing!

[borkky]

Oh, wow!
I am quite new to this topic and have much existing material yet to read! This was a very interesting link.

Yes, it can be challenging to find all the good information surrounding the ciphers. There is a lot of noise to sort through as well. My initial frustration with that challenge led me to make my page of summaries of factual details about the ciphers: http://zodiackillerciphers.com/wiki/ind ... servations It probably is not complete but I try to keep it up to date. Let me know if you notice anything missing!

I had found the wiki and the observation-page, but there's yet alot for me to ingest
Thank you very much for all your work! The wiki is an excellent way of organizing all these theories.

Also, the fbi analysis notes you linked to seems to give additional credibility to that the message was cut horizontally (the suggest cut in two and rearranged, but they did not discover that there appears to be a different thing going on in the center. Also they noted

Additionally, there is far greater randomness for rows versus columns. This rules out any form of columnar or diagnal transpositions (a big step forward).

Which also strengthen the case.

It came to my attention the poster moonrock has been quouted, and his posts was deleted? Has anyone had a look at https://archive.org/web/ or https://archive.fo/ for cached content? Did (s)he post other substantial ideas that was lost when messages was deleted?

[Jarlve]

I wonder how P15/P19 fits in. So far we assume that z340 was first transposed in clear text and then substituted. If you combine this with the current assumption that lines 1-6 and 15-20 were encoded with key "A" and the middle (i. e. lines 7-14) with key "B", we are facing a problem: You cannot solve the individual parts separately, but must know both the transposition and the key. If that is the case, I currently have no idea how to solve such a cipher automatically.

Convert the cipher to numbered by appearance. You have 63 symbols numbered from 1 to 63, numerically add 63 to the symbols of the middle 8 rows to make these distinct from the other rows/key. Then transpose. You can do it with AZdecrypt, under functions, manipulation, raise periodic, from: 103, to: 238. 6-grams should do the job.

[doranchak]

It came to my attention the poster moonrock has been quouted, and his posts was deleted? Has anyone had a look at https://archive.org/web/ or https://archive.fo/ for cached content? Did (s)he post other substantial ideas that was lost when messages was deleted?

Moonrock posted some great ideas about types of homophone cycles. He returned to talk about them starting here:

viewtopic.php?p=56464#p56464

[borkky]

Moonrock posted some great ideas about types of homophone cycles. He returned to talk about them starting here:

viewtopic.php?p=56464#p56464

Thanks for pointing me in the right direction

[smokie treats]

Yes he may have originally intended to send three different parts of the message and the idea of multiple keys makes sense with that. They would either have to be two keys that were nearly identical but for a few symbols if he transposed the plaintext first, otherwise we wouldn't have P19. Or P19 is a phantom spike caused by the two keys, but I don't know if this is possible.

I have messed around with this idea before. Untranspose at P19, identify the symbols originally in the top 6 and bottom 6 rows ( but how do we know that is really what it is? ), then cut and paste and try to solve the blue area by rows. The rows not connecting to each other. It didn't work for me but I don't recall trying P15. I wouldn't want to discourage experimentation here. It could be top 5 bottom 5, or something else. The three parts could originally have been cast in a different shape.

EDIT:

discussion.12.10.2017.png

Again, the font is so awesome and thanks Largo.

It might be a good idea to take a closer look at the symbols unique to the top 6 and bottom 6 rows and see if they are commonly parts of the more frequent isomorphic "cycle" patterns. We did that a little bit recently.