Zodiackillersite.com

Kind of new looking into this case intently. I know this has come up before so my apologies if this has been hashed over numerous times. Isn't it possible, and seems very likely to me, that this guy was full of you know what and there is no message in these ciphers. He just randomly threw a bunch of symbols on these papers just playing his game. I know the first had a message, but it wasn't all that complicated to solve. Either there's no message, just a random collection of symbols or the encryption is so deep and personal to just him that it would be impossible for anyone to solve. People have combed over it for 50 years and found some patterns that are highly improbable, but isn't that just statistics? If I draw a random 340 symbols or I just throw them in the air and let them fall wherever and thousands of folks and super computers sift through it, aren't you going to find some patterns that are highly improbable? I mean the improbable and unlikely really happen all the time. They're just random, naturally occurring patterns. Feel free to correct me on my statistics. Never took any classes in it, just a very basic understanding of it.

Hi..

first welcome to this board. Regarding the cipher, a while ago I had analyzed the cipher with a specific encryption tool (not cryptool but another one, you may look into cryptool first if you're interested in that stuff). It had shown similar characteristics like the 408. Additional hints could be e.g. the existence of double letters and trigrams (which rather wouldn't occur when simply shuffling symbols).

The reasons why the 340 is harder to crack is the following:

- shorter cipher text
- more homophones used
- homophones rather not used in a repeating order ('sequence' or 'cycle' of homophones used..was almost the case in the 408, see file below)
- less repeating bi-/trigrams, no repeating 4- or 5-gram at all (accidentially & due to a higher amount of homophones)

Although those are just details, all this together makes the 340 way harder to crack than the 408 cipher. Of course it is still possible that it could have no content, however there is a better chance that it has at least some message, like the 408.



Theoretically possible variations to place 26 letters on 54/63 homophones:

408 cipher: 2.56 e+76
340 cipher: 1.39 e+89

Therefore the 340 has approximately 5,429 billion more potential variations to place alphabetical letters on its homophones than the 408. You can also compare the amount of symbols compared to the amount of homophones (average symbols per homophone):

408 cipher: 408/54 = 7.6
340 cipher: 340/63 = 5.4

leading to the idea that while the 408 is nearly impossible to 'read' (it was solved without a computer..), the 340 comes up with only 71.4% of such 'readable' information. Therefore no wonder that the 340, on a first look, appears like being an absolute farrago, even more than the (solvable) 408..

A final comment on that: Due to this higher complexity of the 340 there, imo, is no need to look for different or additional encryption methods such as anagramming, line shuffling etc..the 340, as a homophone substitution cipher, indeed works very well...regarding the 340, Zodiac (Z) is still crackproof.

You may want to read the third but last post here to get more information about my approach towards the 340:

viewtopic.php?f=81&t=907&p=43766&hilit=fccp#p43766

Currently I run approximately 200-400m variations a day with a self-made Python program, coming up with interesting results (single cleartext words, however lists of 200-300 for certain parts of the cipher). If those results can be cross-checked usefully with other parts of the cipher, there could be a chance to crack it. But it's still a long way to go..

QT

Quicktrader wrote:Hi..

first welcome to this board. Regarding the cipher, a while ago I had analyzed the cipher with a specific encryption tool (not cryptool but another one, you may look into cryptool first if you're interested in that stuff). It had shown similar characteristics like the 408. Additional hints could be e.g. the existence of double letters and trigrams (which rather wouldn't occur when simply shuffling symbols).

The reasons why the 340 is harder to crack is the following:

- shorter cipher text
- more homophones used
- homophones rather not used in a repeating order ('sequence' or 'cycle' of homophones used..was almost the case in the 408, see file below)
- less repeating bi-/trigrams, no repeating 4- or 5-gram at all (accidentially & due to a higher amount of homophones)

Although those are just details, all this together makes the 340 way harder to crack than the 408 cipher. Of course it is still possible that it could have no content, however there is a better chance that it has at least some message, like the 408.

408str11.jpg

Theoretically possible variations to place 26 letters on 54/63 homophones:

408 cipher: 2.56 e+76
340 cipher: 1.39 e+89

Therefore the 340 has approximately 5,429 billion more potential variations to place alphabetical letters on its homophones than the 408. You can also compare the amount of symbols compared to the amount of homophones (average symbols per homophone):

408 cipher: 408/54 = 7.6
340 cipher: 340/63 = 5.4

leading to the idea that while the 408 is nearly impossible to 'read' (it was solved without a computer..), the 340 comes up with only 71.4% of such 'readable' information. Therefore no wonder that the 340, on a first look, appears like being an absolute farrago, even more than the (solvable) 408..

A final comment on that: Due to this higher complexity of the 340 there, imo, is no need to look for different or additional encryption methods such as anagramming, line shuffling etc..the 340, as a homophone substitution cipher, indeed works very well...regarding the 340, Zodiac (Z) is still crackproof.

You may want to read the third but last post here to get more information about my approach towards the 340:

viewtopic.php?f=81&t=907&p=43766&hilit=fccp#p43766

Currently I run approximately 200-400m variations a day with a self-made Python program, coming up with interesting results (single cleartext words, however lists of 200-300 for certain parts of the cipher). If those results can be cross-checked usefully with other parts of the cipher, there could be a chance to crack it. But it's still a long way to go..

QT

Thanks QT. Not sure my level of competence in this area is in the same ballpark with you and some of the others on here. Might take me a while to get to where I can just vaguely understand some of the concepts yall are throwing around. At the risk of sounding like a real bumpkin yall are talking some real MIT/Cambridge sounding stuff. So, you and many others of your expertise are convinced there is a message in there? I will look into some of the things you mentioned and some other beginners concepts in this ciphering business. Appreciate the information.

Welcome, MrNemo.

Many people ask this, it is a smart question. Ofcourse we have not devoted our time to a random collection of symbols.

The 408 is encoded with sequential homophonic substitution. This type of encoding has a very strong statistical signature to it. The odds that random symbols would exhibit such statistics are astronomically small. The 340, as measured, uses the same type of encoding, though more random, and therefore is extremely unlikely to be a bunch of random symbols. This is not because of some individual patterns here and there, mostly all the symbols are engaged in this type of encoding.

Quicktrader does mention some features why the 340 is harder to crack, though these are trivial to modern automatic solvers. It is far more likely that Zodiac layered different encoding/encryption methods. There is good support for of a specific kind of transposition though it seems to have been complicated by something. Both the FBI and NSA believe the 340 contains a message.

It could possibly be hoax such that there is a random plaintext beneath it but at least the ciphertext itself is genuine.

Just learned that the first homophone substitution with both, vowels and consonants, being substituted was created in 1512..nevertheless it might be solvable 'easily' but guess you need ton's of memory for it..zodiackiller.com is an interesting site as well for newcomers, to look into some of the cases.

QT

Jarlve wrote:Welcome, MrNemo.

Many people ask this, it is a smart question. Ofcourse we have not devoted our time to a random collection of symbols.

The 408 is encoded with sequential homophonic substitution. This type of encoding has a very strong statistical signature to it. The odds that random symbols would exhibit such statistics are astronomically small. The 340, as measured, uses the same type of encoding, though more random, and therefore is extremely unlikely to be a bunch of random symbols. This is not because of some individual patterns here and there, mostly all the symbols are engaged in this type of encoding.

Quicktrader does mention some features why the 340 is harder to crack, though these are trivial to modern automatic solvers. It is far more likely that Zodiac layered different encoding/encryption methods. There is good support for of a specific kind of transposition though it seems to have been complicated by something. Both the FBI and NSA believe the 340 contains a message.

It could possibly be hoax such that there is a random plaintext beneath it but at least the ciphertext itself is genuine.

Thank you sir! It's great to have some folks that are so well versed in this stuff to sort of bounce some of the fundamental, beginners type questions off. Your welcome is very much appreciated. So many forums on different subjects just seem to be full of folks just itching to argue and ridicule someone. I'll be back and hopefully I'll have some recreational reading time to get a little more familiar with this code breaking stuff

Welcome MrNemo!

The most important things have already been said, but I would like to add something else (if you don't mind my mistakes in my writing...I am not a native speaker):
You may check the following: Count the number of columns that contain no repeated symbols. Then do the same with rows.

This image illustrates it very well:
http://zodiackillerciphers.com/images/z ... olumns.jpg

You can see that a horizontal substitution was most likely performed. This means that the plain text letters have been replaced horizontally by the encryption symbols. A completely random cipher would not behave like this. However, as Jarlve said, this is not a proof that the underlying plaintext makes sense.

You can also take a look at the following page, there you will find all observations and their probabilities:
http://zodiackillerciphers.com/wiki/ind ... servations

Largo wrote:Welcome MrNemo!

The most important things have already been said, but I would like to add something else (if you don't mind my mistakes in my writing...I am not a native speaker):
You may check the following: Count the number of columns that contain no repeated symbols. Then do the same with rows.

This image illustrates it very well:
http://zodiackillerciphers.com/images/z ... olumns.jpg

You can see that a horizontal substitution was most likely performed. This means that the plain text letters have been replaced horizontally by the encryption symbols. A completely random cipher would not behave like this. However, as Jarlve said, this is not a proof that the underlying plaintext makes sense.

You can also take a look at the following page, there you will find all observations and their probabilities:
http://zodiackillerciphers.com/wiki/ind ... servations

Thank you very much sir! Appreciate the response and guidance. And yes, you're English is outstanding. Unfortunately, probably better than most Americans. Certainly better along than my Russian is. Myenya PaRusski ochyen plokho. Spaseeba