AK, where is there a 340 cipher tool that you can fill in the boxes? I see Dave’s 340 webtoy site is not working.
By the way, if you guys are correct, the TH almost has to be followed by a vowel(provided Z didnt have an odd spelling error)

I just used Dave’s webtoy 30 minutes age. It should be working.

http://zodiackillerciphers.com/webtoy/

I agree, very likely that any "TH" would be followed by a vowel, as in THE, THEM, THING, THAT, etc. Unless the "TH" is at the end of a word, like WITH, SLOTH, etc.

In my proposed solution the "TH" combos are part of the words THEM, SLEUTH(EO), THESE (or could be THOSE), and THE.

AK, where is there a 340 cipher tool that you can fill in the boxes? I see Dave’s 340 webtoy site is not working.
By the way, if you guys are correct, the TH almost has to be followed by a vowel(provided Z didnt have an odd spelling error)

Hello Mr. Morf, I know that the question is not for me, sorry for the intrusion.
I think you already know this program, found in Dave’s own website:
http://www.softpedia.com/get/Security/D … ypto.shtml
I found it pretty cool, shame that my English is very bad in the field
cheers
Marcelo

AK, where is there a 340 cipher tool that you can fill in the boxes? I see Dave’s 340 webtoy site is not working.
By the way, if you guys are correct, the TH almost has to be followed by a vowel(provided Z didnt have an odd spelling error)

I just used Dave’s webtoy 30 minutes age. It should be working.

http://zodiackillerciphers.com/webtoy/

I agree, very likely that any "TH" would be followed by a vowel, as in THE, THEM, THING, THAT, etc. Unless the "TH" is at the end of a word, like WITH, SLOTH, etc.

In my proposed solution the "TH" combos are part of the words THEM, SLEUTH(EO), THESE (or could be THOSE), and THE.

Unless followed by an "R". THROUGH THROW THROAT THRILLING THREATEN etc.

Thanks up2. That is a good point. Though I think "R" may be the only consonant that can follow TH in typical English?

If so, then it still holds true that anything following TH is probably an E, A, O or I. Unless the TH is at the end of a word like WITH or SLEUTH. In which case, would it usually/always be a vowel that precedes a TH combo?

But I agree R should be included, I didn’t think of it.

Good points.

One thing we know for sure, no matter what(barring a spelling error by Z), a word with a double L must have a vowel as the letter before it…. ALL,ILL,SILLY,BILL,HILL,etc,etc,unless, its a word that ends in L followed by a word starting with L

In AK’s & QT’s solution, the ‘L’ symbol MUST BE A VOWEl, no matter what(column 16)

A much smaller list of words where TH isn’t followed by A, E, I, O, U, or R:

thwack
thwart
thy
thyme
thymidine
thymocyte
thymus
thynne
thyroglobulin
thyroid
thyroxine
thyself

I was waiting for Dave to chime in

So what’s your take Dave??

My opinion is that doing this kind of "expected letter frequency" analysis is doomed to fail. Correct me if I’m wrong, but it still makes these assumptions about the cipher text:

1) The message is written left to right, up to down.
2) The plain text is in English.
3) The plain text is a valid, comprehendible message.
4) The message is enciphered using simple homophonic substitution (like the 408 was)

I think the reason the cipher hasn’t been solved is because one or more of those assumptions are wrong. The evidence for this is the numerous other statistical code breaking approaches that have been applied, successfully to other ciphers, but unsuccessfully to the 340.

I believe that even when you make test ciphers, with known and valid messages, that closely resemble the 340, they will be easily broken with known techniques, but the Zodiac’s 340 will remain unbroken, until someone figures out what the correct assumptions are.

Drats! Thrawted again! Thy well played the game sir.

Though in a code I doubt Zodiac would use a lot of those fancy words, he used none in his letters.

I think I would say it is most likely he would use words like THE, THEM, THESE, THOSE, THING, THAT…and much less likely he would use words like thymidine, thymocyte, thymus, thynne, thyroglobulin.

I play the percentages..

Doranchak said: My opinion is that doing this kind of "expected letter frequency" analysis is doomed to fail. Correct me if I’m wrong, but it still makes these assumptions about the cipher text:

1) The message is written left to right, up to down [At least in part, in the Raw GS the words LIST and BOMBS appear diagonally, the words DUEL, LEASH, TIES appear vertically- AK]
2) The plain text is in English. [Mostly if not all – AK]
3) The plain text is a valid, comprehendible message. [At least in part, likely most of the first 8 to 10 lines – AK]
4) The message is enciphered using simple homophonic substitution (like the 408 was) [At least in part, likely the first 8 to 10 lines – AK]

Assuming that Zodiac was a bragger, who liked having people pay attention to him, why would he make a cipher that would be so hard to solve? It’s even possible that he provided the police with a code key for the first cipher to help them along. Wonder why he wouldn’t do so for the 340?

Those are definitely some big challenges by the way Dave.

My opinion is that doing this kind of "expected letter frequency" analysis is doomed to fail. Correct me if I’m wrong, but it still makes these assumptions about the cipher text:

1) The message is written left to right, up to down.
2) The plain text is in English.
3) The plain text is a valid, comprehendible message.
4) The message is enciphered using simple homophonic substitution (like the 408 was)

I think the reason the cipher hasn’t been solved is because one or more of those assumptions are wrong. The evidence for this is the numerous other statistical code breaking approaches that have been applied, successfully to other ciphers, but unsuccessfully to the 340.

I believe that even when you make test ciphers, with known and valid messages, that closely resemble the 340, they will be easily broken with known techniques, but the Zodiac’s 340 will remain unbroken, until someone figures out what the correct assumptions are.

Excellent point, Dave. The ZKDekrypto would have cut through the 408 cipher like butter if it was around at the time this was sent despite spelling and/or substitution errors. It really hasn’t come close, to my knowledge, with the 340 so SOMETHING is fundamentally different. I would vote for #1 and/or #4 above and hope that whatever curveball Z threw in could be GUESSED. I think it would be interesting if the ZKDekrypto could perhaps rearrange the 340 symbols by different approaches (like reading backwards, diagonally from corners, spiral etc.) to see if anything shakes out.

I, when I have time, I variants with altered columns or letters that contexts with Z as THE MIKADO and etc.

Doranchak said: My opinion is that doing this kind of "expected letter frequency" analysis is doomed to fail. Correct me if I’m wrong, but it still makes these assumptions about the cipher text:

1) The message is written left to right, up to down
AK: Mostly or at least in part, in the Raw GS the words LIST and BOMBS appear diagonally, the words DUEL, LEASH, TIES appear vertically.
2) The plain text is in English.
AK: Mostly if not all. What else would it be in, German? English seems a safe assumption.
3) The plain text is a valid, comprehendible message.
AK: At least in part, likely most of the first 8 to 10 lines.
4) The message is enciphered using simple homophonic substitution (like the 408 was)
AK: At least the first 8 to 10 lines, if not more.

AK: If we rules out the 4 assumptions, what are we left with? Wouldn’t any amateur effort to solve the code likely be pretty useless? And if we rule out those assumptions, it means Z created a cipher almost impossible/impossible to solve, which to me seems pointless. Yes, after the first cipher I think Z definitely wanted to make something a lot more difficult. But in my opinion there is a message (as the FBI says is likely) and at least part of it – the first 8 to 10 lines – is more or less a solvable homophonic substitution cipher.

Of course at some point, perhaps after half or all of it is solved as a normal HSC, there is at least ONE OTHER STEP that must be taken. Maybe a Caesar shift of varying degrees, 0-3-6-9-12-15. Maybe a re-arrangement. Maybe use of vertical, diagonal and backwards elements. Maybe some tight to moderate anagram use. Maybe some other factor. Maybe two or more of those mentioned.

I agree with what Michael Cole of http://www.zodiacrevisited.com said at Doranchaks site here in a comment:

I think most of us who have spent a nontrivial amount of time trying to solve the 340 by way of computer program have come to the conclusion that the cipher is something more than a standard homophonic substitution cipher (HSC). I personally believe that the cipher is fundamentally a HSC, but that the killer introduced one or more complexities that have thus far frustrated all those who have taken up the challenge.

The problem I see is that there are no good way to infer what the additional complexities actually are. The possibilities are infinite. If somebody is actually going to solve the 340, they are going to need a fair bit of luck in order to essentially “guess” one or more of the complexities.

http://www.zodiackillerciphers.com/?p=490#comments

Sure what QuickTrader is trying to do, or what I am trying to do in building on it, may well be wrong. But isn’t it worth trying? If not we might as well just give up.
——————

QT based on your previous work I think you showed why the "+" symbol likely solves as "L". Using that, and your idea on TH with my application of it, and my solve of S99 as SEE, this what you get.

I think this has the weight of both your statistical analysis and some word solves, and would provide an interesting starting point for someone to try to take it a step further, either through computer applications or pen and paper. Not saying this is absolutely right, but gotta start somewhere, and this has some solid justifications to at least be a working hypothesis.

AK-

I don’t know if you have ever used ZKDecrypto, but one of the features it includes is the ability to "lock-in" certain chosen letters and let it solve around them. Just as a test, I locked in the letters that you have chosen
above, and ran the program multiple times to see what it would come up with. The absolute best that I found, after many many tries, was this:

Which, although it contains quite a few words and vaguely looks like English, doesn’t really seem like much of a solution.

I try to avoid bringing up ZKD much, lest I sound like one of those parents with a "My child is an honor student" bumper sticker. But the truth is, ZKD is really very good at what it does. And as Dave mentioned, normal
homophonic ciphers like the 408 are broken fairly easily by programs now. Mike Cole’s PERL program has also independently solved it, and it has been done by several people as college work. Normal homophonic ciphers
are just not that hard to solve with computers now.

So I agree with the points Dave has made. With the 340, something else is going on. What it is, I do not know. And I don’t mean to sound discouraging, but I don’t think that the approach here is going to work.

-glurk

I have what may be a really dumb question, and since this thread is active I will throw it in here.

Could it decipher into a foreign language?