MY 340

[smithy Profile]

Adam, in respect to what you said here about Excel, if you type a single quote before an operator ('- in a cell and '+ in a cell for instance), that should show you what you're after. (A quick aside, have I misunderstood how you apply a period length? Shouldn't that be "FAFTXI"?)

Anyway, in any case I feel I've followed you reasonably, (thinking "Playfair" to myself), as far as #3, and I'm chewing on "there is no short-cut I could find to examine the z340 and determine accurately a period length" right now.

I'm also wondering how good was Z. with ciphers, and whether he would have managed this encoding scheme or one like it.
Not without errors hmm?

[ALisowsky Profile]

Adam, in respect to what you said here about Excel, if you type a single quote before an operator ('- in a cell and '+ in a cell for instance), that should show you what you're after

Haha, it's about time I learned that! Thanks Smithy.

(A quick aside, have I misunderstood how you apply a period length? Shouldn't that be "FAFTXI"?)

Hmm, I'll walk you through how I got mine:
IDECID
211121
445344
FAFTXT

Were you using a polybius square that was unmixed when you enciphered?

Anyway, in any case I feel I've followed you reasonably, (thinking "Playfair" to myself), as far as #3, and I'm chewing on "there is no short-cut I could find to examine the z340 and determine accurately a period length" right now.

I'm also wondering how good was Z. with ciphers, and whether he would have managed this encoding scheme or one like it.
Not without errors hmm?

Yeah, there are quite a few published articles detailing some good ways to decrypt this type of cipher, but those methods are nullified when you add in the homophonic sub :(

[smithy Profile]

Adam, Excel's such fun.
I was using a mixed mind when looking at your material last night, no excuses!
This thread needs an unmixed Polybius though, I can only think that changing screens made me read 24 somehow, which is a bit pathetic.



While the hunt for periodicity proceeds, I'll try not to be depressed about the idea of changing periodicity within the 340, or my own instant encipherment error, and so on. Gulp.

[ALisowsky Profile]

Unlike the playfair or the vigenere or the caeser shift discussed previously in this thread, we can't necessarily rule anything out using the methodology I've suggested.. however it does make a 'trial and error' approach feasible to find the period if you use my method. Organizing a known 340 cipher (in this case my340) by correct period and examining the rate of naturals, then organizing the 340 into n-period length, we should see similar numbers if we have the z340 organized correctly. the only catch is that we have to assume some naturals may be hidden due to the homophonic sub. But the good part is, as I mentioned above, the more high-frequency symbols (which are more likely to produce the naturals) still receive less unique symbols assigned to it.

For instance with the example I posted above, we can see 1 natural appearing in 36 characters (encoded with period=6 for the first example) and 11 naturals appearing throughout my 340 when encoded with a period length of 9. This gives us a probability of roughly what to expect if the z340 is a BIFID and we have it organized correctly (keeping in mind of course that some naturals may be hidden). I guess what I am getting at is that if someone runs the 340 using my above approach at a period of 7, and generates 6-8 naturals from that, I would consider that to be a good 'hit', and certainly does not rule out a bifid. However, if we examine all reasonable period lengths up to a certain length (i do not think this method would work beyond a period length of ~15), and they all reveal no naturals, then I do think this would support that a bifid was not used on the z340 prior to a homophonic sub.

Thanks for posting up the square!

Adam

[smithy Profile]

Adam, posting the square was the least I could do!
Now, on another thread I'm being accused of delegating, so I'm very reluctant to say this - but - aren't you going to try different period lengths yourself......???

[Pittsburgh Phil Profile]

I don't know how the hell you guys deal with these ciphers, it totally goes over my head. But I applaud you efforts.

[ALisowsky Profile]

After trying a couple periods with no results, I decided to try something different. I broke the 340 into periods according to where the periods (or dots) occur in the 340. When I did this, and again looked for naturals, I found this:



I circled the naturals which occurred but also circled potential homophone pairs. By no means conclusive, but IMO worthy of further investigation.

Thoughts?

Adam

[traveller1st Profile]

I'll wait for the experienced guys to add something useful but out of curiosity what periods did you try in the end?

[ALisowsky Profile]

traveller1st,

Prior this one, I tried period length =8, 10 for the entire message (fixed period lengths)

Just to clarify a bit more on this after re-reading my post: since the period appears at character 30, i basically treated the symbols that preceded it as one period. In other words, I am assuming for the sake of this example that the period occurring in the 340= indicator of where a bifid period ends.

The logical assumption for this type of varied period length would be he organized the plaintext by statement/sentence/phrase. For instance, if zodiac wrote, "I like killing people because it is so much fun", and chose to encipher that phrase with bifid, he would get: hgbhmmlnlaatbrtnqgsquzqdmzueklxytnrxlx (which displays one natural when factored). Then, perhaps, he decided to assign that last x in the phrase with a period when applying the homophonic sub, to keep track of what he was doing.

It's all purely speculative of course, merely a possibility that can't be ruled out with my method.

-Adam

[traveller1st Profile]

I like the idea. It's neat and it's 'neat' lol.

I'm not a cipher guy but I'm wondering from your diagram does the use of the periods give you those results.

I notice that on the first section you end with a period as it appears on the cipher but you also include it at the start of second section as it ends the first and starts the second section.

However you start the third section with no period whereas if you were to follow the pattern from the first two it should start with the period that also ends the second section. The same for sections 4 and 5.

Section 6 you are back to starting with the period that also ends the 5th section.

The seventh section doesn't start with a period but it does end with one which it doesn't do in the cipher.

Just thought I'd point that out. If there's something I'm not understanding (very possible) then just ignore all that lol.