Here are the cycle stats. I couldn’t exactly replicate the 340, which has a few more high scoring cycles. Otherwise the numbers are comparable, and the randomization numbers are almost dead on.

But here is the weird thing, the top scoring cycles:

Let’s take it from the top.

Cycle 5-12. There is the familiar pattern, but it is FALSE. On odd numbered rows, 5 maps to B and 12 maps to E.

Cycle 54-56. There is the familiar pattern, but it is TRUE. On odd numbered rows, they map to T, and on even numbered rows, they map to G.

Cycle 13-47. The familiar pattern again. 13 maps to R on odd rows, and 47 maps to R on even rows! I don’t even know how to explain that one yet.

Cycle 55-56. The familiar, not quite perfect pattern again. It is TRUE. They map to T on odd rows, and to G on even rows.

Cycle 11-35. It is FALSE. On odd rows, 11 maps to E and 35 maps to N. On even rows, 11 maps to R and 35 maps to A.

So this is very painful to say the least and I don’t know exactly what to make of it.

Note again the cycle stats. At the 512 range, the message has 14 cycles. But if I scramble completely, I would only get from 0.7 to 2.7 cycles at the 512 range.

And for the top perfect cycle, 24-32, it is FALSE.

Before I turn in tonight, I will randomize the new message once and show the top scoring cycles.

I marked the cycles, TRUE, mixed, or FALSE.

Most of them are FALSE.

Now, here is one randomization of the same message:

There are not very many high scoring cycles at all. And the numbers are very dead on to 340 randomization stats.

With the experiment, I was able to create a LOT of false cycles.

Here are the 340 top cycles:

With the experiment, the 340 has 11 cycles that score over 520. The alternated key for row has 16 cycles that score over 520, even though most of them are false. And if I randomize that message, I get only two cycles that score 516+.

I do not know how to explain all of the false cycles. But if Zodiac used more than one key, we are screwed.

Thanks for another cipher smokie.

The message solves on the even rows and scores a bit higher than the 340 (because the message on the even rows scores higher than the message on the uneven rows so it becomes dominant).

erawlwayaratcrbcr
ebecauseitissomec
usheadreadershose
enkillindwildsome
regunsbeelidorhnh
semanisthemostmen
tlethinaaanydstar
tokillsomethindsi
infangulatedsuear
lingewheretceitis
rinaondslegenetrd
titsrounlocksoits
aguntreadunonfsct
ntofitisthatwheni
orremaryonenodeaa
theromiceandallth
reetinwayandmaryo
ecomemysloresisil
yabstainlthalanal
becauseyouwilltna

Then I expanded the symbols on the unvenen rows into a new set and got an almost perfect solve.

slikekillingpeopl
ebecauseitissomuc
hfunitismorefurth
ankillingwildgame
intheforrestbecau
semanisthemostdan
gerousanimalofall
tokillsomethinggi
vesmethemostthril
lingexperenceitis
evenbetterthanget
tingyourcocksoffw
ithagirlthebestpa
rtofitisthatwheni
tieswillbereborni
nparadiceandallth
eshavekilledwillb
ecomemyslavesiwil
lnotgiveyoumyname
becauseyouwilltry

The same was done for the 340 with rows and columns, resulting scores were rather poor in comparison to smokie’s 5th cipher and columns did quite a bit better than rows.

340 even columns expanded:

urdmanicalthecosm
edistheworkforusg
overanalorthehand
idatingrelatindre
amanalysthemicall
andgenesiselement
shirtwasarthefact
thesubsticebandco
mpanynamemorthegu
esthenisitatomove
andwillbobodycare
epandstricalongse
asonthisthesasand
thesisisaforegoni
tsthisamericalisi
trolasusedifulthe
dronehemonentsthe
ingrestratinowtha
tmostpolocktorema
ilonedonalentbuyi

But perhaps the scheme is harder when done for columns?

But perhaps the scheme is harder when done for columns?

What do you mean? I was thinking that maybe Zodiac two different keys and changed with every letter. The 17 columns would hide the scheme because if one key has unique symbols, they wouldn’t line up with each other vertically in columns. It would be more difficult to spot. So, I guess, checkerboard is what I was thinking.

But your solution to the 340 looks very interesting. I see pairs of words that seem to sort of naturally go together, like "chemicallandgenes," which could be "chemicals and genes". I see "companyname," which could be something "my name". I see "andwillbobodycare". I see the words "thisamerica".

Well I guess it isn’t a totally coherent solution but maybe a good start. Congratulations on your new solver!

urd manical the cosmedis the work for us goveran alor the hand
i dating relatin dream analys them i call and genesis element
shirt was ar the fact the substice band compauny name morthegu
es then is it a to move and will bobody careep and stri calongse
as on this the sas and y the sis is a foregon its this america l is i
trolas used iful the drone he monents the ingrestrati now that
most polock to remail one don alen t buy i

Interesting when ya break it down into little bits. Not sure if it helps to put a theory of how z built the code.

Ps you need to read it fast a few times..

Then try it slow like a poem with breaks ? maybe

@smokie and Mr lowe, with the expansion of the even columns into a new symbol set the cipher becomes 120 symbols. Meaning that it is much more probably for readable fragments to appear. I personally don’t give much credit to it. Thanks for splitting up the solve into words Mr lowe.

@smokie, you used 2 keys, one for uneven rows and another for even rows. Something like that can also exist per column, but I believe it could be harder to crack (untested). Checkerboard is another possibility that may or not be harder than the rows thing. daikon noted some odd things about symbols at uneven versus even positions in the 340: viewtopic.php?f=81&t=2625

Yup i am with you on that jarlve,.. Not looking for a solve but a clue? or pattern in how he may have built the code. .

Something simple like a verse from genesis, it written in poetic form, a line from a play, line swap, repetition, name drop, date, place or how different areas have runs of readable text where others is gobbidycrap.
What I am afraid of is you guys may be looking for the perfect solve and may have a clue pass you by because you don’t see a partial solve (million to one) I know, but our human thought processes can sometimes spot the diamond.

Jarlve, I understand that long strings of expanded symbols cause the solver to find anything.

Lowe, I agree that gems may be missed. I wonder about that sometimes.

And so I am thinking about the idea that maybe Zodiac may have cycled his key. Like a "Vigenere" cipher maybe.

I read daikon’s post. Most of the symbols that are left out of the odd or even category are low count. 37 appears 7 times and 38 appears 5 times. But none of the groups really cycle with each other.

But that gave me an idea. Maybe break down the 340 into odds and evens, and see if there are an extraordinary number of perfect or high scoring cycles. Or the message could be broken down into more than two groups. Like three or four or whatever. See what happens.

Last night when making two keys intending them to cycle at each row, I used a shift of 13. And with initial keys, some of the high frequency plaintext lined up with low frequency or no frequency plaintext when making the shift. So I had to distribute the ciphertext more evenly to make sure that I would always have a ciphertext to use when making the shift.

Now that I have my spreadsheet set up, I can make a checkerboard shift ( only two keys used at odd and even numbered ciphertext positions respectively ) pretty quickly. That is shifting at columns in my view, because there are an odd number of columns in the message.

I think that I am going to do that first to see what happens with cycles, etc. And I am also considering making a straight 1:1 for each plaintext, just to see if there are naturally occurring cycles in the English language.

So that is three new cycle projects for me now. Last night I was ready to retire from the 340 entirely. But your post, Jarlve, gave me inspiration!

Here is the checkerboard message.

smokie6

The key for odds:

A 1 2 3 4
B 5 6
C 7 8
D 9 10
E 11 12 13 14
F 15 16
G 17 18
H 19 20 21 22
I 23 24 25 26
J
K 27 28
L 29 30 31 32
M 33 34 35
N 36 37 38 39
O 40 41 42 43
P 44 45
Q
R 46 47 48 49
S 50 51 52 53
T 54 55 56 57
U 58 59
V 60
W 61
X 62
Y 63
Z

The key for evens (EDITED AFTER FINDING MISTAKES):

A 54 55 56 57
B 5 6
C 7 8
D 9 10
E 1 2 3 4
F 15 16
G 17 18
H 11 12 13 14
I 19 20 21 22
J
K 27 28
L 23 24 25 26
M 33 34 35
N 29 30 31 32
O 36 37 38 39
P 44 45
Q
R 40 41 42 43
S 46 47 48 49
T 50 51 52 53
U 58 59
V 60
W 61
X 62
Y 63
Z

I didn’t shift the entire key. I shifted only high frequency plaintext A E H I L N O R S and T to the next high frequency plaintext on the key.

Here is the message (EDITED AFTER FINDING MISTAKES):

smokie6

23 23 24 27 11 28 25 24 31 22 36 17 44 2 40 45 32
3 5 4 7 54 58 46 11 19 54 20 51 48 41 33 59 8
19 15 58 30 25 51 26 49 34 38 46 2 16 59 38 52 20
55 39 27 23 23 30 20 36 18 61 21 31 9 17 56 35 3
26 30 57 13 14 15 43 41 48 1 50 50 6 2 7 57 58
47 13 33 1 31 23 48 55 14 14 34 40 49 56 10 2 32
18 1 49 37 59 46 3 29 24 35 4 26 42 16 1 23 30
53 43 28 25 25 32 47 40 33 12 50 19 22 37 17 18 19
60 3 52 34 14 51 20 1 35 37 53 52 57 13 46 20 29
24 25 31 17 2 62 44 13 41 14 32 8 1 26 50 23 46
12 60 13 29 5 4 55 52 11 42 57 14 2 30 18 2 54
51 24 31 17 63 42 58 49 40 43 7 27 47 40 15 16 61
25 52 19 56 18 22 47 25 57 12 13 6 14 48 54 45 4
42 55 37 15 19 56 20 53 53 21 54 54 61 22 1 39 21
9 22 12 19 61 20 32 23 5 3 49 4 6 38 46 29 25
30 44 55 47 56 10 22 8 1 4 31 9 54 30 25 55 11
12 19 20 55 60 3 28 20 32 23 14 10 61 21 30 25 5
1 7 39 33 2 34 63 50 26 3 60 13 47 26 61 23 23
30 32 40 52 17 20 60 4 63 37 59 35 63 29 4 33 11
6 12 8 1 58 52 3 63 38 59 61 25 25 32 53 48 63

The solution is I like killing people etc.

I have no idea what the cycles will look like, but am trying to emulate 340 stats.

Here are the cycle stats for the checkerboard message. They are similar to the 340, but not dead on. It looks like randomizing the checkerboard message results in a few more higher scoring cycles.

But here are the cycles that score above 256. Most of them are FALSE. However, the number of cycles that score above 256 is almost exactly the same as the 340, with the checkerboard at 33 and the 340 at 34.

So, with cycling the key, I was able to make about the same number of high scoring cycles.

That’s all I have to report tonight. I need to work backwards now, and break the message down into odds and evens and show the cycles. I haven’t figured out how to do that yet, but I will practice with the checkerboard message. Then on to the 340 with different breakdowns into twos, threes, etc. See if there are a bunch of perfect cycles or better stats at some point by doing that.

Thanks for yet another cipher smokie! I haven’t been able to solve it so far, though I confirmed that it is solveable with cribbing "ilikekilling". I tested the full message and the even and uneven parts with my non-repeats measurement, normalized by counts. Higher score is typically more cyclic.

340:
Full message: 4462
Uneven symbols: 4258
Even symbols: 3412

408:
Full message: 4498
Uneven symbols: 3715
Even symbols: 4641

smokie6:
Full message: 3917
Uneven symbols: 4920
Even symbols: 5792

So it doesn’t seem actual for the 340 because the full message scores higher than any of the parts. While for the smokie6 both parts score (much) higher than the full message.

Well I don’t know if it is necessary to try to solve it. Don’t feel compelled to do it.

But I was wondering about using cycle stats to detect the use of key cycling. I wonder if we could develop a simple way to do that. Do your numbers reveal a two key cipher because the odd / even stats are higher than for the whole? What about three, four or five key ciphers?

The numbers reveal a 2 key (even/uneven) for the smokie6 but not for the 340. Here are the results for 3 and 4 parts.

340:
Full message: 4462

3 parts:
part 1: 3775
part 2: 3234
part 3: 3279

4 parts:
part 1: 2952
part 2: 2248
part 3: 3784
part 4: 2988

For the 4 parts there is either not enough data to work with or something interesting might be going on. I’m going to run a test.

I made another mystery cipher that is in line with what we are currently exploring: m5p1.txt.

@smokie, I’m not sure if you will be able to adapt your spreadsheet to this kind of stuff.

Edit: I’ve been able to solve my own cipher (expanded to 148 symbols!) and will try the same on the 340.

O.k. thanks a lot, Jarlve. I will do what I can regardless and fairly soon.

Of course, if you solve the 340 in the meantime, don’t wait for my results before you show us the solution.