Just speculation, but maybe the + means to rearrange symbols that are to the left, right, above and below the +. That’s why it is a +. But that doesn’t explain Row 2 and Row 3 .
Here is a new message. All of the cycles are randomized from 2% to 30% step 2% from Row 6 down. There is one 1:1 substitute with a count of 9. Originally, the message started with only 42 bigram repeats; I made the key to distribute the symbols as flatly as possible. There are a few uncommon words, one is really uncommon. One word is spelled backwards and has a null before and after it. There is one wildcard symbol that I used to mask only two bigram repeats by substituting the wildcard for one of the two symbols, and then flipping the symbol on the right and left of the wildcard. They are on the same row. See if your hot spot program can find that stuff.
Here is the key:
A 1 2 3 4
B 5
C 6 7
D 8 9
E 10 11 12 13 14
F 16
G 17 18
H 19 20 21 22
I 25 24 25 26 27
J
K 28
L 29 30 31
M 32
N 33 34 35 36
O 37 38 39 40 41
P 42 43
Q
R 44 45 46
S 47 48 49 50
T 51 52 53 54 55
U 56 57
V 58
W 59 60
X 61
Y 62
Z
Here is the message:
47 32 37 28 25 10 48 19 38 56 49 11 20 1 50 51 21
12 5 44 39 59 33 45 13 6 29 57 47 14 2 30 48 40
28 34 41 60 35 3 49 31 37 61 38 50 7 10 29 11 47
46 12 6 30 56 48 4 52 22 13 49 42 24 8 14 44 19
1 50 2 58 10 36 39 32 40 57 47 5 25 53 11 17 45
41 59 48 54 37 5 12 55 60 13 14 33 49 26 61 51 38
52 59 10 34 53 62 32 27 31 29 25 32 11 54 12 46 50
3 35 9 30 24 58 13 47 16 39 44 40 36 14 55 41 51
60 37 62 10 4 45 48 52 20 11 61 48 46 12 8 25 43
50 61 7 1 33 44 2 35 18 13 16 45 38 32 31 26 17
21 54 54 39 9 3 46 28 5 44 40 59 35 5 56 55 51
22 14 52 19 41 45 4 61 1 36 8 2 5 9 37 32 10
33 32 3 62 34 37 51 35 11 6 12 47 48 1 46 27 29
62 5 13 54 20 14 49 1 32 10 7 38 30 40 44 32 2
46 28 25 36 18 50 41 33 51 22 11 8 37 46 49 3 31
48 24 38 63 9 16 51 21 13 52 19 39 61 63 44 30 40
41 28 49 30 27 28 14 1 58 26 37 31 24 33 2 36 8
57 36 29 25 28 10 38 53 20 11 45 50 42 26 9 12 46
47 55 20 13 44 14 3 45 11 39 34 30 62 48 25 61 11
62 12 49 26 34 50 55 10 4 8 40 16 14 27 17 22 55
I’m curious what you guys think about these patterns and how they might related to the encoding schemes you’ve been experimenting with:
I dunno. I have been marking some of the dead zones and trying to figure out if some of the symbols there are the same. I wonder if the scores that make these cells shade light or dark correlate with high or low scoring cycles. I see one of your boxes that shows dark, the last one that you posted upper left row 2 and 3, but that’s it.
I think that maybe Zodiac did "something" to make the dark areas. Especially Row 2 and Row 3 and Row 16. What about symbol 54 at Column 4, Row 11? Why is that so dark all by itself? Maybe uncommon words or misspellings, like you said. We probably need to do some testing with the cell shading program and see what makes stuff dark or light. I have contemplated that the + symbol serves to reverse and then forward sections in between + symbols. That he may have masked his bigrams more where the dark areas are. The message is so relatively short. I have made some test messages, and can only get a few sentences with 340 symbols.
I dunno.
Thanks for creating another cipher smokie. Here is the map of smokie3. The message solves really well so I think the n-grams will have picked accurately up on the stuff you did. I corrected the last 4 squares but they will not be very accurate since there is normalization going on.
smokie3:
In the following images I colored all symbols in the 340 involved with the darker line according to the regions smokie selected in his last image, the 408 is brought up on the right side for some comparison. All but the "B" symbol are first time occurences in the cipher. They occur 33 times in the first 10 rows and 42 times in the last 10 rows, not a huge discrepancy but for the 408 it is the other way around. Going from row 4 to 8 there are not many of these symbols sitting around, but the 408 has similar rows. The "B" symbol likes to group with the other selected symbols.
I think that maybe Zodiac did "something" to make the dark areas. Especially Row 2 and Row 3 and Row 16. What about symbol 54 at Column 4, Row 11?
Either it’s a clue to what is going on with the 340 or it’s a fluke. I’ve run a test where all combinations of up to 4 symbols are removed from the 340 and symbol 54 sure likes to pop up in the top results. What if the wildcards are not as black and white as we think them to be. And that for instance all (or a larger set then previously assumed) symbols have some wildcards in them. For example the "+" symbol could have 12 instances where it binds to just one letter and in 12 other instances it is a wildcard/poly/null. It could explain why we are not solving it with the expansion thing. Could you make such a cipher with about 50 individual wildcard symbols?
Since the 408 also has some individual polyalphabetic symbols in them maybe the 340 is just more of the same, but to a much higher degree?
1. There is an uncommon word on Row 3; it s darker there.
2. At the end of Row 9/ beginning of Row 10, I spelled a word backwards and added an "x" to the beginning and end of the word (the message started at 338 letters so I needed to put them somewhere anyway). I find it interesting that the dark spot on Row 10 comes immediately after the gibberish word.
3. I masked two bigrams on Row 16 by replacing them with a wildcard and also exchanging the symbols to the right and left of the wildcard. The one on the left is really dark; it’s the darkest spot on the whole message. The one on the right was darker than average.
4. No explanation for the dark area on Row 5.
I’ve run a test where all combinations of up to 4 symbols are removed from the 340 and symbol 54 sure likes to pop up in the top results. What if the wildcards are not as black and white as we think them to be. And that for instance all (or a larger set then previously assumed) symbols have some wildcards in them. For example the "+" symbol could have 12 instances where it binds to just one letter and in 12 other instances it is a wildcard/poly/null. It could explain why we are not solving it with the expansion thing. Could you make such a cipher with about 50 individual wildcard symbols? Since the 408 also has some individual polyalphabetic symbols in them maybe the 340 is just more of the same, but to a much higher degree?
Yes, I have suspected that some of the symbols may have multiple purposes.
19 20 20 19 19 19 19 19 19 19 19 19 19 20 19 19 20 19 20 19 20 19 20 19 19 20 19 20 19 19 20 19 19 20 20 19
That’s the most cycling that 19 does with any other symbol. It’s symbol 20, or the B. However, the 340 has 35 cycles with eight consecutive alternations, with up to 20 of them being false. If I randomize the 340, Symbol 19 may do that with one of the other high count symbols just because of the high count. On the other hand, the example above does look a bit man made to me. It’s just a subjective thing to some extent.
I have considered that symbol 5 could be a wildcard for the top one third of the message, and cycle with symbol 29 in the bottom two thirds of the message.
5 5 29 5 5 5 5 29 5 29 5 29 5 29 5 29 5
I have considered that he may have cycled the use of the symbols. He may have had some method to shift some of the cycles. To shift symbols in and out of cycles. A mechanical way of some sort.
Do you want me to make a message like that, not trying to mask bigrams, but just a more mixed use of wildcards? Blend them into the cycles a bit and then tell you what the wildcards are? Use about 50 total count. Is that what you are asking for?
Maybe we should think of a way to shift or cycle the cycles a little bit and try it to see how it feels. Maybe there aren’t any wildcards used in the way that I have been thinking that they were used. Maybe he shifted the cycles somehow as he went along, causing some of the symbols to be polyalphabetic. Are the two examples above evidence of this? Tell me what you think.
Just a rough draft of an idea. Maybe he had a key, and somewhere, perhaps at mid point, he shifted one line or row of the key. Here is a rough draft of what I am thinking. Something like this:
4. No explanation for the dark area on Row 5.
Just uncommon string of letters.
Do you want me to make a message like that, not trying to mask bigrams, but just a more mixed use of wildcards? Blend them into the cycles a bit and then tell you what the wildcards are? Use about 50 total count. Is that what you are asking for?
Yes, if you please.
Maybe we should think of a way to shift or cycle the cycles a little bit and try it to see how it feels. Maybe there aren’t any wildcards used in the way that I have been thinking that they were used. Maybe he shifted the cycles somehow as he went along, causing some of the symbols to be polyalphabetic. Are the two examples above evidence of this? Tell me what you think.
These cycles for 19 and 20, 5 and 29 are interesting and indeed seem deliberate. The idea of the cipher being polyalphabetic is certainly worth exploring. If actual it’s very much the question how exactly it was done, since there are many ways to go about it. I’ll start thinking about creating some polyalphabetic ciphers, give me… quite a while.
K, give me a bit to make the message. I am going to look at 29. See if anything cycles with 29 before 5 cycles with 29, and then stops. It may be difficult to trace a shift with cycles that have many symbols and randomization, but some of the two symbol cycles with low randomization may be traceable. I want to look at that a bit first.
EDIT: Here, try this when you get the chance. This is the key for the first half:
A 1 2 3 4
B 5
C 6 7
D 8
E 9 10 11 12 13
F 14 15 16
G 17 18
H 19 20 21 22
I 25 24 25 26 27
J 61
K 28
L 29 30 31
M 32
N 33 34 35
O 37 38 39 40
P 41 42 43
Q
R 44 45 46
S 47 48 49 50
T 51 52 53 54 55
U 56 57
V 58
W 59 60
X
Y 62
Z 63
Here is the slightly different key for the second half:
A 1 2 3 4
B 5
C 6 7
D 9
E 8 10 11 12 13
F 14 15 16
G 17 18
H 19 20 21 22
I 25 24 25 26 27
J 61
K 28
L 29 30 31
M 32
N 37 34
O 33 38 39 40
P 41 42 43
Q
R 44 45 46
S 47 48 49 50
T 51 52 53 54 55
U 56 57
V 58
W 59 60
X
Y 62
Z 63
Here is the message:
51 19 9 44 10 8 17 29 1 45 11 37 14 52 20 12 53
38 46 6 21 30 25 18 22 54 24 33 17 55 19 13 25 34
51 9 44 26 39 45 40 15 53 20 10 5 31 37 7 28 21
38 56 47 11 48 22 39 59 12 8 32 13 53 35 9 60 40
46 49 54 37 16 32 62 2 41 42 44 10 20 11 35 50 27
38 33 47 45 12 3 29 25 63 13 8 55 21 9 43 24 46
4 51 10 48 59 11 44 12 25 34 41 39 49 50 13 47 48
26 40 34 35 14 52 19 35 19 38 57 49 10 1 33 8 50
53 39 45 11 47 54 20 12 46 13 60 2 47 55 21 9 6
3 49 28 40 15 7 37 18 34 2 7 51 22 10 35 13 60
12 45 13 52 19 8 42 38 35 28 1 37 9 5 36 10 2
9 1 50 5 11 16 39 46 12 4 34 9 60 20 4 53 51
13 35 14 40 30 9 27 34 7 46 8 2 49 10 9 35 62
21 33 44 45 38 35 37 39 55 1 48 25 17 34 40 16 4
37 62 43 44 24 47 33 34 11 45 25 6 38 56 31 9 39
37 30 62 61 57 9 17 13 51 22 1 52 2 30 31 19 2
9 41 11 46 26 50 20 8 9 36 34 9 32 35 21 10 1
44 53 47 32 40 54 10 32 12 48 33 46 13 29 62 35 19
2 51 27 19 4 9 37 38 52 5 8 11 34 53 20 11 46
12 54 39 42 13 45 27 48 21 60 24 55 36 51 19 8 32
Randomization of symbol selection increases from 2% to 30% starting at Row 6. The polyalphabetic symbols are 8, 9, 33 and 37. The wildcards used to mask bigram repeats, primarily in the second half, are 35 and 36, but symbol 35 also maps to N in the first half. Total count for all of those symbols is 54.
Smokie
I have another idea about how to figure out what the cycles are. Another hill climbing approach.
Start with 26 cycles and arrange the symbols any way that you want. Score the cycles and total up the scores. Pick a random symbol and move it. If the total score improves, keep the random change. If not, move a new random symbol. Keep doing that until the total score cannot be improved upon.
Thanks for the cipher smokie. It solves very roughly without expanding any symbols and much better so with. My early conclusion is that it’s plausible, but more of this has to be going on with the 340, for it not to solve.
The hill climber process you are describing is good but prone to get stuck in an early local maxima, because it has not enough exploration. There are three ideas/concepts which you typically have to combine to some degree, in some way, to get the best results.
The first is exploitation, which is hill climibing at its core and it basicly means you always make the best possible move (which typically will let the algorithm become stuck in an early local maxima since it can’t go back).
The second idea is exploration, which can be compared to a random walk in a given searchspace (which will always find the right solution but will take millions of years).
And the third idea (meta-algorithm) are simply random restarts, which means that for your case you simply randomize the symbol distribution over the cycles you start with. And you restart with a new randomization once every so much time/iterations/until it gets stuck.
Thanks for getting back to me on that, and especially for the explanation about hill climbing. I have been using exploitation with my horizontal hillclimber.
Anyway, I will be thinking of a way to do what I described above with Excel. What I should have done is gotten a degree in computer programming. But in any case, consider my idea at some point in the future, and how it could be adapted or used.
I spent some time looking at cycles last night. Some cycles are consistent from Row 1 to Row 20. Other high scoring cycles are disrupted at different areas of the message. Some in the middle, and some in other areas. Some cycles look deliberate even though they are not perfect. There are cycles similar to A B B A B B A B B A and so on.
Sometimes A will cycle with B, and B will cycle with C, but A and C don’t cycle together. And then C of course will cycle with D and E and so on. I have considered a Caesar shift of cycles at different points in the message, other forms of cycling the cycles, and even a mathematical formula to determine symbol selection within a cycle.
19 20 20 19 19 19 19 19 19 19 19 19 19 20 19 19 20 19 20 19 20 19 20 19 19 20 19 20 19 19 20 19 19 20 20 19
Just look at the cycle with 19 and 20. Makes me think of doranchak’s discussion of duplicated highly improbable random cycles in the prior thread. But it also looks like there is some type of shifting happening. Almost like following a mathematical curve where 20 is dominant at first, then 19 is very dominant for a while, then 19 and 20 are balanced, then looks like 20 is becoming dominant toward the end just like at the beginning. Maybe instead of wildcards, whether random or deliberate, the shifting of cycles or symbol selection takes care of everything to make the message unsolvable. The dude liked to cycle stuff, so it would seem to make sense that he could have used some type of two-directional cycling.
Anyway, thanks for working with me on the wildcard idea. Good luck with your new solver!
I have considered a Caesar shift of cycles at different points in the message, other forms of cycling the cycles, and even a mathematical formula to determine symbol selection within a cycle.
Something like that could very well be, or some change from row to row. It’s something that I first considered when comparing my non-repeat metric between the 408 and the 340. It’s a simple metric which relates highly to the cycles.
In the following images the red line is of interest, note the obvious difference between the 408 and the 340. The 340 oddly peaks at 17 and then drops off very sharply. This is somewhat reproducable with other schemes and blind luck (wildcards etc) but it’s still quite pronounced and unusual in the 340. Even after removing the dreaded "+" symbol, 17 remains the highest peak. This is why I believe some kind of row per row alteration of the cycles is worth looking into (and the other things you mentioned).
408:
340:
By the way smokie, if you can find the time, can you please share the plaintexts of your 3rd and 4th cipher, or did I miss them? Thanks. I have some translations from AZdecrypt but they are not perfect.
Jarlve …Could it have been built like a crossword puzzle similar to his slaves paradice by fire Etc letter.. With filler confusing the crap out of us.. Apparently there was this, I don’t know the date of it nor have I seen it. But it would be interesting. Thomas Gardiner, Zodiac Killer, crossword puzzle feature
It may explain some of your findings
Here is one.
s m o k i e s h o u s e h a s t h
e b r o w n r e c l u s e a l s o
k n o w n a s l o x o s c e l e s
r e c l u s a t h e s p i d e r h
a s a v e n o m o u s b i t e g r
o w s t o b e t w e e n s i x t o
t w e n t y m i l l i m e t e r s
a n d l i v e s f o r o n e t o t
w o y e a r s t h e x s r e d i p
s x c a n r a n g e f r o m l i g
h t t o d a r k b r o w n b u t t
h e t h o r a x a n d a b d o m e
n m a y n o t n e c e s s a r i l
y b e t h e s a m e c o l o r m a
r k i n g s o n t h e d o r s a l
s i d e o f t h e t h o r a x l o
o k s l i k e a v i o l i n a n d
u n l i k e o t h e r s p i d e r
s t h e r e a r e o n l y s i x e
y e s i n s t e a d o f e i g h t
And here is the latest one, it is an excerpt from Treasure Island.
T H E r e d g l a r e o f t h e t
o r c h l i g h t i n g t h e i n
t e r i o r o f t h e b l o c k h
o u s e s h o w e d m e t h e w o
r s t o f m y a p p r e h e n s i
o n s r e a l i z e d T h e p i r
a t e s w e r e i n p o s s e s s
i o n o f t h e h o u s e a n d s
t o r e s t h e r e w a s t h e c
a s k o f c o g n a c t h e r e w
e r e t h e p o r k a n d b r e a
d a s b e f o r e a n d w h a t t
e n f o l d i n c r e a s e d m y
h o r r o r n o t a s i g n o f a
n y p r i s o n e r I c o u l d o
n l y j u d g e t h a t a l l h a
d p e r i s h e d a n d m y h e a
r t s m o t e m e s o r e l y t h
a t I h a d n o t b e e n t h e r
e t o p e r i s h w i t h t h e m
