I will try to match the frequency profile of the 340 as closely as possible and see if that changes things.

Virtually the same.

Take the following with a grain of salt but my testing revealed that the sequential randomness in the first half of the 340 is about 15% and about 35% in the second half.

Here’s a 170 character cipher with a normal plain text and about 15% sequential encoding randomness:

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

JO,XPG.BW*[E4>=D
]0(/B1'8Z[^%$?OQ(
KN<B;1SNM>UY1&6>[
AXB!@"4&CN0+.#12#
(8$C@XV*L1?AHIS6D
]<TNO3+1[K7)FNBRP
XEN*A1[W>B(IT@PD
ZNGOM]150=8+[IBQ
X!NG*2'198?-<6>[;
W(/%#D@4^L>@N:1KI

I like that and am pondering symbol selection as some type of function. I think it may be a function of something that causes the pivots, but I don’t know what exactly. A function of previous symbols in certain relative positions and the letter to be encoded? It is just interesting to me that the first repeated symbol is at position 19, and the pivot nodes, if you trace back at P39 you get to position 1. Sorry that I don’t have anything better than some ponderings right now. I am stumped after failing to produce two pivots with any cipher, but still trying to think of something.

Excuse my language but basically the pivots are the most retarded observation in a cipher, ever. I would still like to say that there is a 50% chance that Zodiac put these in on purpose, it sort of lines up with what he did in the last line of the 408. But then the whole period 39 thing makes up the other 50% in where it could be a genuine encoding artifact of some sort. It is unreal how perfectly twisted this is.

Language excused and it is perfectly twisted.

I’ve tried several different ways to create a cipher on the fly to come up with something somewhat close in structure to the z340.

My current "best" algorithm goes like this.

I set a cap at 3 symbols per letter.
I have four colums. The first with the letter and then three columns for symbols matching that letter.
If a letter got less than three symbols I create a new symbol in the next column and use it.
If a letter has three letters I use the letter in the next column in relation to a global column postion. If the column is empty I go back to the first column. I then update the global position. Same position for all letters.

For my tests this gives the worst results with the tools but the ciphers are still more or less decrypted. Seems a bit harder if I reduce the number of letters too.