Zodiackillersite.com

I was commuting to work this morning and had an idea. Draft a 338 letter message LRTB into a 26 x 13 rectangle, and encode with a 6 letter long Vigenere key LRTB. Then read off the letters TBLR and transcribe { EDIT: LRTB } into a 17 x 20 rectangle. Then encode with an inefficient homophonic key. This should produce period 78 unigram repeats. I said before that I couldn't see how it would create period 19 bigram repeats, earlier in this thread. But I wonder if generating a lot of messages with a homophonic key that maps all 26 letters would create more frequent false positive period x bigram repeat spikes.

EDIT: I counted 2173 period 6 unigram repeats in all 100 of Jarlve's plaintext messages. So that is an average of 22 for each message. The letter E is most frequent, coming in at 502 of the 2173.

Perhaps some variation of this could create a lot of period 39 repeats and pivots. Maybe with a certain type of six letter keyword.

Then untransposing the TBLR would reveal a keyword length of 6, not? Have you checked that?

I am going to right now.

Here is the 340 untransposed LRTB at period 13.

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

It moved the spike to a period of 6. There are not a lot of other little spikes at multiples of 6 though, except period 72.

Hey smokie,

I have updated AZdecrypt (1.04) and its thread with a vigenère solver available for download. You should be able to explore the hypothesis now.