Removing the 2nd line from the 340 increases the score more than other lines. I’m totally not sure if it could be related to why the 340 doesn’t solve but perhaps filler is inserted here and there. I would guess not per line since I’ve checked up to 6 lines deep.

By the way, these images are mesmerizing!

Edit: please scroll down or go to the next page to see the update pictures.

340:
http://i62.tinypic.com/312af5i.png

408:
http://i57.tinypic.com/v8iiwo.png

smokie2:
http://i58.tinypic.com/2vtcd1u.png

Jarlve . Could he have placed filler in all of column 1 or column 17 Going down, before he even started and then a small amount at the end to round it off.

Jarlve, that is really very interesting. It looks like row 2 has a spot where words infrequently appear. And row 16. It looks like some of the symbols may be associated with dead zones. That is really cool. I wonder how hot spot and dead zone statistics will compare to other statistics. I will get back to you soon with answers to all of your questions!

It is also interesting that some dead zones of the 408 correspond to areas that have misspellings!

I also added smokie’s 2nd wildcard cipher which kinda solves without expanding any wildcards, it may be useful to see if there is any relation. If yes, we may be able to pick more likely wildcard candidates.

Jarlve. Could he have placed filler in all of column 1 or column 17 Going down, before he even started and then a small amount at the end to round it off.

I tried it but nothing came up. By the way, there are always 4 black spots at the end of every cipher because of the 5-grams so ignore the color of the 4 last characters.

Could you give some specific information about the hobbit cipher also, perfect cycles? 1:1 substitutes? wildcard symbols number? Thanks!

See the edited 8th post at: viewtopic.php?f=81&t=2617&start=110 for hobbit cycles (many are perfect), 1:1 substitutes, and wildcard numbers.
Answers to other questions coming soon.

Does the cell shading for the hot spots and dead zones track the starting position of the words? What about every position. I mean you could have a variable array of x(340), and for every word just add one to all of the corresponding letter positions. I am saying because it looks like that with Smokie2: The Hobbit message. Some of the masked bigrams appear in your cell shading, but offset by a few spaces. Some do not show up at all; there are false dead zones. I will have to look closer; maybe this is a break between two words, a low frequency symbol or letter, who knows? I noticed that the 408 has a dead zone for the letter "x" in "experence." Can this new approach help us find high and low frequency letters?

This new approach is exciting to say the least, and may tell us something about the 340!

The picture now shows the score per symbol? . . . What do you think about symbols 7 and 54?

The picture shows how many two symbol cycles for each symbol.

For instance, with symbol 7, the highest scoring two symbol cycle is 2048, or an ABABABABABAB. If I randomize the 340 ten times, find the mean number of 2048 cycles and the standard deviation, and then add those together, I get 0.6. The 340 has two 2048 cycles, and symbol 7 is in one of them. In other words, it is possible that symol 7 is a wildcard, but not probable because of the high scoring cycle that it is in. And see the entire column for symbol 7. It is just in a lot of cycles with other symbols.

Try symbol 54 if you want to. That symbol is in five two symbol cycles scoring 64, or an ABABABA. Odds are, however, that one or two of those cycles are true. 5-(47.3/72*5)=1.72. Roughly speaking.

I have found that count correlates with number of cycles found. A high count 1:1 can have a lot of small false cycles with the other 62 symbols.

.

Does the cell shading for the hot spots and dead zones track the starting position of the words? What about every position. I mean you could have a variable array of x(340), and for every word just add one to all of the corresponding letter positions. I am saying because it looks like that with Smokie2: The Hobbit message. Some of the masked bigrams appear in your cell shading, but offset by a few spaces.

I did as you asked (add to all of the corresponding letter positions) and removed the offset. But I kept in the original system to give some extra weight because otherwise "experence" in the 408 became almost totally blue. It’s looking very nice now!

Edit: for the 408 you can see that it’s no longer the letter "x" but now the letter "e" which is supposed to be the letter "i" in "experience". Like it should be.

340:

408:

smokie2:

I have found that count correlates with number of cycles found. A high count 1:1 can have a lot of small false cycles with the other 62 symbols.

Perhaps normalize by count? The reason why I asked about 7 and 54 is because I’m currently running a test for nulls and these symbols along with 19 are very often in the top results.

That looks totally awesome. I love it!!!

Here is the new version of Smokie 2 next to the wildcards.

Perhaps normalize by count? The reason why I asked about 7 and 54 is because I’m currently running a test for nulls and these symbols along with 19 are very often in the top results.

Here are the symbols scored by dividing the number of consecutive alternations by the total count of the two symbols. Sorted by score from high to low on the left; sorted by count / score on the right.

Symbols 7 and 54 are both cyclic. Some of it could be random, but not likely all.

I also added smokie’s 2nd wildcard cipher which kinda solves without expanding any wildcards, it may be useful to see if there is any relation. If yes, we may be able to pick more likely wildcard candidates.

Jarlve. Could he have placed filler in all of column 1 or column 17 Going down, before he even started and then a small amount at the end to round it off.

I tried it but nothing came up. By the way, there are always 4 black spots at the end of every cipher because of the 5-grams so ignore the color of the 4 last characters.

Thanks jarlve for answering all my questions in such a polite manner.. I hope you and the team of guru cipher stuff crack this ..
I have another thought. I am at jfk airport about to embark on a 24+ hr journey home.. I will put it to you after I have checked it out further .. It’s just an idea how he may have constructed the 340.. Cheers

The side by side comparison shows a high correlation for the hobbit message, especially on lines where several wildcards are grouped. Though it is not unexpected since the message solves without expanding any wildcards. With the purple haze message, which doesn’t solve without expanding wildcards, the color map may therefore not correlate very well. Relating this back to the 340, the color map doesn’t seem to correlate well with any of the suggested wildcards. So the solve may be too far away from the solution. But there is a very dark long line going from the 2nd to the 3rd row, it is strange.

Yes, I agree. I was thinking that possible explanations for dark spots could include misspelled words, like doranchak pointed out, uncommon words, like "hobbit" or "nor," intentional gibberish (see the last dark line of the 408), or something else done to confuse a code breaker. I tried to compare the +’s, and it looks like they might be a bit on the dark side, on average. Also see the last two or three symbols, they are really dark. So I think that is probably filler. It’s not because it is at the end of the message. Otherwise, the first few symbols would be dark too. You are right, though, the second half of row two and first part of row three is really dark and strange. If Zodiac did that, then I am guessing that he did more of whatever he did there than anywhere else. Maybe he used a system to mask bigrams similar but different than the system that I tried. I mean, maybe the + is a signal to rearrange the symbols near the + somehow. Or maybe the symbols at the 2nd row dark spot do that. Do any of those symbols correlate with other areas that are comparatively dark? I spent a little while staring at it last night trying to correlate things, but didn’t get very far. If you want to try an experiment of any kind, let me know. I could make a message with a few misspelled words, uncommon words, low frequency letters, put a section of gibberish in there, and then mask a few bigrams in a pretty severe way. See if you can spot any of it. Or something else or nothing else. I am also thinking about the cycles.

If you want to try an experiment of any kind, let me know. I could make a message with a few misspelled words, uncommon words, low frequency letters, put a section of gibberish in there, and then mask a few bigrams in a pretty severe way.

By all means go ahead, make a mystery cipher if you please. By the way, I believe there is currently an error with the calculation of the last 4 squares. I’ll work on updating the images to correct this. If anything, intentional gibberish/filler is the best explanation for the line 2-3 thing. But I would like to compare these results to daikon’s solver with 6-grams when he gets back, perhaps it comes up with a different solve.