Zodiackillersite.com

This thread will serve as a catalog for my work on the ciphers, in particular the unsolved Zodiac 340 cipher. Mostly everything will be made available through Google Documents in a new format with example visuals, clear descriptions, downloadable test ciphers and results. I do not wish to claim that any of my work is conclusive. I have no training in any scientific branch and my methods can be considered as experimental and self-thaught. My goal is to make contributions towards solving the Zodiac 340 cipher and to inspire others to do same.

Programs:

- AZdecrypt 0.992 BETA: fast automatic multi-threaded (homophonic) substitution (manipulation/transposition) solver for one or many ciphers.
- Examine: example of an experimental interactive (homophonic) substitution transposition solver.

My other work:

- Plaintext library: 100 handpicked plaintexts having 340 characters each and randomized counterparts with the same letter frequencies.
- A mathematical cipher? I attempt to mimic homophonic substitution with a mathematical approach and somewhat succeed.
- Sharp constrast is added to a high resolution image of the Zodiac 340 cipher in attempt to reveal how the symbols are constructed.
- The exact middle of each symbol in the Zodiac 340 and 408 ciphers is determined and plotted onto a map to reveal a few interesting things about how the cipher was possibly constructed physically.

Jarlve, thanks for doing this. Your writeups are very useful and informative. I will work on linking to these results via my wiki.

Thank you doranchak, it is much appreciated.

Jarlve wrote:and to inspire others to do same

Well, you inspire me to keep going and not give up on this thing.

Great job on your work so far! I still haven't tried your solver but I promise myself to do it soon.

_pi wrote:Well, you inspire me to keep going and not give up on this thing.

Great job on your work so far! I still haven't tried your solver but I promise myself to do it soon.

That's nice to hear!

Let me know If you have any issues with the program and I'll see what I can do.

Great work.

Thanks zydeco,

AZdecrypt095 is finished and available for download AZdecrypt095. I have not yet updated the documentation.

The solver has moved up to using 5-grams from practical cryptography. Thanks to daikon for making me aware of that.

The program has lost about half of it's speed but it's a much more powerful solver now. There seem to be some things that it excels at, for instance, reasonable short strings of ciphertext with not too many symbols. An example of this is given by a solve of the first 8 rows of the 408 (multiplicity 0.375). Which was achieved by changing the Ioc target weight to 2 and many attempts. It now supports up to 200 unique symbols and ciphertexts up to 10.000 characters.

Code: Select all

ilikekillingpeopl
ebecauseitissomuc
hfunitismorefurth
ankillingwillgame
intheforcestbecau
semaristhemoathan
gertueanamalofall
tokillsomethinggi


The program works great out of the box but a better solve can be made on some ciphers by changing the Ioc target weight to 2 or higher. A good example of such a cipher is daikon's 3rd cipher (included with the program).

Have fun with it and if there's anything, let me know.

Edit: broken links are fixed now.

Jarlve, would you be willing to share AZdecrypt's source code with me? I am a web developer and am very interested in porting it to a web application. Ideally, I would host the web app somewhere, and anyone could go to the web site and submit cryptograms for it to solve.

Great idea doranchak!

I added 2 new tests, row flips in a 17 by 20 and 20 by 17 grid. I also added some of my other work on this forum to the My other work section. I'm now starting work on 3 new tests, exploring the possibility of the cipher having a number of rows, columns or symbols as filler up to a given multiplicity limit. Quite an exciting prospect with my new solver!

Download AZdecrypt096.

I wasn't planning on releasing another version so soon but I needed some additional features for some of my upcoming tests.

- 10% to 15% faster.
- Removed alphabet support.
- About double the recovery rate for most ciphers.
- Fixed in and/or deflation of scores within the same solution.
- Pressing the "x" button to stop the program working is much more responsive.
- Changed the system "keys per ciphers" to "iterations per cipher", one iteration is about 300.000 keys.
- Changed the output file format to include multiplicity. Example: (15637_654_132_1.txt = score_ioc_multiplicity_number.txt).
- Normalization by multiplicity (not perfect, but will do for the multiplicty range the program is able to handle). There was too much score inflation for unsolved ciphers with high multiplicity.