MY 340

[ALisowsky Profile]

Hey guys,

In an effort to improve ZKD and potentially test future orientations methods/tools, etc, I decided to create a 340 character (63 unique symbol) cipher that matches nearly exactly the symbol frequency count of the z340 (in other words, there are 24 + symbols, 12 B symbols, etc etc). The difference is, for this one we know the answer

I have written it using doranchak's ASCII representations... but I have not yet had a chance to run it through ZKD to see how it solves. Anyway, here it is:

(;+Hc^BE&pPN5FB*U
cYW+T2b4FD<OK+RyT
+I+JjR8I1C3#LBHlB
*F/^2W.zGO7LBV4>K
O<X+;y31(d)MFzW/M
_Sj+lpfUC)l.DF+TI
zOJ*zHVI6zP.2A+U*
GZ>+(&1#cJkNpvB2V
BFOORIw+Z>Bz68@Y9
pTB2C+#lcq5M*RZMI
p<b+EUckC+l.zFC5I
k#cG25dp_-^Rct+z7
J8M2(#Gc^pMpFHOR-
+D2K94bBI)L^5tLpd
WpAFpYlOk+NRyKYBR
9k4N.^(D+%dOZ<;*-
K6<(t+z7S+GFcOL5.
SV_Lqly%G:y)f2+T-
(f+I-4XBKd5<9:)t+
S8fN+>WI4MEPcUV/K


And here's the frequency table:

Please let me know if you guys need any further info about it!

Adam

[doranchak Profile Website]

Why wait until then? Just go to Subway.

[AuthUser Profile]

AL,

Thanks for making this cipher. It's always nice to have something that can be solved.
It took about 30 seconds with ZKD.

There are other tests ciphers included in the ZKD package that you might be interested
in looking at as well.

I have the solution, but I'll wait until later to post it.

Regards,
Auth

Edit: I should add that ZKD did not solve the cipher completely. It was solved to about
90-95% complete, and I entered the remainder of the characters by hand.

[rod Profile]

ZKD did not entirely solve it for me. ZKD solved part of it and then I tried to correct it.
I'm having some difficulties at line 7, i'm getting "TOMATCHSOTWATEECA"... did you add some errors into the cipher to make it zodiac like?

[glurk Profile Website]

doranchak-

I like how your comment only makes sense after solving the thing, LOL!

rod-

You will find that there are some symbols (not many) that represent more than one letter, as Zodiac did in the 408, probably intentionally.

-glurk

P.S. - Note that ZKD was never intended to be a press-it-and-forget-it cipher solving panacea. It's just a tool meant to be used interactively to help (a lot) in solving homophonic ciphers. I'd be surprised if it ever found a fully 100% solution to a cipher, unless it was a very easy one. But that was never the point. Once you get to know the program and the tricks that you can use, it's really quite powerful.

As Auth noted, there are quite a few example ciphers included with ZKD. I have the key/solution to all of them, but purposely did not include them with the program. They range from fairly easy to very, very difficult. 340.tonyb2.example.txt is probably the most difficult, but it is solvable.

[ALisowsky Profile]

Ok, next step for those of you willing!

1. We know MY340 is easily picked apart by ZKD... next I've encoded the same plaintext with a vigenere:

Bkmpowllgupyqgktjq
cnxqpggmjtbyxskekll
uorxzbukhumpxxhbrjm
oigthvvrntpabrolllkm
pentmaprtntjrxhmgug
khkgkkzbbstakgyhapn
zplkntnzmvzmvqzvkvd
rzalbbuezerallbbkqwtb
xxaprukpovtghthtlvtik
wjqcnxyigzapacububv
chbtqrbrmguflvgohuqq
ugvbpuesmpzlsisklhvqo
duwjnhdbbvkvxrxefacke
sxnxtkqpkbdwhrwhtfuep
srzhtmazbvvvgfocamkf
qasrhngkksqsklsiikldqcr
fhsrsxmzryaziajppkukl

The key should be fairly obvious. If could: test drive the above and post the IoC freq to ensure we have a vigenere sig showing up? Did ZDK solve it?

2. Next, I have encoded the vigenere with a 63-char homophonic substitution: (interesting side note I did not think through: when using homophones AFTER vignere, due to more evenly spaced char freq, each letter on average has less homophones)
And here is the cipher:

B+LR^1.4pK()kp+J6q8IyM(_pYfOB)zV%;+
*4G^DyWBS+XKYEzy2NFfLA:pO<c5DIH
(dBD^*.4+ZRCIOLTEFHIJf>z2LpKp+<+_
++WNBlOU+j)2TRIb(.+IJIWZcWL5kW
c&P#>Wd.9BGCWCFT.*BB+k1ONzzU(
DK+EAcOp2H2O*cH-+/6M8Iy):jWTRd
7GBKBc729OMFN>ZpGt4P_^2SkMGpc
9RG;lY(W*V-V+4<PM^#S16I2#BB5%
czFzCtU8+;lzIyO+q(+N#/<F/2OtSCRlF
bXJZUW95ccp3^7dL&3MUV><@p++lk
V+.l--+*#M8Ft<VFlzYWF)Tb-dfRR+K+4

Could someone again please: see if it is solved by ZDK and post the IoC to see if after the nasty homophonic sub is added, if we still get a signature.

3. Ok, the fun part. Using the key below, on the cipher encoded with both vigenere AND homophonic (the cipher from step 2), merge many of the homophones together and re-run/post the IoC. For the sake of 'research', woudl you keep track of how many mergers you made before running the IoC?



(note: I created this key in Excel.. therefore, the *minus* means -, the *plus* means +, etc. Due to the default font used, the lower case L and upper case I look the same. The lower case L is represented in the key as l (L))

THE PURPOSE: If Z decided to encode with a shift prior to homophonic sub, can correct homophone merging at least reveal the vig/caeser shift?
Thanks in advance for anyone taking the time to do this, my notebook with ZKD is currently in repairs and I am on a work computer which can't download programs.

Adam

[doranchak Profile Website]

The key should be fairly obvious. If could: test drive the above and post the IoC freq to ensure we have a vigenere sig showing up? Did ZDK solve it?

ZDK wouldn't have any problem with it. But I easily cracked it with this:

http://smurfoncrack.com/pygenere/pygenere.php

[Zabagliona Profile]

The name of that one alone makes it worth using!!!!

[doranchak Profile Website]

The "nasty" vigenere+homophonic encoding still doesn't mask the keylength which clearly appears in the IoC scores.

First, the IoC counts for the original plain homophonic:



IoC counts for the Vigenere:



IoC counts for the Vigenere+Homophonic:



The dual-encoding doesn't hide the IoC peak intervals. It only suppresses the expected IoC scores for English plaintext (which is around 0.0667). I didn't try to make ZKD solve the cipher because I think the heuristic won't work since it is expecting only the single homophonic encoding.

By contrast, the IoC scores for the Zodiac's 340 are quite flat:



I'm curious to know if it is possible to construct a valid Vigenere+Homophonic cipher that thoroughly suppresses the IoC peaks AND has the same multiplicity as Zodiac's 340.

[glurk Profile Website]

I'm curious as to why so many people keep thinking that out of the MANY cipher types, Zodiac used Vigenère! If there is some reason people keep focusing on Vigenère, well, it confuses me...

-glurk