By the way, I started a new page to collect AZdecrypt "tips and tricks" here: http://zodiackillerciphers.com/wiki/ind … and_tricks
I want to add more stuff to it, like examples of ways to solve other cipher types, and other instructions and info you’ve given in other forum posts. Can you recommend some other of your forum posts for me to link to from there?

I’ve noticed, thanks.

I wonder if you will need to adjust the solver to take the fixed N values into account.
Do you want me to post the plaintext, and the plaintext with modified shift applied for that last cipher, just to rule out any mistakes in encipherment?

I prefer not having the plain texts right now. I want to make a project out of solving your ciphers but may not have the time for it right now since I am gaming at the moment. Next month there will be a new World of Warcraft expansion which may keep me busy until next year. Just saying that I may not have the time to tackle this project right now.

My apologies, I was messing around with Substitution + sparse polyalphabetism with 8-grams_english_beijinghouse_v6.txt loaded without adjusting any settings.

As you have stated, the larger amount of freedom generates more words but clearly not Z speak:

Oh, I see, that solver works great without the need to alter settings but it also needs to run for a very long time for harder problems. What are the specs of your PC? CPU & RAM?

Allot of clean words indeed.

I prefer not having the plain texts right now. I want to make a project out of solving your ciphers but may not have the time for it right now since I am gaming at the moment. Next month there will be a new World of Warcraft expansion which may keep me busy until next year. Just saying that I may not have the time to tackle this project right now.

Sounds like a fun break! Enjoy yourself.
I’ve been feeling the urge to start gaming again too.

Oh, I see, that solver works great without the need to alter settings but it also needs to run for a very long time for harder problems. What are the specs of your PC? CPU & RAM?

Allot of clean words indeed.

Nothing impressive CPU wise, been needing to upgrade for quite some time just can’t seem to decide which way to go.

Processor AMD Phenom(tm) 11 X4 965 Processor, 3400 Mhz, 4 Core(s), 4 Logical Processors

16 GB of ram Windows 10 x64

ASUSTeK M4A79XTD EVO

As you can see, in need of a serious upgrade

Solve hasn’t come through yet, will need to investigate what the problem is.

@doranchak, half the planet is gaming nowadays.

@4on4off, it was a very good PC back then. Still okay if not doing heavy tasks I guess.

Solve hasn’t come through yet, will need to investigate what the problem is.

I encoded the plain text doranchak gave me with homophonic substitution and here are the solve times:

Solver: substitution + polyphones [user] using 2 letters per symbol and Louie’s 8-grams

26 symbols (no homophonic substitution): 0 seconds
35 symbols: 7 seconds
44 symbols: 45 seconds
53 symbols: 12 minutes
63 symbols: 8 minutes

So at least it is solvable and within a reasonable time.

@doranchak, is it normal homophonic substitution or are there also errors introduced etc?

Solve hasn’t come through yet, will need to investigate what the problem is.
@doranchak, half the planet is gaming nowadays.

True. I’ve been solving sudokus on my phone almost every day, maybe that counts.

Solve hasn’t come through yet, will need to investigate what the problem is.

I encoded the plain text doranchak gave me with homophonic substitution and here are the solve times:

Solver: substitution + polyphones [user] using 2 letters per symbol and Louie’s 8-grams

26 symbols (no homophonic substitution): 0 seconds
35 symbols: 7 seconds
44 symbols: 45 seconds
53 symbols: 12 minutes
63 symbols: 8 minutes

So at least it is solvable and within a reasonable time.

@doranchak, is it normal homophonic substitution or are there also errors introduced etc?

It should be normal. Let me check. Here’s the shifted plaintext again:

QLILCWQEBPLUBIPBE
RIBVPRPWYBZXQLKWK
HJILOQKLPMXUSLVBY
LGHCLDIODPHVCLOSF
FQROBVLOOLRNLQJJI
XVPBVWKHVHDFSHXZL
ABODKDBQLQEHZLRAF
XUSBUIRUXIIEROAHU
PXOHPRLWXEOBWRCUD
JHPKBDAPJDVYBDSSO
LTXWRFROQBUPXQAFH
QWUHPLCIODJHVLQFP
YBUVPXFKQLYBGHPFU
HGWEDWGBPFDKBOPXQ
AFDUYHUPZLXIABABU
WQEBFUFKYHQQLYBKH
PVXKGBKGBXSLXOQRY
OBDNRSQKHJLKRQRKB
XKGIBBYIHQHVVZEFZ
KZEDUXFQHULWHPRPQ

When I put it in AZdecrypt with your settings, the solution it finds has a few wrong letters but overall is about 93% accurate.

Here is the homophonic encipherment of the shifted plaintext again:

y2_HfpyY|LH8|_LKY
>_MUL>L7+|&S:Hlcl
P9_V#:q;R3S4%2UK@
VE*f2(b1WR*Uf;X<O
Oy/DKj2##H/d;y99_
5jB|U7qTUP(O%JA&V
NK1Wq(MyVtYZ&H/GO
S8<M8_/8^_bk/#NT4
L^DJB>Hc5Y#F->f4W
9JLlMWGB9WU@F(<%D
26A-/O>D:F8RStGOJ
t78*B2f_D(9Pj;:OB
@|4UB5Olt;@FCPLO4
*.7Y(7E|LO(zM#R5y
NOW8+J8L&HSbG|GM4
-yYFO8Ol+P:t;+Kl*
RUSlEMzCKA);51:>+
#F(d><tzP9;l>:/lF
AzCbK|+bJ:*UU&kO&
l&YW8AO:T4H-ZB>Bt

To double check, I derived symbol-to-plaintext assignments from the above cipher text and shifted plaintext, and got this key (including assignment counts):

OF  15 
lK  10 
UV  10 
8U  10 
LP   9 
>R   9 
:Q   9 
|B   8 
_I   8 
HL   8 
BP   8 
;L   8 
(D   8 
yQ   7 
tQ   7 
YE   7 
WD   7 
MB   7 
KB   7 
FB   7 
9J   7 
4U   7 
/R   7 
&Z   7 
#O   7 
SX   6 
PH   6 
JH   6 
2L   6 
+Y   6 
*H   6 
fC   5 
bI   5 
RP   5 
GA   5 
DO   5 
AX   5 
7W   5 
5X   5 
zK   4 
VL   4 
@Y   4 
<S   4 
-W   4 
qK   3 
jV   3 
TH   3 
NA   3 
EG   3 
CG   3 
1O   3 
%S   3 
kE   2 
dN   2 
cW   2 
^X   2 
ZH   2 
pW   1 
XO   1 
6T   1 
3M   1 
.G   1 
)S   1 

Here are the homophones for each plaintext letter:

A GN
B FKM|
C f
D (W
E Yk
F O
G .CE
H *JPTZ
I _b
J 9
K lqz
L 2;HV
M 3
N d
O #1DX
P BLR
Q :ty
R />
S %)<
T 6
U 48
V Uj
W -7cp
X 5AS^
Y +@
Z &

As far as I can tell, I don’t see anything irregular with the encipherment.

Thus, a proper approach would be the change to solver to have fixed N values, with an additional hill-climber on top to figure out the values of N.

I found some time to implement this approach. Running all ciphers through with 6-grams though it seems that 5-grams may be enough.

Here’s one solve and will share more later:

d;y+.Ec5@<N>3;.7P
7Dtd||y:7TF2q>):8
<5UjkYW21S+6^D-^8
_/)L.Nk.5X);.P*DX
<EGRz.ONEK+J9yYVK
7;YcRpX75p/yl.<:b
*;9l/1P1>L.qLN.jD
_MzR+U#F8HO3.#(YE
-19y^Gz.NjYXRW3(5
&(5<D)):P1f%*qETf
OZNSC1Vt+2<+Dj/XA
B8dW)N8c<(WR.//N@
Yd%f/YB.U(L.jMl.1
7%d51M_R-X+1lR1D@
(KV9j(/#/-4^5&WtO
9fV97;&RF/1V>y6)Y
7l.U6N7KK/Jp%VS<V
y:WS*WddFK/1^UT3q
*(ydqXp/NE%3W).P*
GXW8G.+HN><TB)NVK

Score: 22939.04 IOC: 0.0635 Multiplicity: 0.2764 Minutes: 18.51
Repeats: MEDICAL THEYARE ANDARE OUTOF TTHEY ANCE (2) THE (2)
PC-cycles: 197

Symbols: 94

SAMEFOLKMETHISFED
ERHAPPENEDBYCHANC
ECLEARLYTHEYAREAC
QUAINTANCESANDARE
WORKINGTOGETHERTO
WARDSOMECOMMONENV
ISHOULDTHINKITFER
YLIKELYJUDGINGFRO
MTHEIRINTERESTINC
ANCERSANDTOXICOLO
GYTHATTHEYWEREMED
ICALSTUDENTSNUMBE
RSFOURIFLFINELOFT
EXACTLYSEEMTOSTRE
NGTHENMYMEDICALHY
POTHESASBUTTHEYAR
EONLYTWOOUTOFTHET
ENTHATSABOUTALLIC
ANMAKEOUTOFITANDA
RETURNEDTHELISTTO

I found some time to implement this approach. Running all ciphers through with 6-grams though it seems that 5-grams may be enough.
Here’s one solve and will share more later

Nice work, Jarlve! Did you already run it on Z340 to rule out this scheme for it?

a very close solve and easy google find from The Darrow Enigma,
not a lot wrong with it ..

I found some time to implement this approach. Running all ciphers through with 6-grams though it seems that 5-grams may be enough.
Here’s one solve and will share more later

Nice work, Jarlve! Did you already run it on Z340 to rule out this scheme for it?

No, still working on cracking all 10 of your test ciphers. Some of them have resisted and the solver will probably need some additional optimization. Since it is a lightly modified version of the polyphones solver I hope to push the optimizations (if they will work) to that solver as well. The polyphones solver is from late 2015 and pretty much remained unchanged until now, so I hope to give it a nice update.

Do you have a name for this encryption? It makes me think of Vigenère with a 2 letter keyword applied aperiodically/randomly instead.

I wonder if there are some unbroken ciphers out there which have this scheme or a variation thereof.

No, still working on cracking all 10 of your test ciphers. Some of them have resisted and the solver will probably need some additional optimization. Since it is a lightly modified version of the polyphones solver I hope to push the optimizations (if they will work) to that solver as well. The polyphones solver is from late 2015 and pretty much remained unchanged until now, so I hope to give it a nice update.

That’s great! I look forward to those updates.

Do you have a name for this encryption? It makes me think of Vigenère with a 2 letter keyword applied aperiodically/randomly instead.

I called it "arbitrary shift cipher" or "arbitrary caesar shift cipher" but that may not adequately capture the fixed nature of the shift amount.

"Random direction Caesar cipher"?

If you made Vigenere that had random directions, it would be rather evil, don’t you think? For example, given the key (2,6,10,4), you could shift ahead or behind by each value, instead of just in one direction.

The other variation would be to randomly toggle between two values altogether. For example, instead of randomly picking between -2 and +2, you could pick between -3 and +5.

I wonder if there are some unbroken ciphers out there which have this scheme or a variation thereof.

That would be fascinating to discover. Perhaps it is already documented somewhere and we just don’t know how to search for it.

That’s allot of words. Will probably call it "random shifts" of "fixed shifts" or just "shifts" for AZdecrypt.

The other variation would be to randomly toggle between two values altogether. For example, instead of randomly picking between -2 and +2, you could pick between -3 and +5.

Oh. So your test ciphers are all of the form -2 +2 or -3 +3 or -4 +4 etc? My hill-climber currently looks for all combinations, such as -3 +5, except for those where both shifts point to the same letter.

That would be fascinating to discover.

I thought of Kryptos 4 and the Feynman ciphers, but perhaps with more than 2 shifts. When the new solver is optimized I want to try it.

That’s allot of words. Will probably call it "random shifts" of "fixed shifts" or just "shifts" for AZdecrypt.

You could call it "Hafer shifts", since a man named Charles Hafer came up with the scheme for his diary entries. Bob Bogart of the ACA discovered them and I think he is not aware of any other usage of the scheme (he’s a crypto enthusiast and also works for the NSA).

The other variation would be to randomly toggle between two values altogether. For example, instead of randomly picking between -2 and +2, you could pick between -3 and +5.

Oh. So your test ciphers are all of the form -2 +2 or -3 +3 or -4 +4 etc? My hill-climber currently looks for all combinations, such as -3 +5, except for those where both shifts point to the same letter.

Correct – I picked a random fixed shift value N, and then while encoding, picked -N or +N at random.

That would be fascinating to discover.

I thought of Kryptos 4 and the Feynman ciphers, but perhaps with more than 2 shifts. When the new solver is optimized I want to try it.

Cool – good luck!

Here’s a thing I threw together quick to encode these types. I think it works, LOL!

http://bardstowncable.net/~xenex/hsc/

-glurk

Here’s a thing I threw together quick to encode these types. I think it works, LOL!

http://bardstowncable.net/~xenex/hsc/

-glurk

That’s really cool, glurk – thanks for posting that!