So I made this one up to see if it would work.  Keep in mind, there’s not much stego with text because it was typically assumed that once you had what looked like enciphered text then it was the same as just having the ciphertext.  Au contraire mon frere, it must be possible to hide a plaintext or ciphertext within a section of text, even if it looks like ciphertext and the masking text will thwart any efforts to solve it until it has been removed.

Now we just have to come up with logical methods that could have been used on K4.

This one I like to call “The Every-Other Null Cipher”.  Let it be known throughout the land that I, KryptosFan, have invented this one.  The idea is simple.  Every other letter is important and every other letter besides the important ones is null text.  Now by itself, it is not a very strong system:

Kqrxywpjtmovsyiksgtzoxonhpaarid…well, it’s not very clear right off the bat but look again:

Kqrxywpjtmovsyiksgtzoxonhpaarid

or “Kryptos is too hard”.

The beauty of The Every-Other Null Cipher or EONC is that you can pick whatever letters you want for the in-between bits.  I’d recommend looking at the frequency of your plaintext and then deliberately evening them out with some q’s, x’s and z’s but don’t forget to make it a little random and put some doubles in.  An even better solution would be to pull letters out of a hat (or use a random letter generator).  Even though my brief, un-bolded example was a little bit gibberish, okayyyy, it was a lot of bit gibberish, it’s still not secure enough.  You get some dreamy-eyed cryptoanalyst and they’ll inevitably see the pattern.  So how do you make it harder?  Simply encrypt your message with the method of choice from any of about a million-jillion substitutions or transpositions and then apply the Every-Other Null Cipher to make it just a lot harder.  You can even change whether it starts with the first or second letter of the final string.

Sha-zaam!  We just came up with a previously undescribed, still very logical and easy to use masking method that is totally possible to help unlock K4.  So, what did I get when I tried it?

Starting with the O:
OKUXGUBOIBWLVQRGSOWQJSEZWTKUIWNBYVTZPWDZTCIKHAECR

Starting with the B:
BROOHLSLFBFRQPNKSTTSQSKZAJLDAIFNPTMFKGKXJDGUUUKA

I did this with pencil and paper last night so excuse me if I may have made a mistake.  I dont’ see much that would help us straight away but that is pretty much to be expected if someone actually wanted this to be hard.  Two things come to mind: are they substitution ciphers possibly and since the O-string has 49 characters, could it be some kind of matrix transposition?

For my analysis, I used Decrypto 8.5 in Patristocrat mode.  It’s free and online so why not?

Starting with the O:
YH EMBERY PROVING BAYOND ALSO THE POUR WITSKOFSTJPHQZLJG
Possible real plaintext:
“blah Embery proving beyond also the four blah”

Starting with the B:
YOU UNDEDLY LOCAPIERRECEIVFHDTFWL PARK LIMIGHTM XXXI F
Possible real plaintext:
“You undoubtedly/undecidedly locate/located Pierre/Received blah park limit blah XXXI

Well good sirs, it seems we are left with some tasty morsels but the method is still imperfect.  Perhaps a little googling of the phrases perchance may reveal their origin, unless of course it is an original Sanbornism.

Google Results:
Not much for the O-string
Not much for the other, either

C’est la vie, ne c’est pa?

Well, the Every-Other Null Cipher showed promise for a hidden substitution cipher but no solution.

What about ye olde transposition potential of the O-string?

Well, here’s what they look like in rows of 7×7:

O-K-U-X-G-U-B
O-I-B-W-L-V-Q
R-G-S-O-W-Q-J
S-E-Z-W-T-K-U
I-W-N-B-Y-V-T
Z-P-W-D-Z-T-C
I-K-H-A-E-C-R

I looked it over a couple times trying some different route tranpositions and even considered the possibility of keyed-columnar but it just didn’t seem that likely.  Not likely enough to consider trying all of the possibilities or double transpositions.

Unmask with the Every-Other Null Cipher and then run a Substitution analysis and then transpose you ask?  Great idea but if you briefly consider how to add the sort-of plaintext I was able to get, it doesn’t seem like it would fix the gobbledygook.

Final Conclusion?
A worthy first attempt.  The fact that you could get any sort of plaintext out of the text sort of implies that while it’s not as simple as every other, it is perhaps possible to use the English words we found as clues to perhaps the real orientation or extraction.

Possibly…

It’s at least worth a shot.

Kryptos Fan