Ran autokey with all words in 173K list as priming key. Using KRYPTOS alphabets as in sections 1 & 2. Grading results on occurence of singles, digrams and trigrams. Repeated with plain alphabets. No interesting results.

Message 77 Tried a ciphertext autokey: cipher(k) = cipher(k-1) + plaintext(k). (David Wilson)

Ragbaby

Running key

Gromark

{Scryer, 2 June 2004} This is an interesting possibility because Scheidt wanted to use systems that would not give away state secrets, but still be challenging. An article on the Gromark (Gronsfeld with running key) was published in Cryptologia shortly before Kryptos was designed -- since Cryptologia is unclassified, there'd be no question of giving away secrets. Without the n-digit primer Gromark can be quite challenging but not impossible to decrypt; however, if some of the digits we see in the other decryptions are the primer, then it would be considerably easier. We'd just need to identify which digits those are.

I've tried shotgun hillclimbing attacks with a broad range of primers, with no luck.

Arguing against this hypothesis is that Gromark does not exhibit the one clear (at least to me) statistical "phenomenon" in K4: the doubled letters appearing at multiples of 7.

Lorenz Cipher

http://www.codesandciphers.org.uk/lorenz/

One-Time Pad

(Elonka) Okay, so how about we come up with a sample solution and pad, and see what Sanborn says? :)