Comment on No Transposition in K4 of Kryptos by Mr. Gallow

There is a distinct chance that transposition is used on a segment of known text to generate the key/pad. In general I’ve noticed that the keyed vigenere produces stronger Friedman IC/ kappa-plaintext values the closer the key gets to Sanborn’s alphabet. In fact most of the pads that I’ve generated that match the expected letter/digraph distribution when dragged averaged around 1.3 for the IC, and .045 for the KP. Not perfect by any means, but considering that most randomly generated pads produce an average IC of around 1 and KP of about .033, and K1-3 produce around a 1.7 IC and .066 KP that’s significant. Even only considering letter distribution seems to raise the values by a few points.
Some observations:
1. While Sanborn has stated that NYPVTT decodes to BERLIN, it’s worth noting that he hasn’t distinguished weather it’s plaintext or keytext… technically plaintext and ciphertext can decode keytext.
2. Most of the previous clues Jim released have pertained to mistakes or purposeful alterations to the text i.e. the extra L, xLAYERTWO, the Times attachement was released with reference to questions on typos. Who’s to say that this clue isn’t the same case… Sanborn, but he hasn’t.
3. The digraph frequencies of the plaintexts are quite a bit higher than that of standard English.
There are roughly 4 times the expected THs. While I know that Sanborn has stated that knowing the solutions of K1-3 are not required to solve K4, I do recall that Scheidt stated that the first three sections give you the alphabets.
4. Considering that steganograpy is statedly used likely means that the primary decryption wont have as high of a IC/KP since most cryptanalysts state that steganography comes most often before encryption.
5. KryptosFan has the best links.