unforum

[mapmaker]

I've (still) been looking at this, and somehow just today noticed there was another repeated phrase, this one nine letters: FNAJXQPHK. Now if we were to still presume Vigenere, we've got a problem - the key would have to be two letters, assuming both OXGIVQFVER and FNAJXQPHK were repeated non-accidentally. *Sigh* All in all, still no luck. I'm almost positive the key can't be just two letters, BTW, because there's not much in the way of repeated trigrams, which one would expect for such a low key-length.

Map

[rowan]

[mapmaker]

Well, the concept as I understand it is that the most common reason phrases are repeated in Vigenere ciphers is that the same part of the key overlaps the same word, phrase, or subsection of a word.

In order for the keyword and the plaintext to be in sync, the key has to go through complete revolution before getting to the next plaintext. So if the two plaintext phrases were 15 letters apart, the key would have to be either 3, 5, or 15 letters.

So the distances between OXGIVQFVERs are 240 - many possible key lengths. But the distances between FNAJXQPHKs are 58 - which factors into 2 * 29. Only common factor is two - that's why I said there had to be only a key length of two.

[Abraxas]

[Animac]

Just tried all of the 676 combinations required for a two character key in Vigenere on the text as it appears. No result I'm afraid guys. But at points it appears like words could appear if the grids are put together differently. I just put them alongside each other, but maybe a LRLRLRLRLRLRLR approach would yield better results. But I can't try it right now as i am in work!

[Nalerenn]

I tried the LRLRLR thing, but gave up after a few lines. It was just random garbage I was getting. I may have done something wrong, though, so I'll work through the rest of it and post it up. Someone might get something out of it.

[Geekkake]

Has anyone considered attempting to decrypt the message top-down, right to left, as the Japanese read it? It is, after all, a Tokyo division memo. I don't have the ability to attempt it right now, being that I'm at work, but it's a thought.

[theShaggy]

You know what, I think you have something there. I don't know if it would make more sense layout-wise if approached that way, but it looks like maybe it would.

I just don't have any idea what kind of a key to use

[Geekkake]

[NewLurker]

No major breakthroughs here, but a few notes on the discussion so far:

*To those using rybya/orgyba as a key, it seems unlikely that the text--already decoded with rot-13 as 'subject: belton'--would be being re-used in its ciphertext form for a key at this latter point.

*The repeated strings OXGIVQFVER and FNAJXQPHK seem to preclude any possibility of this cipher being intended to be read any way other than forward or backwards with the two pages being counted as left and right aspects of the same code. (using any LRLR... combo, reading down columns, or employing columnar tranposition or playfair would break these repeated sections of ciphertext apart: the chances of these two strings appearing by coincidence is very minimal.)

*mapmaker, your work on discovering possible key lengths is admirable, though I think you were right the first time when you said the repeat of OXGIVQFVER is at a 280 interval as opposed to the 240 you used in your later calculations. That said, I don't think this changes the result of the only common factor with 58 being 2. Ultimately this point is just a matter of pedanticism.

How can we procede from here? Are some colums of letters meant to be left out? Could a double keyed vig cipher give such repeats as are in evidence? Is it possible that this sheet is intended to be a key source rather than a cipher in itself (might OXGIVQFVER and FNAJXQPHK be keys to some other relevant cipher)?

As my name implies, I'm a lurker and I'm pretty new to this so feel free to cut me down appropriately.

[mapmaker]