| Viewing Single Post From: The Index Of Coincidence - The Chi Test, The Kappa | |
|---|---|
| jdege | May 30 2008, 04:49 PM |
Elite member
  ![]](http://209.85.122.85/static/1/pip_r.png)
|
It's not a very general technique, though. Using chi to match the different alphabets to each other depends on being able to shift the ciphertext component of the alphabet. Which means you know the sequence of the ciphertext components. Which you do in Vig, Beaufort, Variant, and Quag I, but which you don't in Gronsfeld or Quag II-IV. Something I noticed, last night, reading MILCRYPT I. I'd skimmed it before, skipping over the sections on standard and reverse standard alphabets as being too obvious to be worth wasting time on. This time, I actually read those sections. They discussed running down the alphabets - taking the first ten letters of the ciphertext, and the writing the successive letters below until the plaintext appeared:
Then they discussed doing the same for a reverse alphabet - by converting the ciphertext using any arbitrary reverse alphabet. The result will be encrypted with some standard alphabet, which can then be walked down in the normal manner. Which got me thinking - wouldn't this work for any known cipher component sequence? Porta uses a non-standard alphabet, so it can't be shifted in the normal manner. But it always uses the same sequence. Would simply decrypting with any Porta sequence result in standard cipher component sequences that can be chi-tested? Would it be possible, with the Gronsfeld and the Quag II, to reconstruct enough of the cipher component sequence, to do a trial decryption, and to see if the result is close enough to standard for it to be chi-tested? |
| When cryptography is outlawed, bayl bhgynjf jvyy unir cevinpl. | |
 |
|
| The Index Of Coincidence - The Chi Test, The Kappa · General | |
3:44 PM Nov 27
