I have an encrypted string based on an "alphabet" of 29 runes and access to successfully decrypted (runes->English) plaintext. Testing for a Vigenere cipher, I tried a Friedman attack (aka Kappa test, Index of coincidence test) to try and find a probable key length and obtained the following results for the index of coincidence (IoC):
- IoC of decrypted plaintext: 0.0616742 (c.f. 0.065 for plain english)
- IoC of encrypted text: 0.0332257
- IoC of random choice: 1/29=0.0344828 (c.f. 1/26 for plain english)
(Mostly adapting the procedure outlined in method 3 of https://www.it.uu.se/edu/course/homepage/security/vt08/labs/lab1.pdf)
As you can see, the encrypted text has a lower IoC than random choice which I suspect is due to an engineered bias against doubled runes in the encrypted text.
Is this indicative that a Vigenere cipher was not used? Is such a bias characteristic of any other classical ciphers?
(Note: the IoC above were obtained from a string of ~250 encrypted runes and ~850 decrypted runes from a separate plaintext)