Cryptology - Unofficial St. Mary's College of California Web Site

Chapter 6
Decrypting Monoalphabetic
Ciphers
We begin with the unsolved cipher from Section 5.5:
KNHHXKK QS PXTDQSB YQFJ NSISCYS HQEJXUK QK LXTKNUXP AO FJXKX RCNU
FJQSBK QS FJX CUPXU STLXP EXUKXVXUTSHX HTUXRND LXFJCPK CR TSTDOKQK
QSFNQFQCS DNHI FJX TAQDQFO TF DXTKF FC UXTP FJX DTSBNTBX CR FJX
CUQBQSTD QK VXUO PXKQUTADX ANF SCF XKKXSFQTD
The ciphertext has as its frequency chart
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
4 5 10 10 2 17 0 6 2 9 17 3 0 9 4 7 18 4 16 16 12 2 0 27 2 0
We have previously determined that this is not a Caesar cipher (as none of
the usual hill and valley patterns fit) so we are treating it as a monoalphabetic
cipher of unknown type. The most common letters are C, D, F, J, K, Q, S,
T and X, so these are probably the substitutes for etaoinshr. But which is
which and how can we decide
Just as ciphers with word spacing are generally easier to decrypt than those
without, long ciphers are generally easier to decrypt than short ones. And for
a simple reason: the longer the cipher is, the more accurate the information
given by the frequency count is likely to be. If we are given a 100, 000 letter
message enciphered with a monoalphabetic substitution (without other trick!),
we can be quite confident that exactly one of the ciphertext letters will occur
almost 13% of the time, and this letter stands for e, that the next most common
letter will appear very near 9% of the time and will stand for t, etc. Very little
thought will be needed to decrypt the cipher. It is the medium length ciphers of
25 to 100 letters, usually without proper word division, that give the beginning
codebreaker some difficulty. We will concentrate on those in this chapter.
89