I found the following table quite interesting:
Occurences distance char-pair
6862 6 .-.
5028 1 e-e
4509 1 i-i
2145 3 o-o
1837 7 y-y
1583 6 h-h
965 5 c-c
951 7 d-d
901 7 a-a
844 7 k-k
494 5 l-l
409 3 r-r
381 7 t-t
363 7 q-q
279 6 s-s
271 7 n-n
53 8 p-p
17 5 m-m
7 7 f-f
Besides ee ii oo rr (which can occur within one token), VMS seems to consist of tokens with an average length of 6 (in a peak range of 5-7 characters.These IMHO suggests, that spaces are token separators, the frequency of a space (.) pair is exactly 6.
Claus