A key to understand the VMS

[nickpelling]

If you take the three (EVA) letters 'e', 'o', 'l', and count the occurrences of their six permutations (in the Takahashi transcription), I think you get the following:

eol - 961
ole - 39
oel - 0
leo - 2
loe - 0
elo - 0

These numbers strongly imply that Voynichese has very strong letter adjacency rules in play: and it is surely far from coincidental that both eol and ole are the only two permutations where l immediately follows o. (For reference, 'ol' occurs 5507 times in the same transcription.)

Without any doubt, these numbers are inconsistent with auto-copying if you are trying to argue that errors introduced during autocopying are the systematic source of the random variation within Voynichese.