How perfect is Voynichese?

[ololololo]

It is worth noting that if you combine anagrams with substitution, you will get a fairly secure cipher, no worse than a one-time pad... 

Uh, no, not even close. A one-time pad is theorethically perfect, it's way harder to decrypt even than, say, RSA encryption. Anagrams are just moderately difficult and simple substitution won't change the frequency distribution of characters of the original language so it does not add much to anagrams.

It will be possible to solve rather complicated anagrams, but is it possible for a person who does not have any data about the cipher itself to do this?
If the VMS cipher is exactly what I'm talking about in this thread, then even after getting the encryption algorithm (substitution with anagrams), then we also have to work manually with sorting out the internal rules of this cipher. And we know very little about them...
And don't think that I consider Voynichese to be just anagrams. For reference, there is a separate thread about the numeric cipher, where I outlined some assumptions.