I have a long list of statistical features somewhere...
But let's start with a few remarks:
"Entropy is invariant under simple substitution ciphers." This is not a feature of Voynichese.
Transliterations are subjective and unreliable. Glyphs are often difficult to categorize/recognize: the differences between them are small and range on a continuum. Spaces are even more unreliable: every transliteration makes different, inconsistent choices. So some statistics (especially on rare words and patterns) are very unreliable. On the contrary ciphers of the 15th century typically use very distinct symbols to avoid interpretation errors. This is an important feature of the text: it is impossible to transliterate reliably.
But let's start with a few remarks:
"Entropy is invariant under simple substitution ciphers." This is not a feature of Voynichese.
Transliterations are subjective and unreliable. Glyphs are often difficult to categorize/recognize: the differences between them are small and range on a continuum. Spaces are even more unreliable: every transliteration makes different, inconsistent choices. So some statistics (especially on rare words and patterns) are very unreliable. On the contrary ciphers of the 15th century typically use very distinct symbols to avoid interpretation errors. This is an important feature of the text: it is impossible to transliterate reliably.
![[+]](https://www.voynich.ninja/images/collapse_collapsed.png)
