An Essay on Entropy: what is it, and why is it so important?

[Koen G]

I am beginning a very long-term project of proof-reading the transcription to make corrections and to resolve many of the uncertain readings. Might take as long as a year, but I think it will be well worth the effort.

Woah! The LFD master transcription, sounds like a dream. 

If I may dare to make a prediction, I would guess that even correcting lots of errors won't change the troublesome h2 statistic a lot. It is a fairly broad statistic that can take quite a beating before it changes. And one of your corrections that shifts the results one way may be offset by another. I could be proven wrong though, which would be a very pleasant surprise. And either way there are lots of other analyses where having an improved transcription will make a significant difference.

Ruby: Yes, I'm afraid you misunderstand my purpose. I did not invent VM entropy research, in fact I learned most of what I know about it from others, like Anton and Rene. Of course you are right that statistical analyses are tools. And in this case, the tool reveals a troublesome side of the VM: its conditional character entropy is much too low compared to any known text. I did not invent this either, but I struggled to wrap my mind around it since I saw how important others found these observations. 

The results of my own research are mostly negative. I demonstrated that accounting for an EVA bias is not sufficient to fix everything, and explained why this is the case. Then I explained that I explored if Voynichese could statistically be a verbose cipher, but again I did not reach any results yet that I find satisfactory: the stats aren't quite right yet, and the process makes words much too short.

However, looking for a VM solution without taking into account the entropy problem is like driving around blindfolded.