"Is the Voynich a Hoax?" — Five test say NO

[oshfdk]

I didn't really check the paper, I just would like to comment on the general premise. Unless the paper provides a deciphering which produces a sensible plaintext, it's impossible to use a statistical test as an argument that the manuscript is NOT a hoax. No statistical test ever will be able to show that the manuscript is not a hoax.

One of the easiest way to create a plausible hoax is to get a real text as the source. For example, I can take each second letter (including spaces, skipping punctuation) of the beginning of this very post and produce a hoax text, which probably will pass a lot of statistical tests:

IDDTRAL HC H AE UTWUDLK OCMETO H EEA RMS NESTEPPRPOIE EIHRN HC RDCSASNIL LITX T MOSBET S TTSIA ETA NAGMN HTTEMNSRP SNTAHA OSAITCLTS VRWL EAL OSO HTTEMNSRP SNTAHA N FTEESETWYT RAEAPASBEOXI OGTARA ETA H OREFREAPEICNTK AHSCN ETRICUIGSAE KPIGPNTAINFTEBGNIGO HSEYPS N RDC OXTX HC RBBYWL ASALTO TTSIA ET

It has some repeating words (SNTAHA which is made from "iS NoT A HoAx"), it would have some topic clustering. It would have weird statistics and the original text may or may not be recoverable, but I guess it would look more like a meaningful text than not, because it is based on a meaningful text. Yet, it is not designed to convey any meaning, just to give an impression of a meaningful text, which makes it a hoax.