All of you.
Some time ago, I presented an article about testing the Stroke theory by trying to decompose the VM text into “words” consisting of two “syllables” each where each syllable would represent one plaintext letter in simple substitution. This method stalled at rendering roughly 3/4 of the ciphertext volume with a set of 2×25 syllables/”letters”.
Then I worked the other way around and tried to synthesize VM text by taking plaintext and translating it into ciphertext by the Strokes algorithm to get an idea what the ciphertext would look like.
And here is where the betrayal comes into play.
Not one of you suggested that I could subject the pseudo-ciphertext (“PCT”) created this way to the stroke analysis of the first step above and see what would happen.
Namely, if the PCT could be decomposed again 100% into two syllable sets, then my analytical algorithm was correct, but the VM obviously was concocted in a different way, which would render the Stroke Theory not obsolete, but mostly obsolete. If, on the other hand, only a fraction of the PCT would be recognised as consitutent of two syllable sets (like, say 3/4 by volume), then this would point to a general defect in the decomposition algorithm, and the game would be afoot.
Not one of you felt the duty to take my responsibility to think for myself upon your shoulders for nearly a year. Shame on you.