Mysteries Department
Sign inRequest access
Run · Voynich Manuscript

T41 generation 2: lattice EM followed by trigram key polish (element solver), two rounds; branch killed

Lattice EM decipherment: sign boundaries and key learned together, with a structure-preserving null: Held-out dictionary coverage (bar 0.7 x true plaintext's) / key accuracy by use.

Fails the control gate; polishing the EM key with M-T40's trigram element solver on the EM's Viterbi segmentation does not raise key accuracy and pushes the decode towards vowel runs (the solver's frequency term). Branch killed; the manuscript was not decoded.control: coverage 0.239 (bar 0.488), key 0.306 after polish (EM alone 0.319); pseudo-Latin nulls 0.234 and 0.2480

Result

RunMeasuredResultAgainstDecisionDate
RunHeld-out dictionary coverage (bar 0.7 x true plaintext's) / key accuracy by usecontrol: coverage 0.239 (bar 0.488), key 0.306 after polish (EM alone 0.319); pseudo-Latin nulls 0.234 and 0.248Generation 1 (EM alone): coverage 0.324, key 0.364Fails the control gate; polishing the EM key with M-T40's trigram element solver on the EM's Viterbi segmentation does not raise key accuracy and pushes the decode towards vowel runs (the solver's frequency term). Branch killed; the manuscript was not decoded.

What it used and found

Parameters

pw = 0.25, alpha = 0.05, restarts = 40, settings = T39 best Herbal A (as g1a), permutations = 10, em_iterations = 60, hybrid_rounds = 2, solver_iterations = 600000, change_from_generation_1 = added HYBRID stage: Viterbi segmentation from EM, framework/elem_solver.solve (trigram element model, as M-T40) initialised from the EM key, then EM re-fitted from the polished key

Cite it

Mysteries Department. (2026). T41 generation 2: lattice EM followed by trigram key polish (element solver), two rounds; branch killed [Run RUN0DCFZ7A7Z2R2]. https://mysteriesdept.com/r/RUN0DCFZ7A7Z2R2
Private beta Downloads are for beta members

We're in a private beta while we get things right. Everything here is free to read; downloading datasets, sources and code is open to members for now.

It takes a couple of minutes. We'll reply by email.