Run · Voynich Manuscript

T44 generation 1, rung R1: 5,000 restarts, 40 rounds of batched basin hopping and 40 of parallel tempering; key 74 % recovered, coverage 0.44 under the bar of 0.483

t44_gen1_R1.zip is the generation archive: the code as it ran, t44_g1r1.json, the best state (t44_g1r1_control_best.npz), every restart's likelihood (t44_g1r1_control_lls.npz), the ceiling outputs and the combined log. Tempering acceptance per level fell from 0.37 (hottest) to 0.12 (coldest) and exchange rates were 0.33-0.66, so the ladder mixed; it still did not beat greedy hopping.

Fails the control gate on coverage alone (0.440 < 0.483; key 0.736 passes its half). Likelihood: batched hopping is the best search (-40,675, 160 nats short of the true key's basin and still improving at round 40); tempering from the same states plateaued at -40,700 after 10 rounds; restarts alone saturate (the best of 5,000 is only 60 nats above the best of 400). Generation 2 doubles the hop rounds and adds the element trigram model (polish of the best states and trigram Viterbi decode), which the ceiling study shows is needed from R3 anyway.restarts (5,000): best -40,788.8 (key 0.652; quantiles 0/10/50/90/99 %: -42,107 / -41,736 / -41,490 / -41,314 / -41,078); basin hopping from the top 16 (40 rounds x 1,024 proposals): -40,674.8 (key 0.736); parallel tempering from the same states (40 rounds x 1,024 chains, 8 temperatures): -40,698.6 (key 0.666). Decode of the best state: coverage 0.440 (permuted keys mean 0.266, max 0.371, p 0.005), character similarity 0.638, key 0.736, null words 0.154, held-out ll/glyph -1.9368. 'aterescontigititeporiseriasatundentiticunepiciqueiam...' for 'aberenoncontigitcumteporisarboresabundentibipunicaeficiqueiam...'0

Result

Run	Measured	Result	Against	Decision	Date
Run	Best training log-likelihood per search; held-out key accuracy by use, dictionary coverage (bar 0.7 x plaintext's = 0.483), character similarity, key-permutation p (200 permutations)	restarts (5,000): best -40,788.8 (key 0.652; quantiles 0/10/50/90/99 %: -42,107 / -41,736 / -41,490 / -41,314 / -41,078); basin hopping from the top 16 (40 rounds x 1,024 proposals): -40,674.8 (key 0.736); parallel tempering from the same states (40 rounds x 1,024 chains, 8 temperatures): -40,698.6 (key 0.666). Decode of the best state: coverage 0.440 (permuted keys mean 0.266, max 0.371, p 0.005), character similarity 0.638, key 0.736, null words 0.154, held-out ll/glyph -1.9368. 'aterescontigititeporiseriasatundentiticunepiciqueiam...' for 'aberenoncontigitcumteporisarboresabundentibipunicaeficiqueiam...'	Session 12: best of 400 restarts -40,852 (key 0.623), 250 hops -40,836 (key 0.691, coverage 0.407) · EM from the true key -40,514 on this inventory (coverage 0.518, key 0.778)	Fails the control gate on coverage alone (0.440 < 0.483; key 0.736 passes its half). Likelihood: batched hopping is the best search (-40,675, 160 nats short of the true key's basin and still improving at round 40); tempering from the same states plateaued at -40,700 after 10 rounds; restarts alone saturate (the best of 5,000 is only 60 nats above the best of 400). Generation 2 doubles the hop rounds and adds the element trigram model (polish of the best states and trigram Viterbi decode), which the ceiling study shows is needed from R3 anyway.

What it used and found

Lattice EM on CUDA with batched restarts, basin hopping and parallel tempering, run up a ladder of planted controls (T44)

Method

MTH

Herbal A and Herbal B paragraph text: training and held-out pages

Dataset

DAT

Latin genre texts

Dataset

DAT

Naibbe cipher tables and example plaintexts (greshko/naibbe-cipher, commit 719f2f4c)

Dataset

DAT

Token table

Dataset

DAT

Parameters

pw = 0.25, rung = R1, topk = 16, alpha = 0.05, chunk = 1024, dtype = float32, signs = 871, pt_beta = geometric 0.005 .. 1, elements = 71, restarts = 5000, settings = {'K': 50, 'h_c': 1, 'h_v': 1, 'lam': 0, 'Ssyl': 0, 'abbrev': 0, 'p_copy': 0, 'p_stay': 0.9, 'window': 5, 'novowel': 0, 'p_break': 0.5, 'spacing': 'stream', 'simplify': 1, 'positional': 0}, hop_pairs = 3, pt_levels = 8, pt_rounds = 40, hop_rounds = 40, permutations = 200, em_iterations = 60, restart_seeds = 100000-104999, PYTHONHASHSEED = 0, hop_em_iterations = 30, pt_chains_per_level = 128, hop_proposals_per_round = 1024

Cite it

Mysteries Department. (2026). T44 generation 1, rung R1: 5,000 restarts, 40 rounds of batched basin hopping and 40 of parallel tempering; key 74 % recovered, coverage 0.44 under the bar of 0.483 [Run RUN0DCMW01CAK37]. https://mysteriesdept.com/r/RUN0DCMW01CAK37

@misc{mysteriesdept_run0dcmw01cak37,
  title     = {T44 generation 1, rung R1: 5,000 restarts, 40 rounds of batched basin hopping and 40 of parallel tempering; key 74 % recovered, coverage 0.44 under the bar of 0.483},
  author    = {{Mysteries Department}},
  year      = {2026},
  note      = {Run, catalog key RUN0DCMW01CAK37},
  url       = {https://mysteriesdept.com/r/RUN0DCMW01CAK37}
}