Experiment 218
A13.4 (Stage 2, step 3, issue #75). Since 2026-10-18 four experiments kept K4 readings with all 24 known letters in place: 0207 (periodic key on the sculpture's tables, up to 3 slips), 0212 (ciphertext autokey and progressive key, 2 slips or a lost step), 0213 (plaintext autokey, 2 broken links or a lost step) and 0214 (a move then the table, 2 slips). None kept its copies' readings. Here each frozen attack.py's own per-text search is called on K4 and on the same 20 seeded copies it used (tools/a13_check.py's pass 4), every reading it returns goes through the battery, and K4 is ranked per experiment and test. Pre-registered 2026-10-21 in reports/a13-structure-check.md: 56 cells, flag at p < 0.05/56, candidates replicated sequentially (seeds 4113 on). New in this run: masked planted messages with slips through 0207's search.
What the method does with the 24 known letters.
The searches that allowed for one or two enciphering mistakes produced readings with every known letter right but gibberish elsewhere. If the messagethe message The 97 letters of English that Jim Sanborn wrote and enciphered. was disguised before it was enciphered, the right one would look like gibberish too. We rerun those searches exactly, on the carving and on the same scrambled copiesa shuffled control The same experiment run on the carved letters shuffled into a random order. they used, and test every reading for 14 kinds of hidden pattern; the carving counts only if it beats all the copies. We also hide disguised messages with mistakes of our own and check the tests notice them.
Words in the margin are defined on the glossary page.
The carved letters go in at the top; the reading comes out at the bottom.
A13.4: every reading the slip-tolerant searches 0207, 0212, 0213 and 0214 let through, K4's and their copies', through 14 pattern tests.
The readings were regenerated (all 80 saved K4 candidates among them), then the plant step stopped: 0216's masking routine, loaded from its own folder, looked for Carter's book there.
0219: the routine's root set to the repository, the plant step tested on its own first, the same run repeated.
What this experiment was built on, from the lineage page: 15 experiments before it, back to the root of its line. Its number is in bold.
This experiment produced no candidate: the search ruled everything out before a full text was built. Its product is the bounded negative under Result and bounds below.
Redo the method on a few letters with pencil and paper.
The same steps run by the code on the real letters.
Family, key space, the constraint under test, and the clues it reads.
A13 with A1 (knowledge/12-assumptions.md): a masked plaintext (C4(a), Scheidt's 'I masked the English language'; C19, X and Q stops) enciphered with a slip or two (C17(b), Sanborn's 'sloppy'). The readings are those A1's exact searches let through: 0207 (8 table classes, key lengths 1-26, k <= 3), 0212 (ciphertext autokey on 487,734 table setups, k <= 2 or a lost step; progressive key, k <= 2), 0213 (plaintext autokey lags 1-52, 2 breaks or a lost step), 0214 (1,847,590 landings of 0110-0114 and the type-1 library, both orders, k <= 2). Statistic: per experiment and test, the best value over a text's readings ('?' read as unknown, the 24 known letters set as the candidates set them); K4's rank among its 20 copies, ties against K4. 0199, 0201-0206 and 0208-0211 kept no K4 reading (nothing fits K4, or a free alphabet leaves it undetermined): noted, not rerun.
Written before the run; the result decides it.
K4 strictly above every copy in about 56/21 = 2.7 cells, none replicating; no flag. Every saved K4 candidate of the four experiments is among the regenerated readings. The masked plants through 0207 are found by their mask's own test in most cases, because 0207's readings at lengths 13 and 20 are fixed by the known letters and the planted reading itself is among them. Zero candidates pass gate 4.
The harness's record: candidates, gates, loss, and the negative result this experiment leaves behind.
The run stopped with an error: FileNotFoundError: [Errno 2] No such file or directory: '/Users/alexisrondeau/Workshop/cryptos/experiments/0216-a13-the-saved-readings-against-their-own-search-on-fresh-copies-rerun/knowledge/sources/carter-1923-tomb-of-tutankhamen-vol1.txt'
How this experiment sits among the others is on the lineage page; its family's open hypotheses are on the families page.