“F1.1 at key length 26 decided with design 2 (K4 against its own copies); design 2 carried to two free alphabets (Quagmire IV) and its power measured at F1.2's lengths 8, 13 and 16”
F1.1 Beaufort 26 and F1.2. 0140 found design 2 has power at length 26 too (5 of 6 planted texts beat all their own copies), so K4 at 26 now runs against 12 fitting shuffles of itself. In the same run design 2 is carried to two independent alphabets (Quagmire IV: C(c) = P(p) + k), the case F1.2 leaves at lengths 8, 13 and 16 under 0136's bound: compiled tempering over both alphabets, every move keeping the known-letter checks, each run started from an exact witness (0123's decider) randomised by 400 check-keeping swaps; power measured on planted texts against their own fitting shuffles.
Two pieces. First, the carved textthe carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. is run through the new search for a scrambled alphabet with a 26-letter key, against scrambled copies of itself. Second, the search is extended to two scrambled alphabets, one for the messagethe message The 97 letters of English that Jim Sanborn wrote and enciphered. and one for the carved letters, and tested on planted English at the three key lengths where English could in principle decide, before it is allowed to judge the carved text.
Words in the margin are defined on the glossary page.
1.1
The method at work
The carved letters go in at the top; the reading comes out at the bottom.
the idea
Two jobs in one run: K4 at key length 26 against shuffles of itself, and the same search carried to two unknown alphabets, its power measured on planted English at lengths 8, 13 and 16.
what happened
After 28 minutes the two-alphabet part ran out of time while drawing shuffles that fit. One text had no finished search, and that broke the summary, so no result was kept.
the rerun
0142 ran the length-26 job alone. 0143 ran the two-alphabet power test.
2
Best candidate against the known letters
This experiment produced no candidate: the search ruled everything out before a full text was built. Its product is the bounded negative in section 7.
3
Walkthrough, by hand
Redo the method on a few letters with pencil and paper.
Two alphabets: number the message letters by one scrambled order and the carved letters by another; carved place = message place + key value.
Two known letters under one key letter must give the same difference, carved place minus message place: at length 8 sixteen such checks.
Find one pair of orders that satisfies every check exactly, then shuffle it by swaps that keep the checks.
Run sixteen searches side by side over pairs of orders that keep the checks, scoring the whole text as English.
Do the same on planted English and on shuffles of it; count how often the planted text wins.
4
Worked example, generated
The same steps run by the code on the real letters.
5
Hypothesis, for a cryptanalyst
Family, key space, the constraint under test, and the clues it reads.
C27(c), C1(b). L26 one alphabet, Beaufort form (1 check), as 0140 (seed 141): K4 vs 12 shuffles of K4 fitting at 26; measured power 5 of 6 at 6 copies (0140); rule fixed before the run: K4 above all 12 stops the queue for replication, otherwise the cell closes sampled at that power. Quagmire IV (0123's model, two unknown bijections P and C, key length L): 0136 bounds the open bits at 89, 112 and 126 for L 8, 13, 16 against 130 of quadgram evidence; checks 16, 11, 8. Search: replica exchange, 16 replicas, T 0.1 -> 0.002, 10 runs x 625 rounds x 1,000 moves per text (0139's effort); moves swap two places in one alphabet, repaired by a second swap when a check breaks. Power: 4 planted texts per length (Carter windows, seed 1411), each against 5 shuffles of itself fitting that length (0123's decider); null 1/6 per plant; bar fixed before the run: at least 3 of 4 beat all their own copies. A length that passes gets K4 against its own copies in the next run; one that fails stays blocked.
6
Prediction
Written before the run; the result decides it.
K4 inside its copies at 26 (cell closes at power about 0.8). Two alphabets: power falls with the number of checks and the doubled alphabet freedom: at least 3 of 4 at length 8 (16 checks), fewer than 3 at 13 and 16. No reading.
7
Method
Draw 12 shuffles of K4 fitting at 26 (one alphabet, Beaufort form); 4 planted Quagmire IV texts at each of 8, 13, 16 and 5 fitting shuffles of each.
One-alphabet texts: 0139's search. Two-alphabet texts: the same search over both alphabets from a witness start; the compiled decryption checked against Python on every text.
L26: K4's rank. Two alphabets: plants beating all their own copies per length.
8
Result and bounds
The harness's record: candidates, gates, loss, and the negative result this experiment leaves behind.
0candidates returned
0passed every gate
0/24known letters, best
–lowest loss
nonechampion before
The run stopped with an error: TypeError: cannot unpack non-iterable NoneType object
How this experiment sits among the others is on the lineage page; its family's open hypotheses are on the families page.
Rendered from the record on 2026-10-18 12:26 UTC; updated after every experiment.