Experiment 106
0019 tested a monoalphabetic substitution with 23,931 routes. Since then experiments 80-99 built far larger move sets: the type-1 library (27,852), grid paths (43,579), mixed-sense double turns (53,743) and every triple clockwise turn (1.7 million). A one-to-one substitution commutes with any move, so one test covers both orders: the known letters must land on carved letters that a single one-to-one table can explain.
What the method does with the 24 known letters.
The simplest two-step cipher swaps each letter for another by a fixed table and then moves the letters around. Experiment 19 tried this with about 24,000 ways of moving letters. Since then we have built ways of moving letters by the million. For each one we check whether a single swap table can turn the known lettersthe known letters Sanborn has confirmed that positions 22 to 34 of the message read EASTNORTHEAST and positions 64 to 74 read BERLINCLOCK. into the carved lettersthe carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. where they land: the same letter must always become the same letter.
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.
The simplest two-step cipher swaps each letter for another by a fixed table, then moves the letters around. Experiment 19 tried about 24,000 moves; since then the project has built them by the million.
For each move, the same known letter must always land on the same carved letter, and two different known letters never on the same one. E alone, three times among the cribs, is a strong test.
1,835,741 moves in four sets: not one admits a swap table, on K4 or on any of ten scrambled coppers.
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.
Redo the method on a few letters with pencil and paper.
The same steps run by the code on the real letters.
snake rows of 10, read by columns: the known letters and the carved letters where they land
letter places (lands at: carved)
A 23 (23: L), 32 (83: C)
B 64 (37: W)
C 70 (95: C), 73 (78: K)
E 22 (13: B), 31 (92: U), 65 (47: Z)
H 30 (91: A)
I 68 (77: D)
K 74 (68: T)
L 67 (67: V), 71 (96: A)
N 26 (53: K), 69 (86: G)
O 27 (63: B), 72 (87: K)
R 28 (73: P), 66 (57: I)
S 24 (33: S), 33 (74: K)
T 25 (43: S), 29 (82: J), 34 (64: N)
a single swap table needs every row to show one carved letter, different from every other row
Family, key space, the constraint under test, and the clues it reads.
Family 9 (least covered, two experiments), open 'only for routes outside the set (spirals, diagonals, keyed routes)' (families.md); 0018 found substitution-then-transposition the only construction reproducing the width-21 repeats. Model: c = route(sigma(p)) = sigma(route(p)), sigma any permutation of 26 letters. Test: for the 24 known letters at their landing places, equal message letters must land on equal carved letters and different on different. Route sets: the type-1 library (K3's rule, one and two clockwise turns, plain and true-keyword columnar, forward and inverse); 0092's 30 x 29 path pairs at widths 2-96; 0099's mixed-sense double turns; 0083's triple clockwise turns at widths 2-96; all deduplicated. Ten shuffled coppers. Survivors decrypt wherever sigma is fixed and are scored.
Written before the run; the result decides it.
No route of any set survives on K4, and none on the shuffled coppers, as in 0019. Zero candidates.
The harness's record: candidates, gates, loss, and the negative result this experiment leaves behind.
Routes tested: {'type-1 library': 27852, 'grid paths': 42978, 'mixed double turns': 53214, 'triple clockwise turns': 1711697}. Consistent (K4 / controls): none on any text.
What was searched, as the harness recorded it.
How this experiment sits among the others is on the lineage page; its family's open hypotheses are on the families page.