“F6.21, replication: 0203's row-reshuffle family, where K4 led every copy in 6 cells against a null of 1, rerun unchanged against 40 fresh scrambled copies, with a per-width breakdown; a deliberate slip in the keyword columnar (one column read out of its place, or two columns swapped), around every keyword columnar already tested, then or before a repeating key on the table, exact, with planted slips in the same run”
0203 found K4 above every one of its ten scrambled copies in 6 of 52 cells for the letters-reshuffled-inside-rows family (F6.18) with one slip (A 19, 21, 22, 23, 26; B 24), where no copy led the rest in more than one: outside the copies' range, which stops the queue until replicated (order-of-work rule 5). Here the same search, unchanged, on that family only, against 40 fresh scrambled copies (new seed), with the same fixed statistic (cells where K4 exceeds every copy, against each copy's own count against the rest), K4's rank in every cell, and the counts split by row width to show which widths carry any excess. Budget 45 minutes (41 texts plus a plant).
Agents in the war often made one mistake when they copied out a transpositiona grid Writing the message into rows of some width and reading it out in another order, column by column for instance, which shuffles the letters without changing them.: they numbered the keythe secret word The word or string a cipher needs besides the method. word wrongly and read one column at the wrong time, or read two columns in each other's place. Sanborn says he made mistakes on purpose. We take every column shuffle the project has already ruled outdead Said of a period, grid or method when the known letters contradict it: no secret word of that kind can produce the carved letters. and try it again with each such single mistake, followed or preceded by a short repeating key word on the sculpture's tablethe table The right-hand half of the sculpture, 26 rows of the alphabet each shifted one place further, starting from the word KRYPTOS..
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
0203 found K4 ahead of all ten scrambled copies in six cells of one family, where no copy led in more than one: outside the copies' range, which stops the queue until it is replicated.
the check
The same search, unchanged, against forty fresh scrambled copies, with K4's rank in every cell and the counts split by row width.
the bound
K4 leads every copy in one cell only (A21, carried by rows of 9), as ten of the forty copies do in one or two. Chance; F6.21 closed. The plant is found.
2
Experiment lineage
What this experiment was built on, from the lineage page: 8 experiments before it, back to the root of its line. Its number is in bold.
keptreverted, in its family's colour×error or timeoutreverted under today's rulesbuilt on a parent in the same familybuilt on a parent in another family
3
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 under Result and bounds below.
4
Walkthrough, by hand
Redo the method on a few letters with pencil and paper.
Take the key word SHADOW: alphabetical order A D H O S W numbers the columns 5 3 2 6 4 1... read in the order 3, 4, 2, 5, 1, 6.
Make one slip: read column 2 one step early (3, 2, 4, 5, 1, 6), or swap columns 3 and 5.
Write the 97 letters in rows of 6 and copy out the columns in the slipped order.
Encipher on the table with a repeating key, or encipher first.
Follow each known letter to its carved place; known letters that share a key slot must agree.
the worksheet: the first steps by hand, on squared paper
5
Worked example, generated
The same steps run by the code on the real letters.
key word SHADOW: columns read in the order [3, 4, 2, 5, 1, 6]
one slip, the third column read one step early: [3, 2, 4, 5, 1, 6]
place known lands at carved
22 E 36 T
26 N 21 W
34 T 38 T
64 B 43 S
74 K 29 R
6
Hypothesis, for a cryptanalyst
Family, key space, the constraint under test, and the clues it reads.
C1(a), C5, C10 ('mistakes on purpose'; misspellings as security checks in SOE practice, Marks p. 16). Source: reports/marks-between-silk-and-cyanide.md. Model: c = T(sigma'(p)) or sigma'(T(p)), sigma' a keyword columnar (or two passes) whose column reading order differs from the keyword's by one move of one column (g(g-1) orders) or one swap of two non-adjacent columns ((g-1)(g-2)/2; adjacent swaps are moves), in one pass. A skipped or doubled column alone changes the length (97 would need a message of 97 plus or minus a column, or two compensating slips): not searched. Not included: Myszkowski, AMSCO, keyed blanks, triangle and disrupted columnars (not on the item's list). Each slipped route is tested exactly as 0110 tested its routes: periods 1-26, both orders, 8 classes, by slot agreement of the 24 known letters, forward and inverse. Statistic (0134's, fixed): per base family, the shortest key length consistent on K4 against the copies, and the cells (order, length) where K4 exceeds every copy, against each copy's own count. Routes are counted per slip, not deduplicated across words (copies counted the same way).
7
Prediction
Written before the run; the result decides it.
The excess does not replicate: against 40 fresh copies K4 leads every copy in no more cells than a copy does (expected about 52/41, roughly 1 cell), and its rank in 0203's six cells falls inside the copies. The plant is found. Zero candidates.
8
Method
Rebuild each base family's column groups and reading order with the experiments' own builders.
In compiled code, for every base, every slip of the reading order (one column moved, two swapped) in each pass, build the route and its inverse.
Test each with the 0110/0164 periodic-key test, orders A and B, 8 classes, periods 1-26, on K4, ten copies and four planted texts; checked against Python on sample routes.
The fixed statistic per base family; each plant must show its own slip live at its own length.
9
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
F6.18 row reshuffle: 121,696,528 routes, shortest live A 19 B 24 (copies {'A': 18, 'B': 17}), K4 above every copy 1 (null max 2). Plants: F6.18 row reshuffle found True
What was searched, as the harness recorded it.
family
one slip (a column moved or two swapped) in the reading order of every keyword columnar already closed, then or before a periodic key