Experiment 271

“F14.2: every type-1 move of 0129's library (27,852), both orders, key lengths 1-26, then a table with two dictionary keyword alphabets (208,064 each, standard and KRYPTOS among them) and a repeating key word; K4 against 50 shuffles, plants”

F14.2, the full run Alexis decided (#102). 0244's exact pair index (every plain and cipher keyword alphabet matched through the differences the known letters force) run after or before every move of 0129's type-1 library, at every key length 1-26, both forms. Faster than the design's estimate: which known places share a key slot depends only on the setting, so the plain-side index is built once per setting and K4 and its 50 shuffles are looked up in it; a difference table D[x, y] over all alphabets makes each check one contiguous row. Checked: with the identity move it reproduces 0244's fit counts on K4 (desk, 2026-11-04).

revertedstatus
experiment 244parents
7.5 htime, all phases
0candidates
0/24known letters
unclocked
1

The experiment method

What the method does with the 24 known letters.

Sanborn said he wrote the messagethe message The 97 letters of English that Jim Sanborn wrote and enciphered. into a grida 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. and read it out another way, then used the cipher table. Scheidt said the English was disguised first, and that the system is worked with key words. Here the grid step is any of about 28,000 ways of reading a grid, and both alphabets of the tablethe table The right-hand half of the sculpture, 26 rows of the alphabet each shifted one place further, starting from the word KRYPTOS. are scrambled by English words. For each recipe, the 24 known lettersthe known letters Sanborn has confirmed that positions 22 to 34 of the message read EASTNORTHEAST and positions 64 to 74 read BERLINCLOCK. tell us exactly which pairs of word-scrambled alphabets could work; then we check whether the keythe secret word The word or string a cipher needs besides the method. they force is a word. Fifty scrambled copiesa shuffled control The same experiment run on the carved letters shuffled into a random order. of K4the carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. go through the same test, so we know how many matches chance gives.

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.

OBKRUOXOGHULBSOLIFBBWEASTNORTHEASTOTWTQSJQSSEKZZWATJKLUDIAWINFBBERLINCLOCKWGDKZXTJCDIGKUHUAUEKCARmove, then the table (A): does a pair of the 208,064 keyword alphabets fit K4 after any of 27,852 moves, key lengths 1-26?1101426no pair fitspairs fit, as on the shufflesthe table, then the move (B):11526key words among the fitting pairs: K4 0, the 50 shuffles at most 0; plants found 12 of 12EASTNORTHEASTBERLINCLOCKIN: THE COPPER, ALL 97 LETTERS, THE 24 KNOWN PLACES IN ACCENTOUT: NO READING; NO KEY WORD FITS K4
the idea

F14.2: any of 27,852 remembered grid moves, then K1's table with both alphabets scrambled by dictionary words and a key word.

the check

1,448,304 settings per text, every pair of keyword alphabets tested exactly by the known letters; 50 shuffles alike.

the bound

No key word on K4 or any shuffle; plants 12/12. A flaw left 3.1% of settings unchecked (length 10's lone fit is one): rerun as F14.2a.

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.

kept reverted, in its family's colour ×error or timeout reverted under today's rules built on a parent in the same family built 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.

  1. Write the message in rows of 9 and read the columns in a key word's order (one of the moves).
  2. For a key length, say 12, follow each of the 24 known letters to its carved place and its key slot.
  3. Two known letters in one slot must be the same distance apart on the plain alphabet as their carved letters are on the cipher alphabet.
  4. File every word-scrambled alphabet by those distances, plain side and cipher side, and pair up the matching files.
  5. For each matching pair, read off the key letters the known letters force and look for a word that gives them.
  6. Do the same for 50 scrambled copies of K4 and compare the counts.
the worksheet: the first steps by hand, on squared paper
order, lengthK4shuffles, medianmaxpA10 Beaufort1000.02A10 added1010.04A14 Beaufort6270.10A18 Beaufort3,4453,0533,7250.08A18 added3,4723,0483,7430.08A23 Beaufort7,2406,6207,9790.16A23 added7,3446,5747,9600.12A24 Beaufort6,8536,3667,2620.16A24 added6,8306,3597,2800.18A26 Beaufort7,6816,6038,0230.08A26 added7,6586,6248,0280.10B22 added1861602080.14B24 Beaufort23,28623,14923,5400.20B24 added23,28623,14923,5400.20Smallest p 0.02 of 100 cells, about what chance gives. The A10 fits are one setting where the first six checks repeatknown letters: its pairs were counted, not verified (F14.2a reruns 41,986 such settings).SETTINGS WHERE A KEYWORD-ALPHABET PAIR FITS, THE CELLS WITH K4 NEAREST THE TOP (ALL 100 CELLS IN THE RECORD)
5

Worked example, generated

The same steps run by the code on the real letters.

F14.2: a remembered route, then a table with two word-scrambled alphabets and a key word.

  move       write the message in rows, read the columns in a key word's order (one of 27,852 moves)
  checks     two known letters that share a key slot after the move: their gap on the plain alphabet
             equals their carved letters' gap on the cipher alphabet (added form)
  index      every keyword alphabet filed by its gaps, plain side and cipher side; matching files pair up
  key word   the pair forces the key at the covered slots; a word of the key's length must give it
  50 scrambled copies of K4 through the same steps give chance's counts.
6

Hypothesis, for a cryptanalyst

Family, key space, the constraint under test, and the clues it reads.

F14 (Scheidt's mask keyed by words; C4(a), C5; 03-scheidt-said.md 'masked the English language', 'executed with the correct key word/s') with Sanborn's matrix step first or last (C1(a): 'I used a matrix code, and then... this [tableau]'; 'double matrixes'): a remembered route (keyword columnar, K3's rule, turns, rail fence, grid paths: 0129's 27,852 distinct moves and their inverses) and a Quagmire IV table whose alphabets are keyword alphabets from the dictionary, keyed by a repeating word. F6.1 and F6.2 decided free alphabets around these moves at the copies' rate; keyword alphabets cost about 35 bits for the pair, so the known letters decide them where 5 or more same-slot checks fall (chance pairs 4.3 x 10^10 / 26^c per form: 3,600 at c = 5, 140 at 6, 5 at 7, 0.2 at 8). A fitted pair forces the key at the covered slots; a dictionary or puzzle word of length L (any rotation, numbered on standard or KRYPTOS, any shift) must give it. Settings with 4 or fewer checks are counted only (chance pairs 10^5 and more). Prior attempts (reports/Kryptos K4 prior attempts database.md): moves with periodic keys on the standard and KRYPTOS tables (keithadler, nitsuah, Zekios) and free-alphabet Quagmires without a move; a move with two dictionary keyword alphabets is not covered. Clue entries C1(a), C4(a), C5, C27(c).

7

Prediction

Written before the run; the result decides it.

K4's fitting pairs and key-word matches per (order, key length) inside the 50 shuffles' range; plants found at every band; no reading English. Zero candidates kept.

8

Method

  1. Settings: 0129's route library (27,852 moves) x order (A move then table, B table then move) x key length 1-26; for each, the 24 known places' carved places and key slots, and the same-slot checks (first place of a slot against each other).
  2. Per setting and form, the plain-side key of every alphabet on the first min(c, 6) checks, sorted; per text, the cipher-side keys looked up (a bitmap first, then exact), the remaining checks verified on the candidate pairs. c <= 4: the number of fitting pairs counted exactly, not listed.
  3. Key word (L <= 24, 3 or more covered slots): each fitted pair's forced key differences matched against every rotation of every dictionary and puzzle word of length L, numbered on standard or KRYPTOS (and on the pair's own alphabets when 20 or fewer pairs fit); a match fixes the whole key and so the reading.
  4. Texts: K4 and 50 shuffles (seed 1420), every setting; 12 plants (English with the known words, a random move, order, two dictionary alphabets and a dictionary key word; 4 in each length band 6-10, 11-18, 19-24, on settings with 5 or more checks), each searched at its own setting with the same code.
  5. Verdict: K4's fitting pairs (settings with 5+ checks) and key-word matches per (order, length) ranked among the shuffles; K4's key-word readings submitted as candidates, the shuffles' best as controls.
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

F14.2: 27852 moves x 2 orders x 26 lengths x 2 forms, 208,064^2 keyword alphabet pairs each; key-word matches K4 0, copies median 0.0, max 0; smallest per-class p 0.02 over 100 classes; plants found 12/12.

What was searched, as the harness recorded it.

item
F14.2
routes
27,852
orders
A move then table, B table then move
key lengths
1, 26
forms
added, Beaufort
alphabets
208,064
alphabet pairs per setting and form
43,290,628,096
settings per text
1,448,304
listed at checks
5
settings listed k4
1,363,532
key word lengths
1, 24
key numbering
standard or KRYPTOS (and the pair's own alphabets when 20 or fewer pairs fit), any shift
self check identity vs 0244
16/V
5, 5
16/B
0, 0
19/V
0, 0
19/B
0, 0
23/V
5, 5
23/B
1, 1
per class
A1B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A1V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A2B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A2V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A3B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A3V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A4B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A4V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A5B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A5V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A6B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A6V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A7B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A7V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
2
p at or above
1
A8B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A8V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A9B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
2
p at or above
1
A9V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
2
p at or above
1
A10B
settings listed
27,852
k4 settings with fit
1
copies median
0
copies max
0
p at or above
0.02
A10V
settings listed
27,852
k4 settings with fit
1
copies median
0
copies max
1
p at or above
0.039
A11B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A11V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A12B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A12V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
A13B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
2
p at or above
1
A13V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
A14B
settings listed
27,852
k4 settings with fit
6
copies median
2
copies max
7
p at or above
0.098
A14V
settings listed
27,852
k4 settings with fit
0
copies median
2
copies max
25
p at or above
1
A15B
settings listed
27,852
k4 settings with fit
36
copies median
32
copies max
65
p at or above
0.235
A15V
settings listed
27,852
k4 settings with fit
30
copies median
33.5
copies max
59
p at or above
0.667
A16B
settings listed
27,852
k4 settings with fit
511
copies median
513
copies max
638
p at or above
0.549
A16V
settings listed
27,852
k4 settings with fit
545
copies median
510
copies max
678
p at or above
0.333
A17B
settings listed
27,852
k4 settings with fit
1,991
copies median
1,958
copies max
2,359
p at or above
0.431
A17V
settings listed
27,852
k4 settings with fit
1,949
copies median
1,946
copies max
2,324
p at or above
0.49
A18B
settings listed
27,852
k4 settings with fit
3,445
copies median
3,053
copies max
3,725
p at or above
0.078
A18V
settings listed
27,852
k4 settings with fit
3,472
copies median
3,048
copies max
3,743
p at or above
0.078
A19B
settings listed
27,852
k4 settings with fit
4,731
copies median
4,616
copies max
5,258
p at or above
0.392
A19V
settings listed
27,852
k4 settings with fit
4,701
copies median
4,578.5
copies max
5,324
p at or above
0.412
A20B
settings listed
26,243
k4 settings with fit
4,366
copies median
4,117.5
copies max
4,814
p at or above
0.255
A20V
settings listed
26,243
k4 settings with fit
4,342
copies median
4,098
copies max
4,755
p at or above
0.235
A21B
settings listed
26,366
k4 settings with fit
5,140
copies median
5,348.5
copies max
6,192
p at or above
0.706
A21V
settings listed
26,366
k4 settings with fit
5,072
copies median
5,362
copies max
6,191
p at or above
0.725
A22B
settings listed
24,447
k4 settings with fit
5,503
copies median
5,382
copies max
6,053
p at or above
0.412
A22V
settings listed
24,447
k4 settings with fit
5,401
copies median
5,375
copies max
6,095
p at or above
0.471
A23B
settings listed
24,535
k4 settings with fit
7,240
copies median
6,619.5
copies max
7,979
p at or above
0.157
A23V
settings listed
24,535
k4 settings with fit
7,344
copies median
6,574
copies max
7,960
p at or above
0.118
A24B
settings listed
23,226
k4 settings with fit
6,853
copies median
6,365.5
copies max
7,262
p at or above
0.157
A24V
settings listed
23,226
k4 settings with fit
6,830
copies median
6,359
copies max
7,280
p at or above
0.176
A25B
settings listed
21,230
k4 settings with fit
5,428
copies median
6,383.5
copies max
7,503
p at or above
0.98
A25V
settings listed
21,230
k4 settings with fit
5,453
copies median
6,382.5
copies max
7,495
p at or above
0.98
A26B
settings listed
19,849
k4 settings with fit
7,681
copies median
6,603
copies max
8,023
p at or above
0.078
A26V
settings listed
19,849
k4 settings with fit
7,658
copies median
6,624
copies max
8,028
p at or above
0.098
B1B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B1V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B2B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B2V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B3B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B3V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B4B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B4V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B5B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B5V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B6B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B6V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B7B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B7V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B8B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B8V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B9B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
6
p at or above
1
B9V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
4
p at or above
1
B10B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B10V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B11B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B11V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B12B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B12V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
0
p at or above
1
B13B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
2
p at or above
1
B13V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B14B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B14V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B15B
settings listed
27,852
k4 settings with fit
118
copies median
106
copies max
136
p at or above
0.255
B15V
settings listed
27,852
k4 settings with fit
88
copies median
112
copies max
144
p at or above
1
B16B
settings listed
27,852
k4 settings with fit
2,839
copies median
2,772.5
copies max
2,938
p at or above
0.216
B16V
settings listed
27,852
k4 settings with fit
2,802
copies median
2,771
copies max
2,937
p at or above
0.373
B17B
settings listed
27,852
k4 settings with fit
725
copies median
811.5
copies max
896
p at or above
0.98
B17V
settings listed
27,852
k4 settings with fit
725
copies median
813
copies max
897
p at or above
0.98
B18B
settings listed
27,852
k4 settings with fit
18,658
copies median
18,859.5
copies max
19,437
p at or above
0.765
B18V
settings listed
27,852
k4 settings with fit
18,838
copies median
18,843
copies max
19,386
p at or above
0.529
B19B
settings listed
27,852
k4 settings with fit
119
copies median
109
copies max
136
p at or above
0.235
B19V
settings listed
27,852
k4 settings with fit
108
copies median
112
copies max
151
p at or above
0.588
B20B
settings listed
27,852
k4 settings with fit
1
copies median
0
copies max
2
p at or above
0.275
B20V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
2
p at or above
1
B21B
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B21V
settings listed
27,852
k4 settings with fit
0
copies median
0
copies max
1
p at or above
1
B22B
settings listed
27,852
k4 settings with fit
150
copies median
163.5
copies max
201
p at or above
0.824
B22V
settings listed
27,852
k4 settings with fit
186
copies median
160
copies max
208
p at or above
0.137
B23B
settings listed
27,852
k4 settings with fit
18,230
copies median
18,289.5
copies max
18,756
p at or above
0.667
B23V
settings listed
27,852
k4 settings with fit
18,215
copies median
18,285
copies max
18,716
p at or above
0.667
B24B
settings listed
27,852
k4 settings with fit
23,286
copies median
23,149
copies max
23,540
p at or above
0.196
B24V
settings listed
27,852
k4 settings with fit
23,286
copies median
23,149
copies max
23,540
p at or above
0.196
smallest class p
0.02
classes
100
key word matches
k4
0
copies median
0
copies max
0
p at or above
1
k4 best readings
none
plants
  1. L
    9
    o
    A
    route
    two turns cw 69,61
    checks
    15
    found
    yes
    matches at setting
    1
  2. L
    7
    o
    B
    route
    K3 rule a=38 b=26 inverse
    checks
    17
    found
    yes
    matches at setting
    1
  3. L
    7
    o
    B
    route
    K3 rule a=74 b=13
    checks
    17
    found
    yes
    matches at setting
    1
  4. L
    10
    o
    A
    route
    K3 rule a=47 b=58 inverse
    checks
    16
    found
    yes
    matches at setting
    1
  5. L
    14
    o
    B
    route
    K3 rule a=47 b=3 inverse
    checks
    11
    found
    yes
    matches at setting
    1
  6. L
    15
    o
    A
    route
    two turns cw 30,94 inverse
    checks
    18
    found
    yes
    matches at setting
    1
  7. L
    13
    o
    A
    route
    two turns cw 31,77 inverse
    checks
    13
    found
    yes
    matches at setting
    1
  8. L
    15
    o
    B
    route
    K3 rule a=68 b=21 inverse
    checks
    9
    found
    yes
    matches at setting
    1
  9. L
    19
    o
    A
    route
    two turns cw 88,51
    checks
    11
    found
    yes
    matches at setting
    1
  10. L
    22
    o
    B
    route
    keyword LANGLEY (repeats dropped) inverse
    checks
    9
    found
    yes
    matches at setting
    3
  11. L
    23
    o
    B
    route
    two turns cw 94,89
    checks
    7
    found
    yes
    matches at setting
    2
  12. L
    21
    o
    A
    route
    columnar 47
    checks
    7
    found
    yes
    matches at setting
    1
plants found
12
plants total
12
texts
51
exhaustive
yes
core seconds
597,255
compute seconds
11
resumed units
697
cost estimate
1.45 million settings x 51 texts, about 0.2 s per setting for all texts and both forms (desk 2026-11-04) plus the key-word step: about 80-120 core-hours; AWS 32 cores about 3-4 h, about $2-3

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 2026-11-05 17:14 UTC; updated after every experiment.
view on GitHub