Experiment 53

“The repeating secret word again, on tables rebuilt from the words of the solved messages”

Experiments 20 and 48 rebuilt the table's alphabets only from the puzzle's own vocabulary list. Here the alphabets are rebuilt from every word of the three solved messages and from the words Sanborn's 2025 letter points at (EGYPT, WALL, CAIRO, MESSAGE, DELIVER and others), and the crib test is rerun at every period from 1 to 52, with shuffled copies of the copper as the control.

revertedstatus
experiment 20parents
86 stime, all phases
0candidates
0/24known letters
1

The experiment method

What the method does with the 24 known letters.

The first two messages used 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. with a secret word repeated over and over. The table itself can be rebuilt from a word, and so far we rebuilt it only from the puzzle's own word list. The solved messages are full of other words, and Sanborn's last letter mentions Egypt and a message being delivered, so we rebuild the table from all of those and ask again, for every length of secret word up to 52, whether 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. can agree. Scrambled copies of the copper get the same test, so we can tell real agreement from luck.

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.

OBKRUOXOGHULBSOLIFBBWEASTNORTHEASTOTWTQSJQSSEKZZWATJKLUDIAWINFBBERLINCLOCKWGDKZXTJCDIGKUHUAUEKCARIN: THE COPPER, ALL 97 LETTERS, THE 24 KNOWN PLACES IN ACCENTSHADING for the message, INVISIBLE for the copper, a secret word of 7: places 22 and 29 share a slotEASTNORTHEASTBERLINCLOCK22 needs key 1, 29 needs key 0: length 7 is dead for this pairsettings consistent at each length: the copper (ochre) against the mean of ten shuffled coppers (grey)2526272829303132lengths 1 to 24 and 33 to 49: none on either; heights on a log scaleOUT: NO READING; NOTHING BEYOND CHANCE, NO PUZZLE WORD FITSEASTNORTHEASTBERLINCLOCK
the tables

Every word in the K1 to K3 solutions, plus the 2025 words, was turned into a keyword alphabet: 102 tables, 10,404 pairs, three ways of using them, secret words of length 1 to 52.

the check

Drawn for SHADING and INVISIBLE: places 22 and 29 are seven apart, so a 7-letter word gives them the same key letter, and they need different ones.

the bound

At every length the copper has about as many consistent settings as a shuffled copper, and no forced pattern matches a puzzle word: the bigger vocabulary changes nothing.

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.

  1. Take the word SHADING from K1 and build its alphabet: SHADINGBCEFJKLMOPQRTUVWXYZ.
  2. Take INVISIBLE from K2 for the carved letters: INVSBLEACDFGHJKMOPQRTUWXYZ.
  3. Place 22 holds E under carved F: counting from 0, E is 9 in the first alphabet and F is 10 in the second, so Vigenere needs key 1.
  4. With a secret word of 7 letters, places 22 and 29 share a slot, so place 29 (T under carved R) must also need key 1.
  5. T is 19 in the first alphabet and R is 19 in the second, so it needs 0: a contradiction, and length 7 is dead for this pair.
  6. Repeat for every pair, mode and length.
the worksheet: the first steps by hand, on squared paper
SHADING FOR THE MESSAGE, INVISIBLE FOR THE COPPER, VIGENERE, A SECRET WORD OF 7 LETTERSSHADINGINVISIBLESI0HN1AV2DS3IB4NL5GE6BA7CC8ED9FF10JG11KH12LJ13MK14OM15PO16QP17RQ18TR19UT20VU21WW22XX23YY24ZZ25placeknowncarvedslot of 7key = carved - known22E=9F=100129T=19R=190023A=2L=51330H=1N=110Places 22 and 29 share slot 0and need keys 1 and 0; 23 and30 share slot 1 and need 3 and0.One disagreement is enough: length 7 is dead on this pair oftables.
4

Worked example, generated

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

plain  SHADINGBCEFJKLMOPQRTUVWXYZ
cipher INVSBLEACDFGHJKMOPQRTUWXYZ

place  plain  carved  key (Vigenere)  slot at period 7
   22    E      F          1               0
   29    T      R          0               0
   23    A      L          3               1
   30    H      N          0               1
5

Hypothesis, for a cryptanalyst

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

Periodic tableau with separate plain and cipher alphabets (Quagmire family), alphabets keyword-mixed from words outside the 62-word vocabulary: every distinct word of the K1-K3 plaintexts with at least three distinct letters (C7, C30: earlier text may carry the key) and the C13/C14 words EGYPT, CAIRO, LUXOR, WALL, NOVEMBER, MESSAGE, DELIVER, DELIVERING, TRANSMIT, plus standard and KRYPTOS. For each plain-cipher pair and mode (Vigenere, Beaufort, variant Beaufort) the 24 cribs force 24 key indices; period P is consistent when every two cribs sharing a slot force the same index. Count consistent (pair, mode) cases per period on K4 and on ten shuffles of the copper with the same pool. At consistent periods up to 29, read the forced slots in the standard, KRYPTOS and plain alphabets and match against every word and every K1-K3 window of that length; a full match decrypts and submits. Bounds: complete for this pool and periods 1-52.

6

Prediction

Written before the run; the result decides it.

Periods 1-26 and 30-52 are consistent on K4 no more often than on the shuffles (near zero below 21); 27-29 are consistent for most pairs, as before, and no forced fragment matches a word or a K1-K3 window. Zero candidates.

7

Method

  1. Build the word pool and its keyword alphabets, dropping duplicates and those already in the vocabulary pool.
  2. For every plain and cipher alphabet and mode, compute the 24 forced key indices.
  3. For each period, check the slot-sharing pairs.
  4. Repeat on ten shuffles.
  5. At consistent periods up to 29, match forced fragments against words and K1-K3 windows; decrypt and submit matches.
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
34champion before

102 alphabets, 10404 pairs, 3 modes, periods 1-52. Consistent cases on K4 by period (shown where K4 or a control has any): 25: 0 (controls mean 0.4, max 2), 26: 2139 (controls mean 1166.2, max 1554), 27: 31212 (controls mean 31212.0, max 31212), 28: 31212 (controls mean 31212.0, max 31212), 29: 31212 (controls mean 31212.0, max 31212), 30: 1339 (controls mean 1163.2, max 1699), 31: 47 (controls mean 38.1, max 59), 32: 6 (controls mean 0.5, max 3), 50: 0 (controls mean 0.4, max 2), 51: 10 (controls mean 26.1, max 74), 52: 2139 (controls mean 1166.2, max 1554). Word matches: 0.

What was searched, as the harness recorded it.

family
periodic tableau, alphabets keyword-mixed from K1-K3 words and C13/C14 words
alphabets
102
pairs
10,404
modes
3
periods
1-52
controls
10
consistent by period k4
25
0
26
2,139
27
31,212
28
31,212
29
31,212
30
1,339
31
47
32
6
50
0
51
10
52
2,139
consistent by period controls mean
25
0.4
26
1,166.2
27
31,212
28
31,212
29
31,212
30
1,163.2
31
38.1
32
0.5
50
0.4
51
26.1
52
1,166.2
consistent by period controls max
25
2
26
1,554
27
31,212
28
31,212
29
31,212
30
1,699
31
59
32
3
50
2
51
74
52
1,554
word matches
0

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.
view on GitHub