Experiment 170
F12.4. Stein: 'the process of discovery is often intuitive as well as analytic.' Before any search code, K4 was written out at the widths that matter with the known letters marked and looked at; these six patterns were noted and are tested here exactly as written, each against 1,000 letter-preserving shuffles: (1) the most copies of one letter in one column at width 7 (K: 4 of 8 at places 32, 46, 53, 74); (2) K, the commonest letter, in two adjacent width-7 columns (7 of 8: places 3, 32, 46, 53, 74, 87, 94); (3) the number of repeated bigrams (10: HU, SO, FB, GK, SS, QS, EK, KZ, TJ, DI); (4) how many of their spacings share the factor 4 (5 of 10: 20, 28, 32, 44, 48), Kasiski's test; (5) five of the six S's within 12 places (33-44: SSOTWTQSJQSS); (6) one letter three times at a fixed spacing of 2 or 3 (U at 88, 90, 92). An observation outside every shuffle becomes a new item.
What the method does with the 24 known letters.
We wrote the carved lettersthe carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. out in rows of different widths, as Stein did with pencil and paper, and noted what caught the eye: the letter K lining up in sevens, five S's crowded together, three U's two apart, letter pairs that repeat. Each is then checked against a thousand shufflings of the same letters; the eye finds patterns in random lettersa shuffled control The same experiment run on the carved letters shuffled into a random order. too.
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.
Stein: 'the process of discovery is often intuitive as well as analytic.' K4 written out in rows of 7, 8, 13, 14, 21, 24 and 31, the known letters marked, and every pattern noted before any code.
Six patterns: the letter K lining up in two columns of seven, five S's crowded together, repeated letter pairs and their spacings, three U's two apart; each counted on a thousand shufflings of K4's letters.
Nothing beats every shuffling; the strongest two (1 in 250) were picked out by eye among many possible patterns. No new item.
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.
K4 in rows of 7 (K in capitals, everything else lower case):
o b K r u o x
o g h u l b s
o l i f b b w
f l r v q q p
r n g K s s o
t w t q s j q
s s e K z z w
a t j K l u d
i a w i n f b
n y p v t t m
z f p K w g d
K z x t j c d
i g K u h u a
u e K c a r
K at places 3, 32, 46, 53, 74, 78, 87, 94
Family, key space, the constraint under test, and the clues it reads.
Stein 1999 (knowledge/sources/stein-1999-cracking-the-courtyard-crypto.txt): looking before attacking; F12.2 (0168) measured the fixed battery. These six are by-eye findings, so their selection is post hoc: the multiple-comparison count is the number of patterns a reader could have noticed, larger than six; a pattern counts only if outside every shuffle. Statistic definitions are in the method and fixed before the run.
Written before the run; the result decides it.
None of the six lies outside every shuffle (eye-found patterns in 97 letters are usually chance). Zero candidates.
The harness's record: candidates, gates, loss, and the negative result this experiment leaves behind.
(1) most of one letter in one width-7 column: K4 4, p 0.204; (2) commonest letter's share in two adjacent width-7 columns: K4 0.875, p 0.004; (3) repeated bigrams: K4 10, p 0.04; (4) repeated-bigram spacings divisible by 4: K4 5, p 0.021; (5) shortest window holding all but one copy of a 6-copy letter: K4 12, p 0.004; (6) most equal letters at a fixed step of 2 or 3: K4 3, p 0.198
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.