Experiment 236
F13.4c (issue #85). F1.5's design 2 (0140's compiled parallel tempering) is run with no known letter: every key slot free, no check-keeping, the alphabet and key climbed together by quadgrams, a second form (added) beside 0140's Beaufort. A probe before this run (one tempering run per plant) read 54 of 97 letters of a plant at length 5 and nothing at 10 or 15, so power comes first, in the same run: 8 plants at each of the lengths 2-8, 10, 13, 16, 19 (no known words; half X stops; half each form), each against 6 own shuffles climbed the same way (null 1/7), 3 tempering runs per text, best reading rescored by MIX8m. Then K4 and 50 shuffles at every length 1-19 in both forms; the statistic is the largest z-score over the cells the plants decide (6 of 8). Cost stated before the run: 616 + 1,938 climbs x about 10 s, about 7 core-hours.
What the method does with the 24 known letters.
The same as before, but now 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.'s alphabet is also unknown: any of the 400 million million million million orders of the 26 letters. Without the clue wordsthe known letters Sanborn has confirmed that positions 22 to 34 of the message read EASTNORTHEAST and positions 64 to 74 read BERLINCLOCK. the computer has to guess the alphabet and the secret wordthe secret word The word or string a cipher needs besides the method. together from how English the result looks. First we find out, with hidden test messages, up to what word length that can work at all; only there do we compare K4the carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. with shuffles of itself.
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.
F13.4c: K1-K2's system with the alphabet unknown, solved from the carved letters alone: alphabet and key guessed together.
Power first: planted messages at lengths 2-19, each against its own shuffles; then K4 against 50 shuffles where the plants are found.
Plants are found to length 5; there K4 sits inside its shuffles (p 0.37). Lengths 6-19 cannot be decided this way.
What this experiment was built on, from the lineage page: 12 experiments before it, back to the root of its line. Its number is in bold.
Laid out as the copper carries K4: OBKR ends the line that finishes K3, then three lines of 31. Left, as carved, the 24 known places in accent. Right, the candidate: a known place it gets right is an accent cell, one it misses is outlined. This is the candidate with most known letters right; none passed every gate.
in: the copper, always the same
OBKR ends the line that finishes K3, then three lines of 31
out: the candidate
checked: the known letters in place, and the reading turns back into the copper
The loss of the best candidate, part by part. Zero is solved.
| Known letters missed | 19 of 24, ten points each | 190.00 |
| English | the 73 free letters score -5.36; English text scores -3.6 | 1.76 |
| Key | 114 bits to write it down (key words are cheap, arbitrary letters 4.7 bits each); each 60 beyond 60 costs a point | 0.90 |
| Loss | of 240, the worst possible | 192.66 80.3% |
Redo the method on a few letters with pencil and paper.
The same steps run by the code on the real letters.
A planted message from K3's own words, a shuffled alphabet and a key of 3, Beaufort form, no known letter:
alphabet, true CKEMJNZWSYGDRPVFBHOAQULXIT
carved TUMWQZNVALJNNTWAEBRQETUMIURULMRRRMJMPPVLXYNYMTQBEVNXGASHZDCMBGCTBLPGZNVYBQBTUMRQPUNWTUWAGELCULK
alphabet, found TUYVAIMHKQPEWRJLDZXBSCNGFO
reading THEJOTAIRESCAKINGFROMTHECHAOBERCASSEDFJEDLAMETOFLICKERMUTPREHENTLYDETAILHOFTHEROHOWITHINEMERDETRN
letters right: 81 of 97
Family, key space, the constraint under test, and the clues it reads.
F1.1 (Quagmire III with an unknown order: K1-K2's system with Sanborn's 'slightly different' chart, C27(c)) read without the 24 known letters: the hedge for A1/A2 on the family closed exactly at 1-22, 24, 25 (0119) and reopened at Beaufort 23 and 26 (0228, 0229). F13.1 puts one free alphabet decidable by quadgrams to length 19 (80 bits for the alphabet plus 4.7 per key letter, under 173), which is necessary, not sufficient: a climb must also find the key, and with no known letter design 2 has no anchor. Pre-registered: a cell (length, form) is decided when the planted length bracketing it from above found at least 6 of 8 (lengths below the shortest failing planted length); S = max z over decided cells against 50 shuffles; K4 flags at p < 0.05. If no length is decided the item is blocked by power with the numbers.
Written before the run; the result decides it.
Plants found at lengths 2-4, some at 5-6, none from 8: the decided range ends near 5. K4 inside its shuffles in that range (p above 0.05). F13.4c closed for the short lengths, blocked from about 6 by power.
The harness's record: candidates, gates, loss, and the negative result this experiment leaves behind.
All 10 candidates, lowest loss first.
plant power L2 8/8, L3 8/8, L4 7/8, L5 7/8, L6 4/8, L7 5/8, L8 3/8, L10 3/8, L13 2/8, L16 4/8, L19 3/8; decided to length 5; K4 statistic 1.67, p 0.373
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.