“F6.1: one free alphabet after or before every type-1 move, key lengths 6-10, every setting K4 fits decided by D1's one-run decider, with 20 shuffled stand-ins through the same procedure”
F6.1 at lengths 6-10, reopened by D2 (0252: 7,443 settings, one-run power 31 of 48 on plants drawn from K4's own fitting settings). 0253's decide() unchanged (one tempering run of 0252's climb on K4, then shuffles the exact decider fits at the same setting climbed one at a time, stop at the first at or above, at most 12; no escalation, so K4 and the stand-ins are decided alike), now on 0129's move library: every (move, order, length 6-10, form) that fits K4. Twenty full shuffles of K4 stand in for it (0254/0255: a count over settings needs texts put through the whole procedure), each on 300 of its own fitting settings.
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., read it out another way, and then used the cipher table. With his own scrambled alphabet in 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. and a short key wordthe secret word The word or string a cipher needs besides the method., thousands of such recipes agree with 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.. For each one we search for the most English-looking reading of K4the carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. and do the same on scrambled copiesa shuffled control The same experiment run on the carved letters shuffled into a random order.; twenty scrambled stand-ins for K4 go through the very same test, so we know what 'K4 wins often' would mean.
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
A grid move, then the table with one scrambled alphabet and a key of 6-10 letters: 7,443 such recipes fit the 24 known letters.
the check
Every recipe searched once on K4 and on shuffles that fit it, up to 12; 20 shuffled stand-ins through the same test.
the bound
K4 beat all 12 copies in 570 recipes (7.7%); 8 of 20 stand-ins at or above. Plants 6/12 clean, 5/12 Sanborn-style: closed at that power.
2
Experiment lineage
What this experiment was built on, from the lineage page: 19 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.
Pick a recipe: write the message in rows of 9 and read the columns in the order of a key word (the move), then encipher on the table with an unknown scrambled alphabet and a 7-letter key.
Undo the move on paper for the 24 known letters: each lands on a carved place and a key slot.
Check that some alphabet makes every pair of known letters in one slot agree.
Search alphabets and keys for the most English-looking reading of K4 under the recipe.
Shuffle K4 until the shuffle also passes step 3, search it the same way; stop when a shuffle reads as well as K4, at most 12 shuffles.
Do steps 4-5 for every recipe K4 fits, and for twenty scrambled stand-ins of K4; compare how often each wins all 12.
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.
F6.1: one free alphabet around a move, key lengths 6-10, decided setting by setting.
setting columns of a 9-wide grid read in key-word order, then the table, key length 7, Beaufort
slot map each carved letter's key slot and message place follow from the move
K4 gets one tempering search under that setting; shuffles of K4 that fit the same 24 checks get the same,
one at a time, until a shuffle reads as well as K4 (at most 12).
Twenty shuffled stand-ins for K4 go through the whole procedure: K4's rate of wins must stand out from theirs.
6
Hypothesis, for a cryptanalyst
Family, key space, the constraint under test, and the clues it reads.
F6.1 (C1(a): 'a matrix code... then went back to this table'; C27(c): a 'slightly different' tableau): one free alphabet on both sides (Quagmire III with an unknown order), a periodic key of length 6-10, after or before one move of 0129's type-1 library (27,852 routes: columnar, keyword columnar, K3's rule, turns, rail fence, grid paths...), Beaufort or added form. 0129/0130 found no fit below 6 and fits from 6 at the copies' rate, so the known letters cannot decide these settings and English must. Decider: D1's one run of 0140's tempering generalised to any slot map and reading order (0252), copies fitted exactly at the same setting (0253). Clue entries C1(a), C27(c), C11.
7
Prediction
Written before the run; the result decides it.
K4's rate of settings above all 12 copies inside the 20 stand-ins' spread (p above 0.05); plants (24, on settings K4 fits) found about 14 of 24; no reading English. Zero candidates.
8
Method
Settings: 0129's route library (27,852 moves) x order (A move then table, B table then move) x key length 6-10 x form (added, Beaufort); each decided exactly on K4 by 0252's fits(); the fitting ones are K4's cell.
Per setting (one job, own seed): 0253's decide() without escalation: one tempering run on K4, then shuffles that fit the same setting climbed one at a time until one reaches K4, at most 12.
Stand-ins: 20 shuffles of K4 (seed 2680), each with its own fitting settings found the same way, 300 drawn and decided alike.
Plants: 12 clean and 12 Sanborn-style messages on settings K4 fits, random alphabet and key; found if above all 12.
Verdict: K4's share of settings above all 12 copies, ranked among the stand-ins' shares.
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
F6.1 lengths 6-10: 7443 settings fit K4; K4 above all 12 copies in 570 (7.7%); 20 stand-ins 3.0-13.3%, 8 at or above K4 (p 0.43); plants clean [6, 12], Sanborn-style [5, 12].
What was searched, as the harness recorded it.
item
F6.1 lengths 6-10 (reopened by D2, 0252)
decider
0253's decide() unchanged, without escalation: one run of 0252's generalised tempering (16 x 625 x 1000) on the target, then shuffles the exact decider fits at the same setting, climbed one at a time, stop at the first at or above, at most 12
every type-1 move of 0129's library, both orders, key lengths 6-10, both forms, that K4 fits on one free alphabet; each decided at D1's one-run power
resumed units
1,028
cost estimate
about 7,443 K4 settings (0252) and 6,000 stand-in settings x about 4.1 one-run climbs (H_12 copies) at 17-25 core-s, plus 21 exact scans of the library (about 18 core-minutes each) and 24 plants: about 300-400 core-hours; on AWS c8g.8xlarge (32 cores, $0.60/h): about 10-13 h, about $6-8
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.