“Hand padding at each grid stage: the blank cells in any of eight places, dropped or carried as fillers into the next grid, one or two stages, then or before a repeating key on the table”
F6.4 (padding applied separately at each grid stage; nulls dropped between griddings). 0082 tried blanks at the start of the first row or the end of the last, and carried as fillers, for one and two clockwise turns and short-row-on-top columnar. The list of hand rules (NOTEBOOK 2026-10-16) adds six places for the blank cells: start of the last row, end of the first row, and down the first or last column from the top or from the bottom; each grid of a two-stage move chooses its own place, its own reading (clockwise turn, or Sanborn's rows-then-columns-down) and whether the first grid's blanks vanish or travel as filler letters to be struck out at the end. One stage: 8 places x 4 readings x widths 2-96. Two stages: (8 places x 2 readings x widths 2-48)^2 x drop or keep.
97 letters never fill 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. exactly, so a person writing them in rows leaves a few empty squares somewhere: at the end, at the start, or down one side. With two grids, the empty squares of the first can disappear or be copied along as dummy letters and crossed out at the end. We try every such choice, for one and two grids, each followed or preceded by a repeating secret wordthe secret word The word or string a cipher needs besides the method. on the sculpture's tables, and check them against 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. and scrambled copies of the carved textthe carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page..
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
97 letters never fill a grid, so a hand worker leaves empty squares: at the end or start of a row, or down a side column. With two grids, the first grid's empty squares can vanish or travel on as dummy letters, struck out at the end.
the check
Eight places for the empty squares in each grid, drop or keep, one grid (4 readings, widths 2-96) or two (clockwise turn or Sanborn's matrix, widths 2-48): 874,572 distinct routes, each with a repeating key on the sculpture's tables.
the bound
No key shorter than 21 fits K4 after any route; the coppers reach 20. K4 is above every copper in 0 of 52 cells, a copper 0-2 against the rest.
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.
Write the 97 letters in rows of 8: 13 rows hold 104 squares, so 7 stay empty; leave them at the bottom of the last column and the six above it (place P6).
Read the columns left first, top to bottom (Sanborn's matrix), and copy each empty square out as a dummy letter (keep): 104 symbols, the E of place 22 now 68th among the real letters.
(In drop mode the empty squares are simply skipped.)
Write the 104 symbols in rows of 9: 12 rows hold 108 squares, so leave the first 4 squares of the top row empty (P2); read the columns from the bottom up, left first (a quarter turn), and strike out every dummy: the E lands at place 90, where the copper has U.
Note where each known letter has landed and the carved letter there; with a key of length p on the table, two known letters in one slot must force the same key letter.
Repeat for every place, reading, width, drop or keep, the key before or after, and on scrambled copies.
the worksheet: the first steps by hand, on squared paper
4
Worked example, generated
The same steps run by the code on the real letters.
rows of 8, blanks down the last column from the bottom, read down the columns left first, blanks kept as dummies;
then rows of 9, blanks at the start of the first row, read up the columns left first; dummies struck out
place letter after grid 1 after grid 2 copper there
22 E 68 90 U
23 A 81 34 S
24 S 94 76 G
64 B 22 84 D
65 E 35 29 R
74 K 62 26 Q
5
Hypothesis, for a cryptanalyst
Family, key space, the constraint under test, and the clues it reads.
C1(a) (Sanborn, CNN 2005, knowledge/sources/cnn-2005-sanborn-interview-transcript.txt: 'you write a line of plain text out, right, in successive rows. And then basically you write it going from top to bottom... double matrixes and triple matrixes'), C2(a), C17(b) ('whether it's 97, whether it's 100... I seem to like to manipulate the information'). 97 is prime, so every hand grid has h w - 97 blank cells; research note combined_methods_and_sevens.md: classical practice leaves the last row short, completes with nulls, or fixes blank cells; every coded search pads once at the ends (gap 6: per-stage padding). Places for b blanks: end of the last row (P1, covered by 0008), start of the first row (P2, 0082), start of the last row (P3), end of the first row (P4), last column from the top (P5) or bottom (P6), first column from the top (P7) or bottom (P8); column places need b <= h. Readings: cw (left column first, bottom up: the K3 turn), col (left first, top down: Sanborn's matrix code); for one stage also ccw and colr. Keep mode: the first grid's blank cells are read out as filler letters and fill cells of the second grid, all fillers struck out at the end. Key: periodic tableau 1-26 after (order A) or before (order B) the move, 8 classes of standard/KRYPTOS plain and cipher alphabets and Beaufort-or-not; a period is live if no slot holds two different forced key letters. Exact; ten scrambled coppers.
6
Prediction
Written before the run; the result decides it.
As in 0082 and 0083: nothing live below period 20 beyond the coppers' rate, and K4 above every copper in no more (order, period) cells than the most any single copper manages against the rest. A K4 count beyond that null stops the queue for a replication. Zero candidates.
7
Method
For every width, reading and place of the blanks, the where-map of one stage (keep mode: fillers counted).
Compose every pair at widths 2-48 (both readings, all places, drop or keep), forward and inverse; skip repeats.
For each route, class and text, the key letter each known letter forces at its slot.
Period p (1-26) is live in an order if no slot holds two different forced letters.
Count live settings per period on K4 and ten scrambled coppers; compare cell by cell; list K4's live settings below 20.
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
nonechampion before
874,572 routes (3040 of 3040 first stages) x 8 classes x 2 orders, periods 1-26. K4 settings live below 20: 0 (A), 0 (B); coppers A 0-0, B 0-0. K4 above every copper in 0 of 52 cells; null 0-2. Stop queue: False.
What was searched, as the harness recorded it.
family
hand padding per grid stage: blank cells in one of eight places, dropped or carried as fillers; one or two stages; then or before a periodic tableau key
places
P1
end of last row
P2
start of first row
P3
start of last row
P4
end of first row
P5
last column from the top
P6
last column from the bottom
P7
first column from the top
P8
first column from the bottom
readings
cw
left column first, bottom up
ccw
right first, top down
col
left first, top down
colr
right first, bottom up
singles
8 places x 4 readings x widths 2-96
doubles
(8 places x readings cw/col x widths 2-48)^2 x drop/keep