“The grid by hand: blank cells at the top, or filler letters carried through the second turn, then a repeating key”
97 letters never fill a grid, so a hand turn leaves blank cells. The coded searches (0008, keithadler, nitsuah) put them only at the end of the last row and drop them at once. Here the blanks sit at the start of the first row (the short row at the top), or the second grid is filled with the first grid's blank cells still in it as filler letters that are struck out only at the end, for one and two quarter turns at every width and for column shuffles, each joined to a repeating key on the table before or after the move.
The third messagethe earlier messages The three solved parts of the sculpture. was scrambled by writing it in 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. and turning the grida quarter turn Writing the text in rows of a given width and reading the columns out, each from the bottom up, left column first, as if the grid had been turned on its side.. Its 336 letters fill a 24 by 14 grid exactly. K4the carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page.'s 97 letters fill no grid, so someone working by hand has to leave some squares empty, or fill them with dummy letters. Computer searches so far always left the empty squares at the very end. A person might leave them at the start instead, or carry the dummy letters into a second grid and strike them out only at the end. Each choice moves the letters differently. We try them all, joined to a repeating secret wordthe secret word The word or string a cipher needs besides the method. on 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 see whether any lets a short word fit 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..
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 grid
97 letters fill no grid. By hand the blank cells can go at the start instead of the end, or travel as dummy letters into a second grid and be struck out at the end.
the move
With rows of 10 and the blanks on top, the E of place 22 comes out at 45 instead of 18: each choice is a different scramble.
the bound
120,871 such scrambles with a repeating key of 1 to 26 letters before or after: no period below 22 fits, on the copper or the scrambled coppers. Period 23 runs a little high, spread evenly over every choice: chance.
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 places in rows of 10. Ten rows hold 100 cells, so 3 stay blank; put them at the start of the top row, so place 1 sits in column 3.
Turn the grid a quarter clockwise: read each column from the bottom up, left column first, skipping blanks. Columns 0-2 give 9 letters each and column 3 gives 10, so column 4 starts at letter 38.
Place 22 is in row 2, column 4, the eighth letter read from the bottom of that column: it lands at place 45. With the blanks at the end instead it would land at 18.
The repeating key letter at the landing place must turn E into the copper letter there; work out that key letter for all 24 known letters.
For a key of length p, two known letters whose landing places differ by a multiple of p use the same key letter; if they need different ones, p is dead.
Repeat for every width, blank placement, second turn, filler choice and order, and on ten scrambled coppers.
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.
width 10, 97 letters: 10 rows, 3 blank cells at the start of the top row ('.' = blank, numbers = places)
. . . 1 2 3 4 5 6 7
8 9 10 11 12 13 14 15 16 17
18 19 20 21 22 23 24 25 26 27
...
88 89 90 91 92 93 94 95 96 97
one turn clockwise (columns bottom-up, left first): where the known letters land
place letter blanks on top blanks at the end
22 E 45 18
23 A 55 28
64 B 61 34
65 E 71 44
5
Hypothesis, for a cryptanalyst
Family, key space, the constraint under test, and the clues it reads.
Type 1 of directions.md, gap a1 of the prior-work database ('rotation with per-stage padding (nulls dropped between the two griddings), the short row at the top, or blanks in key-determined cells... all coded searches pad once, at the ends'). Clues C1(a) (02-sanborn-said.md, CNN 2005, 'double matrixes and triple matrixes'), C2(a) (the K3 matrix and its shifts; K3's 336 letters fill 24 x 14 and 8 x 42 exactly, K4's 97 fill nothing), C17(b) ('whether it's 97, whether it's 100... I seem to like to manipulate the information'). Grid of width w and h = ceil(n/w) rows; the h w - n blank cells are the first cells of the top row ('top') or the last of the bottom row ('bottom', the covered case). A quarter turn reads columns bottom-up, left first (cw) or top-down, right first (ccw). Two turns: the first grid's blank cells either vanish ('drop') or travel as filler letters into the second grid and are struck out at the end ('keep'). New routes: one turn cw or ccw with blanks on top, widths 2-96; two cw turns at every width pair 2-96 with blanks top or bottom in each grid and drop or keep (seven of the eight combinations; bottom-bottom-drop is 0008's); plain columnar (both directions) and vocabulary-keyed columnar with the short row on top; all forward and inverse, deduplicated. Key: a periodic tableau key of period 1-26 after the move (key slot = landing place mod p) or before it (slot = place mod p), 8 distinct classes of plain and cipher alphabet and Beaufort-or-not. A period is live if no two known letters in one slot force different key letters. Exact; ten shuffled coppers through the identical search.
6
Prediction
Written before the run; the result decides it.
No route, order and class has a live period below 20 on K4 beyond what the ten shuffled coppers show; the count of live settings at every period 1-26 lies inside the control range, give or take the spread at 25-26. Zero candidates.
7
Method
Build every route: write the 97 places into the grid with the blanks on top or below, turn, and for two turns either drop the blanks or carry them as fillers into the second grid.
Keep one route per distinct reordering.
For each route, class and text, compute the key letter each known letter forces at its landing place.
For each period 1-26 and each order, the period is live if no slot holds two different forced letters.
Count live settings per period on K4 and on ten shuffled coppers.
List K4's live settings at periods 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
120,871 distinct routes from 127,234 specs x 8 classes x 2 orders, periods 1-26. K4 live settings below period 20: 0; per control (A then B): [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]. Periods where K4 is outside the control range: {'A': [23, 24], 'B': []}. Planted text found: True.
What was searched, as the harness recorded it.
family
hand-padded grid turns and short-row-on-top columnar, joined to a periodic tableau key before or after