“F13.4c at key lengths 6-19 (D3): one free alphabet with a periodic key, ciphertext only, K4 against 100 fresh shuffles climbed by 0236's frozen search, closed at the measured power”
D3 (issue #77, Alexis #100: run power-blocked cells at their measured power). 0236 measured its search's power at 6-19 (plants 2-5 of 8) and climbed K4 there, but decided and recorded K4 only at 1-5, where power was 7-8 of 8. Here 0236's per-text search, unchanged (imported from its record), on K4 and 100 fresh shuffles at every length 6-19 in both forms, every cell recorded; plants at 6, 7, 8, 10, 13, 16, 19 redrawn in the same run (8 each, 6 own copies).
revertedstatus
a new rootparents
2.4 htime, all phases
28candidates
3/24known letters
unclocked
1
The experiment method
What the method does with the 24 known letters.
Suppose Sanborn used the cipher table with his own scrambled alphabet and a repeating key of 6 to 19 letters, and suppose we did not know any of the words he gave away. Our search can still find such messages sometimes, two to five times in eight. We run it on K4the carved letters The 97 letters cut into the copper at the CIA, shown at the top of every experiment page. and on 100 scrambled copiesa shuffled control The same experiment run on the carved letters shuffled into a random order. of K4 and ask whether K4 reads better than the copies. Only after that do we look at whether the known words came out in place.
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
F13.4c: one scrambled alphabet with a repeating key of 6-19 letters, solved from the ciphertext alone; 0236 measured this search's power but judged K4 only to length 5.
the check
K4 and 100 fresh shuffles at every length and both forms, 3,220 searches; planted messages redrawn in the same run.
the bound
K4's best cell 11/beaufort, p 0.84; planted messages found 2-5 of 8: closed at that power.
2
Experiment lineage
What this experiment was built on, from the lineage page: 13 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
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
as carved, 97 letters; the 24 known places in accent
ECDMRIPFEIMEHNLSSTTRTVDOHW?OBKR
UOXOGHULBSOLIFBBWFLRVQQPRNGKSSO
TWTQSJQSSEKZZWATJKLUDIAWINFBNYP
VTTMZFPKWGDKZXTJCDIGKUHUAUEKCAR
OBKR ends the line that finishes K3, then three lines of 31
out: the candidate
every place read; known places in accent, misses outlined
ECDMRIPFEIMEHNLSSTTRTVDOHW?BTIS
AGYPINVARREDSDODESPSPELOTHICERE
SFACAFIEROFANESREIVALSTELASRUDI
VAREPSLECOURPOREACBONISTNUSEWHO
checked: the known letters in place, and the reading turns back into the copper
4
How the score is made
The loss of the best candidate, part by part. Zero is solved.
Known letters missed
21 of 24, ten points each
210.00
English
the 73 free letters score -5.05; English text scores -3.6
1.45
Key
137 bits to write it down (key words are cheap, arbitrary letters 4.7 bits each); each 60 beyond 60 costs a point
1.29
Loss
of 240, the worst possible
212.74 88.6%
5
Walkthrough, by hand
Redo the method on a few letters with pencil and paper.
Write a scrambled alphabet and a key of, say, 11 numbers.
To read the ciphertext, take each letter's place in the alphabet, subtract the key number for that position (Beaufort: subtract the place from the key), and read the alphabet at the result.
A computer searches alphabets and keys for the most English-looking reading of K4, without looking at the known words.
The same search runs on 100 scrambled copies of K4.
K4 counts only if it reads better than every copy; then the 24 known letters are checked.
the worksheet: the first steps by hand, on squared paper
6
Worked example, generated
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
7
Hypothesis, for a cryptanalyst
Family, key space, the constraint under test, and the clues it reads.
F13.4c: a periodic key on one free alphabet (Quagmire III with an unknown order, both forms), solved from the ciphertext alone: the search never reads the 24 known letters (program.md, F13); gate 2 checks them at keep time. 0236's design-2 tempering without known letters, 3 runs x 16 replicas x 625 rounds x 1000 moves, MIX8m rescoring (0233). F13.1 puts one free alphabet with a key within reach of 97 letters to length 19 by quadgrams; 0236 found the search's power falls from 7/8 at 5 to 2-5/8 at 6-19. Statistic: per cell, K4's MIX8m against the shuffles' (z); over the 28 cells the largest z against the shuffles' own largest z. Clue entries C27(c), C5(a).
8
Prediction
Written before the run; the result decides it.
K4 inside its shuffles: largest z over the 28 cells with p above 0.05; no cell where K4 beats all 100. Plant power as 0236 (2-5 of 8 at 6-19). K4's readings put at most 5 of the 24 known letters in place (harness gate 2).
9
Method
Texts: K4 and 100 shuffles of K4 (seed 2560).
Each text at every key length 6-19 and both forms (Beaufort, added): 0236's climb(), unchanged.
Plants: 8 per length at 6, 7, 8, 10, 13, 16, 19 (0236's rule: Carter test windows, half with X stops, a random alphabet and key, forms alternating), 6 own shuffles each; found when above all its own copies.
Per cell: K4's rank among the 100 and its z; the statistic is the largest z, its p from the shuffles' own largest z.
10
Result and bounds
The harness's record: candidates, gates, loss, and the negative result this experiment leaves behind.
28candidates returned
0passed every gate
3/24known letters, best
88.6%lowest loss
nonechampion before
All 28 candidates, lowest loss first.
1failed at cribs3/24 known lettersloss 88.6%
ECDMRIPFEIMEHNLSSTTRTVDOHW?TION
TINTSDOUNTAWOSTYBEINOBELULDCUTA
FLARECOUTCLINOGAROSTJOHOSPENIFR
JAYINEYAMEYEEWARHMSEHOFTHATARSP
2failed at cribs3/24 known lettersloss 88.6%
ECDMRIPFEIMEHNLSSTTRTVDOHW?HPER
LYONOODEPATUSESTARESINGATOONIVC
LISECTIVILTHEYPRTLEDGASIENESSFA
SKLLORNHEDUNDNONXIYONDONSHIMSAH
3failed at cribs3/24 known lettersloss 88.6%
ECDMRIPFEIMEHNLSSTTRTVDOHW?BTIS
AGYPINVARREDSDODESPSPELOTHICERE
SFACAFIEROFANESREIVALSTELASRUDI
VAREPSLECOURPOREACBONISTNUSEWHO
4failed at cribs3/24 known lettersloss 88.7%
ECDMRIPFEIMEHNLSSTTRTVDOHW?PLFA
MBUTAMENLLIELADMOWNATERUDAAGNTI
NSGOVEHINDHARLSSEYWERYISSHADRAS
SUNDOTOWNTOAREUNNEUAGEONINAVIJC
5failed at cribs3/24 known lettersloss 88.8%
ECDMRIPFEIMEHNLSSTTRTVDOHW?BYOU
RINICVELOAKILLACIESTUNNVUTHESAF
REINYMGODTOANESIMAFTSWSYSNEYESH
INEMAFSOCOFEEVERMOGPORTABATOMER
6failed at cribs3/24 known lettersloss 88.8%
ECDMRIPFEIMEHNLSSTTRTVDOHW?TNEP
ROCUTGLATTHTHINNIALTHOWSGODEONT
OMSBSWOMBUTYAMPEDESASIDENUESONI
NECALOWKOTINGERSTSHEEPERALCASDG
7failed at cribs3/24 known lettersloss 88.8%
ECDMRIPFEIMEHNLSSTTRTVDOHW?EACT
CINEWDEBSNFRAMALISPORNISTITEATE
NIRECESTRUCEITSSALBELOWNAPALIAA
TSTENABYALEANDYPRIALLEOCTTOMEFT
8failed at cribs3/24 known lettersloss 88.8%
ECDMRIPFEIMEHNLSSTTRTVDOHW?GUEN
MATFIFOUREHOMATIJETABENSURESHIT
LEFEETACRICAWARFSHERATOESKHIRLA
INFULVENALAIRANEONBYNISLUDLESFY
9failed at cribs3/24 known lettersloss 88.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?MENC
ANIEDKMONMTSNOHEALIMAUROLEMILLE
MEUMURDERTHESHNONTHINETHTYMANBI
STOCLONLANTATDELCALEGRANCTIOOKW
10failed at cribs2/24 known lettersloss 92.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?KDIS
LALSMALPANSEMHADTOBYOFFITWHINGC
ESOFGOLINALAMSVONGPEEURROURDMEF
OOODAHMITBYTRGONSKDMIYDSTBALSUP
11failed at cribs2/24 known lettersloss 92.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?FRMT
ODBEEMARHADSHYCRAZELINBNUSSUNTE
NEOFAPMTOIBEHINENTWHINNITESULWE
GLEDAYNSASBOURNAPORSOFGCTATUMWV
12failed at cribs2/24 known lettersloss 92.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?RALA
NCHERUKRHERERGEDELANETANPUCNORA
NSOLDICHZANSHEMICAREHATOEMINGUS
EOFTWOINHEOFNARCYHAILORSUITIOFM
13failed at cribs2/24 known lettersloss 92.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?YKYA
LITSSHIEFBIMTOCKABUCHNAMPSTHASS
EDTRIFYSOTENFEDLCARTOPRIONPUALI
COLLCOMBATAMABEZACSEMERBNILLMSO
14failed at cribs2/24 known lettersloss 93.0%
ECDMRIPFEIMEHNLSSTTRTVDOHW?HSAB
RTWHASCRITTAMESTSTANDTEDAPSALLS
ORETULEMTHORRICONOFFICOOLYESOMN
IESTUNDESSFARHONLIPAWTMADNGANUE
15failed at cribs1/24 known lettersloss 96.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?IROR
CHEINNINCDISTHECASSPAPROUGZETUP
MVIGTENFLDERMANOFENSSOFCULLIACR
ASARDAREKISHKMTOBSONOIMSWRIODIQ
16failed at cribs1/24 known lettersloss 96.9%
ECDMRIPFEIMEHNLSSTTRTVDOHW?OVEO
FKTONATCOMOEYSTOAFEDEMABRWOAEND
TOFFECHSTBAURAABRACHSIDINGIGNAL
ESWIZULANDSKMCCOMSINETIHEUSEDYS
17failed at cribs1/24 known lettersloss 97.0%
ECDMRIPFEIMEHNLSSTTRTVDOHW?MAAH
ISDEGONLEDDANDINWYERULEFMYOUSMS
TORENTFUIEROZONLOSECTISHICEPLOM
NOTTOFTHUSABMATERDAMARDNTPERRAG
18failed at cribs1/24 known lettersloss 97.1%
ECDMRIPFEIMEHNLSSTTRTVDOHW?HITR
YINDOTHLENTSISRYAGINVIDFENCHAUS
HITINCOSNGEVISTTOGANDANUTERANSR
EATPOINASMDUKRAJETHPAADURYPTCOM
19failed at cribs1/24 known lettersloss 97.1%
ECDMRIPFEIMEHNLSSTTRTVDOHW?HLRH
EGARCKATHETURINWANCOULDCANDSMAT
EASIFALSUCOMINSOFPEDINNIGDANDBT
LETANSEHIINOMOUNCODOMEMOUSCOCRI
20failed at cribs1/24 known lettersloss 97.1%
ECDMRIPFEIMEHNLSSTTRTVDOHW?GHIN
ETASHEMIQIVERJANWONARSEIGNOMEMM
ORREWTENIGUALLYOTHINGLOCRASTRAN
ERNANDINTOLITTOTEFODOTOUNUSESHW
21failed at cribs1/24 known lettersloss 97.1%
ECDMRIPFEIMEHNLSSTTRTVDOHW?ARDC
USONRENBUENTSBEYTODERTROBUSLETR
ONTISATIONINUSEARITUTNOTHENDEDO
FCUNDEIBNCCOMOOURYANGTOPPERSGOB
22failed at cribs1/24 known lettersloss 97.1%
ECDMRIPFEIMEHNLSSTTRTVDOHW?RWES
LTVIESAREURNORYINWROPLEANDCRUAR
ENEXTALYIBETANDSPECONGOTURGYBAE
MESAGERATTHETSSTIMONNSOFNOBEITH
23failed at cribs1/24 known lettersloss 97.2%
ECDMRIPFEIMEHNLSSTTRTVDOHW?BEAS
EAMICAPIKEFICOURNTONCEBEDADGELL
EANIOREROULDYOFMBAILEFENORKERTE
LLASATHBACKTEMASHEFIRELYWEILZEL
24failed at cribs1/24 known lettersloss 97.2%
ECDMRIPFEIMEHNLSSTTRTVDOHW?UEEA
XRIDRESMITONHEREIRREPREYEANUSEY
REANTEDGIANINMSTODINISSUMANTINS
TOTACEITIEFANANDHUBEKRATTODUPCH
25failed at cribs0/24 known lettersloss 101.2%
ECDMRIPFEIMEHNLSSTTRTVDOHW?DHOU
NCKSCKNETOKNCMTHROLDERYVKINEMAW
HANYATSSOTHEBAYEDREMLDSONGGALSG
ENETEMIERIGAHMEDIRPFORAYWITHIKC
26failed at cribs0/24 known lettersloss 101.2%
ECDMRIPFEIMEHNLSSTTRTVDOHW?QYNJ
UDGENKAYBAYBODEDWHONHEFERWKCIZY
AWOFUSEWIESTILLUSKEAGOLIONHESFO
RATITREHINDSTHANSRICNOCIPPESSYJ
27failed at cribs0/24 known lettersloss 101.2%
ECDMRIPFEIMEHNLSSTTRTVDOHW?UHES
NTINDBYOLNGSSTHISREDALLVESOGIOS
TSOVLPLORRISHCAFFADTRSAWBURNORG
ITIPSAJEWEMISTERNATHRBLYAURIPFY
28failed at cribs0/24 known lettersloss 101.3%
ECDMRIPFEIMEHNLSSTTRTVDOHW?RKOE
ASTFURIONTRMAYPMEOTHERSFORMOLOV
ERPPTORAGROWLEXLEMOREETSAFALIST
ESFUNEVOZINLOSTALEESONEDAYMUDJU
F13.4c 6-19: K4 statistic 1.47 at 11/beaufort, p 0.842 against 100 shuffles; cells with K4 above all 0 of 28; plant power L6 3/8, L7 5/8, L8 4/8, L10 4/8, L13 5/8, L16 4/8, L19 2/8
What was searched, as the harness recorded it.
item
F13.4c lengths 6-19 (D3)
search reads known letters
no
forms
beaufort, added
search
0236's climb, unchanged (imported): 0140's parallel tempering with no known letter, 3 runs x 16 replicas x 625 rounds x 1000 moves; MIX8m rescoring
k4 copies
100
cells
28
plant power
6
3, 8
7
5, 8
8
4, 8
10
4, 8
13
5, 8
16
4, 8
19
2, 8
plant letters right
6
9, 7, 11, 12, 26, 34, 2, 17
7
9, 7, 7, 2, 6, 2, 24, 10
8
18, 4, 25, 9, 11, 13, 3, 21
10
12, 8, 22, 23, 13, 6, 10, 10
13
7, 6, 16, 10, 13, 2, 12, 4
16
10, 6, 11, 6, 5, 2, 5, 10
19
8, 12, 10, 7, 7, 1, 5, 5
k4 z by cell
6/beaufort
-1.573
6/added
-0.952
7/beaufort
-0.009
7/added
-0.915
8/beaufort
-2.01
8/added
-0.66
9/beaufort
0.697
9/added
-2.769
10/beaufort
-0.565
10/added
0.913
11/beaufort
1.469
11/added
-1.215
12/beaufort
-1.398
12/added
0.615
13/beaufort
-1.176
13/added
0.171
14/beaufort
1.447
14/added
1.377
15/beaufort
0.39
15/added
-0.724
16/beaufort
0.104
16/added
0.104
17/beaufort
-0.413
17/added
-1.743
18/beaufort
-2.717
18/added
0.104
19/beaufort
0.383
19/added
-0.493
k4 copies at or above by cell
6/beaufort
93
6/added
81
7/beaufort
49
7/added
81
8/beaufort
96
8/added
73
9/beaufort
24
9/added
100
10/beaufort
69
10/added
17
11/beaufort
11
11/added
90
12/beaufort
93
12/added
28
13/beaufort
89
13/added
42
14/beaufort
9
14/added
10
15/beaufort
33
15/added
74
16/beaufort
48
16/added
43
17/beaufort
66
17/added
94
18/beaufort
99
18/added
42
19/beaufort
39
19/added
67
k4 cells above all copies
0
k4 statistic
1.469
k4 best cell
11/beaufort
k4 p
0.842
k4 known letters in place by cell
6/beaufort
0
6/added
2
7/beaufort
1
7/added
1
8/beaufort
0
8/added
0
9/beaufort
2
9/added
2
10/beaufort
3
10/added
3
11/beaufort
3
11/added
3
12/beaufort
3
12/added
1
13/beaufort
3
13/added
1
14/beaufort
0
14/added
2
15/beaufort
1
15/added
2
16/beaufort
1
16/added
1
17/beaufort
3
17/added
3
18/beaufort
3
18/added
1
19/beaufort
1
19/added
1
climbs
3,220
exhaustive
no
compute seconds
8,485
modal cost estimate
3220 climbs x about 10 s (0236) = about 9 core-hours, about $0.6 of Modal credit
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.