340 wildcard experiment

[Jarlve]

In the meanwhile I added something to AZdecrypt that detects wildcards. It seems to work like a charm. Here is a solve for smokie's cipher in under 3 minutes with the new intelligence.

That's very impressive! Can you elaborate on how you perform the detection?

It's not much really. I use my cycles measurement and rank each individual symbol. And then just select those that do not score well but also have a random roll that determines the minimum frequency symbols must have. It could be improved upon.

[Jarlve]

First iteration smokie33 vs 340. As you can see the 340 converged on that "cancer" plaintext again, it seems to be come dominant after expanding the "+" symbol. I don't know what to think of it. Yesterday I also ran the smokie33 with the more intelligent option but it didn't break. Either it's picking up too many false positives or the smokie33 is just crazy hard.

Code: Select all

340:

AZdecrypt 0.993 (Reddit corpus 6-grams)

Manipulation operation(s): LengthSymbol(2,2,2,1,[K,+])

Score: 18553.55 Ioc: 0.07160333 Entropy: 3.976389 Chi-square: 39.80777 Characters: 340 Letters: 21

ceepttalbutisnorn
ethestopingleisli
edmaybeconsistica
ntthatwilltothear
rayhimontorbutmus
tweartoningurpris
tocompareadollist
soralotofregionsa
neycancerreadinmi
nitelyaanotheredn
evesometigetanymo
reateationmylater
inandthetoreasawe
torealormalalmman
dediimmageplecons
talliessotallymor
eandsingeartnetis
thinginaboncnssit
trunnehmergoeinga
pyicetunicowithaf

Multiplicity: 0.2705882 Characters: 340 Symbols: 92 Bigrams: 18/1984 Sequential: 5653/2874

HER>pl^VPk|1LTG2d
Np?B(#O%DWY.<*!f)
By:cMIUZGW()L#zHJ
Spp7^l8*V3pOKQR"2
_9M[ztjd|5FP\&4k/
p8R^FlO-*dCkF>2D(
#5]$q%;2UcXGV.zL|
(G2Jfj#O`_NYza@L9
d<MebgZR2FBcyA64'
-zlUVh^JiOp7<FBy-
UmR/5tE|DYBpbTM,O
2<clRJ|*5T4M.n&BF
z69Sy#oN|5FBc(;8R
lGFN^f524b.cV4trs
yBX1*:49CE>VUZ5-u
|c.3zB0(Op^.fMqG2
RcTvL16C<wFlWB|)L
x{)WCzWcPOSHT/()p
|FkdW<7tB_YOB*-Cc
>MDHNpkSzZO8A|=;+


Code: Select all

smokie33:

AZdecrypt 0.993 (Reddit corpus 6-grams)

Manipulation operation(s): LengthSymbol(2,2,2,1,[14,23])

Score: 18513.02 Ioc: 0.06437619 Entropy: 4.093922 Chi-square: 23.30622 Characters: 340 Letters: 21

awyfnnihearsofdso
meteninatipgraded
mostsrealoriendsu
bswhenmathewinttr
ianditwouldmutedi
scryoffogoforcarr
estthesitsoiveals
onalmesstfloorabi
lindormsomebutiha
dtenyourslotingto
findmosthaveallyw
hereredeseasobean
impleareneditandb
eeffellbothentono
tdismuchaneenwasc
hilleitspastedani
mahugeammadovotea
beautyourisputbef
ootaguytentlyhasg
unstudentildrosoe

Multiplicity: 0.2941177 Characters: 340 Symbols: 100 Bigrams: 13/1553 Sequential: 4484/3295

21 46 1 35 10 33 2 50 11 63 30 26 6 27 39 57 51
90 40 34 15 33 4 10 16 54 4 7 41 52 19 47 61 22
48 42 58 54 13 5 17 18 20 59 28 55 24 8 49 57 53
25 64 65 66 15 10 43 16 44 45 17 46 55 67 54 54 52
55 18 91 47 62 56 92 6 38 9 39 48 38 54 40 49 55
26 31 30 1 51 27 35 6 41 68 35 42 28 31 19 69 30
11 26 34 34 70 12 26 2 56 71 6 62 60 24 16 36 13
3 33 93 20 48 61 58 37 54 27 32 51 6 28 21 25 55
32 62 8 47 6 52 43 57 72 48 61 73 38 44 4 45 16
49 56 40 33 74 6 53 94 26 32 6 54 2 8 41 54 59
27 55 95 22 48 42 57 54 75 21 76 15 19 9 77 1 46
50 78 30 61 52 24 47 17 13 12 18 13 6 25 79 80 10
2 48 7 96 15 16 52 17 8 24 39 97 44 18 10 49 81
12 40 35 27 40 36 9 25 42 44 45 17 10 56 42 33 3
54 39 4 37 82 53 98 83 16 10 11 17 10 46 18 13 31
84 2 20 32 15 55 54 58 7 16 57 54 17 47 16 85 99
48 18 50 38 41 100 16 43 43 18 49 51 60 6 54 24 21
86 17 18 38 56 1 6 87 52 2 57 7 88 56 25 17 35
3 6 54 16 89 38 1 56 17 29 54 32 1 45 18 37 41
53 33 26 56 23 47 14 10 54 2 9 39 52 3 57 6 15


[smokie treats]

Smokie33 is probably pretty hard. At some point I can convert one poly symbol to homophonic at a time.

[Jarlve]

Second iteration smokie33 vs 340. The 340 changed its "plaintext". No solve for both.

Code: Select all

340:

AZdecrypt 0.993 (Reddit corpus 6-grams)

Manipulation operation(s): LengthSymbol(2,2,2,1,[O,z])

Score: 18863.29 Ioc: 0.06964254 Entropy: 3.955365 Chi-square: 48.76514 Characters: 340 Letters: 20

auimerfetitsayhes
leantisoneouswith
noadownthethairas
seeifrewereallike
syouforsthatwayis
eeifarrowsoiament
ihadtohendtheusat
thestrickslowplay
ssolditieandonlyy
ourneifsomeisanoo
neishoutnonedyour
esdristwhyyoucana
llysoillthandthei
rhalftheydudeyoud
ontswayyoumenthou
tduringthefutothe
idywaslosparentha
ttheonedtosaysthe
taisesionsownwood
monaleisntmentchr

Multiplicity: 0.3205882 Characters: 340 Symbols: 109 Bigrams: 16/1843 Sequential: 5791/3704

HER>pl^VPk|1LTG2d
Np!B(#}%DWY.<*sf)
By:cM"UZGW()L#†HJ
Spp7^l8*V3p~$'Ru2
_9M,‡tjd|5FP0&4k/
p8R^Fl-*dCkF>2D(
#5=vq%;2UcXGV.ˆL|
(G2Jfj#€?_NY‰I@L9
d<MQb[ZR2FBcyA64w
-ŠlUV\^J]p7<FBy-
U`R/5tE|DYBpbTMx‚
2<clRJ|*5T4M.a&BF
‹69Sy#eN|5FBc(;8R
lGFN^f524b.cV4tgh
yBX1*:49CE>VUZ5-i
|c.3ŒB{(ƒp^.fMqG2
RcTmL16C<nFlWB|)L
or)WCWcP„SHT/()p
|FkdW<7tB_Y…B*-Cc
>MDHNpkSzZO8A|K;+


Code: Select all

smokie33:

AZdecrypt 0.993 (Reddit corpus 6-grams)

Manipulation operation(s): LengthSymbol(2,2,2,1,[15,54])

Score: 18929.38 Ioc: 0.06578171 Entropy: 4.070371 Chi-square: 25.32544 Characters: 340 Letters: 21

utitsrawhileotdrs
helordsandestinas
moralliesimhertru
ckingspandithinkt
heynotgoandmaketh
eylisttosetomyill
hellybeatgoodearl
yresmarsitssomuch
sornotpromanandda
tteryouresofarswi
thasmorefuckingit
wolatenilbelockis
amethatiredonesto
betterncondistory
addsmuchashistely
passwhereareinaga
mewashappetsdoneu
pieationtarentcit
yofawaitingsidess
uretingstandtyrog

Multiplicity: 0.3529412 Characters: 340 Symbols: 120 Bigrams: 9/1431 Sequential: 4949/4476

21 46 1 35 10 33 2 50 11 63 30 26 6 27 39 57 51
90 40 34 101 33 4 10 16 106 4 7 41 52 19 47 61 22
48 42 58 107 13 5 17 18 20 59 28 55 24 8 49 57 53
25 64 65 66 102 10 43 16 44 45 17 46 55 67 108 109 52
55 18 91 47 62 56 92 6 38 9 39 48 38 110 40 49 55
26 31 30 1 51 27 35 6 41 68 35 42 28 31 19 69 30
11 26 34 34 70 12 26 2 56 71 6 62 60 24 16 36 13
3 33 93 20 48 61 58 37 111 27 32 51 6 28 21 25 55
32 62 8 47 6 52 43 57 72 48 61 73 38 44 4 45 16
49 56 40 33 74 6 53 94 26 32 6 112 2 8 41 113 59
27 55 95 22 48 42 57 114 75 21 76 103 19 9 77 1 46
50 78 30 61 52 24 47 17 13 12 18 13 6 25 79 80 10
2 48 7 96 104 16 52 17 8 24 39 97 44 18 10 49 81
12 40 35 27 40 36 9 25 42 44 45 17 10 56 42 33 3
115 39 4 37 82 53 98 83 16 10 11 17 10 46 18 13 31
84 2 20 32 105 55 116 58 7 16 57 117 17 47 16 85 99
48 18 50 38 41 100 16 43 43 18 49 51 60 6 118 24 21
86 17 18 38 56 1 6 87 52 2 57 7 88 56 25 17 35
3 6 119 16 89 38 1 56 17 29 120 32 1 45 18 37 41
53 33 26 56 23 47 14 10 54 2 9 39 52 3 57 6 15


[Mr lowe]

utitsr a while otdrshelo rdsandesti n as
moral lies imher trucking spand i think t
hey not go and make the y list to set omy ill
hell y beat good early resmars its so much
sor not pro man and datter your e so far s wi
th as more fucking it wolatenil be lock is
ame that i red one sto better ncon distory
adds much as his tely pass where are inage
me was happets done upieation tarentcit
yofa waiting sides sure ting stand tyrog

[smokie treats]

Knowing the answer, it is intriguing to watch this.

43 of the positions in the most recent smokie33 iteration have the correct plaintext. There aren't any complete words, but there is one seven letter word that is close, with the same number of syllables and a pair of low frequency letters.

[doranchak]

I'm trying this approach for smokie33:

Generate all possible selections of 3 symbols from smokie33's alphabet. It has 62 distinct symbols, so there are (62 choose 3) = 37,820 ways to pick 3 different symbols.

For each of the selections, generate the resulting expanded ciphertext (each occurrence of any of the 3 symbols is replaced by a brand new symbol). So now there are 37,8280 new cipher texts.

Run them through azdecrypt, in an order based on the number of times the 3 symbols appear in the original cipher text. That way, more impactful symbols are explored first.

It's gonna take several more days if I let it keep running, but here's the best solution found so far:

Code: Select all

smokie33 19734 14 23 54 sum 56

Score: 21493.75 Ioc: 0.06689224 Chi2: 25.63647 H: 4.049193 M: 0.3382353 C: 340 S: 115

yesernsatedinmapo
recongrougestbeal
lareathetalwhispu
tinformonthewwhat
weredtonisalivesw
ildsomensreallbad
ticchristandwhora
instlarrymoonlytw
odientmpplayingto
steninunionisisla
mwillapplyjobsuse
andathehareanther
slefoothihadnersf
reemerstanthrtani
magroupforthreeal
istoowereoptheoff
leaiscommesownwhy
theitsnotspectthe
inforistheposters
unitrepressatipno

21 46 1 35 10 33 2 50 11 63 30 26 6 27 39 57 51
91 40 34 15 33 4 10 16 103 4 7 41 52 19 47 61 22
48 42 58 104 13 5 17 18 20 59 28 55 24 8 49 57 53
25 64 65 66 15 10 43 16 44 45 17 46 55 67 105 106 52
55 18 92 47 62 56 93 6 38 9 39 48 38 107 40 49 55
26 31 30 1 51 27 35 6 41 68 35 42 28 31 19 69 30
11 26 34 34 70 12 26 2 56 71 6 62 60 24 16 36 13
3 33 94 20 48 61 58 37 108 27 32 51 6 28 21 25 55
32 62 8 47 6 52 43 57 72 48 61 73 38 44 4 45 16
49 56 40 33 74 6 53 95 26 32 6 109 2 8 41 110 59
27 55 96 22 48 42 57 111 75 21 76 15 19 9 77 1 46
50 78 30 61 52 24 47 17 13 12 18 13 6 25 79 80 10
2 48 7 97 15 16 52 17 8 24 39 98 44 18 10 49 81
12 40 35 27 40 36 9 25 42 44 45 17 10 56 42 33 3
112 39 4 37 82 53 99 83 16 10 11 17 10 46 18 13 31
84 2 20 32 15 55 113 58 7 16 57 114 17 47 16 85 100
48 18 50 38 41 101 16 43 43 18 49 51 60 6 115 24 21
86 17 18 38 56 1 6 87 52 2 57 7 88 56 25 17 35
3 6 116 16 89 38 1 56 17 29 117 32 1 45 18 37 41
53 33 26 56 102 47 90 10 118 2 9 39 52 3 57 6 15

My word finder sees these words: PLAYING, POSTERS, REPRESS, ANTHERS, CHRISTA, CHRIST, INFORM, LAYING, POSTER, ROUGES, ANTHER, PLAYIN, THREE, GROUP, MONTH, UNION, PRESS, LIVES, STAND, APPLY, CHRIS, ALIVE, FORTH, LARRY, ISLAM, ROUGE, RAINS, WILDS, OMENS, OSTEN, INFOR, NITRE, THETA, COMME, HYTHE, ONTHE, CONGR, LAYIN, SPECT, FORIS, BEALL, STANT, PUTIN, THEIN, RECON, WILLA, POSTE, SATED, ANIMA, OSTER, THEIT, WWHAT, TWERE, REPRE

The word finder sees these words in Jarlve's solution: AWAITING, TRUCKING, WAITING, FUCKING, RALLIES, RUCKING, BETTER, ALLIES, DEARLY, AWHILE, FUCKIN, ROMANA, WAITIN, MAKETH, MUCHAS, REDONE, DESTIN, PANDIT, WHERE, THINK, WHILE, EARLY, STORY, SIDES, STAND, TIRED, MORAL, LORDS, ROMAN, ARENT, KINGS, TRUCK, AWAIT, ANDES, REINA, GOODE, ANAND, YOURE, BETTE, SANDE, ATION, ETTER, LYRES, REDON, ALLIE, SOFAR, CONDI, ISTEL, MUCHA, TINGS, ATTER

[doranchak]

I've been wanting to write a "fuzzy word" match routine for a while so I just made one. It prints out all fuzzy word matches for a given plaintext, within a range of edit distances. "Edit distance" is the number of modifications needed to make the "fuzzy" word match the "exact" word.

For example: STREAMERS has an edit distance of 2 from SFREEMERS. Modification 1: Change F to T. Modification 2: Change E to A.

Here are sample fuzzy word matches from my plaintext for smokie33, max edit distance of 2:

Code: Select all

CHRISTIANS [CHRISTAN] [CHRISTAND]
INFORMANTS [INFORMONT] [INFORMONTH]
FORTYTHREE [FORTHREE]
CHRISTENED [CHRISTAND]
DEPRESSANT [EPRESSAT] [REPRESSAT]
ANTICHRIST [ADTICCHRIST]
CHRISTIAAN [CHRISTAN]
CHRISTIANA [CHRISTAN] [CHRISTAND]
CHRISTIANE [CHRISTAN] [CHRISTAND]
CHRISTIANI [CHRISTAN] [CHRISTAND]
CHRISTIAN [CHRISTA] [HRISTAN] [CHRISTAN] [CCHRISTAN] [CHRISTAND]
CHRISTMAS [CHRISTA] [CHRISTAN]
REPRESENT [REPRESSAT]
SINGAPORE [INMAPORE] [DINMAPORE]
UNIONISTS [UNIONIS] [UNIONISI] [NIONISIS] [UNIONISIS] [NUNIONISIS] [UNIONISISL]
CHRISTINE [CHRISTAN] [CHRISTAND]
REMEMBERS [REEMERS]
REPORTERS [EPOSTERS] [HEPOSTERS]
CHRISTINA [CHRISTA] [CHRISTAN] [CHRISTAND]
SHATTERED [HATWERED] [WHATWERED]
INFORMING [INFORMON] [INFORMONT]
EXPORTERS [EPOSTERS]
CESSATION [ESSATIPN] [RESSATIPN]
INFORMANT [INFORMON] [NFORMONT] [INFORMONT] [TINFORMONT] [INFORMONTH]
MENSTRUAL [MENSREAL]
REPRESSED [REPRESS] [REPRESSA] [REPRESSAT]
POLLSTERS [POSTERS]
WEATHERED [WHATWERED]
STREAMERS [SFREEMERS]
HERNIATED [ERNSATED] [SERNSATED]
REPRESSOR [REPRESS] [REPRESSA] [REPRESSAT]
CHATTERED [HATWERED] [WHATWERED]
DISPUTING [ISPUTIN] [HISPUTIN] [ISPUTINF] [HISPUTINF]
PLAYTHING [PLAYING]
GILCHRIST [ICCHRIST] [TICCHRIST]
FOURMONTH [FORMONT] [ORMONTH] [FORMONTH] [NFORMONTH] [FORMONTHE]
CONGRUOUS [CONGROU] [CONGROUG]
REPROCESS [REPRESS]
UNIONISED [UNIONIS] [UNIONISI] [UNIONISIS]
CHRISTMAN [CHRISTA] [HRISTAN] [CHRISTAN] [CCHRISTAN] [CHRISTAND]
REPLAYING [PLAYING] [PPLAYING] [MPPLAYING]
REEMERGES [REEMERS]
ROQUEFORT [ROUPFORT]
GROUPWORK [GROUPFOR] [GROUPFORT]
THREEGOAL [THREEAL]
IMPOSTERS [POSTERS] [EPOSTERS] [HEPOSTERS]
REPRESSES [REPRESS] [REPRESSA] [REPRESSAT]
SCREAMERS [SFREEMERS]
LAYINGOUT [LAYINGT] [LAYINGTOST]
HARDSTAND [HRISTAND]
REGALISTS [REEALIST] [REEALISTO]
WRISTBAND [RISTAND] [HRISTAND]
INUNCTION [INUNION]
INTOURIST [INFORIST]
FOSTERSON [OSTERSUN] [POSTERSUN]
EBERSTADT [EMERSTANT]


Full list is here: http://zodiackillerciphers.com/fuzzy-wo ... anchak.txt

Here are sample fuzzy word matches from Jarlve's plaintext for smokie33, max edit distance of 2:

Code: Select all

EXPIRATION [EUPIEATION]
MORALITIES [MORALLIES]
CONSISTORY [CONDISTOR] [ONDISTORY] [CONDISTORY] [NCONDISTORY] [CONDISTORYA]
SUPINATION [UPIEATION] [EUPIEATION]
CORALLITES [ORALLIES] [MORALLIES]
PAINTINGS [AITINGS] [WAITINGS]
CONDUCTOR [CONDISTOR]
GREETINGS [RETINGS] [URETINGS]
TIREDNESS [TIREDONES] [TIREDONEST]
SURVEYING [SURETING]
DISTORTED [DISTORYAD]
ROMANIANS [ROMANAN] [ROMANAND]
OSTEOTOMY [OSETOMY]
PREDATION [PIEATION]
ITERATION [IEATION]
TRICKLING [TRUCKING]
REFUNDING [REFUCKING]
SECRETING [SURETING] [SSURETING]
DESTINIES [DESTINAS]
ONEPARENT [ONTARENT]
DUCKLINGS [UCKINGS] [RUCKINGS]
CORALLINE [ORALLIE] [MORALLIE]
PRIVATION [PIEATION]
WHEREFORE [WHEREARE]
ROUMANIAN [ROMANAN]
URINATION [UPIEATION]
KINGSLAND [KINGSPAN] [INGSPAND] [INGSTAND] [KINGSPAND] [TINGSTAND] [CKINGSPAND] [KINGSPANDI]
LINEATION [IEATION] [PIEATION]
EXPIATION [EUPIEATION]
BETTERING [BETTERN] [BETTERNC]
UNIMATION [UPIEATION]
REFUELING [REFUCKING]
ONESTROKE [ONESTOBE]
ANDESITIC [ANDESTI] [ANDESTIN]
RECORDIST [RNCONDIST]
SHELFOLDS [SHELORDS]
BETTERTON [BETTERN] [BETTERNCON]
TOUCHINGS [TRUCKINGS]
PLICATION [PIEATION]
RAMILLIES [RALLIES]
ONESTOREY [ONESTOBE] [ONESTOBET]
CONDICTIO [CONDISTO]
TIMEZONES [TIREDONES]


Full list is here: http://zodiackillerciphers.com/fuzzy-words-jarlve.txt

You can see from these results that the range of possible words is enormous when you consider all the ways to make small "fixes" to the plaintext.

[Jarlve]

Knowing the answer, it is intriguing to watch this.

43 of the positions in the most recent smokie33 iteration have the correct plaintext. There aren't any complete words, but there is one seven letter word that is close, with the same number of syllables and a pair of low frequency letters.

It's cool to have this kind of feedback! If this test fails to deliver the correct plaintext I'll rerun it with increased settings.

[Jarlve]

I've been wanting to write a "fuzzy word" match routine for a while so I just made one. It prints out all fuzzy word matches for a given plaintext, within a range of edit distances. "Edit distance" is the number of modifications needed to make the "fuzzy" word match the "exact" word.

I'm interested to see what it could turn up.