Starting a branch off thread at:
solution to smokie27c (has it been solved previously?):
Signing the diver to go on, he followed him round the vessel’s stern.
The sand on the other side was high and one could climb on board, but
Lister shrank from the dark alleyway that led to the engine-room. For
all that, he went in and saw the diver had opened the jambed door. When
he reached the ledge a flash from the other’s electric lamp pierced the
gloom and he tried to forget his throbbing head and looked about.
Sparkling b
The actual decrypt from my program shows many errors, but enough so that the text may be found on the Internet:
SIANINGTHEDIVERTOAOINHEFALLOWEDHIMROUNDTHEVASSELSSTERNAHESANDINTHEATHERSADAWASHIGHANDONECOULDCLIMBINPAARDBUTLISTERSHRANKFROMAHEDARKALLESWASTHATLEDAOTHEENAINERIAMFORALLAHATHEWENTANANDSAWTHEDIVARHADOPENEDTHECAMPEDDIORWHENHAREACHEDAHELEDGEAFLASHFROMTHEOTHERSELECTRICLAMPPIERCEDTHEALIAMANDHEARIEDTOFORGETHISTHRIBBANGHEADANDLOOKEDABOUTSPARKLINGB
Hey george4096,
It is possible that we never tried. AZdecrypt does solve it, similar to your own to decrypt. Do you have your own program?
Score: 22494.53 Ioc: 0.07145 Multiplicity: 0.17647 N-grams: 397 PC-cycles: 3518 SIANINGTHEDISERTO AOINHEFALLOWEDHIM ROUNDTHESASSELSST ERNAHESANDINTHEAT HERSADAWASHIGHAND ONECOULDCLIMBINPA ARDBUTLISTERSHRAN KFROMAHEDORKALLEY WAYTHATLEDAOTHEEN AINERIAMFORALLAHO THEWENTANANDSAWTH EDISARHADOPENEDTH ECAMPEDDIORWHENHA REACHEDAHELEDGEAF LASHFROMTHEOTHERS ELECTRICLAMPPIERC EDTHEALIAMANDHEAR IEDTOFORGETHISTHR IBBANGHEADANDLOOK EDABOUTSPARKLINGB
Hi Jarve – thanks for your answer. Yes I wrote my own program. It is based on simulated annealing and log 6-grams.
I found the solution to a couple more, which appear in the list David published and have an empty ‘solution” column. Does that mean they are indeed unsolved?
One that you solve and I can’t is this one
baloony: ,CurN[|dI…
which seems to have plenty of intentional mistakes, so the 6-gram approach failed.
A couple of questions, if I may:
– What is the corpus of plaintext from which you created the 5-grams?
– What is your formula for computing 5-grams/scoring function? Are you scaling to log? Taking the Index of Coincidence into consideration? Which line(s) in your source code (thanks for sharing it!)?
Many thanks!
George Lasry
George! Welcome. So glad you are here.
For anyone who doesn’t know, George is a very accomplished codebreaker who has broken many unsolved classical cipher texts:
https://scholar.google.com/scholar?hl=e … asry&btnG=
I have been to several of his fascinating presentations at NSA’s Cryptologic History Symposium.
I found the solution to a couple more, which appear in the list David published and have an empty ‘solution” column. Does that mean they are indeed unsolved?
Sometimes it just means I failed to update the spreadsheet with solutions that are floating around. (i.e., personal negligence)
One that you solve and I can’t is this one
baloony: ,CurN[|dI…
which seems to have plenty of intentional mistakes, so the 6-gram approach failed.
I noticed the URL linked to that cipher is obsolete. Here is the correct URL, so you can see some context for that cipher:
https://www.tapatalk.com/groups/zodiack … 9-s30.html
Hey George,
Cool to have you here!
One that you solve and I can’t is this one
baloony: ,CurN[|dI..
Do you have a copy of that cipher? Can’t seem to find it.
– What is the corpus of plaintext from which you created the 5-grams?
The "5-grams_english_practicalcryptography_wortschatz" are from http://practicalcryptography.com/cryptanalysis/ and they used the Wortschatz corpus. The "5-grams_english_jarlve_reddit" are from Reddit. My n-gram file format: n-gram size, language, author, corpus.
Reddit corpus: https://files.pushshift.io/reddit/comments/
– What is your formula for computing 5-grams/scoring function? Are you scaling to log? Taking the Index of Coincidence into consideration? Which line(s) in your source code (thanks for sharing it!)?
Just like ZKDecrypto, the n-gram frequencies are converted to log like this: log(n-gram frequency)*10 or log(n-gram frequency)*100. The ioc is taken into consideration and all individual n-grams are simply summed. My formula is something like this: total_ngram_score/1+(ioc*iocweight). So that the total_ngram_score is divided by the ioc such that a higher ioc will result in a lower score for the cipher.
I noticed the URL linked to that cipher is obsolete. Here is the correct URL, so you can see some context for that cipher:
Ah yes. I have this cipher as tonyb2. I think it is a low-level nomenclator and it may take a while for the solver to find the solution.
,CurN[?dISMT^yhvo
sdWSPFYwptc+gbfdB
zOR`yA+0ZIG-VYa6+
_C-UroLZD60Zfu4Jb
sdq4{pikGFBJtTc:z
S+:OS]mNjxlIJJWgu
9vP[C,4Yy7UXr`hJs
FclXfN@wtVRKTa8-N
B_oDxZrZsUdliqXj4
dIWn@JSdl9p,XT`kT
U-UlFa[-zX@JhkYq+
b@26gGRdfOyDpcA+V
i20ZBu,dE_oZbT^vj
IGtT`C2ULZn*YW_Pz
ufSU9rUo[sbZFTUWh
w6BGRI_taD-qVrNis
p+nmoWxFVjVtu9T`O
E_,1n^oW[gYJrTy{C
sHhaL]?cR_o-V+ZlD
ibTUdXjq-SgFZfzZl
Score: 22246.91 Ioc: 0.06196 Multiplicity: 0.20882
N-grams: 367 PC-cycles: 4073
TDROVEYWASQUITESL
OWLSSOASNOTEDRAWA
TTENTHEYCARAMANTE
LDAPOLICITYCARSBR
OWNSINEGROABOUTIT
SEITSFIVESHABBLDR
ESSEDTSATYPHONEBO
OTHHAVESOMEFUNDAV
ALLISCOCOPWHENHES
WALKEBSWHENTHUNGU
PAPHONEATHEBEGANE
RESTDREWATTINTHEM
ESYCARTWILLCRUISE
AROUNDSPICKWALLST
RASPEOPLEORCOUPLE
STAREALONIANMOVEO
NEKILLSOMEMOREUNT
ILTOKILLEDABOUTID
OPENIFYTELLAMECHI
ERUPWHENASDOCATCH
Hi jarve many thx for the details.
Unrelated: in June there is an event on historical crypto – HistoCrypt 2019 which this year will take place in Mons Belgium. Any chance you might be able to attend? Will be happy to meet in person. Some other folks from our team have our expressed interest in your work.
Thanks again,
George
Unrelated: in June there is an event on historical crypto – HistoCrypt 2019 which this year will take place in Mons Belgium. Any chance you might be able to attend? Will be happy to meet in person. Some other folks from our team have our expressed interest in your work.
It’s not something I want to do. Thanks for your offer however, it is appreciated.
623+86kL4mEX6PBKR Tz1cwzF7cn3I91xSQ MbZgncUJd4NEYG+xa YAfW9xbCYXegjIglr SJUrVI6SUkib9yEgc DL+TC+QHGWczOK+sv V6jltEqTZDBi4lHYv zFx9sM+8Wc4M6uBUd KWfHe3lhqB5N9ZOmS tgzvElTzVW+1X+MEA 6NP+yC9nNwMfhDkAc J9td95RVrB4z6MrcU MhOPxTSrI4fL9oaq8 Bg2zVGr3L95WEQ+Pp aZ+vcCSKtP4LzwSqy Au3U9hQtWAN+vwFfG dSCnEJ+5YiN+wPwhX NvKAj9TGQV9OGgudi rB+vbLcxnsjWktdz8 asRS+PEb30PZKGlvA
if this is easy to solve, i’d be interested in what statistic gives it away. i’m trying to think of things a lazy amateur would do to make a cipher more difficult.
It looks like you may have stacked a rearrangement of columns in a 17 by 20 grid on top of a rearrangement of rows in a 68 by 5 grid or vice versa. Though that does not jive with your saying off:
i’m trying to think of things a lazy amateur would do to make a cipher more difficult.
it’s a homophonic substitution with an additional "rule" that i think makes it look more complicated than it is.
it’s a homophonic substitution with an additional "rule" that i think makes it look more complicated than it is.
Okay. I can’t tell what the extra rule is. There seems to be more than 50% cycle randomization so I wonder if your extra rule caused that or if you were just very loose with the cycling. Feel free to tell me what you did and well done.
the "+", "9" and "z" symbols are special. find the mathematical difference (subtraction) in the two symbols after a special symbol to equal the decoded letter –
e.g. "z5y"
"z"=Special Character, "5"decodes to W (23rd letter), "y" decodes to V (22nd letter) = 23-22 = 1 = "A", so "z5y" = "A" even though 5 by itself is "W" and y by itself is "V".
it’s dumb, but it’s easy for an amateur to do and i think that kind of "trick" would fit zodiac’s definition of "fair play". i think it would throw off cycling and bigrams too?
i don’t think he literally did the above, but when i look at the 340 and see two "+" symbols on row 4 next to each other, there’s no way he’d give an obvious clue like that. the "+" has to mean something else, so my guess is he created some internal rule that throws off the solvers.
623+86kL4mEX6PBKR THE BESTWAYTOENC Tz1cwzF7cn3I91xSQ R YP TAMES SAG MbZgncUJd4NEYG+xa EBUTMAKKITHARD E
thanks for your time with this. i have a lot of respect for the lengths you go to solve the 340.
Thanks for sharing what you did masootz. Nifty and original encoding scheme, pretty hard too.
i think it would throw off cycling and bigrams too?
It did. It really looked like what I described earlier. And bigrams are funny too. Something to ponder about for sure.
I have added it to my to do list to make a solver that can handle it. A verbose element basically.
