Chief Delphi - Cryptology

Chief Delphi (http://www.chiefdelphi.com/forums/index.php)

- Chit-Chat (http://www.chiefdelphi.com/forums/forumdisplay.php?f=14)

- - Cryptology (http://www.chiefdelphi.com/forums/showthread.php?t=35759)

My friend gave me a code to crack. He's offering 100$ to the first person to crack it. I don't expect him to pay, but I want to see the look on his face. It's a single letter substitution code, and I don't think he's smart enough to try anything too involved. Then again it may be a load of crap, but I doubt it. Any help directly and/or decryption programs for such a code would be appreciated.

Quote:

plowpn bqn fuul kgcuuwf fzqfgop zqw oqw owb zirl wn lgrwl buuconac nwg sr

Have at you.

Single letter as in one in each word or one letter solves all? :D

u is a vowel.

w is probably a vowel.

q is probably a vowel.

yeah thats really confusing haha

If the U in it means another vowel, it will be e. E is the most commonly used letter in the english language, and there are a lot of those Us. Also, e is often used in douldes like the Us in the encryption are.

Before we go about this, could you tell us how much, if any, experience your friend has with cryptology?
[deleted, formerly addressed the difference between public and private encryption keys (not relevant)] Because the string is short, most likely a few possibilities will work.

In short, what I'm saying is that, while I'll give it a crack (and I'll give it to some friends that are much more computer-fluent to let them have a crack at it), if your friend has a good idea of what their doing, this might take a while [obscure reference to a coding sequence no longer relevant] More info concerning the code might shorten the process.

An option to decrypt this is to use a dictionary-style attack at the cyphertext. This is to say that you find a list of all 8-letter words and then see which ones match the rules for "buuconac". those rules would be:
-the second and third letters are the same (only in this case becuase it's a letter-letter substitution code)
-the fourth and last letters are the same (again, only in this case)
-no other letters are the same
-cyphertext "a" occurs once in the string, so it is likely to not be a common letter (listed in order of most frequent to least frequent, in the english language, the letter order is "E T A O N I S R H D L U F C M W P G Y B V K X J Q Z", source: Enchanted Mind, Cryptography )

Attached is my results (so far) after searching (visually and briefly, so there might be minor errors) through the 2000 most common words in the english language for some of the longer cyphertext strings. Hope this helps! (I'll post my results later after I make some real progress)

This is turning out better than I expected. So far as I know, it started as a joke and this friend of mine said, "gee, I think I'll make one of those too!" In otherwords, I expect no tricks of any sort. Most likely, he just substitued (consistantly) one letter for another. A dictionary attack is what I have in mind but I didn't have any software to do so. I also believe he only used standard english, meaning no strange spellings etc.

Thanks to all who are trying! Especially mgreenley.

I was looking at the information in the thread to date when it hit me that there are a few more clues to decoding this that I'd like to point out to everyone else to save the rest of you some time.
-any repeated characters can allow you to quickly decode the message after you have one letter figured out, but consecutive repeats are more valuble in the initial process
-bqn,zqw,oqw, owb and nwg are all different 3-letter words. Once I finish up some more work, I'll check them out, but remember that when you're essentially brute-force decrypting anything, playing hunches is fine. If you have a guess for one letter, re-write the line below the original and work from there for no less than 10 mins, but no more than 1 hour unless you have a reason to. That bwn ends in an "n" (see a later point for "n")
-These all being different is signifigant because you can compare the cypher-text to clear-text to get a list of possibilities.
-"zirl" has at least one vowel in it. An assuption, but a pretty safe one at that.
-"sr" also should have a vowel in it. Again, a pretty safe assumption.
-"wn and sr" has a limited number or sensible combinations, "is, an, no, be, if, of" is a good list to start from when guessing. If this is a complete sentance, sr has an even more limited number of posibilities. Odds are w is indeed a vowel, but unless you're certain, always qualify your statements.
Important -I messed up in my original .doc (hehe :o sorry), the information in it is correct, but the original string has one error and two omissions (that's what happens when you're working on two AP projects and a health project while you're typing.
-if we assume u, w and q to be vowels as stated by others in this thread, "kgcuuwf" has a three-vowel string, which narrows the search
-the cypher-text "n" is most likely a consonant

And those projects are calling me again, so I've got to get back to work. Good luck with this in the meantime!

P.S. google has a few word-lists if you dredge through the muck (Some useful URL's are below), but I couldn't find a program either, you just have to sit down and dedicate and hour or so to just check one sequence for possibilities and once you get two or three sequences, you can compare to narrow them down from those lists. (i.e. I put all those words in my .doc list in the possible first, then on the second pass when it was just them, eliminated a few.)

URLs:
http://dictionary.new-frontier.info/w/By_word_length_(longest)
http://www1.harenet.ne.jp/~waring/vo...sts/1-2000.txt
http://www.enchantedmind.com/html/cr...ptography.html
http://www.multicentric.com/wapi/mct...NTRIX&ObjID=84 (program, didn't feel like figuring out)
http://www.sedl.org/reading/topics/exception.html (use the syllable list)

Can you at least get a confirmation that it is:

A) Most importantly, Appropriate for these forums if someone does solve it
B)Not as important but would be nice to know if it was an actual statement, or sentence, or phrase and not just a bunch of random words.

I do not think this is a direct-substituion cipher. Reason: There are many 3-letter words and they do not seem to fit in together. "q" appears as the second letter in 3 of those words. I cannot think of any 3 (distinct) three-letter words that share the same middle letter. Additionally, q appears only once in the rest of the cipher.

This leads me to believe that this code was generated using some other method. I think this falls under the Vigenere Cipher (Polyalphabetic cipher), where a keyword is used to shift the letters.

If u is a vowel it is probably not u, i, y, or a. You don't find many doubles of these in the English language. This leaves u to be o or e. Bemis or Ein are probably right.

Either s or r is probably a vowel
If s a vowel r is probably not
If r is a vowel s is probably not

Either w or n is a probably a vowel
If w is a vowel n is probably not
If n is a vowel w is probably not.

Quote:

Posted by DarkJedi613
u is a vowel.
w is probably a vowel.
q is probably a vowel

I am going to say I think it more likely that w is not a vowel. Your are unlikely to get a double vowel then another vowel in a word like in "kgcuuwf." This makes it very likely that n is a vowel.

That part above is almost logic now we are in the area of true guesswork.

P is likely to not be a vowel. Most words ending in a vowel have do not have a second vowel in front of them. Fewer still of these start in the same letter. like in "plowpn."

Something in "zirl" is a vowel. Things would be simpler if that was the r (also seen in "sr") but that is no guarantee.

Large words are divided into syllables. The basic rules for that are divide between two non-vowels or if there is no place where you can do that after a vowel and before a non-vowel. There are not many multi-syllable words with syllables more than 5 letters long. There are also not many single syllable words longer than 7 letters. This makes it very likely that "fzqfgop" is a two-syllable word. If p is not a vowel and q might be a vowel then it is fairly likely that f is not a vowel.

If w is not a vowel then you can say a lot about "oqw" and "owb." Either o is a vowel or o and/or b are vowels. When you look at "plowpn" "buuconac", "bqn" it is more likely that o is a vowel than b. B would force a triple vowel in "buuconac" but o looks like it could work in syllables. Q looks like a vowel from other words.

We should have 6 things that fit vowel rules: a e i o u and then y is the "could be" vowel that only matters when we find it at the end of the word.

u is a vowel
s or r is probably a vowel
n is probably a vowel
q is probably a vowel
o is probably a vowel

r and n could be y but they also appear in the middle of words so that isn't likely. Those above are probably the five vowels. Also one of them appears in each word if you use r instead of s. This makes the final vowels u, r, n, q, and o.

Quote:

Originally Posted by elrabin

That is what I originally believed it to be as well, but the original post does say that it is a single letter-letter substitution. If it hadn't said that, I would've thought that the coder had just used one of the online enigma machine-like applets to write the sequence.

Are you sure he didn't pound his head on the keyboard and just send you on a wild goose chase?

I did some playing around with this and made something to pattern-match against /usr/share/dict/words, and there are very very few words that some of those scrambled words can make and the words they do make don't fit very well into the rest of it. Is there a possibility that the words are reversed? And ranking by popularity in the crypt, the order of letters is: W(13%) U(10%) N(8%) O L(6%) F G Q B(5%) C P R Z A(1%) I K S. There aren't enough letters to get that exact so many of the percentages are the same so it's not quite exact. If you guys need a boost, here are the possibilities I found for 'buuconac'.

base word: 01123452
match: aggerose (01123452)
match: alleyite (01123452)
match: ammelide (01123452)
match: ammeline (01123452)
match: annexive (01123452)
match: annexure (01123452)
match: appetize (01123452)
match: apprizer (01123452)
match: approver (01123452)
match: Arretine (01123452)
match: assemble (01123452)
match: bookmark (01123452)
match: bookrack (01123452)
match: coonskin (01123452)
match: deerhair (01123452)
match: Errantia (01123452)
match: Essenize (01123452)
match: foothalt (01123452)
match: footrest (01123452)
match: irrelate (01123452)
match: keelhaul (01123452)
match: Obbenite (01123452)
match: occlusal (01123452)
match: ollenite (01123452)
match: Ossetine (01123452)
match: rootfast (01123452)
match: rootwalt (01123452)
match: seedbird (01123452)
match: Ubbenite (01123452)
match: unnestle (01123452)
match: woodbind (01123452)
match: woodgeld (01123452)
match: woodland (01123452)
match: woodshed (01123452)
match: woodyard (01123452)

It seems like there's a good chance U = O according to all those matches. The most common words seem to be using it.