[Aaron]

The Kryptos Sculpture was made in 1990 by the CIA as a challenge to all code crackers, and it is definitely solvable though the methods to do so are unclear.

Go here for the full scoop: http://www.elonka.com/kryptos/

[bluewolf]

Pretty interesting one. Wonder if it'll net you a job if you solve it (whether or not you want one LOL)?

Has anyone else played around with this yet? I think the first thing to do would be to break it into sections - since several sections have been deciphered already - to see what's left.

Apparently since part of the panel is a take-off on the Vigenere cipher, one of the sections must obviously use that to decode it. 3 out of 4 sections were broken using pen and paper. This means 3 of the sections are probably able to be reduced to frequency analysis or something not so advanced. Remember, this is a sculpture to commemorate the *history* of cryptography, so the 'old' methods would probably be sentimentally represented.

Anyone else interested in this one?

[Aaron]

                                  ?OBKR
        UOXOGHULBSOLIFBBWFLRVQQPRNGKSSO
        TWTQSJQSSEKZZWATJKLUDIAWINFBNYP
        VTTMZFPKWGDKZXTJCDIGKUHUAUEKCAR

Looks like it's finally almost completed! Nice to see that the page is constantly updated with news and stuff. Now to crack it...

[bluewolf]

That last section has been reputed to be the toughest to crack. Even the author says that it was meant to be the hardest one. He's a bit guarded about his words when he talks about it now. In the beginning, he was less careful about what he said in interviews. I guess at some point he realized the magnitude of the puzzle he created and didn't want to spoil it by being glib.

I've gone through all the information I could find about this on the web - including reading the interviews. The first two parts were pretty straightforward. There was a keyword which altered the table of the sculpture in a specific way. One of the best sites I've found comes from a group called the Realm of Twelve. The site includes some REALLY nice visualizations of the first two parts of the code:
http://realmoftwelve.fateback.com/speculation/k1solution.html  [which is an awesome animated demonstration]
http://realmoftwelve.fateback.com/speculation/k1solution.html

The third section also has an animated solution:
http://realmoftwelve.fateback.com/speculation/k3solution.html

Now, although it's been cracked, I think the third part was cracked in a way that wasn't expected by the author. I think the part that's missing is that Part 3 was cracked without the clue that should lead to Part 4. It would be like cracking 1 and 2 by frequency analysis instead of finding the palimpsest and abcissa keywords.

And we must remember that the author is also (and foremost) an artist. He didn't bury any treasure. He wasn't hiding any BigSecret that would be dangerous to reveal (a la DaVinci Code/Jesus Papers). The whole point of the code was that it complemented and became part of the work of art. Works of art are meant to be beautiful. All these pieces are supposed to come together beautifully. So I doubt that the keywords were meant to dangle in the breeze as a leftover byproduct. I also doubt that the intentional misspellings in the deciphered code were meant to be disregarded. And there appears to be some other clue in the 3rd part that has been overlooked due to the way that section was decoded. And the raised letters are admitted by the author to be significant. YA_R How are they to be combined in all this? Since the raised letters are found in the third part, perhaps they are a clue to the correct [intended] way to decipher Part 3?

I think an analysis of the 3rd part might be a wise move. We have the plaintext and cyphertext. How could they be put together differently? Is there another method that could be used to go from one to the other? Maybe that would reveal the key needed to crack Part 4?

Another note about Part 4... From the interviews, it seems that Part 4 was intentionally coded with a method that would counteract any attempt to analyze it using frequency analysis. There are not many methods that will do this...

[Aaron]

Perhaps it's related to keyboard placement or somesuch? A bunch of letters used to graphically represent a message would certainly foil any frequency analysis.

[bluewolf]

Actually, if you think about it, keyboard placement wouldn't eliminate the frequency. Even using the first two methods of encryption - it didn't eliminate frequency analysis. It really has to be something that makes the letters non-letters so that the frequency of the letters in the plaintext are disguised. That's why I think it might be along the lines of an XOR operation. It's mathematical, but can still be reproduced if you have the key... Also think of things like Raid 5 in a computer. If you have two stripes and a parity stripe, you can rebuild any failure of one drive.

Imagine the Ds as data and the Ps as the parity information.

DDP
PDD
DPD
DDP

If any of the Ps went away, you could regenerate them by calculating the parity. If any of the Ds are missing, you can regenerate them using the other D and the parity information. [Each column represents one hard disk] Therefore if any column disappeared, you could regenerate all the Ds and Ps by using the above, but you couldn't use just one column to figure out what information is in the other two. Now imagine that the left column is the plaintext and the right column is the cyphertext. The middle column is the key that we don't know or understand.

Perhaps the middle column represents the clues teased from the first three sections? And that is used to transform the plaintext into cyphertext using an XOR operation?

For those that don't know about XOR ....  XOR is a mathematical operation that takes two binary numbers as input and produces an output. If the two numbers are the same, the output is 0. If the two numbers are different, the output is 1.

0 - 0  >> 0
0 - 1 >> 1
1 - 1 >> 0

I think this would definitely mask the letter frequency and is a robust method of encryption. I'm sure that his tutoring in crypto must have covered this. Now.... one  of the things about this method is that you must first convert the message to binary. Looking at the pages on ASCII to binary, an XOR operation would have yielded incorrect results (unable to be displayed). So he might have developed his own binary system for the 26 letters. And if it was a 5-bit block (32 possibilities), it would have given him 6 extra representations to leave as dead/null - to further the degree of difficulty. Naturally, all zeros and all ones would probably not be used --- so there may be 4 nulls in the translation to binary.

[Aaron]

I was a little bit unclear with what I said... I meant that either one letter actually stands for more than one letter or a group of letters actually boils down to equal one letter. I was just remembering this one anime where a bunch of gibberish, when typed out on a keyboard, resulted in every 5 letters or so actually outlining a a bigger "letter" on the keyboard. Type out edcftgb ujm and you'll see what I mean.

Your idea has a lot of merit too, however.

[ACE Tech]

In the final string of characters, every letter of the alphabet is used at least once.  How often does THAT happen...?

The OBKR might be end-of-word delimiters or, even dummy digits inserted randomly, leaving 22 valid characters.  Notice the sequence ends in "R"...

[Aaron]

Nice catch! Perhaps they encoded "The quick brown fox jumps over the lazy dog" in as complicated a way as possible, who knows. XD Delimiters or dummy characters are also possible, yes.

[ACE Tech]

However, that WOULD happen in a base₂₂ alpha system,* if the sample was somewhat long. I am not sure my computer can correctly convert a 75(?)-digit number if that was the case.

The next attack I would suggest:

1. Remove the OBKR characters,
2. Convert remaining sequence from base₂₂ (where A=0, ..., Z=21) to decimal.  (You may spot intelligence at this point.)
3. If you don't, convert your decimal number to base₂₇ ("_"=0, A=1, ..., Z=26)
4. If base₂₇ doesn't work, I would convert to binary and check at 5-8 bit groupings, beginning your groups at the radix point going left, because you won't have leading zeroes in your Most Significant [bit cluster].

*Base shifting would also disguise letter frequency, too, because the final output letters would really not be letters, at all, but rather numbers that look like letters, like alphas in an alphanumeric serial code.  As such, they would be devoid of any linguistic value and appear quite random.

[escher7]

The Bluewolf links are dead.

[Aaron]

http://web.archive.org/web/20100731082718/http://realmoftwelve.kryptos.info/speculation/k1solution.html
http://web.archive.org/web/20100726071622/http://realmoftwelve.kryptos.info/speculation/k2solution.html
http://web.archive.org/web/20100725042239/http://realmoftwelve.kryptos.info/speculation/k3solution.html

I'm impressed, even the pictures are archived.

[tommietheturtle]

Kryptos means hidden in Greek, The Sculpture...(in my opinion) it is a piece of artwork and is designed to hide the element of human curiosity from the final solution. The solution to Part 4 of the sculpture is hidden from view...in the absence/illusion of light. My guess is that the morse codes found around the sculpture are the final clues when factored in with the other three solutions. Its like a riddle. My first guess tho was that this sculpture had multiple purposes as does all art. First there are many literary transpositions, most important of all is the first sentence. That is the plot, The morse codes are partially hidden and partially complete. "T is your position" alluding to "what? is you position". Second Mistakes were made and hints were given away, "berlin=NYPVTT" My last and final interest is to measure the the shadow cast by the statue based on a four point compass to SEE if the shadow falls onto a CLUE or object during the day and at night. The obvious answer to PART 4 has to do with BERLIN and all the other solutions references. The key is the cipher! Also the four quadrants (squares) of the sculpture could reference to a geometric cipher, I was always intrigued by this statue/sculpture and its reference to the word hidden! (Its a Geometric Cipher)(Something only an Artist could make) on a side note I myself use it as a (public key) HINT! Its was meant to be private. No public eyes! anyways thats were I'm at I haven't bothered finding which geometric cipher was used i'm guessing a private one. 

[tommietheturtle]

Its as E-ZZ as A-B-C.

[Aaron]

Welcome to the forums!

I wasn't sure what you meant by geometric cipher until I saw the Pigpen cipher in that other link. I believe that's the kind of cipher used in Dorabella as well, due to the clearly angled shapes. I wonder what maps to what...

Surely there's a map hidden somewhere.