Crack this Code!

[~N9NE~]

A fun little exercise I came across another forum. It was posted on the FBI website, it seems they've been doing it for awhile. I was the first to break it on the other forum

Good luck!

If someone breaks it, maybe somone could post another code we can try to break or something? Or I'll post another FBI one.

 

[AltF4]

Looks like a plain old monoalphabetic substitution cipher. Trivially broken. Let me see if I can find any of my old cipher crackers around...

 

[~N9NE~]

pen and paper >>>

If anyone does get the solution post it in spoiler tags so people can still give it a go the traditional way.

 

[AltF4]

Here is the ciphertext in non-******** pictures. It took me quite a while to write all the symbols down and map them to an english symbol, though. It will probably help you if you want to try cracking this on paper.

ABCD CD EFG BHGIJDA KFCL DE MHGN CA CD VEOPJGQ SCAB TEGJ ABHO ASCRJ HD THOU SEGID HD EFG VHDA RGUWAEPGHWBCR RBHVVJOPJN CA HVDE MJHAFGJD H DJS DJA EM WCRAEPGHT DUTXEVD XHDJI EO OHAEYJ HTJGCRHO DUTXEVD HOI TEACMD MEG JHRB RBHGHRAJG EM ABJ HVWBHXJAN ABHOZD MEG WHGACRCWHACOPQ HOI ZFIED EO RGHRZCOP ABJ REIJN

Solution:

Spoiler

THIS IS OUR HARDEST QUIZ SO FAR. IT IS LONGER, WITH MORE THAN TWICE AS MANY WORDS AS OUR LAST CRYPTOGRAPHIC CHALLENGE. IT ALSO FEATURES A NEW SET OF PICTOGRAM SYMBOLS BASED ON NATIVE AMERICAN SYMBOLS AND MOTIFS FOR EACH CHARACTER OF THE ALPHABET. THANKS FOR PARTICIPATING, AND KUDOS ON CRACKING THE CODE.

EDIT:

FYI: This is why the NSA handles cryptography and not the FBI.

 

[GoldShadow]

Solution:

Spoiler

This is our hardest quiz so far. It is longer, of course, with more than twice as many words as our last cryptographic challenge. It also features a new set of pictogram symbols based on Native American symbols and motifs for each character of the alphabet. Thanks for participating, and kudos on cracking the code.

Aw, Alt already got it with his fancy computer stuff!

But yeah, it wasn't too hard. Instead of transcribing symbols onto paper and then assigning them letters, I just used MS Paint to copy symbols and make a little key/legend, and replaced symbols on the actual message with copied/pasted letters.

I figured out some common/distinct vowels (A, O, I) based on word structure (where the symbols appeared in words) and word length. From there, the rest was easy.

 

[AltF4]

For those who care: This is called a substitution cipher. It has three parts:

The Ciphertext: This is the scrambled message. It's what you start out with.

The Plaintext: This is the solution. It is the message in plain english.

The Key: The key is the mapping between ciphertext to plaintext.


This is a monoalphabetic substitution cipher, which is one of the easiest kinds. It has one single mapping as the key. So, for instance, we the letter 'F' in the ciphertext might map to 'L' in the plaintext. This mapping will remain constant for the entire message.


Substitution ciphers are cracked by using statistical analysis. English is highly non-random, as anyone who has watched Wheel of Fortune will know. The frequencies for the english letters are as follows:

So you tally up all of the ciphertext letters and look at their frequency distribution and try to match it up to the corresponding English distribution. (IE: The most frequent letter will be E, then T, etc...) More specifically, you want to minimize the total difference between distributions. Which is something easily done with a computer program.

The statistical attack gets more potent the longer the ciphertext is. This puzzle seems to imply that making it longer makes it harder, but it's actually the other way around.

 

[Pikaville]

Nice job AltF4!

The government will be calling you for your services any day now.

 

[Crimson King]

For those who care: This is called a substitution cipher. It has three parts:

The Ciphertext: This is the scrambled message. It's what you start out with.

The Plaintext: This is the solution. It is the message in plain english.

The Key: The key is the mapping between ciphertext to plaintext.


This is a monoalphabetic substitution cipher, which is one of the easiest kinds. It has one single mapping as the key. So, for instance, we the letter 'F' in the ciphertext might map to 'L' in the plaintext. This mapping will remain constant for the entire message.


Substitution ciphers are cracked by using statistical analysis. English is highly non-random, as anyone who has watched Wheel of Fortune will know. The frequencies for the english letters are as follows:

So you tally up all of the ciphertext letters and look at their frequency distribution and try to match it up to the corresponding English distribution. (IE: The most frequent letter will be E, then T, etc...) More specifically, you want to minimize the total difference between distributions. Which is something easily done with a computer program.

The statistical attack gets more potent the longer the ciphertext is. This puzzle seems to imply that making it longer makes it harder, but it's actually the other way around.

Where can I get a program like that, Alt? I use Mac. Also, where can I learn more about cracking codes?

 

[Tom]

KILLYOURPARENTSKILLYOURPARENTSKILLYOURPARENTS

hmm i dont think i cracked it right...

 

[AltF4]

Where can I get a program like that, Alt? I use Mac. Also, where can I learn more about cracking codes?

Honestly, Wikipedia is the best place you're going to get a beginner's intro to cryptography. Other than that, nearly anything by Bruce Schneier is really good.

As for programs, I made a bunch a year or two back for a class. But I seemed to have lost them now. Pitty. I'll probably re-write them pretty soon. When I do, I'll post the code.

 

[OlimarFan]

You know, this reminds me of Professor Layton...

Tee hee... It's always nice to have a little challenge! (Well for me, anyway).
Would you folks be interested into cracking another code, or solving a puzzle?

Kind regards,
olimarfan

 

[~N9NE~]

Definitely would Olimar

 

[AltF4]

I just found this web app that's pretty good.

http://www.blisstonia.com/software/WebDecrypto/index.php

I'm in the process of writing my own right now, too. It'll be ready to post soon. Right now all it does is the analysis as I described above. Which totally works for long texts, but shorter ones it trips up on since the frequencies aren't precise enough.

I need to add in extra features to make it more accurate. Like recognising that single character words must be A, I, or O. And adding in an entropy checked mutation. That will be cool.

 

[OlimarFan]

Definitely would Olimar

Sweet. Here is a little puzzle, courtesy to Professor Layton: Diabolical Box

20 -> 4 -> 16 -> 37 -> 58 -> ?? -> 145 -> 42 -> 20 -> 4 -> 16 -> 37 -> 58 -> ??
-> 145 -> 42 -> 20 -> ...

Which number should go into the space with the question mark? Note that this pattern is a number cycle.

I hope I haven't made it too confusing.
Kind regards,
olimarfan

P.S. Which online chess database do you use to play matches with?

 

[hyperhopper]

Looks like a plain old monoalphabetic substitution cipher. Trivially broken. Let me see if I can find any of my old cipher crackers around...

cipher crackers? can i have a link to one pretty plz?

 

[Crimson King]

Alt, question:

I went to this recreational site for cryptograms. Using the statistical stuff, I was wrong every time. The articles I read said the most common letter is E, the most common double letter combo is TH, and the most common triple letter combo is THE, but it failed all the Cryptograms.

Where am I going wrong?

 

[AltF4]

Well, you also have to consider that when people write cryptograms, they sometimes tend to intentionally write them so as to disobey the usual tendencies of English.

A good example is Gadsby: Champion of Youth. It's an entire 260 page book written without using the letter 'e' a single time. Impressive!

It also throws a bit of a wrench into statistical analysis if you were expecting a lot of 'e's as is typical. This is not, of course, to say that statistical analysis is impossible, nor effective. It just means you have to recalibrate it. In the process of avoiding the letter 'e', Gadsby overuses other letters and phrases.

Without getting into a big lecture about Entropy and Information Theory, natural languages will always be susceptible to statistical analysis under substitution ciphers. It's because they perform no Diffusion when encrypting. There are two aspects to cryptographic algorithms: Confusion and Diffusion. Substitution ciphers only perform Confusion.


So if you try decoding under the assumption that "THE" is the most common 3 letter word, and it doesn't work out, change your assumptions.

 

[~N9NE~]

A good example is Gadsby: Champion of Youth. It's an entire 260 page book written without using the letter 'e' a single time. Impressive!

That is unbelievable.

I just used the letter e three times in three words, and they didn't use it for 260 pages??

lol

 

[AltF4]

I nearly forgot!

Here is a link to the source of a real quick and dirty frequency analyser I wrote a week or two ago. It is pretty shallow and needs improvement. And I'm pretty sure there's still an annoying bug on one of the inputs I tried. But whatever. I'll look at it again sometime. But since school's going to start in not long I'll probably be too busy.

SOURCE CODE

BTW: Written in NetBeans on Ubuntu 9.10, gcc 4.4.1

Also: I didn't really document that file very well. To use to program, just pass in a single string parameter as the ciphertext and let it do its magic. It tends to work better on larger inputs, and even then not very well yet. I only spent an hour or two on it, give me a break!