All ciphers are crackable? BULLSPIT!!!

[ironizer]

I've heard people say that all ciphers (speaking of substitution/translation) are crackable, especially with brute force (supercomputers). I bet 1000 bucks I can make a code that will be impossible to "crack". Here's how it will work.

 

I will make a small book that will explain how to encode/decode the gibberish. It will have what letter translates to what, etc. Then, I will write a paragraph, page, (an excerpt from a book, or anything in the english language) with my code. I'll give you both the english text and the code. Then i'll give you the code to a new paragraph, and you have to figure out what it says.

.... you won't be able to. But: someone with the little book i make will be able to easily translate the code into the english language.

 

I want to know how this cracking works. How would it eve be possible to attempt to guess my code's pattern?

[Cap'n Refsmmat]

Depends on how your small book works. If it says "A is always Q" and so on, it'd be insanely easy to decode. That's how those "cryptograms" in the newspaper work. If you had a more complex polyalphabetic cipher, it'd take more complex mathematics and a properly trained cryptographer to beat you.

 

However, one-time pads have been in fact proven to be unbreakable. They work by using a random key the same length as the message. As long as the key is secure, the message is too. (And you can't re-use the key.)

[ironizer]

It would be a pretty much like "A is Q" except i have some kick-butt tricks up my sleeves to hide the patterns.

I don't know what you mean by that key, but mine you would use the "book" every time you want to encode something. The problem is that if somebody gets a hold of that "book", they can decode any message very easily. I want make a bet with a cryptographer to see if they can beat my code. Problem is, I don't know any cryptographers.

 

Also, brute force and math analysis would do you no good. Computers don't have intelligence, and they would not be able to think the same way I do.

[Cap'n Refsmmat]

What sort of tricks? A computer could brute-force it and eventually find a message that makes sense.

 

You could try posting a message and the ciphertext and let us have at it.

[ironizer]

What sort of tricks? A computer could brute-force it and eventually find a message that makes sense.

 

You could try posting a message and the ciphertext and let us have at it.

 

I think the problem with brute force is that a computer cannot "notice" patterns. Computers can't just try combos to see if they make sense. My messages will be hidden in the code, so only someone who thinks the same way I do would be able to figure it out.

 

I haven't actually made the code yet, I just have it in my head. It'll probably take me a few days to do it, maybe longer.

[Snail13579]

well, with matrices you could always have a formula for it and code/decode ONLY if you have the right matrix coder.

[Aeternus]

I'd like to point you to

 

http://en.wikipedia.org/wiki/Rubber_hose_cryptanalysis

 

One of the most fun and effective methods

 

If you explain a bit more what you are trying to do, then people can tell you whether your cryptosystem is likely to be any good (this isn't to say that one needs to know what cipher you are using to cryptanalyse it, only that it's much easier to narrow down the field and explain principles without wasting time on a pointless exercise). The fact that you are trying to rely on security through obscurity seems rather ominous of a bad cipher in any case though ( See http://en.wikipedia.org/wiki/Kerckhoffs%27_principle ) .

[SH3RL0CK]

One necessary item required to crack any code is having sufficient samples to solve. There has to be a greater quantity of code than possible answers, otherwise you can only solve for the possible answers and not the actual solution.

 

For example, if the only length of code you have is the character "#" and you have absolutely nothing else, it could be a letter, a word, or a command. If you know it stands for a letter of the English alphabet, then you can only narrow it down to 26 possibilities. However, if you also knew this was a one letter English word, then the possibilities narrow down to the letters A, I or O. Furthermore if you knew it were a pronoun, then you know it is the letter I.

 

As pointed out by an earlier poster (thank you Cap'n Refsmmat), with a one-pass code, the variables contained within the key exceed the amount of code available. This is not to say that a clever cryptoanalysis would be unable to decipher some information from this code however. Its very length provides a clue to the content of the message, as might the source of the code, the time and conditions the code was sent, etc.

[DJBruce]

Everything is crackable it is just a matter of how much time and effort you are willing to put into cracking it. Im in no way saying that I could crack it but as history shows us everything can and most likely will eventually be cracked.

[John Cuthber]

OTP is uncrackable, even by brute force.

http://en.wikipedia.org/wiki/One_time_pad

Beating the answer out of someone is often the easy way to crack a code but, in principle, OTP is perfect. It's just not very practical.

[Mr Skeptic]

OTP is uncrackable, even by brute force.

http://en.wikipedia.org/wiki/One_time_pad

Beating the answer out of someone is often the easy way to crack a code but, in principle, OTP is perfect. It's just not very practical.

 

Wasn't this one of the options with quantum cryptography? You could use the quantum system to exchange a one time pad, then use that to encrypt your message.

[John Cuthber]

I think you are right; but I'm not sure we agree on the meaning of "practical".

[hobz]

I think you are right; but I'm not sure we agree on the meaning of "practical".

Good one

 

Simon Singh has a good and accessible introduction to coding. Check out http://en.wikipedia.org/wiki/The_Code_Book

 

In this book, Singh describes an encryption method similar to what you propose (if I understand you correctly), and how to analyze such an encrypted message. Much of the work involved in "cracking" the code, could be performed by a computer such as comparing the frequency of letters in the encrypted message to that of the standard language in which the message was written.

[Air]

Provided the initial coding is not changed in the process, all codes should be decryptable. For instance:

 

If '[math]D = T[/math]' then it would need to be kept as that. If changed in the process and unnotified, then you will not be able to decipher.

[insane_alien]

that only works for direct mapping encoding.

 

enigma for instance would not always return T for D.

 

the sequence DDDDDDDDDD might come out TXCAKJDJEM where as direct mapping would come out TTTTTTTTTT

[John Cuthber]

OTP Remains uncrackable.

Why bother with the debate?