Cryptanalysis of Vigenere cipher: not just how, but why it works
Вставка
- Опубліковано 28 сер 2020
- The Vigenere cipher, dating from the 1500's, was still used during the US civil war. We introduce the cipher and explain a standard method of cryptanalysis based on frequency analysis and the geometry of vectors. We focus on visual intuition to explain why it works.
The only background needed is some familiarity with vectors and probabilities.
For more on this method:
Introduction_to_Cryptography_with_Coding_Theory by Wade Trappe and Lawrence C. Washington, Second Edition, Section 2.3. Here the method is exactly as in the video, because this is the text I teach my course from.
Secret_History:_The_Story_of_Cryptology by Craig P. Bauer, Section 3.3. Here the method is done with more mathematical sophistication (so it looks slightly different). This is a very valuable and detailed textbook for anyone interested in old ciphers and their cryptanalysis.
I also have some cryptanalysis tools here you can use to implement the method (which I do with my students in the classroom): crypto.katestange.net/cryptan... - Наука та технологія
Awesome video. Good job!
Note: At 9:48, if anyone is wondering if there is a cos θ in higher dimensions, the Cauchy Schwarz inequality handles just that.
What a brilliant video, you have literally explained it 100 times better than my lecturer! Thanks for the clarification.
Awesome. Been looking for a good vid about this. Glad found this.
i don’t know why youtube suggested me this video but this is what i’ve been looking for without knowing that i’ve been looking for it..
Awesome video! First time learning about this and it makes so much sense! Thank you!!!!!
Very well explained. Thanks!
Best Informational Video I ever watched,
Excelent video! you just saved a students life
Easiest to understand so far... Thanks
Really fun video, much appreciated
Amazing video about decrypting vigenere cipher
Thank you so much, this is so much easier now
Awesome video.
Super Helpful
Very helpful video
Amazing
the matching up of various shifts reminds me of the Fourier transform
that's a good intuition!
Great!
Really useful tutorial. Is the probability that two squares are the same; the same as the Index Of Coincidence? What would you do if there was no clear peak between the number of coincidences, or if a low shift such as 3 peaks, shift 6 doesn't peak and 9 peaks again? If you couldn't tell, this is what I seem to be facing at the moment.
This is epic by the way. Will this work if the cipher alphabet is in ASCII?
We've been trying to identify the cipher type of a ciphertext written in ASCII and index of coincidence/other tests fail because of the ciphertext containing letters, numbers, and symbols. We're not sure how to approach identification of cipher species due to this
great vid
Great video. I'm currently working on a solver for the Keyed Vigenere Cipher (Quagmire III), which is a bit more challenging. Do you have any recommendations for that?
My text is ciphered from the entire text of the lord of the rings which is ciphered from the text of "a pickle for the knowing ones" which is ciphered from the wheel of time series, which is ciphred from the eye of argon. (It's a grocery list)
That is a great explanation i could have saved 4-5 hours .
eargasm
So does anybody know what it means if there are 2 big numbers in a row? I’m doing a kinda small ciphertext but last time it happened with a much larger one. Is it just a …coincidence?
Could you hypothetically use the plaintext to be encoded as the key to encode itself?
Interesting. If I understand you correctly, I think you'd end up essentially doubling every letter, so A=0 goes to A=0, B=1 goes to C=2, C=2 goes to E=4 etc. This would be an example of what's called an affine cipher (in this case, multiplication by 2). One problem is that the alphabet has an even number of letters, so you'd have no way to get odd numbers out. You'd have repeats like A=0 and N=13 both going to A=0. This would make decoding a little bit of a guessing game.
This is basically a polyphonic substitution cipher.
It sounds easy when the key is shorter than the plain text and repeats. What if when the key has the same length as the plain text. I generate the key with a pseudo random generator where the seed is the master key. And then the usual mod26 for coding and decoding. I tested it with free42, a HP42 reimplementation available for all platforms.
When the key has the same length as the plain text this is called a "one time pad" and is unbreakable. this is because for every message, there is a key that could take it to any other message. So without knowledge of the key, any ciphertext could have feasibly come from any message you could think of.
The trick is making sure your random number generator is secure. Since it is not "true" randomness, you could theoretically find patterns in the randomness and exploit them.
Also, keeping track of the current state of the random number generator between different messages is fairly impractical, so often instead you'll use a pseudo random function where both the key and some other seed value such as the current time or a previous message is used.
Total is 7
I rather look at small repeated words, THE, AND...
omg ur voice
i blundered
kreygasm
nope
i lost
Sorry, but with the steps given in the video I was not able to reproduce this approach and therefore I also didn't get valid results.
Please provide a more detailed explanation or a real world example to proof that it works.
I'm sorry you weren't able to get it to work. But I appreciate the request for more resources, and I've added some textbooks and tools to the description of the video.
@@ProofofConceptMath Thank you for your reply and for providing further information on the technique. I will definetly check those out. Btw: I've edited my initial comment. Maybe it was a bit harsh. Sorry about that.
I'm divorced and single