Using Chi-Square for Vigenere Cipher

Question

I am writing a program which will use Chi-Square to determine which is the correct keyword for a ciphertext via Vigenere Cipher.

I came across a website that describes the Chi-Square statistics in a rather simplistic manner and had also written some code to achieve my goals. However, the actual keyword does not seem to be the lowest chi-square value as according to the website. I'm not sure if the method as describe is correct or not.

Below is the weblink: http://practicalcryptography.com/cryptanalysis/text-characterisation/chi-squared-statistic/

Would someone be able to provide some advice on this?

Thanks.

Answer 1

This method isn't deterministic. That is, just because a keyword has the lowest chi-square value, that doesn't mean it's the keyword. All it means that there is a good chance that it is the keyword. It's not a guarantee. You should look at the first several likely options to see if the keyword lies there. If you want to automate this, you could compare the possible keywords against known words using some dictionary file.