Qbasicnews.com

Interesting. I made something that compares English letter frequencies and it has about half the letters correct. It's possible to make it more correct by adding in letter frequencies for common words, etc. I don't want to post it yet though, 'cause half correct letters still look like mumbo jumbo. Maybe when/if I put in the 2-letter word frequencies...

I found a cryptogram solver which generates a random key, compares the frequencies to plaintext, and then modifies the random key until it gets the best match. Then it generates another random key, and so on, until the plaintext is mostly revealed, at which time you halt the program and copy the answer. Basically a genetic algorithm.