On: You are dangerously bad at cryptography

[Estragon]

Regarding statistical biases, you can always pipe expected distributions through the test and compare to the expected transformed distribution using chi-squared.