I assigned points to each letter based off of how often it is used. Rare letters are a lot of points, and common letters are low. I forget the formula but I end up looking for a score in a certain range based off of different texts I have thrown in to calibrate it. Its pretty good at narrowing down large possibilities. However given K1-3 have X’s and Q’s thrown in liberally, I had to use a wider range which results in a lot more false positives.
↧