histogram.cpp The starting point of any statistical attack on a substitution cipher is to first find the probability distribution of each of the characters that occur. We can approximate this for long inputs by analyzing the relative frequency of each letter. For example try this command. ./hist