In the last blog post, Robert described why we bootstrap our experimentation data. While that may or may not have…