Celarix comments on What epsilon do you subtract from “certainty” in your own probability estimates?