If you actually try to flesh out the reasoning behind the distance estimation strategy it will turn out 100x more convincing than the analogous argument for bayesian probability estimates. (and for any bayesians reading this, please don't multiply the probability by 100)