The Frequentist
Professor Fisher arrives at The Posterior. The bartender introduces Professor Fisher, who distrusts priors. Dr. Bayes asks what he believes before the evidence arrives.
Behind the joke.
Frequentist methods avoid formal prior distributions and judge a procedure by how it behaves over many repeated samples. Bayesian methods state prior beliefs openly and update them with data. Dr. Bayes's jab is that nobody starts empty-handed: choosing a model, a test and a standard of proof are assumptions too. Declaring that you believe nothing is itself a position.
The transcript.
For readers who prefer dialogue without zooming in.
Read the dialogue ↘
Bartender: Dr. Bayes, meet Professor Fisher.
Dr. Bayes: Extraordinary.
Professor Fisher: I don't trust priors.
Dr. Bayes: That sounds like one.
Dr. Bayes: And what do you believe before the evidence arrives?
Professor Fisher: I trust only what repeated evidence can prove.
Professor Fisher: Nothing.
Dr. Bayes: That's still a prior. Just a smug one.