On the Absurdity of the Bayesian Debate
----------------------------------------


From a pragmatic point of view, the Bayesian debate
only serves to confuse the apprentice statistician.  The statistician
is forced to decide beforehand whether to take a Bayesian approach, and
if so, how to choose the so-called prior, or whether to take a
non-Bayesian approach and be condemned to a fate worse than death (as
the Bayesians would have you believe).  We believe that most of the
confusion is caused by the actual names assigned to concepts, such as
"prior", "posterior" and even "estimate".  These names bring with them
misleading connotations.

 From a theoretical perspective, we claim both Bayesians and
non-Bayesians are equally wrong.  Indeed, estimation is an ill-posed
problem.  If you make one set of assumptions, you may believe the
Bayesian school of thought is correct.  If you make a different set of
assumptions, the non-Bayesians look favourable.  We offer a broader
point of view, which we hope exposes the absurdity of the Bayesian
debate.

Bio: http://rsise.anu.edu.au/~jmanton