“Reality is not always probable, or likely.”― Jorge Luis Borges
As I am supposed to discuss Teddy Seidenfeld‘s talk at the Bayes, Fiducial and Frequentist conference in Harvard today [the snow happened last time!], I started last week [while driving to Wales] reading some related papers of his. Which is great as I had never managed to get through the Dutch book arguments, including those in Jim’s book.
The paper by Mark Schervish, Teddy Seidenfeld, and Jay Kadane is defining coherence as the inability to bet against the predictive statements based on the procedure. A definition that sounds like a self-fulfilling prophecy to me as it involves a probability measure over the parameter space. Furthermore, the notion of turning inference, which aims at scientific validation, into a leisure, no-added-value, and somewhat ethically dodgy like gambling, does not agree with my notion of a validation for a theory. That is, not as a compelling reason for adopting a Bayesian approach. Not that I have suddenly switched to the other [darker] side, but I do not feel those arguments helping in any way, because of this dodgy image associated with gambling. (Pardon my French, but each time I read about escrows, I think of escrocs, or crooks, which reinforces this image! Actually, this name derives from the Old French escroue, but the modern meaning of écroué is sent to jail, which brings us back to the same feeling…)
Furthermore, it sounds like both a weak notion, since it implies an almost sure loss for the bookmaker, plus coherency holds for any prior distribution, including Dirac masses!, and a frequentist one, in that it looks at all possible values of the parameter (in a statistical framework). It also turns errors into monetary losses, taking them at face value. Which sounds also very formal to me.
But the most fundamental problem I have with this approach is that, from a Bayesian perspective, it does not bring any evaluation or ranking of priors, and in particular does not help in selecting or eliminating some. By behaving like a minimax principle, it does not condition on the data and hence does not evaluate the predictive properties of the model in terms of the data, e.g. by comparing pseudo-data with real data.
While I see no reason to argue in favour of p-values or minimax decision rules, I am at a loss in understanding the examples in How to not gamble if you must. In the first case, i.e., when dismissing the α-level most powerful test in the simple vs. simple hypothesis testing case, the argument (in Example 4) starts from the classical (Neyman-Pearsonist) statistician favouring the 0.05-level test over others. Which sounds absurd, as this level corresponds to a given loss function, which cannot be compared with another loss function. Even though the authors chose to rephrase the dilemma in terms of a single 0-1 loss function and then turn the classical solution into the choice of an implicit variance-dependent prior. Plus force the poor Pearsonist to make a wager represented by the risk difference. The whole sequence of choices sounds both very convoluted and far away from the usual practice of a classical statistician… Similarly, when attacking [in Section 5.2] the minimax estimator in the Bernoulli case (for the corresponding proper prior depending on the sample size n), this minimax estimator is admissible under quadratic loss and still a Dutch book argument applies, which in my opinion definitely argues against the Dutch book reasoning. The way to produce such a domination result is to mix two Bernoulli estimation problems for two different sample sizes but the same parameter value, in which case there exist [other] choices of Beta priors and a convex combination of the risks functions that lead to this domination. But this example [Example 6] mostly exposes the artificial nature of the argument: when estimating the very same probability θ, what is the relevance of adding the risks or errors resulting from using two estimators for two different sample sizes. Of the very same probability θ. I insist on the very same because when instead estimating two [independent] values of θ, there cannot be a Stein effect for the Bernoulli probability estimation problem, that is, any aggregation of admissible estimators remains admissible. (And yes it definitely sounds like an exercise in frequentist decision theory!)
On Monday, I took part in a celebration of the remarkable career of a former colleague of mine in Rouen, Gérard Grancher, who is retiring after a life-long position as CNRS engineer in the department of maths of the University of Rouen, a job title that tells very little about the numerous facets of his interactions with mathematics, from his handling of all informatics aspects in the laboratory to his support of all colleagues there, including fresh PhD students like me in 1985!, to his direction of the CNRS lab in 2006 and 2007 at a time of deep division and mistrust, to his numerous collaborations on statistical projects with local actors, to his Norman federalism in bringing the maths departments of Caen and Rouen into a regional federation, to an unceasing activism to promote maths in colleges and high schools and science fairs all around Normandy, to his contributions to professional training in statistics for CNRS agents, and much, much more… Which explains why the science auditorium of the University of Rouen was packed with mathematicians and high schools maths teachers and friends! (The poster of the day was made by Gérard’s accomplices in vulgarisation, Élise Janvresse and Thierry Delarue, based on a sample of points randomly drawn from Gérard’s picture, maybe using a determinantal process, and the construction of a travelling salesman path over those points.)
This was a great day with mostly vulgarisation talks (including one about Rasmus’ socks..!) and reminiscences about Gérard’s carreer at Rouen. As I had left the university in 2000 to move to Paris-Dauphine, this was a moving day as well, as I met with old friends I had not seen for ages, including our common PhD advisor, Jean-Pierre Raoult.
