The analysis is sound, but he completely missed the may report to update his priors.
On Wed, Oct 9, 2013 at 9:59 AM, Jones Beene <jone...@pacbell.net> wrote: > http://ieet.org/index.php/IEET/more/josling20131008 > > In probability theory and statistics, Bayes theorem (rule) is a useful > model > of "conditional probabilities". It is a way to judge improbable future > events which derives from the more basic axioms of probability but it > discounts the scientific objections in favor of actual results - but of > course, assumes some degree of honesty. > > According to this video, Rossi may have almost doubled the prior > probability > of LENR in the eyes of statisticians. (it is still very low). > > Where is "Blaze," the probability guru, when we need him? > > I agree with one commenter that you can't really use linear thinking to > predict a non-linear event like a paradigm change... but that does not keep > betters from trying to predict almost anything. >
