Earlier quoted context omitted.
the point was that "the unpopular, unconvincing hypothesis is right" need be treated as a possibility only if the probability is high enough, and knowing that the vast majority are wrong is an important bayesian input in properly estimating that probability
help me out. historically, at what point in bayesian probability estimation did "there is no such thing as ether" theorists go from crank to lets pay attention to these guys
anyways, the threshold is likely when the hypothesis becomes convincing enough to a sufficient proportion or number of experts in the field