Comment Meaningless Math (Score 3, Insightful) 684
If Jimbo tells you that there's a 1% chance that your tire will go flat if you don't fix it, that's not 1% if Jimbo is wrong 50% of the time. At best, it's 50.5%.
But you assume that Jimbo's being wrong means that the probability of failure is 100%! It's not necessarily. In fact, Jimbo might be wrong in that the probability of a flat tire is actually 0% -- in which case, his being wrong has helped you. If this is the case, then the total probability is 0.5%, much better than 1%. This is the best case; 50.5% is the worst case, and neither is "more likely", because we don't know what the conditional probabilities are. It's this fallacious reasoning -- that if the theory is wrong, the probability of the event must be greater -- that make this article technically true, but useless. We cannot handpick these probabilities. From the TFA (not the abstract):
The other unknown term in equation (1), P(X|not A) [read: the probability of the catastrophe given we're wrong], is generally even more difficult to evaluate, but lets suppose that in the current example, we think it highly unlikely that the event will occur even if the argument is not sound, and that we also treat this probability as one in a thousand.
(emphasis and comment mine). I disagree. This probability is impossible to evaluate, and so this paper means nothing.