Taking the bet in isolation, it would only make economic sense to make any bet at even odds if the probability of the event occurring is > 50%. If Martin Rees

Taking the bet in isolation, it would only make economic sense to make any bet at even odds if the probability of the event occurring is > 50%. If Martin Rees was honestly asserting that the probability of a "bioterror / bioerror" event in that 4-year window was >50%, I think he was clearly wrong, and falling victim to the "availability bias" that Pinker describes. Now, this may be a situation where the sides think they're talking about different things: Pinker may be arguing that the probability is <50%, while Rees may be arguing that the probability, while low, is non-negligible. In which case, in my opinion, they'd both be right: But Rees' bet (again in isolation as a bet) isn't rational (since the expected payoff is negative), while Pinker's is.