alt Hacker NewsYour comment reads like someone with no experience in medical trials - neither pharmaceutical nor medical devices. Death is not the only negative outcome. Nor is a procedure that is commonly followed by death necessarily a bad procedure - many cancer treatments for example fall into this category.
You’re not understanding the point of the hypothetical, it was to illustrate a scenario in that most people would consider evidential and point out that by definition the opposite outcome must also be evidential even if that outcome was far more likely. It’s to highlight an unintuitive aspect of probability formalized as the bayes ratio. The strength of the evidence is dependent on the priors but it exists. Anecdotal evidence is not zero evidence it is weak evidence and with care a lot of anecdotal evidence can be combined to create stronger forms of evidence. Rounding it down to zero or up to one are both incorrect.