Yesterday I told the problem to a smart non-math-geek friend, and he totally couldn’t relate to this “only reasonable interpretation”. He completely understood the argument leading to 13⁄27, but just couldn’t understand why do we assume that the presenter is a randomly chosen member of the population he claims himself to be a member of. That sounded like a completely baseless assumption to him, that leads to factually incorrect results. He even understood that assuming it is our only choice if we want to get a well-defined math problem, and it is the only way to utilize all the information presented to us in the puzzle. But all this was not enough to convince him that he should assume something so stupid.
Yesterday I told the problem to a smart non-math-geek friend, and he totally couldn’t relate to this “only reasonable interpretation”. He completely understood the argument leading to 13⁄27, but just couldn’t understand why do we assume that the presenter is a randomly chosen member of the population he claims himself to be a member of. That sounded like a completely baseless assumption to him, that leads to factually incorrect results. He even understood that assuming it is our only choice if we want to get a well-defined math problem, and it is the only way to utilize all the information presented to us in the puzzle. But all this was not enough to convince him that he should assume something so stupid.
For me, the eye opener was this outstanding paper by E.T. Jaynes:
http://bayes.wustl.edu/etj/articles/well.pdf
IMO this describes the essence of the difference between the Bayesian and frequentist philosophy way better than any amount of colorful polygons. ;)