Comment by zellyn
2 years ago
The most interesting question this has always raised for me is whether it implies you should always try two doctors/dentists/therapists/etc. before making a choice. Seems like it would be just as powerful “in real life“.
There's a whole area of math in "optimal stopping" which aims at exactly this question: how many dentists should you try before you pick your favorite?
For example: https://en.wikipedia.org/wiki/Secretary_problem
That problem is always sort of intuitively misleading. The success criterion is defined as picking the single best option from the field. But that's not how real life works. You don't have to pick the single best option for a job or spouse or dentist or whatever; you'll likely be satisfied with anything in the top ten or quintile or somewhere around there. It's interesting mathematically, but it's seldom really applicable to real life.
Well, this strategy would only help you find ones that are not over loaded? And probably only works if you can change choices over time?
That said, sampling, in general, works surprisingly well in all things.
That reminds me of a great book -- "Algorithms to Live By" -- which examines precisely this kind of question. Useful, educational, and entertaining = highly recommended.