Optimal stopping §

Date created: 2021-09-25

It’s a type of decision-making problem where you have multiple options to choose from and want to pick the optimal one. When should I stop looking and commit to the best option? Some common examples are hiring (which candidate should I pick?), dating, selling a house.

In some cases you have a lot of information and can be very specific with your strategy, for example if you are selling your house and know the range of possible offers as well as the cost of waiting. But in dating or hiring you do not have such information so you need to use a strategy called “look, then leap”. The idea is that you commit to only “looking” for a certain amount of time to get a grasp of the possible options out there, and at a certain point you “leap” and commit to the best option that comes along. Usually the optimal stopping in cases like this is around 37%. This can be after 37% of applicants to a job position, or after 37% of expected lifetime. However this number varies depending on whether or not you can go back to a previous option.

---

References §

• Chapter 1 of Algorithms to live by