The Scam 1992 and the Optimal Stopping problem
Finally completed watching Harshad Mehta series that has trumped Game of Thrones rating on IMDB and levelled Breaking Bad - Wow! that's saying something
Let me say that despite all his smartness, like most of us, he didn't know "when to stop" either.
He too faced the classical secretary problem or as many call it a parking problem or the marriage dilemma.
Here's a refresher for those who don't know - Say we are hunting for a parking spot in a busy area, we often struggle to find the ideal parking space, such that I have to walk less and pay less. The problem always is we stop too early (walk a lot) or too late (u-turn and start again)
Other examples include renting a place, right occasion to open that expensive wine bottle, right time to buy a stock, no. of features in a sprint so to finish in 2 weeks, etc etc.
Towards the end, Harshad says "mujhe 2-3 crore se zyada zaroorat nahi hai (I don't need more than a few crores)" but thats retrospection. Only wish he knew it before hand.
Well obviously, the problem is simple to explain but devilish to solve.
Many mathematicians have suggested to use Rule of 37%. The 37% rule does not tell which one to pick or where to stop but how many options to consider before zero-ing down. To simplify, we set a predetermined time for 'looking' (i.e. we don't choose any parking spot no matter how impressive it is), then leap (commit to the next best spot available). Its also called Look-Then-Leap rule.
How much time to look - 37% and then leap.
So say if you have shortlisted 100 candidates for a position and the goal is to hire single best person for the job. Here, each candidate stands 1% chance of hiring. Now you can interview either all 100 (time is a factor, plus the risk that all the candidates will wait) but mathematicians say you should interview the first 37% and commit to the next best candidate.
References: Algorithms to Live By: The Computer Science of Human Decisions
Book by Brian Christian and Thomas L. Griffiths