Interview candidates one at a time — when do you stop and hire?
You're hiring for one position and will interview candidates one at a time in a random order. After each interview you must accept or reject on the spot — no calling anyone back later, and no peeking ahead at candidates you haven't met yet. What strategy gives you the best chance of ending up with the single best candidate of them all?
Reveal the answer
Reject the first roughly 37% of candidates outright (37% being 1/e), then hire the very next one who beats everyone you've seen so far. This '37% rule' picks the best candidate about 37% of the time — far better than chance, and provably optimal. Mathematicians formalised it in the early 1960s, and Martin Gardner popularised it in his February 1960 Scientific American column as the 'Googol' problem.