← Back to Paradoxes

The Secretary Problem

The 37% Rule — How the number e optimizes life decisions

The Hiring Game

20
Explore (reject first 37%)
Exploit (select if best so far)
Click "Start" to begin interviewing candidates
0
Interviewed
7
Threshold (n/e)
-
Best So Far
-
Current

The Paradox

The Rules

  • Interview n candidates one at a time
  • After each, you must immediately decide
  • If you reject, you can never recall them
  • Goal: Select the single BEST candidate

The Dilemma

Too eager? You might hire someone good, missing the best who comes later.

Too patient? You might reject the best, settling for a worse candidate.

37%

The magic threshold

r* = n/e ≈ 0.368n
Reject the first n/e candidates, then select the first one better than all you've seen

Monte Carlo Verification

1000
Success Rate (selected the BEST)
36.8%
0 trials | Expected: 1/e ≈ 36.79%

Optimal (37%)

36.8%
Reject n/e, then pick first best

Random

1/n
Pick uniformly at random

First One

1/n
Always pick first candidate

Why 1/e?

Let P(r) = probability of success when rejecting first r candidates.

P(r) = (r/n) × [1/r + 1/(r+1) + ... + 1/(n-1)]

As n→∞, maximizing P(r) gives r = n/e

And P(n/e) → 1/e ≈ 36.79%

The Beautiful Coincidence

The optimal threshold is 1/e of the candidates.

The success probability is also 1/e.

The same transcendental number appears in both!

Real-World Applications

Dating & Marriage

If you plan to date ~10 people seriously, the 37% rule says: date the first 3-4 without committing, then marry the next person who's better than all of them.

House Hunting

Viewing 20 houses? Look at the first 7 to calibrate your expectations, then make an offer on the first one that beats them all.

Hiring Employees

The original problem! With 100 applicants, interview 37 to set your benchmark, then hire the next standout.

Parking Spots

Looking for parking? Pass the first 37% of spots to learn what's typical, then take the first good one.

Selling Your House

Expecting 20 offers? Use the first 7 to gauge the market, then accept the first offer that beats them all.

When to Stop Searching

Any optimal stopping problem! Whether it's job offers, college choices, or vacation rentals—the 37% rule applies.

The Deeper Insight

The secretary problem reveals that gathering information has value, but so does acting on it. The 37% rule perfectly balances exploration and exploitation.

"Look at 37% of your options, then leap at the first one that exceeds your benchmark."

Also known as the Marriage Problem, Dowry Problem, or Best Choice Problem. Analyzed by Lindley (1961) and Dynkin (1963).