The Monty Hall Problem

Let's Make a Deal

In September 1990, Marilyn vos Savant — listed in the Guinness Book of World Records for the highest recorded IQ — answered a reader's question in her Parade magazine column. The response would generate roughly 10,000 letters, most of them angry, many from professional mathematicians. She was right. They were wrong. And the story of why they were wrong is more interesting than the puzzle itself.

Here's the question, more or less as it appeared. You're on a game show. The host, Monty Hall, shows you three doors. Behind one is a car. Behind the other two are goats. You pick a door — say, Door 1. Monty, who knows what's behind every door, opens another door — say, Door 3 — and shows you a goat. Then he asks: "Do you want to switch to Door 2?"

Should you switch?

If you're like most people, your gut says it doesn't matter. Two doors left, one car, one goat — it's fifty-fifty. Switching, staying, who cares? Flip a coin.

Your gut is wrong. You should always switch. Switching wins two-thirds of the time. Staying wins only one-third.

And if you don't believe me, you're in excellent company. Paul Erdős — one of the most prolific mathematicians in history — didn't believe it either. Not until someone showed him a computer simulation.1

The Letters

When vos Savant published her answer ("Yes, you should switch"), the response was volcanic. She received an estimated 10,000 letters. About a thousand of them came from people with PhDs. Here's a representative sample:

The anger wasn't subtle. These were credentialed experts telling the smartest woman in America she didn't understand middle-school probability. They were so confident, and so wrong, that the whole affair became a case study in how even trained minds fail at conditional probability.2

The Math That Shouldn't Be Hard (But Is)

Let's walk through it carefully. You pick Door 1. At the moment you pick, the probability the car is behind each door is straightforward:

Now here's the key insight. When you picked Door 1, you effectively split the universe into two groups:

• Your door — probability 1/3
• The other doors — probability 2/3 (combined)

Nothing Monty does can change the probability that you initially picked correctly. You had a 1-in-3 chance, and that's frozen in time. But look at what Monty does do: he opens one of the "other" doors and shows you a goat. He's required to show you a goat — he never opens the door with the car. So the 2/3 probability that lived in "the other doors" now concentrates entirely on the one remaining door.

Switching gets you the car 2/3 of the time. Staying gets it 1/3. That's it. That's the whole proof.

But if it still feels wrong, you're not alone. Let me try a different angle.

The "But It's 50/50 Now!" Objection

The objection always sounds the same: "Once Monty opens a door, there are two doors left. One has a car, one has a goat. That's 50/50. History doesn't matter."

The mistake is treating the problem as if you'd just wandered in off the street and seen two closed doors. But you didn't wander in. You picked a door. And Monty's choice of which door to open was constrained by your choice. He couldn't open your door. He couldn't open the car door. That constraint is where the information lives.

Think of it this way. There are only three scenarios, all equally likely:

Count them up. Switching wins in two out of three equally likely scenarios. Staying wins in one. There's your 2/3.

Don't believe the logic? Try it yourself.

The Hundred-Door Gut Punch

If three doors still feel ambiguous, scale it up. Imagine 100 doors. One car, ninety-nine goats. You pick Door 1. Monty — who knows everything — opens 98 other doors, all goats. He leaves just one other door closed. Door 47, say.

Now he asks: "Want to switch?"

Suddenly the answer is obvious. Your original pick had a 1-in-100 chance. The other 99 doors collectively had a 99/100 chance. Monty just showed you which 98 of those 99 are losers. All that probability — 99% of it — is sitting behind that one remaining door.3

Of course you switch. You'd be insane not to. And the logic is exactly the same with three doors. The numbers are just less dramatic.

Monty Is a Bayesian Machine

Here's what's really going on, mathematically. When Monty opens a door, he's giving you information. Not about your door — about the others. And information, in probability, means you should update your beliefs.

This is Bayesian reasoning. You start with a prior (each door is 1/3). You observe evidence (Monty opens Door 3 and it's a goat). You compute a posterior. The posterior probability that the car is behind Door 2 — given that Monty opened Door 3 — is 2/3.4

The "50/50" crowd makes a specific error: they treat Monty's action as if it were random. If a random earthquake knocked Door 3 open and there happened to be a goat behind it — then it would be 50/50. Because an earthquake doesn't know where the car is. But Monty does know. His action is constrained by the truth, and constraints carry information.

This is why conditional probability trips people up. It's not enough to count outcomes. You have to account for the process that generated what you observe. The same observation — "Door 3 has a goat" — carries different information depending on why you're seeing it.

Where You've Seen This Before (Without Knowing It)

Medical Testing

You get a positive result on a screening test for a rare disease. The test is 99% accurate. Do you have the disease? Almost certainly not — if the disease affects 1 in 10,000 people, most positive results are false positives. The prior matters enormously, just like your initial 1/3 in Monty Hall. The test result (like Monty's reveal) is information, but it doesn't overwrite your prior — it updates it.5

Poker

Every card that hits the table is a Monty Hall reveal. When the flop comes and you see three cards, you're not just looking at what's there — you're updating your beliefs about what's in your opponents' hands. Good poker players intuitively do Bayesian updating every hand. Bad ones think in terms of "what are the odds now?" without connecting to what happened before.6

Deal or No Deal

The entire show is a prolonged Monty Hall problem. You pick a briefcase. Cases get opened. The banker makes offers. Every opened case is information. But here's the twist: in Deal or No Deal, the cases are opened randomly (by the contestant, who doesn't know the values). So unlike Monty Hall, the reveals don't systematically avoid the big prize. The information content is different — which is exactly the point.7

Why Your Brain Hates This

The Monty Hall problem isn't hard because the math is hard. The math is arithmetic. It's hard because your brain has a deep commitment to a heuristic that usually works: when you see two options, assign them equal probability.

This is sometimes called the "principle of indifference," and it's ancient — Laplace wrote about it in the 18th century. In the absence of information, it's reasonable. But the Monty Hall problem is not the absence of information. Monty gave you information, and your brain just… threw it away.

The broader lesson: equal-looking is not the same as equally likely. Two remaining doors look the same. But they arrived at "remaining" via very different paths. Your door survived because you picked it. The other door survived because Monty avoided it. Those are not the same process, and they produce different probabilities.8

Sit with that for a second. Two-thirds of the time, you initially picked a goat. When you did, Monty was forced to open the only other goat door, leaving the car behind the remaining door. Switching wins every single time you initially picked wrong — which is most of the time.

So here's the real lesson of the Monty Hall problem: it's not a puzzle about doors. It's a puzzle about information — who has it, how it flows, and what happens when you ignore it. The PhDs who wrote angry letters to vos Savant weren't bad at math. They were bad at noticing that Monty's constrained choice was data. They looked at two doors and saw symmetry where there was none.

And that, friends, is a mistake you can make anywhere. In the doctor's office, at the poker table, in the courtroom. Whenever someone shows you something and you forget to ask why they showed you that and not something else, you're ignoring the Monty Hall principle. The goat behind the opened door isn't just a goat. It's a message.

Listen to it. Switch.