How the Monty Hall Problem Defies Human Intuition

The Column That Made 10,000 People Write Angry Letters

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 question described a game show scenario: three doors, one car, two goats. You pick a door, the host opens another door revealing a goat, and you're offered the chance to switch. Vos Savant said you should always switch. Within weeks, she received approximately 10,000 letters, nearly 1,000 from PhD holders, telling her she was wrong. She wasn't.

The Setup and the Surprising Answer

The problem is named after Monty Hall, host of the American game show Let's Make a Deal, which aired from 1963 to 1977. Here's the exact scenario:

• Three closed doors stand before you. Behind one is a car; behind the other two are goats.
• You choose Door 1.
• The host, who knows what's behind every door, opens Door 3 to reveal a goat.
• The host asks: do you want to switch to Door 2?

Most people say it doesn't matter—they see two remaining doors and assume a 50/50 chance. The actual probability of winning by switching is 2/3. Staying wins only 1/3 of the time.

Why Switching Wins: The Probability Breakdown

The key insight is that your initial choice has a 1/3 probability of being correct. That probability doesn't change when the host opens a door, because the host's action is not random—he always reveals a goat. The remaining 2/3 probability concentrates entirely on the door you didn't pick.

Three equally likely scenarios. Staying wins in one. Switching wins in two. The math is settled.

The Bayesian Interpretation

Bayes' theorem provides the formal framework. Let C_i represent the event that the car is behind door i. You pick door 1, and the host opens door 3.

Before the host acts, P(C₁) = P(C₂) = P(C₃) = 1/3. The probability the host opens door 3 given the car is behind door 1 is 1/2 (he could open either goat door). Given the car is behind door 2, the host must open door 3—probability 1. Given the car is behind door 3, the host cannot open door 3—probability 0.

After updating with Bayes' theorem, door 2 holds 2/3 probability. The host's constrained action transfers information to the switcher.

Why Smart People Get It Wrong

The error is not about math skills. It's about cognitive architecture. Humans struggle with conditional probability because our brains evolved to assess immediate physical threats, not sequential probability updates. Several specific biases collide in this problem:

• Equiprobability bias: When faced with two remaining options, people default to assuming 50/50 odds regardless of prior information
• Endowment effect: Once you've "chosen" a door, switching feels like giving up something you own
• Failure to track constraints: The host's knowledge and forced behavior change the probability landscape, but people treat the door opening as a random event
• Anchoring: The initial 1/3 probability feels like it should update to 1/2 when one option is removed

Paul Erdős, one of the most prolific mathematicians in history, reportedly refused to believe the answer until he saw a computer simulation. He was not alone.

Simulation Proof: Run It a Thousand Times

The most convincing argument requires no equations. Write a simple program—or use coins and cups—that simulates the game 1,000 times. Track wins for the "always stay" strategy versus "always switch." Stay wins roughly 333 times. Switch wins roughly 667 times. Every simulation, without exception, converges on these proportions.

This empirical approach is what finally convinced many of vos Savant's critics. After her column ran follow-up explanations, schoolteachers across the country ran classroom simulations. The letters of apology started arriving.

The Assumptions That Matter

The 2/3 answer depends on specific conditions that people often overlook:

• The host always knows where the car is
• The host always opens a door with a goat
• The host always offers the switch option
• The car placement is uniformly random

Change any assumption and the answer changes. If the host opens a door randomly and happens to reveal a goat, switching and staying each give 1/2. The host's knowledge is the engine of the entire puzzle.

Related Problems: Bertrand's Box and Three Prisoners

The Monty Hall problem belongs to a family of probability puzzles that exploit the same conditional reasoning gap. Joseph Bertrand posed his box paradox in 1889: three boxes each contain two coins (gold-gold, gold-silver, silver-silver). You draw a gold coin from a random box. The probability you're holding the gold-gold box is 2/3, not 1/2. The logic mirrors Monty Hall exactly.

The Three Prisoners problem, published by Martin Gardner in 1959, presents the same structure in a different costume. Three prisoners await execution; one will be pardoned. A guard reveals which of the other two will die. The revealed prisoner's survival probability transfers to the unmentioned prisoner—just as the opened door's probability transfers to the unchosen door.

Steve Selvin first formulated the Monty Hall version in a 1975 letter to The American Statistician. It was vos Savant who brought it to millions of non-mathematicians fifteen years later.

What the Problem Teaches About Reasoning

The Monty Hall problem endures in probability courses, job interviews, and popular science writing because it exposes something uncomfortable: human intuition about uncertainty is systematically flawed. We confuse "I don't know which one" with "they're equally likely." We treat new information as irrelevant when it arrives through constrained channels.

Entire fields depend on getting these updates right. Medical diagnostics, criminal forensics, financial risk modeling—all require the Bayesian reasoning that the Monty Hall problem tests. A doctor who confuses prior probability with posterior probability misinterprets test results. A juror who ignores base rates convicts the innocent.

The correct answer to the Monty Hall problem is trivially simple once understood. The hard part is accepting that your gut was wrong—and that a columnist in Parade magazine understood probability better than a thousand PhDs who wrote to tell her otherwise.