Go backward to The Tower of Hanoi

Go up to Mathematically Interesting Games

Go down to first subsection The Mathematics Behind the Game

Switch to graphical version (better pictures & formulas)

Go to University of Toronto Mathematics Network Home Page

You get to pick one of the doors, and you will win what is behind it.

However, the way the game works is that the door you pick does not get opened immediately. Instead, the host (Monty Hall) will open one of the other doors to reveal a goat. He will then give you a chance to change your mind: you can switch and pick the other closed door instead, or stay with your original choice.

Which of these two strategies gives you the better chance of winning the car? This simple question recently caused quite a storm of mathematical controversy!

Why don't you play the game a few times to get a sense of which strategy is better, then come back to this page to read more about the controversy and see if you can resolve it.

This question was asked of Marilyn von Savant, who has a column in a Sunday magazine section in many newspapers. She replied that it was better to switch; there was only a 1/3 chance of your original guess being right, and a 2/3 chance that it was wrong, in which case the host has shown you how to "switch, and win." According to Marilyn, your odds of winning are 1 in 3 if you stick with your original choice, 2 in 3 if you win.

Her column prompted a flood of angry letters, some of them from mathematicians, telling Marilyn to brush up on her basic probability. These letters said that the fact that a third door had been opened to reveal a goat didn't alter the fact that the car was still equally likely to be behind either of the two remaining closed doors. According to these letters, your odds of winning are 1 in 2 if you stick with your original choice, and also 1 in 2 if you switch.

Who is right? Or are neither of them right?

See if you can come up with a mathematical justification for one of
these opinions, *and* see if you can figure out what's wrong with
the other one.

To help you, you can try playing the game a few times on our computer simulation, to gather some experimental evidence. But, because the game is random, just playing a few times won't tell you much. For example, flipping a coin 5 times and getting 4 heads is not that unusual (happens with a 5/32 probability), even though the probability of getting a head is only 50% not 80%! So you may have to play many times before your observed proportion of wins gets anywhere close to the predicted probabilities.

Once you've come up with the best mathematical explanation you can, and checked it out against the experimental evidence, you can move on to the next section to read a discussion of the mathematics behind the game.

This page last updated: May 26, 1998

Original Web Site Creator / Mathematical Content Developer: Philip Spencer

Current Network Coordinator and Contact Person: Joel Chan - mathnet@math.toronto.edu

Navigation Panel: Previous | Up | Down to First Subsection | Graphical Version | U of T Math Network Home