WEBVTT
00:00:02.200 --> 00:00:08.160
In this video, we’ll be looking at the famous Monty Hall problem, which was based on a TV game show called Let’s Make a Deal.
00:00:09.400 --> 00:00:12.160
The show started in the United States in 1963.
00:00:12.160 --> 00:00:15.800
But it’s been running in various formats around the world ever since.
00:00:16.280 --> 00:00:18.480
It was originally hosted by a chap called Monty Hall.
00:00:18.840 --> 00:00:20.360
And that’s how this problem got its name.
00:00:22.000 --> 00:00:25.920
The format is that during the show, people are picked to make trades to win a prize.
00:00:26.240 --> 00:00:33.160
And they usually end up with a chance to trade a small prize for a much bigger more valuable one, hidden behind a curtain or a door or in a box.
00:00:34.160 --> 00:00:46.000
The thing is they may end up getting a fabulous prize, like a new car, or they may end up with a dud prize, like a pile of fake money or in item of dodgy clothing or even a live animal which would be very tricky to take home.
00:00:46.560 --> 00:00:48.400
These dud prizes are called zonks.
00:00:49.640 --> 00:00:56.280
Now the Monty Hall problem takes one of these trade scenarios and asks if you can come up with a strategy to help you to decide how to play.
00:00:56.760 --> 00:00:57.560
It goes like this.
00:00:58.520 --> 00:01:00.760
You’re invited to play the big deal of the day.
00:01:01.200 --> 00:01:09.400
The host stands you in front of three closed doors and tells you that behind one of them is the brand new car of your dreams and behind the other two are zonks, in this case goats.
00:01:10.640 --> 00:01:12.560
The doors are numbered one, two, and three.
00:01:12.560 --> 00:01:17.880
And all you have to do is choose one of the doors and you’ll get the prize or zonk that lives behind that door.
00:01:18.200 --> 00:01:18.560
Easy!
00:01:20.080 --> 00:01:21.440
So you make your choice.
00:01:21.440 --> 00:01:22.840
For example, you say, “Door Two!”
00:01:23.400 --> 00:01:25.560
Now the host knows which door the car is behind.
00:01:26.160 --> 00:01:35.880
And instead of letting you open door two, he dramatically opens one of the other doors to reveal a goat calmly chewing on some grass [goat bleating].
00:01:36.800 --> 00:01:42.440
Now here’s the twist, he says, “Okay, would you like to stick with door two or would you like to switch to the other closed door?”
00:01:44.000 --> 00:01:46.960
Well can you change your odds of winning the car by sticking or switching?
00:01:47.480 --> 00:01:48.400
Does it make a difference?
00:01:48.920 --> 00:01:50.160
There are two doors left to choose from.
00:01:50.880 --> 00:01:52.520
And one has a goat, and the other has a car.
00:01:52.880 --> 00:01:53.720
It’s 50/50, right?
00:01:55.000 --> 00:01:59.360
Well pause the video now and think about what you’d do before I explain the maths.
00:02:05.320 --> 00:02:08.520
Okay, this problem caused a great deal of controversy when it was first published.
00:02:08.720 --> 00:02:11.080
And lots of people disagreed about the best strategy.
00:02:11.560 --> 00:02:12.960
So let’s go through it step by step.
00:02:14.320 --> 00:02:16.280
It seems that you have a choice of two doors.
00:02:16.640 --> 00:02:17.360
And one’s got a goat.
00:02:17.600 --> 00:02:18.680
And one’s got a car behind it.
00:02:19.360 --> 00:02:21.960
Surely it’s obvious that it makes no difference which one you pick.
00:02:21.960 --> 00:02:25.280
They’re both equally likely to have the car or the goat behind them.
00:02:26.760 --> 00:02:29.360
But this isn’t taking the whole story into account.
00:02:30.640 --> 00:02:34.800
When you announced your choice of door, the host opened a different door that had a goat behind it.
00:02:36.200 --> 00:02:38.120
Now the host knows where the car is.
00:02:38.440 --> 00:02:43.360
So if you chose the door with the car, then they could open either of the other doors to reveal a goat.
00:02:44.480 --> 00:02:50.840
However, if you chose the door with a goat behind it, then the host had to be very careful to pick the other door with a goat behind it to open.
00:02:52.600 --> 00:03:01.200
And this sequence of events has left us in the situation where if you were right about which door had the car behind it to start with, then if you switch doors you will now be wrong.
00:03:02.640 --> 00:03:07.760
And if you were wrong about which door had a car behind it to start with, then if you switch doors you’ll now be right.
00:03:09.120 --> 00:03:19.120
At the beginning, when you are faced with three doors and you chose one at random, there was a 33 and a third percent chance you’d get the car and a 66 and two-thirds percent chance that you get a goat.
00:03:20.240 --> 00:03:21.960
One of the three doors has a car behind it.
00:03:22.280 --> 00:03:24.720
And the other two out of the three have goats.
00:03:26.120 --> 00:03:32.640
That means if you were to star in the show every week and stick with your original choice every time, then you’d get a car one-third of the time.
00:03:34.000 --> 00:03:41.920
However, if you were to star in the show every week and switch your choice, then the times that you are originally wrong about which door hid the car become times you get the car.
00:03:42.200 --> 00:03:43.200
That’s two-thirds.
00:03:44.560 --> 00:03:48.880
Now the times that you were originally right about which door hid the car become times you get the goat.
00:03:49.200 --> 00:03:50.080
That’s one-third.
00:03:51.280 --> 00:03:54.600
So the best strategy for getting the car statistically is to switch.
00:03:54.920 --> 00:04:00.120
You’re increasing your chances of winning a car from 33 and a third percent to 66 and two-thirds percent.
00:04:01.120 --> 00:04:03.600
That’s a 100-percent increase in your chances of winning.
00:04:04.160 --> 00:04:06.160
But we can talk about percentages another time.
00:04:07.360 --> 00:04:09.760
Now let’s use a probability tree to write all of this down.
00:04:10.920 --> 00:04:14.560
First let’s consider the strategy where you stick with the door that you originally chose.
00:04:15.000 --> 00:04:16.240
First you got to choose the door.
00:04:17.240 --> 00:04:19.720
Now in a third of cases, that door will have a car behind it.
00:04:20.680 --> 00:04:23.160
And two-thirds of cases, it will have a goat behind it.
00:04:24.480 --> 00:04:33.960
Now given that you pick the car first time, if you stick with that choice, there’s a probability of one you’ll keep the car and there’s a probability of zero that you’ll get a goat.
00:04:35.160 --> 00:04:45.920
But if the door you originally chose had a goat behind it and you stick with that choice, there’s a zero probability that you’ll end up with a car and absolute certainty that you’ll end up with a goat.
00:04:46.120 --> 00:04:56.800
Now if we multiplied these conditional probabilities together along the branches, we find that the probability that we started off choosing a car and ended up with a car on this strategy of sticking is a third.
00:04:57.520 --> 00:05:01.040
The probability of starting off choosing a car and ending up with that goat is zero.
00:05:01.480 --> 00:05:04.560
The probability of starting up with a goat ending up with a car is zero.
00:05:04.560 --> 00:05:07.600
And the probability of starting off with a goat ending up with a goat is two-thirds.
00:05:08.440 --> 00:05:14.520
This means the overall probability of ending up with a car is one-third plus zero, which is a third.
00:05:15.440 --> 00:05:21.280
And the overall probability of ending up with a goat is zero plus two-thirds, which is two-thirds.
00:05:22.120 --> 00:05:23.520
So let’s make a note of that over here.
00:05:23.920 --> 00:05:29.200
If we stick, the probability of getting a car is a third and the probability of getting a goat is two-thirds.
00:05:30.480 --> 00:05:32.640
Now let’s consider the strategy of switching.
00:05:33.560 --> 00:05:39.960
When you choose a door at random, there’s still a probability of one-third of that being a car behind it and two-thirds of that being a goat behind it.
00:05:41.960 --> 00:05:51.640
Now because the host revealed the other door, it means if the door that we chose had a car behind it and we switch, we’ll now definitely have a goat and we definitely won’t get the car.
00:05:53.000 --> 00:05:59.840
And if we chose the door with a goat behind it and we switch, we definitely will get the car and we definitely won’t end up with a goat.
00:06:01.320 --> 00:06:12.280
So with the switching strategy, the probability of ending up with a car is zero plus two-thirds and the probability of ending up with a goat is one-third plus zero, which is one-third.
00:06:13.080 --> 00:06:17.800
So the strategy of switching leaves you with the best probability of ending up with a car.
00:06:19.000 --> 00:06:21.120
Now does this strategy mean that you’ll always win the car?
00:06:21.840 --> 00:06:22.560
Well no it doesn’t.
00:06:22.880 --> 00:06:27.040
But it does mean that you win one two- thirds of the time rather than only one-third of the time.
00:06:27.480 --> 00:06:28.480
So that’s gotta be worth doing.
00:06:29.600 --> 00:06:37.400
To sum it all up then, a little bit of mathematical thinking can help you devise the best strategy to win a car [goat bleating].