• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Monty Hall Problem

I agree with Sol. Earlier, I calculated (post #400) that a clueless Monty that opens a random door that is not the player's pick gives equal chances of winning the game regardless of keep or switch. I thought this was fairly obvious, but since it appears that some people won't be convinced without a program and I've a few minutes to spare, here's an ugly MATLAB script:
Code:
trials  = 2e5; cars = 0; valid_trials = 0; for n=1:trials
 car    = 1+floor(3*rand); player = 1+floor(3*rand);  %
 monty  = player;           % Monty guesses until his pick
 while ( monty == player )  % is different from the player's
  monty = 1+floor(3*rand);  % (equivalent to picking a random
 end                        %  door that is not the player's)
 if ( monty ~= car )        % Car not revealed
  valid_trials = valid_trials+1;
  % Player switches to the other door
  for k=1:3 if (k~=player & k~=monty) guess2 = k; end; end
  if ( guess2 == car ) cars = cars+1; end
 end
end; valid_trials, cars/valid_trials
The result was 133172 trials in which Monty did not accidentally reveal the car, 49.82% of the player won by switching.

Yes... but what would happen if he DID reveal the car as he would do 1 out of 3 times when choosing blindly? Do you still get to choose or is the choice suddenly taken away?

If you don't know whether Monty chose blindly or not, you still know that you only have a 1 in 3 chance of having chosen correctly the first time.
 
If Monty always opens a door, always chooses the door randomly, and has revealed a goat, then you have a 50% chance of already having chosen the door with the car.

Sure, in 1/3 of all possible situations you will have already chosen the car, but you will have chosen it in fully half of this restricted set of situations.

It could be that this is not what anyone's arguing (I think several people are arguing several different things) but I thought I'd offer it just in case this is where a misunderstanding is happening.
 
Yes... but what would happen if he DID reveal the car as he would do 1 out of 3 times when choosing blindly? Do you still get to choose or is the choice suddenly taken away?

It's taken away; that iteration doesn't count because it's stipulated that Monty reveals a goat.
 
Last edited:
Can I clarify something articulett?

If you accept that there is only one car behind three doors and one door is open revealing a car, what are the odds that there is a car behind either of the two closed doors?

0%-- but then the game is over. If you are told to pick a number between 1 and 3 you have a 1 in 3 chance of picking the right number. Once the correct number is revealed-- you have a 0% chance or a 100% chance of having picked the right number. If you do a multitude of these trials... you will be in the 100% category 1/3 of the time and in the 0% category 2/3 of the time. What you are really doing in the alternating scenarios is changing whether the hosts offers you a chance to switch or not. If he always offers you the switch... and he reveals the car-- you switch 100% of the time. If he doesn't (2/3 of the time) it doesn't matter what you do. This still means, that overall-- if always given the chance to switch-- you increase your odds by doing so.

The real question is-- if the host reveals the car-- do you still get to switch? Staying with your first choice in every case never increases your odds more than 1 in 3 for winning the car. If you don't know the hosts reasons or you have no additional information, and you want to improve your odds of winning the car, then the only way to increase you odds is to switch if given the opportunity.

Hindsight is always 20/20-- all possibilities have collapsed to 0 or 1.
 
Last edited:
Yes... but what would happen if he DID reveal the car as he would do 1 out of 3 times when choosing blindly? Do you still get to choose or is the choice suddenly taken away?
I've covered two cases of this back in post #400. If the choice is taken away completely, i.e., that Monty revealing the car counts as an automatic loss, then the chances of winning are 1/3 regardless of whether one resolves to keep or switch on a goat. If the choice is not taken away and the player plays it competently (i.e., whenever Monty reveals the car, go for that door), then this counts as an automatic win and the chances of winning are 2/3 regardless of whether one resolves to keep or switch on a goat.
Conclusion: resolving to keep or switch on a goat give the same probability of winning if Monty picks a random door that is not the player's.

I suppose there is a third case that I neglected, that of an incompetent player who is allowed to pick again even if the car is revealed but who resolves to pick not to pick a revealed car. However, I don't think that's what Sol was talking about, and I didn't think that's what you were talking about either.

If you don't know whether Monty chose blindly or not, you still know that you only have a 1 in 3 chance of having chosen correctly the first time.
Edit: Hmm... I think I see where you're going with this. You're really are talking about a player who keeps the same door even if Monty reveals a car. Yes, you're right--probability of winning the game for such a player is 1/3, although I don't believe that's what Sol disputed (but he can clarify his own position better himself).
 
Last edited:
Yes... but what would happen if he DID reveal the car as he would do 1 out of 3 times when choosing blindly? Do you still get to choose or is the choice suddenly taken away?

If you don't know whether Monty chose blindly or not, you still know that you only have a 1 in 3 chance of having chosen correctly the first time.

If you don't know whether Monty is choosing randomly AND he opens a door to reveals a goat results in two scenarios:

1. Monty chose randomly: The possibly for for each remaining door is 1 in 2. Changing doesn't matter.

2. Monty knowingly chose a goat door: Switching increases your win chance to 2 in 3.

I think your objection is with removing the cases in which Monty randomly opens the door with the car from the possible ways you could lose in the original game and you consider that changing the game.
 
If Monty always opens a door, always chooses the door randomly, and has revealed a goat, then you have a 50% chance of already having chosen the door with the car.

Sure, in 1/3 of all possible situations you will have already chosen the car, but you will have chosen it in fully half of this restricted set of situations.

It could be that this is not what anyone's arguing (I think several people are arguing several different things) but I thought I'd offer it just in case this is where a misunderstanding is happening.

This is what I've been trying to say, but nowhere near as clearly as you have.
 
It's taken away; that iteration doesn't count because it's stipulated that Monty reveals a goat.

Correct. So what this scenario is actually doing is changing the odds of Monty asking you if you want to switch or not... it's not changing the fact that staying with your choice gives you a 1 in 3 chance of winning a car. It just means that you have a 1 in 3 chance of not having been offered a choice and a 1 in 3 chance of having the car behind the door monty didn't reveal... yes, your odds increase to 50% by switching... but your odds of having chosen the correct door the first time are still 1 in 3. What you are really getting in this scenario is information that you are not in the situation where the Host has revealed a car. You are in the situation you'd expect to be in 2/3 of the time... half of the time you are in such a situation (which will be 2/3 of the time), you will have the car (you'll be the 1 in 3)-- half of the time you won't. You've eliminated the possibility that the host will be taking away your choice by revealing the car in the blind guess.

Information only tells you whether it's smarter to switch or not. It doesn't change the fact that staying with your first choice gives you a 1 in 3 chance of winning... varying host motivations just give you more information as to whether you might be that 1 in 3-- If you know the host only asks people to switch if they have the car-- then-- if you are asked to switch you don't-- you KNOW you are the 1 in 3. But, this means that you won't be given such an opportunity 2 of 3 times... and so you still never increase you odds of winning to more than 1 out of 3. You've just collapsed the possibilities so that you've become the 1 in 3.
 
Last edited:
Correct. So what this scenario is actually doing is changing the odds of Monty asking you if you want to switch or not... it's not changing the fact that staying with your choice gives you a 1 in 3 chance of winning a car. It just means that you have a 1 in 3 chance of not having been offered a choice and a 1 in 3 chance of having the car behind the door monty didn't reveal... yes, your odds increase to 50% by switching... but your odds of having chosen the correct door the first time are still 1 in 3.


We're assuming that Monty always asks.

Your odds of having chosen the correct door are 1 in 3 over the entire range of possibilities, yes, but if you consider only the subset of those possibilities where Monty always gives a choice, chooses randomly, and has revealed a goat, then you'll have chosen the correct door in half of those.
 
Last edited:
I think anyone can see that if you always stick with your current choice--you will have 1 in 3 chance of winning-- the only way to increase those odds is to change your choice after a goat is shown. The host can always show a goat. The host does not always need to offer you the switch. [...] The salient part of this exercise is that your odds if you stay are not more than 1 in 3-- ever.

No. The host could be using the strategy of only offering a switch if the car is behind your currently chosen door. In that case, the odds if you stay go up to 1 and the odds if you switch go down to 0.
 
Maybe I'm misunderstanding someone's position here, but this is the way this dispute appears to me:
Articulett: chance of winning via sticking with the first choice no matter what happens is just P(first pick is car).
Others: incorporating the added information once an event has happened gives the conditional probability P(first pick is car|info), which is not equal to P(first pick is car)
I don't understand why there is such a case of miscommunication. Those two claims are perfectly consistent.
 
Last edited:
Yes... I was derailed by Claus' claim that your odds change when you pick a certain door.

They don't. The thing that actually is changing in the alternate scenarios is whether the host offers you a choice in changing. You always have a 1/3 chance of winning by staying with the original door-- when offered the option of switching after being shown a goat, you are just getting more information as to whether your odds increase by switching. Unless you have knowledge that the host only offers prizes if he knows you have the car, you can increase your odds of getting the car to 1 in 2 or 2 in 3 by switching. Therefore, if you have no additional knowledge or constraints on the problem-- you increase your odds of getting the car by switching. In fact, the only way you CAN increase your odds is by switching if given the choice. Claus just has a need to prove me wrong whenever he can... even if it means arguing a straw man version or tangent of what I actually said. He inferred that I was mistaken or stupid regarding my claim. I was, however, correct.
 
Last edited:
No. The host could be using the strategy of only offering a switch if the car is behind your currently chosen door. In that case, the odds if you stay go up to 1 and the odds if you switch go down to 0.

Yes... but then the host would only be offering you that choice 1/3 of the time. If the host was using that strategy-- then 2/3 of the time you would never even be offered the choice. If you have no knowledge that the host does this... then the only knowledge you have is that he's shown you a goat (which he can always do) and he's offered you a choice (which he can always do.) Unless you have reason to suspect otherwise (a host only giving you the choice because you picked the car)... you can only increase you odds by changing. Sure, in 100% of the cases where you know why the host gave you the choice-- you will lose by changing. But you will also lose 2/3 of the time because he will never offer you the choice-- he'll just reveal that you didn't pick it the first time.

Yes, you are correct... but if you don't know that the host is using that strategy... then you can only hypothesize that it's the 1 in 9 possibility that the host IS using that strategy and you have managed to choose the right door which is why he's offered you the choice. That's the only info. that you are getting without having prior knowledge or reason to suspect such a thing. If you think this is the strategy and you are right-- you will win every time he offers you the choice-- he just won't offer you the choice if you are wrong. If you think this is the strategy and you are wrong you will only win 1/3 of the time and lose 2/3 of the time.
 
Last edited:
Articulett claimed that the odds the car is behind the door you pick were always 1/3 and never change, even in this scenario. That is false. The odds are perfectly well-defined, and if the door is opened and a goat is revealed they change to 1/2. If the door is opened and a car is revealed they change to 0.

There is no need to discard the games where the car is revealed, but including them has no effect on the odds in cases where a goat is revealed.

No. If you include the games where the cars are revealed and you still allow the choice-- then the person wins by switching 100% of the time. If a goat is revealed, you win by switching 50% of the time... If you don't know which if these scenarios you are dealing with... then you have a 1/3 chance of winning by keeping your choice and a 2/3 chance by switching. If you know that the host always offers a chance to change your choice-- you have a 50% chance of winning by changing your choice if you see a goat and a 100% chance of winning by seeing a car. Your odds of your first choice being the car do not change from 1 in 3. You are only getting info. as to whether you can improve those odds by switching... that's it. If you know the host's choice is random and that he will offer you a choice either way... you still are better off switching... because 1/2 of the time you will DEFINITELY have the car. And 1/2 the time you will have 50% of a chance of getting the car by switching. IF he doesn't offer the choice when he reveals the car-- you still have the 50% odds left if he reveals the goat-- and 0% chance if he reveals the car. What you are really getting is information about whether and how your odds increase if you are given the opportunity to switch. You don't increase your odds of winning the car more than 1/3 by keeping that choice-- you just don't get offered the chance to switch as often in the alternate scenarios.
 
Last edited:
Yes, you are correct... but if you don't know that the host is using that strategy... then you can only hypothesize that it's the 1 in 9 possibility that the host IS using that strategy and you have managed to choose the right door which is why he's offered you the choice. That's the only info. that you are getting without having prior knowledge or reason to suspect such a thing. If you think this is the strategy and you are right-- you will win every time he offers you the choice-- he just won't offer you the choice if you are wrong. If you think this is the strategy and you are wrong you will only win 1/3 of the time and lose 2/3 of the time.

Now I see what was worrying me in what you said earlier. You said
The only way to increase those odds is to switch if given the opportunity.

It would be clearer to say
"The only way to change those odds is to switch if given the opportunity".​

Obviously if your strategy is to choose a door and never to switch, you will, in the long run, win 1/3 of the games. But if your strategy is always to switch when given the chance, you may increase or decrease your odds, depending on what strategy Monty is using.

For instance, if Monty (unknown to you) only gives the opportunity of switching when there is a car behind your chosen door, the strategy of switching if given the choice will ensure that you never win: you have decreased your odds from 1/3 to 0.

A real-life game show host will probably use a mixture of strategies, to keep up the suspense and make sure that nobody can work out how to cheat him. You can only work out if it's better for you to switch or to keep your door if you know precisely what mixture of strategies he uses.
 
Just to be clear articulett, if Monty randomly opens a door and it reveals a goat and he ALWAYS gives you the option to switch do agree that the probability of the car being behind your originally chosen door is now 1 in 2?

Giving Monty the choice of offering or refusing a switch obviously changes the problem.

Edit: The last sentence is incorrect. The odds are 1 in 2 regardless. I confused myself by thinking about Monty having the choice to either open a door or not.


No, if Monty always reveals a goat and always gives you a chance to switch
then you double your odds by switching from the 1/3 you had originally to 2/3. The odds are not 50/50 despite the fact that there are two choices. All he's done is increased the odds in favor of switching. Really. Do the math. If he chooses blindly then he's going to be taking away your choice when he chooses the car (forcing you to lose) and/or revealing the correct answer --and letting you switch guaranteeing you'll win. This will happen 1/3 of the time. If he chooses the goat (2/3 of the time)-- then you have information that tells you that 1 of the two options left is the car... that is the 50-50 thing. It would be expected to happen 2/3 of the time. See?

Knowing Monty's motives only tells you 2 things-- what are your odds of being offered a choice... and what are your odds that you increase your odds by switching. There's 2 factors... not one. They are co-dependent. In the scenario it's stipulated that he always shows a goat and you always get a choice. In that case, switching doubles your odds.

I think everyone can understand that if you keep your first choice every time no matter what you know about Monty's motives or whether you are offered a chance to change or not-- you will have 1 in 3 chance of winning a car (good odds). The question then becomes is there any way to increase your odds of winning the car. The only way to increase your odds is to change if given the choice. If you are always given a choice (even when you picked a goat or the host reveals a car) then you win 2/3 of the time by changing. No matter what the motives of the host.

If you aren't always given a choice, then it's a different game. But your odds don't increase from the 1 in 3 regarding the original choice. The only way you can increase those odds is by changing if offered the choice. Changing doesn't always increase your odds-- but if the host always shows a goat and always offers a switch, you will always increase your odds by changing. Increasing your odds does not mean that you are guaranteed to win... it just means that doing this consistently or amongst many trials or huge numbers of people-- you will win twice as much as if you don't switch.

I think that should suffice. This is true--basic probabilities. Ask any expert on the subject or run the figures yourself.
 
Last edited:
No, if Monty always reveals a goat and always gives you a chance to switch then your double your odds by switching from the 1/3 you had originally to 2/3.

That is the standard version of the problem.

The odds are not 50/50 despite the fact that there are two choices.

They are 50/50 if Monty chooses randomly AND has revealed a goat, because his randomly choosing a goat happens 2/3 of the time, and your 1/3 is half of that.
 
Last edited:
Maybe I'm misunderstanding someone's position here, but this is the way this dispute appears to me:
Articulett: chance of winning via sticking with the first choice no matter what happens is just P(first pick is car).
Others: incorporating the added information once an event has happened gives the conditional probability P(first pick is car|info), which is not equal to P(first pick is car)
I don't understand why there is such a case of miscommunication. Those two claims are perfectly consistent.

You're absolutely right. For me, the Monty Hall problem is clearly asking a question about the chances of winning at the point where Monty offers the switch. What I find interesting is that the distribution of probabilities between the two remaining doors does not depend solely on the condition that Monty revealed a goat, it also depends on how he came to the decision of which door to open.
 
Last edited:

Back
Top Bottom