How smart are you 4

The explanation in the video makes your head spin. And the explanation carter80 quotes makes you head spin right off. Here's the simple explanation.

Say you play the game 99 times.
We will call the door you initially chose: your-door. (It neednt always be the same door.)
We will call the door the host opens: host-door. (Again, not always the same door.)
We will call the last door: last-door.
Obviously, the car is behind your-door 33 times out of the 99. So if you stay with your-door, you win the car 33 times - and lose 66 times.
Now, the car is never behind host-door. So, whenever the car is not behind your-door, it must be behind last-door. In other words, the car is behind your-door 33 times, and behind last-door 66 times.
So switch, baby, switch.
 
The key is that Monty ALWAYS open a door that contains nothing, and he knows which door has the car. His action is therefore informative and alters the probabilities.

If you pick door A and Monty opens door B the odds are as follows:

1/3rd of the time the car is behind door A, and Monty randomly picks door B or C, 50% chance that he will pick either one. Monty is telling you nothing in this case.

2/3rds of the time the car is behind door B or C, and Monty has NO CHOICE in what door he opens, by opening B he is, 2/3rds of the time, telling you the car is behind C.

Thus in 2/3 cases you win by switching.

This is entirely counter-intuitive but it is nevertheless correct.
 
2/3rds of the time the car is behind door B or C, and Monty has NO CHOICE in what door he opens, by opening B he is, 2/3rds of the time, telling you the car is behind C.

Thus in 2/3 cases you win by switching.

This is entirely counter-intuitive but it is nevertheless correct.
That is incorrect.

Why does the total 1/3 probability of door B get assigned to door C?

There is also a 2/3 probability the prize is behind doors A and C as there is for doors B and C. Hence, once B is eliminated, as spheroid said, the 33% chance of door B gets assigned equally to doors A and C, i.e. 50% each.
 
That is incorrect.

Why does the total 1/3 probability of door B get assigned to door C?

There is also a 2/3 probability the prize is behind doors A and C as there is for doors B and C. Hence, once B is eliminated, as spheroid said, the 33% chance of door B gets assigned equally to doors A and C, i.e. 50% each.

You are ignoring the new information given to you when Monty opens a door. The rules of the game are that Monty ONLY opens a door with no car behind it, it is not random. There is not a 33% chance that Monty will reveal the car when he opens his door, he knows where the car is, and he never does that, therefore his pick contains information.

In the case where the door you picked does not have the car Monty has no choice about what door he opens, one of the other two doors has the car and he is not allowed to open that door. Therefore he is giving you new information in 2 of 3 cases.

When you pick door A you have a 1 in 3 chance of getting the right door.

1 chance in 3 that you got it right, 2 chances in 3 that you got it wrong.

In the 1 in 3 case the odds that Monty picks door B are 50/50, as both B and C are empty.

In the 2 in 3 case the odds that Monty picks door B are 100% as he has no choice, he HAS to pick the empty door. Alternately, if the car is behind door B, he has no choice but to open door C. Either way, there is a 2 in 3 chance that you failed to pick the right door on your first try, and in that case Monty tells you where the car is.

So 1 in 3 times your first pick is the correct answer and his answer is random, and 2 in 3 times your pick was wrong and he is telling you where the car is. You are better off taking the 2 in 3 odds of being right than in taking the 1 in 3 odds of being right, so it is better to switch.

To be fair in post #1 YF did not emphasize enough that Monty only ever opens a door that has no car behind it, but that is how his show on TV actually worked, and a critical part of the famous "Monty Hall problem" in statistics, which is what this thread is about.
 
You are ignoring the new information given to you when Monty opens a door. The rules of the game are that Monty ONLY opens a door with no car behind it, it is not random. There is not a 33% chance that Monty will reveal the car when he opens his door, he knows where the car is, and he never does that, therefore his pick contains information.
I disagree.

No matter what door you pick, Monty will always be able to open one door that does not contain the car. Thus, when he does so, he has not given you any useful information at all, sine you already knew in advance that he would do so.
 
The probilities are not really linked. They are discrete events. If I flip a coin ten times in a row and get heads every time there is no greater chance on the eleventh throw that I'll get a tail on that eleventh flip. It is true that ON AVERAGE after many, many tosses I'll get a fifty-fifty mix but for any individual flip there is no force acting on the coin to make it any particular result.
 
that is the reason we pay for this hobby we want a sure thing and not have to play the odds.
 
The key is that Monty ALWAYS open a door that contains nothing, and he knows which door has the car. His action is therefore informative and alters the probabilities.

If you pick door A and Monty opens door B the odds are as follows:

1/3rd of the time the car is behind door A, and Monty randomly picks door B or C, 50% chance that he will pick either one. Monty is telling you nothing in this case.

2/3rds of the time the car is behind door B or C, and Monty has NO CHOICE in what door he opens, by opening B he is, 2/3rds of the time, telling you the car is behind C.

Thus in 2/3 cases you win by switching.

This is entirely counter-intuitive but it is nevertheless correct.
bingo
swapping allows you to make a hit in two out of three cases because the host helps you to do so
video clearly explains it

flipping a coin has nothing to do with it
 
Fuji is correct, you should always switch. If the host RANDOMLY opened one door, that would be a different matter. This is an old mathematical puzzle.

Here is the DEFINING explanation:


There are nine and only nine possibilities, you choose A - then 3 possible winners, you choose B, then 3,,,etc. ...but lets look at just 3:

Prize ///////////// Your Choice /////////////// What will host uncover? ////////////// You WILL switch to a

A ************** A **************** B or C (both losers, makes no difference) ***** loser
A ********** B **************** C (remember, he will ONLY uncover a loser) winner
A ************ C ****************** B ****** " ************* winner

you can repeat for if B or c is the winner, but you get the point. Only if you chose correctly ooriginally will it be wrong to change, but you only get it right originally one third of the time.

Try it yourself, using a die, where , say, 1 and 2 represents A, 3 and 4 represents B etc.
 
I disagree.

No matter what door you pick, Monty will always be able to open one door that does not contain the car. Thus, when he does so, he has not given you any useful information at all, sine you already knew in advance that he would do so.

No new information? Oh REALLY? lol. Suppose then that there are two of you lined up and the host is going to tell you which of you is about to be executed. I don't know about you, but I would find that to be VERY new information.

Knowing that information is about to be DIVULGED is most assuredly NOT the same as knowing what that information ACTUALLY IS! PU LEAZE!

Like I said, its very easy to do an experiment your self. Hell, if I had some way of posting an Excel spreadsheet, with a few simple macros I could make an Excel version of it to prove my point.
 
There was an old thread on this one and I was just as confused as others. Think on a more macro scale. The effect is counterintuitive because you are using the worst case.

You buy a lottery ticket.....all other "winning combinations" after the draw are eliminated except one. Would you go for that ticket. Your odds with your ticket are the original 14 billion to one, the odds with the other ticket are slightly higher.... :)
 
There was an old thread on this one and I was just as confused as others. Think on a more macro scale. The effect is counterintuitive because you are using the worst case.

You buy a lottery ticket.....all other "winning combinations" after the draw are eliminated except one. Would you go for that ticket. Your odds with your ticket are the original 14 billion to one, the odds with the other ticket are slightly higher....

this was good, but wait, somebody is gonna dispute this too...ok let me try... well you already knew that there this whole bunch of loser tickets out there, so eliminating them does not help you! how is that?:)
 
No matter what door you pick, Monty will always be able to open one door that does not contain the car. Thus, when he does so, he has not given you any useful information at all, sine you already knew in advance that he would do so.

You knew in advance that he would do so, but you did not know which door it was going to be.

There is a 1 in 3 chance the car is behind the door you picked, and a 2 in 3 chance it is behind one of the other doors. By switching you get access to the full 2 in 3 odds as he eliminated one of the two for you.

Maybe it will help you to think of switching as getting two pick two doors instead of only one as Monty tosses out the losing door for you.
 
So let me get this straight....... Me and Fuji are on the stage.... Fuji picks door A, and I pick door B..... Monty reveals to each of us that Door c is empty and offers booth of us a switch.... Who should switch?


The point here is obvious.... you now have a new game entirely. It started over. He's asking you to choose A or B. 50/50

here.... I'll prove it mathimatically. there is one car. there are 2 doors.

1 divided by 1 = .5

your odds of getting this correct is now half.

Cheers! :D
 
Poker, go through my exercise with a die, it will soon become obvious that switching is much better.
 
So let me get this straight....... Me and Fuji are on the stage.... Fuji picks door A, and I pick door B..... Monty reveals to each of us that Door c is empty and offers booth of us a switch.... Who should switch?

What happens if the car is behind door C? In your example Monty decides nothing and offers no information, it's possible the door he opens will have the car, thus the odds are indeed 50/50. However in the situation discussed on the thread Monty decides which door to open based on knowing where the car is. By doing so he is essentially offering you a chance to pick two doors in one.
 
Quite right Fuji, poker ignores the fact that his example is NOT the same. The original setup was that Monty AWLAYS OPENS A NON_WINNING DOOR. This 'winnowing' effect is what increases the odds. But in Poker's example, if you follow those SAME rules, then Monty won't be able to open the door when neither has picked a winner.
 
Here's my entry for the shortest-possible, completely-convincing, way of putting the argument.

Play the game 99 times, under the conditions as presented. On the average, over the 99 games, the prize is behind your door 33 times, is behind the host door zero times, and is behind the last door 66 times. So switch.
 
Your logic escapes me, is it not the same odds if he does not switch ?

Anyways, you are incorrect
If you use Bayes’s rule to update, you should change. Mathematically, it is indisputable. However, it is one of those things that I personally have never found intuitive even though I have learnt (and even taught) Bayes rule. Statisticians are forever trying to come up with new ways of explaining why this unintuitive result is intuitive.
There are also a couple of other interesting paradoxes of a similar nature (but have little to do with Bayes rule). These include the Allais paradox and Ellsberg paradox.
They are all interesting because the fact they are not intuitive to most people (including many who understand the math) raises questions regarding whether or not people really are rational when they make decisions under conditions of uncertainty.