Re^8: Marilyn Vos Savant's Monty Hall problem (common sense)

Although your three scenarios may be pedantically valid, only the first one makes any real sense.

If there are goats behind two of the doors, then Monty opening the door with the car behind it would make for a really stupid game show. He might as well just open the door you chose since both actions tell you that you picked a goat door.

I suppose you could argue that the problem as stated doesn't make it absolutely clear that you *know* that there are two goat doors (or that there are *always* two goat doors).

But as stated, the only reasonable interpretation is the first one (that makes sense for a game show, that fits common sense, not that fits a mathematician's standard for precision).

As for the real Monty saying that he sometimes opened the "car" door, this is true. And that didn't make the real game show really stupid (yes, I'm sure many regarded the show as plenty stupid, but I'm talking stupid beyond sense, not just a matter of taste) because the real show always (or almost always, I haven't personally reviewed every single episode) had one "goat" door (a joke "prize", and no, it wasn't always a goat), one "car" door (a major prize), and one "appliance" door (a minor prize).

That makes for a more complex statistics problem (as it should, being a real game show). But it has little more to do with the problem as stated other than being the likely inspiration for it and, later, the inspiration for its retitling as well.

Also note that Monty had other options besides opening one of the doors. He could have a prize wheeled out and offer you that in exchange for the door. He often pulled cash out of his pocket and offered you that (perhaps in addition to the prize he just wheeled out). It wouldn't have made much sense to call the show "Let's make a deal" without such negotiations. (:

- tye

Comment on Re^8: Marilyn Vos Savant's Monty Hall problem (common sense)

Replies are listed 'Best First'.
Re^9: Marilyn Vos Savant's Monty Hall problem (common sense) by tilly (Archbishop) on Aug 24, 2004 at 15:35 UTC
I didn't model it that way, but who is to say (given the problem statement) that Monty wouldn't choose to open your door first if you didn't choose the car? In that case the analysis comes out to what I showed. Indeed in the slight variation of this game known as the shell game, it is indeed traditional to show people right off the bat that they got it wrong. (It is also traditional to remove/re-add the pea by sleight of hand so that the chump playing always loses..but let's ignore that fact.) Indeed that precedent suggests that there might indeed be room for a game show where the host tries to make life hard for the players. Indeed the existence of reality shows suggests that the traditional game show format is not the only possible one with an audience. Furthermore even if you do (based on cultural knowledge) find only one possible assumption natural, it is still good to analyze the problem clearly enough to know that that is an assumption. Because while implicit assumptions often are right, every so often you'll run across a problem that you can only get straight by being able to find and correct a wrong assumption that you made. (In my experience, "every so often" comes by pretty frequently...)	[reply]

Replies are listed 'Best First'.

Re^9: Marilyn Vos Savant's Monty Hall problem (common sense)
by tilly (Archbishop) on Aug 24, 2004 at 15:35 UTC

wouldn't

Indeed in the slight variation of this game known as the shell game, it is indeed traditional to show people right off the bat that they got it wrong. (It is also traditional to remove/re-add the pea by sleight of hand so that the chump playing always loses..but let's ignore that fact.) Indeed that precedent suggests that there might indeed be room for a game show where the host tries to make life hard for the players. Indeed the existence of reality shows suggests that the traditional game show format is not the only possible one with an audience.

Furthermore even if you do (based on cultural knowledge) find only one possible assumption natural, it is still good to analyze the problem clearly enough to know that that is an assumption. Because while implicit assumptions often are right, every so often you'll run across a problem that you can only get straight by being able to find and correct a wrong assumption that you made. (In my experience, "every so often" comes by pretty frequently...)

[reply]