home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.puzzles
- Path: sparky!uunet!spool.mu.edu!news.nd.edu!mentor.cc.purdue.edu!seaman.cc.purdue.edu!ags
- From: ags@seaman.cc.purdue.edu (Dave Seaman)
- Subject: Re: Revised Monty Hall problem
- Message-ID: <C07rIx.C42@mentor.cc.purdue.edu>
- Sender: news@mentor.cc.purdue.edu (USENET News)
- Organization: Purdue University
- References: <1i01n8INN1bb@iskut.ucs.ubc.ca>
- Date: Sat, 2 Jan 1993 06:32:57 GMT
- Lines: 59
-
- In article <1i01n8INN1bb@iskut.ucs.ubc.ca> leon@unixg.ubc.ca (Leon ter Beek) writes:
- >There are five doors, behind two of them a car, and behind each of the
- >other three doors a goat. You pick a door and the quizmaster (who knows
- >what's behind the doors) opens a door with a goat behind it.
- >1) Should you change doors?
-
- I assume Monty always opens an unchosen door and reveals a goat, as in the
- original puzzle.
-
- Without switching, you have a 2/5 chance of winning a car.
- If you switch, and if your original choice was a car, then only one of the
- three doors that you could switch to has a car behind it. If your original
- choice was a goat, then your probability of winning by switching rises to 2/3.
- Since your first choice is a car 2/5 of the time, your probability of winning a
- car by using the always-switch strategy is 2/5 * 1/3 + 3/5 * 2/3 = 8/15, which
- is greater than 2/5. Therefore, you should switch.
-
- >2) Why is the sum of the probability of changing and not changing not
- >equal to one?
-
- Because the probability of winning without switching is 2/5, and
- the probability of winning by switching is 8/15. The sum of these is 14/15, and
- 14/15 is not one. That is why the sum of the probabilities is not one.
-
- Oh, I see. You think this is a contradiction, and you want an explanation of
- why it isn't.
-
- It's like the famous bellhop problem that concludes by asking where the other
- dollar went. You wave your hands and add up some probabilities, and then you
- ask why the sum is not 1.
-
- If you add up the probabilities of disjoint events that cover (allmost all of)
- the space, then you will get one. In the three-door problem, you have
-
- probability of winning by switching + probability of losing by
- switching = 1 [Rule 1]
-
- and also, because the two events are identical, you have
-
- probability of losing by switching = probability of winning by not
- switching [Rule 2]
-
- from which it follows that
-
- probability of winning by switching + probability of winning by not
- switching = 1. [Rule 3]
-
- In the five-door game, Rule 1 continues to hold. However, Rule 2
- is false. It is possible for your original choice to be a car, and
- for you to switch and wind up with the other car. It is also possible
- for your first choice to be a goat, and for you to switch and wind
- up with another goat. Rule 2 fails because the two events described
- are not identical. Since Rule 3 was derived from Rules 1 and 2,
- there is no reason to expect Rule 3 to hold for the five-door
- problem. Hence, no contradiction.
-
- --
- Dave Seaman
- ags@seaman.cc.purdue.edu
-