Showing posts with label Monty Hall problem. Show all posts
Showing posts with label Monty Hall problem. Show all posts

Monday, October 28, 2019

The Monty Hall Puzzle… when it would be best NOT to switch



                     


Start the week with a little 'reverse' puzzle-thinking... There are lots of variations of the Monty Hall “paradox,” but as probably all readers here know in its standard form it is statistically better (indeed twice as good) for the contestant to ‘switch’ doors in the usual game as to stay with their initially chosen door. But one can make a small change in the set-up to make it better for the contestant to stick with their chosen door. 

Instead of the usual premise of say a car behind one door, and billy goats behind the other two, consider the case of a goat behind one door and cars behind the other two (continue to assume the contestant prefers a car... even though billy goats can be downright appealing… and host Monty still knows what’s behind all 3 doors).

Now, contestant Sheldon picks say Door #1, and Monty in turn opens Door #3 revealing a car… should Sheldon switch to Door #2 or remain with #1 (or, does it make any difference)?
The quick math shows that at the outset Sheldon had a 2/3 chance of selecting a car by random chance, better than the chance that either one of the remaining doors hide a car, and thus Sheldon is better off staying with his 2/3 first pick in this instance (essentially, exactly the reverse argument of the normal set-up).
(Though, IF, as a prior, we know that Sheldon is a telepathic extraterrestrial, or one with X-ray vision, than that could alter the odds considerably…;)




Wednesday, October 1, 2014

Bayes-mania


Seeing an awful lot written on Bayesian ideas in the last year (and week!).
Jason Rosenhouse uses the Monty Hall problem and a NY Times article as a launching point for a discussion of the subtlety of Bayes here:

http://scienceblogs.com/evolutionblog/2014/10/01/a-poor-description-of-the-monty-hall-problem/

Rosenhouse takes the Times' article to task, and ends simply with:
"Applying statistics correctly is hard, even for people with professional training in the subject. But the problems are found in the complexity of real-life situations, and not in the underlying philosophical approaches to probability and statistics."
(Rosenhouse's 2009 book on the The Monty Hall problem is great, by the way, if you've never read it -- yes, an entire volume on that one problem)

Meanwhile, The Guardian also presented Bayesian statistics this week: http://tinyurl.com/ojjazs4

And in other statistics briefs, Jeff Leek tries to come to the defense of much-maligned p-values here:
http://simplystatistics.org/2014/09/30/you-think-p-values-are-bad-i-say-show-me-the-data/ 

Thursday, August 4, 2011

Goat, Goat, Car...

Sol Lederman is encouraging all to visit Jeremy Jones' new site explaining the Monty Hall problem. Even if you know this problem (and solution) well, it is entertaining to see how Jeremy has pieced it all together (be sure to check out all 3 sections -- play/explanation/history -- of his site):

http://www.stayorswitch.com/

While I'm on the subject I may as well again direct any who are truly enamored (obsessed?) of this classic problem to be sure and read Jason Rosenhouse's "The Monty Hall Problem" which covers it in all its nuances and variations.

Sunday, June 26, 2011

Revisiting the Most Re-Revisited Problem in Popular Mathematics... Again

A recent 14-minute podcast from Scientific American on... guess what... the Monty Hall Problem:

http://www.scientificamerican.com/podcast/episode.cfm?id=lets-make-a-probabilistic-deal-a-fr-11-06-25

Nothing new here, but if someone prefers a verbal/vocal description of the puzzle over a written description this podcast may fill the bill.

Jason Rosenhouse's book-length treatment of the puzzle, "The Monty Hall Problem," remains a great (and thorough) source of information on this classic.

Wednesday, June 2, 2010

Monty Hall Redux

Jason Rosenhouse's delectable volume devoted entirely to the 'Monty Hall problem':

http://tinyurl.com/y927aou

(I think the first three, and last three chapters are especially good.)