Reasoning under uncertainty #9 - The Monty Hall problem

This presentation of the Monty Hall problem will focus on the problem as it's originally and often stated. The setup presented will turn out to be insufficient to uniquely define the data probabilities. Thus personal interpretation and probabilities would be necessary for a complete evaluation. This presentation will not go that far, but will present some archetypal Monty Hall strategies.

I refer to two video's in this presentation, namely websnarf's video on the subject:
http://www.youtube.com/watch?v=o2L_2psS9uI
and Prepoceros' video:
http://www.youtube.com/watch?v=1uS4Sa3nIFc

A more full presentation can be found by combining the full manuscript:
http://folk.uio.no/trondr/uncert9_manuscript.pdf
and the slides:
http://folk.uio.no/trondr/uncert9_slides.pdf

Category:

Howto & Style

Tags:

• Bayesian
• Bayes
• data
• probability
• reasoning
• model
• Monty
• Hall

[BrunoTheQuestionable]

Most strange - I almost suggested that you tackle the Monty Hall problem, when I watched the Monty Hall videos a few weeks ago. Its another useful example of updating probability based on new data.

[trondreitan]

Thanks. I almost skipped this video, since as it's usually stated, the treatment must be much to involved to give a simple glimpse into probabilistic reasoning. It's very far from the easiest example. When it's state properly the principles handles so far quickly yield results, though. And these results are so that some people might find counter-intuitive.

[giovea]

or to "holes " alike , hehehe

[giovea]

THE PARADOX- THERE ARE NOT 2 GOATS THE SAME, hihihi

[giovea]

anyway make it a white and black goat not get confuse and unfold it again. you will see post-test and pre test chance beeing the same at the point of reconsidering. that was the gap. iluusion of 2 goats, being the same

[giovea]

Errors:

#1- how the number relates to prize is irrelevant and you dont know and can't know!

#2-if showman knows and always opens a goat door it's also irrelevant too (it's his concern and the branches possibilities remain secret to you).

SO If you left with 2 doors and 1 shot is simple 1/2. Rest is pos-test probability of game developing, not contestant options in the time of switching.

THat's it