Let O be the event that Monty Hall opens door b Now the question is Is PAO PCO

Let o be the event that monty hall opens door b now

This preview shows page 25 - 35 out of 38 pages.

Let O be the event that Monty Hall opens door b. Now the question is: Is P(A|O) ≤ P(C|O)?
Image of page 25

Subscribe to view the full document.

The Monty Hall problem (5) According to Bayes’ rule (definition of conditional probability), it holds that: 𝑃𝑃 𝐴𝐴 𝑂𝑂 = 𝑃𝑃 𝑂𝑂 𝐴𝐴 ∗ 𝑃𝑃 ( 𝐴𝐴 ) 𝑃𝑃 ( 𝑂𝑂 ) 𝑃𝑃 𝐶𝐶 𝑂𝑂 = 𝑃𝑃 ( 𝑂𝑂𝑂𝐶𝐶 ) ∗ 𝑃𝑃 ( 𝐶𝐶 ) 𝑃𝑃 ( 𝑂𝑂 ) We know that P(A)=P(B)=P(C)=1/3 Hence, all we need to know are the (conditional) probabilities for the event O.
Image of page 26
The Monty Hall problem (6) What are the conditional probabilities for the event O (i.e., the event that Monty will open door b)? Remember 1: we assume we have picked door a. P(O|A)=1/2 (if the car is behind a, then Monty opens b or c.) P(O|B)= 0 (if the car is behind b, then Monty never opens b.) P(O|C)=1 (if the car is behind c, then Monty will open b.) What is left is the unconditional probability P(O).
Image of page 27

Subscribe to view the full document.

The Monty Hall problem (7) To calculate P(O), note that the prize is only behind one door (i.e., A, B, C are disjoint events), hence: P(O)= P(O and A) + P(O and B) + P(O and C) The general multiplication rule gives: P(O)=P(A)*P(O|A)+P(B)*P(O|B)+P(C)*P(O|C) P(O)=1/3 * ½ +1/3* 0 +1/3* 1 P(O)=1/2
Image of page 28
The Monty Hall Problem (8) Insert the numbers into the definition of conditional probability: 𝑃𝑃 𝐴𝐴 𝑂𝑂 = 𝑃𝑃 𝑂𝑂 𝐴𝐴 ∗ 𝑃𝑃 ( 𝐴𝐴 ) 𝑃𝑃 ( 𝑂𝑂 ) = 1 3 1 2 1 2 = 1/3 𝑃𝑃 𝐶𝐶 𝑂𝑂 = 𝑃𝑃 ( 𝑂𝑂𝑂𝐶𝐶 ) ∗ 𝑃𝑃 ( 𝐶𝐶 ) 𝑃𝑃 ( 𝑂𝑂 ) = 1 3 1 1 2 = 2 3 That’s it! So, when you have picked door #1, always switch to door #3 when door #2 is opened.
Image of page 29

Subscribe to view the full document.

Back to last week: MC-answer guesses Last week, we have guessed MC- answers. By calculating the sample space and calculating the number of possibilities for a certain outcome we were (in principle) able to calculate the probability of an event. E.g. S=4^8=65536 P(X=8) = 1/65536 The probabilities of these events can be calculated in a more efficient way using the binomial setting.
Image of page 30
Binomial probabilities (1) Binomial distributions are models for some categorical variables, typically representing the number of successes in a series of n trials. The observations must meet these requirements: The total number of observations n is fixed in advance. The outcomes of all n observations are statistically independent. Each observation falls into just one of two categories: success and failure. All n observations have the same probability of success, p .
Image of page 31

Subscribe to view the full document.

Binomial probabilities (2) The distribution of the count X of successes in the binomial setting is the binomial distribution with parameters n and p: B ( n,p ). The parameter n is the total number of observations. The parameter p is the probability of success on each observation. The count of successes X can be any whole number between 0 and n .
Image of page 32
MC-answer guesses (once more) Binomial probabilities can be calculated as follows: So, in the coin toss example of last week: k n k p p n k k X P = = ) 1 ( ) ( 31 . 0 75 . 0 * 25 . 0 * 2 8 ) 2 ( 6 2 = = = X P X 0 1 2 3 4 5 6 7 8 P(X) 0.10 0.27 0.31 0.21 0.09 0.02 0.003 0.0003 0.00002
Image of page 33

Subscribe to view the full document.