47
Some Nomenclature
A Partition of the Sample Space: A1, A2, …, An
• P(Ai): “Prior Probability” of Ai
the unconditional probability of Ai
• B: Some event
• P(Ai|B): “Posterior Probability” of Ai
Bayesian Update of probability of Ai
48
Example: Which Box?
From Textbook: Pg. 47-50, Examples 1 and 2
• Consider the following three boxes.
• I mix the boxes, and pick one at random. Then I pick a
ball at random from the chosen box and show to you. You
now have to guess the box.
a) Which box should you choose if the ball I show you is
white?
49
Which box? (Continued)
b) Suppose you know that I pick a box each time at random
and suppose you adopt the strategy of guessing the box
with highest posterior probability given the observed colour,
as in part (a).
About what proportion of the time do you expect to be right
over the long run?
50
Which box? (Continued, modified from
Ex. 7, Pg. 55)
c) Suppose you use the guessing strategy found in b), but I
was in fact randomising the choice of the box each time
with probabilities (1/4,1/2,1/4), instead of (1/3,1/3,1/3). Now
how would your strategy perform over the long run?
d) Suppose you knew that I was either randomising with
probabilities (1/3,1/3,1/3) or with probabilities (1/4,1/2,1/4).
How could you learn which I was doing? How should you
respond, and how would your response perform over the
long run?
Now try Exercise 8, Pg. 55 to find out about the optimal
situation. (Homework)
51
Independence
• Two events A and B are called independent, if
P(A | B) P(A | B) P(A)
or equivalently P(B | A) P(B | A) P(B).
• Notice that all of the above reduce to one simple formula
P(A B) P(A)P(B)
52
Examples
• Randomly drawing from a pack of cards, getting a king is
independent of getting a spade.
Notes
1. Mutually exclusive (non-null) events are never
independent.
2. (Non-null) Sub-events are never independent of the
corresponding super-events, and vice-versa
53
Example (Review Exercise #4, text)
There are two boxes.
• Box 1 contains 2 red balls and 3 black balls.
• Box 2 contains 8 red balls and 12 black balls.
One of the two boxes is picked at random, and then a ball
is picked at random from the box.
a) Is the colour of the ball independent of which box is
chosen?
b) What if there were 10 black balls rather than 12 in Box 2,
but the other numbers were the same?
54
Mutual Independence versus Pairwise
Independence
Example (#18, Problem Set)
Two six faced fair dice are thrown. Define events
• A = Die 1 shows six,
• B = Die 2 shows six and
• C = Both dice show the same face.
Check: A, B, C are pairwise independent but not mutually
independent.
Qn. Which conditions ensure independence of A, B and C,
and why?
55
Condition for Independence
• For three events A, B and C, in addition to pairwise
independence, we also need
P(A | BC) P(A | BC) P(A | BC) P(A | BC) P(A)
or a similar description for B or C.
• Together with pairwise independence, all of the above
conditions reduce to one single additional condition:
P(A B C) P(A)P(B)P(C)
• I leave the general case to you as an exercise.
56
Example (#19, Problem Set)
57
Example: #22, Problem Set
• The valves in the image below develop faults and stop
working independently with probability p each.
a) What is the probability that gas is flowing from A to B?
b) Given that gas is flowing from A to B, what is the
probability that V1 has developed a fault?
c) Given that gas is not flowing from A to B, what is the
probability that V1 has not developed a fault?
