AMS 507

Lecture 5

Chapter Three
Conditional Probability and Independence
3.2 Conditional Probabilities

Definition: If P( F ) > 0, the conditional probability of E given F, denoted by P(E|F), is

P( EF )
P( E | F ) = .
P( F )
This definition satisfies the axioms of probability theory.

Example 1. From the set of all families with two children, a family is selected at random
and found to have a girl. What is the probability that the other child of the family is a
girl? Assume that in a two-child family all sex distributions are equally probable.
Answer: 1/3.

Example 2. From the set of all families with two children, a child is selected at random
and is found to be a girl. What is the probability that the second child of this girl’s family
is also a girl? Assume that in a two-child family all sex distributions are equally probably.
Answer: ½.

Law of Multiplication: P( EF ) = P( F ) P( E | F ).

The (General) Multiplication Rule:

If P( E1 E 2  E N ) > 0, then
P( E1 E 2 E3  E n −1 E n ) = P( E1 ) P( E 2 | E1 ) P( E3 | E 2 E1 )  P( E n | E1 E 2  E n −1 ).

Example 4 (2g in text). An ordinary deck of 52 playing cards is randomly divided into 4
piles of 13 cards each. Compute the probability that each pile has exactly one ace.

Example: The gambler’s ruin problem.

3.3. Bayes’ Formula

Law of Total Probability

Let F be an event with P(F)>0 and P(Fc)>0. Then for any event E,
P( E ) = P( E | F ) P( F ) + P( E | F c ) P( F c ).
Definition: Let {F1 , F2 ,  , Fn } be a set of nonempty subsets of the sample space S of an
n
experiment. If the events F1 , F2 ,  , Fn are mutually exclusive and  i =1
Fi = S , the set
{F1 , F2 ,  , Fn } is called a partition of S.

Generalized Law of Total Probability

If {F1 , F2 ,  , Fn } is a partition of the sample space of an experiment and P(Fi)>0 for
i = 1,2,  , n , then for any event E of S,
n
P( E ) = ∑ P( E | Fi ) P( Fi ).
i =1

Bayes’ Theorem

Let {F1 , F2 ,  , Fn } be a partition of the sample space S of an experiment. If

for i = 1, 2, , n, P( Fi ) > 0, then for any event E of S with P(E)>0,
P( E | F j ) P( F j )
P( F j | E ) = .
P( E | F1 ) P( F1 ) + P( E | F2 ) P( F2 ) + + P( E | Fn ) P( Fn )

Example 3c. In answering a question on a multiple-choice test, a student either knows the
answer or guesses. Let p be the probability that the student knows the answer and 1-p the
probability that the student guesses. Assume that a student who guesses at the answer will
be correct with probability 1/m, where m is the number of multiple-choice alternatives.
What is the conditional probability that a student knew the answer to a question, given
that he or she answered it correctly?

Example 3g. In the world bridge championships held in Buenos Aires in May 1965 the
famous British bridge partnership of Terrence Reese and Boris Schapiro was accused of
cheating by using a system of finger signals that could indicate the number of hearts held
by the players. Reese and Schapiro denied the accusation, and eventually a hearing was
held by the British bridge league. The hearing was in the form of a legal proceeding with
a prosecuting and defense team, both having the power to call and cross-examine
witnesses. During the course of these proceedings the prosecutor examined specific hands
played by Reese and Schapiro and claimed that their playing in these hands was
consistent with the hypothesis that they were guilty of having illicit knowledge of the
heart suit. At this point, the defense attorney pointed out that their playoff these hands
was also perfectly consistent with their standard line of play. However, the prosecution
then argues that as long as their play was consistent with the hypothesis of guilt, then it
must be counted as evidence toward this hypothesis. How can Bayes’ theorem be used to
assist in judging this question?