Additional Points From Last Class
In the deck of cards example, the experimental set up allowed
us to observe N(A) and n, and since they are equiprobable
N(A)
events, we can apply the formula P(A) = n . However, the mail
delivery example, P(A) = 0.6, hence not equiprobable events.
Interpreting N(A)/ n as a relative frequency holds for both
equiprobable events, and non-equiprobable events.
The axioms do NOT completely determine an assignment of
probabilities to events, they only serve to rule out assignments
that are inconsistent with out intuitive notion of probability.
C. Gravel (Carleton)
STAT 3502
January 2015
1 / 21
Chapter 2 - Probability
2.3 Counting Techniques
2.3 Counting Techniques
Recall...
For an experiment with equally likely events, we can reduce the task
of computing probabilities to counting.
P(A) =
N(A)
n
The probability of of compound event A is the number of outcomes
contained in event A, N(A), divided by the total number of outcomes
in the sample space, n.
If we cannot list all outcomes easily due to an extremely large
number of possible outcomes, n, we can exploit general counting
rules to compute probabilities.
In this section, we will derive and apply these rules.
C. Gravel (Carleton)
STAT 3502
January 2015
2 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 2
Ordered Pairs
Consider pairs of events such that if O1 and O2 are objects, then the
pair (O1 , O2 ) is different from the pair (O2 , O1 ). Pairs of this nature
can be called, ordered pairs.
Examples
Imagine you want to go see a movie and have dinner with a
friend. There are two movie theatres in town and each is close
to two restaurants.
If we denote the movie theatres as M , = 1, 2, and the
restaurants as Rj , j = 1, 2, 3, 4, then the ordered pairs are (M , Rj )
where 6= j and M and Rj are in the same area of town.
Additional example: 2.18, p. 65
C. Gravel (Carleton)
STAT 3502
January 2015
3 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 3
Proposition: Product Rule for Ordered Pairs (p.65)
If the first element or object of an ordered pair can be selected in n1
ways, and for each of these n1 ways the second element of the pair
can be selected in n2 ways, then the number of pairs is n1 n2 .
Another way to think of this is,
Consider a 2 stage process - the first stage must be complete
for the second stage to be performed.
There are n1 ways to perform the 1st process.
There are n2 ways to perform the 2nd process for each of the 1st
processes n1 ways.
Thus, there are n1 n2 ways to perform the process in sequence.
Note: This can rule be generalized to k items, see p. 66.
C. Gravel (Carleton)
STAT 3502
January 2015
4 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 3a
What if we chose to eat before the movie?
As long as there are the same number of choices of the second
element for each of the first elements, the product rule is valid.
Note, that this example can be considered in both orders, that may
note be the case. For example, a process that occurs in stages in
which the second stage depends on the results of the first stage.
C. Gravel (Carleton)
STAT 3502
January 2015
5 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 4
C. Gravel (Carleton)
STAT 3502
January 2015
6 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 4
C. Gravel (Carleton)
STAT 3502
January 2015
7 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 4
Consider a group of n distinct objects. How many ways are there to
select a subset of size k from the group?
Example
A businessman wishes to open a second franchise in Ottawa and the
owner wants to transfer 3 of his existing 10 staff members to the
new store. How many different ways can we select these 3
individuals? Note: Assume they are all of equal rank in the old store.
Does order matter?
C. Gravel (Carleton)
STAT 3502
January 2015
8 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 5
Definition: Permutations and Combinations
A permutation is an ordered subset.
A combination is an unordered subset.
If the 3 employees will be ranked manager, assistant manager and
cashier in the new store, then order matters and we are interested
in the number of permutations.
If not, we are interested in the number of combinations.
C. Gravel (Carleton)
STAT 3502
January 2015
9 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 6
Counting Permutations
The number of permutations of size k that can be formed from n
objects in a group is,
Pk,n =
n!
(n k)!
= n(n 1)(n 2) ... (n (k 2))(n (k 1)).
Counting Combinations
The number of combinations of size k that can be formed from n
objects in a group is,
n
n!
Ck,n =
=
.
k
k!(n k)!
C. Gravel (Carleton)
STAT 3502
January 2015
10 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 7
From our previous example,
the number of ways we can form groups of 3 workers out of 10
in which we will rank them as manager, assistant manager and
10!
= 10 9 8 = 720.
cashier is P3,10 = 103!
the number of ways we can form groups of 3 workers out of 10
where
rank doesnt matter is
10
10!
=
= 1098
= 720
= 120.
3
3!(103)!
321
6
C. Gravel (Carleton)
STAT 3502
January 2015
11 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 8
C. Gravel (Carleton)
STAT 3502
January 2015
12 / 21
Chapter 2 - Probability
2.3 Counting Techniques
Counting Techniques - 9
C. Gravel (Carleton)
STAT 3502
January 2015
13 / 21
Chapter 2 - Probability
2.4 Conditional Probability
2.4 Conditional Probability
Suppose that we are interested in the determining the
probability of event A, but we have the additional information,
event B has occurred".
For example, the overall polling results for the 2015 Canadian
election versus the same results conditioned on a particular
province.
Conditional Probability
The conditional probability of A given that B has occurred is
P(A|B) =
C. Gravel (Carleton)
P(A B)
P(B)
STAT 3502
January 2015
14 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 2
We are frequently interested in joint probabilities, P(A B), when
only P(A|B) and P(B) are available.
The Multiplication Rule
P(A B) = P(A|B) P(B).
C. Gravel (Carleton)
STAT 3502
January 2015
15 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 3
Example: Advertising Effectiveness
Suppose 100 people are surveyed upon entering a store as to:
i) did they see the stores advertisement?
ii) did they make a purchase?
What is the probability that a person who did not see the ad
made a purchase?
What is the probability that a person who did see the ad made a
purchase?
Survey Results:
Sixty people saw the ad.
Forty customers made purchases and thirty of those saw the ad.
C. Gravel (Carleton)
STAT 3502
January 2015
16 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 4
An easy way to visualize and solve conditional probabilities is to use
a probability tree.
Example: Electronic Device Repair
An electronic device contains 2 parts that are easy to remove
and replace. If the device fails, the probability that Part 1 needs
to be replaced is 50%. However, if Part 1 needs to be replaced
Part 2 may also need to be replaced with probability 70%.
Alternatively, if Part 1 is functional, Part 2 needs to be replaced
with probability 10 %.
i) What is the probability that the device failing will result in both
Part 1 and 2 needing to be replaced?
ii) What is the probability that Part 2 will be replaced?
C. Gravel (Carleton)
STAT 3502
January 2015
17 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 5
Law of Total Probability, p. 78
Let A1 , ..., Ak be mutually exclusive, exhaustive events. Then, for
any other event B,
P(B)
P(B|A1 )P(A1 ) + ... + P(B|Ak )P(Ak )
k
X
=
P(B|A )P(A ).
=1
Recall, mutually exclusive means that no two events have common
outcomes, or, A Aj = , for all , j such that 6= j.
And, note that exhaustive means that at least one event, A must
occur, or, A1 A2 ... Ak = S (the sample space).
This will allow us to solve P(B) from example 3 on the previous slide.
C. Gravel (Carleton)
STAT 3502
January 2015
18 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 6
Bayes Theorem, p. 79
Let A1 , ..., Ak be mutually exclusive, exhaustive events with prior
probabilities P(A ), = 1, ..., k. Then, for any other event B, the
posterior probability of Aj given that B has occurred is,
P(Aj |B) =
P(Aj B)
P(B)
P(B|Aj )P(Aj )
= Pk
,
P(B|A )P(A )
=1
for each j = 1, ..., k.
Prior probabilities are initial probabilities of occurrence,
observed without any other information.
However, we may be interested in the posterior probability, or
what happens to the probabilities of the events of interest
given that new information, B, is observed.
C. Gravel (Carleton)
STAT 3502
January 2015
19 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 7
C. Gravel (Carleton)
STAT 3502
January 2015
20 / 21
Chapter 2 - Probability
2.4 Conditional Probability
Conditional Probability - 8
C. Gravel (Carleton)
STAT 3502
January 2015
21 / 21
