Arkansas Tech University
MATH 3513: Applied Statistics I
Dr. Marcel B. Finan
2.2 Counting and Combinatorics
The major goal of this section is to establish several (combinatorial) tech-
niques for counting large finite sets without actually listing their elements.
These techniques provide effective methods for counting the size of events,
an important concept in probability theory.
The Fundamental Principle of Counting
Sometimes one encounters the question of listing all the outcomes of a cer-
tain experiment. One way for doing that is by constructing a so-called tree
diagram.
Example 2.2.1
List all two-digit numbers that can be constructed from the digits 1,2, and
3.
Solution.
A tree diagram is shown in Figure 2.2.1.
Figure 2.2.1
The different numbers are {11, 12, 13, 21, 22, 23, 31, 32, 33}
Of course, trees are manageable as long as the number of outcomes is not
1
large. If there are many stages to an experiment and several possibilities
at each stage, the tree diagram associated with the experiment would be-
come too large to be manageable. For such problems the counting of the
outcomes is simplified by means of algebraic formulas. The commonly used
formula is the Fundamental Principle of Counting, also known as the
multiplication rule of counting, which states:
Theorem 2.2.1
If a choice consists of k steps, of which the first can be made in n1 ways,
for each of these the second can be made in n2 ways,· · · , and for each of
these the k th can be made in nk ways, then the whole choice can be made in
n1 · n2 · · · · nk ways.
Proof.
In set-theoretic term, we let Si denote the set of outcomes for the ith task,
i = 1, 2, · · · , k. Note that n(Si ) = ni . Then the set of outcomes for the entire
job is the Cartesian product S1 × S2 × · · · × Sk = {(s1 , s2 , · · · , sk ) : si ∈
Si , 1 ≤ i ≤ k}. Thus, we just need to show that
n(S1 × S2 × · · · × Sk ) = n(S1 ) · n(S2 ) · · · n(Sk ).
The proof is by induction on k ≥ 2.
Basis of Induction
Suppose that S1 = {a1 , a2 , · · · , an } and S2 = {b1 , b2 , · · · , bm }. Then
S1 × S2 = {(a1 , b1 ), (a1 , b2 ), · · · , (a1 , bm ),
(a2 , b1 ), (a2 , b2 ), · · · , (a2 , bm ),
(a3 , b1 ), (a3 , b2 ), · · · , (a3 , bm ),
..
.
(an , b1 ), (an , b2 ), · · · , (an , bm )}
Thus, n(S1 × S2 ) = n · m = n(S1 ) · n(S2 ). Hence, the result is true for k = 2.
Induction Hypothesis
Suppose
n(S1 × S2 × · · · × Sk ) = n(S1 ) · n(S2 ) · · · n(Sk ).
2
Induction Conclusion
We must show
n(S1 × S2 × · · · × Sk+1 ) = n(S1 ) · n(S2 ) · · · n(Sk+1 ).
To see this, note that there is a one-to-one correspondence between the sets
S1 ×S2 ×· · ·×Sk+1 and (S1 ×S2 ×· · · Sk )×Sk+1 given by f (s1 , s2 , · · · , sk , sk+1 ) =
((s1 , s2 , · · · , sk ), sk+1 ). Thus, n(S1 × S2 × · · · × Sk+1 ) = n((S1 × S2 × · · · Sk ) ×
Sk+1 ) = n(S1 × S2 × · · · Sk )n(Sk+1 ) ( basis of induction). Now, applying the
induction hypothesis gives
n(S1 × S2 × · · · Sk × Sk+1 ) = n(S1 ) · n(S2 ) · · · n(Sk+1 )
Example 2.2.2
The following three factors were considered in the study of the effectiveness
of a certain cancer treatment:
(i) Medicine (A1 , A2 , A3 , A4 , A5 )
(ii) Dosage Level (Low, Medium, High)
(iii) Dosage Frequency (1,2,3,4 times/day)
Find the number of ways that a cancer patient can be given the medecine?
Solution.
The choice here consists of three stages, that is, k = 3. The first stage, can be
made in n1 = 5 different ways, the second in n2 = 3 different ways, and the
third in n3 = 4 ways. Hence, the number of possible ways a cancer patient
can be given medecine is n1 · n2 · n3 = 5 · 3 · 4 = 60 different ways
Example 2.2.3
How many license-plates with 3 letters followed by 3 digits exist?
Solution.
A 6-step process: (1) Choose the first letter, (2) choose the second letter,
(3) choose the third letter, (4) choose the first digit, (5) choose the second
digit, and (6) choose the third digit. Every step can be done in a number of
ways that does not depend on previous choices, and each license plate can
be specified in this manner. So there are 26 · 26 · 26 · 10 · 10 · 10 = 17, 576, 000
ways
3
Example 2.2.4
How many numbers in the range 1000 - 9999 have no repeated digits?
Solution.
A 4-step process: (1) Choose first digit, (2) choose second digit, (3) choose
third digit, (4) choose fourth digit. Every step can be done in a number
of ways that does not depend on previous choices, and each number can be
specified in this manner. So there are 9 · 9 · 8 · 7 = 4, 536 ways
Example 2.2.5
How many license-plates with 3 letters followed by 3 digits exist if exactly
one of the digits is 1?
Solution.
In this case, we must pick a place for the 1 digit, and then the remaining
digit places must be populated from the digits {0, 2, · · · 9}. A 6-step pro-
cess: (1) Choose the first letter, (2) choose the second letter, (3) choose the
third letter, (4) choose which of three positions the 1 goes, (5) choose the
first of the other digits, and (6) choose the second of the other digits. Every
step can be done in a number of ways that does not depend on previous
choices, and each license plate can be specified in this manner. So there are
26 · 26 · 26 · 3 · 9 · 9 = 4, 270, 968 ways
Permutations
Consider the following problem: In how many ways can 8 horses finish in a
race (assuming there are no ties)? We can look at this problem as a decision
consisting of 8 steps. The first step is the possibility of a horse to finish first
in the race, the second step is the possibility of a horse to finish second, · · · ,
the 8th step is the possibility of a horse to finish 8th in the race. Thus, by
the Fundamental Principle of Counting there are
8 · 7 · 6 · 5 · 4 · 3 · 2 · 1 = 40, 320 ways.
This problem exhibits an example of an ordered arrangement, that is, the
order the objects are arranged is important. Such an ordered arrangement is
called a permutation. Products such as 8 · 7 · 6 · 5 · 4 · 3 · 2 · 1 can be written
in a shorthand notation called factorial. That is, 8 · 7 · 6 · 5 · 4 · 3 · 2 · 1 = 8!
(read “8 factorial”). In general, we define n factorial by
n! = n(n − 1)(n − 2) · · · 3 · 2 · 1, n ≥ 1
4
where n is a whole number. By convention we define
0! = 1
Example 2.2.6
10!
Evaluate the following expressions: (a) 6! (b) 7!
.
Solution.
(a) 6! = 6 · 5 · 4 · 3 · 2 · 1 = 720
(b) 10!
7!
= 10·9·8·7·6·5·4·3·2·1
7·6·5·4·3·2·1
= 10 · 9 · 8 = 720
Using factorials and the Fundamental Principle of Counting, we see that
the number of permutations of n objects is n!.
Example 2.2.7
There are 5! permutations of the 5 letters of the word “rehab.” In how many
of them is h the second letter?
Solution.
Then there are 4 ways to fill the first spot. The second spot is filled by the
letter h. There are 3 ways to fill the third, 2 to fill the fourth, and one way
to fill the fifth. There are 4! such permutations
Example 2.2.8
Five different books are on a shelf. In how many different ways could you
arrange them?
Solution.
The five books can be arranged in 5 · 4 · 3 · 2 · 1 = 5! = 120 ways
Counting Permutations
We next consider the permutations of a set of objects taken from a larger
set. Suppose we have n items. How many ordered arrangements of k items
can we form from these n items? The number of permutations is denoted
by P (n, k). The n refers to the number of different items and the k refers to
the number of them appearing in each arrangement. A formula for P (n, k)
is given next.
5
Theorem 2.2.2
For any non-negative integer n and 0 ≤ k ≤ n we have
n!
P (n, k) = .
(n − k)!
Proof.
We can treat a permutation as a decision with k steps. The first step can be
made in n different ways, the second in n − 1 different ways, ..., the k th in
n − k + 1 different ways. Thus, by the Fundamental Principle of Counting
there are n(n − 1) · · · (n − k + 1) k−permutations of n objects. That is,
P (n, k) = n(n − 1) · · · (n − k + 1) = n(n−1)···(n−k+1)(n−k)!
(n−k)!
n!
= (n−k)!
Example 2.2.9
How many license plates are there that start with three letters followed by 4
digits (no repetitions)?
Solution.
The decision consists of two steps. The first is to select the letters and this
can be done in P (26, 3) ways. The second step is to select the digits and
this can be done in P (10, 4) ways. Thus, by the Fundamental Principle of
Counting there are P (26, 3) · P (10, 4) = 78, 624, 000 license plates
Example 2.2.10
How many five-digit zip codes can be made where all digits are different?
The possible digits are the numbers 0 through 9.
Solution.
10!
The answer is P (10, 5) = (10−5)!
= 30, 240 zip codes
Combinations
In a permutation the order of the set of objects or people is taken into ac-
count. However, there are many problems in which we want to know the
number of ways in which k objects can be selected from n distinct objects in
arbitrary order. For example, when selecting a two-person committee from a
club of 10 members the order in the committee is irrelevant. That is choosing
Mr. A and Ms. B in a committee is the same as choosing Ms. B and Mr. A.
A combination is defined as a possible selection of a certain number of objects
taken from a group without regard to order. More precisely, the number of
6
k−element subsets of an n−element set is called the number of combina-
tions of n objects taken k at a time. It is denoted by C(n, k) and is read
“n choose k”. The formula for C(n, k) is given next.
Theorem 2.2.3
If C(n, k) denotes the number of ways in which k objects can be selected
from a set of n distinct objects then
P (n, k) n!
C(n, k) = = .
k! k!(n − k)!
Proof.
Since the number of groups of k elements out of n elements is C(n, k) and
each group can be arranged in k! ways, we have P (n, k) = k!C(n, k). It
follows that
P (n, k) n!
C(n, k) = =
k! k!(n − k)!
n
An alternative notation for C(n, k) is . We define C(n, k) = 0 if k < 0
k
or k > n.
Example 2.2.11
A jury consisting of 2 women and 3 men is to be selected from a group of 5
women and 7 men. In how many different ways can this be done? Suppose
that either Steve or Harry must be selected but not both, then in how many
ways this jury can be formed?
Solution.
There are C(5, 2) · C(7, 3) = 350 possible jury combinations consisting of 2
women and 3 men. Now, if we suppose that Steve and Harry can not serve
together then the number of jury groups that do not include the two men at
the same time is C(5, 2)C(5, 2)C(2, 1) = 200
Choosing a combination of k objects from a set of n divides the n objects
into two subsets: the k that were chosen and the n − k that were not cho-
sen. Sometimes a set is to be divided up into more than two subsets. The
following theorem covers this situation.
7
Theorem 2.2.4
Suppose we have n objects where there are k1 , k2 , · · · , kr objects that are
identical and k1 + · · · + kr = n. The number of ways to arrange them is
n!
.
k1 !k2 ! · · · kr !
Proof.
For the first group, there are C(n, k1 ) arrangements. For the second group,
there are C(n − k1 , k1 ) arrangements, and so on, for the rth group there
are C(n − k1 − k2 − · · · − kr−1 , kr ) arrangements. Hence, the Fundamental
Principle of Counting, the total number of ways to arrange the n items into
r groups is
C(n, k1 )C(n − k1 , k2 ) · · · C(n − k1 − k2 − · · · − kr−1 , kr )
n! (n − k1 )! (n − k1 − k2 − · · · − kr−1 )!
= ···
k1 !(n − k1 )! k2 !(n − k1 − k2 )! kr !(n − k1 − · · · − kr )!
n!
=
k1 !k2 ! · · · kr !
Example 2.2.12
How many ways to order the letters of MISSISSIPPI?
Solution.
There are 11 letters, one M, four S, four I and 2 P. The number of ways to
11!
order the letters is 1!4!4!2! = 34650 ways
Next, we illustrate the use of counting rules in finding probabilities.
Example 2.2.13
In an actuarial course in probability, an instructor has decided to give his
class a weekly quiz consisting of 5 multiple-choice questions taken from a
pool of previous SOA P/1 exams. Each question has 4 answer choices, of
which 1 is correct and the other 3 are incorrect.
(a) How many answer choices are there?
(b) What is the probability of getting all 5 right answers?
(c) What is the probability of answering exactly 4 questions correctly?
(d) What is the probability of getting at least four answers correctly?
8
Solution.
(a) We can look at this question as a decision consisting of five steps. There
are 4 ways to do each step so that by the Fundamental Principle of Counting
there are
(4)(4)(4)(4)(4) = 1024 possible choices of answers.
(b) There is only one way to answer each question correctly. Using the
Fundamental Principle of Counting there is (1)(1)(1)(1)(1) = 1 way to answer
all 5 questions correctly out of 1024 possible answer choices. Hence,
1
P(all 5 right) = 1024
(c) The following table lists all possible responses that involve exactly 4 right
answers where R stands for right and W stands for a wrong answer
Five Responses Number of ways to fill out the test
WRRRR (3)(1)(1)(1)(1) = 3
RWRRR (1)(3)(1)(1)(1) = 3
RRWRR (1)(1)(3)(1)(1) = 3
RRRWR (1)(1)(1)(3)(1) = 3
RRRRW (1)(1)(1)(1)(3) = 3
So there are 15 ways out of the 1024 possible ways that result in 4 right
answers and 1 wrong answer so that
15
P(4 right,1 wrong) = 1024
≈ 1.5%
(d) “At least 4” means you can get either 4 right and 1 wrong or all 5 right.
Thus,
P(at least 4 right) =P(4R, 1W ) + P (5R)
15 1
= +
1024 1024
16
= ≈ 0.016
1024
