Discrete Probability Distributions

Mahbub Latif, PhD

February 2025
Plan
The binomial distribution
The Geometric and Negative Binomial Distributions
The Hypergeometric Distribution
The Poisson Distribution

2
Examples of some discrete random variables

X1 = number of heads in tossing 35 tosses of a coin

What is the probability to observe 20 heads in 35 tosses of a coin?
P (X1 = 20) =?

3
Examples of some discrete random variables

X2 = number of tosses needed to observe the head for the first time
What is the probability to observe the head at 5th toss for the first time?
P (X2 = 5) =?

X3 = number of tosses needed to observe the head for the third time
What is the probability to observe the head at 7th toss for the third time?
P (X3 = 7) =?

4
Examples of some discrete random variables

Suppose a box has 80 balls of which 60 are red. A sample of 10 balls are
drawn from the box without replacement.
X4 = number of red balls in the selected 10 balls
What is the probability that 3 of the 10 balls are red?
P (X4 = 3) =?

5
Examples of some discrete random variables

X5 = number of "Premio" model Toyota car crosses an intersection

between 10AM to 11AM.
What is the probability that 3 Premio model cars will cross the intersection
between 10AM to 11AM?
P (X5 = 3) =?

6
Introduction

Discrete probability distributions allow the concise description of many

commonly occurring phenomena
The probability mass functions of these distributions are described by
formulas that depend upon some parameter values
The corresponding expectations and variances are functions of these
parameters

7
Introduction

The following distributions will be discussed in this chapter:

the binomial
the geometric
the negative binomial
the hypergeometric
the Poisson distribution

8
The Binomial Distribution

9
Bernoulli Random Variables

A Bernoulli random variable takes one of two values and it is related to an

experiment that has only two possible outcomes (Bernoulli trial)
E.g. tossing of a coin, whether an item is defective or not, etc.
In general, such outcomes are labelled as "success" or "failure"
Typically, a Bernoulli distributed random variable is labeled as either 0 or 1
E.g. experiment with tossing a coin, the random variable takes the value 1
if head appears and takes 0 if tail appears

10
Bernoulli Random Variables

The Bernoulli distribution has only one parameter p, the probability of

success and so the probability of failure is (1 − p), 0 ≤ p ≤ 1

Let X is the number of success in one binary trial, the probability mass
function of X follows a Bernoulli distribution with parameter p, and the
corresponding probability mass function
x 1−x
P (X = x) = p (1 − p) , x = 0, 1

1 1−1
P (X = 1) = p (1 − p) = p

P (X = 0) =?

11
Bernoulli Random Variables

The expectation of a Bernoulli distributed random variable X

1

E(X) = ∑ xP (X = x)

x=0

= 1 × P (X = 1) + 0 × P (X = 0) = p

The variance of the random variable is

V ar(X) = p(1 − p)

12
Experiment with tossing three coins

p → probability of observing a head in a single toss

X → number of heads observed in tossing three coins
outcome X prob X prob
HHH 3 p
3
0 (1 − p)
3 0
p (1 − p)
3

HHT 2 2
p (1 − p) 1 3p(1 − p)
2
3p(1 − p)
2

HTH 2 2
p (1 − p) 2 2
3p (1 − p)
2
3p (1 − p)

THH 2 2
p (1 − p) 3 p
3 3
p (1 − p)
0

HTT 1 p(1 − p)
2

THT 1 p(1 − p)
2

TTH 1 p(1 − p)
2

TTT 0 (1 − p)
3

13
Experiment with tossing three coins

Distribution of X, the number of heads in tossing three coins

X prob
0 (1 − p)
3 0
p (1 − p)
3 3 0
( ) p (1 − p)
0
3

1 3p(1 − p)
2
3p(1 − p)
2 3
( ) p(1 − p)
1
2
n n!
( ) =
2 2
3p (1 − p)
2
3p (1 − p)
3 2
( ) p (1 − p)
2
x x! (n−x)!

3 p
3 3
p (1 − p)
0 3 3
( ) p (1 − p)
3
0
3 3! 3×2×1
( ) = = = 3
2 2!(3−2)! (2×1)(1)

0! = 1

What would be the distribution of the number of heads in tossing four

coins?
14
Binomial distribution

Consider an experiment consisting of n independent Bernoulli trials and

each trial has a constant probability p of success
The total number of successes X of such n Bernoulli trials follows a
binomial distribution with parameters n and p
X ∼ B(n, p)

15
Binomial distribution

The probability mass function of X ∼ B(n, p)

n x n−x
P (X = x) = ( ) p (1 − p) , x = 0, 1, 2, … , n
x

It can be shown
n
n
x n−x
∑ ( ) p (1 − p) = 1
x
x=0

16
Binomial distribution

Expected value and variance of X

μ = E(X) = ∑ x P (X = x) = np
x

2 2 2
σ = V (X) = E(X ) − μ = np(1 − p)

17
The shape of binomial distribution depends on the value of p, binomial
distribution is symmetric for p = 0.5 and skewed for p ≠ 0.5
18
Binomial distribution

For X ∼ B(8, 0.5)

P (X ≥ 7) = P (X = 7) + P (X = 8)

8 7 1
8 8 0
= ( )(0.5) (0.5) + ( )(0.5) (0.5)
7 8

= 0.03125 + 0.00391 = 0.03516 = P (X ≤ 1)

19
Binomial distribution

For X ∼ B(8, 0.75)

P (X ≥ 7) = P (X = 7) + P (X = 8)

8 7 1
8 8 0
= ( )(0.75) (0.75) + ( )(0.75) (0.75)
7 8

= 0.26697 + 0.10011 = 0.36708 ≠ P (X ≤ 1)

20
Example 17 (Milk Container Contents)

The probability of 0.261 that a milk container is underweight. Suppose that

the milk containers are shipped to retail outlets in boxes of 20 containers.
What is the distribution of the number of underweight containers in a box?

21
Example 17 (Milk Container Contents)

What is the distribution of the number of underweight containers in a box?

The milk contents of two different milk containers are assumed to be
independent of each other
Each milk container can be classified as either "underweight" or "non-
underweight" (Bernoulli trial) and assume probability of being underweight
0.261 is constant over different containers
The number of underweight containers X in a box of 20 follows a binomial
distribution with parameters n = 20 and p = 0.261, i.e. X ∼ B(20, 0.261)

22
Example 17 (Milk Container Contents)

The expected number of underweight containers in a box

E(X) = np = 20 × 0.261 = 5.22 ≃ 6

The variance
V ar(X) = np(1 − p)

= (20)(0.261)(1 − 0.261)

= 3.86

Find the value of standard deviation of X

23
Example 17 (Milk Container Contents)

We have X ∼ B(20, 0.261)

What is the probability that a box contains exactly seven underweight

containers?
The number of underweight containter in a box of 20 containers X
follows a binomial distribution with parameters n = 20 and p = 0.261
20 7 13
P (X = 7) = ( ) 0.261 (1 − 0.261)
7

20! 7 13
= 0.261 (1 − 0.261) = 0.125
7!(20 − 7)!

24
Example 17 (Milk Container Contents)

We have X ∼ B(20, 0.261)

What is the probability that a box contains no more than three underweight
containers?
P (X ≤ 3) = P (X = 0) + P (X = 1) + P (X = 2) + P (X = 3)

20 0 20
20 1 19
= ( ) 0.261 (1 − 0.261) + ( ) 0.261 (1 − 0.261)
0 1

20 2 18
20 3 17
+ ( ) 0.261 (1 − 0.261) + ( ) 0.261 (1 − 0.261)
2 3

= 0.1935

25
Homework 3A

3.1.1 Suppose that X ∼ B(10, 0.12) . Calculate:

(a) P (X = 3) ,

(b) P (X ≤ 2) ,

(c) E(X)

(d) V ar(X)

26
Homework 3A

3.1.4

An archer hits a bull's-eye with a probability of 0.09, and the results of different
attempts can be taken to be independent of each other.
If the archer shoots nine arrows, calculate the probability that:
(a) Exactly two arrows score bull's-eyes.
(b) At least two arrows score bull's-eyes.
(c) What is the expected number of bull's-eyes scored?

27
Homework 3A

3.1.9

A company receives 60% of its orders over the Internet. Within a collection of
18 independently placed orders, what is the probability that
(a) between eight and ten of the orders are received over the Internet?
(b) no more than four of the orders are received over the Internet?

28
The Geometric and Negative Binomial
Distributions

29
The Geometric Distribution

The number of trials needed to observe the first success in a sequence of

independent Bernoulli trials with a constant success probability p has a
geometric distribution with parameter p
Let X be the number of trials required to get the first success and X follows
a geometric distribution with parameter p, i.e., X ∼ G(p)

The probability mass function of X

x−1
P (X = x) = (1 − p) p

x = 1, 2, 3, …

30
The Geometric Distribution

The cumulative distribution function

x
P (X ≤ x) = 1 − (1 − p)

Expected value
1
E(X) = ∑ P (X = x)x =
p
x

Variance
1 − p
V ar(X) =
2
p

31
Negative binomial distribution

Geometric distribution deals with the number of trials needed to observe

the success (e.g. head in a coin tossing experiment) for the first time
The number of trials needed to observe the rth (r ≥ 1) success is the
quantity of interest for a negative binomial distribution

32
Negative binomial distribution

The negative binomial distribution has two parameters: r and p

p → the probability of success
for r = 1 negative binomial distribution reduces to a geometric
distribution
Let X be the number of trials required to get the rth (r ≥ 1) success

33
Negative binomial distribution

The probability mass function of X, which follows a negative binomial

distribution with parameters r and p
x − 1
x−r r
P (X = x) = ( ) (1 − p) p , x = r, r + 1, r + 2, …
r − 1

This probability mass function deals with the fact that the rth success
occurs on the xth trial, so there are (x − 1) success in (r − 1) trials, and
corresponding probability can be obtained using the mass function of
binomial distribution

34
Negative binomial distribution

Expectation and variance

r
E(X) =
p

r(1 − p)
V ar(X) =
2
p

35
Example 25 (Telephone Ticket Sales)

Telephone ticket sales for a popular event are handled by a bank of

telephone salespersons who start accepting calls at a specified time.
In order to get through to an operator, a caller has to be lucky enough to
place a call at just the time when a salesperson has become free from a
previous client. Suppose that the chance of this is 0.1.
What is the distribution of the number of calls that a person needs to make
until a salesperson is reached?

36
Example 25 (Telephone Ticket Sales)

X → number of calls that a person needs to make until a salesperson is

reached
p = 0.1 → probability of reaching a salesperson and the probability of not
reaching a salesperson is 0.9 (= 1 − p)

X follows a geometric distribution with p = 0.1

the expected number of calls needed to get through to a salesperson
1 1
E(X) = = = 10
p 0.1

37
Example 25 (Telephone Ticket Sales)

The probability that a caller gets through on the fifth attempt

5−1
P (X = 5) = (1 − p) p

4
= (0.9) (0.1)

= 0.06561

38
Example 25 (Telephone Ticket Sales)

The probability that 15 or more calls are needed to reach the salesman
P (X ≥ 15) = 1 − P (X ≤ 14)
x
= 1 − (1 − p)

14
= 1 − (1 − .9) = .229

39
Example 12 (Personnel Recruitment)

Suppose that a company wishes to hire three new workers and that each
applicant interviewed has a probability of 0.6 of being found acceptable.
What is the distribution of the total number of applicants that the company
needs to interview?

40
Example 12 (Personnel Recruitment)

What is the distribution of the total number of applicants that the company
needs to interview?
An applicant could be either a "success" or a "failure", and the probability of
success p = 0.6
The random variable X is defined as the number of success needed to
select three new workers, where the probability of selecting a single worker
is 0.6
X follows a negative binomial distribution with parameters r = 3 and
p = 0.6

41
Example 12 (Personnel Recruitment)

Expected number of interviews needed to select three new workers is

r 3
E(X) = = = 5
p 0.6

42
Example 12 (Personnel Recruitment)

The probability that exactly six applicants need to be interviewed to select

three workers is
6 − 1
6−3 3
P (X = 6) = ( ) (1 − 0.6) (0.6)
3 − 1

5 3 3
= ( ) (0.4) (0.6) = 0.138
2

43
Example 12 (Personnel Recruitment)

If the company has a budget that allows up to six applicants to be

interviewed, then the probability that the budget is sufficient
P (X ≤ 6) = P (X = 3) + P (X = 4) + P (X = 5) + P (X = 6)

2 0 3
3 1 3
= ( ) (0.4) (0.6) + ( ) (0.4) (0.6)
2 2

4 5
2 3 3 3
+ ( ) (0.4) (0.6) + ( ) (0.4) (0.6)
2 2

= .216 + .259 + .207 + .138 = .820

44
Hoework 3B

3.2.1

If X has a geometric distribution with parameter p = 0.7, calculate:

P (X = 4), P (X = 1), P (X ≤ 5), P (X ≥ 8)

45
Hoework 3B

3.2.9

A company receives 60% of its orders over the Internet. Within a certain
period of time:
(a) What is the probability that the fifth order received is the first Internet
order?
(b) What is the probability that the eighth order received is the fourth
Internet order?

46
Hoework 3B

3.2.5 An archer hits a bull's-eye with a probability of 0.09, and the results of
different attempts can be taken to be independent of each other.
If the archer shoots a series of arrows, what is the probability that the first
bull's-eye is scored with the fourth arrow?
What is the probability that the third bull's-eye is scored with the tenth
arrow?
What is the expected number of arrows shot before the first bull's-eye is
scored?
What is the expected number of arrows shot before the third bull's-eye is
scored? 47
The Hypergeometric Distribution

48
The Hypergeometric Distribution

Consider a collection of N items of which r are of a certain kind, e.g.

r of the N items as being "defective",
a box containing N balls of which exactly r are red, etc.
The probability that a randomly selected items is of the special kind is
p = r/N

The experiment is to select n items from N of which r items are of special

kind

49
The Hypergeometric Distribution

The random variable of interest

X → the number of items of the special kind in n selected items

50
The Hypergeometric Distribution

Either sampling "with replacement" or "without replacement" can be used to

select n items from N
If n items are chosen with replacement
the distribution of X follows a binomial distribution with parameters n
and p = r/N

Selection with replacement ensures that the probability of selecting an

item of special kind remains the same over trials

51
The Hypergeometric Distribution

If n items are chosen without replacement, the distribution of X follows

hypergeometric distribution with parameters n, r, and N
The probability mass function of hypergeometric distribution
r N −r
( ) × ( )
x n−x
P (X = x) = max{0, n + r − N } ≤ x ≤ min{n, r}
N
( )
n

52
The Hypergeometric Distribution

The expected value

nr
E(X) = = np
N

It represents the distribution of the number of items of a certain kind in a

random sample of size n drawn without replacement from a population of
size N that contains r items of this kind

53
Example 17 (Milk Container Contents)

Suppose that milk is shipped to retail outlets in boxes that hold 16 milk
containers.
One particular box, which happens to contain 6 underweight containers, is
opened for inspection, and 5 containers are chosen at random.
What is the distribution of the number of underweight milk containers in the
sample chosen by the inspector?

54
Example 17 (Milk Container Contents)

What is the distribution of the number of underweight milk containers in the

sample chosen by the inspector?
The number of underweight milk containers in the sample X follows
hypergeometric distribution with parameters
n = 5, r = 6, N = 16

The probability that the inspector chooses two underweight containers

6 16−6
( ) × ( )
2 5−2
P (X = 2) = = 0.412
16
( )
6

55
Homework 3C

3.3.1

Let X have a hypergeometric distribution with N = 11, r = 6, and n = 7.

Calculate: P (X = 4), P (X = 5), and P (X ≤ 3)

3.3.2

A committee consists of eight right-wing members and seven left-wing

members. A subcommittee is formed by randomly choosing five of the
committee members.
Calculate probability mass function of the different values of the number of
right-wing members serving on the subcommittee.
56
Homework 3C

3.3.3

A box contains 17 balls of which 10 are red and 7 are blue. A sample of 5
balls is chosen at random and placed in a jar.
Calculate the probability that: (a) The jar contains exactly 3 red balls, (b) The
jar contains exactly 1 red ball, (c) The jar contains more blueballs than red
balls.

57
The Poisson Distribution

58
The Poisson distribution

The Poisson distribution is used to analyse a random variable that counts

the number of "events" that occur within certain specified boundaries
Examples of Poisson random variables
the number of typos in a page
the number of radioactive particles emitted by a substance
the number of telephone calls received by an operator within a certain
time limit
This distribution is named after the French mathematician Simeon Denis
Poisson (1781–1840) 59
The Poisson distribution

A random variable X follows a Poisson distribution with parameter λ (> 0)

has the following probability mass function

−λ x
e λ
P (X = x) = , x = 0, 1, 2, …
x!

The Poisson distribution is useful to model the number of times that a

certain event occurs per unit of time, distance, or volume
For a Poisson random variable X ∼ P o(λ)

E(X) = V ar(X) = λ

60
The Poisson distribution

The Poisson distribution can be used to approximate the B(n, p)

distribution when
1. n is very large (larger than 150, say) and
2. the success probability p is very small (smaller than 0.01, say)
A parameter value of λ = np should be used for the Poisson distribution, so
that it has the same expected value as the binomial distribution

61
Example 3 (Software errors)

Suppose that the number of errors in a piece of software has a Poisson

distribution with parameter λ = 3.
1. What is the probability that the software has no errors?
2. What is the probability that there are three or more errors in the software?

62
Example 3 (Software errors)

Suppose that the number of errors in a piece of software X has a Poisson

distribution with parameter λ = 3, i.e. X ∼ P o(3).

What is the probability that the software has no errors?

−λ 0
e λ
−3
P (X = 0) = = e = 0.0498
0!

63
Example 3 (Software errors)

Suppose that the number of errors in a piece of software X has a Poisson

distribution with parameter λ = 3, i.e. X ∼ P o(3).

What is the probability that there are three or more errors in the
software?
P (X ≥ 3) = 1 − P (X < 3)

= 1 − P (X = 0) − P (X = 1) − P (X = 2)

−3
= 1 − e [1 + 3 + 9/2]

= 1 − 0.0498(17/2) = 0.5767

64
Homework 3D

3.4.1

If X ∼ P o(3.2) , calculate:

P (X = 1)

P (X ≤ 3)

P (X = 0 | X ≤ 3)

65
Homework 3D

3.4.4

The number of cracks in a ceramic tile has a Poisson distribution with a mean
of λ = 2.4.
What is the probability that a tile has no cracks?
What is the probability that a tile has four or more cracks?

66
Homework 3D

3.4.8

In a scanning process, the number of misrecorded pieces of information has

a Poisson distribution with parameter λ = 9.2.
What is the probability that there are between six and ten misrecorded
pieces of information?
What is the probability that there are no more than four misrecorded
pieces of information?

67
Practice problems

3.8.1

An integrated circuit manufacturer produces wafers that contain 18 chips. Each

chip has a probability of 0.085 of not being placed quite correctly on the wafer.
Find the probability that a wafer contains at least three incorrectly placed
chips.
What is the probability that a wafer contains no more than one incorrectly
placed chip?
What is the expected number of incorrectly placed chips on a wafer?

68
Practice problems

3.8.4

A company's toll-free complaints line receives an average of about 40 calls per

hour. Use the Poisson distribution to estimate the probability that in one
minute there are
no calls
exactly one call
three or more calls

69
Practice problems

3.8.9

Suppose that a box contains 40 items of which 4 are defective. If a random

sample of 5 items is chosen, what is the probability that it contains no more
than 1 defective item?

70
Practice problems

3.8.10

If a fair die is rolled 22 times, what is the probability that a 6 is obtained

exactly 3 times?
If a fair die is rolled, what is the probability that the third time that a 6 is
obtained is on the tenth roll?
If a fair coin is tossed 11 times, what is the probability that three or fewer
heads are obtained?

71