Chapter 4 Discrete Distributions

Bernoulli Distribution
Consider flipping a coin that has probability 𝜃 of coming up heads and probability 1 − 𝜃 of
coming up tails, where 0 < 𝜃 < 1. Let 𝑋 = 1 if the coin is heads, while 𝑋 = 0 if the coin is tails.
Then 𝑃𝑋 (1) = P(X = 1) = θ, while 𝑃𝑋 (0) = P(X = 0) = 1 − θ . Or

𝑃𝑋 (𝑥) = 𝜃 𝑥 (1 − 𝜃)𝑛−𝑥 𝑥 = 0, 1
The random variable X is said to have the Bernoulli distribution; we write this as
𝑋 ~𝐵𝑒𝑟𝑛𝑜𝑢𝑙𝑙𝑖(𝜃).
Bernoulli distributions arise anytime we have a response variable that takes only two possible
values, and we label one of these outcomes as 1 and the other as 0. For example, 1 could
correspond to success and 0 to failure of some quality test applied to an item produced in a
manufacturing process. In this case, 𝜃 is the proportion of manufactured items that will pass the
test. Alternatively, we could randomly select an individual from a population and record a 1 if
the individual is female and a 0 if the individual is male. In this case, 𝜃 is the proportion of
females in the population.
Example1: The prevalence of a certain disease in the general population is 10%. If
we randomly select a person from this population, we can have only two possible
outcomes (diseased or healthy person). We call one of these outcomes ( diseased
person) success and the other (healthy person), a failure. The probability of
success (p) or diseased person is 10% or 0.1. So, the probability of failure (q) or
healthy person = 1-p = 1-0.1 = 0.9.

Binomial Distribution
Consider flipping n coins, each of which has (independent) probability 𝜃 of coming up heads,
and probability 1 − 𝜃 of coming up tails (0 < 𝜃 < 1). Let X be the total number of heads
showing. We see that 𝑥 = 0, 1, 2, … , 𝑛 and
𝑛
𝑃𝑋 (𝑥) = 𝑃𝑋 (𝑋 = 𝑥) = ቀ ቁ 𝜃 𝑥 (1 − 𝜃)𝑛−𝑥 𝑥 = 0, 1, 2, . . . , 𝑛
𝑥
The random variable X is said to have the Binomial distribution; we write this as
𝑋~𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙(𝑛, 𝜃).
The binomial distribution is applicable to any situation involving n independent performances of
a random system for each performance, we are recording whether a particular event has
occurred, called a success, or has not occurred, called a failure. If we denote the event in
question by 𝐴 and put 𝜃 = 𝑃(𝐴). We have that the number of successes in the n performances
is distributed Binomial (𝑛, 𝜃).
Example 1 We could be testing light bulbs produced by a manufacturer, and 𝜃 is the probability
that a bulb works when we test it. Then the number of bulbs that work in a batch of n is
distributed Binomial (𝑛, 𝜃).
Example 2 If a basketball player has probability of 𝜃 getting his shots results in goal, then the
number of shots obtained in n attempts is distributed Binomial (𝑛, 𝜃).

Question 1: Find the binomial distribution of getting a six in three tosses of an unbiased dice.

Answer: Let X be the random variable of getting six. Then X can be 0, 1, 2, 3.

Here, n = 3

p = Probability of getting a six in a toss = ⅙

q = Probability of not getting a six in a toss = 1 – ⅙ = ⅚

𝑛 3
P(X = 0) = ቀ ቁ pr q(n – r) = ቀ ቁ (⅙)0 (⅚)3 – 0
𝑟 0

= 1 × 1 × 125/216 = 125/216

𝑛 3
P(X = 1) = ቀ ቁ pr q(n – r) = ቀ ቁ (⅙)1 (⅚)3 – 1
𝑟 1
= 3 × ⅙ × 25/36 = 25/72

𝑛 3
P(X = 2) = ቀ ቁ pr q(n – r) = ቀ ቁ (⅙)2 (⅚)3 – 2
𝑟 1

= 3 × 1/36 × ⅚ = 5/72

𝑛 3
P(X = 3) = ቀ ቁ pr q(n – r) = ቀ ቁ (⅙)3 (⅚)3 – 3
𝑟 3

= 1 × 1/216 × 1 = 1/216

Question 2: Find the probability of getting at least 5 times heads on tossing an unbiased coin
for 6 times by using the binomial distribution.

Answer:

p = P(getting an head in a single toss) = ½

q = P(not getting an head in a single toss) = ½

X = successfully getting a head

P(X ≥ 5) = P(getting at least 5 heads) = P(X = 5) + P(X = 6)

6 6
= ቀ ቁ (½)5 (½)(6 – 5) + ቀ ቁ (½)6 (½)6 – 6
5 6

= 6 × (½)6 + 1 × (½)6 = 7/24.

Hence, the probability of getting at least 5 heads is 7/24.

Question 3: On average, every one out of 10 telephones is found busy. Six telephone numbers
are selected at random. Find the probability that four of them will be busy.

Answer: Let

X: event of getting a busy phone number

P: probability of getting a phone number busy = 1/10

q: probability of not getting a phone number busy= 9/10

6
The required probability = P(X = 4) = ቀ ቁ p4 q(6 – 4)
4
= 15 × (1/10)4 × (9/10)2 = 15 × 81/106 = 0.001215.

Question 4: When a biased coin is tossed, the probability of getting a head 3 times more than
the probability of getting a tail. Find the probability distribution for getting a tail, if the coin is
tossed twice.

Answer: Let the probability of getting a tail be p, then the probability of getting a head will be
3p. Now, p + 3p = 1 ⇒ p = ¼

q = P(not getting a tail) = 1 – ¼ = ¾

X = event of getting a tail in a toss. Then, possible values of x will be 0, 1, 2

P(X = 0) = C(2, 0) p0 q2 – 0 = 1 × 1 × (¾)2= 9/16

P(X = 1) = C(2, 1) p1 q2 – 1= 2 × (¼) × (¾) = ⅜

P(X = 2) = C(2, 2) p2 q2 – 2= 1 × (¼)2 × 1 = 1/16

Some questions
1. Find the binomial distribution of getting an even number if an unbiased dice is thrown
thrice.

2. Three cards are drawn with replacement from a well-shuffled deck of 52 cards. Find the
probability distribution of the number of aces drawn. Also, find the mean and variance
of the distribution.

3. The probability of a trainee archer hitting the target is ¼. If he takes 7 shots, what is the
probability of his hitting the target at least twice?

4. The random variable X has binomial distribution Binomial(20, 0.2). Determine each of
the following. A) P(X = 5) B) P(X < 13) C) P(X > 10) D) P(8 < X <14)

5. The probability of Mathew being late to school is 0.15. Find the probability that in next
25 mornings he will arrive late to school:

a) on exactly 6 occasions.
 b) on less than 6 occasions.

c) at least twice

Geometric Distribution
Consider repeatedly flipping a coin that has probability 𝜃 of coming up heads and probability
1 − 𝜃 of coming up tails, where again 0 < 𝜃 < 1. Let X be the number of tails that appear
before the first head. Then for 𝑘 ≥ 0, 𝑋 = 𝑘 if and only if the coin shows exactly 𝑘 tails
followed by a head. The probability of this is equal to (1 − 𝜃)𝑘 𝜃 . Hence,

𝑃 (𝑋 = 𝑘) = (1 − 𝜃)𝑘 𝜃 𝑘 = 0, 1, 2, 3, …
The random variable X is said to have the Geometric distribution; we write this as
𝑋~ 𝐺𝑒𝑜𝑚𝑒𝑡𝑟𝑖𝑐(𝜃).

The geometric distribution applies whenever we are counting the number of failures until the
first success for independent performances of a random system where the occurrence of some
event is considered a success. For example, the number of light bulbs tested that work until the
first bulb that does not (a working bulb is considered a “failure” for the test) and the number of
shots without a hit until the first goal for the basketball player both follow the geometric
distribution.

Example1: A researcher is waiting outside of a library to ask people if they support a certain law.
The probability that a given person supports the law is p = 0.2. What is the probability that the
fourth person the researcher talks to is the first person to support the law?
Answer: The number of “failures” until the first success (the number of people who don’t
support the law until the first person supports it) is 3. Thus, using the Geometric Distribution
with p =0.2 and x = 3 failures, we find that P(X =3)= 0.10240.
Example2: Matthew is a high school basketball player and a 75% free throw shooter. Calculate
what will be the probability that Matthew makes his first goal on his fifth shot?
Answer: The distribution we are working with is a geometric distribution with a success
probability of 0.75, or 𝑋~𝐺(0.75). Using the probability formula: P(X=5)= (1−0.75)5-1(0.75)=
(0.25)4(0.75)= 0.00293 ≈ 0.003

The probability that the first success is on the third trial is approximately 0.003.

Example3: Suppose you are playing a game of darts. The probability of success is 0.4. Calculate
what will be the probability that you will hit the dartboard on the third try?
Answer: We are looking for the first success, thus, geometric distribution has to be used.
p = 0.4
P(X = x) = (1 - p)x−1p
P(X = 3) = (1 - 0.4)3-1(0.4) = (0.6)2 (0.4) = 0.144
Probability (you will hit the dartboard on the third try)= 0.144

Some questions:
1. Calculate the probability density of geometric distribution if the value of p is 0.42; x = 1,
2, 3, …
2. If a patient is waiting for a suitable blood donor and the probability that the selected
donor will be a match is 0.2, then calculate what will be the expected number of donors
who will be tested until a match is found including the matched donor.
3. A light bulb manufacturing factory finds 3 in every 60 light bulbs defective. Calculate
what will be the probability that the first defective light bulb with be found when the 6th
one is tested?

Negative-Binomial Distribution
Generalizing the Geometric Distribution, consider repeatedly flipping a coin that has probability
θ of coming up heads and probability 1 − 𝜃 of coming up tails. Let r be a positive integer, and X
be the number of tails appearing before the 𝑟 − 𝑡ℎ head. Then for 𝑘 ≥ 0, 𝑋 = 𝑘 if and only if
the coin shows exactly 𝑟 − 1 heads (and 𝑘 tails) on the first 𝑟 − 1 + 𝑘 flips, and then shows a
head on the (𝑟 + 𝑘)-th flip. The probability of this is equal to

𝑟−1+𝑘 𝑟
𝑃 (𝑋 = 𝑘) = ቀ ቁ 𝜃 (1 − 𝜃)𝑘 𝑘 = 0, 1, 2, 3, …
𝑘
The random variable X is said to have the Negative-Binomial (𝑟, 𝜃) distribution; we write this as
𝑋~ 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 − 𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙(𝑟, 𝜃). Of course, the special case r=1 corresponds to the Geometric
distribution. So, in terms of our notation, we have that 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 − 𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙(1, 𝜃) =
𝐺𝑒𝑜𝑚𝑒𝑡𝑟𝑖𝑐(𝜃).
The 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 − 𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙 (𝑟, 𝜃) distribution applies whenever we are counting the number of
failures until the rth success for independent performances of a random system where the
occurrence of some event is considered a success. For example, the number of light bulbs
tested that work until the third bulb that does not and the number of shots without a hit until
the fifth goal for the basketball player both follow the negative-binomial distribution.
Poisson Distribution
A Poisson distribution is a discrete probability distribution. It gives the probability of an event
happening a certain number of times (k) within a given interval of time or space. We say that a
random variable X has the 𝑃𝑜𝑖𝑠𝑠𝑜𝑛(𝜆) distribution and write 𝑋 ~ 𝑃𝑜𝑖𝑠𝑠𝑜𝑛(𝜆). The probability
function of this variable is:

𝜆𝑥
𝑃(𝑋 = 𝑥 ) = 𝑒 −𝜆 𝑥 = 0, 1, 2, 3, …
𝑥!
𝜆: mean number of successes that occur during a specific interval.
𝑥: number of successes
𝑒: a constant equal to approximately 2.71

These are examples of events that may be described as Poisson processes:

1. My computer crashes on average once every 2 months.

2. Hospital emergencies receive on average 5 serious cases every 24 hours.
3. The number of cars passing through a point, on a small road, is on average 4 cars every
30 minutes.
4. I receive on average 3 e-mails every 2 hours.
5. Customers make on average 10 calls every hour to the customer service.
A Poisson experiment is an experiment that has the following properties:

1. The number of successes in the experiment can be counted.

2. The mean number of successes that occurs during a specific interval of time (or space) is
known.
3. Each outcome is independent.
4. The probability that a success will occur is proportional to the size of the interval.
Foe example, suppose a particular hospital experiences an average of 10 births per hour. This is
a Poisson experiment because it has the following four properties:

• The number of successes in the experiment can be counted – We can count the number
of births.
• The mean number of successes that occurs during a specific interval of time is known –
It is known that an average of 10 births per hour occur.
• Each outcome is independent – The probability that one mother gives birth during a
given hour is independent of the probability of another mother giving birth.
• The probability that a success will occur is proportional to the size of the interval – the
longer the interval of time, the higher the probability that a birth will occur.

When λ is low, the distribution is much longer on the right side of its peak than its left (i.e., it is
strongly right-skewed).
Question1: It is known that a certain website makes 10 sales per hour. In each hour, what is the
probability that the site makes exactly 8 sales?

Answer: Using the Poisson Distribution with λ = 10 and x = 8, we find that P(X =8) = 0.1126.

Question2: It is known that a certain realtor makes an average of 5 sales per month. In a given
month, what is the probability that she makes more than 7 sales?

Answer: Using the Poisson Distribution with λ = 5 and x = 7, we find that P(X >7) = 0.13337.

Question3: It is known that a certain hospital experience 4 births per hour. In a given hour,
what is the probability that 4 or less births occur?

Answer: Using the Poisson Distribution with λ = 4 and x = 4, we find that P(X ≤ 4) = 0.62884.

Question4: A statistician records the number of cars that approach an intersection. He finds
that an average of 1.6 cars approach the intersection every minute. Assuming the number of
cars that approach this intersection follows a Poisson distribution, what is the probability that 3
or more cars will approach the intersection within a minute?

Answer: For this problem, 𝜆 = 1.6. The goal of this problem is to find 𝑃(𝑋 ≥ 3), the probability
that there are 3 or more cars approaching the intersection within a minute. Since there is no
upper limit on the value of k, this probability cannot be computed directly. However, its
complement, 𝑃(𝑋 ≤ 2), can be computed to give: 𝑃(𝑋 ≥ 3)

1.60
𝑃(𝑥 = 0) = 𝑒 −1.6 ≈ 0.202
0!

1.61
𝑃(𝑥 = 1) = 𝑒 −1.6 ≈ 0.323
1!

1.62
𝑃(𝑥 = 2) = 𝑒 −1.6 2!
≈ 0.258

⇒ 𝑝(𝑥 ≤ 2) = 𝑝(𝑥 = 0) + 𝑝(𝑥 = 1) + 𝑝(𝑥 = 2) ≈ 0.783

⇒ 𝑝(𝑥 ≥ 3) = 1 − 𝑝(𝑥 ≤ 2) ≈ 0.217.

Question5: A random variable X has a Poisson distribution with parameter λ such that P (X = 1)
= (0.2) P (X = 2). Find P (X = 0).

Answer: For the Poisson distribution, the probability function is defined as:

P (X =x) = (e– λ λx)/x!, where λ is a parameter.

Given that, P (x = 1) = (0.2) P (X = 2)

(e– λ λ1)/1! = (0.2)(e– λ λ2)/2!

⇒ λ = λ2/ 10 ⇒ λ = 10

Now, substitute λ = 10, in the formula, we get: P (X = 0) = (e– λ λ0)/0! = e-10

The Connection Between the Poisson and Binomial Distributions

The Poisson distribution is a limiting case of a Binomial distribution when the number
of trials, n, gets very large and p, the probability of success, is small. As a rule of
thumb, if

𝑛 ≥ 100 𝑎𝑛𝑑 𝑛𝑝 ≤ 10

the Poisson distribution (taking λ=np) can provide a very good approximation to the
binomial distribution.

This is particularly useful as calculating the combinations inherent in the probability

formula associated with the binomial distribution can become difficult when n is
large.
Key Differences Between Binomial and Poisson Distribution
The differences between binomial and Poisson distribution can be drawn clearly on the
following grounds:
1. The binomial distribution is one in which the probability of repeated number of trials is
studied. A probability distribution that gives the count of a number of independent
events occur randomly within a given period, is called probability distribution.
2. Binomial Distribution is bi-parametric, i.e. it is featured by two parameters n and p
whereas Poisson distribution is uni-parametric, i.e. characterised by a single parameter
m.
3. There are a fixed number of attempts in the binomial distribution. On the other hand, an
unlimited number of trials are there in a Poisson distribution.
4. The success probability is constant in binomial distribution but in Poisson distribution,
there are an extremely small number of success chances.
5. In a binomial distribution, there are only two possible outcomes, i.e. success or failure.
Conversely, there are an unlimited number of possible outcomes in the case of Poisson
distribution.
6. In binomial distribution Mean > Variance while in Poisson distribution mean = variance.

Some questions
1. A fast-food restaurant gets an average of 2.8 customers approaching the register
every minute. Assuming the number of customers approaching the register per
minute follows a Poisson distribution, what is the probability that 4 customers
approach the register in the next minute?

2. Suppose a particular hospital experiences an average of 10 births per hour. Now,

answer the following questions:

What is the probability that more than 12 births occur in each hour?

What is the probability that less than 5 births occur in each hour?

What is the probability that between 8 to 11 births occur in each hour?

 Summary
• A random variable X is discrete if ∑𝒙 𝑷(𝑿 = 𝒙) = 𝟏, i.e., if all its probability comes from
being equal to particular values.

• A discrete random variable X takes on only a finite, or countable, number of distinct

values.

Important discrete distributions include the Bernoulli, binomial, geometric, negative binomial,
and Poisson distributions.

• Binomial: Number of successes in n trials.

• Geometric: number of failures before the first success.
• Negative binomial: number of failures before the Xth success.
• Poisson: number of successes in a given interval of time or space.