Chapitre 3

Discrete Random Variables

3.1 General Notions on Discrete Random Variables

Definition 3.1 (Random Variable). A random variable is a variable that can take on
different real values to which it is possible to assign a probability.

Definition 3.2 (Real-valued Random variable). A real-valued random variable, often

denoted by X (or some other upper-case letter), is a function mapping a probability space
(ω, P ) into the real line R. In formal terms, we write

X :Ω −→ X(Ω) ⊂ R
ω −→ X(ω)

Definition 3.3 (Domain of a random variable). The set X(Ω) of possible realizations
that the random variable X can take is called the domain of the random variable X.

Definition 3.4 (Discrete random variable ). A random variable is said to be discrete if

its domain has either a finite or an infinite but countable number of elements. In this
case, the domain X(Ω) is given by X(Ω) = {x1 , x2 , . . . , xk } with the assumption that
x1 < x2 < · · · < xk .

3.2 Probability Distributions

Definition 3.5 (Probability Distribution Function). The probability distribution function
(p.d.f ) of a discrete random variable X is a list of each possible value of X (i.e. X(Ω)) to-
gether with the probabilities pi = P (X = xi ) that X takes the value xi for i ∈ {1, 2, . . . , k}.

The probability distribution is given by an explicit formula or in tabular form as

follows :

1
USTHB: ING INFO 1st year Probability and Statistics

xi x1 x2 ... xk
pi p1 p2 ... pk

The probabilities in the probability distribution function of a discrete random variable

X must satisfy the following two conditions :
• Each probability pi must be between 0 and 1 :

0 ≤ pi ≤ 1;

• The sum of all the possible probabilities is 1 :

i=k
X
pi = 1
i=1

Definition 3.6 (Cumulative Distribution Function). If X is a discrete random variable,

the cumulative distribution function is given by

FX :R −→ [0, 1]
X
x −→ FX (x) = P (X ≤ x) = P (X = xi )
xi ≤x

which yields to

0 if x < x1 ;











 p1 if x1 ≤ x < x2 ;



p1 + p 2 if x2 ≤ x < x3 ;






FX (x) = ..
 .



p1 + p 2 + · · · + p i if xi ≤ x < xi+1 ;







 ..



 .



 1 if x ≥ xk .


Remark 3.1. An analogy can be made with descriptive statistics. We find again the notion
of cumulative frequencies : ficum = f1 + f2 + · · · + fi .

Properties 3.1. A cumulative distribution function has these four properties :

1. 0 ≤ FX (x) ≤ 1, ∀x ∈ R ;

2
USTHB: ING INFO 1st year Probability and Statistics

2. FX (x) is an increasing function ;

3. FX (x) is a right-hand continuous function ;
4. lim FX (x) = 0 and lim FX (x) = 1.
x→−∞ x→+∞

Furthermore, and in general, we can show that for any discrete random variable X and
a, b ∈ R, we have :
• P (a < X ≤ b) = P (X ≤ b) − P (X ≤ a) = FX (b) − FX (a);
• P (a ≤ X ≤ b) = P (X ≤ b) − P (X < a) = FX (b) − FX (a) + P (X = a) ;
• P (a ≤ X < b) = P (X < b) − P (X < a) = FX (b) − FX (a) − P (X = b) + P (X = a);
• P (a < X < b) = P (X < b) − P (X ≤ a) = FX (b) − FX (a) + P (X = b).

3.3 Position and Dispersion Parameters

3.3.1 Mathematical Expectation

Definition 3.7 (Mathematical Expectation). The mathematical expectation of a discrete
random variable X, with domain X(Ω) = {x1 , x2 , . . . , xk }, is the number denoted E(X)
and defined by
Xk k
X
E(X) = xi p i = xi P (X = xi ).
i=1 i=1

Remark 3.2.
• The expectation E(X) is the (theoretical) mean of the random variable X ;
k
• An analogy can be made with the descriptive statistics : X = xi fi which is the
P
i=1
observed mean of X, calculated on a sample.

Properties 3.2. For any random variables X and Y and any real numbers a and b, we
have
• E(a) = a ;
• E(aX) = aE(X) ;
• E(aX + b) = aE(X) + b ;
• E(aX + bY ) = aE(X) + bE(Y ).

3.3.2 Variance and Standard Deviation

Definition 3.8 (Variance). The variance of a discrete random variable X with domain
X(Ω) = {x1 , x2 , . . . , xk } is the positive number denoted V (X) and defined by
X
V (X) = E(X − E(X))2 = (xi − E(X))2 P (X = xi ).
xi ∈X(Ω)

3
USTHB: ING INFO 1st year Probability and Statistics

Definition 3.9 (Standard Deviation). The standard deviation of the random variable X,
denoted σX (or σ(X)), is given by
p
σX = V (X).

Definition 3.10 (Koening formula). We show that

V (X) = E(X 2 ) − E 2 (X)

where
k
X
2
E(X ) = x2i P (X = xi ).
i=1

Properties 3.3. For any random variables X and Y and any real numbers a and b, we
have
• V (a) = 0 ;
• V (aX) = a2 V (X) ;
• V (aX + b) = a2 V (X) =⇒ σ(aX + b) = |a|σ(X) ;
• V (aX + bY ) = a2 V (X) + b2 V (Y ), if X and Y are independent.

3.4 Independence of two random variables

Definition 3.11 (Independence of two real random variables). Two real random variables
X and Y are said to be independent if and only if the probability distribution function of
the dual variable (X, Y ) is equal to the product of the probability distribution functions of
the marginal variables X and Y , that is :

FX,Y (x, y) = FX (x) × FY (y) ∀x ∈ X(Ω), and ∀y ∈ Y (Ω),

i.e.

P (X ≤ x, Y ≤ y) = P (X ≤ x) × P (Y ≤ y) ∀x ∈ X(Ω), and ∀y ∈ Y (Ω).

Definition 3.12. Two discrete random variables X and Y are said to be independent if
and only if

P (X = x, Y = y) = P (X = x) × P (Y = y) ∀x ∈ X(Ω), and ∀y ∈ Y (Ω).

4
USTHB: ING INFO 1st year Probability and Statistics

3.5 Example
A random experiment consists in drawing (simultaneously) 2 balls from an urn contai-
ning 2 red balls (R) and 3 white balls (W). Let X be the random variable associating to
each draw the number of white balls obtained.
1. What is the number of possible drawings ?
2. Determine the probability distribution function of X.
3. Calculate the cumulative distribution function.
4. Calculate the expectation, variance and standard deviation of X.
5. Assume Y = 5X − 3. Calculate E(Y ), V (Y ), and σ(Y ).
Solution :

1. Ω = {Combinations of 2 balls among 5}. Then, the number of possible draws is

equal to |Ω| = C52 = 10.
2. The probability distribution function of X :
(a) We have X(Ω) = {0, 1, 2}
C22 1
(b) P (X = 0)=P ("to have 2 R balls and 0 W balls") = C52
= 10
C21 C31 6
P (X = 1)=P ("to have 1 R balls and 1 W balls") = C52
= 10
C32 3
P (X = 2)=P ("to have 0 R balls and 2 W balls") = C52
= 10

P
xi 0 1 2
1 6 3
pi 10 10 10
1

This probability distribution function can be summarized by the following for-

mula :

C22−k C3k
P (X = k) = P ("to have 2 − k R balls and k W balls") = , ∀k ∈ {0, 1, 2}.
C52

3. The cumulative distribution function is given by





 0 if x < 0;




 1

if 0 ≤ x < 1;
10
FX (x) =
7
if 1 ≤ x < 2;


10






1 if x ≥ 2.



5
USTHB: ING INFO 1st year Probability and Statistics

4. Calculation of the expectation, the variance and the standard deviation of the
random variable X using the statistical table
P
xi 0 1 2
1 6 3
pi 10 10 10
1
6 6 12
xi p i 0 10 10
E(X) = 10
6 12 18
x2i pi 0 10 10
E(X 2 ) = 10

3
P 12
(a) E(X) = xi p i =
i=1 10
3 18
(b) V (x) = E(X 2 ) − E 2 (X) = x2i pi − E(X)2 = − 1.22 = 0.36
P
i=1 10
p √
(c) σ(X) = V (X) = 0.36 = 0.6

5. Calculation of the expectation, the variance and the standard deviation of the
random variable Y = 5X − 3
(a) E(Y ) = E(5X − 3) = 5E(X) − 3 = 3
(b) V (Y ) = V (5X − 3) = 52 V (Y ) = 9
√
(c) σ(Y ) = 9 = 3

6
Chapitre 4

Common Discrete Probability

Distributions

In the previous chapter, we saw that a random variable X defined on a domain X(Ω)
is associated with a function P (X = x), called the probability distribution function of
X. Obviously, the probability distribution functions existing in practice are extremely
numerous. However, some of them have received particular attention because of their
usefulness in practical applications. Therefore, they have been given special names.

4.1 Discrete Uniform Distribution

Definition 4.1 (Discrete Uniform Distribution). A random variable X is said to have a
uniform discrete distribution if

1. X(Ω) = {x1 , x2 , . . . , xk } ;
1
2. P (X = xi ) = for i ∈ {1, 2, . . . , k}.
|X(Ω)|
Therefore, we note
X ∼ U{x1 ,x2 ,...,xk } .

Description : This distribution models a random experiment with equiprobable out-

comes.

Properties 4.1. The random variable X ∼ U{x1 ,x2 ,...,xk } has an expectation and a variance
which are given by
Pk Pk
xi x2
E(X) = , and V (X) = i=1 i − E(X)2 .
i=1
|X(Ω)| |X(Ω)|

7
USTHB: ING INFO 1st year Probability and Statistics

Remark 4.1. In the special case of a random variable X with a discrete uniform distri-
bution on the set {1, . . . , n}, we write :

X ∼ U{1,2,...,n} .

Furthermore, we have :

n+1 n2 − 1
E(X) = , and V (X) = .
2 12

Example 4.1. Consider the experiment of rolling a fair six-sided die. Let X be the random
variable corresponding to the number of the revealed face. We have

1
X(Ω) = {1, 2, 3, 4, 5, 6} and P (X = i) = for i ∈ X(Ω).
6

Thus, we can write :

X ∼ U{1,2,3,4,5,6} .

Moreover, we have :
7 35
E(X) = , and V (X) = .
2 12

4.2 Bernoulli Distribution

Definition 4.2 (Bernoulli Distribution). A random variable X is said to have a Bernoulli
distribution with parameter p if
1. X(Ω) = {0, 1} ;
2. P (X = x) = px (1 − p)1−x for x ∈ {0, 1}. That is to say, P (X = 1) = p and
P (X = 0) = 1 − p.
Therefore, we note
X ∼ B(p).

Description : A Bernoulli trial is a random experiment whose only possible outcomes

are the two events {X = 1} and {X = 0}, called success and failure respectively.

Properties 4.2. The random variable X ∼ B(p) has an expectation and a variance which
are given by
E(X) = p, and V (X) = p(1 − p).

Example 4.2. In a fair coin toss, there are two distinct outcomes, "Heads" and "Tails".
We can represent "Heads" as 0 and "Tails" as 1. We have

1 1
P (X = 0) = P ("Heads") = and P (X = 1) = P (”T ails”) = .
2 2

8
USTHB: ING INFO 1st year Probability and Statistics

Thus, we can write  

1
X∼B .
2
This gives
1 1
E(X) = , and V (X) = .
2 4

4.3 Binomial Distribution

Definition 4.3 (Binomial Distribution). A random variable X is said to have a Binomial
distribution with parameters n and p if
1. X(Ω) = {0, 1, . . . , n} ;
2. P (X = x) = Cnx px (1 − p)n−x for x ∈ X(Ω).
Therefore, we note
X ∼ B(n, p).

Description : This distribution models a random experiment in which a Bernoulli trial of

parameter p is performed identically and independently n times, where p is the probability
of success.
A random variable X that has Binomial distribution with parameters n and p represent the
number of successes obtained in a series of independently repeated Bernoulli experiments
with probability p of success.

Properties 4.3. The random variable X ∼ B(n, p) has an expectation and a variance
which are given by
E(X) = np, and V (X) = np(1 − p).

Remark 4.2.
• Bernoulli distribution is a particular binomial distribution for which n = 1.
• X can be seen as the sum of n independent random variables Y1 , Y2 , . . . , Yn of
Bernoulli distribution of parameter p each :
n
X
X= Yi , whereYi ∼ B(n, p) for i = 1, 2, . . . , n.
i=1

Example 4.3. A well-balanced coin is tossed 10 times in succession and the obtained
results were observed. Let X be the random variable that counts the number of Heads
obtained in the sequence of 10 tosses. Then
 
1
X ∼ B 10, .
2

9
USTHB: ING INFO 1st year Probability and Statistics

Now, consider the event A : "obtain exactly 3 Heads". Then

 3  7
3 1 1 15
P (A) = P (X = 3) = C10 = .
2 2 128

Example 4.4. 8 balls are drawn successively from an urn containing 3 white balls and
7 black balls. Let X be the random variable that counts the number of white balls drawn.
Then  
3
X ∼ B 8, .
10
Thus, the probability that, for example, the three white balls are all drawn is :
 3  5
3 7
P (X = 3) = C83 = 0.254.
10 10

4.4 Geometric Distribution

Definition 4.4 (Geometric Distribution). A random variable X is said to have a Geo-
metric distribution with parameter p if
1. X(Ω) = N∗ ;
2. P (X = x) = p(1 − p)x−1 for x ∈ X(Ω).
Therefore, we note
X ∼ G(p).

Description :The geometric distribution models the rank of the first success by repeating
a Bernoulli trial (whose probability of success is p) identically and independently until to
infinity (theoretically).

Properties 4.4. The random variable X ∼ G(p) has an expectation and a variance which
are given by
1 1−p
E(X) = , and V (X) = .
p p2
Example 4.5. A fair die is rolled continuously until the number "six" is obtained. X is
the random variable representing the number of throws required to obtain "six". In this
1
case, the random variable X follows a geometric distribution with parameter , and we
6
write
1
X ∼ G( ).
6
Now, we need to calculate the probability that the first time we get "six" is on the 10th
draw :  9
1 1
P (X = 10) = 1− .
6 6

10
USTHB: ING INFO 1st year Probability and Statistics

Example 4.6. An urn contains 4 white balls and 6 black balls. We make random draws
with replacement, and note X the random variable that counts the number of draws requi-
red to obtain a white ball for the first time. We then have :

2
X ∼ G( ).
5

We now need to calculate the probability that the first time we get a white ball is on the
20th draw :  19
2 2
P (X = 20) = 1− .
5 5

4.5 Poisson Distribution

Definition 4.5 (Poisson Distribution). A random variable X is said to have a Poisson
distribution with parameter λ > 0 if

1. X(Ω) = N ;
λx
2. P (X = x) = e−λ for x ∈ X(Ω).
x!
Therefore, we note
X ∼ P(λ).

Description :Poisson distribution describes the probability of an event occurring during

a fixed time interval, when the probability of the occurrence of that event is very low.
This is the distribution that best models frequency (number of customers at a checkout
in a day, number of calls received at a call center, etc.).

Properties 4.5. The random variable X ∼ P(λ) has an expectation and a variance which
are given by
E(X) = V (X) = λ.

Approximation of the binomial distribution by the Poisson distribution

The binomial distribution depends on two parameters n and p. When n becomes large,
calculating its probabilities becomes very tedious. Provided that
 
n > 30 n > 50
or
np < 5 np < 0.1,

the binomial distribution B(n, p) can be approximated by the Poisson distribution with
parameter λ = np i.e. P (= np).

11
USTHB: ING INFO 1st year Probability and Statistics

Example 4.7. Let X be a random variable having a binomial distribution with parame-
ters n = 51 and p = 0.09 (B(51, 0.09)). Let Y be a random variable having a Poisson
distribution with parameter λ = np = 4.59 (P(4.59)).

k 0 1 2 3 4 5
P (X = k) 0.008 0.041 0.102 0.164 0.195 0.181
P (Y = k) 0.010 0.047 0.107 0.164 0.188 0.172

Approximation of the binomial distribution by a Poisson distribution gives probability

values that are almost identical, within 10−2 .

12