BASIC PROBABILITY CONCEPTS

TABLE OF CONTENTS

Introduction 1

Basic Probability Concepts 1

Acronyms 6

Bibliography 6

INTRODUCTION
The objective of this reading is to clarify any queries regarding probability theory that may arise during
the course.

BASIC PROBABILITY CONCEPTS

“Insanity is doing the same thing over and over and expecting different results.” This phrase is often
attributed to Albert Einstein, although he actually never said it. As a genius, he undoubtedly knew that,
although many of the processes we perform in our daily lives are deterministic, i.e., if we repeat them under
the same conditions we will get the same results, many other processes are eminently random by nature,
even though we may repeat them in exactly the same way, their results will vary. For example, that is the
case with rolling a die: the result is not always the same.

Probability theory is a branch of mathematics that analyzes the behavior of random phenomena through
probability calculations (Montero Alonso, 2007). We have already provided one example of a random
phenomenon; another one would be flipping a coin to see whether you get heads or tails. The controlled
repetition of a random phenomenon is known as a random experiment. Each of these possible outcomes
for a random experiment is called an elementary event (e), and all together, they make up the sample
space (S). For example, the sample space for flipping a coin is:
S={heads , tails}

1
In the case of a die, each of the 6 possible outcomes you can get when you roll the die is a distinct elementary
event. On the other hand, an event (A) refers to the set made up of one or more elementary events, such
as rolling the die and getting an outcome greater than 3 or rolling it and getting a prime number.

Events that cannot occur simultaneously are called incompatible or mutually exclusive events, such
as expecting to get a 1 and a 2 at the same time when you roll a die. In addition to the concept of
incompatibility, we have complementarity. Two events are complementary when the nonoccurrence of one
inevitably implies the occurrence of the other. For example, if you flip a coin and you do not get heads, then
you will necessarily get tails, so the tails event is complementary to the heads event and vice versa.

Now that we have defined the concepts of random experiment, elementary event, sample space, event, and
incompatible and complementary events, we can talk about the probability of an event. Laplace stated
that the probability of an event (A) is the ratio of the number of favorable outcomes (elementary
events that A comprises) to the total number of possible outcomes (elementary events in sample space
S), supposing that each outcome has the same probability of occurring, as in the case of the die or coin.

P(A) = number of favorable outcomes

number of possible outcomes

Therefore, the probability of flipping a coin and getting heads is:

P(A) = 1 = 0.50
2

While the probability of getting more than 4 (i.e., 5 or 6) when rolling a die is:

P(A)= 2 = 1 ≈ 0.33
6 3

At this point, we could ask: What if instead of flipping one coin we flipped two; what is the probability of getting
heads on both? What we get on one coin does not influence the other, so these two events are said to be
independent and, according to probability theory, the probability of an event conditional to another event of
which it is independent (event A would be getting heads on the first coin and event B would be also getting heads
on the second coin) is equal to the product of each of them if they occurred separately.

P(A B) = P(A)•P(B) = P(B A)

U U

Therefore, the probability of getting heads on one coin, conditional on the other coin having gotten heads as
well, or in other words, the probability that both coins will get heads is:

P(A B) = P(A)•P(B) = 1 • 1 = 1 = 0.25

U

2 2

2
To continue, we must introduce the concept of random variables. Given a random experiment with its
corresponding probability space1 (S, Ą, P), a random variable (X) is an application that assigns a real
value (x) to each elementary event (e). Expressed mathematically, X:S -> R . Basically, the random variable
associates a number with each possible outcome of a random experiment. Going back to the two coins,
we could use a random variable to assign the number of heads obtained to the coin flip. Therefore, and
since the sample space is S={tails and tails , tails and heads, heads and tails, heads and heads}, the random
variable would be as follows:

X(tails and tails)=0

X(tails and heads)=1
X(heads and tails)=1
X(heads and heads)=2

In this example, the random variable is discrete because it only adopts certain isolated values within a
given range. However, if the random variable were able to adopt any value in a certain range, it would be
continuous.

An example of a continuous random variable could be the millimeters of accumulated rain in a year in a
given region.

After defining a random variable, the next step is to define its distribution function (or probability distribution
function). This function returns how probable it is that the random variable does not exceed a certain value.
In mathematical terms, this would be: FX(x)=P(X≤x). Therefore, the random variable distribution function
for flipping two coins would be:

Fx(x<0) = P(X≤x<0) ≤ P(X<0) = 0

Fx(x<0) = P(X≤0) = P(tails and tails) = 1 = 0.25

Fx (x=1) = P(X≤1) = P(X=0) + P(X=1) =

= P(tails and tails) + (P(tails and heads) + P(heads and tails)) = 1 + 2 = 3 =0.75

Fx (x=2) = P(X≤2) = P(X=0) + P(X=1) + P(X=2)=

=P(tails and tails) + (P(tails and heads) + P(heads and tails)) + P(heads and heads) = 1
+2+1=4=1

Fx (x>2) = P(X≤x>2) = P(X≤2) = 1

In the first case, this probability is null because the random variable has been defined in such a way that
its value cannot be less than 0. In the last case, this probability is 1 because the maximum value that the
random variable can adopt is 2, so the probability that the random variable is less than a number greater
than 2 is total. Figure 1 represents this distribution function graphically.

1Un espacio probabilístico lo integran el espacio muestral, E, de los posibles resultados del experimento aleatorio; Ą, que son
todos los sucesos elementales y no elementales que pueden darse; y la función de probabilidad , P, que a cada suceso posible le
asigna su probabilidad correspondiente.

3
 1

0,75

FX(X) 0,5

0,25

-2 -1 0 1 2 3 4

Figure 1. Distribution Function Associated with Flipping Two Coins

The graphical representation of the distribution function enables us to observe three conditions that
characterize it:

The distribution function is monotonically increasing. In other words, when x

increases, FX(x) also increases or stays the same, but it never decreases.

The distribution function is continuous to the right. Setting aside mathematical

rigor for a moment, in this case this means that when there is a vertical jump
for a certain value of x, FX(x) always adopts the greatest of its two possible
values.

In addition to being monotonically increasing, the distribution function is

delimited by 0 and 1, and it complies with
lim Fx (x) = 0 y lim Fx (x) = 1
X →∞ X →∞

As we have seen, the distribution function returns the nonexceedance probability of a certain value. If
we want to know the probability that the random variable will fall within a certain range, we just need to
calculate the difference between its value for the upper end and lower end of that range. For example, the
probability that we will get heads on several coins is:

P (heads,heads) = P(x=2) = Fx(x=2) - Fx(x=1) = P(x≤2) - P(x≤1) = 1 - 3 = 1 = 0.25

4 4

The concept of nonexceedance probability is widely used in disaster risk management for characterizing, for
example, the probability of floods of different magnitudes in a waterway.

A very common example of a distribution function is binomial distribution.

A binomial distribution is a discrete probability distribution that returns the number of favorable outcomes
obtained by repeating a random experiment n times with a probability, p, of success. To indicate that a
random variable follows a binomial distribution, the convention is to write:

X→B(n,p)

4
To calculate the probability that a certain number of successes will occur, we use the expression of its
probability function, which is:

P(X=x) = n! px (1-p)(n-x)
x! (n-x)!

In this function, x is the number of favorable outcomes, so it can only take integer values between 0 and
n. With that expression, we can, for example, calculate the probability that we will get a certain number of
heads when we flip the coin 20 times by considering n=20 and p=0.5. Figure 2 shows the values obtained
in this example.

0,15

P(X=x) 0,1

0,05

0 5 10 15 20

x
Figure 2. Binomial Distribution for n=20 and p=0.5

As an example, binomial distribution can be used to analyze the probability of failure of, for example, the
gates of the spillway in a flood control dam (iPresas Risk Analysis , 2019) or a valve in a drinking water
supply network.

Before completing this brief review of the basic concepts of probability theory, we must talk about normal
or Gauss distribution. This is the most famous continuous distribution because it allows for a faithful
mathematical representation of many distributions present in nature. We can indicate that a continuous
random variable is modeled using a normal distribution using the following formula:

X → N(μ,σ)

In this formula, μ is the mean value of the distribution and σ a parameter known as standard deviation,
which provides information on how far the values of the distribution are separated from the mean value and
which only adopts values greater than 0. Its probability density function (the equivalent in a continuous
distribution to the probability function for discrete variables) is:

(x-μ)2
-2σ2
f(x)= 1 e
σ√2π

5
When the mean of the normal distribution is zero and its standard deviation equals the unit, it is known as the
standard normal distribution. Figure 3 represents the density function of a standard normal distribution,
i.e., one whose parameters are μ=0 y σ=1:

0,5

0,4

F(x) 0,3

0,2

0,1

-4 -3 -2 -1 0 1 2 3 4

Figure 3. Density Function of a Normal Distribution with μ=0 and μ=0 y σ=1.

Many biological, psychological, or social phenomena or variables have a normal distribution.

For example, the most common probability distributions in hydrology are normal, log-normal, Gumbel,
log-Gumbel, and Pearson type III (Ven Te Chow, 1988). In the case of seismic analysis, the magnitude and
intensity of earthquakes are usually determined by means of extreme value distribution functions, such as
Gumbel or Poisson distributions.

ACRÓNIMOS

P Probability

S Sample Space

X Random Variable

​BIBLIOGRAPHY

​iPresas Risk Analysis. 2019. Manual de usuario del software iPresas. Valencia (Spain): iPresas.

Montero Alonso, M. Á. 2007. Apuntes de Estadística II. Granada (Spain): Ed. Vicerrectorado de Planificación,
Calidad y Evaluación Docente, Universidad de Granada.

Ven Te Chow, D. R. 1988. Applied Hydrology. McGraw Hill.