# Definition:Random Variable/Discrete

## Contents

## Definition

Let $\mathcal E$ be an experiment with a probability space $\left({\Omega, \Sigma, \Pr}\right)$.

A **discrete random variable** on $\left({\Omega, \Sigma, \Pr}\right)$ is a mapping $X: \Omega \to \R$ such that:

- $(1): \quad$ The image of $X$ is a countable subset of $\R$
- $(2): \quad$ $\forall x \in \R: \left\{{\omega \in \Omega: X \left({\omega}\right) = x}\right\} \in \Sigma$

Alternatively (and meaning exactly the same thing), the second condition can be written as:

- $(2)': \quad$ $\forall x \in \R: X^{-1} \left({x}\right) \in \Sigma$

where $X^{-1} \left({x}\right)$ denotes the preimage of $x$.

Note that if $x \in \R$ is not the image of any elementary event $\omega$, then $X^{-1} \left({x}\right) = \varnothing$ and of course by definition of event space as a sigma-algebra, $\varnothing \in \Sigma$.

Note that a discrete random variable also fulfils the conditions for it to be a random variable.

## Also known as

The image $\operatorname{Im} \left({X}\right)$ of $X$ is often denoted $\Omega_X$.

## Discussion

The meaning of condition $(2)$ in this context can be explained as follows:

Suppose $X$ is a discrete random variable. Then it takes values in $\R$. But we don't know what the actual value of $X$ is going to be, since the outcome of $\mathcal E$ involves chance.

What we *can* do, though, is determine the probability that $X$ takes any particular value $x$.

To do this, we note that $X$ has the value $x$ if and only if the outcome of $\mathcal E$ lies in the subset of $\Omega$ which is mapped to $x$.

But for any such element $x$ of the image of $X$, the preimage of $x$ is an element of $\Sigma$.

## Sources

- 1986: Geoffrey Grimmett and Dominic Welsh:
*Probability: An Introduction*... (previous) ... (next): $\S 2.1$: Probability mass functions