Example of a discrete random variable using the measure-theoretic definition

measurable-functionsmeasure-theoryprobability theoryrandom variables

Context

I'm trying to wrap my head around the measure-theoretic definition of a random variable. I put forward this example in the hopes that someone can verify that my understanding is correct, or else indicate how it is incorrect.

I have done a non-exhaustive search on this site. I have found 1 helpful, but it still does not give me full comprehension.

Problem

In the figure below, I offer a circular wheel with uniform density. The experiment is to spin the wheel and record what color is aligned with the rightmost vertice of the black triangle.

An experiment consists of spinning the wheel and seeing what color aligns with the triangle

By $\Omega$, I denote the sample space, which I enumerate with three outcomes as
$$\Omega = \left\{\text{red}, \text{blue}, \text{green} \right\}.$$

By $\mathcal{F}$, I denote the $\sigma$-algebra, which I enumerate with $2^3$ events as
$$\mathcal{F} = \left\{\varnothing, \left\{\text{red}\right\}, \left\{\text{blue}\right\}, \left\{\text{green}\right\}, \left\{\text{red}, \text{green}\right\}, \left\{\text{red}, \text{blue}\right\}, \left\{\text{green}, \text{blue}\right\}, \left\{\text{red}, \text{green}, \text{blue}\right\} \right\}.$$

By $P$, I denote the probability measure, which I enumerate
\begin{align}
P(\varnothing)
&=
0,
\\
P(\left\{\text{red}\right\})
&=
1/6,
\\
P(\left\{\text{blue}\right\})
&=
3/6,
\\
P(\left\{\text{green}\right\})
&=
2/6,
\\
P(\left\{\text{red}, \text{blue}\right\})
&=
4/6,
\\
P(\left\{\text{red}, \text{green}\right\})
&=
3/6,
\\
P(\left\{\text{green}, \text{blue}\right\})
&=
5/6,~\text{and}
\\
P(\left\{\text{red}, \text{green}, \text{blue}\right\} )
&=
6/6.
\end{align}

By the triplet $(\Omega ,{\mathcal {F}},P)$, I denote the probability space.

By $E$, I denote the set $\left\{1,2,3\right\}.$

By $\mathcal{E}$, I denote a $\sigma$-algebra that I give as
$$\mathcal{E} = \left\{\varnothing,\{1\},\{2,3\},\{2\},\{1,3\},\{3\},\{1,2\}, \{1,2,3\}\right\}.$$

By the tuple $(E ,{\mathcal {E}})$, I denote the measurable space.

By the measurable function $ X\colon \Omega \to E$, I denote an $(E,{\mathcal {E}})$-valued random variable that I define as
\begin{align}
X(\omega) =
\begin{cases}
1,&~\text{if}~\omega = \text{red};
\\
2,&~\text{if}~\omega = \text{blue};~\text{or}
\\
3,&~\text{if}~\omega = \text{green}.
\end{cases}
\end{align}

For every subset $B\in {\mathcal {E}}$, I denote its preimage as $X^{-1}(B)$, where $X^{-1}(B)=\{\omega :X(\omega )\in B\} $.

I want to check that for every subset in $\mathcal {E} $, its preimage $X^{-1}(B)$ is in $\mathcal {F}$.

\begin{align}
X^{-1}(\varnothing)
&= \{\omega : X(\omega)\in \varnothing\} = \varnothing
\\
X^{-1}(\{1\})
&= \{\omega : X(\omega)\in \{1\}\} = \{\text{red} \}
\\
X^{-1}(\{2\})
&= \{\omega : X(\omega)\in \{2\}\} = \{\text{blue} \}
\\
X^{-1}(\{3\})
&= \{\omega : X(\omega)\in \{3\}\} = \{\text{green} \}
\\
X^{-1}(\{1,2\})
&= \{\omega : X(\omega)\in \{1,2\}\} = \{\text{red},\text{blue}\}
\\
X^{-1}(\{1,3\})
&= \{\omega : X(\omega)\in \{1,3\}\} = \{\text{red},\text{green} \}
\\
X^{-1}(\{2,3\})
&= \{\omega : X(\omega)\in \{2,3\}\} = \{\text{blue},\text{green} \}
\\
X^{-1}(\{1,2,3\})
&= \{\omega : X(\omega)\in \{1,2,3\}\} = \{\text{red},\text{blue},\text{green} \}
\end{align}

Biibliography

1 Intuitively, how should I think of Measurable Functions?

Best Answer

I'm pretty sure that my example is verified as correct. Since I could not find a lucid example on-line, I am leaving this question up for pedagogical purposes.

Related Question