Understanding Probability Spaces

Hello! Welcome back to our course.

Introduction

In our last lesson, we explored the concept of a σ-algebra ($\mathcal{F}$), establishing it as the necessary structure for defining a consistent set of "events" in a probability model. We saw that for uncountable sample spaces, we can't assign a probability to every conceivable subset without running into contradictions. The σ-algebra provides the solution by specifying a well-behaved collection of subsets—the measurable sets—that we can work with.

Today, we will build directly on that foundation to construct the central object of modern probability theory: the probability space. The learning outcome for this lesson is to define a probability space as the triple $(\Omega ,\mathcal{F},P)$ and to explain the role of each of its three components.

By the end of this lesson, you will understand how these three elements—the sample space, the σ-algebra of events, and the probability measure—fit together to provide the rigorous framework that underpins the statistical, econometric, and financial models you have encountered in your career.

The Probability Triple: $(\Omega ,\mathcal{F},P)$

A probability space is the mathematical formalization of a random process. It consists of three components, often called the "probability triple".

As the reading states, the three components are:

1. $\Omega$: The Sample Space
2. $\mathcal{F}$: The σ-algebra of Events
3. $P$: The Probability Measure

Let's examine each one in detail.

1. The Sample Space ($\Omega$)

The sample space, denoted by the Greek letter Omega ($\Omega$), is the set of all possible elementary outcomes of a random experiment. This concept should be familiar from your prior studies in probability and statistics.

• Definition: $\Omega$ is a non-empty set containing every possible result of the phenomenon being modeled.

For example:

• Single Coin Toss: $\Omega =\{H,T\}$ (a finite set)
• Number of emails arriving in an hour: $\Omega =\{0,1,2,\dots \}$ (a countably infinite set)
• A stock's return over one day: $\Omega =\mathbb{R}$ (an uncountably infinite set)

The sample space simply defines the "universe" of what can happen. It contains no information about the likelihood of any particular outcome.

2. The σ-algebra of Events ($\mathcal{F}$)

This is the component we focused on in the last lesson. The σ-algebra $\mathcal{F}$ (sometimes denoted with a script 'A' or 'F') is a collection of subsets of $\Omega$. These subsets are called events.

• Role: $\mathcal{F}$ specifies which combinations of outcomes we are able and willing to assign a probability to. It defines the "questions" we can ask about the experiment.

As a brief recap, a collection of subsets $\mathcal{F}$ is a σ-algebra if:

1. $\Omega \in \mathcal{F}$ (The certain event is an event).
2. If $A\in \mathcal{F}$, then its complement $A^c\in \mathcal{F}$ (If we can ask about an event, we can ask about it not happening).
3. If $A_1,A_2,\cdots \in \mathcal{F}$, then their countable union ${\bigcup }_{i=1}^{\infty }{A}_i\in \mathcal{F}$ (If we can ask about a series of events, we can ask about at least one of them happening).

For a dice roll where $\Omega =\{1,2,3,4,5,6\}$, the event "the roll is even" is the set $\{2,4,6\}$. This set would be an element of $\mathcal{F}$. The σ-algebra ensures that if $\{2,4,6\}$ is a valid event, then so is its complement, $\{1,3,5\}$ ("the roll is odd"), and their union, $\Omega$.

3. The Probability Measure ($P$)

The final component, the probability measure $P$, is the function that assigns a specific probability (a number between 0 and 1) to each event in the σ-algebra $\mathcal{F}$.

• Role: $P$ quantifies the likelihood of each event defined in $\mathcal{F}$. It is a function with the signature $P:\mathcal{F}\to [0,1]$.

For $P$ to be a valid probability measure, it must satisfy three axioms, famously laid out by Andrey Kolmogorov.

To summarize the axioms, a function $P$ is a probability measure if:

1. Non-negativity: For any event $A\in \mathcal{F}$, $P(A)\ge 0$.
2. Normalization: $P(\Omega )=1$. The probability of the entire sample space (the "sure event") is 1.
3. Countable Additivity: For any countable collection of pairwise disjoint events $A_1,A_2,\dots$ in $\mathcal{F}$ (meaning $A_i\cap A_j=\varnothing$ for $i\ne j$), the following holds: $$P\left(\bigcup _{i=1}^{\infty }{A}_i\right)=\sum _{i=1}^{\infty }P(A_i)$$

The first two axioms are straightforward. The third, countable additivity, is the cornerstone of measure-theoretic probability. In introductory courses, you likely used finite additivity ($P(A\cup B)=P(A)+P(B)$ for disjoint A and B). Countable additivity extends this to countably infinite sequences of disjoint events. This is not just a technicality; it is essential for proving foundational results in statistics and finance, such as the Laws of Large Numbers and theorems in stochastic calculus, which often involve limits of sequences of events.

Putting It All Together

The probability space $(\Omega ,\mathcal{F},P)$ provides a complete, self-consistent model of a random phenomenon.

• $\Omega$ tells us what can happen.
• $\mathcal{F}$ tells us what events we can measure the probability of.
• $P$ tells us how likely each of those events is.

The following short video provides a great summary of how a general measure space is interpreted as a probability space, using a simple coin flip as an example.

Example: A Single Stock Price

Let's connect this to a familiar domain: modeling the price of a stock, say AAPL, at the close of the next trading day.

1. Sample Space $\Omega$: The set of all possible prices. We could model this as the set of non-negative real numbers, $\Omega =[0,\infty )$. This is an uncountably infinite set.

2. σ-algebra $\mathcal{F}$: Since $\Omega$ is uncountable, we cannot use the power set. We need a σ-algebra that contains the events we care about but excludes pathological sets. The standard choice is the Borel σ-algebra, $\mathcal{B}([0,\infty ))$. This is the smallest σ-algebra containing all intervals of the form $[a,b]$. It allows us to ask questions like:

   • What is the probability the price is below $200? (Event: $[0,200)$)
   • What is the probability the price is between $210 and $215? (Event: $[210,215]$)
   • What is the probability the price is exactly $212.34? (Event: $\{212.34\}$)

3. Probability Measure $P$: This is the function that assigns probabilities to these Borel sets. In practice, this measure is defined by a probability distribution. For instance, in the Black-Scholes model, the stock price is assumed to follow a log-normal distribution. The probability measure $P$ would then be the one corresponding to that specific log-normal distribution. For an event $A=[a,b]$, the probability would be calculated via the integral of the probability density function (PDF):

   $$P(A)=P([a,b])={\int }_a^{b}f(x)dx$$

   where $f(x)$ is the log-normal PDF. This integral is a Lebesgue integral, a topic we will touch on later.

This example shows how the abstract triple $(\Omega ,\mathcal{F},P)$ provides the formal underpinning for the practical models you've used in finance. The model implicitly assumes this structure to be well-defined.

Conclusion

In this lesson, we have formally defined the probability space as the triple $(\Omega ,\mathcal{F},P)$. Let's recap the role of each component:

• Sample Space ($\Omega$): The set of all possible elementary outcomes. It defines the scope of the random experiment.
• σ-algebra ($\mathcal{F}$): A collection of subsets of $\Omega$ called events. It constitutes the set of all "reasonable questions" one can ask about the outcome, and serves as the domain for the probability measure.
• Probability Measure ($P$): A function that maps each event in $\mathcal{F}$ to a value in $[0,1]$, satisfying the axioms of non-negativity, normalization, and, crucially, countable additivity.

You now see how the σ-algebra, which might have seemed like an abstract constraint, is a vital part of a complete system. Without it, the probability measure $P$ would not have a well-defined domain for uncountable sample spaces, and the entire structure of modern probability would collapse.

Preview of the Next Lesson:
With the probability space firmly established, our next step is to understand how we get from abstract outcomes in $\Omega$ to the numerical values we work with in statistics (e.g., the number on a die, the return of a stock). This is achieved through a random variable, which in measure theory is formally defined as a measurable function. In the next lesson, we will explore what a measurable function is and why this formal definition is so important.