8 Kolmogorov: Ξ©, π, β
8.1 π΅ The formal frame
Probability becomes rigorous when three objects are specified:
\[ (\Omega, \mathcal{A}, \mathbb{P}) \]
These symbols are the skeleton of the theory.
Ξ©is the sample spaceπis the collection of eventsβis the probability measure
Together, they say:
- what may happen
- what may meaningfully be asked
- how likelihood is assigned
8.2 π° Ξ© β the space of possible outcomes
Ξ© is the universe of outcomes under the chosen model.
Examples:
- coin toss:
Ξ© = {Head, Tail} - die roll:
Ξ© = {1,2,3,4,5,6} - waiting time:
Ξ© = [0,\infty) - measurement:
Ξ© = \mathbb{R}
This is the first act of modeling: define the space of possible resolution.
8.3 βοΈ π β the events we are allowed to talk about
Not every subset is always manageable, especially in continuous spaces.
So Kolmogorov introduces a selected family of subsets:
\[ \mathcal{A} \subseteq \mathcal{P}(\Omega) \]
called a Ο-algebra.
A Ο-algebra is a collection of events such that:
Ξ© β π- if
Event β π, thenEvent^c β π - if
Event_1, Event_2, Event_3, \dots β π, then
\[ \bigcup_{n=1}^{\infty} Event_n \in \mathcal{A} \]
From these rules, intersections and differences also become allowed.
βοΈ This may seem technical, but the idea is simple:
πis the family of questions stable under logic and limits.
Probability needs not only possibilities, but a disciplined language of events.
8.4 β― β β probability as measure
Now comes β.
\[ \mathbb{P}: \mathcal{A} \to [0,1] \]
This means:
- each allowed event receives a number between
0and1
and that assignment must satisfy the Kolmogorov axioms:
8.4.1 1. Non-negativity
For every Event β π,
\[ \mathbb{P}(Event) \ge 0 \]
8.4.2 2. Normalization
The whole sample space has probability 1:
\[ \mathbb{P}(\Omega) = 1 \]
8.4.3 3. Countable additivity
If Event_1, Event_2, Event_3, \dots are pairwise disjoint, then
\[ \mathbb{P}\left( \bigcup_{n=1}^{\infty} Event_n \right) = \sum_{n=1}^{\infty} \mathbb{P}(Event_n) \]
This is the heart of the theory.
Probability respects decomposition.
8.5 π‘ Why this structure is beautiful
Kolmogorov does not tell us what reality βreally is.β
He does something subtler and more powerful.
He tells us what a coherent theory of uncertainty must look like.
A probability space is not a final metaphysics.
It is a formal contract between:
- possibility
- logic
- measure
That is why the structure feels inevitable.
Once one has:
- outcomes
- events
- stable composition of events
- numerical assignment
the axioms almost force themselves.
8.6 πͺ Example: a fair die
Let
\[ \Omega = \{1,2,3,4,5,6\} \]
Let π be all subsets of Ξ©.
Then define β by:
\[ \mathbb{P}(Event) = \frac{|Event|}{6} \]
for every event Event β Ξ©.
Then:
β({2,4,6}) = 3/6 = 1/2β({1}) = 1/6β(\Omega) = 1
This is the finite case.
In continuous spaces the idea remains, but counting is replaced by measure.
8.7 β οΈ Final lesson
The move from everyday chance to formal probability is the move from:
- vague possibility
to - structured possibility
The Kolmogorov framework gives probability its modern form because it joins:
- the ontology of outcomes:
Ξ© - the logic of events:
π - the arithmetic of uncertainty:
β
And with that, probability becomes ready for everything that follows:
- random variables
- distributions
- expectation
- convergence
- stochastic processes
All of them begin here.