7 Finite and Infinite Sample Spaces

7.1 🏵 Some worlds can be counted, others only approached

Not all sample spaces look alike.

Some are finite:

\[ \Omega = \{Head, Tail\} \]

Some are countably infinite:

\[ \Omega = \mathbb{N} \]

Some are continuous:

\[ \Omega = \mathbb{R} \]

This matters because probability behaves differently in each case.

Finite sample spaces are the most intuitive.

A die, a card draw, a finite survey response set: these all give a small visible universe of possibilities.

In finite spaces:

These are the ideal starting point for intuition.

Sometimes the possible outcomes never end, yet they can still be counted one by one.

Examples:

These live naturally in spaces like:

\[ \Omega = \{0,1,2,3,\dots\} \]

This is already a step beyond pure enumeration, because the space is endless, but still discrete.

Then there are outcomes that vary across intervals.

Examples:

Here the natural model is not ℕ, but ℝ, or some interval inside it.

This changes everything.

In a continuous space, one cannot usually assign a positive probability to each individual point in the same way as in a finite space.

Instead, probability lives on sets of values:

⚜️ This is a crucial transition:

in finite spaces, probability often feels like counting.
in continuous spaces, probability becomes measure.

Reality may be continuous while instruments are discrete.
Or reality may be modeled as discrete while the underlying process is continuous.

A thermometer prints a finite decimal expansion.
The model may still use real numbers.

So the choice between finite and infinite spaces is not only about ontology.
It is also about resolution, convenience, and explanatory power.

Once the sample space becomes infinite, intuitive “just assign values to outcomes” reasoning begins to fail.

One needs a more careful theory of:

This is the road to Kolmogorov.

The infinite is where probability stops being only common sense and becomes a true mathematical architecture.