the probability space

A probability space is the complete package that specifies a random situation. It has three parts, often written as the triple (Omega, F, P). Omega is the sample space, the set of all possible outcomes. F is the collection of events — the subsets of Omega to which we are willing to attach probabilities. And P is the probability measure, the rule that assigns each event in F a number obeying Kolmogorov's axioms. Together these answer three questions: what can happen, which questions we may ask, and how likely each answer is.

An everyday picture: setting up a probability space is like fully describing a game before anyone plays. Omega lists every way the game can turn out; F is the menu of bets you are allowed to place (each bet is really a yes-or-no question, i.e. an event); and P is the agreed odds on every allowed bet, set so the odds never contradict each other. For a single die the package is plain — Omega = {1,...,6}, F is all subsets of those faces, and P spreads probability 1/6 on each — but writing it as a triple makes explicit the machinery we usually leave unspoken.

Why bother naming the second part, F, at all? For finite or countable Omega you can let F be every possible subset and never think about it again. The subtlety appears only on continuous spaces, where it turns out you cannot consistently assign a probability to literally every subset; you must restrict F to a well-behaved family of events (technically a sigma-algebra). At this informal level the point is just that a probability space bundles outcomes, events, and a consistent assignment of chances into one self-contained object — the formal construction of F belongs to the measure-theory field.