What we mean by randomness
Probability begins with a deceptively simple idea: a situation whose outcome we cannot predict with certainty, but whose set of possible outcomes we can describe. We call this a random experiment — not because a lab is involved, but because, before it runs, more than one result is genuinely on the table. Flipping a coin, rolling a die, waiting to see how many emails arrive before lunch: each is a random experiment in this sense.
Honesty matters here. "Random" does not mean lawless or causeless. A tossed coin obeys physics perfectly; if we knew every force and angle precisely we could in principle predict the face. We call it random because, in practice, we lack that information and the outcome is exquisitely sensitive to details we cannot measure. Probability is the mathematics of reasoning carefully under exactly that kind of uncertainty — it is not a claim that the universe rolls dice.
The sample space: listing everything that could happen
The first move in any probability problem is to write down the stage. The sample space, usually written as the Greek letter Omega, is the set of all possible outcomes of the experiment — and crucially, the outcomes must be distinct and exhaustive: exactly one of them happens each time. Each single result inside it is an outcome (a sample point). For one die, Omega = {1, 2, 3, 4, 5, 6}. For one coin, Omega = {heads, tails}.
The sample space is a choice, not a fact handed to you, and choosing it well is half the battle. Suppose you toss two coins. If you only care how many heads appear, you might use Omega = {0, 1, 2}. But if you care about the order, the cleaner space is Omega = {HH, HT, TH, TT}, with four equally natural outcomes. The second description is richer: from it you can recover the first (one head means HT or TH), but not the reverse. A good rule of thumb is to choose the finest description that still answers your question — you can always group outcomes together later.
Events: the questions we can ask
An outcome is what actually happens; an event is a question we can answer once we see the outcome. Formally, an event is simply a subset of the sample space — a collection of outcomes we have grouped together because they share a property we care about. For a die, "the result is even" is the event {2, 4, 6}. "The result is at least 5" is {5, 6}. "The result is a 3" is the single-outcome event {3}, which we call an elementary event.
This subset picture is the single most useful mental image in elementary probability, because it quietly translates the words of everyday language into the operations of set theory. "A and B both happen" becomes the intersection A and B (the outcomes in both subsets). "A or B happens" becomes the union A or B (the outcomes in either). "A does not happen" becomes the complement of A (everything in Omega outside A). We say an event A occurs precisely when the realized outcome lands inside the subset A.
Two special events bookend the collection. The whole sample space Omega is the certain event — whatever happens, the outcome is somewhere in Omega, so it always occurs. The empty set is the impossible event — it contains no outcomes, so it can never occur. The full family of all events you are allowed to talk about, together with how they combine under and, or, and complement, is called the algebra of events, and tidying it up properly is the subject of the very next guide.
When events cannot happen together
Two events are mutually exclusive (or disjoint) when they share no outcomes — their intersection is the empty set, so they cannot both occur on the same run of the experiment. On one die, "the result is even" and "the result is 3" are mutually exclusive: no outcome is both. This is a structural fact about the subsets, visible just by looking at whether they overlap.
Mutual exclusivity matters because it is exactly the condition under which probabilities simply add. If A and B cannot both occur, then P(A or B) = P(A) + P(B), with no correction term. When they can overlap, you must subtract the double-counted middle — that is the inclusion-exclusion idea the next guide develops. For now, just hold the picture: disjoint subsets stack side by side, and their sizes add cleanly.
From sets to numbers: the first taste of probability
All of this set-talk is the skeleton; probability is the flesh we put on it. Probability is a rule that assigns to each event a number between 0 and 1, measuring how strongly we expect that event to occur. The honest, axiom-based definition of that rule is the job of guide 3 in this rung; here we only meet the friendliest special case, the one that gives the right intuition.
When the sample space is finite and we have a symmetry argument that every outcome is equally likely — a fair die, a balanced coin, a well-shuffled deck — the classical definition of probability says: the probability of an event is just the fraction of outcomes that belong to it. In symbols, P(A) = (number of outcomes in A) / (number of outcomes in Omega). Probability becomes counting, which is why the very next rung of this ladder is devoted to counting cleverly.
Omega = {1,2,3,4,5,6} (a fair die, 6 outcomes)
A = "even" = {2,4,6} P(A) = 3/6 = 1/2
B = "at least 5"= {5,6} P(B) = 2/6 = 1/3
A and B = {6} P(A and B) = 1/6
A or B = {2,4,5,6} P(A or B) = 4/6 = 2/3
check: P(A)+P(B)-P(A and B) = 3/6 + 2/6 - 1/6 = 4/6 (they overlap at 6)One warning to carry forward: the classical recipe only works when "equally likely" is genuinely justified, usually by symmetry. It is silent about a bent coin, a loaded die, or "will it rain tomorrow," where no two outcomes are interchangeable. Treating outcomes as equally likely just because you listed them is a classic error — the choice of sample space can make outcomes look uniform when they are not. The deeper question of what a probability number even means when symmetry fails is the subject of the last guide in this rung.