Understanding probability: definitions and core rules

Probability gives a numerical way to describe how likely an event is. A Bureau of Meteorology forecast might say there is a 40 per cent chance of rain in Melbourne tomorrow. That percentage refers to rain within the forecast area during the stated period; it does not mean rain will fall for 40 per cent of the day, and it does not guarantee whether a drop reaches the ground.

At its simplest, every probability problem has three parts: a random experiment, a sample space containing all possible outcomes, and an event that selects some of those outcomes. Before an AFL coin toss at the Melbourne Cricket Ground, the experiment is the toss, the sample space is {Heads, Tails}, and the event might be Heads. For a fair toss, each outcome has probability (1/2).

Calculations can be theoretical or empirical. Theoretical probability uses the structure of an experiment, such as counting the six faces of a die. Empirical probability comes from observed results, such as recording which team wins 20 rounds of a local rugby competition. A good method keeps these ideas distinct and states any assumptions before reaching a number.

What probability actually measures

A random experiment is any activity with an uncertain result that can be repeated under stated conditions. The sample space is the complete set of possible outcomes, while an event is any subset of interest. Drawing a numbered ball from a cup, rolling two dice or choosing a respondent for a survey all fit this framework.

Probabilities range from 0 to 1, or 0 to 100 per cent. Zero means an event cannot occur under the model; 1 means it must occur. A probability of 0.25 says one quarter of the equally likely cases are favourable. It is a statement within a defined model, not a guarantee about the future.

Australian speech often uses chance, likelihood and odds loosely. In maths, odds express a comparison between favourable and unfavourable outcomes, while probability expresses the favourable share of all possible outcomes. If someone says a result is “not odds”, that is informal judgement, not a calculated probability.

Building a sample space and calculating a value

Start by listing outcomes without leaving gaps or counting one outcome twice. For two fair six-sided dice, write ordered pairs such as (1,1) and (1,2), giving 36 outcomes. If the event is a total of 7, there are six favourable pairs, so (P(\text{total 7})=6/36=1/6). Order matters because (2,5) and (5,2) are different rolls.

The shortcut (P(A)=\frac{\text{favourable outcomes}}{\text{equally likely outcomes}}) works only when all outcomes in the sample space are equally likely. A biased spinner, unequal lottery ticket sales or a game with different scoring chances cannot be handled by counting alone. Such cases need weights, careful data or a different method.

For empirical probability, divide successful observations by the total number of observations. If a coin lands heads 7 times in 20 tosses, the relative frequency is (7/20=0.35). More trials can improve an estimate under stable conditions, but they do not guarantee the result will settle near one half if the coin is actually biased.

Combining events with the right rule

Use the addition rule when an event combines two possibilities with “or”. If the events are mutually exclusive, meaning they cannot occur together, then (P(A\cup B)=P(A)+P(B)). For example, a single raffle ticket can carry either a blue marker or a green marker, but not both. If overlap is possible, subtract it: (P(A\cup B)=P(A)+P(B)-P(A\cap B)).

For combinations involving “and”, the multiplication rule is (P(A\cap B)=P(A)P(B\mid A)). The vertical bar means “given”. In a cup with four red and two blue tokens, (P(\text{blue, then blue without replacement})=2/6\times1/5=1/15). If A and B are independent, (P(B\mid A)=P(B)), so the product becomes (P(A)P(B)).

Conditional probability reverses the direction of attention: (P(A\mid B)=P(A\cap B)/P(B)). It asks how likely A is within the restricted group B. Interpretation also depends on context, so figures should not be detached from the community or conditions that produced them; this reflection on cultural insight is a useful reminder to read evidence in context.

Practical checks for Australian learners

Everyday examples need the same care as textbook examples. A rain forecast, an AFL tipping competition and a TAB price all use uncertainty, but they describe different models. Newsagents also sell lottery products under different names across states, including Oz Lotteries and Tatts products, and prize-sharing chances can change with ticket sales.

Check the setup before calculating:

The wording often signals the operation. “Either” or “or” suggests a union, while “and” may require an intersection. Conditional wording such as “given that” changes the relevant sample space.

Check the rule and arithmetic:

After calculating, ask whether the answer lies between 0 and 1 and whether the event matches the question. For a one-in-five chance, the probability is 0.2, the percentage is 20 per cent and the odds against are 4 to 1. Keeping these forms separate prevents a common source of errors.

A repeatable method for solving the next problem

A reliable solution begins with language. Rewrite the question in your own words, identify the random experiment and define the event. Then construct the sample space, preferably as a table, tree or ordered list. This step often reveals missing cases before algebra or a calculator is needed.

Next, choose the rule that matches the event structure: complement for “not”, addition for “or”, multiplication or conditional probability for “and”, and independence only when justified. A diagram or online homework walkthrough can help verify the setup, but the final explanation should show why each step follows.

Finally, substitute carefully, simplify the fraction or decimal, and interpret the result in context. Check whether it is plausible and label it as a probability, percentage or odds ratio. The next concrete step is one practice question, a one-sentence definition of its random experiment, and a completed sample space before any formula is chosen.