What is probability? Probability is a mathematical way of describing how likely an event is to happen. It ranges from 0 to 1, or from 0% to 100%. A probability of 0 means an event is impossible under the model; a probability of 1 means it is certain. Values between them express different degrees of likelihood.
Probability is not a promise about what will happen next. A fair coin has a probability of one-half of landing heads on a single toss, but it can land tails several times in a row. Probability describes a model of uncertainty over many possible outcomes. Good reasoning asks what the possible outcomes are, whether they are equally likely, and what assumptions the calculation depends on.
Quick notes: Probability of an event = number of favourable outcomes ÷ number of equally likely outcomes, when the outcomes are finite and equally likely. The complement rule is P(not A) = 1 − P(A). For mutually exclusive events, probabilities add. For independent events, probabilities multiply. Conditional probability changes the calculation when information is known. Probability is not the same as odds, frequency or certainty.
What probability measures
Probability assigns a number to an event within a specified model. An event is a set of possible outcomes. For a standard six-sided fair die, the sample space is {1, 2, 3, 4, 5, 6}. The event ‘roll an even number’ contains {2, 4, 6}. If all six faces are equally likely, three of the six outcomes are favourable.
P(even number) = 3/6 = 1/2 = 0.5 = 50%
The result depends on the assumptions. If the die is biased or weighted, its six faces may not be equally likely. The sample space remains useful, but the simple count of favourable outcomes is no longer sufficient unless each elementary outcome has equal probability.
The sample space: start by listing what can happen
A sample space is the complete set of possible outcomes for an experiment. Before calculating probability, define the experiment and list the relevant outcomes. A coin toss has two outcomes if the model is heads or tails. A two-coin experiment has four ordered outcomes: HH, HT, TH and TT.
A frequent error is to count outcomes inconsistently. If order matters in the experiment, HT and TH are different outcomes. If order does not matter for the question, they may belong to the same event, but they still come from distinct underlying outcomes in the sample space.
Classical probability: equally likely outcomes
When a finite sample space contains equally likely outcomes, probability can be calculated by counting favourable outcomes and dividing by the total number of outcomes. This is the familiar school formula. It is powerful for fair dice, fair coins and carefully specified random selections.
P(A) = favourable outcomes / total equally likely outcomes
For example, a bag contains 3 red counters and 2 blue counters. One counter is selected uniformly at random. P(red) = 3/5, while P(blue) = 2/5. The probability that the counter is not red is 2/5, which agrees with the complement rule.
The complement rule
The complement of event A is the event that A does not occur. The event and its complement cover all possibilities and cannot happen together. Therefore their probabilities add to 1.
P(not A) = 1 − P(A)
If P(rain) = 0.3, then P(no rain) = 0.7
under the same defined forecast model.
The complement rule is especially useful when it is easier to calculate the chance of something not happening. For example, the probability of at least one six in several die rolls can be found by calculating one minus the probability of no sixes.
Addition rule: when events are combined
For two events A and B, the general addition rule is P(A or B) = P(A) + P(B) − P(A and B). The overlap is subtracted once because it would otherwise be counted twice.
If A and B are mutually exclusive, they cannot occur together, so P(A and B) = 0. The rule becomes P(A or B) = P(A) + P(B). Do not add probabilities automatically if the events can overlap.
Multiplication rule and independent events
Two events are independent when knowing that one has occurred does not change the probability of the other, under the model. For independent events A and B, P(A and B) = P(A) × P(B). A sequence of fair coin tosses is a familiar idealised example.
P(A and B) = P(A) × P(B), when A and B are independent
The first coin landing heads does not change the probability that the second fair coin lands heads. The probability of two heads is 1/2 × 1/2 = 1/4. Independence is an assumption about how the experiment is generated, not a statement that events look unrelated.
Conditional probability: when new information matters
Conditional probability describes the chance of event A occurring when we already know that event B has occurred. It is written P(A | B), read ‘the probability of A given B’. The information changes the relevant sample space or probability calculation.
P(A | B) = P(A and B) / P(B), when P(B) > 0
Suppose a class has 20 students, 12 of whom take Biology, and 8 of whom take both Biology and Chemistry. If one student is known to take Biology, the probability that the student also takes Chemistry is 8/12 = 2/3. The denominator is 12 because the condition restricts attention to the Biology students.
Conditional probability is important in medical testing, weather forecasts, quality control, machine learning and everyday decisions. The key question is always: what do we already know, and how does that information change the set of relevant cases?
Probability versus odds
Probability and odds describe related ideas in different forms. If an event has probability 3/5, the odds in favour are 3:2, because there are three favourable parts for every two unfavourable parts. Odds against are 2:3. A statement of ‘3 to 2 odds’ should not be read as a probability of 3/2.
In betting contexts, quoted odds may also reflect the bookmaker’s margin and payout convention. Always identify whether the phrase refers to mathematical odds, fractional odds, decimal odds or a different pricing system.
Experimental probability: use observed results
Experimental probability estimates an event’s probability using observed frequency. If a coin is tossed 100 times and heads appears 53 times, the experimental relative frequency is 53/100 = 0.53. It is an estimate based on those trials, not proof that the coin’s true probability is exactly 0.53.
As the number of independent repeated trials increases, relative frequency tends to settle around the underlying probability in many standard models. This is related to the law of large numbers. It does not mean every short sequence must look balanced or that the next outcome is determined by previous outcomes.
Expected value: probability and long-run average
Expected value is the probability-weighted average of the possible numerical outcomes. It describes the long-run average result per repetition under the model, not necessarily what happens in a single trial.
Expected value = Σ [outcome × probability of outcome]
Imagine a fictional game that pays $10 with probability 0.2 and $0 with probability 0.8. Its expected payout is $10 × 0.2 + $0 × 0.8 = $2 per play. This does not mean the player receives $2 each time. They receive either $10 or $0 in this simplified game; $2 is the average over many repetitions.
Probability and risk: useful but not identical
Probability is one component of risk. A low-probability event can still matter if its consequences are severe. Risk analysis therefore considers likelihood, impact, exposure, uncertainty and the quality of available information. A flood with a small annual probability can require serious planning if the consequences would be disastrous.
For everyday decisions, avoid treating a single probability estimate as the whole story. Ask how the estimate was produced, what time period it covers, what the consequences would be, and whether the assumptions still apply.
Common probability mistakes
- ‘A fair coin landed heads five times, so tails is due.’ The next toss remains one-half for a fair independent coin.
- ‘A 70% chance means it will happen.’ It means the event is assigned 0.7 probability under a stated forecast or model.
- ‘The outcomes I listed are equally likely.’ Listing outcomes does not make them equally likely; check the experiment.
- ‘Add any two probabilities.’ If events overlap, subtract the overlap; if they are mutually exclusive, simple addition applies.
- ‘Independent and mutually exclusive mean the same thing.’ They are different concepts; two non-zero-probability mutually exclusive events are not independent.
- ‘Expected value is the outcome I should expect on one try.’ It is a probability-weighted average, not a guaranteed single result.
A seven-day probability learning plan
- Day 1: Define experiment, outcome, sample space and event.
- Day 2: Calculate probabilities for equally likely outcomes.
- Day 3: Use the complement rule on ‘at least one’ questions.
- Day 4: Compare mutually exclusive and overlapping events.
- Day 5: Calculate combined probabilities for independent and dependent events.
- Day 6: Compare theoretical probability with experimental frequency.
- Day 7: Explain a real-world risk estimate and state its assumptions.
Frequently asked questions about probability
What is probability in simple words?
Probability is a number that describes how likely an event is, from 0 for impossible to 1 for certain.
Can probability be greater than 1?
No. A valid probability is between 0 and 1 inclusive, or between 0% and 100%.
What does a 50% chance mean?
It means probability 0.5 under the specified model. It does not guarantee that exactly half of a small number of trials will succeed.
What is the difference between theoretical and experimental probability?
Theoretical probability is calculated from a model; experimental probability is estimated from observed outcomes.
What is conditional probability?
It is the probability of an event given that another event is already known to have occurred.
What does independent mean in probability?
Two events are independent when the occurrence of one does not change the probability of the other under the model.
Why is probability useful?
It helps people reason about uncertain events in science, medicine, engineering, weather, finance, games and daily decisions.
Continue learning at eduKateSG
For school mathematics applications, read Simple and Compound Probability Tables and Bayes’ Theorem and Conditional Probability. Those pages extend the foundations explained here.
Sources and further reading
The OpenStax Introduction to Probability and NIST statistical resources provide further study. The examples on this page are fictional teaching scenarios; probabilities only have meaning when the experiment, assumptions and outcomes are clearly specified.
