Why use this calculator
Probability is quoted in at least four incompatible ways, and mixing them up is what produces bad decisions. A scientist writes 0.3, a newspaper writes 30 per cent, a bookmaker writes 7 to 3 against and a betting exchange writes 3.33, yet all four describe the same event.
This calculator shows every form side by side so a figure can be moved between them without arithmetic slips. It also answers the question single-event probability cannot: what happens when you repeat the experiment. A one in six chance feels unlikely until you roll four times, at which point it is more likely to happen than not, and that gap between intuition and arithmetic is where most people go wrong.
How to use it
- Enter the number of favourable outcomes, meaning the results that count as a success for the question you are asking: two aces in a deck, three defective units in a batch, one winning ticket.
- Enter the total number of equally likely outcomes next, counting the unfavourable ones as well, so a standard deck is 52 and a fair die is 6.
- If the outcomes are not equally likely, this simple ratio does not apply and you need the underlying probabilities instead.
- Finally set how many independent trials you plan to run, leaving it at one for a single event, and the calculator will work out the chance of seeing the outcome at least once and the number of times to expect it.
What each result means
- Probability
- The chance of the event on a single trial, on a scale from 0 (impossible) to 1 (certain).
- As a percentage
- The same probability multiplied by one hundred, which is how most reports quote it.
- As a fraction
- The raw counts you entered, shown so you can check the outcome space is right.
- Odds in favour
- Favourable outcomes compared with unfavourable ones rather than with the total, reduced to its simplest whole-number ratio.
- Odds against
- The same ratio inverted, which is the direction bookmakers and insurers normally quote.
- Fair decimal odds
- The payout per unit staked that would break even in the long run, with no margin added.
- Probability it does not happen
- The complement of the event. A probability and its complement always add to exactly one.
- At least once across the trials
- The chance of seeing the outcome one or more times over all the trials you set.
- Every trial succeeds
- The chance that the outcome occurs on every single trial, which falls very fast.
- Expected occurrences
- The long-run average number of successes, which need not be a whole number.
The formula
P(A) = favourable outcomes / total outcomes, a dimensionless number between 0 and 1
The complement
P(not A) = 1 − P(A)
Odds in favour are favourable : unfavourable, and fair decimal odds are 1 / P(A), which is the break-even payout per unit staked. Across n independent trials the chance of never seeing the outcome is (1 − P)ⁿ, so the chance of at least one occurrence is 1 − (1 − P)ⁿ, the chance that every trial succeeds is Pⁿ, and the expected number of occurrences is n × P. Independence means one trial's result does not alter the probability of the next.
How this number is used in practice
Reliability engineering runs on the at-least-once calculation. If a component fails on one per cent of flights, the fleet is not exposed to a one per cent risk; across 200 flights the chance of at least one failure is about 87 per cent. Maintenance intervals, redundancy requirements and spares holdings all come out of that arithmetic.
The same structure drives quality sampling: pulling 20 units from a batch that is five per cent defective gives roughly a 64 per cent chance of catching at least one defect, which tells an inspector immediately whether the sample plan is strong enough to be worth running.
Gamblers and insurers work the same numbers from opposite sides. Fair decimal odds are one divided by the probability, so a true one in four shot is 4.00. A bookmaker offering 3.50 is charging a margin of roughly twelve per cent, and comparing quoted odds against the fair figure is how value is judged.
Insurers price the mirror image: the expected number of claims is trials times probability, and the premium is that expectation plus loading for costs, capital and uncertainty. Both rely on independence, which is exactly what fails in a flood or a pandemic, when correlated events arrive together.
Worked examples
Three favourable outcomes out of ten, one attempt
Probability 0.3000, 30.00 per cent, odds of 3 : 7 in favour and fair decimal odds of 3.33.
Rolling at least one six in four throws of a fair die
51.77 per cent, which is why the bet was historically a profitable one.
Drawing an ace from a full deck, five draws with replacement
A single-draw probability of 0.0769 and a 32.98 per cent chance of at least one ace.
When people use it
- Converting between probability, percentage, fractional odds and decimal odds
- Working out the chance a rare fault appears at least once over many runs
- Sizing an inspection sample so a known defect rate is likely to be caught
- Checking whether quoted betting odds are better or worse than the fair price
- Estimating the expected number of events for staffing or stock planning
Tips
- Probability runs from 0 to 1 and a percentage from 0 to 100. Anything outside those ranges means the outcome space was counted wrongly.
- Independent events multiply: P(A and B) = P(A) × P(B). Mutually exclusive events add: P(A or B) = P(A) + P(B).
- The chance of at least one occurrence is easiest via the complement: work out the chance of none, then subtract from one.
- Dice, coins and shuffled cards have no memory. A run of losses does not make the next trial any more likely to win.
- Fair decimal odds are one divided by the probability. Any quoted price below that figure carries the house margin.
- This assumes every listed outcome is equally likely. Weighted or conditional situations need Bayes' theorem instead of a simple ratio.
What it assumes, and where it stops being accurate
- All listed outcomes are equally likely, which holds for fair dice and shuffled decks but not for most real-world events.
- Trials are independent and the probability is identical each time, so sampling without replacement is not modelled.
- Favourable outcomes are capped at the total, since a subset cannot be larger than the set it comes from.
- Odds are reduced using rounded whole numbers, so fractional counts are displayed to the nearest integer.
- Fair decimal odds exclude any bookmaker margin, commission or tax, so real quoted prices will always be shorter.
Terms used on this page
- Complement
- The probability that an event does not occur, equal to one minus the probability that it does.
- Independent events
- Events where the outcome of one leaves the probability of the other unchanged.
- Decimal odds
- Total return per unit staked, including the stake; the reciprocal of the probability when fair.
- Expected value
- The long-run average outcome, found by multiplying each result by its probability.
Questions people ask
- What is the difference between probability and odds?
- Probability compares favourable outcomes with the total, so three in ten is 0.3. Odds compare favourable outcomes with unfavourable ones, so the same event is 3 to 7 in favour or 7 to 3 against. They carry identical information, but odds have no upper bound while probability can never exceed one.
- Why is the chance of at least one so much higher than the single-trial chance?
- Because every trial gets its own opportunity to succeed. The chance of missing every time is the single-trial miss probability multiplied by itself once per trial, and that shrinks quickly. A one in six chance repeated four times misses only about 48 per cent of the time, so it happens at least once more often than not.
- Does a losing streak make a win more likely next time?
- No. That belief is the gambler's fallacy. Independent trials have no memory, so a fair coin that has landed on heads ten times running still has a fifty per cent chance of heads next throw. What does happen over time is that the proportion of wins drifts toward the true probability, simply because later results outnumber the early run.
- How do I handle events that are not equally likely?
- This calculator assumes a uniform outcome space, so weighted situations need the underlying probabilities supplied directly. For conditional questions, where knowing one fact changes the chance of another, use Bayes' theorem. The at-least-once and expected-count arithmetic still applies once you have the correct single-trial probability.
- What does expected occurrences mean if it is not a whole number?
- It is a long-run average, not a prediction for any single run. An expectation of 0.67 defects in a sample means that repeating the sampling many times would average two thirds of a defect per sample. The individual results are always whole numbers; the average across them need not be.
- Is drawing cards without replacement handled here?
- No. Removing a card changes both the favourable and total counts for the next draw, so the trials are no longer independent. For small numbers of draws from a large population the difference is minor, but for anything like a five-card hand you need the hypergeometric distribution or a combinations calculation.
Use this with
Probability Calculator answers one part of the question. These pick up where it stops: Z-Score Calculator, Standard Deviation Calculator, Sample Size Calculator and Random Number Generator.
Where these figures come from
This calculation is jurisdiction-neutral arithmetic. No country-specific rules are applied.
- NIST/SEMATECH e-Handbook of Statistical Methods. Reference treatment of descriptive statistics and hypothesis testing
- Royal Statistical Society. Professional guidance on statistical practice and reporting
Probability Calculator produces an estimate for general information. It is not financial, tax, medical, legal or engineering advice. read the full disclaimer. Spotted a wrong number? Report it and it gets fixed. The form picks up this calculator and the inputs you used.