Types of Probability Distributions
Every probability distribution has a job: the normal for measurements, the binomial for counts of successes, the Poisson for rare events, and so on. This guide maps the main distributions, explains when each applies, and links every one to a free calculator on this site.
A probability distribution is a recipe for how chance spreads out. Pick a distribution, and it tells you which outcomes are likely, which are rare, and how the values cluster. Pick the wrong one, and every probability you compute after it is wrong too.
Discrete or Continuous? The First Split
Every distribution on this page is one of two kinds. A discrete distribution deals in separate, countable values: 0, 1, 2, 3 successes, never 1.5. The binomial, Poisson, geometric, hypergeometric and negative binomial are discrete. A continuous distribution deals in measurements that can take any value in a range: a height of 172.4 cm, a waiting time of 3.7 minutes. The normal, t, chi-square, F, exponential, uniform and lognormal are continuous.
That split is the first question to ask. If you are counting things, start with the discrete half of this page. If you are measuring things, start with the continuous half.
Which Distribution? A Quick Decision Table
| What are you working with? | Use this distribution |
|---|---|
| A measurement that can take any value (height, weight, score) | Normal |
| A z-score, or a normal value converted to standard units | Standard normal (z) |
| A count of successes in a fixed number of trials | Binomial |
| A count of rare events in a fixed interval of time or space | Poisson |
| A mean from a small sample, with the population SD unknown | t |
| A test statistic for counts, or a sample variance | Chi-square |
| A ratio of two variances | F |
The other six distributions on this page β exponential, uniform, geometric, hypergeometric, lognormal and negative binomial β cover the jobs the seven above do not: waiting times, equal chances, and counts that run until something happens.
The Continuous Distributions
Normal distribution
The bell curve for measurements: heights, weights, test scores, errors. Parameters are the mean ΞΌ and the standard deviation Ο. Use it whenever data piles up symmetrically around a centre.
Try the normal distribution calculator β
Standard normal (z) distribution
The normal distribution standardised to a mean of 0 and a standard deviation of 1. Every normal value converts to a z-score on this scale, so one table answers every normal probability question.
Try the z-table calculator β
t-distribution
The bell curve for small samples. It looks like the normal but with heavier tails, and it flattens towards the normal as the degrees of freedom grow. Its one parameter is the degrees of freedom. Use it for t-tests and for confidence intervals for a mean when the population standard deviation is unknown.
Try the t-distribution calculator β
Chi-square distribution
The distribution of a sum of squared standard normals, so it only ever takes positive values and is skewed right. Its one parameter is the degrees of freedom. Use it for chi-square tests of counts and for inference about a variance.
Try the chi-square test calculator β
F-distribution
The distribution of a ratio of two variances, so it compares two sources of spread. Its parameters are the two degrees of freedom, df1 and df2. Use it for the F-test of two variances and for one-way ANOVA.
Try the F-distribution calculator β
The Discrete Distributions
Binomial distribution
Counts of successes in a fixed number of independent trials, each with the same success probability. Parameters are n (trials) and p (success probability). Use it for pass/fail counts, coin flips, and anything that asks “how many out of n?”.
Try the binomial probability calculator β
Poisson distribution
Counts of rare events in a fixed interval of time or space: calls per hour, typos per page, accidents per week. Its one parameter is Ξ», the mean rate. Use it when events arrive independently at a steady average rate.
Try the Poisson distribution calculator β
Geometric distribution
The number of trials until the first success: the first head, the first sale, the first working prototype. Its one parameter is p, the success probability on each trial. Use it when the question is “how long until it first happens?” rather than “how many in n tries?”.
Try the geometric distribution calculator β
Negative binomial distribution
The number of trials until the r-th success β the geometric distribution generalised from one success to r. Its parameters are r (successes wanted) and p (success probability per trial). Use it when the question is “how many tries until r successes?” rather than “how many successes in n tries?”.
Try the negative binomial distribution calculator β
Hypergeometric distribution
Counts of successes when you draw without replacement from a finite population: drawing cards from a deck, or sampling items from a batch for quality control. Its parameters are N (population size), K (successes in the population) and n (draws). Use it whenever the population is small enough that each draw changes the odds.
Try the hypergeometric distribution calculator β
Exponential distribution
The waiting time until the next event, when events arrive at a steady rate: the time until the next bus, the next failure, the next call. Its one parameter is Ξ», the rate. Use it for lifetimes and gaps between events. It is the continuous cousin of the Poisson: Poisson counts the events, exponential times the gaps.
Try the exponential distribution calculator β
Uniform distribution
Every value in a range equally likely: a random number between 0 and 10, a bus arriving at any moment in the next 20 minutes. Its parameters are a and b, the ends of the range. Use it when there is no reason for any value in the range to be more likely than any other.
Try the uniform distribution calculator β
Lognormal distribution
A distribution for positive, right-skewed measurements β incomes, house prices, concentrations β where the logarithm of the value is normal. Its parameters are ΞΌ and Ο, the mean and standard deviation of that logarithm. Use it when effects multiply rather than add, so the data stretches to the right.
Try the lognormal distribution calculator β
Frequently Asked Questions
What is the difference between a discrete and a continuous distribution?
A discrete distribution deals in separate, countable values β 0, 1, 2, 3 successes, never 1.5. The binomial, Poisson, geometric, hypergeometric and negative binomial are discrete. A continuous distribution deals in measurements that can take any value in a range β a height of 172.4 cm, a waiting time of 3.7 minutes. The normal, t, chi-square, F, exponential, uniform and lognormal are continuous.
Which distribution do I use for counts?
Counts of successes in a fixed number of trials take the binomial. Counts of rare events in a fixed interval take the Poisson. Counts of trials until a success take the geometric (first success) or the negative binomial (r-th success). Counts drawn without replacement from a finite population take the hypergeometric.
When does the normal approximation apply?
When the sample is large, the binomial and the Poisson both look increasingly normal: the binomial when np and n(1 β p) are both above about 10, the Poisson when Ξ» is above about 10. That is the central limit theorem at work β sums and means of many values tend towards the normal shape, whatever the pieces look like. See the central limit theorem explained guide for why.
Check Tools
Try the distributions on this page with the site’s free calculators:
Normal Distribution Calculator
Probabilities, percentiles and the bell curve for any mean and standard deviation.
Try the normal distribution calculator β
Binomial Probability Calculator
Exact probabilities for counts of successes in n trials.
Try the binomial probability calculator β
Poisson Distribution Calculator
Probabilities for counts of rare events at a given rate.
Try the Poisson distribution calculator β
F-Distribution Calculator
F probabilities, critical values and the F-test for two variances.
Try the F-distribution calculator β
Explore All Statistical Tools
See every calculator on the site in one place.
Check Articles
Go deeper with these guides:
Central Limit Theorem Explained
Why sample means form a bell curve, whatever the population looks like.
Binomial Distribution Examples
Worked binomial counts, from coin flips to pass rates.
Poisson Distribution Examples
Worked Poisson counts for rare events over time and space.