Probability Distribution Calculator
Computes probability density (normal) or probability mass (binomial, Poisson), plus the cumulative distribution function, for three commonly used distributions in statistics coursework:
- Normal: continuous, defined by mean (μ) and standard deviation (σ).
- Binomial: discrete, the probability of exactly k successes in n independent trials with success probability p.
- Poisson: discrete, models the number of events in a fixed interval given an average rate λ.
Choosing the right distribution
The distribution is a model of how your data is generated, and picking it is the part that requires thought.
Binomial applies to a fixed number of independent trials, each with the same probability of success — 20 coin flips, 500 emails each with a 3% open rate. It answers "how many successes out of n".
Poisson applies to counts of events in a fixed interval when events are independent and occur at a constant average rate — arrivals per hour, defects per batch, requests per second. It has no upper bound, which is what distinguishes it from binomial. When n is large and p small, binomial converges to Poisson with λ = np, which is why the two are often interchangeable for rare events.
Normal applies to continuous measurements clustering around a mean, and matters far beyond data that is genuinely normal: the central limit theorem means sample means tend towards normality whatever the underlying distribution, which is what makes t-tests and confidence intervals work.
PDF, PMF and CDF
For discrete distributions the PMF gives the probability of exactly one outcome — P(X = 5) is a real, meaningful number. For continuous distributions the PDF is a density, not a probability: P(X = 5) is zero for a continuous variable, and only intervals have probability. The density's height is only meaningful relative to other heights.
The CDF gives P(X ≤ x), and is usually what a question actually needs. "At most 3 defects" is the CDF at 3. "More than 3" is 1 minus that. "Between 2 and 5" is the difference of two CDF values.
Accuracy
The normal CDF has no closed-form integral and is computed here with a numerical approximation accurate to roughly seven decimal places — far beyond what any coursework or applied use requires. Binomial and Poisson are computed exactly, though very large parameters can reach the limits of floating-point arithmetic.