Distribution Lab
Nineteen probability distributions as working objects: compute probabilities and quantiles, build test distributions from their components, find the power of a test, and fit models to a simulated department whose true processes are hidden.
Distribution bench
Drag the circles on the horizontal axis to move a and b. The shaded area is the probability in the box at left.
Constructions
Each test distribution is defined by a recipe. Here the recipe is executed draw by draw and the histogram of results is compared with the exact density.
One draw, step by step
Test desk
Department simulator
Each event below follows a true distribution that you cannot see. You observe twelve past outcomes, choose a family and parameters for each event, and simulate thousands of years. Then compare your model with reality.
Years with problems
One simulated year
Exact densities and distribution functions are computed in the browser and tested against R to about eight significant digits. Source: src/lib/distributions.mjs.