independence

Shows independence as a geometric/product rule: event A is a vertical strip and event B is a horizontal strip inside a sample-space rectangle, so the overlap area (A∩B) equals P(A)·P(B). A second panel compares P(A) to P(A|B) (and P(A|not B)) to emphasize that conditioning gives no new information when A and B are independent.

canvasclick to interact

t=0s

practical uses

01.Checking whether two sensors/errors can be modeled independently in a system model
02.Simplifying calculations in probability by replacing joint probabilities with products of marginals
03.Validating assumptions in A/B testing or randomized experiments (treatment independent of assignment)

technical notes

Responsive two-panel layout using scale=Math.min(w,h)/baseSize with 4px snapping for a blocky aesthetic. Time-based 4.2s cycle reveals marginals, then joint-as-product, then conditional equality; uses the provided ease(t) for smooth phase transitions. Pure Canvas2D drawing (rectangles, stroke/fill, monospace labels) on black with GREEN/GREEN_DIM accents.

← greedy-algorithms maximum-likelihood-estimation →