HomeStatisticsA/B Test Lab: Peeking, SRM & Bandits

🧪 A/B Test Lab: Peeking, SRM & Bandits

Watch two live A/B variants collect visitors while a p-value trace, a sample-ratio-mismatch alarm, and a Thompson Sampling bandit show why peeking inflates false positives and when a bandit beats a fixed split.

Statistics3DModerate60 FPS
ab-testing-pitfalls-peeking-sample-ratio-mismatch-bandits ↗ Open standalone
⚙ Under the hood

Watch two live variants collect visitors while a p-value trace, a sample-ratio-mismatch alarm, and a Thompson Sampling bandit reveal why peeking inflates false positives and when a bandit beats a fixed A/B split.

Three.jsA/B testinghypothesis testingp-valuesample ratio mismatchmulti-armed banditThompson samplingstatistics

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install

What did you find?

Add reproduction steps (optional)