Bayesian A/B testing

Bayesian A/B — see which variant really wins, and how sure you can be.

Most A/B test tools hand you a p-value and a yes/no verdict. Bayesian A/B answers the
question you actually care about: "What is the probability that B is better than A?" —
with a clear winner, the probability it is best, and the believable range for every
conversion rate. It runs fully offline, holds no accounts, and never touches the network.

Type in the trials and conversions for each variant and the app instantly draws the
posterior belief curve for each one, ranks them, and tells you in plain language whether
you have a convincing winner yet.

WHAT YOU GET
• Clear verdict — the best variant, its probability of being best, and the uplift over
the runner-up.
• Posterior curves — overlaid Beta distributions so you can see the winner separate as
data grows. Less overlap means a clearer winner.
• Probability to be best — estimated by Monte Carlo sampling from each posterior.
• 95% credible intervals — the believable range for every variant's true rate.
• Up to 4 variants — compare A/B/C/D in one test.
• Flexible input — type numbers directly, or use sliders with − / + buttons.

LEARN AS YOU GO
• Concept tab explains, in plain words, what a Bayesian A/B test is and why a curve of
plausible rates beats a single p-value.
• The maths, rendered beautifully with KaTeX and bundled offline: the Beta posterior and
the probability that B beats A.

SIMULATE & EXPLORE
• Simulate tab — set a base rate, a lift and a sample size, then watch the winner pull
clear as visitors pile up. Set the lift to zero to see how identical variants behave.
• Examples — load ready-made tests with one tap: a clear winner, a too-close-to-call
result, a three-variant test, and a big-sample / small-lift case.

PRIVATE & LIGHTWEIGHT
• 100% offline — no sign-in, no tracking, no data leaves your device.
• Your inputs are saved automatically, so you never lose a test.
• Tiny download, no bloated chart or math libraries.

Whether you run marketing experiments, optimize a product funnel, or are just learning
Bayesian statistics, Bayesian A/B turns raw conversion counts into a decision you can
trust.