Project
Roulette

A/B Test Calculator

Is that lift real, or did you just stop looking at a lucky moment?

IntermediateWeb appData viz8–18hAI-buildable 4/5no setupworth shipping

A significance calculator with sample size planning and a plain-language verdict. Its real job is to push back on the most common mistake in experimentation: peeking at results and stopping the moment they look good.

UI
Logic
Useful

Hosting · Free forever

Deploys free on GitHub Pages, Cloudflare Pages or Vercel Hobby.

Deploy it and share a link.

Core features · 7

  • Enter visitors and conversions per variant
  • Significance and confidence interval are calculated
  • A plain-language verdict explains what the numbers mean
  • Required sample size calculates from a target effect size
  • A warning appears when the sample is below the planned size
  • The distribution of both variants is plotted with overlap shown
  • The scenario encodes into a shareable URL

The interesting part

  • Implementing the statistics correctly, including the normal approximation's limits
  • Explaining significance without reinforcing common misinterpretations
  • Discouraging peeking without making the tool feel obstructive

Once it works

  • Add Bayesian analysis alongside frequentist
  • Add multiple variant comparison with correction
  • Add sequential testing that permits valid early stopping

Suggested stack

reacttypescripttailwindrecharts

Track it