NoiseFloor

NoiseFloor is a statistics toolkit for telling real change from noise. You send the numbers; it returns the test, the assumptions behind it and the limits of what it shows. With NoiseFloor you can: - get a peek-safe verdict on an A/B test; - check whether a metric really changed, or whether a move is explained by sample size; - review several metrics at once without false discoveries piling up; - get an empirical interval for the next value, and replay how past forecasts scored; - review market series for noise, shared factors and evidence quality, and triage repeated headlines. Results do not establish causality, investment performance or certainty about future outcomes. NoiseFloor works on public, synthetic or explicitly authorized data and computes only on what you send. Every tool is read-only, and no account, API key or payment is needed; public use is rate limited.