Augmented Dickey-Fuller (ADF) Test
A statistical test for whether a time series wanders over time (non-stationary) or reverts to a stable pattern (stationary).
Many statistical models assume a series' average and variability don't systematically drift over time — that property is called stationarity. Interest rates and exchange rates usually don't have it: they trend and wander over years. The Augmented Dickey-Fuller test checks whether a series is stationary or not.
A low p-value (below 0.05 on this site) means the test rejects the idea that the series is a random wandering path, i.e. the series is judged stationary. A high p-value means it isn't — in which case this site differences the series (looks at day-to-day changes instead of levels) before feeding it to a model.
See the results for every series on the diagnostics page.