Accuracy

Honest about how often we're right.

Most weather apps don't tell you how often they're wrong. We do. Every three hours we snapshot what each of our five forecasts said. Every hour we fetch actual readings from UK Met Office stations. Then we score the forecasts against what actually happened. The numbers below are the result.

Loading scores

How this is measured

Plain-English methodology

What we measure

Four numbers per source:

Temperature error — the average gap between forecast and actual, in °C. Lower is better. A score of 0.8°C means the forecast was, on average, just under a degree off.
Wind error — the average gap between forecast wind speed and actual, in km/h. Wind is harder to forecast than temperature because it's locally gusty, so the bands here are more generous than temperature.
Rain hit-rate — when the source said it would rain (≥50% probability), did it? When it said no rain, did it stay dry? The hit-rate is the percentage of times those calls matched reality.
Condition match — did the forecast's category (sunny, cloudy, rainy, etc.) match what we observed? Percentage of times they agreed.

What we use as ground truth

We pull observations from Open-Meteo's Historical Weather Archive, which blends ground-station readings with ECMWF's ERA5 reanalysis to give an hourly observation record for any point in the UK. We monitor 20 locations spread across the UK so the scores reflect performance nationally, not just one corner.

What we don't do

We don't cherry-pick. We don't drop bad runs. We don't weight scores by what makes any one source look better. The default window is the last 14 days; you can change it with a ?days=N URL parameter (up to 90).