Actual wind vs the forecast you would have had
For each hour, the latest forecast published before the decision time. Scroll or drag the slider to zoom; click to open that day below.
Error over time
Forecast minus actual. Above zero: forecast too high (you would be short). Click to open that day.
Day drill-down
Every hour of one day. Opens on the worst day; pick another or click a chart above.
Does deciding later help?
Typical miss and the 90% buffer for different decision times on the same dates. Click a bar to use that time.
Does forecast age matter?
Error grouped by how many hours before the target hour the forecast was published.
How big are the misses?
Number of hours per 0.5 GW error band (labels in GW).
Which hours are worst?
Typical miss and bias by hour of day (UTC).
About this analysis
What the numbers mean and how they were made.
- The question
- If you had to plan the day before, using the latest wind forecast available at your decision time, how far off would it have been?
- Data
- Elexon national GB wind: actual output (FUELHH, half-hourly) and the day-ahead wind forecast (WINDFOR, hourly). Delivery 2 January to 1 February 2026.
- Decision time
- A time on the day before delivery, in UK time (default 09:00). For each hour we take the newest forecast published at or before that time. A forecast published later is never used.
- Matching
- Forecasts are hourly and actuals half-hourly, so the two half-hours inside each forecast hour are averaged before comparing.
- Error
- Forecast minus actual. Positive means the forecast promised more wind than came, so you would have been short.
- Missing data
- Hours with no forecast before the decision time, or no actual, are skipped and counted under coverage, never filled in.
- Times
- Charts and tables are in UTC. In January UTC and UK time are the same.