Scorecards
The same prospective verification reports delivered to subscribers — per-region accuracy, hit rates and capture reliability, day after day
Latest period at a glance
Latest period at a glance — every tile compares the newest report period against the one before it. MAE is the average miss; bias is the direction of the miss; interval and bucket are hit rates your subscribers can verify themselves against the official daily high.
Region × period matrix
Region × period matrix — one cell per region per report period, shaded by the selected metric, split across the three capture hours (08:00 / 12:00 / 16:00 local). Green is good on MAE and coverage; on bias, both extremes shade toward red.
MAE by checkpoint hour
MAE by checkpoint hour — how far off the published high was at each of the three capture times. Later checkpoints know more about the day, so error typically shrinks from H08 to H16.
Systematic bias
Systematic bias — weighted signed error per capture hour. Consistently above zero means forecasts run warm; below zero means cool. A bias that persists for days is a calibration problem, not bad luck.
Forecast outcome quality
Forecast outcome quality — how often the realised high landed inside the issued interval, and how often it hit the exact bucket. These are the two numbers a market reader cares about most.
Capture reliability
Capture reliability — the share of eligible checkpoints actually captured. Accuracy means nothing if the checkpoint never got recorded; this is where silent pipeline gaps show up first.
Near-miss quality
Near-miss quality — how often the realised high landed within 1°C of the published centre. The difference between a wrong forecast and a nearly right one.
Report logs
Report logs — every generated report for the selection. Click any row to expand its per-hour breakdown (eligible, captured, missing, unresolved, within-1°C). ⚠ marks thin periods under 8 paired samples.