Accuracy ledger
This view preserves forecast evidence from fully completed San Francisco/Fly runs. When a seven-day window closes, its observed outcome can be attached without changing the stored distribution or verdict. Go and Don’t are scored; Hold stays explicitly unscored. Legacy binary rows remain labelled legacy.
This is a scheduled seven-day weather benchmark. It does not measure the accuracy of your personalized trip, photography light or activity conditions. Outcomes use archived weather-model data, not reports from travelers.
Loading the published method…
Forecast rows—
Verified—
Incorrect—
Loading retained evidence…
| Called | Destination | Window | Strong / Borderline / Poor | Lead-adjusted decisive support | Verdict | Observed | Result | Version |
|---|