No promises — measurements. Every day, yesterday’s forecast is compared with what the satellite actually observed. Here is the method, and the numbers — including the ones we miss.
No promises — measurements. Every day, yesterday’s forecast is compared with what the satellite actually observed. Here is the method, and the numbers — including the ones we miss.
100% of our "clean water" forecasts (calm season, 2026-05-17 → 2026-06-17) proved correct (over 2362); the rare alerts stay flagged low-confidence until the data confirms them.
Period: 2026-05-17 → 2026-06-17
100% of our "clean water" forecasts (calm season, 2026-05-17 → 2026-06-17) proved correct (over 2362); the rare alerts stay flagged low-confidence until the data confirms them.
From May 17 to June 17, 2026, 3,420 forecasts were checked against satellite observations across 20 beaches.
Accuracy is measured on our longest-running network — 20 Caribbean beaches (Martinique & Guadeloupe), where the verification archive is deepest. The exact same model (pipeline v3) powers the Riviera Maya forecasts.
A forecast counts as a “hit” when the predicted status (clean, moderate, avoid) matches the status the satellite observed on the target day. Mean absolute error on the AFAI index: 0.044.
| Horizon | Hit rate | Checks | Displayed confidence |
|---|---|---|---|
| Day +1 | 76% | 620 | 50% |
| Day +2 | 72% | 600 | 37% |
| Day +3 | 72% | 580 | 30% |
| Day +4 | 76% | 560 | 12% |
| Day +5 | 82% | 540 | 9% |
| Day +6 | 88% | 520 | 6% |
The model also publishes its own confidence: high at day +1, low at day +6. When it is low, read the forecast as a trend, not a certainty.
8 of 20 beaches score above 90%. The hardest ones to predict — published anyway:
Exposed beaches whose state hovers around a threshold (clean / moderate) remain the hardest to predict.