Track record
Two kinds of evidence, kept apart. Test results: how each published model scored on years held back from its development, in tests written down before the data was scored. Live record: how the forecasts we actually published scored once their windows closed. If a live model is not scored yet, this page says so directly. It also shows any scoring backlogs.
2,871
forecasts recorded
Every published forecast, frozen with its inputs.
2,729
closed and scored
Windows that have ended and been scored against what happened.
5
waiting to be scored
Windows that have ended but are not scored yet.
0
breaks in the ledgers
Each entry carries the previous entry’s hash, so an edit to history would show here.
Needs attention
eq_operational_S2_gear1_v2-ecb87f020f99tornado_v3-05a06c843c87forecast_id removed (they were hashed before that field was added); 0 mismatchesEvery model version that has published a number, including retired ones. A score is shown only once 10 or more events have been observed for that version: with fewer, a model that always says “no” would look as good as a skilful one. Earthquake forecasts are issued several times a day for 30 days ahead, so their windows overlap: each earthquake is counted once, and a score also needs two windows that do not overlap.
| Hazard | Model version | Status | Forecasts closed | Events observed | Score |
|---|---|---|---|---|---|
| Earthquake | eq_coherence_v1_0 | retired | 668 (6 non-overlapping) | 74 | ranking 66.2%; information gain per M6+ earthquake per window vs a uniform map −15.29 (averaged over overlapping windows) |
| Tornado | tornado_storm_v1_0 | retired | 233,600 | 634 | ranking 74.2%; Brier skill vs climatology −8.45 |
| Tornado | tornado_v3-05a06c843c87 | published | 15,294 | 30 | ranking 99.1%; Brier skill vs climatology 0.14 |
| Tornado | tornado_v3-9005905f08b7 | retired | 807 | 0 | too few events to score yet (need 10) |
| Hurricane | hurricane_ri_stack_v1-35938d6c09b7 | published | 105 | 18 | ranking 91.9%; Brier score 0.10 |
| Hurricane | hurricane_ri_v8_2 | retired | 33 | 6 | too few events to score yet (need 10) |
| Hurricane | hurricane_ri_v8_3 | published | 34 | 4 | too few events to score yet (need 10) |
| Hurricane | jtwc_warning_ingest | retired | 5 | 0 | too few events to score yet (need 10) |
Calibration compares what was forecast with what happened. The published tornado model’s 2025 test, split by how high the forecast was. Live calibration is measured once enough forecasts with events have closed.
| Forecast range | Storm observations | Average forecast | Tornado followed | 95% interval |
|---|---|---|---|---|
| 0%–0.1% | 1,340,592 | <0.01% | 0.01% | <0.01%–0.01% |
| 0.1%–0.2% | 44,834 | 0.14% | 0.13% | 0.10%–0.17% |
| 0.2%–0.5% | 39,904 | 0.32% | 0.29% | 0.24%–0.35% |
| 0.5%–1.0% | 18,840 | 0.70% | 0.58% | 0.48%–0.70% |
| 1.0%–2.0% | 11,724 | 1.40% | 1.17% | 0.99%–1.38% |
| 2.0%–5.0% | 8,184 | 3.09% | 2.30% | 1.99%–2.64% |
| 5.0%–10.0% | 3,231 | 7.03% | 6.96% | 6.12%–7.88% |
| 10.0%–20.0% | 1,595 | 13.76% | 16.99% | 15.21%–18.89% |
| 20.0%–30.0% | 430 | 24.49% | 21.86% | 18.15%–25.95% |
| 30.0%–50.0% | 641 | 37.65% | 34.63% | 31.03%–38.38% |
| 50.0%–100% | 2 | 51.27% | 50.00% | 6.08%–93.92% |
In a well-calibrated row the two middle numbers agree. Every breakdown of the tornado test