Calibration scorecard
We publish our forecasts and our score, including when the score is bad. Authority is earned, not declared.
How this is scored
Every resolved forecast is scored with the Brier score: the squared distance between the probability we declared and what actually happened. 0 is perfect, 0.25 is what you get by always saying 50/50, and 1 is being confidently wrong. A human resolves each forecast against the evidence; nothing auto-resolves.
Where we stand
Below the threshold we do not claim to be calibrated — a handful of forecasts cannot establish a track record. Until then this page is a commitment, not a credential.
Our score is currently WORSE than always saying 50/50. We are publishing that rather than hiding it.
Every resolved forecast
| Forecast | We said | It happened | Brier | Horizon |
|---|---|---|---|---|
| APARIENCIA: Antes del 2026-08-01 ocurre un episodio público visible de escalada Irán–EEUU/Israel (declaraciones, movimiento militar, o incidente naval | 0.55 | yes | 0.2025 | 2026-08-01 |
| SUSTANCIA: Antes del 2026-08-01 se materializa un cambio sustantivo y verificable (ataque cinético directo de envergadura, ruptura formal de negociaci | 0.20 | yes | 0.6400 | 2026-08-01 |
| TESIS: Antes del 2026-08-01 los hechos confirman la tesis de que el episodio marca un giro estratégico duradero (guerra abierta sostenida o reconfigur | 0.12 | yes | 0.7744 | 2026-08-01 |
What is not here yet
Forecasts whose horizon has not arrived are not shown as scores, because they are not results. The falsifiers stated in advance on each brief are the same discipline applied per piece.