Calibrated forecasts
The resolution ledger is our product and our proof: every forecast we have ever issued, scored against what actually happened. No deletions, no do-overs.
The record
A forecast you cannot score is an opinion with confidence theater.
When we say 70%, it happens about seven times in ten. That is the whole product.
Recently resolved
| Question | Domain | Forecast | Outcome | Brier |
|---|---|---|---|---|
| EU carbon price closes above €90 in Q2 | Energy | 0.72 | Yes | 0.078 |
| Panama Canal daily transits back to 36 by June | Supply chain | 0.31 | No | 0.096 |
| US CPI YoY prints below 2.6% in May | Macro | 0.58 | Yes | 0.176 |
| Cat-3+ hurricane makes US landfall before Aug 1 | Climate | 0.22 | No | 0.048 |
| ECB cuts at the July meeting | Macro | 0.84 | Yes | 0.026 |
| Baltic Dry Index above 2,000 on Jun 30 | Shipping | 0.45 | No | 0.203 |
Illustrative sample. The full ledger — every question, forecast history, and resolution source — is queryable via the World Model API.
Methodology
Every resolved forecast gets a Brier score — squared error against the outcome. Lower is better; 0.084 is the fleet average.
Questions cannot be edited or removed after issue. Embarrassing misses stay on the ledger next to the wins.
Every question names its resolution source at issue time — the print, the index, the official record.
Run your first forecast for free, then pay as you go when the decisions get bigger.