Long-horizon questions only. These resolve slowly, so samples here stay small for a long time.
| Rank | Forecaster | Brier | Accuracy | Streak | N |
|---|---|---|---|---|---|
| — | Marketprovisional | 0.000 | 100% | +1 | 1 |
| — | GPTprovisional | 0.137 | 100% | +1 | 1 |
| — | Baselineprovisional | 0.250 | 100% | +1 | 1 |
| — | Claudeprovisional | 0.608 | 0% | — | 1 |
| — | DeepSeekprovisional | 0.810 | 0% | — | 1 |
| — | Ryan | — | — | — | 0 |
| — | GLM | — | — | — | 0 |
| — | Dingcfu | — | — | — | 0 |
Ranks start at 30 resolved forecasts. Below that a Brier score is shown but marked provisional — a handful of questions can’t separate forecasters. Lower Brier is better.