The accountability funnel
How many resolved predictions survive to a real verdict
(
confirmed/refuted).Resolved calls — the actual verdicts
Every graded prediction: what PIE called, the
event probability it assigned, and the open-source evidence that confirmed or
refuted it. Newest first.
Why predictions go ungraded
Ungraded records need a valid analyst review or dated evidence.
Missing evidence does not establish whether a forecast was correct.
Grading coverage by impact tier
Graded (green) vs ungraded (red), split by the stakes of the call.
Reported probability distribution
Histogram of predicted probability across all resolved forecasts.
Reliability diagram
Each point is a probability bucket: x = predicted, y = observed
frequency. Points should sit on the dashed line. Bubble size = sample count.
Calibration by domain and impact
Calibration broken out by domain and impact tier: how often
each slice’s statements resolve true (
confirmed share), how the stated
probabilities are (Brier, lower = better), and whether the
engine over- or under-states its confidence in that slice.
Methodology: groups computed in-browser from the same
prediction_outcomes
payload; per-domain Brier prefers the resolver’s published
calibration_scores where available. Slices with fewer than 30 graded
calls are flagged low sample and make no confidence claim.
Grading rate by domain
Historical volume and reviewed coverage by domain.
| Domain | Total | Graded | Rate |
|---|