PERMANENT PERFORMANCE LEDGER

THE RECEIPTS,
NOT THE HYPE.

Every eligible call is judged by the published exact-stage rule. The founding edition locked that rule before official results; every later edition must lock it before the forecast itself. Pending calls never enter accuracy or calibration.

1EDITIONS

108LOCKED CALLS

0VERIFIED RESULTS

0%RESULT COVERAGE

—CALIBRATION ERROR

—BRIER SCORE

EDITION LEDGER

FESTIVAL BY FESTIVAL.

Coverage includes eligible assignments and documented exclusions. Accuracy includes eligible calls only.

2026 · ORLANDO2026 — EDC Orlando

COVERAGE0%ELIGIBLE0ACCURACYPENDINGOPEN

CONFIDENCE CALIBRATION

DID THE NUMBERS MEAN WHAT THEY SAID?

Each model version is calibrated as its own cohort, so legacy confidence never contaminates a later standard. Calibration error is the weighted confidence gap; Brier score measures probability error from 0 to 1.

MODEL COHORTFounding legacy model1 EDITION · 108 CALLS · 0 SCORED

CALIBRATION ERROR —BRIER SCORE —

50–59%0/13 SCORED

60–69%0/63 SCORED

70–79%0/6 SCORED

80–89%0/5 SCORED

90–100%0/21 SCORED

STAGE PERFORMANCE

WHERE THE MODEL HITS—AND MISSES.

AWAITING THE FIRST OFFICIAL STAGE RELEASE

This section activates automatically when verified eligible results are added.

MEASUREMENT POLICYNo pending call is treated as a miss. No cancellation is treated as a prediction failure.

Original forecasts and confidence values remain immutable. Corrections to official sources must be documented separately.

READ THE STANDARD