Public forecast documentation
Forecast accuracy: evidence and limits
The dated audit reports two aggregate statistics, while retained historical rows support their arithmetic without establishing a prespecified, current accuracy evaluation by horizon or market regime.
Historical audit snapshot created 2026-09-10 — not a live/current forecast
Aggregate performance evidence, kept separate
This is the central page for the forecast-performance evidence that is currently supported. “Accuracy” is not one generic number here: direction performance, probability quality, price-error measurement, and interval coverage must remain distinct.
The table contains only the two aggregate measurements supplied by the dated audit. It does not create rows beyond the retained observations or turn row counts into unsupported current-performance claims.
| Metric | Value | Sample N | Evaluation period | Interpretation | Limitation |
|---|---|---|---|---|---|
| Direction Brier score | 0.250491 | n=40 (retained resolved SOL rows) | Observed target span: 2026-09-07T07:14:30Z–2026-09-10T10:59:20Z; not a prespecified evaluation window | The retained 40 resolved rows reproduce 0.2504913345, displayed here as 0.250491; this is slightly above the 0.25 reference for a constant 50% binary forecast. | This row-derived snapshot is contextual evidence, not a demonstration of directional skill or predictive superiority. |
| Observed coverage of nominal 80% bands | 95% observed; 80% nominal | n=40 (retained resolved SOL rows) | Observed target span: 2026-09-07T07:14:30Z–2026-09-10T10:59:20Z; not a prespecified evaluation window | 38 of 40 retained resolved outcomes fell inside the published bands (95%), above the nominal target in this snapshot. | Without interval widths, horizon-level coverage, or a prespecified evaluation design, higher coverage cannot establish useful accuracy. |
Evidence inventory
- Evaluation period
- Observed target span 2026-09-07T07:14:30Z–2026-09-10T10:59:20Z; not a prespecified evaluation window.
- Number of forecasts
- 56 total SOL forecast rows in the retained historical capture; not a complete-history claim.
- Number resolved
- 40 resolved SOL rows in the retained historical capture.
- Horizons evaluated
- The retained resolved rows show 6h n=13, 12h n=12, 24h n=9, 48h n=6; these are observed counts, not a complete-history claim.
- Direction performance
- A directional Brier score of 0.250491 is supplied; a directional hit-rate or classification-accuracy value is not, and no production threshold rule is documented.
- Baseline comparison
- The constant 50% binary reference scores 0.25; this is a contextual baseline, not a holdout benchmark.
- Interval coverage
- Nominal 80% and observed 95% are supplied; interval widths and horizon-level coverage are not.
- Normalized error
- No normalized-error value is published here. Read the normalized forecast-error definition for the observed deployed-card calculation.
Choose the metric owner
For the score's formula and baseline, read the direction Brier-score explanation. For probability and interval reliability, read the calibration evidence explanation.
Readers looking for retrospective proof should start with forecast history and provenance; the retained rows are point-in-time evidence, not an immutable row-level archive.
For the captured price-error calculation, read the normalized forecast-error definition.
Performance evidence limits
The retained September 10, 2026 audit evidence contains 56 total SOL forecast rows and 40 resolved rows. Their observed target span is 2026-09-07T07:14:30Z–2026-09-10T10:59:20Z; this is row-derived context, not a prespecified evaluation window.
The figures below are a dated audit snapshot, not a live feed. The retained rows support arithmetic for the displayed Brier and coverage values, but they are point-in-time API evidence rather than an immutable research dataset. The snapshot cannot establish profitability, statistical significance, non-overlapping holdout performance, or predictive superiority.
Data provenance and export status
The statistics shown here come only from the forecast audit snapshot created 2026-09-10. Current values remain on the external live product. The local scheduled-worker forecast table is a distinct implementation and is not represented as that external production feed.
Citation-ready public export status: NOT READY. The retained historical evidence contains 56 total SOL forecast rows and 40 resolved rows, with resolved horizon counts of 6h n=13, 12h n=12, 24h n=9, 48h n=6 and observed target span 2026-09-07T07:14:30Z–2026-09-10T10:59:20Z. These are point-in-time observations, not an immutable issued-and-resolved archive, complete history, or prespecified evaluation window. A current public feed may expose rows, but it is not an immutable versioned research archive. Metrics calculated from a changing feed are not automatically comparable to this dated snapshot. A research export would need versioned records with forecast ID, asset and venue, UTC issue and target timestamps, horizon, point prediction, interval bounds, direction probability, model version, outcome source, observed price, resolved-at time, scoring-rule version, and content hash. No research export is available from these pages.
The existing Dataset JSON-LD preserves the audit date rather than implying freshness.