Numerai Classic model score distributions
Per-round distributions of individual Classic models’ final CORR20 and CORR60 scores, with median, standard deviation and roster coverage.
Coverage & dates
- Record date range
- to
- Date meaning
- Round opening dates in UTC. Each score horizon becomes final independently. The range can include gaps; it does not imply complete observations throughout.
- Collected
- · 01:31:58 UTC
- Prepared by
- nmrdash, from the public sources below.
Sources & methods
Per-round median of available individual models' final CORR20 and CORR60, with a band equal to median plus/minus one population standard deviation of those same scores. Models have equal weight. No rolling average or warm-up period. Missing scores are excluded, never zero-filled. Partial published rosters are included with returned/submitted and scored-model counts disclosed; these distributions may differ from the full population. Each metric has independent finality. Dates are round opening dates.
The chart uses equal model weights and median ±1 population standard deviation. Available scores from partial rosters are included with coverage disclosed; these can differ from the full population. No rolling average is applied.
Source fields: Numerai roundDetails.models.v2Corr20 / corr60; independent final score clocks from public model submissionScores (integration_test, with a second model for missing historical rounds); roundScoreConfigs and round target.
JSON fields
Paths use [] for array items and * for metric keys such as corr20 and corr60. Absent fields or null values mean unavailable information; a numeric zero remains a real observation.
| Field | Meaning | Unit |
|---|---|---|
fetched_at | Collection timestamp in UTC. This does not establish the observation date or score finality. | Timestamp |
definition_version | Version of the dataset’s measurement and export definition. | Text |
points[].round | Numerai Classic round number. | See definition |
points[].time | Round opening timestamp in UTC. | See definition |
points[].round_target | Target universe configured for the round. | See definition |
points[].total_models | Expected submitted model count, when known. | models |
points[].returned_models | Number of model records returned by the source. | models |
points[].roster_complete | Whether the collected model roster is complete. | See definition |
points[].metrics.*.score_name | Official metric name; * selects corr20 or corr60. | See definition |
points[].metrics.*.version | Metric definition version. | See definition |
points[].metrics.*.final_date | Final score observation date. | See definition |
points[].metrics.*.scored_models | Number of available model scores in this distribution. | models |
points[].metrics.*.median | Median of available individual model final scores. | CORR |
points[].metrics.*.std_dev | Population standard deviation of those same model scores. | CORR |
points[].metrics.*.lower | Median minus one population standard deviation. | CORR |
points[].metrics.*.upper | Median plus one population standard deviation. | CORR |
points[].metrics.*.missing_reason | Explanation for an unavailable score distribution. | See definition |
Download
Download model-performance.json · application/json
Definition version: numerai-classic-round-model-performance-v1. Decimal strings preserve exact amounts; use decimal arithmetic when working with NMR or USD totals.
See the release manifest for this file’s checksum and byte count. The download is public and needs no authentication or JavaScript.
Interactive charts can load newer snapshots. Compare collection and observation dates before combining a download with chart values or another release.