Race intelligence

Every number needs a source and a state.

The methodology is deliberately public so athletes can tell verified evidence from a planning model.

Evidence states

The label is part of the statistic, not a footnote.

observedMeasured from an authorised sample
modelledA planning estimate, never presented as fact
provisionalUseful but still awaiting final verification
pendingThe source or sample is not yet sufficient
suppressedHidden to protect a small cohort

Publication rules

Systematic imports require an open licence, explicit permission or first-party consent.

Rights

Allowed sources

Open-licensed, permission-granted and consented first-party aggregates.

Explore →
Thresholds

Minimum samples

50 overall and 20 per subgroup before observed values are published.

Explore →
Corrections

Traceable changes

Every correction records the field, reason, source and publication date.

Explore →

Translation safety

English is the factual source. A translated article is indexed only after review against the exact source-content fingerprint.

Freshness fails closed.

If the English title, search intent, race link, factual sections or FAQ changes, the translation becomes stale automatically. It is removed from language alternates and sitemaps, and its generated page is withdrawn until the translation is reviewed again.

From source to published figure

A number is published only after its source, rights, sample and evidence state are recorded.

1 · Source

Record provenance

Capture the source, date, permission or licence and the field being supported.

2 · Classify

Assign the state

Label the value observed, modelled, provisional, pending or suppressed before presentation.

3 · Threshold

Check the sample

Observed aggregates require at least 50 records overall and 20 in a published subgroup.

4 · Publish

Keep label and value together

The evidence state, source scope and relevant limitation travel with the number.

5 · Correct

Record the change

A correction states the field, reason, source and publication date in the data changelog.

What the data can and cannot say

Race Intelligence separates description from prediction so an estimate is not presented as a verified fact.

Evidence boundary.

Observed values describe the authorised sample. Modelled values are planning estimates. Neither guarantees an individual finish time, course condition or race-day outcome. Mutable facts such as routes, cut-offs, aid stations and schedules must be checked against the current organiser guide.

Methodology reviewed . See the data changelog for traceable corrections.