Data & Sources

XC-SPI is built from public evidence. This page states the source hierarchy, cutoff rule, and missing-data policy used to decide what can enter a weekly ranking.

Source hierarchy

A Direct primary evidence. Official NCAA/TFRRS championship results, current university athletics roster, official transfer/professional announcement, official meet result or timing file.

B Strong indirect evidence. Program-level continuity statement, official track roster where fall XC roster is incomplete, or other strong official evidence that does not completely settle championship availability.

C Unresolved or secondary evidence. A conflict or weakly supported status that must be quarantined, widened for uncertainty, or resolved before it can improve a team projection.

The preseason cutoff

Week 0 is frozen to public information available through August 28, 2026. Later roster news belongs in the later weekly model; it is not silently read backward into the preseason calculation.

Evidence weights are not literal probabilities

For Week 0, XC-SPI uses Confidence/shrinkage weights: 0.99 for verified current membership or transfer-in, 0.85 for probable return/program continuity without equally strong fall-specific verification, 0.70 for materially unresolved cases, and 0 for verified unavailability. These values are launch evidence-confidence/shrinkage rules, not claims that an athlete literally has an 85% chance of racing at NCAA.

Race and environmental data

Weekly records attach race distance, venue/course identity, start-time weather, race-window precipitation, antecedent precipitation, surface/drainage evidence, and source grade. Weather and ground data are modeled as contextual evidence with uncertainty rather than deterministic stopwatch corrections.

What missing data means

No public injury inference is made from an early absence. A missing roster page is not treated as proof that a runner left. A transfer or newcomer without directly comparable 2025 NCAA cross-country scoring evidence receives the common conservative prior rather than an athlete-specific invented value. For a 2025 NCAA individual qualifier, the preseason base uses the athlete’s hypothetical NCAA team-scoring-sequence position rather than raw overall finish. Missingness should make the output less certain, not more imaginative.

Quality control

Before release, the Week 0 dataset was checked for arithmetic consistency, NCAA scoring-rule implementation, statistical specification, and source provenance. Unresolved evidence remains explicitly down-weighted rather than promoted to verified status. The working audit files are retained as internal project records rather than published as reader-facing downloads.