Data should earn our trust.
AIONORA verifies historical coverage, dates and source rights before using sports results for research. A successful download is not proof of a complete league season or a reliable betting forecast.
What we checked
Source: OpenFootball CC0 GitHub repository; 18 original files and two cross-checked extended files selected (plus one unavailable league-season path). Numbers describe repository records, not independently certified official match coverage. A source inventory and hashes were archived for reproducibility.
| Competition | Season | Results / listed | Observation |
|---|---|---|---|
| Premier League | 2024/25 | 380 / 380 | All listed results scored |
| Premier League | 2025/26 | 380 / 380 | Full-source recovery |
| Bundesliga | 2025/26 | 306 / 306 | Full-source recovery |
| La Liga | 2025/26 | 365 / 380 | Incomplete |
| Serie A | 2025/26 | 344 / 380 | Incomplete |
| Ligue 1 | 2025/26 | 282 / 306 | Incomplete |
| Greek league | 2025/26 | 61 / 182 | Substantially incomplete |
Two complete 2025/26 variants recovered 39 missing scores after 353/353 England and 294/294 Germany overlap checks with zero score conflicts. Club-name crosswalks (20 England, 18 Germany) prevent duplicate team identities. These extended archives also contain historical post-match lineups, which must not be misrepresented as available pre-kickoff. The 2026/27 Premier League archive's newest scored match was dated 20 September 2026. It is not suitable as a real-time October feed; future fixtures are never allowed into historical results training.
Daily collection status
Checking saved research snapshot...
A configured automatic daily job is not proof of successful provider collection, private archiving or model certification.
What the new controls prevent
Research-only data importer and benchmark have passed 28 new tests; the 111 pre-existing self-optimization tests remain green. Unscored games, invalid scores, duplicates, future results, missing source identity and unverifiable kickoffs are never promoted into a Certified CORE probability.
Source rights: OpenFootball CC0 is usable for historical score research. Commercial tennis sources and NBA-origin betting statistics need separate usage rights; neither has been silently ingested.
Neither this audit nor the simple research score benchmark is proof of model accuracy, profitable edge or a betting recommendation.