Per-match Data Completeness Score exposed as an API field and a coverage timeline
Not every match in 400,000+ historical fixtures has the same depth: some have full lineups, shotmaps and per-player stats, older ones may only have a scoreline. Every match will now carry an honest completeness score and a breakdown of exactly which blocks are present, missing or internally inconsistent — so you know before you query whether a season is safe to train a model on. The same information powers a coverage timeline showing depth by competition and season.
Comments
1
Bzzoiro · Admin
Sep 23, 2026 03:25
Hi, thanks - and this one I have some sympathy for, because the problem you describe is real.
Depth genuinely does vary across the archive. An old fixture may be a scoreline and nothing else, while a recent one has lineups, a shotmap and per-player stats. Today you find that out by asking and getting nulls.
One piece of it already exists: events carry a has_xg flag, so at least that dimension is answerable without a trial request.
A full completeness score across every dimension is a bigger commitment than it looks, because the score has to stay honest as coverage changes, and a stale completeness score is worse than none. Not saying no forever, but not something I'll promise a date on.
- Bzzoiro
Log in to comment.