Data health

The state of our own pipeline.

A site that publishes a record should be honest about the plumbing underneath it. This page is generated from the same build that produces every other page.

Rows in file

8,394

as delivered by the collector

Usable

8,393

parsed and validated

Quarantined

rejected as malformed and excluded

xG coverage

86.5%

of played matches carry expected goals

1 rows were rejected on this build. They are malformed — typically a record torn in half, which happens when the data file is copied while the collector is still writing to it. They have been excluded from every number on this site rather than quietly used. This page exists so that a degraded pipeline is visible rather than hidden.

Coverage

Leagues currently in the data: 2. Bundesliga, Bundesliga, Champions League, Championship, Eredivisie, Europa League, La Liga, La Liga 2, Ligue 1, Ligue 2, Major League Soccer, Premier League, Primeira Liga, Scottish Premiership, Serie A, Serie B.

Ratings carried forward across 9889 xG matches on file (1354 this season, 2026/27) across 11 leagues.

Scheduled fixtures ahead: 3,480. Played matches on record: 4,913, of which 4,248 carry expected goals.

Collector jobs

Every scheduled pull records its own exit code when it finishes. This table is that record, unedited.

Job	Last run	State
expand	18 hours ago	ok
morning	17 hours ago	ok
picks	7 hours ago	ok
players	35 min ago	ok
postmatch	5 min ago	ok
prematch	1 min ago	ok
refresh	8 hours ago	ok
xg-backfill	23 days ago	ok

Phase 2 watch: corners and cards

We capture corner and card odds for every tracked fixture and the real counts at full time, but we have not built a corners or cards model. Our rules say a separate model must be built, backtested and calibration-checked before a single pick in those markets is published. The gate for a first read is 500 paired fixtures (pre-match odds plus the full-time count); a confident read needs about a season. This is that sample filling up, in public.

Market	Paired	With a close	To first read	Where from
Corners	911	911		since 2026-08-23 · Championship 133, La Liga 2 115, La Liga 98, Serie A 96
Cards	779	767		since 2026-08-23 · Championship 133, La Liga 2 114, La Liga 98, Serie A 94

Capture began 23 Aug 2026 and cannot be backfilled — closing prices only exist if you were there before kick-off. Nothing here is a tip; the discipline tables are colour, not a model.