How Are Football Match Stats Collected? From the Stadium to Your Screen

For every Premier League match, three Opta-trained analysts work the live data feed:

• Coder 1 — possession + passing. Codes every pass: who, to whom, success/failure, body part, pass length category.
• Coder 2 — defensive actions. Codes every tackle, interception, clearance, foul.
• Coder 3 — shots + complex events. Codes shots (with location coordinates, body part, xG inputs), set-piece deliveries, and edge cases.
• Quality assurance. A fourth analyst reviews edge cases in real time and corrects any miscodes flagged by automated checks.

Two major tracking technologies provide the off-ball data:

• Camera-based tracking (Second Spectrum, Hawk-Eye). Multiple high-resolution cameras around the stadium track every player and the ball at 25 frames per second. Computer vision converts the video into x,y coordinates.
• Sensor-based tracking (Catapult, STATSports). GPS pods worn by players record position. More common in training; less in matches.
• StatsBomb 360. Combines camera tracking with their event data — provides player positions at the moment of every key on-ball event.

Three layers of verification:

• Real-time peer review. The fourth analyst on each match catches flagged issues during play.
• Post-match scrub. Within 1-2 hours of full-time, a senior analyst reviews the entire match and corrects any errors.
• Cross-comparison. Different data providers (Opta, StatsBomb, Wyscout) collect overlapping events. When their numbers disagree, both flag it for review.

Match data flows to multiple endpoints:

• Broadcasters. Sky, TNT, BT Sport, etc. receive live event data for graphics overlays during matches.
• Football apps. Sofascore, Fotmob, Fbref all license event data from Opta or its competitors.
• Clubs. Use the data for opposition analysis, recruitment, training planning.
• Bookmakers. Live in-play odds adjust based on event data within seconds of each event.
• Public + paid analytics. FBref publishes free; StatsBomb and Opta Pro charge for deeper access.

Latency between match action and data on your screen is typically:

• Live broadcast graphics. ~1-2 seconds.
• Sofascore / Fotmob match views. ~5-15 seconds depending on event type.
• Bookmaker odds updates. ~3-10 seconds depending on the bet type.
• FBref deep stats. ~30 minutes after full-time (real-time isn't their priority).
• Cross-checked / archived data. 24-48 hours for fully reviewed canonical figures.

Three industry topics:

• Cost. A Premier League match's data collection costs ~£10,000-30,000 in human-analyst labour and tracking equipment. The total Premier League match-coverage market is worth £100m+ annually.
• Player consent. Players are tracked extensively; data is used in performance reviews, contract negotiations, and recruitment models. Pro Football Association (PFA) has been pushing for clearer player consent and data-use rules.
• Lower-tier coverage gaps. Premier League and Championship are fully covered. Lower English tiers, women's football below WSL, and most non-European leagues have partial or no coverage. This affects scouting equity.

Three principles for fans and amateur analysts:

• Trust archived numbers more than live numbers. Live event data has higher error rates; the canonical figures (24-48 hours post-match) are more reliable.
• Cross-reference between providers. If an Opta number and a StatsBomb number disagree, the disagreement is usually a methodology difference, not an error.
• Check the context. A "20 shots" stat is meaningless without xG context; a "70% possession" stat is meaningless without progression context.