The Short Answer
Football event data is a row by row record of everything that happens on the ball in a match: every pass, shot, tackle and touch, logged as its own event with a timestamp, the player and team, an XY pitch location, and qualifiers that describe how it happened. Where match statistics give you totals, event data gives you the sequence, so you can rebuild the match action by action.
What is football event data?
Event data is the atomic layer of football data. A match is not stored as a summary but as a long list of individual on-ball actions, in the order they happened, each one a row in a table. A typical match produces somewhere between 1,500 and 2,500 events depending on the provider and how finely actions are split. In Dribble's own record of Crystal Palace against Arsenal on 24 May 2026 there are 1,574 events, and 920 of them are passes. Shots, the events most summaries are built around, account for just 25.
That ratio is the point. Match statistics, the numbers on a results page, are aggregates: passes completed, shots taken, possession share. Every one of those totals was counted from event rows like these. If you hold the events, you can recount the totals any way you like, filter them by minute, by player, by pitch zone, or rebuild the match as a sequence. If you only hold the totals, none of that is possible.
What does one event row actually contain?
Take one real row end to end: event 727 of that Palace against Arsenal match, a shot by Mikel Merino in the 79th minute.
- When. Period 2, at 79 minutes 0 seconds. Every row carries a period and a clock time, so the full match can be replayed in order.
- What. The event type is MISS, one of a fixed vocabulary of types (PASS, SHOT outcomes like GOAL, MISS, ATTEMPT_SAVED and POST, TACKLE, CLEARANCE, and so on), with an outcome flag.
- Who. Team Arsenal, player Mikel Merino.
- Where. x 94.8, y 48.8: the spot the shot was struck from, on the 0 to 100 pitch grid explained below.
- How. Body part left foot, and two fields most summaries never show: goalmouth_y 49.6 and goalmouth_z 58.3, where the shot crossed the plane of the goal, across and up. A miss struck from that close, crossing that close to the centre of the goal, tells a very different story from the word MISS on a scoresheet.
- Context. A qualifiers field, covered in the next section, recording that the chance was assisted, came from the small box centre zone, and links back to the pass that created it.
Not every buyer needs every field, which is why the table is tiered. In Dribble's match_events, the Starter shape carries 19 columns, Pro 25 and Max 31. The six Max-only columns are period_id, goalmouth_y, goalmouth_z, blocked_x, blocked_y and qualifiers_json: the period split, where shots ended up, where they were blocked, and the full qualifier set. The pricing page sets out what each tier holds.
What are qualifiers in event data?
A qualifier is a named attribute attached to an event that describes how it happened, beyond the type itself. Event types are deliberately coarse: PASS covers a five yard sideways ball and a defence splitting through ball. Qualifiers carry the difference.
In Merino's row, the qualifiers include SMALL_BOX_CENTRE (where the shot was taken from), ZONE with the value Center, LEFT_FOOTED, ASSISTED, and RELATED_EVENT_ID with the value 726, which is the event ID of the pass that set the shot up. That last one matters more than it looks: it is the thread that stitches a match back together, from a shot to the assist, and from the assist to the pass before it.
Qualifiers are also where derived measures ride. In Dribble's tables, the xG values travel as qualifiers on the shot event: qualifier 321 holds pre-shot xG, the value of the chance before the strike, and 322 holds post-shot xG, often called xGOT, the value of the shot as actually struck. Our explainer on what xG is and how it is calculated works through both in full.
Providers differ in how qualifiers are delivered: some flatten the common ones into their own columns (Dribble does this for fields like is_big_chance and is_through_ball), others leave everything in a qualifier list. Either way, when you compare two event datasets, the qualifier set is where the real depth difference hides.
What do the XY coordinates mean?
Every event carries a location on a standardised pitch grid: x runs 0 to 100 from one goal line to the other, y runs 0 to 100 from one touchline to the other, and the attacking direction is normalised so the team in possession is always playing left to right. A shot at x 94.8, like Merino's, was struck deep in the attacking end, a few metres from goal. A pass received at y 50 was received in the centre channel; at y 5 or y 95, hugging a touchline.
Passes carry a second pair of coordinates, end_x and end_y, for where the ball was received, plus a pass length in metres and a pass angle in radians computed from the two points. With start and end points on every pass, you can rebuild a team's whole passing network, measure how direct they are, or find every switch of play without watching a minute of video. Coordinates are what turn a list of actions into a map, and they are the input underneath almost every modern metric: pitch control, packing, expected threat and xG all start from where events happened.
How is event data different from match statistics?
Statistics are what you get when someone else has already counted the events for you, their way. The counting is convenient, but it is a one way door. A stats feed will tell you a team attempted 14 shots; it cannot tell you that 9 of them came in the twenty minutes after they went behind, from an average position 25 metres out, unless the provider chose to cut the data exactly that way in advance.
Event data hands you the counting. The same 1,574 rows from the Palace and Arsenal match can answer questions the stats page never asked: how a team's pass map changed after a substitution, which player received most often between the lines, or whether a striker's shots are getting closer to goal over a run of matches. The trade is work. Events need tooling and some fluency to query, which is why the two products sit at different prices and why serious modelling starts from events. If you are choosing between the two, our guide to football data APIs walks through how event feeds are delivered and what to check before you buy.
Frequently asked questions
How many events are there in a football match? Typically 1,500 to 2,500 on-ball events, depending on the provider and how finely actions are split. Passes are the large majority: 920 of the 1,574 events in the match used above.
Can event data be joined to other tables? Yes. Every event carries match, team and player IDs, so events join to fixtures, lineups and player tables. That is how per player event totals and pass networks are built.
Do all providers record the same events? The core vocabulary is similar across providers, but definitions, qualifier sets and coordinate conventions differ in detail. Compare like with like, and check the qualifier depth before you switch a model between feeds.
The event counts and the walked row in this article come from Dribble's Max tier sample of Crystal Palace v Arsenal, 24 May 2026 (1,574 events), published in full on Kaggle for anyone who wants to inspect a real event file before buying anything.