Introducing the next generation of Pressure Index

What changed in this update

The most important clarification is conceptual:
– The current Pressure Index is a relative scoring‑pressure metric that evaluates which team generated the strongest scoring threat over a recent window of play.
– The updated version improves how pressure is computed by introducing a machine learning model. This model estimates the probability of each team scoring a goal within 5-minute time windows, based on the in-game events that occurred during each window.

How it works

1) Rolling five‑minute scoring window (goals excluded)

The model looks at match activity every five minutes, excluding goals. It then identifies which team posed the greater scoring threat in that period.

2) Weighted events, not simple counting

Attacking indicators (for example, shots, dangerous attacks, and other attacking signals) are not treated equally. They are weighted by the model so that actions historically more connected to goal‑scoring contribute more to the score.

3) Relative output (not a calibrated probability)

Pressure Index values represent a normalised difference between the scoring probabilities of each team (home vs. away), designed to reflect competitive advantage — not a literal ‘X% chance of a goal in the next minute’.

4) Descriptive, not predictive

The Pressure Index is descriptive, not predictive — it reflects what is happening in the game every 5 minutes, not what is likely to happen next.

How to interpret the Pressure Index values

Pressure shows advantage, not certainty

A higher Pressure Index means one team had a stronger run of meaningful attacking threat in the most recent evaluation window. It does not mean a goal is imminent, and it should not be interpreted as an absolute probability.

One team positive at a time

The Pressure Index is designed as a relative measure: in a given moment, only one team has positive pressure, while the other is at (or near) zero.

Key improvements vs the previous Pressure Index

A scoring model rather than a descriptive counter

The updated version is designed to better reflect chance-quality signals by weighting events rather than simply accumulating “pressure actions”.

Cleaner scaling for visualisation

The updated model is bounded and more consistent for charting (compared with earlier behaviour where values could be much larger), improving usability in visual dashboards. The pressure index now scales naturally between 0% and 100% for each team.

Better handling of trend timelines

Trend handling has improved over time (period/minute separation, injury time clarity), making pressure plotting over a match safer and less ambiguous.

Technical implementation (Sportmonks API)

How to retrieve Pressure Index data

Pressure Index is exposed as a trend include on fixture endpoints.
– Include name: pressure
– Pattern: request a fixture and enrich the response with the pressure include.

Example request:

GET https://api.sportmonks.com/v3/football/fixtures/{fixture_id}?api_token=YOUR_TOKEN&include=pressure

Response shape (high level)

The API returns a pressure array containing time-stamped records. Each record includes:
– fixture_id
– participant_id (team)
– minute
– pressure (numeric value)

Working with the timeline: minutes, periods, and sorting

Pressure is trend data. For reliable plotting and analysis, sort records by match time.

Recommended approach:

1. Group by participant_id (team)
2. Sort by period and minute (API 3 separates these to avoid half‑overlap issues)

This helps you handle the first half / second half / added time correctly when visualising or computing derived metrics.

Data coverage and availability

Pressure Index requires sufficient live data coverage. If there is insufficient underlying data, treat pressure values as unavailable rather than inferring them.

Practical guidance:
– Expect limited availability for some lower‑tier fixtures with sparse tracking.
– Implement a UI status indicator (for example: “insufficient data coverage”) when pressure cannot be computed.

Known limitations (what to expect in real usage)

Trend gaps

Trend-type data can show gaps when values remain unchanged for a while or when coverage is low. In practice, this means you might see pressure points at minute 39 and then the next update later.

Scheduled maintenance / brief delays

There may be rare short interruptions due to scheduled system tasks; applications should use retry logic and handle missing pressure gracefully.

Roadmap: what’s coming next

We plan to release a live prediction model in the future. The current Pressure Index should be understood as a relative scoring pressure signal over a recent five‑minute window, not a short-horizon goal probability forecast.

Get started

If you’re already using Sportmonks, you can start retrieving Pressure Index trend data via the pressure included on fixtures.

If you’re building visualisations, consider plotting pressure per team over time and annotating the graph with key match events (goals, red cards, substitutions) for context.

About Sportmonks

Sportmonks delivers sports data through APIs for developers building football applications, including livescores, match events, statistics, and trend-based analytics like Pressure Index.