A Glossary of Physical Metrics in Football: High-Speed Running, Sprints, Accelerations, PlayerLoad and Metabolic Power Explained

Football physical metrics are measurements of how a player moves rather than what a player does with the ball: how far, how fast, how often they accelerate or brake, and how much strain that movement creates. They come from tracking systems, not from event logs like the ones behind shots or passes on RubiScore and similar match pages.

This glossary groups the main terms by the question each one answers, from the hardware that collects the data to the workload ratios that sports scientists build on top of it. Event-based match data, such as scores, lineups and live statistics, can be followed at https://rubiscore.com, while the physical layer described here usually sits with clubs, leagues and specialist tracking providers.

Where Does Physical Data Come From?

Every physical number depends on the system that captured it, so the source terms come first.

Optical tracking. Fixed stadium cameras follow every player and the ball, producing x and y coordinates many times per second. Research on the Premier League, for example, has used TRACAB camera systems to study running across several seasons.
EPTS (electronic performance and tracking systems). FIFA's umbrella term for devices that track player movement, covering both wearable units and optical systems. The IFAB allowed wearable technology in official matches in 2015, and FIFA launched its quality programme for wearable systems in 2017, widening it in 2019 to performance tests for optical devices too.
GPS units. Wearable devices, usually worn in a vest between the shoulder blades, that estimate position and speed from satellite signals. They are standard in training, where stadium cameras are absent.
Accelerometer or IMU. An inertial measurement unit inside the wearable that records acceleration on three axes. It captures jumps, collisions and changes of direction that positional data can miss.
Broadcast tracking. Positions estimated from the television feed with computer vision. It covers many more leagues, but players who leave the camera frame create gaps, which affects metrics that rely on continuous data.
Sampling rate. How many position readings a system takes per second, expressed in hertz. Higher rates generally capture short, sharp movements such as accelerations more faithfully.

What Do the Volume Terms Measure?

Volume terms describe how much movement occurred, without saying how hard it was.

Total distance. The ground covered across a match or session. In the 2010 Serie A study by Osgnach and colleagues, outfield players averaged just under 11 km per match, a figure in line with the commonly cited range of roughly 10 to 12 km.
Metres per minute. Total distance divided by minutes played. It lets analysts compare a starter with a substitute who played 25 minutes, and it is the simplest measure of running density.
Ball-in-play adjustment. Some analyses express running per minute of ball in play rather than per minute on the clock. A match with long stoppages produces less running, and this adjustment removes much of that effect.
Minutes played. Not a physical metric itself, but the denominator for most of them. Without it, volume numbers mostly reflect selection.

What Do the Speed-Band Terms Mean?

Speed-band terms split distance by how fast the player was moving. The boundaries are the single biggest source of confusion in this vocabulary.

Speed zones. Bands such as walking, jogging, running, high-speed running and sprinting. Each provider and study sets its own boundaries, so two "sprint" figures may not describe the same thing.
High-speed running distance (HSR). Distance covered above a speed threshold. A common research convention for the Premier League is above 5.5 metres per second, which equals 19.8 km/h.
High-intensity running distance. In some studies a band inside HSR, for example between 5.5 and 7 metres per second. Other sources use the term loosely for everything above the high-speed threshold, so the definition must be checked.
Sprint distance. Distance covered above the sprint threshold, commonly 7 metres per second, or 25.2 km/h, in Premier League research.
Sprint count or efforts. The number of separate runs that cross the sprint threshold, usually with a minimum duration. Counts describe repeated explosive work better than distance.
Top speed or peak velocity. The fastest speed recorded in a match or period. Single peaks are sensitive to measurement error, which is why one provider, SkillCorner, built a percentile-based alternative called PSV-99 for broadcast data, using the 99th percentile of a player's high-speed activity rather than the single highest value, before adding a separate peak velocity metric.
Relative thresholds. Speed bands set from each player's own capacity, such as a percentage of their peak speed, instead of one figure for everyone. A 2024 Premier League study compared both approaches and found that relative bands can record more high-intensity running for some players than generic cut-offs do.

What Do the Acceleration and Mechanical-Load Terms Capture?

Speed alone misses much of football's strain. A player who stops, turns and restarts five times in ten metres may never reach a high-speed band.

Accelerations and decelerations. Counts or distances of efforts where speed rises or falls faster than a set rate, measured in metres per second squared. Reviews note that thresholds differ widely across studies, from below 2 to 4 metres per second squared, and that counts above a threshold are less reliable than average values.
High-intensity efforts. A combined count of sprints plus hard accelerations and decelerations. Definitions are provider-specific, so the components should be listed before the total is compared.
PlayerLoad. Catapult's accelerometer metric, described by the company as the sum of accelerations across all three axes during movement, scaled by dividing by 100. It is an arbitrary unit meant for comparing relative workload, and Catapult warns that absolute values depend partly on running style. A 2020 review in the Journal of Human Kinetics argued that the literature is inconsistent about what the number actually measures.
Metabolic power. An estimate of the energy cost of movement per kilogram per second, expressed in watts per kilogram. The approach, developed by di Prampero and applied to football by Osgnach and colleagues in 2010, treats acceleration as running uphill, so a sharp burst from standing can register as demanding even at modest speed.
HMLD (high metabolic load distance). STATSports' metric for distance covered while metabolic power exceeds about 25.5 W/kg. The company says this roughly matches steady running at 5.5 metres per second, or significant acceleration and deceleration, so HMLD folds high-speed running and speed changes into one figure.

What Is the Difference Between External and Internal Load?

Everything above measures external load, the work a player performs. Sports science pairs it with internal load, the body's response to that work.

External load. Distance, speed bands, accelerations and accelerometer metrics. It answers what the player did.
Internal load. Physiological and perceptual responses, mainly heart rate and perceived exertion. It answers how hard that work was for this player on this day.
Session RPE. A method popularised by Foster and colleagues in 2001: the player rates the session's intensity on a modified 0 to 10 scale, and the rating is multiplied by the session's duration in minutes. A 90-minute session rated 5 gives 450 arbitrary units.
Heart-rate load. Measures such as time spent in heart-rate zones, or TRIMP-style scores that weight minutes by heart-rate intensity. Heart rate lags behind short sprints, so it suits sessions better than single efforts.
Acute to chronic workload ratio (ACWR). Recent load, often the last week, divided by longer-term load, often a rolling four-week average. It was proposed as a way to flag sudden spikes in workload, but later work raised statistical concerns, including mathematical coupling when the recent week sits inside the longer window, so it should be treated as a monitoring prompt rather than an injury predictor.

How Do These Terms Fit Together?

The vocabulary works as a ladder, with each rung adding a judgement to the one below.

Raw signal. Positions from cameras or GPS, and accelerations from an IMU.
Derived movement. Speed and acceleration calculated from those signals.
Thresholded counts. HSR, sprint distance, sprint efforts and acceleration counts, each depending on a chosen cut-off.
Modelled load. PlayerLoad, metabolic power and HMLD, each resting on a formula or physiological assumption.
Workload management. Session RPE, heart-rate load and ACWR, which combine physical output with time and the player's response.

Most public coverage stops at rung three. Most club decisions are made at rungs four and five. A headline about a player's "sprints" and a sports scientist's view of that player's load can therefore describe very different things.

Physical numbers also sit beside, not inside, the event data most fans see. A live-score platform such as RubiScore tracks goals, cards, substitutions and match statistics. A substitution or a red card on that event timeline often explains a sudden change in a team's running more convincingly than fitness does.

What Are the Common Reading Pitfalls?

Comparing across providers. A sprint threshold of 7 metres per second in one dataset and a lower one in another makes their sprint figures incompatible.
Mixing optical, GPS and broadcast data. Each system measures with its own error, so the same player can post different totals from different systems in the same match.
Ignoring minutes and game state. Totals rise with time on the pitch, and teams chasing a game often run differently from teams protecting a lead.
Treating more as better. Covering more distance is not the same as playing well. Positioning that saves running can be a sign of quality.
Comparing absolute PlayerLoad between players. The vendor itself frames the unit as relative, with individual baselines the safer comparison.
Reading one peak speed as a true maximum. A single reading may reflect a spike in the data. Percentile-based or repeated measurements are more robust.
Assuming a workload ratio predicts injury. ACWR can prompt a conversation about load, but its predictive value is disputed.

A short check covers most of these. Ask which system collected the data, which thresholds define each band, whether the figure is per match, per minute or per minute of ball in play, and whether the comparison is against the player's own baseline or someone else's. With those four answers, a physical metric becomes a description of movement that can be read in context, rather than a number that looks precise and says little.