GPS tracking in team sports: metrics, meaning, and limits

Walk into any professional field-sport training ground and you will see the same small black pod sitting in a pouch high on each athlete's back. That is the GPS unit, and over the last fifteen years it has become the default instrument for measuring external load in football, rugby, AFL, and every invasion sport played on a pitch. It is the closest thing the field has to a standard, and like any standard it is most useful to people who understand exactly what it can and cannot do.

Modern units are not pure GPS any more. They pull from multiple satellite constellations, so the right term is GNSS, and they pair the position fix with an onboard accelerometer, gyroscope, and magnetometer. The position data tells you where the athlete moved and how fast; the inertial sensors fill in the work the satellite cannot see, such as the load of a tackle or a jump.

The core metrics

Total distance

The simplest output, and the most trustworthy. Total distance is the full ground covered in a session, and it is the headline volume number. Across the evidence, GPS units of every sampling rate track total distance with acceptable accuracy and good reliability, which is why it is the metric coaches lean on for overall workload. Its weakness is that it is blind to intensity: ten kilometres of jogging and ten kilometres with twenty sprints in it read the same on this one number, which is why total distance is never used alone.

High-speed running

High-speed running is the distance covered above a velocity threshold, often set somewhere around 5.5 metres per second, though the exact cut-off varies by sport and provider. This is the metric that captures the demanding work, the sprints and hard runs that drive both adaptation and hamstring risk. It is also where accuracy starts to degrade. The brief review of GPS validity in team sport found that units measure total distance and peak speed acceptably, but become less reliable for higher-intensity activities, so a high-speed figure should be read as a good indicator of trend rather than a precise count of metres.

One practical consequence: because thresholds differ between systems, a high-speed running number from one provider is not directly comparable with another. Comparisons only hold within a single, consistent setup.

Accelerations and decelerations

Sprinting in a straight line is only part of the demand in team sport. Much of the real cost is in the changes of pace: the explosive accelerations out of a standing start and the hard decelerations that brake a player at speed. Decelerations in particular load the muscles eccentrically and carry their own injury cost, so the counts of high accelerations and decelerations are watched closely.

These are also the metrics where GPS is least at home. Rapid changes of direction degrade position accuracy, and lower-sampling-rate units in particular struggle with the short, sharp efforts that make up so much of an invasion sport. Higher-frequency 10 Hz units handle them better, which is one reason they have become the practical standard, but acceleration counts should still be treated as a sensitive but noisy signal, best read as a trend over weeks rather than a hard total from a single session.

Body load from the accelerometer

To capture the work the satellite misses, units derive a composite load metric from the onboard accelerometer, the best-known being PlayerLoad, formalised by Boyd and colleagues using the Australian Institute of Sport's microtechnology. It sums the rate of change of acceleration across the three axes, so it counts the contacts, jumps, and collisions that move the body without covering ground. It is the metric that makes GPS useful in the contact and court sports where distance alone undersells the demand. The caveat is that these composite scores are device-specific and not standardised across vendors, so a load number means something only relative to itself.

How the sampling rate changes the picture

Not all units are equal, and the difference comes down to how many times per second the device fixes a position. Early units sampled at 1 Hz or 5 Hz; the modern standard is 10 Hz. The evidence is consistent: 10 Hz units are the most valid and reliable to date, clearly outperforming slower ones for the high-intensity, change-of-direction movements that define team sport, while pushing beyond to 15 Hz adds no further benefit. The practical takeaway for a programme buying or comparing systems is that 10 Hz is the floor for trusting the high-intensity metrics, and any analysis mixing data from different sampling rates is comparing things that are not the same.

Reading the data honestly

The most important habit with GPS data is to know which numbers to trust at face value and which to read as trends. Total distance and peak speed are solid. High-speed running, accelerations, and decelerations are sensitive and valuable, but noisier, and they are best interpreted as a direction of travel across a block of sessions rather than a precise count from any single one.

Three rules keep a programme out of trouble:

• Compare like with like. Same provider, same firmware, same speed thresholds, or the comparison is meaningless.
• Watch trends, not single sessions. The signal lives in how a metric moves over weeks against an athlete's own baseline.
• Never read external load alone. A distance figure is only half the story until it sits next to the internal cost.

Where GPS fits in the bigger picture

GPS measures the external load: the work done. The 2017 consensus statement on monitoring training loads is clear that external load is only meaningful when paired with internal load, the physiological cost to the individual athlete. A big distance day means one thing for a fresh athlete and another for a fatigued one, and the GPS unit cannot tell them apart on its own.

That is the case for reading GPS inside a connected athlete record rather than a standalone vendor dashboard. When the external-load stream sits alongside recovery scores, wellness check-ins, and any open medical note on the same athlete, a load spike is interpreted in context: is this athlete prepared for it, and how are they recovering from it? That is the model behind physical device integration and Strong's data and insights platform, where GPS becomes one connected layer of the picture rather than a number stranded on its own screen.

Where to go next

For the wider field and how teams turn these signals into decisions, read the sports performance analytics hub. If you want the plain-language foundation first, start with what sports performance analytics is.