GPS, heart rate or RPE: how to measure a footballer’s physical load without drowning in data

Training ends. Two central midfielders have almost the same total distance and a similar high-speed volume. One has a clearly higher average heart rate. The other rates the session 8 out of 10 for perceived hardness; the first rates it 5.

Who had the greater load? The right answer: the question is too broad.

The first dataset describes external load — the work the footballer performed. The second and third show different sides of internal load — the body’s response to that work. That split sits at the core of modern training monitoring: external load is the stimulus; the individual psychophysiological response forms internal load.

So GPS, heart rate and RPE do not so much compete as look at the session from different angles.

First define what you mean by load

In football practice the word “load” is often used for several different things at once.

• “The player covered 8 kilometres” — volume of work done.
• “He spent 18 minutes above a target HR zone” — physiological response.
• “The session felt like 8 out of 10” — subjective effort rating.
• “His ACWR rose” — already a mathematical transform of accumulated values over chosen periods.

Mixing those levels is dangerous. A systematic review of football literature found dozens of internal and external load measures and separately flagged inconsistent definitions and classifications across studies.

For staff the principle is simple: question first — metric second.

If you need to know whether a full-back got the required high-speed running exposure, movement data are the better lens. If you care about the physiological cost of aerobic work, heart rate adds another layer. If the task is how hard the session felt for that player, RPE appears.

GPS/GNSS: the best answer to “what did the player do?”

In football “GPS” is often used as a catch-all for wearable positioning systems, even though modern devices may use multiple satellite systems — GNSS.

Their main value is quantifying player movement. Depending on the system, staff can get total distance, distance in speed bands, HSR, sprint distance, speed, accelerations and decelerations, plus inertial-sensor metrics.

But “objective” does not mean “error-free”. FIFA runs a dedicated Electronic Performance and Tracking Systems (EPTS) testing programme and evaluates accuracy for position and speed across speed ranges.

That matters in practice: two columns labelled “HSR” are not necessarily identical if they come from different devices, different processing algorithms or different speed thresholds.

Where GPS is especially useful. Take MD-3 — three days before a match. Staff want to check whether wide players got the required speed stimulus. Average heart rate will not answer the main question: a player can take a high cardiovascular load in an intense small-sided game and barely reach top speeds.

GPS shows the missing part: whether high-speed bouts and sprint exposures actually happened. The 2026 sided-games consensus notes that such games help several aerobic and football tasks, but are less suited on their own for developing high-speed qualities.

Practical takeaway: high exercise intensity does not guarantee the required top-speed exposure.

Heart rate: not how much the player worked, but the cardiovascular cost

Heart rate answers a different question. Two players can cover the same running volume with different internal responses. Fitness, accumulated fatigue, environment, health status and the nature of the work all matter.

HR metrics include average and maximum values, time in individual zones and derivatives such as TRIMP. In elite football studies, HR-based metrics relate to training-load characteristics but do not become a carbon copy of external measures.

That is why heart rate should not replace GPS. A short acceleration, deceleration, change of direction or sprint creates a mechanical stimulus faster than the cardiovascular system fully reflects it in one number. Conversely, relatively modest locomotor work can carry a high physiological cost.

The main heart-rate error is treating low HR as proof of low load overall. It mainly describes one side of the physiological response. For aerobic interval work it can be central. For short sprints, strength work or aggressive deceleration series, heart rate alone is not enough.

RPE: a subjective metric you should not turn into a “bad objective” one

RPE — rating of perceived exertion — is a subjective rating of how hard the load felt. “Subjective” is sometimes treated as a flaw. Here subjectivity is the object of measurement.

Staff are not asking “how many metres did you run?” but “how hard was this work for you?”

Football often uses session-RPE, or sRPE: RPE multiplied by session duration. Example: 60 minutes × RPE 7 = 420 arbitrary sRPE units.

The number is convenient for tracking individual dynamics and weekly internal load. Alone it says nothing about what created that hardness. 420 units can come from a long moderate session or a shorter, very hard one.

Recent data also show that RPE’s relationship with HR and external metrics depends on exercise format. Authors recommend interpreting RPE with objective measures and task context. So RPE is valuable not despite subjectivity, but because it captures the individual cost of the work done.

GPS vs heart rate vs RPE: who wins?

There is no universal winner.

Why ACWR should not become a “injury traffic light”

Acute:Chronic Workload Ratio compares a recent “acute” load with a longer “chronic” load. Intuitively the idea is attractive: if current load differs sharply from habit, notice it.

The problem starts when the ratio becomes a universal safety boundary. Methodological critique argues that a causal link between managing ACWR and reducing injuries is not established, and that the ratio structure itself creates statistical and interpretive issues.

Evidence is mixed: one semi-professional team study found a weak link with non-contact injuries; a 2025 meta-analysis found associations but urged caution because of heterogeneity and calculation differences.

So the useful practical reading is modest: ACWR can flag a change in load, but it is not a diagnosis and not a predictor of who will get injured. A sharp change is a reason to ask questions, not an automatic reason to pull a player from training.

The most useful signal appears when metrics diverge

The main value of combined monitoring appears not when every metric says the same thing. It appears when they diverge.

Imagine a player with a normal total distance, normal HSR, clearly higher heart rate and RPE 9 instead of the usual 6. External work did not rise; internal cost did. That is not a diagnosis or proof of fatigue. But it gives staff a reason to check context: recovery, previous loads, sleep, wellness, temperature, illness, exercise design and data quality.

The reverse is also informative: high sprint distance, many accelerations, moderate average HR and RPE 5. Calling the session “easy” from heart rate alone would be wrong — it may have created a substantial speed and mechanical stimulus.

How to build a minimal monitoring system without dozens of metrics

Do not start by buying the maximum number of sensors. Start with the decisions staff want to make.

Level 1. No GPS. Minimal workable system: duration + RPE + sRPE + session context. The key condition is a consistent collection procedure.

Level 2. GPS/GNSS available. Add a small set of external metrics tied to your preparation model: total distance, HSR, sprint distance, accelerations and decelerations. You do not need every export column in the daily report.

Level 3. Need a physiological layer. Add heart rate for some sessions — especially where you must estimate the internal cost of aerobic and interval work.

Level 4. Compare, do not collapse. Do not turn metres, heartbeats and RPE into one magic score. Analyse the chain: external stimulus → internal response → subsequent player dynamics.

Five rules without which good sensors produce bad data

• Use stable definitions. If HSR starts at one speed today and another next month, the time series stops being comparable.
• Collect RPE the same way every time. Same scale, same moment after the session, same instruction to the player — otherwise ratings are not comparable.
• Do not conclude from one number. If GPS, heart rate and RPE diverge, check context and data quality first — do not change the plan automatically.
• Separate training from matches and playing time from the full event. 15 minutes and 90 minutes are fundamentally different exposures.
• Compare a player mainly to themselves. Position, role, format, starter/substitute status and microcycle structure strongly shape load metrics.

What staff should measure in the end

Do not ask: “GPS, heart rate or RPE — which is more accurate?” Ask:

• What did the player do? — an external-load metric.
• How did they respond? — internal-load metrics.
• How hard did the work feel? — RPE.
• How different is today’s picture from their own baseline and recent weeks? — ACWR.

Good monitoring does not remove uncertainty. It makes uncertainty visible. The sooner staff stop hunting for one number that supposedly describes a footballer completely, the more useful sensors, surveys and training data become.