Real-Time Data Tracking

The single most important conceptual distinction in modern athlete monitoring is between external load: what the athlete did, measured in physical terms (kilometres covered, sprints, accelerations, weight lifted, repetitions), and internal load: the biological cost the athlete experienced in doing it (heart rate, perceived exertion, hormonal response). External load is the dose prescribed by the coach. Internal load is the body’s response to that dose.

Research evidence. Impellizzeri, Marcora, and Coutts (2019) revisited their original framework fifteen years on and emphasised that monitoring systems built only on external load metrics miss the individual variability that determines training adaptation and injury risk. Two athletes performing the same prescribed session experience different internal loads depending on fitness, fatigue, sleep, environment, motivation, and a dozen other factors. The coach who tracks only external metrics is tracking only half the picture.

The framework has direct practical consequences:

• Comparing two athletes’ responses to “the same session” using only external metrics is misleading; you are comparing prescription, not response.
• Tracking internal load over time (sRPE, HR-derived TRIMP) lets you see when a previously well-tolerated dose has become a strain.
• The internal-external comparison is in itself diagnostic: an external load that is unchanged but produces a higher internal load is one of the clearest early signals of accumulated fatigue.

Modern team-monitoring platforms provide real-time feeds that follow a common architecture: GPS and accelerometer signals stream from athlete-worn devices to a base station, are processed locally and/or in the cloud, and surface on the coach’s tablet within seconds. Brand-specific implementations (Catapult OpenField Live, STATSports Apex Live, Polar Team Pro Live, and others) vary in detail; the generic features a coach should expect are:

• Per-athlete live metrics (HR zones, total distance, high-speed running, accelerations/decelerations, PlayerLoad-type composite).
• Customisable thresholds that trigger visual alerts (colour change, audible alarm).
• Drill or period segmentation so post-session analysis aligns with the session plan.
• Export of raw and processed data for further analysis.

The honest reality: not every “live” metric is reliable enough to act on within seconds. Heart rate, distance, and impacts update with sufficient stability. Composite metrics with smoothing windows (some PlayerLoad variants, some asymmetry indices) are more meaningful as session totals than as live readouts. As a rule of thumb: in real time, trust simple metrics; trust composites at session debrief.

Most dashboards use a recurring set of visualisations. Coaches who can answer the same question (“what is this picture telling me?”) across the five families are dashboard-literate.

Time series. The horizontal axis is time; the vertical axis is a metric (e.g. heart rate). Strengths: shows trends, peaks, recoveries. Watch for: y-axis scaling that exaggerates or suppresses variation.

Heat map. Spatial intensity, most often used in field sport (zones of the pitch where activity was concentrated). Strengths: spatial coverage, tactical insight. Watch for: heat maps include both running with the ball and running without; they do not reflect tactical intent on their own.

Radar (spider) chart. Multi-axis comparison of metrics for one athlete or one session, often used for athlete profiles. Strengths: at-a-glance multidimensional view. Watch for: arbitrary axis scaling that visually distorts apparent strengths.

Box plot. Distribution of a metric across the squad, with median, quartiles, and outliers. Strengths: surfaces outliers immediately. Watch for: small squad sizes make box plots unstable; treat with caution under n = 10.

Scatter plot. Two metrics plotted against each other to surface relationships (e.g. weekly sRPE versus session HR). Strengths: shows correlation and outliers in two-dimensional space. Watch for: correlation is not causation; never act on a scatter pattern alone.

Before any decision is made on a metric, the coach checks the data. The most common quality issues:

• GPS dropout indoors and in dense environments. Indoor halls or stadiums with heavy steel structures produce intermittent satellite signal. A drill where a player apparently covered 50 metres in 10 minutes was almost certainly indoors; treat the metric as missing.
• Motion artefact in optical HR. Wrist-based PPG sensors give erratic readings during gripping tasks, vibrations, and short high-intensity bursts. If a player’s HR shows a 20-bpm jump and drop within five seconds, scepticism is warranted.
• Sensor drift. Long sessions, hot conditions, or sweat-saturated chest straps produce gradually distorted readings. Compare end-of-session HR with start-of-session HR at similar effort.
• Time-sync issues across devices. When data from multiple devices are aligned, small clock offsets can confuse derived metrics.
• Missing or partial files. A coach who learns to spot a 73-minute file labelled “90 minutes” before computing weekly load is a coach whose weekly reports are more trustworthy.

The discipline: a metric that looks too good, or too bad, to be true probably is.

Just because a coach can see a metric live does not mean the athlete should see it live. Three questions:

What? Some metrics are actionable in real time (HR for pacing, distance covered for substitution decisions). Others are not (HRV, sleep: these are pre-session inputs). Share what is actionable.

When? During a drill, the athlete is executing; metric talk is a distraction. Between drills or at the end of a block, brief metric-anchored feedback (“your HR didn’t come down between reps; rest longer next time”) is well placed. Trend feedback (“your HRV has been declining for a week”) belongs in a one-to-one debrief, not on the touchline.

How? Frame metrics as information, not verdicts. “Your sprint count was high today, let’s protect tomorrow” lands differently from “You overworked.” Negative-news framing scripts are covered in Module 5.

Worked computation. A 60-minute training session for an athlete returns:

• Total distance: 6,800 m
• High-speed running (>19.8 km/h): 720 m
• Average HR: 152 bpm; Max HR observed: 188 bpm; Resting HR: 56 bpm
• RPE (Borg CR-10): 7

Computed metrics:

• sRPE = RPE × duration (min) = 7 × 60 = 420 arbitrary units (AU)
• TRIMP (Banister, simple form) = duration × HR fraction × weighting, where HR fraction = (HR_avg − HR_rest) / (HR_max − HR_rest). HR fraction = (152 − 56) / (188 − 56) = 96 / 132 = 0.727. Using a typical weighting factor that accounts for non-linear HR response (Banister’s exponential form), TRIMP for this session approximates 94 AU. (See Banister, 1991, for the full equation; commercial implementations vary in the exponential factor.)
• ACWR. Suppose this athlete’s chronic load (28-day rolling average of weekly sRPE) is 2,400 AU. This week’s acute load (7-day total sRPE) is 3,300 AU. ACWR = 3,300 / 2,400 = 1.375. This sits at the upper edge of the often-cited “sweet spot” (≈ 0.8-1.3) and warrants attention; whether it warrants intervention depends on the athlete’s individual history and the surrounding context (see Module 4 for full treatment, including critiques of the framework).

Practical tip. The first time you compute ACWR by hand, it feels laborious. A spreadsheet template (Template A4) automates this for a squad. The point of doing it by hand once is to internalise what the number means, not to make hand calculation a habit.