Progress over seasons

Define progress before opening the dashboard

Progress is not synonymous with a larger number. More running distance may reflect longer playing time. A higher peak speed may reflect a clearer opportunity to sprint. A lower heart rate may accompany improved efficiency, but it can also reflect a different task or an unreliable signal. Begin by defining the performance capability you want to improve.

For one athlete, the objective might be tolerating repeated high-speed efforts within a planned programme. For another, it might be completing a standard submaximal task with a different internal response. A wearable can help observe aspects of these objectives, but the measure must match the intended adaptation. Otherwise, the reporting system rewards whatever is easiest to count.

Build a baseline from comparable observations

A baseline is a reference distribution, not a single best day. Collect repeated observations using a consistent protocol and enough contextual information to identify meaningful differences between sessions. There is no universal number of days that makes every baseline trustworthy. The required history depends on the variability of the measure, the athlete and the decision.

Record device identity, placement, software version, surface, task instructions and relevant environmental conditions. Document interruptions and deviations instead of quietly discarding inconvenient results. If a test is intended to be submaximal, keep the external task comparable. If effort varies freely, the internal response becomes harder to interpret as a change in capacity.

Source note: the Australian Institute of Sport and UK Sports Institute are examples of applied performance-support institutions. Their broader context is useful, but a baseline protocol still needs to be justified for your own setting.

Separate three different questions

First, could the difference be explained by measurement error? Second, could it be ordinary day-to-day variation? Third, is it large enough to matter for the performance decision? These are related but distinct questions. A change can be detectable without being useful, and a potentially useful small change can remain difficult to distinguish confidently from noise.

Estimate typical variation under your own protocol where feasible. Inspect individual observations as well as summaries, and be cautious when the baseline contains very few sessions. Statistical uncertainty should become more visible when data are sparse, not disappear behind a decimal place. Avoid presenting a universal percentage change as meaningful for every metric and athlete.

Precision on a screen is not certainty about adaptation.

Keep opportunity separate from capacity

Competition data are shaped by tactics, opposition, playing time and role. A forward making fewer high-speed runs may have received fewer opportunities rather than lost physical capacity. Dividing totals by minutes can add context, but it does not remove the influence of match demands or substitution patterns. Rate measures and absolute exposure answer different questions.

Triangulate match observations with planned training and suitable performance tests. A standardised sprint assessment, for example, asks a different question from the fastest moment recorded in a match. Do not substitute one for the other simply because both report speed. Explain what each observation contributes and where its interpretation stops.

Treat equipment changes as a break in the story

Replacing hardware, updating firmware or switching an analysis platform can alter output without any biological change. Preserve the old definitions and export enough metadata to understand the transition. If practical, compare systems during a planned overlap period using the same tasks. An overlap can reveal differences, but it does not guarantee interchangeability.

Keep original files where your governance arrangements allow, alongside processed summaries and a change log. Avoid silently recalculating historical values under new definitions and then presenting the result as an unchanged record. When comparability is uncertain, show the boundary in the chart and explain it in plain language.

Use recovery measures as supporting evidence

Resting heart rate, heart-rate variability and sleep estimates may contribute context, but collection conditions and processing matter. A single low recovery score does not establish that an athlete cannot train. Consumer sleep estimates should not be treated as diagnostic sleep studies. Discuss persistent sleep concerns, unusual symptoms or unexplained deterioration with qualified professionals.

Source note: the American Academy of Sleep Medicine position statements distinguish consumer sleep technology from clinical assessment. Device-derived sleep stages and proprietary recovery scores require their own validation.

Review the decision, not just the chart

A monitoring system should help staff learn whether their decisions were useful. Record the expected response when changing a programme, the observations you will review and the time frame appropriate to the adaptation. If the expected improvement does not appear, reconsider both the training hypothesis and the measurement process.

Keep the athlete involved in that review. Feeling more confident during repeated efforts, recovering better between sessions or executing a skill under fatigue can add important context. These accounts do not invalidate objective measurement; they help establish whether the measured change corresponds to something meaningful in the athlete’s working life.