Measuring in Running: What a Stopwatch Sees, What a Watch Sees, and What a Coach Sees

The Problem of Measuring Twice and Getting Two Truths

An athlete joins my program with a lab-tested VO2max of 58 ml/kg/min, measured eight months ago during a graded exercise test with a gas analyzer. This week, his watch tells him his estimated VO2max is 61. Neither number is "wrong." They're simply measuring different things, at different times, with different methods — and the athlete, reasonably, doesn't know which one to trust.

This happens constantly in modern running. We have more capacity to measure than ever before, but very little clarity about what each measurement actually means, who should interpret it, and what specific decision it's supposed to inform. That's what this article is about: the difference between measuring in the field and measuring in the lab, between technology measuring and the coach measuring, and why the right question isn't "which data point is more accurate" but "which data point is useful for what I actually need to decide."

The Lab: Precision, But a Snapshot

The lab has something no watch has: control. Constant temperature, a standardized protocol, a calibrated gas analyzer, capillary lactate drawn from real blood rather than estimated by an algorithm. When you measure VO2max, lactate threshold, or running economy in a lab, you're measuring with a level of precision the field simply can't match.

But that precision comes at a cost: it's a snapshot, not a film. It's captured on a single day, under conditions that don't resemble race day — no wind, no uneven surface, none of the accumulated fatigue from an 80 km training week, none of the emotional variable of actually competing. As the review by Siebers and Bizjak points out [PMID:41922043], exercise biomarkers are powerful tools for understanding an athlete's physiological state, but their interpretation depends heavily on the context in which they're collected: the same biomarker can mean something entirely different at rest, during acute fatigue, or in chronic overtraining. A lab number stripped of the training context surrounding it is an incomplete piece of data.

What's more, the lab measures what's measurable with expensive instruments — not necessarily what most influences race outcomes. VO2max explains only a fraction of the variance in endurance performance; running economy, the ability to sustain a pace relative to threshold, and psychological tolerance for discomfort explain a good part of the rest, and those are far harder to capture with a gas analyzer than by simply watching an athlete run six repeats of 1000m on the track.

The Field: Context, But Noise

The field — the track, the road, the GPS watch — measures under real conditions, which is both its great advantage and its great problem. A watch estimates VO2max from the relationship between pace, heart rate, and how that relationship shifts over time. It's a statistical model trained on population-level data, not a direct measurement of that particular athlete's gas exchange. That's why two athletes with nearly identical physiology can get different VO2max estimates from the same watch model, simply because of how they run or how the algorithm interprets their heart rate signal.

That doesn't make the watch useless. It makes it useful for something else: trends, not absolute values. If an athlete's estimated VO2max drops consistently over three weeks while training volume stays constant, that's information — even if the absolute number itself is imprecise. Resting heart rate, heart rate variability (HRV), and pace at a constant perceived effort are field metrics that, tracked over time by the same person with the same device, say more about the direction of change than about the exact value.

The stopwatch, meanwhile, is the most honest instrument in all of athletics. It doesn't estimate anything — it measures elapsed time between two events. A 3:05 for 1000m is a 3:05, with no statistical model involved. The stopwatch's limitation isn't its precision; it's its scope: it doesn't tell you why the athlete ran that pace, how it felt, or whether their stride broke down on the last lap.

What the Coach Sees That No Sensor Does

Here, for me, is the central point of this article. Neither the lab nor the watch captures what a coach sees standing on the track: the running mechanics that start to break down on rep five of eight, the face of an athlete who is genuinely at their limit versus one who's faking being at their limit, the subtle shift in cadence fifteen meters before the watch registers any anomaly, the breathing that turns ragged in a way that's different from that athlete's usual fatigue pattern.

This isn't anti-technology romanticism. It's a practical observation: field data collected by someone physically present carries a qualitative dimension that no current sensor replaces. A coach who watches an athlete train week after week builds a mental model of what that person looks like when they're well, when they're at a healthy limit, and when they're entering risk territory. That model is built over hundreds of hours of direct observation, and in my experience, it's the earliest warning sign there is — earlier than any drop in HRV, earlier than any fatigue biomarker.

This matters especially in the context of cardiovascular health. The review by Frandsen and Aaroe on the cardiac effects of prolonged endurance exercise [PMID:40667749] found that both young and older athletes can show acute cardiac adaptations after prolonged efforts, and that the pattern of those adaptations doesn't always correlate simply with training level. A lab value taken months earlier doesn't capture that. A watch can record heart rate, but it doesn't necessarily correctly interpret an early arrhythmia or an unusual recovery pattern. The clinical eye — or at least the trained eye of a coach who knows the athlete's history — remains irreplaceable when it comes to deciding whether a data point warrants medical attention. This is especially relevant given that marathon mortality, while rare, is associated in a meaningful proportion of cases with undiagnosed underlying cardiac causes [PMID:31321073]. No watch replaces a medical evaluation when warning signs appear; what it can do is trigger the alert that leads someone to seek that evaluation in time.

Should You Measure the Athlete, or Measure the Coach?

A question I get often is whether technology should measure the athlete or whether the coach should "measure" using their own judgment — and honestly, framing it as an either/or is the mistake. These are two data sources measuring different things that, combined, correct each other.

The athlete, measured by technology, provides objective, continuous data: pace, heart rate, distance, sleep, HRV. It's high-frequency, low-cost information — the watch measures every single day without anyone needing to be present. Its weakness is that it lacks narrative context: it doesn't know whether that slow pace happened because the athlete was tired, because they were running uphill, or because they had an argument with their partner before heading out the door.

The coach, observing the athlete, provides low-frequency but high-value narrative context: understanding the why behind the data. The weakness here is that the coach can't be present every single day, and their memory and perception — however well-trained — are subjective and prone to bias. A coach can, without meaning to, see what they expect to see.

In practice, the combination works like this: technology generates the alert ("this athlete has had below-baseline HRV for ten straight days"), and the coach decides what that alert means in the context of that specific person ("they've been sleeping poorly because of a job change — this isn't overtraining, but we're easing off intensity this week anyway"). Neither replaces the other. Data without human interpretation is noise; human interpretation without data is unverified intuition.

Why It Matters Who Collects the Data

Here's an idea that doesn't get discussed enough: data changes meaning depending on who collects it and when. A 1000m split timed by the coach on the track, watching the running form live, is not the same data point as that identical split recorded by the athlete's GPS watch — even if the final number is exactly the same. The coach who was there knows whether that split was achieved with a clean stride or with compensations that are announcing an injury three weeks out. The watch only knows that 1000 meters passed between two points in a given amount of time.

This has a direct application in training design. For example, when comparing high-intensity training modalities, a study on cycling versus running HIIT for fat loss and gut microbiota composition in overweight men [PMID:38233990] found differences between modalities that can only be explained by filling in the picture with contextual variables — adherence, perceived effort, associated eating patterns — that no power or heart rate sensor captures on its own. The lab data (body composition, metabolic biomarkers) needed behavioral context to become interpretable.

Practical Recommendations

Use the lab to calibrate, not to govern. A graded exercise test once or twice a year is useful for setting training zones with real precision. Don't replace that calibration exercise with constant watch estimates, but don't ignore the watch's trends between tests either.

Use the watch for trends, not absolute values. If estimated VO2max goes up or down, ask yourself why the trend changed — not whether the number itself is correct.

The stopwatch is still the gold standard for real performance. Track paces, timed live, are the single most reliable piece of data that exists about what an athlete can do today.

Don't substitute data for presence. If you train alone, try to film yourself at least once every few weeks to observe your own mechanics, because no watch captures that qualitative dimension.

When signals are ambiguous, prioritize professional evaluation over self-interpretation of the data, especially in the cardiovascular domain, where the consequences of a misread can be serious [PMID:31321073] [PMID:40667749].

Log context alongside the number. A pace with no note about how it felt, what you ate, or how much you slept is an orphaned data point. The most valuable biomarker or metric is the one that comes with its context attached [PMID:41922043].

The Bottom Line

Measuring well in running isn't about having more sensors. It's about knowing what question you're asking each instrument, and not asking one to answer what only the other can answer. The lab answers: "What is your physiological capacity today, under controlled conditions?" The watch answers: "How is your condition changing over time?" The stopwatch answers: "What can you do right now?" And the coach, present and paying attention, answers the question no instrument can answer alone: "What does all of this mean for this person, at this point in their life and their season?" That last question is, in the end, the one that matters most.

References

• Siebers M, Bizjak DA. Exercise biomarkers. Adv Clin Chem. 2026. [PMID:41922043]
• Couvert A, Goumy L. Effects of a Cycling versus Running HIIT Program on Fat Mass Loss and Gut Microbiota Composition in Men with Overweight/Obesity. Med Sci Sports Exerc. 2024. [PMID:38233990]
• Frandsen J, Aaroe M. Cardiac Effects of Prolonged Endurance Exercise in Young and Older Athletes. Scand J Med Sci Sports. 2025. [PMID:40667749]
• Traish AM, Zitzmann M. The complex and multifactorial relationship between testosterone deficiency (TD), obesity and vascular disease. Rev Endocr Metab Disord. 2015. [PMID:26590935]
• Dayer MJ, Green I. Mortality during marathons: a narrative review of the literature. BMJ Open Sport Exerc Med. 2019. [PMID:31321073]