The watch does not lie, but it also does not know everything
The sports watch on your wrist is one of the most useful tools ever made for those who train. It records pace, heart rate, elevation and builds a history memory could never achieve. The problem is not the device: it is how we read the numbers. Some data are measured with high accuracy. Others are estimates with a large margin of error, useful for tracking trends, but dangerous when treated as absolute truth.
This text separates what the watch measures well from what it merely estimates, based on published validation studies. The idea is simple: use the data as a compass to know the direction, never as a judge to deliver the final verdict.
GPS: good for the route, imprecise at the decimal
GPS distance is reliable for most workouts, but it has real error. Gilgen-Ammann, Schweizer and Wyss tested eight watches from Apple, Coros, Garmin, Polar and Suunto in city, forest and track. The mean absolute percentage error sat between 3.2% and 6.1%, and distance was systematically underestimated in urban areas and forest, where buildings and trees disturb the signal.
Gilgen-Ammann, Schweizer & Wyss, "Accuracy of Distance Recordings in Eight Positioning-Enabled Sport Watches", JMIR mHealth uHealth, 2020
Heart rate: the strap beats the green light
Here is the most important and least understood difference. The optical sensor on the wrist (that green light) measures the heartbeat from the variation of light reflected by the blood. It works well at rest and low intensity. But at high intensity, with the arm swinging, sweat and less blood flow to the skin, it misses more: studies point to deviations of 5 to 15 bpm relative to the electrocardiogram, while the chest strap stays at 1 to 2 bpm.
The American College of Cardiology showed in 2017 that, in a test with several devices, the chest strap was the most accurate, and wrist sensors missed more at high intensity. The wrist sensor tends to "lag" on rapid changes, because it averages over a few seconds. In interval training, that corrupts precisely the moment you want to measure.
VO2max, HRV, sleep and calories: this is estimation
These four metrics are not truly measured by the watch. They are inferred by algorithm from HR, pace and movement. Treating them as exact values is the most common mistake.
Estimated VO2max. The mean percentage error was about 16% in a study with the Apple Watch Series 7, with reliability classified as poor, and 7 to 8% on the Garmin Forerunner 245. In the Garmin study, the error tends to be larger precisely in highly trained athletes, reaching about 10% in that group. The number serves to see whether your fitness rises or falls over the months, not to pin down your real aerobic capacity.
Calories. The most fragile metric of all. In the Stanford study conducted by Shcherbina and colleagues, the best device missed energy expenditure by 27% on average, and the worst reached 93%, even with HR within 5% error. Estimating how many calories you burned remains, in practice, an educated guess.
Shcherbina et al., "Accuracy in Wrist-Worn, Sensor-Based Measurements of Heart Rate and Energy Expenditure in a Diverse Cohort", J. Personalized Medicine, 2017 (Stanford)
Sleep. Watches get 60% to 70% of sleep stages right compared with polysomnography, the reference exam. A well-documented trend is overestimating deep sleep, because the device relies on immobility and low heart rate, which also appear in lighter phases. They work well to see whether you slept little or a lot, poorly to say "you had 42 minutes of deep sleep".
HRV and readiness. Heart rate variability (HRV) reflects the balance of the autonomic nervous system and is one of the most useful estimates. Reviews indicate that the RMSSD trend over days helps monitor recovery, fatigue and overtraining risk. A single day's value matters little; the line over weeks matters a lot.
The warning nobody gives you: the elite is not you
Many apps show numbers from top athletes, or compare your HRV, VO2max and volume with "elite" bands. Careful. One study compared developing elite-level athletes with recreational practitioners and found the former training about 17 hours per week, versus 6 hours for recreational ones. And, despite training nearly triple, the elite reported more sleep, better sleep quality and less stress. They do not handle more training through sheer willpower: they handle it because they have a structured routine, time to recover, and a support structure most of us mortals do not have.
There is a concrete health risk in this. The International Olympic Committee (REDs consensus, Mountjoy and colleagues, 2023) describes how chasing "athlete numbers" without enough available energy sickens men and women: hormonal drop, weak bones, low immunity and worsening performance. The watch does not see that. It shows a green number and you think you are doing well.
How to use the watch without becoming its hostage
The practical rule: trust what is measured (HR by strap, pace, time, distance with a margin) and read what is estimated (VO2max, HRV, sleep, calories) as trend, looking at the curve over weeks, never a single day's data. If the number on the screen contradicts how your body feels, the body takes priority. Poor sleep, persistent pain or unusual tiredness call for rest, even with "readiness" showing green.
