Understanding running watch data without getting lost

You finish your workout, open your watch app, and everything shows up at once: a pace, a heart rate, a VO2max, a "recovery status." All these numbers appear in the same font with the same confidence. Yet they are not worth the same.

There is a simple way to sort them out: ask each number how it was obtained. A distance is measured, a VO2max is estimated from other numbers, and a recovery score is calculated from those estimates. The further a number is from a direct measurement, the more caution you need when reading it.

Three types of data: measured, estimated, interpreted

Sorting each number into the right category already saves you half the bad decisions.

These figures come from specific models, often older ones. Treat them as a rough guide, not a guarantee for your watch.

Distance and pace: what GPS does well, and when it gets it wrong

Distance is the most reliable metric on your watch, but it is not as accurate as many people think. The Swiss Federal Institute of Sport tested 8 watches with GNSS (GPS and its cousins). Average error ranged from 3.2% to 6.1% depending on the watch.

Terrain matters. In this study, distances were generally underestimated and less accurate in urban areas and in forests, because buildings and trees scramble the signal. On a running track, 6 of the 8 watches overestimated the distance, but with good precision. A track is therefore a good place to test your watch: run several laps and compare with the actual distance.

In practice:

• In a forest or between tall buildings, your watch may record less distance than you actually covered, which makes your average pace look a bit slower than it really was.
• Compare workouts on similar routes. Over 10 km (6.2 miles), a 3% gap is 300 m (980 ft). At a pace of 6 min per km (about 9:40 per mile), that shifts your average pace by about 11 seconds per km (about 18 seconds per mile). A difference like that between a city run and a run on open ground says nothing about your fitness.

Wrist heart rate: good at rest, less reliable under effort

The optical sensor on the wrist reads blood flow using light. It is more reliable at rest than during exercise.

For the Apple Watch, a meta-analysis of 82 studies gives a very small average bias of -0.27 beats per minute. But the limits of agreement run from -7.19 to +6.64: even on an accurate measurement, the gap can reach seven beats per minute in either direction. This result only applies to the Apple Watch.

What this means for you:

• To be cautious, base your judgment on the average heart rate of a steady-pace run rather than on precise values taken in the middle of intervals.
• If you want precise values during intervals, try a chest strap and compare it with your watch.

HRV, VO2max, load: estimates to read as a trend

Here, the watch no longer measures directly what it displays. It calculates.

HRV, which varies a lot from one device to another

HRV is the variation in time between two heartbeats. It is used as a recovery indicator. A 2025 study compared the nighttime HRV of five devices against an ECG, across 13 healthy adults and 536 nights. For three of them, the mean absolute error was 5.96% for the Oura Gen 4, 10.52% for the Garmin Fenix 6 and 16.32% for the Polar Grit X Pro. Same heart, same night, very different error levels. So look at your HRV over several nights in a row, not on a single night.

VO2max, a number that moves

VO2max is the maximum amount of oxygen your body can use. Your watch estimates it from your pace and heart rate. In a 2019 study of wrist activity trackers, the mean absolute error was above 10% for every device except the Garmin Forerunner 920XT (7.3%), which yet significantly underestimated VO2max.

Load, an indicator specific to each brand

On Garmin, acute load is a weighted sum of the oxygen your body consumes after exercise (EPOC, excess post-exercise oxygen consumption) over the last few days. The "optimal" range depends on your fitness level and your history. It is a proprietary indicator with no universal threshold: one brand's load cannot be compared with another's.

The error figures come from research; the rule of following the trend comes from coaching practice. With margins above 10%, the trend over several weeks tells you more than a single number, but it is still limited by the quality of the device.

What you can do from your next workout: the 5-minute routine

After your workout, once the watch has synced, go through these points in order: measured data first, then estimated data, then interpretation.

• ✓Time and distance: does the route look like last time? In a forest or in the city, keep some doubt about the distance.
• ✓Pace per kilometer: is it steady, or is there a gap the terrain explains?
• ✓Average heart rate: at a similar pace, is it higher or lower than a few weeks ago?
• ✓One line on how you felt: heavy legs, easy breathing, tiredness from the week.

Once a week, log your VO2max, HRV and load in a notebook or spreadsheet, always on the same day. To read how your heart rate drifts at the end of a long run, the method is in the article on fueling.

When to ignore your watch

A watch does not feel what you feel. If your recovery status says "good" but your legs feel like lead, trust your legs. If the score is poor but you feel fresh, also check how you slept and ate before canceling your workout.

KILYO reads your watch workouts (Garmin, Apple Watch, Coros, Polar, Suunto, Strava…) and analyzes pace, heart rate, volume and consistency. For example, it can show whether, at the same pace, your heart rate changes from one week to the next, and adjust your plan accordingly.

Written with the help of AI.