/

Training

What Your Apple Watch Actually Measures During a Workout

Asier G. Morato

Asier G. Morato

·

·

7

min read

Close-up of a runner's forearm mid-stride with a black sports watch on the wrist, on a sun-dappled dirt path through a pine forest (AI-generated image)

Cover image generated with AI — © Chubby Studio S.L.

KEY TAKEAWAY

During a workout an Apple Watch measures only heart rate, motion and, outdoors, GPS position; distance, pace, calories and VO2 max are calculated from those. Heart rate is accurate on average but less reliable in hard, fast-changing efforts, GPS distance is usually within a few percent, and calories are the least reliable number on the summary.

During a workout, your Apple Watch directly measures only a handful of things: your heart rate, read with light through the skin of your wrist, your movement, from its motion sensors, and your position, from satellites when you are outdoors. Everything else on the summary screen, from distance and pace to calories and VO2 max, is calculated from those signals plus your height, weight, age and sex.

That distinction is the whole story of workout accuracy. A number that comes straight off a sensor can be checked against a lab instrument, and the research on those is reassuring. A number built from a model inherits every assumption the model makes about you, and that is where the large errors live.

What does an Apple Watch actually measure during a workout?

It helps to sort the summary into three layers.

Measured: heart rate, motion and, on outdoor workouts, GPS position. These are raw signals, and each has its own known weak spots, which we cover below.

Derived: distance, pace, splits and time in each heart rate zone. These are arithmetic on measured signals. Outdoor distance is the sum of GPS points; pace is that distance divided by time. They are only as good as the signal underneath, but they add little error of their own.

Estimated: active calories, VO2 max and similar fitness estimates. These come from models that combine your heart rate and movement with assumptions about your body. Two people with identical heart rate traces can get different numbers, and the model cannot see the things that make them different. The further a number sits from a sensor, the less weight it can carry.

How accurate is Apple Watch heart rate during exercise?

Good on average, and less dependable during hard, fast-changing efforts. The largest summary so far, a 2025 meta-analysis by Choe and Kang at the University of Mississippi, pooled 56 studies that compared the Apple Watch with reference equipment. Across 148 heart rate comparisons, the average bias was a fraction of a beat per minute (0.12), and no subgroup they looked at crossed the 10% error mark that researchers use as a validity threshold.

An average near zero still hides individual readings that miss. The same analysis put the limits of agreement at roughly 11 beats per minute either side, which means a single reading can be that far off even when the long-run average is perfect. A 2017 treadmill study by Khushhal and colleagues shows where those misses cluster. Twenty-one men walked, jogged and ran at 4, 7 and 10 km/h wearing an Apple Watch on each wrist and a chest strap as the reference. Agreement was very good while walking (r = 0.97), good while jogging (0.93) and weaker while running (0.81 to 0.86), and the watch recorded fewer readings as the pace went up. Wrist heart rate is most trustworthy when the effort is steady and least trustworthy when it changes fast.

Apple's own guidance points the same way. Rhythmic movement such as running or cycling gives better readings than irregular movement such as tennis or boxing. The watch needs to sit snugly on top of the wrist, cold weather can reduce blood flow in the skin too much for a reading, and some tattoos block the sensor's light. When readings are inconsistent, Apple's own fix is to pair a Bluetooth chest strap, the kind of instrument validation studies often use as their reference.

See what each workout really cost you

FitWoody reads the workouts your Apple Watch records and analyses each one against your own cardiac load, then tracks your training load over weeks next to your sleep and recovery. Comes with a 7-day free trial.

Download the app

★★★★★

4.5 · 732 ratings on the App Store

How accurate is GPS distance and pace?

Usually within a few percent, and more often short than long. In a 2020 study from the Swiss Federal Institute of Sport, Gilgen-Ammann and colleagues tested eight sport watches, including an Apple Watch Series 4, on measured courses in a city, in a forest and on a track. Average errors ranged from 3.2% to 6.1% across the watches. Distances were generally underestimated in the city and the forest, where buildings and trees disturb the satellite signal, and slightly overestimated on the open track.

Two details from that study matter for runners. Running produced larger errors than walking or cycling for most devices, and the authors concluded that recorded distances might come up short by up to 9%. A 10 km route through tall buildings can honestly read as 9.5 km without anything being broken. Pace inherits the same error, and the live pace on your wrist is noisier than the average of a whole kilometre, so judge a session by its splits rather than by any single glance.

Why does indoor distance disagree with the treadmill?

Because indoors there is no GPS, so the watch is no longer measuring distance at all. It is estimating it from your arm and body movement and a stride length it has learned. Apple's calibration guidance is to walk or run outdoors at your normal pace for about 20 minutes in an open area with good GPS reception, and to repeat that at each speed you usually train at. The watch keeps refining your stride length on later outdoor walks and runs, which improves its estimates when GPS is limited or unavailable.

So a gap between the watch and the treadmill console is expected, not a fault in either. If indoor distance matters to you, choose the Indoor Run workout rather than Outdoor Run, calibrate outdoors first, and treat the indoor figure as a good estimate rather than a measurement.

Which workout numbers are estimates, not measurements?

Calories are the clearest case. In the same 56-study meta-analysis, every subgroup for energy expenditure exceeded the 10% error threshold, across walking, running, cycling and mixed sessions. The watch calculates calories from your heart rate and motion together with your height, weight, sex and age, so it is modelling your metabolism rather than measuring it. The calorie total is the least reliable number on your workout summary, and it is not close.

VO2 max is an estimate too, built from how your heart rate responds to pace on outdoor walks, runs and hikes. It is useful as a trend and much less useful as a single value, which we unpack in what a low cardio fitness reading actually means.

Heart rate zones sit in between. The heart rate is measured, but the zone boundaries are a choice. If they come from an age formula rather than your own physiology, time in zone 2 can be precisely measured and still mean the wrong thing, which is the case for setting your own heart rate zones.

Why do two runs at the same pace show different numbers?

Often because the heart rate is right and your body was different. Heat, a short night, a cold coming on or accumulated fatigue all raise the heart rate a given pace costs you. That is not the watch failing; it is the watch measuring the cost of the effort rather than the effort itself. It is also why the same session lands differently from one day to the next, and why we analyse workouts through your personal cardiac load instead of pace alone.

How do you get cleaner data from your watch?

Most of it comes down to a few habits. Wear the watch snugly on top of the wrist during the workout and loosen it afterwards. Pick the workout type that best matches what you are doing, as Apple recommends, and add a custom one when nothing fits. Calibrate outdoors once at your usual speeds. Keep your height, weight and age current in the Health details, since every estimate leans on them. For intervals or sports with irregular movement, use a chest strap if precise heart rate matters to you.

Then read your data the way the research suggests: trust trends across weeks more than any single workout. A bias of a few percent that repeats the same way every session cancels out when you compare this month with last month, which is what training load is built to do.

This is where FitWoody fits today. It reads the workouts your Apple Watch records, analyses each one against your own cardiac load and tells you how it actually affected your body, and it tracks your training load over weeks alongside your sleep and recovery baselines. You can also set your own heart rate zones. An Apple Watch is required, and FitWoody informs you rather than diagnosing anything.

Where to start

Open your last three workouts and sort each number on the summary into measured, derived or estimated. Lean on heart rate, time and outdoor distance. Read pace by the split, not the second. Treat calories and VO2 max as rough trends. The watch is a good instrument; it just measures fewer things than the screen implies.

Frequently asked questions

Is Apple Watch heart rate accurate enough for zone training?

For steady efforts such as easy runs, long runs and steady rides, yes: research pooling dozens of studies found its average heart rate error close to zero. It is less reliable in short, hard intervals and sports with irregular arm movement, where a paired chest strap gives cleaner readings.

Why is my Apple Watch distance different from the treadmill?

Indoors the watch has no GPS, so it estimates distance from your movement and a stride length learned on earlier outdoor walks and runs. A gap with the treadmill is expected. Choosing the Indoor Run workout and calibrating outdoors for about 20 minutes at your usual pace narrows it.

Which Apple Watch workout number is least accurate?

Active calories. A 2025 meta-analysis of 56 studies found energy expenditure error above the 10% validity threshold in every subgroup it examined, because calories are modelled from heart rate, motion and your body details rather than measured.

Sources

Keep reading