Is your Apple Watch lying about your heart rate?

You finish the last hill repeat, bent over, and look at your wrist. It says 171. Was that right? Your training partner swears by a chest strap, and forums call a watch a toy for serious training. Is your Apple Watch lying about your heart rate?

Here is what 27 studies found, the good and the bad. MONS asks you to wear an Apple Watch, so we had a reason to hope for good news. Not all of it is.

01

A chest strap and a watch measure different things.

A chest strap picks up the electrical signal that makes your heart beat. It is the same signal a hospital ECG reads, and it is very hard to fool.

The watch does something harder. It shines a green light into the skin of your wrist and measures how much comes back. Each heartbeat pushes a pulse of blood under the sensor, and the watch counts the pulses. Anything that shakes your wrist or squeezes the blood out of it, like a hard grip, makes those pulses harder to count.

02

On a steady effort, the watch is close.

Most studies put people in a lab, fit a chest strap or an ECG, and compare.

In a Stanford study, 60 people walked, ran and cycled. The Apple Watch was off by a median of 2 percent, the best of the seven wrist devices tested. At the Cleveland Clinic, 50 adults used a treadmill, a bike and a cross-trainer. The chest strap matched the ECG almost perfectly. The Apple Watch came next, ahead of Garmin, TomTom and Fitbit. In a 2026 study of 62 men with a Series 7, most readings fell within about 7 beats of the ECG.

Most head-to-head studies since 2016 found the Apple Watch the most accurate wrist device they tested. Not all of them. In one from 2026, a Fitbit and a Google Pixel Watch both beat the Apple Watch SE.

Read these results with care. Most of the studies were small, done in a lab, with young, healthy and mostly light-skinned people, and many used older watches. Apple's own large study from 2026 has not been checked by outside scientists. Apple ran part of it, and it left out people with tattoos where the watch sits.

03

One reading can be far off, even when the average is right.

Over a whole session, the watch's average is usually close. A single number on the screen is less certain. In a 2017 study on the first Apple Watch, single readings while running could land 20 beats or more above or below the strap. At 150 beats a minute, the screen might show 130 or 170 for a moment. Newer watches did much better, within about 7 beats in the 2026 study.

On the first Apple Watch, one reading could be 20 beats out. The average over a session was close.

04

Fast changes catch the watch out.

In the same 2017 study, the watch read 11 to 12 beats too high in the minutes after people stopped running. Stop-start walking made the error bigger in the 2026 Pixel study too. Some of the studies with good results left out the moments when effort changed, on purpose, and nobody has measured how many seconds the watch falls behind.

Two delays get mixed up here. Your heart itself takes a minute or two to climb when you start a hard effort. That is your heart, not the watch. On top of that, the watch can stay high after you stop. It may also rise late when you speed up, but no study has measured that yet.

05

Gripping, swimming and skin make it harder.

A tight grip. Across all wrist brands, a review of 44 studies found the biggest errors in strength training. For the Apple Watch, one study found it agreed well with a strap in squats, deadlifts and rows, but only moderately or poorly in burpees and a curl to press. Two others found it accurate in strength work. Nobody has tested it during kettlebell swings, cleans or snatches, where a hard grip meets a fast-moving wrist.

Water. We found no study of the Apple Watch during real swimming. Other wrist watches did badly there: one Garmin was off by more than half over 1,000 metres in a 2021 field study. Your heart rate also runs lower in water at the same effort, so ranges from land do not carry over. In the pool, go by how long you could hold the effort.

Darker skin and tattoos. The results on darker skin are mixed. One study found more missing readings, another found the watch read about 8 beats low in a group of six people, and several found no difference. Apple says tattoos can block the sensor's light, and that cold can stop a reading.

06

For base training, the watch is enough.

The base range, the effort of a base run or base ride, is about 25 to 30 beats wide, much wider than the watch's average error. Hold a steady effort for twenty minutes and the watch should tell you which range you were in. Threshold is narrower, about 13 beats, so give the reading a few minutes to settle before you trust it.

For short, hard efforts like VO2max intervals and hill sprints, the number is least reliable, because the effort changes faster than the watch can follow. Read the trend over a few minutes, not the number at one second. Feel the effort while you do it, and look at the heart rate afterwards. If you want closer numbers during intervals, the Apple Watch can pair with a Bluetooth chest strap, and a 2025 study found the pair matched better at high effort and in recovery. Wear the watch snug, above the wrist bone, as Apple advises.

07

Heart rate is a guide, not a measuring stick.

This is why heart rate is not the only way to read your effort, and not the one in charge. You have three more, and you carry them everywhere, with or without a watch:

Heart rate checks what those three tell you, the way filming yourself checks a kettlebell swing. When the watch and your feel agree, you have found the effort. When they disagree in the middle of a hard interval, trust the feel, and look at the watch again once the effort has settled.

That only works if the ranges are set from your own maximum, not from a formula. Finding it with a hill test, and learning what each range feels like, is in Does your heart rate agree with you?

Sources

Everything you need to go anywhere, anytime. By email.

Email me the MONS method in a few short emails, then each new chapter of the book. I can leave with one click. Privacy