By Raji Shah ยท Updated 2026-10-24
Wrist heart-rate sensors are close to a chest strap at rest, during sleep and on many steady runs. They struggle with fast changes in effort, strength training, cold hands and irregular movement, and some people's readings are much worse than others'. This guide covers how the sensors work, what the research says about where they fail, and the fit and gear changes that fix most problems.
How a wrist sensor reads your pulse
Optical sensors use a method called photoplethysmography. LEDs shine light into your skin, and a photodiode measures how much comes back. Each heartbeat briefly raises the volume of blood under the sensor, and because blood absorbs green light strongly, the reflected light dips with every pulse. As Apple puts it, "Blood is red because it reflects red light and absorbs green light." Software then turns that pulsing signal into a heart rate, and tries to filter out everything else that changes the light: arm movement, the band shifting on your skin, and changes in blood flow.
Green light is the usual choice for a reason. In lab tests, green produced a stronger pulse signal than other colors across skin types, and held up better during movement such as hand-waving and jumping. But the pulse signal is still easily swamped by motion, and anything that reduces blood flow at the wrist, such as cold, weakens it further.
How accurate are wrist sensors?
On average, quite accurate, but averages hide a lot. A 2020 meta-analysis of 44 studies and 15 brands found wrist sensors read almost exactly the same as a chest strap or ECG at rest (an average difference of 0.01 bpm), and within about half a beat during treadmill walking and running. They underestimated heart rate by about 4.6 bpm in cycling and 7.3 bpm in resistance training, and in resistance training the gap grew by about 3 bpm for every 10 bpm rise in heart rate.
Brands differ. A systematic review of studies up to 2019 found that in controlled settings, 71% of Apple Watch comparisons were within 3% of the reference, against 51% for Fitbit and 49% for Garmin, with Fitbit tending to read lower than the other two. Most of those studies were small, and the devices are now old. A 2026 review of Apple Watch studies found an average bias of just 0.27 bpm, but individual readings could still be about 7 bpm off either way, and agreement was tighter from the Series 6 onward. A meta-analysis of Fitbit studies found an average underestimate of about 3 bpm, with 95% of readings falling between about 24 bpm too low and 18 bpm too high.
So a wrist sensor is usually close on average over a steady run, but a single reading at a given moment can be well off. That matters most when you use live heart rate to steer a workout.
When wrist sensors fail
- Fast changes in effort. Wrist readings lag behind your heart. In a study of six wearables, optical heart rate trailed an ECG by up to 43 seconds, so the start of an interval or a sudden hill shows up late, on top of the time your heart itself takes to respond. Our heart-rate zones guide covers how to pace short reps instead.
- Cadence lock. A sensor can lock onto the rhythm of your arm swing and report your step rate as your heart rate. Researchers call this signal crossover, and it shows up as a heart rate that sits close to your cadence. Apple's own 2026 accuracy study included cadence lock during running as one of its test conditions.
- Cold. Cold narrows the blood vessels near the skin. Apple says that when you exercise in the cold, "the skin perfusion in your wrist might be too low for the heart rate sensor to get a reading," and COROS says that if your hands feel cold, blood flow is reduced significantly.
- Strength training and grip. Gripping a bar flexes the tendons and muscles under the sensor. Polar recommends a chest strap when there is "pressure or movement in muscles or tendons near the sensor". In one study, an Apple Watch Series 6 agreed well with a reference in squats, deadlifts and rows but only moderately to poorly in curl-to-press and burpees.
- Irregular movement. Apple says rhythmic movement such as running or cycling gives better results than irregular movement such as tennis or boxing.
- Tattoos. Apple says the ink, pattern and saturation of some tattoos can block light from the sensor. In a 2025 study of an arm-worn optical sensor, average error over tattooed skin was 22.9% at rest, mostly because readings dropped to zero at times in 9 of 25 people, and 5.1% while running, against under 5% on nearby bare skin throughout.
- Slow walking. Several studies found wrist sensors less accurate at slow walking speeds than at running speeds.
Does skin tone affect wrist heart rate?
The evidence is mixed. Darker skin contains more melanin, which absorbs visible light, including green. A 2023 review of 10 studies found that four reported lower accuracy in darker skin, four found no effect and two had mixed results. A 2024 meta-analysis found no significant average bias by skin tone, but the spread of errors in people with dark skin was more than twice as wide as in people with light skin, and that group was only 16 people, all from one study. In one 2025 study of 25 people using a Fitbit Charge 5 during cycling, errors were similar at rest, but above 60% of heart-rate reserve they averaged 16.5 bpm in participants with dark skin, against 3.5 bpm in those with light skin. In another cycling study, skin tone didn't affect the accuracy of an Apple Watch or a Garmin, but participants with dark skin, about a third of the group, accounted for more than half of the outlier readings on every device and a disproportionate share of unexplained data gaps on the Apple Watch and an armband.
If your wrist readings are erratic during hard efforts, whatever your skin tone, a chest strap is the reliable fix.
How to get better wrist readings
The manufacturers' own advice is consistent:
- Wear it higher and tighter for workouts. Polar says to wear the watch at least a finger's width above the wrist bone and to slide it further up and tighten the band during training. Garmin says to wear it snug but comfortable, just above the wrist bone, and Apple suggests tightening the band for workouts.
- Warm up before you trust it. Garmin recommends waiting until the heart-rate icon is solid, warming up for 5 to 10 minutes and getting a heart-rate reading before you start, and warming up indoors in cold weather.
- Keep the sensor and skin clean. Garmin advises keeping sunscreen, lotion and insect repellent off the skin under the watch, and rinsing the watch after workouts.
- Avoid tattoos under the sensor if you can, as Polar advises.
- Check for cadence lock. If heart rate sits suspiciously close to your step rate, tighten the band or move the watch, or switch to a chest strap for that run.
Chest straps and arm bands
A chest strap reads the heart's electrical signal rather than blood flow, which is why it's the consumer gold standard for exercise. In a study of 50 adults, a chest strap agreed almost perfectly with an ECG, while wrist devices varied widely. On a trail run, a watch paired with a chest strap averaged 2% error, against 13% for a wrist-worn watch.
Upper-arm and forearm optical sensors usually sit between the two. On the same trail run, a forearm band averaged 6% error, though in the study of 50 adults a forearm band did worse than three of the four wrist devices. In a 2024 study over 24 hours of daily life and exercise, upper-arm sensors consistently outperformed wrist sensors, especially during arm movement and at higher intensities. In another study, a forearm optical sensor read about 1 bpm low on average against a chest strap, against about 7 bpm low for a wrist band. They're a good choice if a strap is uncomfortable.
Most running watches pair with Bluetooth or ANT+ chest straps, and Apple Watch connects to Bluetooth chest straps. WHOOP can't use an external sensor for its own data; our WHOOP chest-strap guide explains what it can and can't do, and our guide to WHOOP heart-rate errors on runs covers band-specific fixes.
Is wrist heart rate good enough?
| Use | Wrist sensor | Better option |
|---|---|---|
| Resting heart rate and sleep | Good | Not needed |
| Easy and steady runs | Usually good, once warmed up | A chest strap if readings spike early or lock to cadence |
| Intervals and hill reps | Often lags or misreads | A chest strap or arm band, and pace or effort for short reps |
| Strength training | Often underestimates | A chest strap |
| Max heart-rate tests and hard zone work | Not reliable enough | A chest strap |
Bottom line
Wrist heart rate is good enough for resting heart rate, sleep and most steady running. It is least reliable when effort changes fast, when you grip or swing your arms irregularly, in the cold, over tattoos, and for some people more than others. Fix the fit first, and use a chest strap or an arm band when the exact number matters.
FAQ
How accurate is wrist heart rate during running?
Close on average: a 2020 meta-analysis found wrist sensors within about half a beat per minute of a chest strap or ECG during treadmill walking and running. Individual readings can still be several beats off, especially at the start of hard efforts and during intervals.
Why is my watch heart rate wrong at the start of a run?
Wrist readings lag behind your heart and can lock onto your cadence until the signal settles, and cold skin weakens the signal. Garmin recommends a 5-to-10-minute warm-up, indoors in cold weather, before you rely on the reading.
Is a chest strap more accurate than a watch?
Yes, during exercise. In one study of 50 adults, a chest strap agreed almost perfectly with an ECG while wrist devices varied widely. Use one for intervals, hard zone work and any effort where the exact number matters.
Do tattoos affect heart-rate sensors?
They can. Apple says some tattoos block light from the sensor, and in a 2025 study of an arm-worn sensor, error over tattooed skin was 22.9% at rest, mostly from readings dropping to zero, and 5.1% while running, against under 5% on bare skin. Wear the sensor on untattooed skin, or use a chest strap.
Does skin tone affect wrist heart-rate accuracy?
The evidence is mixed. Some studies found larger errors in people with darker skin, especially at higher intensities, while others found no difference. If your readings are erratic during hard efforts, a chest strap is the reliable fix.
Are arm bands more accurate than wrist sensors?
Usually, during exercise. In a 2024 study, upper-arm optical sensors consistently outperformed wrist sensors, especially with arm movement and at higher intensities. They're a good middle ground if a chest strap is uncomfortable.