Part 1 of 3 in Paceline's Wearable Accuracy Series. Part 2: How accurate is wearable sleep tracking? · Part 3: VO2max, calories, steps and SpO2

Most modern wearables get resting heart rate close to an ECG. The trouble starts when you exercise hard, and especially when you compare heart rate variability (HRV) between brands. Independent studies show that an Apple Watch, a Garmin, an Oura Ring and a Whoop strap can each look "accurate" in isolation while still reporting different numbers for the same person on the same night. Partly that's sensor error. Partly it's because the brands aren't measuring the same thing.

Why does the same heartbeat read differently on different devices?

Nearly every wrist or finger wearable estimates heart rate with photoplethysmography (PPG): a light shines into the skin and a sensor measures how much comes back as blood volume pulses. A chest strap works differently. It picks up the heart's electrical signal through electrodes, the same basic principle as a clinical ECG.

PPG has three well-documented weak spots, as Duke University researchers summarized in npj Digital Medicine: skin characteristics, motion artifacts, and "signal crossover", where the rhythm of your arm swing or footfall is mistaken for your pulse. Their study found that absolute error during activity was on average 30% higher than at rest. They also found significant differences between devices in how quickly they responded to changes in activity (Bent et al., 2020).

On skin tone the evidence is mixed. Bent et al. found no significant difference, while a Stanford study of seven wrist devices found higher error with darker skin tone, higher BMI, males, and walking (Shcherbina et al., 2017).

Then there's software. Every brand runs its own proprietary algorithms to clean the signal and decide which minutes count. The European INTERLIVE research network notes that differences in how devices are validated "hamper the comparability" of accuracy claims (Mühlen et al., 2021).

How accurate is resting heart rate on Apple Watch, Garmin, Oura and Whoop?

Resting heart rate is where wearables do best:

• A systematic review and meta-analysis of 44 studies (738 effect sizes, 15 brands) found wrist PPG was within −0.01 bpm at rest and −0.40 bpm during sleep, on average, of ECG or a chest strap (Zhang et al., 2020).
• A 2026 living meta-analysis of Apple Watch found resting heart-rate MAPE between 1.69% and 7.2% across studies (Lambe et al., 2026, npj Digital Medicine).
• In a 2025 study, 13 adults wore several devices at once for 536 nights alongside an ECG reference. Nocturnal resting HR error (MAPE) was 1.67% for Oura Gen 3, 1.94% for Oura Gen 4, 2.71% for Polar Grit X Pro and 3.00% for Whoop 4.0 (Dial et al., 2025).

That last study has a telling detail. Garmin was left out of the resting HR analysis because Garmin defines resting heart rate as "the lowest 30 min average in a 24 h period", while the other devices report a sleep-based value. Even when every device is accurate, "resting heart rate" doesn't mean the same thing on every device.

Results also vary by model. In a 2022 sleep-lab study of 53 adults, agreement with ECG for heart rate around sleep (intraclass correlation, where 1.0 is perfect) was 0.99 for Whoop 3.0, 0.96 for Apple Watch Series 6, 0.93 for Polar Vantage V, 0.85 for Oura Gen 2 and 0.41 for the Garmin Forerunner 245 (Miller et al., 2022). The authors note that the Garmin value came from a 3-minute window taken before lights-out, which again shows how much method matters. (Their lab receives research support from Whoop, and the Whoop data were supplied by the manufacturer. The paper discloses both.)

Is wrist heart rate accurate during exercise and HIIT?

Wrist accuracy is fine for steady running and drops off from there.

A Cleveland Clinic study put 50 adults through treadmill, bike and elliptical protocols wearing an ECG, a Polar H7 chest strap and two wrist devices. Across all conditions, agreement with ECG was (Gillinov et al., 2017):

• Polar H7 chest strap: 0.996
• Apple Watch: 0.92
• TomTom Spark: 0.83
• Garmin Forerunner 235: 0.81
• Scosche Rhythm+ (forearm): 0.75
• Fitbit Blaze: 0.67

On the elliptical with moving arm handles, no device reached acceptable agreement. The authors' conclusion: use an electrode chest strap "when accurate HR measurement is imperative."

The 44-study meta-analysis found the same pattern. Treadmill error was small (−0.51 bpm), but wrist devices underestimated heart rate by 7.26 bpm during resistance training and 4.55 bpm during cycling. During resistance training, the gap grew by about 3 bpm for every 10 bpm increase in heart rate (Zhang et al., 2020). So the harder you work, the more the wrist tends to under-read.

For Apple Watch, two recent meta-analyses report average bias near zero but individual limits of agreement as wide as −11.06 to +10.81 bpm (Choe & Kang, 2025; Lambe et al., 2026). A small average error can still mean a big miss on a single workout.

Chest straps hold up at high intensity. A Polar H10 kept 99.4% RR-interval signal quality during high-intensity activities, compared with 89.8% for a clinical Holter monitor (Gilgen-Ammann et al., 2019).

Oura vs Whoop vs Garmin HRV: why don't the numbers match?

HRV is where cross-brand comparison gets hardest, for three separate reasons.

1. Different formulas

Most sleep-focused wearables report RMSSD, a measure of beat-to-beat variation. Apple's HealthKit stores HRV as SDNN, the standard deviation of heartbeat intervals (Apple developer documentation). They're related, but they're different statistics with different typical values.

2. Different time windows

HRV changes through the night, so when you measure it changes the number. The 2025 validation study documented how each device samples (Dial et al., 2025):

• Oura (Gen 3 and Gen 4): 5-minute samples averaged across the entire night
• Garmin Fenix 6: 5-minute windows averaged over the detected sleep period
• Polar Grit X Pro: only the first 4 hours after sleep onset
• Whoop 4.0: a "dynamic average during sleep… weighted towards your last slow wave sleep stage"

A review of HRV norms warns that 24-hour, short-term (~5 min) and ultra-short-term values "are not interchangeable" (Shaffer & Ginsberg, 2017). The same logic applies across brands: a 4-hour average and an all-night average aren't the same number.

3. Different accuracy

Even with the ECG reference matched to each device's window, accuracy varied. Overnight HRV error (MAPE) was 5.96% for Oura Gen 4, 7.15% for Oura Gen 3, 8.17% for Whoop 4.0, 10.52% for Garmin Fenix 6 and 16.32% for Polar Grit X Pro (Dial et al., 2025). The 2022 sleep-lab study found HRV agreement (ICC) of 0.99 for Whoop 3.0 and between 0.24 and 0.69 for the other devices, and all but Whoop tended to overestimate low HRV and underestimate high HRV (Miller et al., 2022).

Also notice that Oura's HRV agreement was modest with a Gen 2 ring in 2022 and excellent with Gen 3/Gen 4 rings in 2025. Hardware and firmware change results, which is why a single study result can't settle "which brand is best" for long.

Heart rate and HRV accuracy: what the studies found

Study (year)	Devices tested	Setting	Key finding
Zhang et al. (2020), meta-analysis	15 brands, 44 studies	Lab and free-living	Mean difference −0.01 bpm at rest; −7.26 bpm resistance training; −4.55 bpm cycling
Gillinov et al. (2017)	Polar H7, Apple Watch, Garmin FR235, Fitbit Blaze, TomTom Spark, Scosche	50 adults, treadmill/bike/elliptical	Chest strap rc 0.996; wrist 0.67–0.92; none accurate on elliptical with arms
Lambe et al. (2026), meta-analysis	Apple Watch	82 studies	HR bias −0.27 bpm; LoA −7.19 to +6.64 bpm
Miller et al. (2022)	Apple S6, Garmin FR245, Polar Vantage V, Oura Gen 2, Whoop 3.0	53 adults, sleep lab	HR ICC 0.41–0.99; HRV ICC 0.24–0.99
Dial et al. (2025)	Oura Gen 3/4, Whoop 4.0, Garmin Fenix 6, Polar Grit X Pro	13 adults, 536 nights	RHR MAPE 1.67–3.00%; HRV MAPE 5.96–16.32%
Stone et al. (2021)	7 devices/apps incl. Oura	5 adults, 148 trials	Oura rMSSD MAPE 6.84%; camera app 112.36%
Gilgen-Ammann et al. (2019)	Polar H10 vs Holter	10 adults, rest to high intensity	99.4% RR signal quality at high intensity

The validation gap: most devices you can buy haven't been independently tested

The research trails what's on people's wrists. Many of the most-cited exercise studies tested 2015–2016 devices. A 2020 review of 158 publications (research through May 2019) found HR accuracy "varies by the manufacturer and device type" (Fuller et al., 2020).

A 2024 umbrella review in Sports Medicine estimated that only about 11% of consumer wearables released to date have been validated for even one biometric, and that the studies that exist represent 3.5% of what a full evaluation would need (Doherty et al., 2024). Independent data on Coros, Ultrahuman and recent Samsung Galaxy Watch models is especially thin for heart rate and HRV. That doesn't mean they're inaccurate. It means nobody independent has checked yet.

What this means: brands' heart numbers aren't directly comparable

Put it together and you get three layers of difference between any two brands:

1. Sensor accuracy, which varies by device, generation, activity and person.
2. Definitions: what counts as "resting" HR, which HRV formula, which slice of the night.
3. Proprietary processing: filters and baselines nobody outside the company can inspect.

So if your Whoop says 62 ms and your partner's Garmin says 48 ms, that tells you almost nothing about who is fitter or better recovered. Any score or leaderboard that ranks people across different devices needs to normalize those readings first, or it's comparing measurements that were never the same.

That's the problem Paceline is built around. People should be able to use whatever device they already own and still get fair, consistent tracking over time.

Habits for more reliable heart rate and HRV data

These are the practices the research supports:

1. Compare yourself to yourself, on one device. Your own trend over weeks means far more than any single reading.
2. Don't compare your Oura HRV to a friend's Whoop HRV. Different formulas and time windows make it apples to oranges.
3. Wear it the same way every day. Snug (not tight), in the same position, on the same wrist or finger. For wrist devices, a finger-width above the wrist bone during workouts helps limit movement.
4. Use a chest strap for intervals, lifting and cycling. That's where wrist PPG under-reads most. Pair the strap to your watch so the workout records strap data.
5. Take HRV at a consistent time. Overnight averages from the same device are the easiest to keep consistent. If you do morning spot checks, do them at the same time, in the same position, before coffee.
6. Judge by weekly averages, not single mornings. Nightly HRV is naturally noisy.
7. When you upgrade devices, expect a step change. Run old and new side by side for a week or two if you can, and reset your mental baseline.

Where gear can help: chest straps and training watches

The single biggest accuracy upgrade for hard workouts is an electrode chest strap. The studies above tested Polar H7 and H10 straps. The Garmin straps in the Paceline Marketplace use the same electrode-based approach, though they weren't the specific models in those studies:

• Garmin HRM 600: a chest strap for intervals, lifting and cycling, where wrist readings drift most.
• Garmin HRM-Fit: a clip-on strap designed to attach to a sports bra.
• Garmin Forerunner 570: a running watch that pairs with a chest strap, so you can use wrist HR day to day and strap HR when it counts.

These are third-party partner brands in the Paceline Marketplace. None of them makes cross-brand numbers comparable. They make your own numbers steadier.

Track the trend, whatever you wear

The most useful thing you can do with heart data is track it consistently over time. The Paceline app connects the wearable you already use and focuses on steady week-to-week habits rather than one-off readings, so a watch, ring or strap all count.

Download the Paceline app

New to the vocabulary? Our heart rate glossary explains the terms.

Next in the series: Part 2: How accurate is wearable sleep tracking, and what's inside readiness and recovery scores?

FAQ

Which is more accurate for heart rate, Apple Watch or Garmin?

In the most-cited head-to-head exercise study, the Apple Watch agreed with ECG slightly better (0.92) than the Garmin Forerunner 235 (0.81), but those were 2015–2016 models (Gillinov et al., 2017). Independent head-to-head data on current models is limited. Both are generally accurate at rest and less accurate during intense or irregular movement.

Is Oura or Whoop more accurate for HRV?

In a 2025 study of 536 nights, Oura Gen 4 (MAPE 5.96%) and Gen 3 (7.15%) were slightly more accurate than Whoop 4.0 (8.17%) for overnight HRV (Dial et al., 2025). The two also calculate HRV over different parts of the night, so their numbers aren't interchangeable even when both are accurate.

Why is my HRV different on two devices worn at the same time?

Different formulas (RMSSD or SDNN), different time windows and different processing. Some of the gap is sensor error, and much of it is definitional.

Sources

1. Zhang Y, et al. Validity of wrist-worn photoplethysmography devices to measure heart rate: a systematic review and meta-analysis. J Sports Sci. 2020;38(17):2021-2034. PubMed 32552580
2. Gillinov S, et al. Variable accuracy of wearable heart rate monitors during aerobic exercise. Med Sci Sports Exerc. 2017;49(8):1697-1703. PubMed 28709155
3. Dial MB, et al. Validation of nocturnal resting heart rate and heart rate variability in consumer wearables. Physiol Rep. 2025;13(16):e70527. PubMed 40834291
4. Miller DJ, Sargent C, Roach GD. A validation of six wearable devices for estimating sleep, heart rate and heart rate variability in healthy adults. Sensors. 2022;22(16):6317. PubMed 36016077
5. Lambe R, et al. The accuracy of Apple Watch measurements: a living systematic review and meta-analysis. npj Digit Med. 2026;9:63. PubMed 41513748
6. Choe JP, Kang M. Apple Watch accuracy in monitoring health metrics: a systematic review and meta-analysis. Physiol Meas. 2025;46(4). PubMed 40199339
7. Bent B, et al. Investigating sources of inaccuracy in wearable optical heart rate sensors. npj Digit Med. 2020;3:18. PubMed 32047863
8. Shcherbina A, et al. Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure in a diverse cohort. J Pers Med. 2017;7(2):3. PubMed 28538708
9. Mühlen JM, et al. Recommendations for determining the validity of consumer wearable heart rate devices: expert statement and checklist of the INTERLIVE Network. Br J Sports Med. 2021;55(14):767-779. PubMed 33397674
10. Stone JD, et al. Assessing the accuracy of popular commercial technologies that measure resting heart rate and heart rate variability. Front Sports Act Living. 2021;3:585870. PubMed 33733234
11. Gilgen-Ammann R, et al. RR interval signal quality of a heart rate monitor and an ECG Holter at rest and during exercise. Eur J Appl Physiol. 2019;119(7):1525-1532. PubMed 31004219
12. Shaffer F, Ginsberg JP. An overview of heart rate variability metrics and norms. Front Public Health. 2017;5:258. PubMed 29034226
13. Fuller D, et al. Reliability and validity of commercially available wearable devices for measuring steps, energy expenditure, and heart rate: systematic review. JMIR Mhealth Uhealth. 2020;8(9):e18694. PubMed 32897239
14. Doherty C, et al. Keeping pace with wearables: a living umbrella review of systematic reviews evaluating the accuracy of consumer wearable technologies in health measurement. Sports Med. 2024;54(11):2907-2926. PubMed 39080098
15. Apple Inc. HealthKit: heartRateVariabilitySDNN (developer documentation). developer.apple.com

This article is for educational purposes only and isn't medical advice. Consumer wearables aren't diagnostic devices. If you have a heart condition or unusual symptoms, talk to a qualified clinician.