The evidence scoreboard
The most accurate fitness trackers, by what has actually been measured
Nine devices, three published studies, and a great deal of honest admission that the record is thinner than anyone in this category lets on.
We earn a commission when you buy through our Amazon links, at no extra cost to you. It never decides a ranking — the rubric that does is published in full, so you can check us rather than take our word for it. How this works.
The short answer
The most accurate device in consumer fitness tracking is a chest strap, not a watch: the Polar H10 is what published research uses as its reference standard. Among devices that have been measured rather than used as the yardstick, the Polar Verity Sense armband recorded a bias under 0.1 bpm against it, and the Oura Ring 76.3% four-stage sleep agreement against clinical polysomnography.
Ranked
Quick picks
| # | Device | Best for | Evidence | Score | Price |
|---|---|---|---|---|---|
| 01 |  Polar H10 The device other people's validation studies use as the reference. That is the strongest recommendation a sensor can have. | Anyone who trains by heart rate | Peer-reviewed validation | 8.9/10 | $104.90Amazon |
| 02 |  Polar Verity Sense The optical armband that measured a bias under 0.1 bpm against a chest strap on a graded treadmill test. The comfortable answer. | Strap-quality data without the strap | Peer-reviewed validation | 8.4/10 | $104.95Amazon |
| 03 |  Oura Ring 4 The only consumer sleep device with a strong independent PSG result behind it — 76.3% four-stage agreement, and the best deep-sleep sensitivity of the three tested. | Sleep staging you can actually cite | Peer-reviewed validation | 8.0/10 | $349.00Amazon |
| 04 |  Garmin Forerunner 970 The most complete running watch Garmin sells, and — counter-intuitively — the cheapest of the pair per hour of GPS you get. | Long-course athletes who want maps | Manufacturer specification | 7.7/10 | $749.99Amazon |
| 05 |  Polar Vantage M3 The pick when positioning accuracy is the thing you actually care about — Polar was the one brand under 5% error in the eight-watch validation study. | Runners who want the best-validated GPS | Peer-reviewed validation | 7.5/10 | $449.90Amazon |
| 06 |  COROS PACE 4 41 stated GPS hours in a 32-gram watch, for a third of the Forerunner 970's money. The endurance-per-dollar winner by a distance. | Long days on a budget | Independent measured test | 7.5/10 | $249.00Amazon |
| 07 |  Garmin HRM 600 The strap to buy if you are already inside Garmin's ecosystem and want running dynamics with it. | Garmin owners wanting running dynamics | Manufacturer specification | 7.4/10 | $159.00Amazon |
| 08 |  Fitbit Charge 6 Cheap, measured, and the weakest of the three devices in the PSG study. We would rather tell you that than not rank it. | Lowest-cost tracker with published data behind it | Peer-reviewed validation | 7.2/10 | $130.14Amazon |
| 09 |  Apple Watch Ultra 3 Excellent positioning hardware and the best screen here, attached to the weakest deep-sleep measurement in the published record. | iPhone owners who train and want one device | Peer-reviewed validation | 6.8/10 | $729.00Amazon |
#ad· Live prices from the Amazon Product API, as of Jul 24, 2026. Where we have no verified live price we show none. Scores come from the published rubric at How we rank — the Evidence column shows what each ranking actually rests on, including where the answer is nothing.
The uncomfortable summary
We went looking for validation studies covering the devices people actually buy. Here is what the record looks like when you lay it out honestly:
- Three consumer sleep trackers have a published head-to-head against clinical polysomnography.
- Two heart-rate devices in our roundups appear in a peer-reviewed accuracy study by name.
- Zero current-generation GPS watches have a published independent distance validation. The good study is from 2020 and tested hardware two generations back.
- Most devices on the market — including several very good ones — have no independent accuracy measurement of any kind.
Every publisher in this category will still happily tell you which tracker is "most accurate". Very few will tell you that the underlying record is this thin. That gap is the reason this site exists.
Why a $105 strap outranks a $750 watch
It looks like a stunt to put a chest strap at the top of a "most accurate fitness tracker" ranking. It is not.
Accuracy is the axis being ranked, and on that axis the H10 is the instrument researchers measure everything else against. Nothing else here holds a comparable position. It measures the heart's electrical activity directly rather than inferring pulse from scattered light, and that is a difference of physical principle rather than of quality tier.
The watches score well on hardware and coverage — a Forerunner 970 does vastly more than an H10 — and the weighted rubric reflects that. But on evidence, which carries the single largest share at 30%, the strap wins outright and the watches score around 5.5 for having detailed specifications and no measurements.
The three findings that matter most
Placement beats brand, for heart rate
The 2025 Sensors study ran three identical Whoop units on wrist, forearm and upper arm simultaneously — removing brand as a variable entirely — and found proximal placements won at every intensity. During burpees, every optical device fell below 0.50 concordance and a wrist-worn Garmin showed a bias of roughly 33 bpm. On a graded treadmill test, the Verity Sense armband recorded a bias under 0.1 bpm with concordance of 0.997.
The practical translation: move the sensor up your arm, do not buy a better watch.
Sleep staging is much weaker than the interface implies
Robbins et al. (2024), n = 35, single-night inpatient polysomnography. Sleep versus wake was solid across the board at 91-93%. Four-stage classification ran 70.9% to 76.3%. Deep-sleep sensitivity ran from 79.5% (Oura) down to 50.5% (Apple Watch Series 8), with the Apple Watch overestimating light sleep by 45 minutes a night.
A spread that wide means the same night on two devices genuinely tells two different stories. Full breakdown here.
Distance error is real, modest, and measured on obsolete hardware
Gilgen-Ammann et al. (2020) measured eight watches against a trundle wheel: 3.2-6.1% error overall, 0.9-4.1% on a track, only Polar under 5%. Every watch tested was single-frequency, so it describes an engineering track record rather than any current model — which is why Polar's evidence score here is 7, not 9.
Where the evidence simply is not
Publishing the gaps is as much the point as publishing the findings:
- No published PSG validation of Garmin sleep staging that we could find, for any Garmin device.
- No independent validation of any current COROS or Suunto device that we could find. The only COROS heart-rate test on record is of the PACE 3, it was unfavorable, and the author flags the dataset as limited.
- No validation of any current-generation optical wrist sensor from Garmin, COROS, Polar or Suunto.
- Two relevant papers we could not read — publisher blocks returned errors on 24 July 2026. Listed in the sources, marked unread, quoted nowhere.
How to read this ranking
The rubric weights evidence at 30%, hardware at 20%, metric coverage at 20%, endurance at 15% and cost per capability at 15%. That is why a single-purpose strap can beat a feature-complete watch, and also why the feature-complete watch does not fall to the bottom: it genuinely does more, and the rubric credits that.
If you weight things differently — and there are good reasons to — the weights are published and every sub-score is expandable on each pick below. Redo it with your own numbers. That is the entire idea.
In full
Every pick, and why
Polar H10
The device other people's validation studies use as the reference. That is the strongest recommendation a sensor can have.
- Sensor
- Electrical (ECG), chest strap
- Connectivity
- Bluetooth LE + ANT+
- Memory
- One session, up to 30 hours
- Battery (stated)
- Up to 400 h, CR2025 coin cell
- Water rating
- WR30
What's good
- Used as the criterion measure in published research — including the 2025 Sensors placement study on this site
- Simultaneous Bluetooth and ANT+, so it pairs to a watch and a bike computer at once
- 400 stated hours on a coin cell you replace rather than recharge
What isn't
- It is a chest strap, and a chest strap is less comfortable than a watch
- The strap wears out well before the sensor does
Skip this if
Skip this if you know yourself well enough to know you will not wear it. An unworn ECG strap measures nothing, and a wrist sensor you actually wear beats it every time.
What the evidence says
Peer-reviewed validationUsed as the reference standard in Sensors (2025), the anatomical-placement study of 28 participants. Being chosen as the criterion is an implicit validation by the researchers, though it means the study cannot measure the H10's own error.
Show how this 8.9 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceResearchers use it as the criterion measure. Nothing else in this category has a stronger evidentiary position. | 30% | 10.0 |
| HardwareElectrical ECG measurement rather than optical — a different and better physical principle for this job. | 20% | 9.5 |
| Metric coverageHeart rate and R-R intervals only; it does not pretend to do anything else. | 20% | 6.5 |
| Endurance400 stated hours on a replaceable coin cell. | 15% | 9.5 |
| Cost per capabilityInexpensive next to any watch, and it outlasts several of them. | 15% | 8.5 |
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Polar H10
Polar Verity Sense
The optical armband that measured a bias under 0.1 bpm against a chest strap on a graded treadmill test. The comfortable answer.
- Sensor
- Optical, forearm or upper-arm band
- Connectivity
- Bluetooth (dual) + ANT+
- Modes
- Live streaming, on-board recording, swim mode
What's good
- Sensors (2025) recorded the smallest bias of any device tested on the graded treadmill protocol — under 0.1 bpm, CCC = 0.997
- Lowest resting error in the same study, at roughly 1.6% MAPE
- Far more comfortable than a chest strap, and it works under a wetsuit
What isn't
- Still an extra device to charge and remember
- Like every optical sensor, it struggled in the same study's burpee protocol — nothing optical handles that well
Skip this if
Skip this if your training is mostly explosive, stop-start work. In the burpee protocol every optical device tested fell below 0.50 concordance, this one included. Buy the H10 for that.
What the evidence says
Peer-reviewed validationDirectly measured in Sensors (2025), n = 28, against a Polar H10 reference: lowest resting MAPE at ~1.6 ± 0.8%, and the smallest bias on the modified Bruce treadmill protocol at under 0.1 bpm (CCC = 0.997).
Show how this 8.4 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceDirectly measured in a peer-reviewed study, with the best result of any optical device tested. | 30% | 9.5 |
| HardwareOptical, but worn on the forearm — the placement the same study found outperforms the wrist. | 20% | 8.5 |
| Metric coverageHeart rate plus on-board recording and a swim mode. | 20% | 7.0 |
| EnduranceRechargeable, with stated endurance comfortably covering long sessions. | 15% | 8.0 |
| Cost per capabilityPriced against a good chest strap, not against a watch. | 15% | 8.5 |
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Polar Verity Sense
Oura Ring 4
The only consumer sleep device with a strong independent PSG result behind it — 76.3% four-stage agreement, and the best deep-sleep sensitivity of the three tested.
- Form
- Ring, worn on the index finger
- Sizing
- Size-specific — the ASIN is per size
- Subscription
- Membership required for full app features
What's good
- Robbins et al. (2024): highest four-stage agreement (76.3%, κ = 0.65) and best deep-sleep sensitivity (79.5%) of the three devices tested against polysomnography
- Matched PSG on seven of eight nightly summary measures, missing only sleep latency by five minutes
- A finger is a better optical site than a wrist for the same reason a forearm is
What isn't
- The membership is mandatory for the full app, so the sticker price is not the price — see our three-year arithmetic
- Rings are sized, so you need the sizing kit first and the ASIN varies by size
Skip this if
Skip this if you lift heavily. A rigid ring on a loaded barbell is uncomfortable at best, and Oura's own guidance is to take it off — a tracker you remove for training is not tracking your training.
What the evidence says
Peer-reviewed validationRobbins et al., Sensors 24(20):6532 (2024), n = 35, single-night inpatient polysomnography, testing the Oura Ring Gen3. Note the study tested Gen3, not the Ring 4. Note also that the manufacturer's press release rounded the four-stage figure to 79%; the paper says 76.3%. We use the paper.
Show how this 8.0 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceThe strongest peer-reviewed sleep result of any consumer device we found — with the honest caveats that it was one night, 35 people, and the previous generation. | 30% | 9.0 |
| HardwareFinger-mounted optical sensing, a better site than the wrist. | 20% | 8.5 |
| Metric coverageSleep, HRV, temperature and readiness; weak on exercise recording. | 20% | 8.0 |
| EnduranceMulti-day battery in normal use. | 15% | 8.0 |
| Cost per capabilityThe mandatory membership makes the true three-year cost materially higher than the sticker. | 15% | 5.5 |
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Oura Ring 4
Garmin Forerunner 970
The most complete running watch Garmin sells, and — counter-intuitively — the cheapest of the pair per hour of GPS you get.
- Case
- 47mm, titanium bezel, AMOLED
- Satellite
- Multi-band GNSS with SatIQ
- Battery (stated)
- 15 days smartwatch / 26 h GPS
- Maps
- Built-in, full color
- Extras
- ECG app, speaker and microphone
What's good
- Multi-band GNSS with SatIQ — the current top tier of consumer positioning hardware
- 26 stated GPS hours is enough for an ironman-distance day with a wide margin
- Built-in maps, which the 570 does not have at any price
What isn't
- The wrist optical sensor is the same class of hardware as the 570's — the premium buys features, not a better heart-rate signal
- 47mm only; there is no smaller case in this model
Skip this if
Skip this if you never race longer than a marathon and never navigate. Maps and titanium are the entire difference, and neither changes a single number the watch records.
What the evidence says
Manufacturer specificationNo independent validation study of the Forerunner 970's optical heart rate has been published that we could find. The specification figures are Garmin's own.
Show how this 7.7 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceSpecs are published and detailed, but no peer-reviewed or independent measured validation of this specific model exists yet. | 30% | 5.5 |
| HardwareMulti-band GNSS with SatIQ and Garmin's current-generation optical sensor — the top hardware tier available. | 20% | 9.5 |
| Metric coverageRunning dynamics, ECG, training load, maps, and a documented export path off the watch. | 20% | 9.5 |
| EnduranceGarmin states 26 hours in GPS mode, 15 days smartwatch. | 15% | 7.5 |
| Cost per capabilityHighest sticker in the range, but the strongest dollars-per-GPS-hour of the two Forerunners — see the arithmetic on the comparison page. | 15% | 7.5 |
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Garmin Forerunner 970
Polar Vantage M3
The pick when positioning accuracy is the thing you actually care about — Polar was the one brand under 5% error in the eight-watch validation study.
- Display
- AMOLED
- Satellite
- Dual-frequency GNSS
- Extras
- Turn-by-turn navigation, offline maps
What's good
- Polar is the only brand that came in under 5% distance error across all conditions in the JMIR eight-watch study
- Dual-frequency positioning plus offline maps at a mid-range price
What isn't
- The validation that flatters Polar is from 2020 and tested older models — it is evidence about the brand's GNSS engineering, not a measurement of this watch
- Smaller ecosystem than Garmin
Skip this if
Skip this if you are deep in Garmin's ecosystem already. Nothing here is enough better to justify rebuilding your training history somewhere else.
What the evidence says
Peer-reviewed validationThe peer-reviewed evidence is brand-level and dated: Gilgen-Ammann et al. (2020) measured eight watches against a trundle wheel and found only Polar's receivers stayed under 5% error. That study tested the Vantage M, Vantage V and V800 — not the M3.
Show how this 7.5 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceThe only brand in this roundup with a favorable peer-reviewed positioning result — but the study is from 2020 and tested earlier models, so it is dated brand evidence, not a measurement of this watch. | 30% | 7.0 |
| HardwareDual-frequency GNSS with offline maps. | 20% | 8.5 |
| Metric coverageStrong training-load and recovery platform; good sensor support. | 20% | 8.0 |
| EnduranceCompetitive for an AMOLED multi-sport watch, per Polar's stated figures. | 15% | 7.0 |
| Cost per capabilityMid-range pricing for dual-frequency plus maps. | 15% | 7.0 |
$499.9910% off
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Polar Vantage M3
COROS PACE 4
41 stated GPS hours in a 32-gram watch, for a third of the Forerunner 970's money. The endurance-per-dollar winner by a distance.
- Display
- 1.2" AMOLED touchscreen
- Weight
- 32 g with nylon band, 40 g with silicone
- Satellite
- All-Systems and Dual-Frequency modes
- Battery (stated)
- 19 days daily use / 41 h GPS (All-Systems)
- Thickness
- 11.8 mm
What's good
- 41 manufacturer-stated GPS hours — more than the Forerunner 970 and the 570 combined
- 32 g is light enough that you stop noticing it, which matters more over 100 miles than any feature
- Dual-frequency positioning at an entry-level price
What isn't
- COROS wrist heart rate is the one area where an independent test has published a specific problem — see the evidence note
- Smaller third-party app ecosystem than Garmin's
Skip this if
Skip this if you train by heart rate off the wrist and will not wear a strap. The one independent test we found flagged COROS wrist HR during exactly the intervals you would be buying it for.
What the evidence says
Independent measured testThe5krunner's measured comparison found the previous-generation PACE 3's wrist heart rate diverged during and after rapid intensity changes. That test was of the PACE 3, not the PACE 4, and the author cautions the dataset was limited. COROS states the PACE 4 has a redesigned sensor; no independent test of the new sensor has been published that we could find.
Show how this 7.5 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceThe only independent measured evidence in the range is a limited-dataset test of the PREVIOUS model, and it was unfavorable. The redesigned sensor is unvalidated. | 30% | 4.5 |
| HardwareDual-frequency GNSS and a redesigned optical sensor, per the COROS spec page. | 20% | 8.0 |
| Metric coverageStrong training metrics; a narrower ecosystem than Garmin for third-party integrations. | 20% | 7.5 |
| Endurance41 stated GPS hours in All-Systems mode is the best in this roundup by a wide margin. | 15% | 10.0 |
| Cost per capabilityComfortably the strongest dollars-per-GPS-hour figure on the site. | 15% | 10.0 |
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to COROS PACE 4
Garmin HRM 600
The strap to buy if you are already inside Garmin's ecosystem and want running dynamics with it.
- Sensor
- Electrical (ECG), chest strap
- Extras
- Running dynamics, HRV data, on-board storage
What's good
- Adds running dynamics that a watch alone cannot compute
- Stores data when out of watch range and syncs later
What isn't
- Costs meaningfully more than the Polar H10 without a published validation advantage
- Its extra metrics only fully surface inside Garmin's own platform
Skip this if
Skip this if you just want accurate heart rate. You would be paying a premium over the H10 for running-dynamics metrics most runners look at twice and never again.
What the evidence says
Manufacturer specificationGarmin publishes specifications but no validation study we could find. It is an electrical chest strap, which is the right physical principle — but 'right principle' is not the same as 'measured accuracy'.
Show how this 7.4 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceCorrect sensor principle, detailed published specs, no independent validation we could find. | 30% | 6.0 |
| HardwareElectrical ECG measurement with on-board storage. | 20% | 9.0 |
| Metric coverageHeart rate, HRV and the full running-dynamics set. | 20% | 8.5 |
| EnduranceGood stated endurance; rechargeable. | 15% | 8.0 |
| Cost per capabilityA clear premium over the H10 without a matching evidence advantage. | 15% | 6.0 |
$169.996% off
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Garmin HRM 600
Fitbit Charge 6
Cheap, measured, and the weakest of the three devices in the PSG study. We would rather tell you that than not rank it.
- Form
- Wrist band
- Battery (stated)
- Up to 7 days
- Extras
- Heart rate on connected gym equipment
What's good
- One of only three devices with a published head-to-head PSG result at all
- Substantially cheaper than every ring here
What isn't
- Robbins et al. (2024): lowest four-stage agreement of the three tested, at 70.9%, and deep-sleep sensitivity of 61.7%
- Overestimated light sleep by 18 minutes and underestimated deep sleep by 15
Skip this if
Skip this if the sleep stages are the point. It was measured, and it came third of three.
What the evidence says
Peer-reviewed validationRobbins et al., Sensors 24(20):6532 (2024) tested the Fitbit Sense 2 — not the Charge 6 — against polysomnography: 70.9% four-stage agreement (κ = 0.55) and 61.7% deep-sleep sensitivity, the weakest of the three devices. Fitbit's staging algorithm is shared across the line, but this remains evidence about a different device.
Show how this 7.2 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceReal peer-reviewed evidence exists for the platform — it is simply unflattering, and it tested a sibling device. | 30% | 6.0 |
| HardwareWrist-worn optical sensing, the weakest of the three placements in the placement study. | 20% | 7.0 |
| Metric coverageBroad everyday metrics; some features sit behind Fitbit Premium. | 20% | 7.5 |
| EnduranceUp to 7 stated days. | 15% | 7.5 |
| Cost per capabilityBy far the cheapest device in this roundup. | 15% | 9.0 |
$159.9519% off
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Fitbit Charge 6
Apple Watch Ultra 3
Excellent positioning hardware and the best screen here, attached to the weakest deep-sleep measurement in the published record.
- Case
- 49mm titanium
- Satellite
- Dual-frequency GNSS
- Connectivity
- GPS + Cellular
What's good
- Dual-frequency positioning and a genuinely excellent display
- Best-in-class as a general smartwatch, by a distance
What isn't
- In the Brigham and Women's PSG study, the Apple Watch's deep-sleep sensitivity was 50.5% — the weakest of the three devices tested
- Battery is short next to every dedicated sports watch here
Skip this if
Skip this if you bought it to fix your sleep data. The peer-reviewed result says that is the one thing it is worst at, and no firmware update changes a published measurement.
What the evidence says
Peer-reviewed validationRobbins et al. (2024) tested the Apple Watch Series 8 against polysomnography: 75.0% four-stage agreement, but deep-sleep sensitivity of only 50.5%, overestimating light sleep by 45 minutes and underestimating deep sleep by 43 minutes. That was the Series 8, not the Ultra 3, and Apple has revised its algorithm since.
Show how this 6.8 was calculated
| Factor | Weight | Score |
|---|---|---|
| EvidenceGenuine peer-reviewed sleep data exists, which most rivals lack — but the result was unfavorable, and it tested an earlier model. | 30% | 6.0 |
| HardwareDual-frequency GNSS and the best display in this group. | 20% | 9.0 |
| Metric coverageVery wide metric coverage; export off-platform is more awkward than Garmin's. | 20% | 8.5 |
| EnduranceMaterially shorter than every dedicated sports watch here. | 15% | 4.0 |
| Cost per capabilityFlagship pricing; the cellular model costs more again. | 15% | 6.0 |
$799.009% off
Price as of Jul 24, 2026. Prices move; Amazon's is the one that counts.
#ad · we may earn a commission from this link to Apple Watch Ultra 3
How this ranking was built
Every score on this page comes from one published rubric, applied identically to every device on the site, with the weights written down and the inputs shown. We have not worn any of these devices and we never say we have. Read the method and check us.
Questions
Frequently asked
What is the most accurate fitness tracker?
For heart rate, a chest strap — the Polar H10 is used as the reference standard in published research. For sleep, the Oura Ring, at 76.3% four-stage agreement with clinical polysomnography. For distance, nobody can say for current models, because no independent validation of current-generation watches has been published that we could find.
Are fitness trackers accurate?
It depends entirely on which number. Resting heart rate and sleep-versus-wake are good across the board. Heart rate during rapid intensity changes on a wrist is poor across all brands measured. Four-stage sleep breakdowns are weaker than the interfaces suggest, at 70.9-76.3% agreement. Distance is within a few percent, worse in cities and forests.
Why is a chest strap ranked above expensive watches?
Because the axis being ranked is accuracy, and the strap measures the heart's electrical activity directly while a watch infers pulse from light through a moving joint. The watches score higher on hardware and metric coverage, and the weighted rubric reflects that — but evidence carries the largest single share, and there the strap wins.
Has anyone independently tested Garmin or COROS accuracy?
Not for current models, as far as we could find. There is no published PSG validation of Garmin sleep staging, and the only independent COROS heart-rate test on record is of the previous-generation PACE 3. That absence is why both score mid-range on evidence rather than high.
Do you test these devices yourselves?
No. We own none of them and run no lab, and we will never claim otherwise. Devices we claim to have tested: zero. What we do instead is read the validation research, compile the published specifications, and show arithmetic you can repeat — all of it linked so you can check us.
Show your working
Sources
Every one of these was read in full before it was cited. Where we could not access a paper, we do not quote its numbers.
- [1]Peer-reviewed validation2025“Impact of Anatomical Placement on the Accuracy of Wearable Heart Rate Monitors During Rest and Various Exercise Intensities”, Sensors (Basel)
n = 28 (14 male, 14 female). Polar H10 chest strap as the reference. Compared Polar Verity Sense (forearm), Garmin Forerunner 55 (wrist) and three simultaneous Whoop 4.0 units at wrist, forearm and upper arm, across rest, cycling warm-up, burpees and a modified Bruce treadmill protocol.
- [2]Peer-reviewed validation2024Robbins et al., “Accuracy of Three Commercial Wearable Devices for Sleep Tracking in Healthy Adults”, Sensors 24(20):6532
Brigham and Women’s Hospital. n = 35 healthy adults, single-night inpatient polysomnography, testing Oura Ring Gen3, Fitbit Sense 2 and Apple Watch Series 8. Authors’ stated limits: one night, lab conditions, healthy adults without insomnia, comparison only during scheduled sleep, and device failures on 6 Apple Watches and 2 Fitbits.
- [3]Peer-reviewed validation2020Gilgen-Ammann, Schweizer & Wyss, “Accuracy of Distance Recordings in Eight Positioning-Enabled Sport Watches: Instrument Validation Study”, JMIR mHealth and uHealth
DOI 10.2196/17118. Eight watches (Apple Watch Series 4, Coros Apex 46mm, Garmin Fenix 5X Plus, Garmin Forerunner 935, Polar Vantage M, Polar Vantage V, Polar V800, Suunto 9 Baro) measured against a trundle wheel accurate to 1 cm, across urban, forest and 400 m track conditions. IMPORTANT: this predates the dual-frequency receivers now standard, so it describes brand engineering history, not any watch currently sold.
- [4]Independent measured test2023the5krunner, measured heart-rate comparison of the COROS PACE 3
An independent measured test, not a peer-reviewed study, and the author explicitly cautions the dataset was limited. Compared against Garmin HRM-Pro Plus and Polar Verity Sense across easy runs, intervals and cycling. Found divergence during and after rapid intensity change. Tests the PACE 3, not the PACE 4.
- [5]Manufacturer specification2026Polar — H10 heart rate sensor published specifications
Manufacturer-stated: electrical (ECG) measurement, Bluetooth LE + ANT+, internal memory for one session up to 30 hours, up to 400 h on a CR2025 coin cell, WR30.
- [6]Peer-reviewed validation2024Svensson et al., “Validity and reliability of the Oura Ring Generation 3 with sleep staging algorithm 2.0 compared to multi-night ambulatory polysomnography”, Sleep Medicine
NOT READ — the publisher returned HTTP 403 to our fetch on 2026-07-24. Reported to cover 96 participants and 421,045 epochs. We quote no figure from it. Readers should also note it is the validation of Oura's own staging algorithm, and that a published comment on it exists.
- [7]Peer-reviewed validation2025Khan et al., “The Oura Ring Versus Medical-Grade Sleep Studies: A Systematic Review and Meta-Analysis”, OTO Open
NOT READ — the publisher returned HTTP 403 to our fetch on 2026-07-24. A pooled analysis would be stronger evidence than any single study, so this is a real gap in our record rather than a formality.
Read next
Related
- Wrist vs chest strap
The placement study in full, at four exercise intensities.
- Sleep accuracy by device
Every published sleep-staging figure in one table.
- How we rank
The weights, the evidence hierarchy, and what we did not do.