Video summary

These Are The Most Accurate Fitness Trackers (Backed By Science): Garmin, Apple, Whoop & Co.!

Main summary

Key takeaways

Product Review

Product(s) covered

This video compares accuracy of major fitness trackers / smartwatches across metrics:

  • Garmin
  • Apple Watch
  • Fitbit
  • Whoop
  • Amazfit
  • Samsung
  • Huawei
  • Coros

It also mentions other brands in passing, such as Suunto and Polar, and references Oura and Withings in some study contexts.

Key focus: measurement accuracy (heart rate, sleep, VO2 max, SpO2, steps, calories)

The reviewer uses two approaches:

  • Personal testing: thousands of training sessions, compared against a chest strap (gold standard) and lab VO2 max testing.
  • Research: cites scientific studies/meta-analyses and claims they “largely align” with the reviewer’s own results.

Heart rate (main segment)

How it’s measured / common limitations

  • Wearables estimate HR using optical sensors (LED + light sensor) to infer blood flow/light absorption.
  • Common sources of error:
    • Wrist movement (especially strength training)
    • Loose fit
    • Cold hands (reduced circulation)
    • Tattoos
    • Sensor lag during rapid HR changes (e.g., interval workouts may show spikes 5–10 seconds late)
  • Chest straps measure electrical signals directly and are near gold standard (ECG-level agreement).

Reported accuracy ranges (from cited studies)

  • Chest straps: about 99% agreement with ECG
  • Optical wrist sensors: typically ~50–80%

Best performers & comparisons

  • Apple Watch is described as a consistent “clear winner” across many years.
    • Strength training example (Apple vs chest strap):
      • Apple Watch Series 11 vs Garmin HRM Pro: curves look very similar; average HR differs by only a few bpm.
      • Apple Watch Ultra 3 and Series 10: similarly consistent.
  • Garmin (flagship models): shown as having more deviation than Apple in the reviewer’s examples.
    • Garmin Fenix 8 Pro: average HR difference nearly 6 bpm (in the example).
    • Another example shows ~7 bpm difference; reviewer suggests this pattern appears across models such as Enduro 3 and Venu 4.
  • Other manufacturers (general conclusion):
    • Beyond Apple, most are described as less accurate.
    • Suunto, Polar, Coros are characterized as performing similarly to Garmin.

Meta-analysis numbers (Wellness Pulse, 2025)

  • Apple Watch average accuracy: 86.31%
  • Fitbit: 73.56%
  • Garmin: 67.73%
  • Overall average across analyzed brands: 67.4%

Takeaway: HR measurement is only moderately accurate overall, with Apple as a high-end outlier.

Additional note

  • Huawei was not included in that specific HR meta-analysis, but the reviewer says Huawei is “pretty accurate” based on personal experience.

Overall HR verdict (from the video)

  • Apple Watch = most accurate for HR (especially during workouts and most scenarios).
  • For highly accurate HR, the reviewer recommends a chest strap.

Sleep tracking

Core caution / limitations

  • The reviewer emphasizes: “You can’t fully trust any manufacturer.”
  • Unlike heart rate, there’s no objective universal standard for sleep stages on-wrist.
  • Vendors use different methods (heart-rate-based, motion-based, or hybrid), and they often don’t clearly explain methodology.
  • Proper verification would require sleep lab testing per device, which is impractical for the reviewer.

2025 Belgian polysomnography (PSG) study: sleep stages

Devices tested:

  • Apple Watch
  • Garmin
  • Fitbit
  • Whoop
  • Withings ScanWatch

Accuracy findings (selected points)

  • No device is truly accurate overall.
  • Common error pattern: sleep stages are often misclassified as light sleep.
Deep sleep detection
  • Whoop: 69.6% (leads)
  • Apple Watch: 50.7%
  • Fitbit: just over 50%
  • Garmin: slightly below (exact value not provided)
REM sleep and awake time
  • Apple Watch best for:
    • REM
    • Awake time
  • Garmin described as quite inaccurate:
    • about 33% REM
    • about 27% awake
Total sleep duration
  • Fitbit best:
    • Fitbit Sense: deviation about 6 minutes
    • Fitbit Charge 5: deviation about 11 minutes
  • Apple Watch: deviation about 19.5 minutes
  • Whoop: 24.5 minutes
  • Garmin: 38.5 minutes (worst)
Sleep quality / efficiency (deviation)
  • Fitbit, Whoop, Apple Watch: low deviations ~2–4%
  • Garmin: ~8%
  • Withings: ~10%
  • Devices tend to overestimate sleep efficiency vs lab results.

Resting HR and HRV (sleep-related health metrics)

  • Resting heart rate accuracy (2025 Ohio study):
    • Oura Ring: 1.67% deviation
    • Whoop: 3%
    • Polar Grit X Pro: 2.71%
  • HRV deviations (same study), larger:
    • ~6–7% (Oura; varies by generation)
    • ~8% (Whoop)
    • ~10.5% (Garmin Fenix 6)
    • ~16% (Polar)
  • Reviewer caveat: small sample size (13 participants).

Sleep tracking verdict (from the video)

  • Sleep stages are not reliably accurate, but sleep efficiency/quality can be estimated reasonably for many devices.
  • Apple Watch is described as very solid.
  • Whoop also performs very well.
  • Practical advice: treat sleep tracking as a rough guideline.

VO2 max measurement

Reviewer’s lab experience (personal test)

  • In lab: VO2 max = 52
  • Wearables in reviewer’s test:
    • Garmin: 53
    • Coros: 52 (exact)
    • Suunto: 51.3
    • Apple Watch and Amazfit Balance 2: ~51
    • Samsung Galaxy Watch 8: 47.5 (bit off)

After training, the reviewer raised VO2 max to 55–60 and says Garmin/Coros/Suunto/Polar estimate accurately (in general).

Studies cited (Apple Watch and Garmin)

  • Apple Watch studies:
    • University of Dublin: average deviation 13.31% (Series 9 / Ultra 2 vs lab)
    • Darmstadt: deviation 15.79% (Series 7)
    • Reviewer note: Apple Watch may underestimate for very fit users and overestimate for less fit users.
  • Garmin studies (better accuracy):
    • University of Würzburg: Forerunner 245 deviation 7.9%
    • 2020 study: Fenix 3 deviation 8.05%
    • 2019 study: Forerunner 920XT deviation 7.3%

VO2 max verdict (from the video)

  • For training-specific metrics like VO2 max, Garmin and traditional sports brands (Coros, Suunto, Polar) are favored.
  • Classic smartwatches (Apple Watch, Samsung Galaxy Watch, Pixel Watch mentioned) are described as generally less accurate for VO2 max.

Steps (step counting)

Findings

  • Reviewer impression: most manufacturers are fairly accurate, with rarely extreme errors.
  • Meta-analysis (step accuracy):
    • Garmin: 82.58% (most accurate)
    • Apple: ~80.1%
    • Fitbit: 77%
    • Polar: ~53%
  • Possible reason for Polar’s lower score: the reviewer suggests Polar may count steps while cycling, which is controversial, but frames it as more activity tracking than pure steps (to avoid “zero steps” on cycling days).

Steps verdict

Step counting differences exist, but it’s described as less controversial than heart rate or sleep stage accuracy.

Calories

Key message: no consensus

  • The reviewer says there is absolutely no consensus between brands.
  • Example: two devices on the same wrist can report drastically different calories (e.g., 400 vs 700 for one workout; 2,200 vs 2,600 for day totals).
  • The reviewer does not rely on calorie estimates.

Meta-analysis numbers (calorie accuracy)

  • Apple Watch: ~71% (best in that analysis, still not high)
  • Jawbone: 65% (not personally tested)
  • Fitbit: 56%
  • Polar: ~50%
  • Garmin: ~48%

Bottom line: none are especially strong.

SpO2 (blood oxygen saturation)

Study highlights and real-world experience

  • 2022 study:
    • Apple Watch absolute deviation: 2.2%
    • Garmin Venu 2S deviation: 5.8%
  • Reviewer personal use:
    • Used Garmin Fenix 7 / Fenix 8 at altitude (e.g., Kilimanjaro, Everest base camp trek).
    • Observed SpO2 dropping as altitude increased.
    • Compared against morning/evening finger pulse oximeter; Fenix watches were “very accurate.”
  • Another study claim:
    • At 4,559 m, average deviation between Garmin and pulse oximeter: 7% (reviewer could not confirm personally)

SpO2 verdict (from the video)

  • SpO2 can track trends with altitude, but absolute accuracy varies by device and conditions.

Overall comparison & recommendation (wrap-up)

“Most accurate overall?”

  • No single brand wins in every category.
  • Reviewer’s summary ranking by metric:
    • Heart rate: Apple Watch best (but chest strap for true accuracy)
    • Sleep: Apple Watch solid; Whoop performs very well
    • VO2 max: Garmin and traditional sports brands (Coros/Suunto/Polar) have the edge
    • Steps: generally fairly accurate; Garmin / Apple / Fitbit look strong in cited analysis
    • Calories: inconsistent—don’t rely blindly
    • SpO2: generally usable for trends; absolute accuracy varies

User-experience advice

  • Don’t rely on metrics blindly; accuracy can vary with:
    • skin tone
    • tattoos
    • other individual factors
  • If you need highly accurate HR, use a chest strap.
  • Treat sleep tracking as rough guidance.

Unique points mentioned (deduplicated)

  1. HR accuracy depends on optical sensor limitations (wrist movement, fit, cold hands, tattoos) and lag during rapid changes.
  2. Chest straps are near gold-standard (near ECG agreement ~99%).
  3. Apple Watch consistently performs best for HR in the reviewer’s long-term testing.
  4. Garmin HR accuracy is generally lower than Apple in the reviewer’s comparisons (example deviations ~6–7 bpm).
  5. 2025 meta-analysis: Apple Watch HR accuracy 86.31%, Fitbit 73.56%, Garmin 67.73%, overall avg 67.4%.
  6. Huawei may be accurate for HR based on personal experience (not included in the HR meta-analysis).
  7. Sleep staging is inherently difficult to measure accurately on-wrist; no device is “truly accurate.”
  8. 2025 Belgian PSG study: deep sleep led by Whoop (69.6%); Apple 50.7%; Garmin lower.
  9. In that study, Apple Watch best for REM and awake time; Garmin notably poorer (~33% REM, ~27% awake).
  10. Total sleep duration: Fitbit most accurate (6–11 min deviations); Garmin worst (38.5 min).
  11. Sleep quality/efficiency deviations: Apple/Fitbit/Whoop low (~2–4%), Garmin higher (~8%), Withings ~10%; devices tend to overestimate efficiency.
  12. Resting HR accuracy (Ohio study): Oura 1.67%, Whoop 3%, Polar 2.71%.
  13. HRV deviations are larger (~6–16%), with small sample size (13 participants).
  14. VO2 max: reviewer’s lab test (lab 52) matched best with Garmin (~53), Coros (52), Suunto (51.3), Apple/Amazfit (~51), Samsung (~47.5).
  15. Claim: Garmin/traditional sports brands estimate VO2 max more accurately than smartwatches.
  16. Apple Watch VO2 max studies show larger deviations (~13–16%).
  17. Garmin VO2 max studies show smaller deviations (~7–8%).
  18. Steps: generally accurate; meta-analysis favors Garmin (82.58%), Apple (~80.1%), Fitbit (77%); Polar lower (~53%) partly due to cycling behavior.
  19. Calories: no consensus and large discrepancies; none are very accurate (Apple ~71% in one analysis; others lower).
  20. SpO2: study deviations Apple ~2.2%, Garmin Venu 2S ~5.8%; reviewer reports altitude trends align well with finger oximeter.
  21. Claimed SpO2 altitude study deviation at 4,559 m around 7% (reviewer couldn’t confirm).
  22. Final recommendation: choose based on your priority metric; don’t blindly trust fitness data; use chest strap for HR; treat sleep tracking as rough guidance.

Speaker / perspective breakdown

  • Primary narrator/reviewer: conducts personal long-term testing; cites and interprets studies; provides final comparisons and recommendations.
  • Music / non-speech segments: no additional opinions.
  • No other speaker voices in the subtitles beyond the narrator.

Original video