Wearable interpretation · Evidence reviewed November 19, 2026
A rising watch VO₂ max estimate is a reason to investigate your progress, not proof that your training worked or that your lifespan increased. The number deserves attention. It also deserves its full name: an estimate made by a particular device, from particular activities, under particular conditions.
That sounds less exciting than a longevity leaderboard. It is considerably more useful. A number without its method can turn an ordinary software change, a different activity or a changed weight entry into a story about your health.
This page replaces our earlier, overly broad treatment of VO₂ max. We have removed unsupported universal accuracy claims, lifespan promises and a one-size-fits-all training prescription. The original publication date is retained. This is a documentation and research review, not a Vora device trial.
First ask which number you have
VO₂ max describes maximal oxygen uptake. Relative values divide oxygen uptake by body mass and are commonly expressed as mL/kg/min. An exercise laboratory can measure respiratory gases during a graded test. A watch uses other observations to estimate the value. Matching units do not make those procedures identical.
| Route | What produces the number | What to retain |
|---|---|---|
| Apple Watch Cardio Fitness | Heart and motion sensors during Outdoor Walk, Outdoor Run or Hiking workouts, together with profile information | Watch/software, activity, estimate date, profile changes and whether older estimates were updated |
| Garmin VO₂ max estimates | Supported activity and device combinations; cycling requires a compatible power meter | Exact model, activity-specific estimate and current device requirements |
| Laboratory indirect calorimetry | Respiratory-gas measurements during a specified exercise protocol | Exercise mode, equipment, protocol and whether criteria for VO₂ max rather than VO₂ peak were met |
Apple's current support page says indoor workouts do not contribute to its Cardio Fitness estimate and lists age, sex, weight, height and heart-rate-affecting medications among its inputs. It also describes an option to update earlier estimates. A changed historical graph therefore needs an explanation before it becomes evidence of changed fitness.
Garmin's requirements depend on the device and activity. Its troubleshooting documentation distinguishes overall, cycling and rowing estimates. Do not transplant a requirement from one watch or sport into every Garmin comparison. Check whether a value is new, carried forward or from a different activity before treating a flat line as stalled progress.
What one recent validation actually found
Lambe and colleagues' 2026 study compared Apple Watch Series 10 running watchOS 11 with treadmill indirect calorimetry. Forty people participated; 35 entered the analysis. Their mean age was 22.7 years and 71% were at or above the 80th fitness percentile. This was a predominantly young, fit sample, not a representative sample of every watch owner.
| Statistic | Reported value | Meaning |
|---|---|---|
| Mean difference | -6.25 | Average underestimation in this sample |
| 95% confidence interval for mean difference | -8.28 to -4.23 | Uncertainty about the mean bias |
| Bland-Altman limits of agreement | -18.23 to 5.73 | Estimated spread of paired differences under this study's method |
| Mean absolute error | 6.79 | Average magnitude of the difference, ignoring its sign |
The study was cross-sectional and could not establish reliability for tracking changes over time. The authors reported no competing interests and funding from Ireland's Health Research Board and Research Ireland. These results belong to the tested hardware, software, sample and protocol. They are not a current-model ranking or a correction factor for your watch.
Do not add 6.25 to your own result
A group average is not your personal error. If your watch says 40, adding the study's average underestimation to obtain 46.25 would produce a precise-looking number without a personal calibration. The wider agreement limits are the warning: individuals did not all differ by the average amount.
The confidence interval for the mean and the limits of agreement answer different questions. Neither is a guaranteed interval containing your true value. Nor should either be converted into a rule that a change smaller than a particular number is always meaningless. A study comparing people at one point in time does not by itself establish the smallest real change within one person.
Our synthetic worksheet makes the distinction concrete. Imagine four laboratory values of 50 and four watch estimates of 44, 56, 44 and 56. The signed errors average to zero; the absolute errors average to 6. An unbiased average can coexist with appreciable individual error. These are invented teaching values, not additional participants or a reanalysis of the published study.
The kilogram in the denominator matters
Here is a separate controlled example. Hold absolute oxygen uptake at 3,000 mL/min. At 75 kg, dividing gives 40 mL/kg/min. At 70 kg, it gives 42.86 mL/kg/min, a 7.14% increase in the relative number while the numerator stays fixed.
This is arithmetic, not a recommendation to lose weight and not a reconstruction of a watch algorithm. Real changes in body mass and training can occur together. The point is narrower: an improvement in a ratio does not establish which part changed. Preserve both units and the weight associated with the observation when comparing laboratory reports. For a watch, record profile edits without assuming the algorithm is simple division.
A useful response to a surprising change
- Check freshness. Find the observation date and qualifying workout. A displayed value can be older than the dashboard showing it.
- Keep the route consistent. Separate running, cycling and laboratory records. Mark device, firmware and algorithm changes.
- Record changed inputs. Keep notes on profile edits and activity conditions. Do not change medication to influence a score.
- Compare with independent observations. Pace or power at a comparable effort can add context. Agreement supports a more coherent story but does not turn a watch into respiratory-gas equipment.
- Match the evidence to the decision. A training diary and a clinical assessment require different certainty. Bring persistent concerns to an appropriate professional instead of using the score to dismiss symptoms.
It is reasonable to follow a consistent series. It is not reasonable to promise that every upward trend is genuine physiology. The question is whether the observation, its context and the decision belong together. For another example of methods changing the interpretation, see our Zone 2 formula comparison.
Download the methods and worked examples
CSV evidence and examples · JSON with sources, units and limitations. The compilation separates published observations, manufacturer documentation and synthetic arithmetic. No estimate has been measured by Vora for this article, and no longitudinal training benefit is claimed.
Publisher disclosure: we make Vora. Vora's proactive AI brings wearable context together with workouts, nutrition and recovery. That context is useful precisely because a single score is incomplete. Check supported integrations for your setup. The reference above is usable without Vora. Corrections can be sent to support@askvora.com; see our editorial standards.