Why Your Smartwatch VO2 Max Is Dropping (Even When Fitness Is Improving): A Validation + Calibration Protocol

A lower smartwatch VO2 max estimate does not prove that your aerobic fitness declined. The value can move because your fitness changed, because the workout conditions changed, or because the device received a different heart-rate, speed, power, or GPS signal.

Treat the estimate as one piece of training context. Check the manufacturer’s requirements, compare similar sessions, and consider heat, altitude, sleep, hydration, illness, and sensor quality before changing training zones.

What a watch VO2 max estimate can tell you

Consumer wearables estimate VO2 max rather than measuring respiratory gases. Apple, Garmin, and Firstbeat use proprietary models with different qualifying activities and input requirements.¹²³

Validation studies show why individual caution matters. A 2024 Apple Watch Series 7 study reported a laboratory mean of 45.88 mL/kg/min and a watch mean of 41.37 mL/kg/min, with 15.79% mean absolute percentage error, an RMSE of 8.85, and an ICC of 0.47.⁴ A 2025 Apple Watch study found mean underestimation of 6.07 mL/kg/min, a mean absolute error of 6.92, and 13.31% mean absolute percentage error.⁵

The INTERLIVE meta-analysis found little average bias for exercise-based wearable algorithms, but wide limits of agreement for individuals. Resting-only algorithms had larger average overestimation and even wider limits.⁶ A small average bias across a group can coexist with a large error for one person.

These studies support using the number as contextual information. They do not establish that every watch accurately tracks longitudinal changes for every athlete.

Apple Cardio Fitness requirements

Apple Watch estimates Cardio Fitness from qualifying outdoor walks, runs, and hikes.¹ Apple says the estimate uses heart and motion sensors together with profile information such as age, sex, weight, height, and medications that affect heart rate. Several qualifying workouts may be needed before an initial estimate appears.

Indoor runs, treadmill sessions, and push workouts do not qualify for Apple Cardio Fitness. A route with repeated stops, poor GPS reception, an inaccurate profile, or a loose watch may make one workout less comparable with another, even when the device still records the session.

Garmin and Firstbeat requirements

Garmin’s requirements depend on the activity and device. Current Garmin guidance generally calls for accurate profile data and sustained work above roughly 70% of maximum heart rate.² A running estimate usually comes from an outdoor GPS activity lasting at least ten minutes, using optical heart rate or a compatible chest strap. A cycling estimate generally requires a power meter and at least 20 minutes of sustained work.

Firstbeat describes its method as examining the relationship between heart rate and speed or power at multiple points, then selecting data judged representative and reliable.³ Terrain, wind, heat, cardiac drift, inaccurate maximum heart rate, GPS error, and power-meter error can therefore affect the inputs without representing a durable change in fitness.

Why estimates can fall without detraining

Heat and hydration

Heat raises cardiovascular strain at a given workload. In a study of 22 runners, participants completed 20 minutes of low-intensity walking before VO2peak testing in environments from 25°C to 45°C. VO2peak fell from 3.77 to 3.13 L/min across those conditions.⁷ That experiment shows an acute heat effect; it did not test a smartwatch algorithm.

Hydration can also change heart rate and performance, but 2% body-mass loss is not a diagnostic boundary. The 2007 ACSM position stand recommends individualized drinking plans intended to avoid excessive dehydration, generally defined there as more than about 2% body-mass loss.⁸

Altitude

Acute altitude exposure lowers oxygen availability. In one study of eight endurance athletes, VO2 max declined from 66 to 55 mL/kg/min between simulated altitudes of 300 and 2,800 meters, an average decrease of about 6.3% per 1,000 meters.⁹ The sample was small and the exposure was acute, so the result should not be treated as a universal correction formula.

Severe sleep loss

A study of eight participants found that 36 hours without sleep reduced time to exhaustion during heavy treadmill exercise by an average of 11%.¹⁰ That is evidence about acute total sleep deprivation, not a precise estimate for one short night or ordinary “sleep debt.”

Fever and illness

Illness can change heart rate, perceived effort, and training performance. An emergency-department analysis found higher heart rate with higher body temperature, but it did not study exercise or wearable VO2 max.¹¹ A retrospective smartwatch study detected physiological deviations around some COVID-19 infections, but only 32 infected participants were identified within a cohort of more than 5,000.¹²

Neither study turns a watch into a diagnostic device. New or concerning symptoms should be assessed on their own merits, not inferred from a VO2 max estimate.

Check heart-rate signal quality

Wrist optical heart rate is useful, but performance varies by device and activity. In one exercise comparison, chest-strap agreement with ECG was higher than Apple Watch agreement, leading the authors to recommend electrode-based chest monitors when accurate heart rate is imperative.¹³

Sartor et al. compared wrist optical monitoring with an ECG-based chest-strap reference in healthy participants and people with coronary artery disease. The wrist device showed high overall agreement, while differences still varied by intensity and population.¹⁴ The study does not establish a separate 0.99 concordance result for the chest strap; the strap was the reference.

A cleaner heart-rate trace improves the input available to an algorithm, but it does not prove that a particular VO2 max estimate is accurate. Confirm whether your device and activity profile support the selected heart-rate source.

Safety first

The comparison sessions below are optional examples for healthy athletes who already tolerate the chosen activity. They are not medical tests and do not require an all-out effort.

Speak with a healthcare professional before a hard exercise test if you have known cardiovascular, metabolic, or renal disease; are pregnant; take medication that affects heart rate; have unexplained exercise intolerance; or are returning after a long break or illness. Stop the session and seek appropriate medical care for chest pain or pressure, fainting, severe or unusual shortness of breath, new palpitations, confusion, or another symptom that feels unsafe.

Do not use a falling watch estimate to test yourself while febrile, acutely ill, unusually fatigued, or injured.

A practical two-session comparison

This is a coaching checklist, not a validated confidence score. Its purpose is to improve comparability, not to diagnose fitness.

Session A: establish a clean reference

1. Confirm that age, sex, height, weight, medications, maximum heart rate, and activity profile are accurate where applicable.
2. Wear the watch snugly and use the same supported heart-rate source you plan to use again.
3. Choose a familiar route with minimal interruptions. Apple users need a qualifying outdoor walk, run, or hike; Garmin users should follow their model’s activity requirements.¹²
4. Complete a normal, submaximal session that already fits your training. Do not add a maximal test solely to produce a watch estimate.
5. Record the watch estimate, heart rate at comparable pace or power, perceived exertion, weather, altitude, sleep, symptoms, and hydration context.

Session B: repeat after normal recovery

Repeat the same supported workout only after you have recovered and conditions are reasonably comparable. A fixed 48-to-96-hour window is unnecessary and may be inappropriate after a hard session or illness.

Compare:

• Whether the watch generated an estimate in both sessions
• Heart rate at similar pace, speed, or power
• Perceived exertion at the same workload
• GPS, power, and heart-rate signal quality
• Heat, altitude, sleep, hydration, and symptoms

Two similar sessions reduce some noise, but they do not validate the device against laboratory gas analysis.

Decide what to do next

Pattern	Reasonable next step
One lower estimate with poor signal or unusual context	Keep the current plan and collect a more comparable session
A small repeated change with stable performance and no symptoms	Monitor the trend without automatically changing zones
A repeated change plus worse pace or power at similar effort	Review training load, recovery, device setup, and recent environment
Persistent performance loss or concerning symptoms	Pause hard training and seek appropriate clinical guidance

This table has no validated 0-to-100 score or universal “ignore/recalibrate” threshold. The appropriate response depends on the magnitude, the supporting performance data, symptoms, and the consequence of changing the plan.

How SensAI uses VO2 max context

SensAI treats connected health and workout data as context for coaching, not as a diagnosis. HRV, resting heart rate, sleep quality, and workout summaries can inform daily readiness summaries and weekly program regeneration based on actual performance and recovery.

SensAI does not run a hidden VO2 max confidence score, automatically validate a watch against a laboratory, or silently recalibrate zones after two sessions. The conversational AI coach uses LLMs and can remember stated injuries, preferences, and constraints. Users can ask for a change through chat or mid-workout quick actions.

Common questions

Why did my Apple Watch VO2 max go down while training improved?

The estimate may have come from a less comparable qualifying workout, or your heart-rate cost may have changed temporarily with heat, altitude, severe sleep loss, hydration, or illness. Apple’s estimate is not a diagnosis, and one lower value cannot identify the cause.¹

Should I trust Garmin or Apple?

Use the device whose requirements match your activity and whose data you can collect consistently. Do not compare the absolute numbers as though the algorithms were interchangeable.¹²³

Does a chest strap improve confidence?

It can improve heart-rate measurement during some activities, especially when wrist optical data is noisy.¹³¹⁴ The effect on a VO2 max estimate depends on whether the manufacturer supports that source and how its algorithm uses the data.

Should I trust smartwatch VO2 max at all?

Use it as one contextual estimate with meaningful individual error. Pair it with comparable pace or power, perceived effort, symptoms, and longer-term training history.⁴⁵⁶

Continue with SensAI

• SensAI FAQ
• Contact SensAI
• Wearable Readiness Score Conflicts: Push, Modify, or Recover?
• Altitude Acclimation with Wearables: A 7-Day SpO2 + HRV + Resting HR Framework
• Data-Driven Deload Week: HRV, Sleep, and Training Load

Bottom line: verify that the workout qualified, compare like with like, and respond to performance and symptoms rather than one watch number.

---

References

Footnotes

1. Apple Support. “Track your cardio fitness levels.” Apple, accessed 2026-07-12. https://support.apple.com/en-us/108790 ↩ ↩² ↩³ ↩⁴ ↩⁵

2. Garmin Support. “What Is VO2 Max Estimate and How Does It Work?” Garmin, accessed 2026-07-12. https://support.garmin.com/en-US/?faq=lWqSVlq3w76z5WoihLy5f8 ↩ ↩² ↩³ ↩⁴

3. Firstbeat Technologies. “Fitness Level (VO2max).” Firstbeat, accessed 2026-07-12. https://www.firstbeat.com/en/science-and-physiology/fitness-level/ ↩ ↩² ↩³

4. Caserman P, et al. “Assessing the Accuracy of Smartwatch-Based Estimation of Maximum Oxygen Uptake Using the Apple Watch Series 7: Validation Study.” JMIR Biomedical Engineering, 2024. https://pmc.ncbi.nlm.nih.gov/articles/PMC11325102/ ↩ ↩²

5. Lambe R, et al. “Investigating the accuracy of Apple Watch VO2 max measurements: A validation study.” PLOS One, 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12080799/ ↩ ↩²

6. Molina-Garcia P, et al. “Validity of Estimating the Maximal Oxygen Consumption by Consumer Wearables: A Systematic Review with Meta-analysis and Expert Statement of the INTERLIVE Network.” Sports Medicine, 2022. https://pubmed.ncbi.nlm.nih.gov/35072942/ ↩ ↩²

7. Arngrimsson SA, et al. “Relation of heart rate to percent VO2 peak during submaximal exercise in the heat.” Journal of Applied Physiology, 2003. https://pubmed.ncbi.nlm.nih.gov/12391114/ ↩

8. American College of Sports Medicine. “Exercise and Fluid Replacement.” Medicine & Science in Sports & Exercise, 2007. https://pubmed.ncbi.nlm.nih.gov/17277604/ ↩

9. Wehrlin JP, Hallen J. “Linear decrease in VO2max and performance with increasing altitude in endurance athletes.” European Journal of Applied Physiology, 2006. https://pubmed.ncbi.nlm.nih.gov/16311764/ ↩

10. Martin BJ. “Effect of sleep deprivation on tolerance of prolonged exercise.” European Journal of Applied Physiology and Occupational Physiology, 1981. https://pubmed.ncbi.nlm.nih.gov/7199438/ ↩

11. Kirschen GW, et al. “Relationship between body temperature and heart rate in adults and children: A local and national study.” American Journal of Emergency Medicine, 2020. https://pubmed.ncbi.nlm.nih.gov/31345594/ ↩

12. Mishra T, et al. “Pre-symptomatic detection of COVID-19 from smartwatch data.” Nature Biomedical Engineering, 2020. https://pmc.ncbi.nlm.nih.gov/articles/PMC9020268/ ↩

13. Gillinov S, et al. “Variable Accuracy of Wearable Heart Rate Monitors during Aerobic Exercise.” Medicine & Science in Sports & Exercise, 2017. https://pubmed.ncbi.nlm.nih.gov/28709155/ ↩ ↩²

14. Sartor F, et al. “Wrist-worn optical and chest strap heart rate comparison in a heterogeneous sample of healthy individuals and in coronary artery disease patients.” BMC Sports Science, Medicine and Rehabilitation, 2018. https://pmc.ncbi.nlm.nih.gov/articles/PMC5984393/ ↩ ↩²