Watches and Rings: What They Measure and What They Invent
Introduction
A sports watch delivers dozens of numbers: calories burned, sleep stages, estimated VO₂ max, recovery score, fitness age. All presented with the same appearance of precision, decimals included.
But those numbers don't have the same backing. Some are reasonable measurements, others are acceptable estimates, and several are proprietary calculations no company publishes or independently validates. Telling them apart changes quite a lot about how you use the device.
A watch measures two or three things well. The rest it calculates, and some of it it invents with style.
What they measure reasonably well
Steps. With moderate error, and consistent enough to follow trends. Useful.
Resting heart rate. Among the best things they do. Measured on waking and watched as a weekly trend, it's a useful indicator: a sustained rise of five beats or more can point to accumulated fatigue, incipient illness, alcohol or poor sleep.
Heart rate during steady effort. Wrist optical sensors work well in running or cycling at a stable pace. Shcherbina and colleagues found low heart-rate error in most devices tested.
Total sleep duration. Approximate but usable. It reasonably detects when you slept little.
What they estimate badly
Calories burned. The least reliable number on the device and the one producing the most wrong decisions. In that same Stanford study, no device tested achieved acceptable error in energy expenditure: all exceeded 20 % error, some considerably more — and that happened even in devices that measured heart rate well.
The reason is that energy expenditure depends on lean mass, efficiency, movement type and other factors a wrist sensor doesn't know. The watch estimates them with a generic model.
The practical consequence is direct: eating back the calories your watch says you burned is the most common way to cancel a deficit without noticing.
Sleep stages. Distinguishing deep sleep from REM requires brain activity; a watch measures movement and pulse. Agreement with polysomnography — the reference method — is poor for stage classification, though reasonable for separating sleep from wake.
VO₂ max. An estimate from the relationship between pace and heart rate. Useful for seeing whether your annual trend rises or falls; not useful as an absolute value or for comparing yourself with someone else.
What they simply invent
The composite scores: recovery, readiness, daily strain, fitness age, "body load".
It isn't that they're false: it's that they're proprietary formulas. Each brand combines its variables with weightings it doesn't publish, and those formulas haven't been independently validated. You're making decisions with a black box whose criteria you can't inspect or challenge.
And they tend to be circular: if you slept badly, the watch lowers your recovery score; the original data — you slept badly — you already had, and it was more actionable.
The risk nobody talks about
There's a documented effect worth attention, especially with sleep. It's called orthosomnia: the obsessive pursuit of perfect sleep as measured by a device ends up making sleep worse. Baron and colleagues described patients whose insomnia worsened through preoccupation with their data, and who trusted the device's score more than how they felt.
The same mechanism appears with recovery: training badly because an app said you were recovered, or skipping a session you felt good for because the number came out low.
The rule that avoids this: when the device and your experience conflict, your experience wins, especially if your performance in the session backs it up.
How to use it well
1. Pick three numbers and hide the rest. Resting heart rate, steps and sleep hours cover almost everything useful. 2. Look at weekly trends, not daily readings. No single day's data should change your plan. 3. Ignore calories. If you're tracking your nutrition, do it with intake and average weight, never with estimated expenditure. 4. Ignore composite scores. Or treat them as curiosity, not instruction. 5. Cross-check with your training log. If your loads are going up, you're recovering fine, whatever the app says. 6. If it makes you anxious, take it off. A week without it is a good experiment.
Common mistakes
Eating back the watch's calories. The costliest error of all.
Making training decisions from a score. It's delegating your judgment to an algorithm you don't know.
Comparing your numbers with someone else's. Different models, different algorithms, different wrists.
Believing a number with decimals is precise. Screen resolution says nothing about accuracy.
Switching devices and comparing the history. The numbers aren't equivalent across brands.
Conclusion
A watch measures pulse, steps and roughly how long you slept well. It estimates energy expenditure and sleep stages badly, and recovery scores are closed formulas with no public validation.
Make one concrete change today: go into settings and turn off the calorie display and the recovery score. Leave resting heart rate and sleep hours visible, and look at them once a week. You'll lose noise and you won't lose any information you were genuinely using.
References
- Shcherbina, A., et al. (2017). Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure in a diverse cohort. Journal of Personalized Medicine, 7(2).
- Baron, K. G., et al. (2017). Orthosomnia: Are some patients taking the quantified self too far? Journal of Clinical Sleep Medicine, 13(2).
- Chinoy, E. D., et al. (2021). Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep, 44(5).
- Fuller, D., et al. (2020). Reliability and validity of commercially available wearable devices for measuring steps, energy expenditure, and heart rate: Systematic review. JMIR mHealth and uHealth, 8(9).
- Buchheit, M. (2014). Monitoring training status with HR measures: Do all roads lead to Rome? Frontiers in Physiology, 5.