A smart ring for the night, a watch for training, a band for recovery, plus three apps each with a different “score”. The wearables market promises that you only need to put the device on and your body will explain itself. Reality is more sober: some quantities are measured very well by sensors on the wrist and finger, others are only estimated from indirect signals, and for a few the number on the display is closer to marketing than to measurement.
This guide does not decide which brand is best. It looks at what each sensor can do according to validation studies, where a ring differs from a watch, and how to choose based on what you actually want from the data.
What every device measures the same way: optical heart rate
The core of rings and watches alike is an optical sensor that shines light into the skin and calculates heart rate from the reflection. At rest, and especially at night, it works reliably. The 2022 study by Miller and colleagues compared six devices with reference ECG during sleep, and night time heart rate came out with a small error on all of them, usually within one to two beats per minute.
The trouble starts with movement. Work by Bent and colleagues in npj Digital Medicine showed that optical sensors err most during activities with irregular arm movement, such as strength training, and that accuracy also varies with the type of activity and skin tone. For interval running or the gym, a chest strap remains more accurate than anything on the wrist. The ring has one more practical problem during sport: it presses when you grip a barbell or handlebars, and many people take it off to train.
Practical conclusion: take resting and night time heart rate seriously from any device. Take heart rate during intense training with a pinch of salt unless you have a chest strap.
Sleep: duration yes, stages only as a guide
This is where marketing and science part ways most. A lab measures sleep from brain activity, eye movements and muscle tone (polysomnography). Neither a ring nor a watch measures any of that. They estimate stages from heart rate, its variability, movement and skin temperature.
The study by Chinoy and colleagues in the journal Sleep compared seven consumer devices with polysomnography. Most devices estimated total sleep time acceptably, but stage detection, meaning light, deep and REM sleep, was markedly weaker. Devices tend to overestimate sleep and underestimate wake during the night. Oura’s own validation (Altini and Kinnunen, 2021) reports agreement with the lab on stages in about 79 percent of windows, which is a lot for a consumer sensor, but it still means that one in five five minute windows is classified differently. The authors are also employees of the manufacturer, which is worth keeping in mind.
Qualitative summary based on Chinoy 2021, Miller 2022, Fuller 2020 and Bent 2020. "High" means an error of a few percent against the reference, "low" errors commonly above 20 percent.
Practical conclusion: follow sleep duration, bedtime regularity and night time heart rate. Treat the share of deep sleep as a trend over weeks, not as a grade for a single night.
HRV: a useful trend, a dangerous table
Heart rate variability (HRV) is the small fluctuation in the intervals between heartbeats. A higher value usually goes with better recovery and a calmer nervous system. Night time measurement from a ring or watch is, according to Miller and colleagues, accurate enough to catch a change in one person.
Two things spoil HRV, though. The first is comparing yourself with others: the value depends heavily on age, genetics and the calculation method, so a “normal HRV” from a table on the internet means nothing. The second is reading single nights: alcohol, a late meal, illness or the menstrual cycle move the value by tens of percent. What makes sense is the seven day average and its direction. When it rises after two weeks without alcohol, you are seeing a real effect. When it drops after three nights of five hours’ sleep, so are you.
Steps, calories, oxygen and ECG
Steps. The systematic review by Fuller and colleagues (JMIR mHealth, 2020) reports that while walking, most devices count steps with an error of a few percent. Steps are therefore one of the most reliable quantities.
Calories. The same review reports that estimates of energy expenditure commonly differ from the reference by more than 20 percent. The device does not know what you eat and calculates burned energy from heart rate and movement using a model. Never use calories from a watch as a precise input to a diet.
Blood oxygen (SpO2). Most rings and watches measure it at night. It is useful for tracking a trend (for example suspected sleep apnoea), but single readings below 90 percent are not a diagnosis. Anyone who regularly sees dips belongs in a sleep clinic.
ECG and atrial fibrillation. Apple Watch and some other watches offer a single lead ECG and irregular rhythm notifications. In the Apple Heart Study (Perez et al., NEJM 2019) with 419,297 participants, 0.5 percent of people received a notification, and among those who then wore an ECG patch, atrial fibrillation was confirmed in 34 percent; the positive predictive value of the irregular pulse notification itself was 84 percent. Rings cannot do ECG yet. For people over fifty or with a family history, this is a reason to choose a watch.
Ring or watch: where they differ
| What you need | Smart ring (Oura, Ultrahuman, Samsung) | Watch (Apple, Garmin, Withings) | Band (Whoop) |
|---|---|---|---|
| Sleep and night heart rate | Most comfortable at night, good skin contact | Accurate, but many people take the watch off at night | Accurate, the strap is worn continuously |
| Training and GPS | No GPS, gets in the way during strength training | GPS, zones, pairing with a chest strap | No GPS, heart rate yes |
| ECG and irregular rhythm | No | Apple, Withings, some Garmin yes | No |
| Skin temperature | Yes, night trend (cycle, oncoming illness) | On newer models yes | Yes |
| Battery life | Usually 4 to 7 days | 1 day (Apple) to 2 weeks (Garmin) | Around a week |
| Subscription | Oura yes, Ultrahuman no | Basics free | Yes, does not work without it |
| Data export | Via Apple Health or Google Health Connect, Oura also an API | Apple Health, Garmin Connect, Withings API | Developer API |
Feature details apply to models on sale in September 2026 and vary between versions. Check the specific model before buying, especially ECG and subscription.
How to choose by your main goal
- I mainly want to sleep better and see recovery. A ring. It is worn 24 hours without the feeling of a display on your wrist, and its night data is the most consistent of the whole range. If a subscription bothers you, pick a brand without one.
- I train and want zones, GPS and accurate heart rate under load. A watch, ideally with a chest strap for intervals. Garmin gives the most data for free, Apple the best integration with the phone, Withings the longest battery life with a classic look.
- I am dealing with my heart, a family history, or I am over fifty. A watch with ECG and irregular rhythm notifications. It is the only feature on the whole list with a large clinical study behind it.
- I want both. A ring at night and a watch by day is a combination many people use. Expect two apps and numbers that will not match; every manufacturer calculates differently.
What to do with the data so that it changes something
The device itself improves nothing. The benefit appears when you connect the data with what you do: skip alcohol and look at resting heart rate and HRV two weeks later, adjust evening light and check your bedtime, add strength training and watch whether resting heart rate falls. And when you put blood work, weight and body composition next to it, you can see whether the change shows up beyond the watch.
That is why it makes sense to have measurements from the watch, ring, scale and lab in one place. One device shows the night. A timeline across all the data shows whether you are doing well in the long run.
FAQ
What people ask
Is a smart ring more accurate than a watch?
For sleep and night time heart rate the ring is usually slightly more accurate, because it fits snugly and the finger is well supplied with blood. During sport the ring often drops out, and a watch with GPS and a chest strap gives better data. It depends on what you want to measure.
Can I trust the sleep stages from a watch or ring?
Only as a rough guide. Devices estimate stages from heart rate, movement and temperature, not from brain activity as a lab does. The best get roughly 60 to 80 percent of windows right. Watch total sleep time, regularity and night time heart rate instead, those are more reliable.
What is HRV and why does it change every day?
Heart rate variability, the small differences between heartbeats. A higher value usually means better recovery. It swings with alcohol, sleep, illness and the menstrual cycle, so it makes sense to follow a seven day average rather than single nights.
Can a watch detect atrial fibrillation?
Apple Watch and some other watches offer a single lead ECG and irregular rhythm notifications. In the Apple Heart Study with more than 400,000 participants, the notification had a positive predictive value of about 84 percent. It is not a diagnosis, but it is a reason to get checked.
How accurate are steps and calories burned?
Most devices count steps with an error of a few percent while walking. Calories burned are the least reliable number, with reviews reporting errors commonly above 20 percent. Treat calories as a rough estimate.
Do I need a subscription?
With some brands, yes: Oura and Whoop have memberships without which data is limited or unavailable. Garmin, Apple and Withings give basic data for free, with only extras paid. Count the subscription into the cost over several years.
How do I turn the data into something useful?
Follow the trend, not single days: resting heart rate, HRV and sleep duration as a seven day average. Expect change after an intervention, such as cutting alcohol or adjusting training. And connect them with blood work and weight to see whether the change shows up elsewhere too.
Sources
What this article is based on
Peer reviewed papers and primary sources the text draws on. Links go to the original.
- Chinoy ED et al.: Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep, 2021Seven devices against lab sleep measurement: sleep duration holds up, stages are weaker.
- Miller DJ et al.: A Validation of Six Wearable Devices for Estimating Sleep, Heart Rate and Heart Rate Variability in Healthy Adults. Sensors, 2022Night time heart rate and HRV of six devices against a reference.
- Altini M, Kinnunen H: The Promise of Sleep: A Multi-Sensor Approach for Accurate Sleep Stage Detection Using the Oura Ring. Sensors, 2021Validation of the Oura sleep staging algorithm; the authors are company employees.
- Bent B et al.: Investigating sources of inaccuracy in wearable optical heart rate sensors. npj Digital Medicine, 2020Optical heart rate: errors during movement, effect of skin tone and activity type.
- Fuller D et al.: Reliability and Validity of Commercially Available Wearable Devices for Measuring Steps, Energy Expenditure, and Heart Rate. JMIR mHealth and uHealth, 2020Systematic review: steps accurate, calories not.
- Perez MV et al.: Large-Scale Assessment of a Smartwatch to Identify Atrial Fibrillation. New England Journal of Medicine, 2019Apple Heart Study, 419,297 participants, positive predictive value of the notification 84%.