Choosing a wearable. What the shape decides, and what it does not.

Start with the question you actually have

Almost everyone buying a wearable is trying to answer one of four questions, and they lead to different devices.

*Am I sleeping badly?* is a question about the night, and the thing that matters is whether you will actually keep the device on while you sleep. *Am I training too hard or not hard enough?* is a question about load over weeks, and it needs something you wear consistently rather than something with the longest sensor list. *Is something changing in my health?* is the question a wearable answers least well, and the section at the foot of this page is about why. *How many steps did I do?* is answered adequately by almost anything, including the phone already in your pocket.

Working out which question is yours narrows the field faster than any feature comparison, because feature lists are written to be long rather than to be relevant.

What the shape decides

The form factor does not decide accuracy. It decides where the sensor sits, whether there is a screen in your life, and how likely you are to still be wearing the thing in three months — which, in practice, matters more than any specification on the box.

Smartwatch

A screen on your wrist: notifications, on-device workout data, maps and satellite positioning on the larger ones. The cost is that it is a computer you are wearing — it wants charging, updating and attention, and it lights up in a dark bedroom. Best when you want the device to be useful while you are wearing it, not only afterwards in an app.

Screen-free band

No display, no notifications, nothing to tap: you wear it and read everything later in the app. Polar describes its Loop in exactly those terms — screen-free, with silent sleep tracking and no on-device controls — and Garmin's CIRQA band reports the same Body Battery reading its watches do without a screen at all. Best when you want the measuring without the second screen.

Ring

Worn on a finger rather than a wrist, which changes where the sensor sits and what gets in its way. Oura's named daily Scores are Sleep, Readiness and Activity, so the ring is read in the app the same way a screen-free band is. Best when a wrist device is not something you will wear at night, or at all.

Basic fitness tracker

Steps, heart rate, a rough activity total. Cheap, simple, and honest about being simple. Worth taking seriously if the question you actually have is whether you moved today — a more expensive device does not answer that question any better.

What all of them are really doing

Underneath the different shapes, consumer wearables are doing something very similar: shining light into your skin and reading what comes back, then estimating everything else from that signal plus movement, and on many devices skin temperature.

This matters because of what it is not. Research on heart rate variability was built on electrical recordings of the heart, and a study published in Frontiers in Physiology in 2025 recorded both at once from the same device on 931 adults: the optical, pulse-based variability came out systematically lower than the electrical measurement on every index tested, and the authors concluded the two should not be treated as interchangeable. That study measured at the chest and upper arm rather than the wrist, so the exact figures do not transfer to a wrist device — but the point does. A wearable is producing a pulse-based estimate, not a medical-grade measurement.

None of that makes the estimate useless. It makes it an estimate, which is a different thing from a measurement, and worth remembering on the morning a number surprises you.

Accuracy, in plain terms.

The honest summary of the published validation research is that these devices are good at some things, mediocre at others, and weakest at exactly the thing they display most confidently.

A laboratory study published in the journal Sleep in 2021 put seven consumer devices next to a full sleep laboratory recording across three nights. Detecting that you were asleep worked well. Detecting that you were awake did not — correctly identifying wake ranged from poor to mediocre across the seven devices, which means they tend to over-count sleep and under-count how broken the night was. On sleep stages the same study found all the devices overestimated light sleep and, on average, failed to correctly identify between 30% and 50% of both deep sleep and REM sleep.

A 2025 meta-analysis in the American Academy of Sleep Medicine's own journal pooled 24 studies and 798 people and found the same direction of error on duration: wrist devices underestimated total sleep time by about 17 minutes on average and overestimated the time spent awake after falling asleep by about 13 minutes. Its authors also named a limit worth carrying into any purchase: manufacturers' algorithms are proprietary and undisclosed, so nobody outside the company can check how any particular number was reached.

Both of those studies tested hardware and software from earlier generations, so neither is evidence about any device on sale today, in either direction. That is itself the finding: independent validation lags the products by years, so nobody can honestly tell you how accurate the device you are about to buy is.

What it costs to keep wearing it

The purchase price is the smaller half of the question. The one that actually decides the cost is: what stops working if you stop paying?

The answers differ sharply and are not always easy to find. Polar states plainly, on its own product page and more than once, that its Loop band has no subscription and that every feature is available from day one. Other companies run a membership under the hardware, and what remains if the membership lapses is not always documented anywhere you can read before buying.

So the question is worth asking directly, before the box arrives rather than after: does the device keep measuring, does the app keep showing history, and can you export what you have already recorded? A device you own outright and a device you are renting access to are different purchases even when the sticker price looks similar.

The questions worth asking

None of these are about sensor counts, and that is deliberate. A longer sensor list has never been shown to produce a more useful morning.

• Will I wear this at night? Everything overnight — sleep, resting heart rate, overnight variability — depends on the answer, and a device that comes off at bedtime answers none of it.
• Do I want a screen? A display is the single biggest difference in how a wearable feels to live with, and it is a preference rather than a quality tier.
• How often does it need charging, and when will I do it? A device that needs charging daily gets charged overnight, which is exactly when it is not on your wrist.
• What happens if I stop paying? Ask before you buy, not after.
• Can I get my history out? Ask whether there is an official export, in a form another app could read.
• What does it claim to detect? Anything advertised as detecting, screening for or diagnosing a condition is either a regulated medical device — which will say so, with the clearances named — or it is over-claiming.
• Does it compare me to me, or to other people? A number read against your own recent weeks tells you what changed. A number read against a crowd tells you where you rank among strangers, which is a different and much less useful fact.

Where Omnisio fits, and where it does not.

This is the paragraph where we are talking about our own product, so here it is in one place rather than seeded through the page.

Omnisio is a screen-free band worn on the wrist day and night, read in an app, and it is sold as a membership: the band is included and is not sold separately, so there is no version where you buy the hardware on the side. What the app does with it is narrow on purpose — one clear answer each morning, read against your own recent weeks rather than a stranger's average, and no score at all on a night that could not be read honestly. The device itself is described in full on the Omnisio Band page, linked in the rail beside this section.

It is not the right choice for everyone reading this. If you want maps and a screen on your wrist mid-run, buy a sports watch. If you want to pay once and own the thing outright, buy something without a membership under it. And if you already wear something else, Omnisio reads the official export files from WHOOP, Oura, Garmin and Apple Health on your own phone — which is worth knowing mostly because it means switching is not the same as starting over, whichever direction you switch in.

What none of them do

No consumer wearable in any of these shapes is a medical device unless it says so and names its clearances. Omnisio is not one: neither the app nor the devices used with it hold FDA, CE-MDR or TITCK medical device clearance, and nothing in it is intended to diagnose, treat, cure or prevent any disease.

That boundary is not a formality. A 2014 review in Sports Medicine, surveying the whole toolkit used to monitor fatigue in athletes, concluded that very few of those markers have strong evidence behind them and that there is still no single definitive marker in the literature. If that is the state of the evidence in laboratories, with trained athletes and standardised conditions, a wrist device on a general population is further from a verdict about your health, not closer to one.

So use one to notice patterns over weeks, and take anything that worries you to a physician. That is the honest ceiling of the category, and a device that promises more than that is telling you something about its marketing rather than about your body.

Where this comes from

Everything factual above traces to one of these. They are the published record, not our summary of it — go and read them if a claim here matters to you.

• Source: Frontiers in Physiology, 2025. Simultaneous optical and electrical recording in 931 adults found optical pulse variability systematically lower on every index tested, and concluded the two measurements are not interchangeable. Recorded at chest and upper arm, not the wrist.
• Source: the journal Sleep, 2021. Seven consumer devices against a full sleep laboratory recording over three nights in 34 healthy young adults: high accuracy detecting sleep, poor accuracy detecting wake, light sleep overestimated by all devices, and 30% to 50% of deep and REM sleep missed on average.
• Source: the American Academy of Sleep Medicine's peer-reviewed journal, 2025. Meta-analysis of 24 studies and 798 participants: total sleep time underestimated by about 17 minutes and time awake after sleep onset overestimated by about 13 minutes, with high variation between studies and undisclosed manufacturer algorithms named as a limit on interpretation.
• Source: Sports Medicine, 2014. Review of load and fatigue monitoring tools in athletes: very few markers have strong evidence supporting their use, and no single definitive marker has been described.
• Product facts about other companies come from those companies' own published pages, read on 8 September 2026, and are used only to say which product is being described.

Questions