Frequently asked questions

For heart rate, a chest strap — the Polar H10 is used as the reference standard in published validation research, which is the strongest evidentiary position a sensor can hold. For sleep staging, the Oura Ring, at 76.3% four-stage agreement against clinical polysomnography. For distance, nobody can honestly say for current models, because no independent validation of current-generation watches has been published that we could find. See the full ranking.

Because an optical sensor infers pulse from light scattered off moving blood, through a joint that moves constantly. A 2025 study running three identical sensors at wrist, forearm and upper arm simultaneously found the wrist worst at every intensity — with limits of agreement of ±10-19 bpm during a cycling warm-up and a bias of around 33 bpm during burpees. It is a placement problem, not a brand problem. Here is the fix.

Yes, and it is a difference of physical principle rather than quality tier. A chest strap measures the heart's electrical activity directly — the same signal an ECG reads. A watch infers pulse optically. At rest the two agree within 1 bpm; the gap widens sharply as intensity rises.

As far up the arm as practical. Forearm and upper-arm placements beat the wrist at every intensity tested in the 2025 placement study, using identical sensors — so the finding is about position, not product. The forearm and upper arm move less and have better perfusion than a wrist.

Moderately. Sleep-versus-wake is reliable at 91-93% agreement with clinical polysomnography. Four-stage classification ran 70.9-76.3% across the three devices ever measured in a published head-to-head, with deep-sleep sensitivity from 79.5% down to 50.5%. Treat the four-color hypnogram as an indication and weekly trends as far more meaningful than any one night. Every published figure.

Best of the three measured devices at telling asleep from awake (93%), and weakest at deep sleep (50.5% sensitivity). In the same study it overestimated light sleep by 45 minutes a night and underestimated deep sleep by 43. That tested a Series 8, and Apple has revised its algorithm since.

Nobody has published a polysomnography validation of Garmin sleep staging that we could find, for any Garmin device. That is not a criticism — it is an empty row in the record, and we would rather show you the empty row than fill it with an opinion.

If your tracks drift in cities or under tree cover, yes. Multi-band receivers hear each satellite on two frequencies and use the disagreement between them to discard signals that bounced off buildings — which is the specific cause of that drift. If you run on open roads, single-band already performs near its best and multi-band buys you very little. The full diagnosis.

The best peer-reviewed measurement found 3.2-6.1% distance error across eight watches against a trundle wheel, falling to 0.9-4.1% on a 400 m track. Errors are consistently worse in cities and forests. That study is from 2020 and tested single-band hardware, so current multi-band watches should do better — though nobody has published a comparable study to confirm it.

Two reasons, and neither is necessarily a fault. Positioning error accumulates over distance. And race courses are measured along the shortest legal line, which almost nobody runs — every time you go wide on a corner you cover real extra distance the watch correctly records.

Depends entirely on the device, and over three years the difference is usually larger than the hardware price. WHOOP is sold as a membership and stops working when you stop paying. Oura requires a membership for the full app. RingConn and Garmin charge nothing ongoing. We work the three-year arithmetic through here.

Not automatically, and on this site newer sometimes ranks lower. The Oura Ring 5 sits below the Ring 4 in our sleep ranking because the peer-reviewed result belongs to an earlier generation — a new sensor with a new algorithm is exactly what a validation exists to test. Newer usually means better hardware and less evidence.

No, and we will never claim to have. We own none of them, run no lab, and accept no review units. Devices we claim to have tested: 0. Everything here comes from published validation studies, manufacturer specification sheets, and arithmetic on live prices — all of it linked so you can check it.

Because you do not have to. The rubric is published with its weights — evidence 30%, hardware 20%, coverage 20%, endurance 15%, cost 15% — and every product page has an expandable panel showing its five sub-scores and the reasoning for each. Every accuracy claim links to its source or says outright that none exists. You can reproduce the ranking, or change the weights and get a different one. The whole method is here.

Affiliate commission, primarily through Amazon Associates. When you buy through a link here we earn a percentage at no extra cost to you. No sponsored posts, no paid placements, no free units. Currently 6 of the 26 devices we rank carry an explicit note that no validation study exists for them, which is not a sales technique. Full disclosure.

They come from Amazon's Product API and are stamped with the retrieval date — currently Jul 24, 2026. If the data is more than 48 hours old the number disappears and the button falls back to "Check price on Amazon". We show nothing rather than a stale figure, and any derived number such as cost per GPS hour vanishes with it.

Of 26 devices ranked on this site, 8 have peer-reviewed evidence behind their headline measurement and 6 have none at all that we could find. The rest sit on manufacturer specifications. That distribution is the honest state of this market, and publishing it is most of what this site is for.

For most people who already own a decent watch: moving the heart-rate sensor off the wrist. It costs about $105, works with the watch you already own, and it is the one area where the published evidence is strong and consistent. It is also, awkwardly for us, the cheapest thing we recommend.

Some devices advertise related features, but no consumer wearable is a diagnostic instrument and we have found no independent validation of those claims. Nobody here is a clinician. If you have symptoms — persistent unrefreshing sleep, witnessed breathing pauses, an irregular pulse — see a doctor rather than buying a different tracker.

Yes. Where the right answer for a buyer is a product sold direct rather than through a retailer we partner with, we name it and link it, marked as non-earning. Buyer-first beats commission — a recommendation we do not believe is worth less than the commission on it.