How we score smart rings

The short version: we read the fine print so the ring's real cost, real evidence and real limits are on the page before you buy. We don't run a lab, and we don't pretend to. Where a fact isn't published, we write “the brand doesn't publish this” instead of guessing.

Researched vs Hands-on: the label on every pick

Iron Mako Media sites follow one standard: “Tested when we can. Researched when we can't. Labeled either way.” Every pick carries one of two labels, and the label links back here.

• Researched means the pick was chosen and scored from the brand's own fee policy, spec sheet, sizing and warranty pages, and from published studies that compared the ring (or its predecessor) with an ECG, a chest strap or a sleep lab. It is not a claim that anyone here has worn the ring.
• Hands-on would mean documented use of that specific product: what was done, for how long, and what we saw, stated on the page.

Right now the answer is simple. We haven't lab-tested these rings; every pick is labeled Researched. When a page describes how a ring behaved on someone's finger, it names the reviewer who reported it (“DC Rainmaker reported…”) rather than presenting it as our own finding.

The scoring rubric

Each product gets a 0–10 sub-score on every line of its category's rubric. The overall score is the weighted mean of those sub-scores, rounded to one decimal. The site's code computes it from the sub-scores, so an overall can never disagree with the numbers printed beside it. Products are only scored against their own category: a chest strap's 9 and a ring's 9 are different judgments.

Hardware price is deliberately not scored. It changes daily and we only show it live from Amazon. The fee a brand makes you pay to use the ring is scored, heavily, because it's published policy and it's the cost this site's readers care about most.

Smart rings

Chest-strap heart rate monitors

Sizing kits, ring covers and charging cases

A worked example

Here are two rings' sub-scores as they appear on the site today, straight from the product records. The weights do the rest.

The first line explains most of the gap. A required membership costs Oura most of its cost-of-ownership points, and that line carries 30% of the weight. Oura wins the evidence line, because its rings have the most published, independent validation of any brand we cover. That is the trade this site exists to show: better-studied ring, recurring fee; or no fee, thinner evidence. The full ranking is in the best smart rings without a subscription.

Scores are visible on the page only. We don't put them into search-engine structured data as review ratings, because they are editorial judgments from documented research, not test results.

How fees are computed

Every subscription figure on the site follows the same four rules.

1. The brand's own US policy page, with a date. Fees come from the brand's support or pricing page, never from a reseller or a review, and every one carries the date we checked it. The current fees were last checked on Sep 19, 2026.
2. The cheapest billing option offered. Where a brand sells an annual plan, multi-year totals use it, because it's the cheaper route. The month-to-month total is shown alongside where it matters.
3. Multiplied by the years you'll wear it. We show one, three and five years. The arithmetic is done by the site's code from the published rate, so no total is ever typed by hand.
4. No discounts, no tax. We ignore any free month included with the ring, promotional pricing, legacy or grandfathered plans, and sales tax. That can overstate the first year slightly. We would rather overstate a fee than hide one.

Optional extras are listed, not added. Ultrahuman sells optional PowerPlugs and Circular plans paid “Circular Coins” for some future features. Neither is required to see your core data, so neither is counted as a required fee. We name them on the ring's page so you can decide.

Hardware prices only come live from Amazon. We never print a ring's price from memory or from a launch announcement, because it goes stale within days. Until a live price feed is connected, buttons read “Check price on Amazon.” When prices do appear, each will carry an “as of” date, and if the feed hasn't been refreshed within 48 hours the number disappears and the button goes back to “Check price on Amazon” rather than show a stale figure.

Every brand's fee policy, side by side, is in smart ring subscription costs, and the Oura arithmetic in full is in what the Oura membership costs.

How we weigh accuracy evidence

The “evidence of accuracy” sub-score asks one question: has this ring, or its direct predecessor, been compared with a reference device (an ECG, a chest strap or a sleep lab's polysomnography), and how much can that comparison be trusted? We rank evidence in this order:

1. Peer-reviewed and independent. Published in a journal, measured against a reference device, with no brand funding and no declared conflicts of interest.
2. Peer-reviewed but brand-funded or brand-affiliated. Still useful, and labeled. For example, Robbins et al. (Sensors, 2024) states it was funded by Oura.
3. Preprints and in-house studies. Counted for little, and always labeled. The only study we found that compares a RingConn ring with a chest strap is a 2026 preprint (Tao et al.) that has not been peer-reviewed, and Ultrahuman's study is an in-house pilot on the Ring AIR.

A maker's marketing figure, such as RingConn's “90.7% Sleep Apnea Monitoring Accuracy” for the Gen 2, is not evidence on this scale. We quote it as the maker's claim. We also weigh how close the studied model is to the one you'd buy: a study of an older generation counts for a newer one only as a predecessor, and the page says so. And we check names. A 2025 sleep-lab study tested a ring called “Circul,” which is not the Circular Ring 2, so we don't count it for Circular.

Missing evidence is scored, not skipped. We found no peer-reviewed validation study of the Samsung Galaxy Ring when we searched in September 2026, and its evidence sub-score reflects that.

Every study we use, with its sample, reference device and funding, is in how accurate are smart rings and smart ring HRV vs a chest strap. If you'd rather check your own ring, here's how to test it against a chest strap at home.

Battery and sizing claims

Battery figures are the maker's claims, and we label them that way. Every claimed figure carries the maker's own conditions. Oura rates the Ring 5 at 6–9 days “depending on ring size and usage” ( Oura); Samsung rates the Galaxy Ring at up to 7 days on sizes 12 and 13 and up to 6 days on smaller sizes (Samsung).

What reviewers saw is attributed to them. DC Rainmaker reported 6–7 days from the Oura Ring 5 (DC Rainmaker). The Gadgeteer's reviewer got about 10 days from the RingConn Gen 3 with alerts on, against RingConn's claim of up to 14 (The Gadgeteer). Those are one person's results on one ring, and we present them that way.

Sizes don't translate between brands, so we never convert them. Oura says its sizes differ slightly from standard US ring sizes and between its own generations (Oura Support). RingConn says its Gen 3 sizes differ from other RingConn rings, standard rings and other smart ring brands (RingConn). No brand publishes an equivalence with another brand's sizes, so our advice is always the same: use the sizer for the ring you're buying. See smart ring battery life and smart ring sizing compared.

What we don't do

• No lab and no testing claims. No page describes a ring as worn, owned or tested by us.
• No paid placement. No brand pays for placement or reviews, and commissions never change a score or the order of a list. See the affiliate disclosure.
• No stale prices. Ring prices appear only from a live feed, with a date, or not at all.
• No star ratings, testimonials or “customers say” blocks. Our scores are editorial and labeled as such.
• No invented credentials. The author's bio says only what he has provided.
• No diagnosis. Rings are wellness devices, and our health content rules set out what we will and won't say about sleep apnea, blood pressure and similar features.
• No guessing. If a brand doesn't publish a figure, the table says so.

How pages are updated

Every page is re-checked on a fixed cadence, and sooner when something changes. This is the schedule we publish and follow:

A page's “Updated” date changes only when its content materially changes: a fee, a score, a pick, a study or a policy. A re-check that finds nothing new doesn't move the date. Each material change adds a dated line to the page's “What changed” list, so you can see what moved and when. Corrections follow the same rule; how to send one is in the editorial policy.

Which sources count

• First choice: the brand's own policy, spec, sizing, support and warranty pages; peer-reviewed validation studies, cited by journal and year; and regulators' own records, such as FDA communications, for anything health-related.
• Second choice: measurements published by established independent reviewers, always credited to the reviewer (“DC Rainmaker reported…”) and never restated as our own finding.
• Not used as evidence: AI-generated listicles, a brand's blog when it describes a rival's product, and forum posts for specs. Owner forums may be quoted as “owners report,” never as fact.

Every page ends with a numbered Sources list, each entry with the date it was checked. If you find a source that contradicts us, send it over.