The Oura Ring lawsuit: 95% or 53%, and why the number was never the point

You wake up feeling fine. Then you open the app, see a poor sleep score and one hour less deep sleep than yesterday, and suddenly you feel tired. You trust the ring more than you trust your own morning, because the ring has percentages and you only have a feeling.

The assumption underneath that habit is reasonable: a device that reports your REM sleep to the minute must be measuring your REM sleep. A lawsuit filed in California last week alleges that assumption is wrong, and the science behind the argument is worth understanding whether or not the case goes anywhere. The truth is stranger than either side's headline number.

What the lawsuit actually claims

On 20 August 2026, a proposed class action was filed against Oura in federal court in San Francisco. The plaintiff, a California resident named Surber, is represented by the Clarkson Law Firm, and the complaint targets one of Oura's biggest selling points: the claim that its rings offer "95% Sleep Staging Accuracy compared to clinical sleep lab", as reported by TechCrunch and Android Authority.

The complaint's core argument is biological, not statistical. Sleep stages are defined by brain activity and eye movement. A clinical sleep study, called polysomnography, records both directly with electrodes on your scalp and sensors near your eyes. A ring on your finger has neither. It reads heart rate, temperature, movement and blood oxygen, then uses AI to infer which stage of sleep those proxies probably correspond to. According to the lawsuit, that inference is closer to guesswork than measurement, with what the complaint calls a coin flip's chance of being correct, per TechSpot.

Two things to hold in mind throughout. First, these are allegations in a proposed class action. Nothing has been proven, the class has not been certified, and Oura disputes the claims. Second, the timing is not incidental: Oura confirmed earlier this year that it has filed confidentially for a US IPO after a valuation of around $11 billion, as reported by Euronews. A lawsuit aimed at the flagship accuracy claim lands harder in that context.

The study behind the 53.18%

The number the complaint leans on comes from a real, independent paper. Herberger and colleagues published a validation study in Scientific Reports in March 2025, testing three finger-worn sleep trackers, including the Oura Ring Gen 3, against full polysomnography in a university hospital sleep laboratory.

The setup: 45 patients, 45 nights, each wearing the rings while wired up to the clinical gold standard. Across the four sleep stages, the Oura ring agreed with the lab 53.18% of the time. On average it underestimated light sleep by about 13.8 minutes and deep sleep by about 6 minutes, and overestimated REM sleep by about 31.6 minutes. The authors reported no funding from Oura and no competing interests, and they made a point that matters more than any single percentage: a decent average across a group can hide large errors for an individual night.

Note what the study population was. These were patients at a sleep clinic, many with suspected sleep disorders, not healthy adults. That is exactly the kind of person most motivated to buy a sleep tracker, and exactly the population where the ring struggled most.

Oura's side, and the study that supports it

Oura responded on 22 August. The company said it stands behind its science, research and accuracy claims, pointed to peer-reviewed validation studies, and said it will defend itself in court, per TechCrunch.

That defence is not empty. Robbins and colleagues at Brigham and Women's Hospital, a Harvard-affiliated sleep laboratory, published a comparison in Sensors in 2024 testing the Oura Ring Gen 3, an Apple Watch and a Fitbit against polysomnography in 35 healthy adults. The Oura ring came out best of the three, with sensitivity for individual sleep stages in the range of 76 to 79.5%, and it was not significantly different from the lab on wake, light, deep or REM estimates.

There is also a bigger-picture synthesis. Khan and colleagues pooled six validation studies covering 388 participants for a 2025 meta-analysis in OTO Open and found the ring's total sleep time was within about 3 minutes of medical-grade measurement on average, with sleep efficiency within about 1.3 percentage points. Their conclusion: comparable accuracy for common summary numbers, but not a replacement for clinical diagnosis.

Where the evidence is genuinely messy

So which is it, 53% or 79%? The honest answer is probably both, and that is the part worth sitting with.

• Different populations. The Brigham study tested healthy adults aged 20 to 50. The Herberger study tested sleep clinic patients. The ring performed well in the first group and poorly in the second. If your sleep is irregular or disordered, the validation data that looks best may not describe you.
• Small samples everywhere. 35 people in one study, 45 nights in the other, and only 31 of those 45 nights produced usable Oura data at all. These are thin foundations for a marketing claim in either direction.
• Averages are not your night. A tracker can be roughly right about a group and meaningfully wrong about your Tuesday. The Herberger authors flagged this explicitly.
• Summary totals are easier than stages. The meta-analysis suggests the ring is genuinely good at total sleep time and sleep efficiency. Stage-by-stage classification, the thing the 95% claim is about, is where the numbers diverge.

None of this tells you whether the lawsuit will succeed. Courts will argue about what "compared to clinical sleep lab" means in an advert. What the literature already tells you is simpler: a finger-worn ring estimates sleep stages from proxies, sometimes well, sometimes not, and it cannot see your brain, as Gizmodo put it in its coverage of the wider wearables gap.

What this means for your training

Here is the contrarian bit: even if your sleep tracker were perfect, the stage percentages would still be the least useful number it gives you. You cannot train deep sleep directly. You can control the inputs. We have written about what the research actually supports on sleep and fat loss and training on poor sleep, and the practical rules are boringly stable:

• Trust duration and consistency over stages. Time asleep and a regular sleep window are the things trackers measure best and the things the training research keys off.
• Do not let a score cancel a session. A bad readiness number is a reason to adjust load, not a reason to skip. How you feel warming up is data too.
• Judge your sleep across weeks, not nights. Single-night stage data is exactly where the error bars are widest.
• Spend the worry on protein, training and bedtime. Three inputs you fully control, all with far stronger evidence behind them than any wearable's staging algorithm.

Where Pocket Fit fits, honestly

Pocket Fit is not a sleep tracker, and this is not the paragraph where we claim to be more accurate than a $300 ring. We do not measure your sleep stages at all, and we think that is the right call, because the lawsuit is ultimately about what happens when a product promises lab-grade certainty from proxy signals.

What Pocket Fit does is hold the controllable inputs in one place. The Body budget keeps a running tally of four deposits: your workout, your streak, your sleep and your nutrition. Sleep counts as one deposit among four, a habit you either kept or did not, not a percentage to be audited. Fuel lets you scan a plate or a barcode and check protein, carbs and fat with no red numbers and no shame. Miss a session and the scheduler reshuffles it into the rest of the week instead of dropping it, and the programme updates week to week from what you log.

Georgi, Pocket Fit's founder, went from 122 kg to competing at The Yard Games, and he did not get there by optimising REM minutes. He got there through the bad weeks, which is why the app exists. The result is decided by what you repeat, not by what you measure.

The Oura Ring lawsuit: common questions

What is the Oura Ring lawsuit about?

A proposed class action filed on 20 August 2026 in federal court in California alleges Oura misled consumers by advertising "95% Sleep Staging Accuracy compared to clinical sleep lab" when, the complaint claims, a finger-worn ring cannot directly measure the brain activity and eye movement that define sleep stages. These are allegations, not findings, and the class has not been certified.

Has Oura responded to the lawsuit?

Yes. On 22 August 2026 Oura said it disputes the allegations, stands behind its science, research and accuracy claims, pointed to peer-reviewed independent validation studies, and said it will defend itself in court.

Is the Oura Ring's sleep tracking accurate or not?

The evidence points both ways. A 2024 Brigham and Women's Hospital study of 35 healthy adults found the ring the best of three consumer trackers, with roughly 76 to 79.5% stage sensitivity against polysomnography. A 2025 Scientific Reports study of 45 sleep clinic patients found only 53.18% four-stage agreement. Healthy, regular sleepers appear to get better numbers than people with disturbed sleep.

Where does the 53.18% figure come from?

From Herberger and colleagues' independent 2025 validation study in Scientific Reports: 45 patients, 45 nights in a hospital sleep laboratory, with the Oura ring underestimating light sleep by about 13.8 minutes, deep sleep by about 6 minutes, and overestimating REM by about 31.6 minutes on average. The authors reported no funding from Oura.

Should I stop using my sleep tracker?

Not necessarily. Trackers are reasonably good at total sleep time and consistency, which are the numbers worth acting on. Treat stage percentages and readiness scores as rough estimates, judge your sleep across weeks rather than single nights, and never let one bad score decide whether you train.