In September 2026 Ultrahuman changed how its ring turns pulse timing into HRV, and members who received the change saw their nightly number fall by about a third. Ultrahuman’s paper on the change, posted on its website on 30 September 2026, describes how the ring builds the number in more detail than the app does. This guide covers the Ring AIR and the Ring PRO, which both use the Ultrahuman app.
Documentation checked: 22 October 2026. Ultrahuman updates its app and algorithms often, so details can change.
What HRV number Ultrahuman shows
The nightly number is Sleep HRV. Ultrahuman’s science team defines it as “the mean RMSSD RMSSD Root mean square of successive differences: an HRV metric built from how much each heartbeat interval differs from the next, in milliseconds. It mainly reflects vagal activity. More about RMSSD of readings retained during sleep” (Ultrahuman Science, 2026). RMSSD is the root mean square of the differences between neighbouring beat intervals, the standard short-term measure (Task Force, 1996). The consumer explainer skips the acronym and says the ring reports “the typical difference between one gap and the next in milliseconds” (Ultrahuman, HRV explainer).
That puts the number in the same family as the RMSSD values from Oura, WHOOP and Polar, and apart from the SDNN SDNN The standard deviation of all NN intervals in a recording, in milliseconds. It measures overall variability, and its value depends strongly on how long the recording lasts. More about SDNN that Apple Health stores for Apple Watch. It still will not match theirs, because each device averages a different part of the night.
When the Ultrahuman Ring measures HRV
The ring measures HRV overnight. An optical sensor reads the pulse in your finger, and an accelerometer tells the ring when your hand is still enough for the timing to be trusted (Ultrahuman, HRV explainer). The ring records about 24 seconds of beat intervals every 5 to 6 minutes. The updated algorithm removes intervals outside the physiological range, replaces outliers with a local median and corrects timing jitter. A quality check then withholds any reading with too few intervals or too many corrected ones (Ultrahuman Science, 2026).
From your finger to Dynamic Recovery
Drawn from Ultrahuman’s own descriptions
- 1 Short readings About 24 seconds of pulse timing every 5 to 6 minutes, from the optical sensor on your finger, when the accelerometer says your hand is still.
- 2 Quality check Intervals outside the normal range are dropped and outliers replaced. A reading with too few intervals or too many corrections is withheld.
- 3 Sleep HRV The mean RMSSD of the readings kept, in milliseconds. In the app: All Sleep HRV, by week, month or year
- 4 HRV Form Last night against your own overnight range, built over your first weeks of wear. In the app: The Recovery card, with a 7-day view
Dynamic Recovery Score, 0 to 100
- HRV Form
- Sleep Index
- Resting heart rate
- Skin temperature
- Stress Rhythm Score
The four overnight contributors are set by the time you wake. Stress Rhythm can move the score by up to 10 points either way during the day. Ultrahuman publishes no weights.
A timeline from falling asleep to the following day. Through the night, a row of short ticks stands for readings of about 24 seconds taken every 5 to 6 minutes; a few are hollow, meaning the quality check withheld them. After waking, a wavy line stands for the Stress Rhythm Score, which follows heart rate and HRV through the day. Below are four steps: readings from the optical sensor on the finger while the hand is still; a quality check that drops out-of-range intervals, replaces outliers and withholds readings with too few intervals or too many corrections; Sleep HRV, the mean RMSSD of the readings kept, shown in All Sleep HRV; and HRV Form, last night against your own overnight range, shown in the Recovery card. Last, the Dynamic Recovery Score from 0 to 100 combines HRV Form, the Sleep Index, resting heart rate, skin temperature and the Stress Rhythm Score; the first four are set by the time you wake, and Stress Rhythm can move the score by up to 10 points either way.
Ultrahuman does not publish the exact start and end of the sleep window or the thresholds of the quality check. Its explainer calls the result “an average across the long, quiet stretches of your sleep rather than a single moment” and argues that sleep is the most repeatable condition most people have (Ultrahuman, HRV explainer).
A whole-night average of short readings is closer to Oura’s method than to WHOOP’s or Polar’s. The model below shades the readings Ultrahuman averages on one simulated night, beside WHOOP’s deep-sleep window and Polar’s first four hours.
Same night, different numbers
A teaching model, not a measurement
Sleep stages lights out to waking
HRV every 5 minutes (RMSSD) 20 to 60 ms
- Ultrahuman
- 47.5 ms RMSSD
A reading of about 24 seconds every 5 to 6 minutes while asleep; the number is the mean RMSSD of the readings that pass a quality check. The model keeps every reading.
Same night, same heart: Ultrahuman 47.5 ms, WHOOP 54.6 ms and Polar 44.8 ms, because each device averages different hours.
The buttons work when JavaScript is on. The window for each device is listed in the description.
Two charts over eight hours of simulated sleep. The top chart is a hypnogram stepping between awake, REM, light and deep sleep, with deep sleep mostly in the first half and REM mostly in the second. The bottom chart is a line of HRV values, one per five minutes, that rise gently through the night and dip during wakeful minutes. One button per device shades the part of the night that device averages. Ultrahuman: 47.5 ms RMSSD. WHOOP: 54.6 ms RMSSD. Polar: 44.8 ms RMSSD. Same night, same heart: Ultrahuman 47.5 ms, WHOOP 54.6 ms and Polar 44.8 ms, because each device averages different hours.
The ring keeps reading during the day. The Stress Rhythm Score “tracks how your heart responds over the course of the day”, analysing heart rate, HRV and resting heart rate Resting heart rate How many times your heart beats per minute while you are fully at rest. For adults it is usually 60 to 100 beats per minute, and lower in well-trained athletes. More about Resting heart rate against your circadian rhythm Circadian rhythm The body's 24-hour clock. Heart rate is lower and vagal HRV higher at night than by day, so readings are only comparable when they are taken at the same time of day. More about Circadian rhythm , and labels you stimulated, relaxed or stressed (Ultrahuman, Ring PRO). Those daytime readings are not part of Sleep HRV.
Why your Ultrahuman HRV dropped in September 2026
In September 2026 Ultrahuman moved some members’ rings to an updated HRV algorithm. Its own analysis, posted on its website and not peer reviewed, compares 1,026 of those members with 4,972 whose rings were updated in the same weeks but kept the old HRV algorithm. For each member, it set the median of the first 14 nights after the update against the median of the 14 nights before (Ultrahuman Science, 2026):
- Sleep HRV fell by about a third. The median change was −33.8%, from 69 ms before the update to 46 ms after it, starting from the first night.
- Resting heart rate held at a median of 57 bpm.
- The comparison group’s Sleep HRV did not change.
- Other numbers moved far less. Recovery fell by a median of 2 points, deep sleep by 5 minutes, and time awake rose by 3 minutes. Sleep score and Stress Rhythm moved by less than 1%.
From the steady heart rate and the unchanged comparison group, Ultrahuman concludes that “the lower values therefore come from the updated computation”, and advises that “nights after the update are best compared with other nights after it”. The paper does not say how you can tell whether your ring has the new algorithm, whether the app resets your baseline Baseline Your own usual HRV level and range, built from weeks of readings taken the same way. Comparing a reading with your baseline says more than comparing it with other people. More about Baseline after the switch, or why deep sleep moved; its chest-strap study did not measure sleep stages.
If your Sleep HRV stepped down by about a third from one night to the next in or after September 2026 while your resting heart rate held, that matches the pattern Ultrahuman reports. Treat the nights before and after as two series, and read the raw milliseconds against post-update nights until you have a few weeks of them. Can a software update change my HRV covers the general case.
HRV Form and the Dynamic Recovery Score
HRV Form is “your heart rate variability baseline, and how far last night departs from it”. It is one of five contributors to the Dynamic Recovery Score Readiness score A single number, often out of 100, that a device builds from your HRV, resting heart rate, sleep and other inputs to suggest how ready you are for the day. The weighting is not published. More about Readiness score , which “summarizes how close your body sits to its baseline rested state on a scale of 0-100” (Ultrahuman, Recovery Score):
- Sleep Index: how long you slept, how efficiently, the stage split and your timing.
- Resting heart rate: overnight, against your own baseline.
- HRV Form: last night’s HRV against your baseline.
- Skin temperature: how far it sits from your usual overnight level.
- Stress Rhythm Score: how heart rate and HRV move across the day.
“Each is graded on its own; then the five combine into the 0 to 100 score.” Ultrahuman publishes no weights. The four overnight contributors are settled when you wake. Stress Rhythm picks up a few hours later and can move the score by up to ten points either way, which is why the number changes during the day. Ultrahuman counts above 85 as optimal and 67 to 85 as high (Ultrahuman, Recovery Score).
What is in a readiness score?
From the makers' own documentation
- HRV : named by the maker HRV Form: your HRV baseline and how far last night departs from it; also the Stress Rhythm Score, which tracks heart rate and HRV across the day.
- Resting heart rate : named by the maker Overnight resting heart rate against your own baseline.
- Breathing rate : not named Not named by the maker.
- Sleep : named by the maker Sleep Index: length, efficiency, stage split and timing.
- Temperature : named by the maker Skin temperature against your usual overnight level.
- Activity or load : not named Not named by the maker.
Ultrahuman's Recovery Score names 4 of the six inputs: HRV, resting heart rate, sleep and temperature. "Each is graded on its own; then the five combine into the 0 to 100 score." No weights are published.
Source: Recovery Score (Ultrahuman)
- HRV : named by the maker HRV Balance: "How has my HRV changed over the past few weeks?" (a two-week average against three months).
- Resting heart rate : named by the maker Resting Heart Rate: "Was my RHR substantially higher or lower than usual?"
- Breathing rate : not named Not named by the maker.
- Sleep : named by the maker Sleep, Sleep Balance, Sleep Regularity and Recovery Index (hours of sleep after the heart rate reached its low).
- Temperature : named by the maker Body Temperature: "Is my body temperature higher or substantially lower than usual?"
- Activity or load : named by the maker Previous Day Activity and Activity Balance (recent load).
Oura's Readiness Score names 5 of the six inputs: HRV, resting heart rate, sleep, temperature and activity or load. "The sum of these daily Readiness Contributors make up your Readiness Score." No weights are published.
A list of scores, one per device, with six inputs for each: HRV, resting heart rate, breathing rate, sleep, temperature and activity or load. Ultrahuman's Recovery Score names 4 of the six inputs: HRV, resting heart rate, sleep and temperature. "Each is graded on its own; then the five combine into the 0 to 100 score." No weights are published. Oura's Readiness Score names 5 of the six inputs: HRV, resting heart rate, sleep, temperature and activity or load. "The sum of these daily Readiness Contributors make up your Readiness Score." No weights are published.
The baseline needs “a few weeks of consistent nightly wear”, and until then Ultrahuman suggests reading the raw milliseconds rather than the verdict (Ultrahuman, HRV explainer). How long until my HRV baseline is reliable explains why it takes that long.
Where to find HRV in the Ultrahuman app
- The Recovery card. Ultrahuman points to “the 7-day HRV Form in your Recovery card” as the place to start (Ultrahuman Blog, training). The Dynamic Recovery card shows how many contributors are in range; tap it for the full list, with the one driving your score named (Ultrahuman, Recovery Score).
- All Sleep HRV. Your nightly values, with weekly, monthly and yearly views (Ultrahuman Blog, training).
- Apple Health. With the integration on, HRV syncs to the Health app under Browse, Heart, Heart Rate Variability (Ultrahuman Blog, Apple Health). Ultrahuman’s guide does not say which statistic it writes there. Apple’s HRV type in HealthKit is defined as SDNN, and the type named for RMSSD that Apple added with iOS 27 has no description yet (Apple Developer). Until Ultrahuman says which value it sends, keep the ring’s numbers in Health apart from Apple Watch readings, and use the Ultrahuman app for your trend.
Where do I find HRV on my device has the same steps for other brands.
Does Ultrahuman need a subscription?
No, according to Ultrahuman: “there are no recurring subscription fees for accessing your data”, and the ring is “a simple, one-time purchase” (Ultrahuman, US store). PowerPlugs, add-on features in the app, come free or paid. Ultrahuman says charges apply “only to medical-grade experiences or partner-developed solutions” and that PowerPlugs work “without altering or limiting any of the core app functionalities” (Ultrahuman, PowerPlugs). Ultrahuman Max is an optional plan, from $9.99 a month or $99.99 a year in the US, that bundles the premium PowerPlugs with UltrahumanX cover for damage, loss, theft and resizing (Ultrahuman, US store).
On 22 October 2026, Ultrahuman’s US store listed the Ring PRO at $479, and its comparison table listed the Ring AIR at $349 (Ultrahuman, US store). The cheapest way to track HRV compares costs across devices.
How accurate is Ultrahuman HRV?
We found no peer-reviewed study that compares the Ultrahuman Ring’s HRV with an ECG ECG A recording of the heart's electrical activity, taken with electrodes on the skin. ECG-based RR intervals are the reference standard for measuring HRV. More about ECG ; we searched PubMed, Europe PMC and Crossref on 22 October 2026. The ring was not in the 2025 home study by Dial and colleagues, the independent overnight comparison most device pages here rely on, and it has no rows in the device validation tracker.
Ultrahuman has published its own comparison (Ultrahuman Science, 2026). Nine adults aged 20 to 40 slept one night each in the ring, running the updated algorithm, and an ECG chest strap. The authors, all Ultrahuman staff, report close agreement for the nightly value and much wider scatter for single 24-second readings. The three nights that read more than 10% high were the nights on which the quality check withheld about half the readings. The paper is on Ultrahuman’s website, not in a journal, and carries no conflict-of-interest statement. On five of the nine nights the strap covered less than half the hours the ring recorded. We will add the ring to the tracker when a peer-reviewed study appears.
For comparison, this is how the devices in the Dial study did against a chest strap. The two rings in it were Oura’s.
Overnight RMSSD against a Polar H10 chest strap
Dial and colleagues, 2025
13 adults, 536 nights at home; differences between Oura and WHOOP in percentage error were not statistically significant.
Source: Dial and colleagues, Physiological Reports, 2025. Every study we have checked, with its numbers, is in the device validation tracker.
A dot chart in two panels, one row per device tested by Dial and colleagues (2025) against a Polar H10 chest strap. The first panel shows concordance (CCC), where 1 is perfect agreement; the second shows the mean absolute percentage error (MAPE), where lower is closer. Oura Ring Gen 4: concordance 0.99, average error 5.96%. Oura Ring Gen 3: concordance 0.97, average error 7.15%. WHOOP 4.0: concordance 0.94, average error 8.17%. Garmin Fenix 6: concordance 0.87, average error 10.52%. Polar Grit X Pro: concordance 0.82, average error 16.32%. 13 adults, 536 nights at home; differences between Oura and WHOOP in percentage error were not statistically significant.
Oura’s result tells you nothing about Ultrahuman’s ring. As Ultrahuman’s paper puts it, “each device’s HRV therefore needs its own comparison against an ECG”.
How to read Ultrahuman’s numbers
- Look at Sleep HRV before the score. Dynamic Recovery mixes five inputs, and Stress Rhythm alone can move it ten points while your overnight HRV stays put.
- Read weeks, not nights. Ultrahuman’s own advice: “Read it weekly, and let single nights go” (Ultrahuman, HRV explainer).
- Mark the update. Keep values from before your ring moved to the new algorithm apart from those after it.
- Wear it every night. Ultrahuman says wearing the ring every night “is what makes the comparison work” (Ultrahuman Blog, training), and your range needs a run of ordinary nights.
- Treat a higher number as a different number. Ultrahuman: “a bigger number is a different number, not a better one.” Compare the ring only with itself. HRV by device lists what each brand measures, and normal HRV explains what published values can tell you.
For other rings and bands, see Oura HRV, WHOOP HRV and WHOOP vs Oura.
When to talk to a doctor
The Ultrahuman Ring is a wellness device, and Ultrahuman calls HRV “a wellness signal, not a medical diagnosis” (Ultrahuman, HRV explainer). Talk to a doctor if low HRV or a run of low recovery scores comes with feeling unwell, lasting fatigue, palpitations, dizziness, fainting or chest pain, whatever the number says.