Two independent validations against a gas analyzer agree wherever their samples overlap: if your true VO2 max. is above 55, the watch reads 3–6 ml/kg/min low. The bottom end appeared in only one of them — below 45 the watch reads about 7% high. Together that is a single picture: the watch pulls everyone toward the middle. On the rating table it is one or two rungs in either direction, so an athlete who belongs in the top band ends up reading the middle one as their own. Here is where the number comes from, why the error depends on your level, and what to do with a number you cannot take at face value.
Download as PDF · 8 pagesVO2 max. is the largest volume of oxygen a person can take up per minute per kilogram of body weight at maximum effort. A lab measures it directly: mask, gas analyzer, a graded test to exhaustion. A watch does not measure it — it estimates it, from the heart rate and pace of an ordinary run, with no maximum effort anywhere in sight.
VO2 max. is the maximum volume of oxygen (in milliliters) you can consume per minute per kilogram of body weight at your maximum performance.
fēnix 8 Owner's Manual, VO2 max.
As of edition v9 of the manual, Garmin was not the one doing the math. The model belonged to Firstbeat, the Finnish company Garmin bought in 2020; the rating scale is borrowed too — a table from the American Cooper Institute, reprinted with permission. For a Garmin metric that is unusual: most of the others say neither whose model computes them nor where their levels came from.
VO2 max. data is provided by Firstbeat Analytics™. VO2 max. analysis is provided with permission from The Cooper Institute®.
fēnix 8 Owner's Manual, trademark section, edition v9
Checked against the manual on November 21, 2026. This credit line has changed. The explainer was written from edition v9, which we read on October 29, 2026. In the current edition v10 (November 2026) the VO2 max. page reads "VO2 max. data is provided by Garmin Human Performance Lab™" — Firstbeat is no longer mentioned there. The Cooper Institute permission survived word for word. Garmin does not say what sits behind the change — a move to an in-house model, or only a different name over the same one. The measurement conditions on that page changed as well: read the sections on indoor running and on run duration below with the v10 correction, marked in place and in the sources.
Garmin has published no validation of its own VO2 max. estimate — at least not in the body of primary sources we went through. The Firstbeat white papers from before 2020 describe the estimate, but those are the model's own documents, not an independent check. What there is instead: two outside studies, both against a gas analyzer in a graded lab test, both on a Forerunner 245.
Düking: 23 recreational runners, three outdoor runs, mean error across the whole group 5.7%. Engel: 35 athletes (mean level 60.1) and two fifteen-minute runs, mean error 7.9% and 7.2%
| True VO2 max. | What the watch showed | Study |
|---|---|---|
| below 45 | reads high by 7.1% | Düking, 2022 |
| 45 to 55 | error 4.1% — in this sample the authors called the watch and the lab interchangeable | Düking, 2022 |
| 55 to 61 | reads low by 6.2% — that is 3.0 ml/kg/min | Düking, 2022 |
| up to 59.8 | error 4.1% and 2.8% across two runs | Engel, 2026 |
| above 59.8 | reads low by 6.3 ml/kg/min, error 10.4% and 9.4% | Engel, 2026 |
Two different samples, different years, different authors — one picture. The error is not flat across the scale: it depends on how strong you are. The watch pulls a strong athlete down — both studies saw that. It lifts a weaker one up — one study saw that: the second sample had no group below 45, its mean level being 60.1. The watch is at its most accurate in the middle, around the level of an ordinary recreational runner. The authors of the second study say it outright: in well-trained athletes the estimate is less valid, and where precision matters the lab is still the answer.
Why three milliliters is two rungs. A gap of three to six milliliters looks like small change until the number is read off the table. For men aged 20–29 the top level starts at 55.4, the next at 51.1, the middle one at 45.4. An athlete whose true value is 55.4 will see 52.4 if the watch reads them 3.0 low — a level below their own; read 6.3 low, they see 49.1, two rungs down. The scale compresses the same way from the other end: a runner whose true value is 44, read 7% high, sees 47.1 and climbs a rung. The arithmetic is ours, done on the published Cooper Institute boundaries — neither Garmin nor the study authors made this conversion.
That leads to one cautious conclusion — again, ours. If the error depends on your level, then your own number today means more set against your own number from last month than against somebody else's rung. But that series is a ruler with a catch: as you get stronger the underestimate grows with you, so the climb on screen looks smaller than the real one. Neither study checked the estimate over time — they had two or three runs each.
The bias does not stay put, either. Garmin builds recovery time on this estimate — the manual says so outright — and recovery time is one of the seven inputs to Training Readiness. An estimate that reads low for a strong athlete travels into the morning readiness score with them.
Garmin's interpretations are built the same way almost everywhere: a number is explained not by a mechanism but by a place in the crowd. For VO2 max. the scale is the Cooper Institute table, split by sex and by six age bands, from 20–29 to 70–79. The levels are tied to percentiles — to the share of people of your sex and age you are ahead of: 95 means you are ahead of 95 people out of 100.
One band out of twelve, as an example. The full tables are in the manual's appendix
| Level | Percentile | Men 20–29, ml/kg/min |
|---|---|---|
| Superior | 95 | from 55.4 |
| Excellent | 80 | from 51.1 |
| Good | 60 | from 45.4 |
| Fair | 40 | from 41.7 |
That frame is worth taking literally. A level answers the question "where do I stand among people of my sex and age" — and does not answer the question "what is happening to me". There is no bridge between the two.
The calculation leans on two signals at once: how much work was done, and what it cost you. Heart rate gives the cost — from the wrist or from a chest strap. Outdoors, GPS gives the work: speed there is measured independently of the body producing it. That is our explanation, not a quote: Garmin does not publish how heart rate and pace are converted into milliliters of oxygen.
Garmin adjusts for conditions on its own. Two thresholds are named: above 22 °C (72 °F) and above 800 m (2,600 ft) of altitude — beyond them both the VO2 max. estimate and training status are adjusted. Whether the thresholds work independently or only together, and in which direction and by how much the number shifts, the manual does not say. One thing it does say: the correction needs weather data from your phone. No "correction applied" marker is described anywhere, so you can only reason from the thresholds and from whether the phone was with you.
Indoor run activities do not generate a VO2 max. estimate in order to preserve the accuracy of your fitness level trend.
fēnix 8 Owner's Manual, VO2 max., edition v9. This sentence is not on the page in the current v10
This section was written from edition v9 — see the November 21 note above. The reason is stated plainly, and it is not "the treadmill is worse". What follows is our explanation. A treadmill announces its own speed, and how honestly it does so depends on its calibration — and the watch has no way of knowing whether anyone has calibrated it. For a single workout that does not matter. For a series built up over years it does: you cannot let a value measured by a different method into it. Garmin's wording here is precisely about the accuracy of the trend.
The consequence is worth saying out loud: the number you see on screen in March, after a winter spent indoors, was taken off you in autumn. The watch does not recalculate it and does not flag it as stale — it simply had nothing to update it from.
Load and training effect are still calculated indoors. In the list of exclusions Garmin gives separately, indoor running is listed in the row for the fitness estimate and running economy — not for load.
From the current edition v10 of the fēnix 8 manual, checked November 21, 2026. In edition v9, which the explainer above was written from, the minimum duration was named differently in two places — "at least 10" and "at least 15 minutes". The runs in both validations above were fifteen minutes long
| What | Condition |
|---|---|
| Running | outdoors with GPS; per the current v10 — at least 15 minutes at no less than 70% of maximum heart rate (v9 said both "10" and "15") |
| Cycling | a power meter and a heart rate monitor, moderate intensity |
| Accuracy conditions on a bike | a ride of at least 20 minutes, heart rate above 70% of maximum, steady power |
| What to avoid on a bike | hilly terrain and riding in a group on someone's wheel: power there is ragged and heart rate cannot keep up with it |
Running and cycling — one number or two?
The manual phrases the requirement as “run outside with GPS or ride with a power meter”, which describes one estimate that either sport can update. It does not say either way whether the device keeps a separate running and cycling value — that is an open question, not a claim of ours. You can settle it on yourself in a couple of days: note the number after an outdoor run, then again after a ride with a power meter.
How far can the number jump after a single workout?
Not published. Garmin names neither a step size nor the weight a new run carries against earlier ones. One practical conclusion follows: a single reading decides nothing, you need a series.
The number dropped after a hard or hot race — have I lost fitness?
The screen does not support that. Above 22 °C and above 800 m Garmin adjusts the estimate itself, and the size of the adjustment is not published; on top of that, an all-out effort is not the condition the model is described for. Read the series, not the last point.
The watch reads low — so am I stronger than it shows?
If your level is above 55, that is how it came out in both samples. But these are group means, not a correction you can subtract for yourself: both studies also reported the spread inside the groups. The correct reading is “my real number is probably higher than the screen”, not “my real number equals the screen plus six”.
A series of explainers, "Garmin metrics", built from primary sources. Published wording, thresholds and limits come with a link to the Garmin manual or to a peer-reviewed paper; our explanations, calculations and examples are marked. Fizkult is not affiliated with Garmin, Firstbeat or The Cooper Institute; the names belong to their owners.
The thing that matters most about the VO2 max. on your watch is not the number but the date it was taken — and that date isn't on the screen, so a value you read in March may have been measured back in October. Fizkult prints VO2 max. with the date of the estimate and its age beside it — "last estimate October 12, 41 days ago" — and labels an old one in words: "old data, possibly out of date by now."
Fizkult reads your Garmin data itself and sends a plain-words breakdown to Telegram every morning: what your sleep, recovery and load are saying and what to pay attention to today. It builds your norm from your own history, and while the days are still too few it says so outright: “I don’t know your usual yet: counted over 9 days out of the 14 I need.” You connect the watch once; don't skip the Historical Data box — without it your norm builds only from new days, and more slowly. First week free, everything included. The bot writes in English if your Telegram is set to English; /language switches it either way.
Connect your watch in Telegram →This is a coach's read, not a medical one. If you don't feel you're recovering, see a doctor.