I still remember my first proper VO2max test. The mask that made you feel like you couldn’t breathe. The prickle of the needle every three minutes, drawing blood for a lactate reading. The poster of Lasse Virén — or maybe it was Martti Vainio — on the wall of the lab at the Paavo Nurmi -keskus, the only thing to look at, and so the only thing to keep my mind off the pain. The clock that refused to move. One of the rewards for figuratively dying on that treadmill, besides my aerobic and anaerobic threshold paces, was a number they called VO2max. Oh, how times have changed.

Because now you don’t have to die on a treadmill for it. You finish an easy run, and before your heart rate has even come down, there it is on your wrist: two digits. 52. Or 54. Or, on a bad day — a hot day, a hilly day, a tired-legs day — 51, and something in your chest sinks a little. The number went down. You went down.

Somewhere between that lab and this wrist, we collectively decided that this number is us. It’s on the club WhatsApp. It’s in the Strava comments. It’s the thing the longevity podcast said would decide when you die. VO2max has become the runner’s credit score — a single figure that supposedly ranks your engine against everyone else’s.

It’s a bad number to worship, though. First, the one your watch shows you is of shaky accuracy. Second, even a perfect, lab-measured VO2max — the kind I once bled for — would be close to useless for deciding what to do at training on Tuesday.

What the number actually is

VO2max is the maximum rate at which your body can take in and use oxygen. Measured properly it means a mask, a treadmill, a ramp test to exhaustion, and a machine sampling your breath — “indirect calorimetry,” the actual gold standard.

Your watch does none of that. It runs a regression model: it watches your heart rate against your pace on a run and guesses what the lab would have said. That’s not a criticism of the maths — it’s just important to be clear that the wrist number is a guess about a ceiling, not a measurement of anything you did.

Why it got so hot

Three engines, none of them about running faster.

Wearables compute it by default and gamify it. Garmin puts a VO2max on essentially every running watch and stacks a whole tower of features on top of it. Polar literally titles its guide “VO2max: The Ultimate Training Blueprint,”¹ calls it “the ‘gold standard’ of cardiorespiratory fitness,” and slots you onto a colour-coded ladder toward “Elite.” Apple Watch shows a “Cardio Fitness” number and will actively notify you that it’s low. When a number appears, unbidden, after every workout, on a hundred million wrists, it becomes important by sheer repetition.

The longevity crowd moralised it. Peter Attia’s line that VO2max is “the single greatest predictor of your lifespan”² turned a training-lab variable into a memento mori. (Cardiologist Eric Topol, it’s worth noting, calls that specific claim false³ — the mortality evidence is largely observational, and a 2025 causal-inference study found that genetically-predicted VO2max is “not causally associated with… longevity.”⁴)

Strava turned it into a scoreboard. As one coaching site puts it, VO2max is now “the ultimate ‘bragging rights’ number… thrown around on club runs and Strava comments like a badge of honour.”⁵ Or, more bluntly, the metric has become the measure of the runner.⁶

Problem one: the watch number is imprecise — and biased exactly where you’re bragging

Independent validation is unflattering. A 2026 study of the Garmin Forerunner 245 (Engel, Masur, Sperlich & Düking, in the European Journal of Applied Physiology)⁷ found an overall mean absolute error around 7–8% — but with a twist that matters. For moderately trained runners it was decent (error ~3–4%). For highly trained runners it was poor: the watch underestimated their VO2max by about 6 ml/kg/min — a 9–10% error, with reliability so low the authors told serious athletes not to trust it.

That pattern shows up everywhere. Wearables tend to underestimate the fit and overestimate the unfit;³ across devices the error runs anywhere from ~5% to ~16%, and it’s sensitive to terrain, heat, hydration, your max-heart-rate setting, and optical-sensor quirks.⁸

Notice what that does to the bragging use-case. The number is least trustworthy for the very people who most want to compare it — trained runners — and it systematically sells them short. Two runners comparing wrist VO2max aren’t comparing engines; they’re comparing two noisy, differently-biased guesses.

To be fair: most of these studies conclude the estimate is fine as a personal trend line — good enough to tell whether your own fitness is drifting up or down over months. That’s the honest ceiling on the number’s accuracy. It is a rough, private trend gauge. It is not a ruler you can hold up next to someone else.

Problem two: even a perfect number isn’t a training input

Here’s the part the hype never mentions. Suppose your watch were exactly right. What would you do with it?

Nothing, really. VO2max describes the size of your aerobic engine. It doesn’t tell you what session to run, what pace to hit, or what to fix. You don’t train differently because the number reads 54 instead of 52. You program running from paces, thresholds, and effort — none of which the VO2max score hands you.

And as a target, it’s a dead end for most of us:

• It plateaus. VO2max rises fast in a beginner (which is exactly why apps love it — it moves quickly and looks like progress), then flattens. After that, performance keeps improving through everything else⁹: threshold, economy, durability. “Your engine may not get much bigger — but you can become dramatically better at using it.”
• It doesn’t separate runners — in either direction. Once people are reasonably fit, VO2max values cluster, and the real gaps open up in lactate threshold (how much of the ceiling you can hold, and for how long) and running economy (how little oxygen it costs you to hold a pace). In a now-classic study of elite 10k runners with near-identical VO2max, VO2max barely correlated with race time (r = −0.12), while running economy explained 65% of it.¹⁰ And it runs the other way too: runners with nearly the same race times can carry very different VO2max values.¹¹ The number won’t pick the winner among similar runners, and it isn’t shared by runners of similar ability. It simply isn’t the thing.
• The fastest version of you can have a lower number. Runners routinely set personal bests while their measured VO2max drifts down, because they got more economical and lifted their threshold. The ceiling fell; the useful space under it grew.

It’s largely a gift anyway — like your max heart rate

Even the ceiling isn’t really yours to take credit for. In the HERITAGE Family Study — 481 people put through the same 20-week programme — the improvement in VO2max was about 47% heritable¹² and clustered in families: gains ran from basically nothing to over a litre a minute, non-responders sitting right next to super-responders. Your baseline number, before you trained at all, is just as strongly inherited. A large slice of your VO2max was handed to you at birth.

Which makes ranking runners by VO2max a bit like ranking them by maximum heart rate. Max HR is a fixed, individual quirk — it drifts down with age, barely moves with training, and, crucially, says nothing about fitness: a max HR of 200 doesn’t make you faster than someone at 180, and nobody sane brags about it. VO2max isn’t quite that useless — unlike max HR it’s genuinely trainable, and it does track fitness across a wide range — but as a number to hold up against another runner it carries the same defect: you’re mostly comparing genetic endowments, not who trained smarter.

And it matters least in the sport everyone’s tracking it in

Running has an unusually large efficiency component — biomechanics, elastic recoil, fibre type — and it varies enormously from one runner to the next. That variance is exactly why VO2max is such a weak divider among runners: it gets swamped by economy. In sports where efficiency sits in a narrower band, aerobic capacity tracks performance more tightly; running is the awkward case. (The raw values happen to be highest in cross-country skiers, then runners, then cyclists — but that’s mostly how much muscle each sport drives at once,¹³ not a league table of who’s fitter.)

And look at who actually wins. Norway currently produces the best distance runners and skiers on earth, and their training is built around lactate, not VO2max: Jakob Ingebrigtsen runs his intervals with a finger-prick lactate meter at trackside, holding a 2–3 mmol/l “double threshold”¹⁴ — chasing a metabolic sweet spot, not an oxygen ceiling. Lactate is itself a surrogate marker, of course — but at least it’s an actionable one: it tells you, right there on the track, to speed up, ease off, or stop. Even at the summit of cross-country skiing, the analysis of Johannes Høsflot Klæbo is that he doesn’t necessarily own the field’s highest VO2max;¹⁵ technique and efficiency carry the rest. The number the influencers push hardest is the one the actual champions bother with least.

The strongest counter-argument (and why it doesn’t rescue the number)

The best case for VO2max is a 2026 Sports Medicine re-analysis of Joyner’s endurance model across 888 people,¹⁶ which found VO2max was the single strongest predictor of lab performance markers — explaining 65–76% of the speed at lactate threshold.

Read the fine print, though. That cohort spans recreational to world-class — and VO2max always wins when you’re allowed to compare a beginner to an Olympian. The authors say plainly that within homogeneous trained groups its discriminative power drops, and that the study shows no causation: nothing in it says raising your VO2max makes you race faster. It supports “VO2max tracks fitness across a huge range.” It does not support “your VO2max is the number to train by.”

What to chase instead

If you want to spend your attention on things that actually change how you train and race, look under the ceiling — some are measurable, some are just qualities you build:

• Threshold pace — roughly the effort you could sustain for about an hour. You don’t have to actually run a punishing hour to find it (and there’s no point if you never race that far): a lab lactate test, a short 20–30-minute time trial, or even just the honest “comfortably hard — can only gasp a word or two” feel all get you close enough. It’s your money metric from 10k to marathon, and it responds to training you can plan.
• Running economy — get more speed per unit of oxygen. Volume, strides, strength, running relaxed.
• Durability — how little you fade in the back third of a long race. Increasingly the thing that separates good from great.
• And, you know, your actual race times. The least gameable metric there is.

None of these fit neatly into a wrist notification, which is precisely why they’re worth more than the one that does.

So keep glancing at the VO2max if you like — as a rough, private line on whether your base is trending the right way over a season. Just stop treating it as a verdict, a ranking, or a plan. It’s an error-prone guess at a ceiling you don’t train by, biased against you the fitter you get. The number was never the point. The running is.

References