Study S3 · August 2026

How much does Garmin move your VO2 max?

Key finding

Garmin's VO2 max estimate changes roughly 16 times per 90 days, and when it moves it moves by a median of 0.2 points. Over a full year, the median athlete's figure shifted 2.0 points — less than the error band of the lab test it is imitating.

Gneta, August 2026 · 42 athletes · 180+ days each

42

athletes

14,919

observed changes

0.2

median step size

What we did

We took every athlete with at least 180 days of VO2 max history — 42 of them — and reconstructed the series day by day. Then we asked the plainest possible questions: how often does this number move, by how much, and how far does it travel over a year.

The cohort is fit. Median current estimate 51.9, with the middle half between 47.7 and 55.5.

What we found

Finding 1

It moves often, in very small steps

Median 16.0 changes per 90 days — call it once every five or six days. The middle half of athletes saw between 10.5 and 25.5 changes in the same window.

Across 14,919 observed changes, the median step was 0.2 points (IQR 0.1–0.5). Nine times in ten the step was 1.0 point or smaller.

Finding 2

A year of training moves it about two points

For the 29 athletes with a clean year-apart comparison, the median absolute change over twelve months was 2.0 points (IQR 1.0–3.2).31% of athletes stayed within a point across the year and 52% within two; nine in ten stayed within 4.6.

Published validation work puts consumer VO2 max estimates in a 5–10% error band against laboratory measurement. For an athlete around 52, that band is roughly 2.6 to 5.2 points wide — which is to say a full year of training produced a change comfortably inside the measurement error.

Finding 3

The lifetime range is much wider than the yearly one

Across each athlete's entire recorded history, the spread between their highest and lowest estimate had a median of 9.9 points (IQR 7.7–13.0).

Ten points of range against two points of yearly movement is worth sitting with. Some of that range is genuine long-run fitness change. Some of it is device upgrades, changed heart-rate straps and firmware revisions — and this data cannot tell you the split.

Limits you should know before citing this

  • The cohort is self-selected. Garmin owners who sought out a third-party analytics tool. Fitter and more data-curious than average, and not a random sample of anyone.
  • Change frequency is an upper bound on granularity. This is an observed-day series. Gaps between syncs mean one recorded step can bundle several underlying updates, so the true update rate is at least what we report, never less.
  • We cannot separate fitness from hardware. A new watch or a different heart-rate sensor can shift the estimate on its own. Long-range figures in particular should not be read as pure physiology.
  • The lab comparison is borrowed, not measured. The 5–10% error band comes from published validation literature, not from our cohort. Nobody in this dataset took a lab test.
  • Six accounts excluded. Near-identical multi-year histories under different Garmin accounts, created inside a known incident window. See the methodology.
  • Observational. No causal claim is made anywhere in this study.

The data

The aggregate output and the script that produced it are public. Every number on this page traces to that file.

View the aggregate data (JSON) →