From @jem_arnold: calculate your 'predicted' VO2max

VO2 = Weight^-0.854 * Height^1.44 * exp(0.424 - 0.346 * Sex - 0.011 * Age)

VO2: ml/kg/min
Sex: M = 0, F = 1
Height: cm
Weight: kg
Age: yrs

I’m just gonna guess that pretty much everybody reading this is going to be WAY above their ‘predicted’ number.


About half of what Garmin says I have. Mmm.

1 Like

‘How to make Garmin VO2 look credible’


Garmin is supposed to be +/- 5% if there’s an accurate observed max HR based on the right effort (i.e. not 220-Age)

I am assuming this is from something like a regression. If so, then users should understand that it’s an estimate of the average VO2max given your weight, height, sex, and age. Average based on whatever population was represented in the sample.

If you want to use this as a prediction for your VO2max, that’s fine, but just be aware that all predictions from things like regression are going to have some prediction error. You may be above average, you may be below average - it is, after all, in the nature of averages. Some people will be way above average. If the sample is based on a general population, then sure, a lot of people here will be quite a bit above average.

From the formula, my predicted average is 41.86 mL O2 / kg min. I’m a 43 year old male. I’m pretty sure that is already a fair bit above the average for men in the general population.

1 Like

About as accurate as 220 - age for max HR. In other words not a lot of use as a predictor for the individual.


Where’s @WindWarrior ? I bet he’s double the population predicted VO2max.

I think it just goes to show how much a little fitness improvement could really help a lot of people. I’m just plugging some different ages into the equation…if it’s true then quality of life for a retiree that’s a standard deviation away from normal must not be very good.

1 Like

Add this equation to the pile…

bmjosem-2022-December-8-4–inline-supplementary-material-3.pdf (157.4 KB)


Does Jem Arnold say where the data from this comes from? Is it “general population”?

There’s quite a large penalty for being female. I do get that females have lower vo2max in general, but I wonder if there’s a double effect of there being less physically active females than males in the general population.

Double? No way. Predicted is 37.1 and last year I peaked at 45.6 on intervals. Which if I magically dropped down to freshman college weight and no change in fitness would be 56.3 :thinking:

1 Like

Spot on Garmin’s estimate for me give or take 28.5%, I’m sure that’s standard differentiation :wink:

1 Like

Prediction: 39


(Based on an accurate Max HR, but likely my peak for the year after a dedicated VO2 block and threshold right after for a final bump…)


@BCM : crushed it!

1 Like

There is really no way to know v02max without going into the lab. Since the power you are producing at maximal aerobic power is a function of your efficiency and your oxygen uptake, you really can’t read too much into vo2max based on HR and power data (unless you already know your efficiency ratio). Garmin does a decent job, but it’s assuming you have average efficiency. I’ve done v02max testing in the lab a couple times and both times I was about 5% under the garmin estimate and my efficiency is a bit higher than average. All that said, v02max (and absolute O2 uptake) are interesting things to know, but they don’t really inform training that much. Your maximum aerobic power is a much more useful metric to track/train and it’s easier to estimate out of the lab.

I agree with this, but at the same time, Garmin is one of the few estimates that has been at least tested against lab data. They quote +/- 5% IF your max HR is correct.

From my perspective, I find it interesting how the Garmin number changes though. I dropped into the mid-high 50’s in the off-season, but doing a dedicated VO2 block followed by a number of hard FTP workouts brings the number up. And most interesting for me, the peak is higher by 2 points for me which shows progress.

Is my VO2 Max changing that much? No, but it does reflect the types of workouts I’m doing and the HR vs. Power modeling they do.

1 Like

Garmin was interesting in the data as last year I was dealing with an injury most of the year. Saw my VO2 number from Garmin get down to 48. Lowest I had seen since I started cycling 9 years ago. So this year when I did a build block the number just kept rising back to the high 50s. Did my VO2 really change or did I just lack the ability to ride to a level that Garmin would be happy again with the calculation.

Thanks for posting! Keep in mind this equation is estimated from a population of “healthy” USA adults, i.e. untrained individuals. With large individual variation of something like ± 20 ml/kg/min around the mean estimated VO2max values.

So trained athletes will tend to be higher than these predicted values, but with very large error predicting any one individual VO2max value.

The twitter thread was just an excuse to create some interesting isocline figures :laughing:


Individual variability (Bland-Altman plots)

A better prediction equation for trained cyclists & runners comes from this one below. This usually gives me more accurate numbers vs lab measurements.


The problem in general with predicting individual values from gadget- or population-based equations is that, they will be reasonably accurate for most people, but you don’t know how accurate it will be for you.

Every athlete/participant who I test in the lab shows me their Garmin estimated VO2max, and most of them are very accurate! Then, every so often an estimate will be waaaaay off in one direction or the other. And I can’t predict if the next athlete to walk into the lab will be the former or the latter.

More on this problem:


No plans to lab test, but my Garmin estimates and WKO estimates always seem to be aligned and reasonable given the power and other data I’ve fed them. And a good bit different from the Apple walking based estimates of vo2max, would probably help if I actually decided to run again.

1 Like

I lab tested many years ago. I’m absolutely confident that I am fitter today (but a decade older), and my Garmin estimate is nearly identical.