Bad VO2max estimates by Apple watch?

[EarthRower]

I always wanted to test my VO2max, but my local university's sports lab has closed for VO2max testing since COVID started. My Apple watch estimated my VO2max at 47.4 mL/kg/min last September when I had just started rowing. Right now, it estimates my VO2max to be 46.5 mL/kg/min. In the meantime, my rowing ability has improved significantly. For example, I could only do a 30min steady state row in September at a pace of 2:14.2 and an average heart rate of 158. Now, I can do 1-hour steady state with relative ease at a pace of 2:12.9 and an average heart rate of 151. My first 2k test a couple of days ago was 7:41.5. Since it was my first test, I don't think I reached VO2max as I still had ~10 bpm HR to tap into at the end of the test. This number would be equivalent to a VO2max of 56.38 mL/kg/min based on the concept2 VO2max calculator, which is about 20% higher than the Apple Watch estimate.

I wonder if anyone has compared their Apple Watch VO2max estimates with actual tests or rowing-based estimates?

[Dangerscouse]

I'm not a fan of watches for detailing information. Some of it is fairly reliable, e.g. SPO2 & steps, as it's simpler science, but others, like VO2 max is, imo, an add-on to make it look more attractive to customers.

I saw a programme about HR measurements on a Fitbit and, to help conserve battery power, it modulates its readings so you're taking a best guess, rather than an accurate indication

[Tsnor]

5% accuracy would be excellent. Each measurement is 47 ml/kg/min +/- 2.2, so functionally identical. That's a surprise since 4-6 weeks of v02max training will move vo2max a lot. Here's some data on garmin and polar watches vs accurate measurements. https://www.runnersworld.com/gear/a2085 ... r-vo2-max/

For fitbit charge 3, there are only a few activates that adjust estimated vo2 max. Running with GPS does it. Rowing workouts don't.

VO2Max is in "ml/kg/min" so if the estimated "maximum ml of O2 per minute" is constant and your weight changes then VO2max is going to move a bit.

I like the C2 vo2max estimator, but suspect that the watches have 10X or 100X or maybe a thousand times more data behind their estimate than the C2 calculator. The C2 one is fun though.

[EarthRower]

5% accuracy would be excellent. Each measurement is 47 ml/kg/min +/- 2.2, so functionally identical. That's a surprise since 4-6 weeks of v02max training will move vo2max a lot. Here's some data on garmin and polar watches vs accurate measurements. https://www.runnersworld.com/gear/a2085 ... r-vo2-max/

For fitbit charge 3, there are only a few activates that adjust estimated vo2 max. Running with GPS does it. Rowing workouts don't.

VO2Max is in "ml/kg/min" so if the estimated "maximum ml of O2 per minute" is constant and your weight changes then VO2max is going to move a bit.

I like the C2 vo2max estimator, but suspect that the watches have 10X or 100X or maybe a thousand times more data behind their estimate than the C2 calculator. The C2 one is fun though.

The apple Watch's VO2max was based on outdoor walking. I suppose there is a large variance in how walking speed and walking heart rate relate to VO2max. Even though Apple has a lot of data, it may not be able to account for the variance. So the estimated VO2max will be accurate for an "average person", but no one is really an "average person". I guess the Concept 2 calculator faces the same problem.

My body weight has decreased by about 2 kg since I started rowing. So, if everything is equal, I should see a slight increase in VO2max just based on the math. However, I don't believe that I have had no improvement in VO2max since I started rowing in September. I am only 40k away from completing my first million meters. I averaged about 60k per week in the last 12 weeks, and I am able to do a lot more than I could when I just started.

[EarthRower]

I'm not a fan of watches for detailing information. Some of it is fairly reliable, e.g. SPO2 & steps, as it's simpler science, but others, like VO2 max is, imo, an add-on to make it look more attractive to customers.

I saw a programme about HR measurements on a Fitbit and, to help conserve battery power, it modulates its readings so you're taking a best guess, rather than an accurate indication

I think heart rate and steps are probably very accurate for Apple watch when in the exercise mode if it is worn relatively tightly. SPO2 is not because the Watch uses reflected light. It underestimates mine by about 3-5% compared to two different FDA-approved fingertip pulse oximeters.

If SPO2 is considered by Apple algorithm for VO2max, it could be another reason for the unreasonable estimates of my VO2max, IMO.

[JaapvanE]

I wonder if anyone has compared their Apple Watch VO2max estimates with actual tests or rowing-based estimates?

As one of the developers of Open Rowing Monitor, I dove into the iner workings of VO2Max, and how it is calculated by most. Please note, the labtest you mention is in fact the golden standard: they push you to the limit to find out how much O2 you can process. It is a tough test, but the data is extremely accurate. All other tests or measurements are approximations of that number.

In essence, there are two approaches to estimate VO2Max without pushing the patient to the edge:

1. Use a known relation between maximum (rowing) effort and VO2Max. In essence, Concept 2 established a relation between an all out timed 2K effort, and VO2Max. There obviously is a relation, HOWEVER, there are quite some factors that could disturb it. For example, people's (lack of) technical rowing capabilities make them much less effective, greatly affecting their 2K time, but their ability to take in O2 and bring it to the muscles (i.e. VO2Max) might be quite OK. So the resulting relation is certainly not bulletproof.
2. Extrapolate the heartrate. Basically, you find the relation for Heartrate and power, extrapolate it to find the maximum power for the maximum heartrate, and then transform that into VO2Max (there are standardised formula's for that for several activities). Biggest issue is that nobody knows their maximum heartrate, so this isn't accurate in any way to start with. Next, again effectiveness is missing from the equation: VO2Max should tell you about your lungs ability to process O2, not how good/effective you are at rowing/walking/cycling.

All in all, both are estimates of that absolute number determined in the lab, but aren't really accurate. So getting two totally different numbers from two sources thus isn't odd. Seeing that number improve is a good sign, although you don't know if this is because you get more effective at rowing/walking/cycling or that your O2 uptake indeed improves. But it is an interesting parameter nonetheless.

[GlennUk]

Seeing that number improve is a good sign, although you don't know if this is because you get more effective at rowing/walking/cycling or that your O2 uptake indeed improves. But it is an interesting parameter nonetheless.

I think this is the key point about data collecting outside of a lab environment.

There are so many variables that make reliable data collection difficult nay impossible for the average user who doesn't have access to lab grade kit.

The best we can hope for is that data we collect and analyse is consistent in its accuracy and precision so that we can monitor change.

Assuming that Apple is always 'wrong' to the same degree for each measurement then its the change that is important and the direction it moves. If we are looking for 'bragging rights' then find a device the gives you the best figures since there is reasonable uncertainty over the values returned by the various devices anyway.

FWIW the only true metric of course, that is not really open to quite so much debate, is how far you can row in a given time, irrespective of what 'the stats' say.

JMHO

[JaapvanE]

FWIW the only true metric of course, that is not really open to quite so much debate, is how far you can row in a given time, irrespective of what 'the stats' say.

In all honesty, combining it with HR is interesting: for example a 2K might still take you 8 minutes, but your average HR (and thus effort) is dropping so your condition or effectiveness is improving.

But I haven't found a decent meteic that behaves like that.

[Sakly]

FWIW the only true metric of course, that is not really open to quite so much debate, is how far you can row in a given time, irrespective of what 'the stats' say.

In all honesty, combining it with HR is interesting: for example a 2K might still take you 8 minutes, but your average HR (and thus effort) is dropping so your condition or effectiveness is improving.

If your condition or effectiveness is improving, you will be faster on the same distance, if you are pushing all out (and this was meant by Glenn). Probably this will be at the same heart rate then.
If you use the same pace, you are right. Average HR will go down, as you don't need the same effort anymore from metabolic point of view.

[nick rockliff]

Over the years, I've undergone blood lactate and V02MAX step tests numerous times. I've always been of the understanding that the absolute can only be increased between 5% % 15% through training and that's it. As you get older it will decrease no matter what you do.

When expressed as mL/(kg·min) the lighter you are the better it will look but the absolute will be what counts.

You V02MAX isn't necessarily at the same point as your MHR either.

I can't understand why people use data from watches and the like and expect them to be correct/accurate?

If you train and your performance improves you know you are doing something right.

[JaapvanE]

I can't understand why people use data from watches and the like and expect them to be correct/accurate?.

I guess because it is a number everybody is talking about, VO2Max sounds scientific, and people are told it is fairly accurate when they tested the watch on highly trained athletes.

FirstBeat, who implements the algorithm for Garmin and others, describes the algorithm here: https://assets.firstbeat.com/firstbeat/ ... 6.2017.pdf There is quite some research behind it, but they also indicate there are too many variables to account for.

[Sakly]

And in a 2k there is only one number of importance - you time finished the race. If you reach that with VO2max of 45 or 54 does not matter

[HornetMaX]

I can't understand why people use data from watches and the like and expect them to be correct/accurate?.

I guess because it is a number everybody is talking about, VO2Max sounds scientific, and people are told it is fairly accurate when they tested the watch on highly trained athletes.

FirstBeat, who implements the algorithm for Garmin and others, describes the algorithm here: https://assets.firstbeat.com/firstbeat/ ... 6.2017.pdf There is quite some research behind it, but they also indicate there are too many variables to account for.

Interesting. But which devices actually allow to estimate VO2max via the Firstbeat protocol on a rower ?

[JaapvanE]

Interesting. But which devices actually allow to estimate VO2max via the Firstbeat protocol on a rower ?

It is possible, according to https://sportcoaching.co.nz/how-does-ga ... e-vo2-max/ the formula is known:

Rowing (Indoor): Theoretical VO2 (ml/kg/min)=(14.72*Power+250.39)/person’s weight

Unit of power = watts (W)
Unit of weight = kilograms (kg)

When compared to the same analysis using C2's approach, you get a similar number (with ORM, I structurally compare the two results if borh deliver valid results).

However, no watch to date has implemented this, as far as I know. In theory, a watch recieves sufficient information to calculate it from a PM5 when it supports recieving the C2 ANT+ interface. But Garmin has not implemented this (yet).

I must say, I'm not certain about the quality of the numbers it produces. My Garmin, based on my walking data, estimates my current VO2Max to be 44 to 45. Based on rowing, ORM estimates my VO2Max around 33 both via C2's 2K method and the above formula from Firstbeat. I'm the first to admit my rowing technique could improve, but I have a hard time believing it is that inefficient.

On a positive note, according to my Garmin, 4 years ago my walking VO2Max was 35 and thanks to rowing I've seen it improve to 47 (now in an easier period, so around 45). To me, the impovement is what counts, not the absolute number.

[HornetMaX]

I tried to look at FirstBeat white papers but they seem very secretive on what they actually use. They do explain that they throw away part of the data if it looks suspicious (e.g. halting at a red light) but other than that they do not say what they do with the resulting data.

I suspect they did something robust for runners (which can run into red-lights, slopes, sand, bridges breaking GPS connection etc) and not requiring a max effort standardized test (e.g. a 2K).

For indoor rowers I'm not convinced you need all that stuff: only thing that may be of interest is to provide an estimate without doing a max effort 2K.
Once you're OK to do a max effort 2K, a simple formula (backed by statistical data) is probably more than enough.