VO2 Max calculation issue

PJohann said:

It doesn't explain why Garmin "merge" my running and biking VO2Max.

Now I must wait 1-2 months until the VO2Max come back to my normal VO2Max, and pray God that Garmin will not merge again both VO2Max)

For me this merge is a kind of "bug" or bad code.

Sorry to belabor the point, but I agree with this 100% (that this behavior is undesirable, regardless of the interpretation of why it’s happening). 

I know I’ve already said this, but my understanding (and the understanding of most ppl who talk about Garmin VO2 Max) is that there are only two VO2 Max values that Garmin tracks:

- “VO2 Max”. Often referred to as “general VO2 Max” or “running VO2 Max” by laypeople on the internet. Yes, in the absence of running data, it is documented that walking activities or all-day heart rate can also be used to calculate this value. I still think that regardless of the source of the data, this VO2 Max is always the basis of the running race predictor. Ofc I can’t prove this, but it’s my intuitive understanding. I will also note that when you open the Connect website and look at the VO2 Max graph, the JSON network response refers to this as generic VO2 Max. (Note that the response only contains data for generic and cycling VO2 Max)

- “Cycling VO2 Max” - self-explanatory

Yes, either of these values can be used for training status, as the documentation states. I don’t think that implies that “VO2 Max” changes its meaning / source depending on what activity data is available.

IOW, I don’t agree that “VO2 Max” is synonymous with dominant VO2 Max, a value that represents either running VO2 Max, walking VO2 Max, all-day HR VO2 Max, or cycling VO2Max at any given time, I think generic/running VO2 Max and cycling VO2 Max are calculated, stored and displayed separately, but apparently they can also influence each other. I think my explanation makes more sense for various reasons, including the wording of the documentation, the display of the graphs in the UI, the display of the numbers on the watch, and the fact that following the big spike in cycling VO2 Max (after you stopped running), “VO2 Max” and “Cycling VO2 Max” were not instantly equal. If “VO2 Max” simply “switches sources” to show “dominant VO2 Max”, then it doesn’t make sense to do any kind of “smoothing” of the value when the source changes.

It seems that you also agree with me that “VO2 Max” means “running VO2 Max”, based on your quote.

One way to test my theory would be for you to look at your running race predictor. Since your “VO2 Max” has skyrocketed, have your predicted running race times gone down significantly? If the answer is yes, that supports my theory that “VO2 Max” is “running VO2 Max”. If the answer is no, that supports the theory that “VO2 Max” is “dominant VO2 Max” (which in this case, means cycling VO2 Max)

I fully admit that I could be wrong, but it seems that my interpretation is simpler and fits my observations better, including the following anecdote.

As I said, I have seen it work the opposite way that you’ve experienced. Rather than seeing VO2 Max changing to match Cycling VO2 Max, I’ve seen the initial value of Cycling VO2 Max set very close to my running VO2 Max (even though my lone cycling w/power activity contained bad power data, as you will see), and this can’t be explained by the “dominant VO2 Max” theory.

I’m primarily a runner and I never cycle with a power meter, except for one time when I cycled indoors and accidentally had Stryd paired as a power meter (I was testing an app.) Clearly using a Stryd running footpod as a power meter for indoor cycling is not a valid use case and should not be expected to generate a correct cycling VO2 Max estimate. In this case my cycling VO2 Max was reported once (there’s only one data point on the graph, and my watch still refers to that reading from 3 years ago by date). It just so happens that the (clearly fake) cycling VO2 Max that was estimate was very close to my running VO2 Max at the time. In runalyze I can see both values, and the difference is 0.30. This doesn’t seem like a coincidence, it seems like Garmin used my generic VO2 Max as a baseline to calculate cycling VO2 Max, especially since I only had one cycling activity with power, and cycling VO2 Max was estimated immediately.

Playing devil’s advocate, it makes sense from the standpoint that if cycling VO2 Max is generally known to be 10% lower than generic/running VO2 Max, then if data is light for one metric, that metric can be estimated from the other metric. e.g. If someone’s generic VO2 Max is estimated as x, then cycling VO2 Max can be estimated as 0.9x. If someone’s cycling VO2 max is estimated as y, then their generic VO2 Max can be estimated x / 0.9. This sounds great, but where it falls apart is when one metric has a large error in its estimate and it’s used to influence the value of the other metric.

I also think my theory can be tested (somewhat) when you start running again.

Will your running VO2 Max instantly change back to its old value (or a value slightly lower than that)? That would support the theory that “VO2 Max” is “dominant VO2 Max” (i.e. it can “switch sources” between running and cycling, depending on available data).

Or will your running VO2 Max gradually *approach* its old value? That would support my theory that “VO2 Max” is really generic/running VO2 Max, and cycling VO2 Max was used to augment/estimate running VO2 Max while you weren’t running.

Another way you can test my theory is if you stop cycling after you start running (and your running VO2 Max falls back to its “real” value). Will the estimated value of cycling VO2 Max fall to approach your running VO2 Max? Personally I would expect cycling VO2 max to just disappear from the graph, the same way it did after my 1 cycling activity with power, but maybe that happened because I only had 1 qualifying cycling activity as opposed to a long history of cycling.

Again it doesn’t matter too much to me as I don’t cycle with a power meter, and I don’t put too much stock in Garmin VO2 Max estimates except for the trends, but I’m curious about how it really works.