How does Garmin compute VO2Max?

I think that many of the discrepancies are due to the fact that the different concepts of VDOT and VO2max are completely confused here.

VDOT is simply a method to predict the possible results at other distances from known results. It was specially developed after it was found by Jack Daniels that runners with the same VO2max still achieve different running results.
( www.coacheseducation.com/.../jack-daniels-nov-00.php )

VO2max is the maximum amount of oxygen that can be absorbed by the body during maximum exertion. The value is a general fitness indicator, but is difficult to use for race time prediction because differences in running efficiency, biomechanics and willpower are not taken into account.

Both values are not interchangeable. Here is an example of a range in which both values can be determined for easy comparison:

Take the Cooper test (Formula: (distance in meters - 504.9) ÷ 44.73):
- If you run 3 km in 12 minutes you will get a VO2max of 55.8.
- If you run 3,219 km (2 miles) in 12 minutes, you will get a VO2max of 60.7.

Using Jack Daniels' VDOT tables (according to the URL above):
- For 3 km in 12 minutes you get a VDOT of 47.9.
- For 2 miles in 12 minutes you get a VDOT of 52.1.

For both distances the VDOT values are only about 85 % of the VO2max values!

The Firstbeat-Whitepaper linked above by HermanB adopts the VDOT values one to one and titles them "Race time prediction based on VO2max" (Table 4 on page 7).
Possibly this confusion of VDOT and VO2max results in the overly optimistic race time prediction. It seems the watches determinate VO2max and then look up with this value in the VDOT table.

---

Runalyze does not estimate VO2max. Their value is more similar to VDOT than to VO2max and is therefore called "Effective VO₂max" (probably because the label "VDOT" is legally owned by Jack Daniels).

th3oretiker_ thanks for the clarification. I've heard this argument on the forums before, as well, but never in so much detail. IIRC someone else said that FirstBeat doesn't measure VO2Max, it estimates VDOT.

So would you agree that it's possible that the FirstBeat method is just estimating/calculating VDOT? Runalyze calculates VDOT based on HR and speed, and FirstBeat calculates ????? based on HR and speed, according to their own whitepaper.

If so, then weight doesn't matter for VDOT (since it's all about speed), and it shouldn't factor into FirstBeat's calculations.

And do you think the ACSM formula calculates relative VO2Max (as claimed), or VDOT?

I found this interesting and relevant:

https://www.letsrun.com/forum/flat_read.php?thread=3802725

"Do not equate VDOT to VO2max. It is NOT VO2max. Also keep in mind VO2max is actually a fairly poor indicator of race times among fit people. Many many factors go into racing, from lactate, to cardiac output, to psychological abilities to push one's self, to stride mechanics, to carbohydrate storage and fat utilization, etc, etc, etc, etc. You get the point. "

...

"VO2max is a physiological measurement, essentially based on your maximal oxygen uptake and your body weight.

VDOT is a measure of how fast you are, it doesn't care how much you weigh or how much oxygen you take in.

Stop acting like the two should magically match up somehow. It's like trying to compare your vertical leap to your blood pressure."

None of these VO2Max (or is it VDOT) estimation methods involve strapping a mask to your face and measuring oxygen uptake. What they actually do sounds to me like estimating VDOT from pace and HR, which should have nothing to do with how much you weigh.

I think Firstbeat calculates a value that should correspond to the VO2max lab result.
As for my understanding, they use speed and apply a few formulas to take error influences and sub-max HR into account so that a value comes out that is statistically close to the measured VO2max lab value for the majority of trained users. I don't know if weight is part of this. Possibly you get a better approximation if you enter your correct weight. If that is so, perhaps even more if weight differs much from average.

I don't know enough about the ACSM formula to evaluate it.

You wrote that you found that the Garmin-reported VO2Max is about 3 units too high fpr you. Do you have a lab result or do you compare it with other "VO2max"-estimators? For me, the VO2max value of my 935 fits quite well to my VO2max lab result. Only the race time predictions on the 935 are much too optimistic, possibly by using an unsuitable table. Runalyzes "effective VO2max" on the other hand fits quiet well to my VDOT, which gives useful race predictions.

I based the “3 units too high” statement on the Garmin race predictor lookup table, as crowd sourced on these forums a few years ago. Others have also said 3-4 too high, iirc. So it could be that the Garmin table is just too optimistic.

It’s worth noting that this table is totally different than what runalyze refers to as its “VO2Max values with equivalent race results” table (which it uses to predict race results), which goes back to what you’re saying about FirstBeat and runalyze not calculating the same things. 

I found this interesting: https://forum.runalyze.com/viewtopic.php?f=35&t=1729 (Runalyze vs Garmin VO2 max?)

There's a bunch of stuff about VDOT and VO2Max, as well as an assertion that Garmin was plugging VO2Max into the Jack Daniels VDOT table and getting a result that was too optimistic.

Then again, here's the VDOT table from Jack Daniels himself: https://www.runnersworld.com/advanced/a20825580/threshold-training-finding-your-vdot/

The Jack Daniels VDOT table and the runalyze table are exactly the same (which makes sense given that runalyze calculates VDOT), but the Garmin table is completely different, which at least points to the fact that Firstbeat doesn't calculate VDOT, but "actual" VO2Max.

---

Thanks for all the food for thought. I still find Lactate Threshold to be the best Garmin predictor of performance for me, although I’m really curious how well the new FirstBeat race predictor works.

The crowd sourced table however does not correspond to the table in the Firstbeat whitepaper (the original but mislabeled VDOT table), which makes this even more complicated. Now as you mention it, my "race predictions" are also still far too optimistic, but a little less than if I just looked up my VO2max in the VDOT table. Maybe the watch results are _values_ based on the VDOT table, but altered with some user specific factors like age and weight.

As far as I know, there is no table where you can use your VO2max and get suitable race predictions. This does not mean that your estimated VO2max value is wrong (compared to the lab result). It is simply the wrong concept for this purpose. For this you need VDOT.

Yes, the last posting there (Runalyze forum) is also from me. ;) I had hoped to get some more input to better understand the topic, but so far there have been no reactions.

th3oretiker_ said:

The crowd sourced table however does not correspond to the table in the Firstbeat whitepaper (the original but mislabeled VDOT table), which makes this even more complicated.

Yes, the table in the FirstBeat whitepaper does match the "standard" Jack Daniels VDOT table (within a few seconds). The few seconds' discrepancy can be explained by the fact that Daniels has updated the table over the years, as more running data has been gathered.

So there's really only two tables:

1) Jack Daniels VDOT/"VO2Max" race predictor tables: runalyze, FirstBeat white paper, and online race predictors

2) Garmin race predictor table

Maybe 2) was just supposed to be adjusted for whatever error/skew was perceived to be built-in to the FirstBeat algorithm. I dunno. Or maybe it's supposed to account for some assumed relationship between "real VO2Max" and VDOT.

th3oretiker_ said:

Maybe the watch results are _values_ based on the VDOT table, but altered with some user specific factors like age and weight.

I don't think so, as my 630 and 935 race predictions have matched the crowd-sourced table exactly, for the past 3 years or so. If we were to look at the old forum thread that was the source of that table, I'm sure we'd find absolutely no deviations/contradictions from what's posted on that blog (meaning that age, gender and weight did not affect he results one bit.)

I'm fairly certain that the old Garmin race predictor tables have always been a straight (Garmin) VO2Max to Race Result lookup. That crowd-sourced blog post would be pretty useless otherwise.

th3oretiker_ said:

As far as I know, there is no table where you can use your VO2max and get suitable race predictions. This does not mean that your estimated VO2max value is wrong (compared to the lab result). It is simply the wrong concept for this purpose. For this you need VDOT.

Which sort of brings us full circle. Why wouldn't FirstBeat calculate VDOT if VDOT is the metric that everyone cares about (since it correlates to race predictions)? Who cares about VO2Max (relative or absolute) as some abstract metric of fitness? I mean it's an interesting concept, but everyone just wants to know how fast they can run....

Thanks again for all the context and nuance! I've actually read (or skimmed) the Jack Daniels article, the various runalyze resources, and articles all over the net several times in the past in trying to figure out the exact correlation between "FirstBeat VO2Max" and "online VO2Max/VDOT-based race predictors". I always had this vague idea that VO2Max and VDOT are not the same, but it doesn't help that most online sources treat them as equivalent or similar.

Maybe it really just is FirstBeat secret sauce that isn't meant to be understood. Maybe I just need to buy a new Garmin and see if the new race predictor is awesome :p. (Nah, I have a ton of training to do, first.)

Your heart rate max can get higher in heat an humidity my heart rate could reach 178 at 3.30 pace running 35.00minutes at the end of a 10 kilometre run in 33 degree Celsius and 65 percent relative humidity. Down south I would run about 34 minutes, 3.24 and hit a near heart rate max of 172 at the end of a 10 kilometre run. The problem is your heart rate can spike higher when you are hot a dehydrated due to a lower stroke volume and arterial pressure which means the heart becomes less efficient and beats faster.