VO2 Max data must be inaccurate

Have a look at this,  can give you a pointer. 

www.ntnu.edu/.../vo2max

https://assets.firstbeat.com/firstbeat/uploads/2017/06/white_paper_VO2max_30.6.2017.pdf

Because the Firstbeat method is sub-maximal by nature, it uses an age-based estimated maximum heart rate (HRmax) in the calculation. Therefore, the error in the HRmax estimation affects the accuracy of the VO2max estimate. Figure 5 shows how much the difference between a person's actual and age- based HRmax affects the VO2max estimation error in the mentioned database of 2690 freely performed workouts. If the HRmax is estimated 15 beats/min too low, the error in the VO2max result is about 9%. Respectively, if the HRmax is estimated 15 beats/min too high, the error in VO2max result is 7%. If the person's real HRmax is known, the VO2max assessment error falls to the 5% level.

The running VO2Max method is based on max HR. If for any reason you do not accurately know your max HR, it becomes less accurate.

The Firstbeat VO2max method has been developed against laboratory measured VO2max values (Figure 3) and validated with different exercise modes. The accuracy of the method when applied for running is 95% (Mean absolute percentage error, MAPE ~5%), based on a database of 2690 freely performed runs from 79 runners whose VO2max was tested four times during their 6-9 -month preparation period for a marathon. In a vast majority of the measurements, the error was below 3.5 ml/kg/min and the error was evenly distributed around the mean value. For perspective, the error in a typical indirect submaximal test is 10-15% and in a direct laboratory test about 5%.

On average, the best case of error was +- 3.5 ml/kg/min.

For 50 ml/kg/min as an actual value, on avergave, best case the value on the watch would be between 46 and 54.

If your max HR is a bit wrong, then it is accurate to somewhere between 7-9%, assuming everything else is good. Everything else being good requires your weight to be accurate and pace information must also be accurate. The Firstbeat algorithm was validated in a lab where all of these things could be known with more accuracy and precision than a watch in the field.

Think about this withing 3.5 ml/kg/min best case accuracy for a moment. If your lab running VO2Max is 50 mL/(kg·min) then you could expect values between 46 and 54 from your Garmin (as whole numbers) assuming that you have a correct max HR, weight, and accurate pace.

That's a pretty significant range in the best case scenario. And that also assumes that your running workouts are similar to the type of input that the algorithm expects. It likes steady state, threshold running. Brick workouts and some kinds of definitely clearly confuse it.

It's also clear to me that different models/generations of Garmin watch will yield different VO2Max estimate results for the same athlete. Those values are not comparable for a single individual, let alone comparable longitudinally across athletes.

I've heard people complaining that Garmin does not display the decimal part of the VO2Max estimate. But really it is only accurate to about 1 significant digit. It should really be dislplaying something more in keeping with the uncertainty.

For our best case example:

VO2Max: approx 46 to 54

But if you factor in a bit more uncertainty from max HR, weight, and pace error it's easy to imagine +-10% is optimistic:

VO2Max: approx 45 to 55

This explains why in the real world running mates of similar ability can have wildly different Garmin VO2Max.

Based on their own research, even when working properly, I don't think it is an actionable metric because the uncertainty is high. I don't think it is a particularly useful feature.

R_Tellis said:

I’ve had Garmin watches that had a Vo2Max feature since the FR220 had the values have always been consistent from one model to the next.

Agree on that. My VO2 has always been +-5 and changes pretty much follows my performance and how I feel when running.

silentvoyager said:

F6 VO2MAX isn't accurate is because it depends on accurate measurement of instant pace, which is horribly inaccurate on F6.

VO2 is calculated over a longer period of time which is not impacted by the problem with instant pace. I started to use a Stryd in July and my VO2 calculations are pretty much the same now as in June.

brian.reiter said:

The algorithm is just lame. 

That may be true but anything except a lab test is. If I compare different algorithms and tables for VO2 estimation based on my runs with Garmin's estimation Garmin gives me the highest.

Funny stuff this. I tried the calculator, based on the same parameters i use in the fenix 6 pro, it gave me a VO2max of 52 and a fitness age of 30 years old. Garmin give everything between 43 and 46 in VO2max, but at it’s best a fitness age of 20 years old. Actually i will turn 65 years in march next year. By the way, the edge 830 is the same. Guess a labratory test is required if you want to know.

TobiasLj said:

brian.reiter said:

The algorithm is just lame. 

That may be true but anything except a lab test is.

Yes.

That said, TrainingPeaks Lactate Threshold detection seems to produce legitimately useful estimates. Stryd machine learning determination of Critical Power also seems pretty reasonable if you are consistently running on good surfaces with Stryd. 

Crucially both of these can be used as an input to prescribe workout duration and intensity. 

By contrast the Garmin VO2max estimate is not very useful unless you start to take the Training Status and recovery recommendations et al from the watch seriously. I wouldn’t recommend doing that. 

I’m happy I found this forum to know I’m not the only one. I know I shouldn’t take it seriously but it still hurts my feels when I feel like I’m getting stronger it takes my little points away. 

Also, any other trail runners out there that have found starting a run on an uphill, so slower pace and higher heart rate, really makes my garmin think I’m out of shape. It will tell me my performance condition is down and then take vo2max points away from me. Even running the uphill, it doesn’t care, all it sees is my pace and heart rate. I just want my watch to be proud of me