Traveling along the trails, sidewalks, and main streets of the towns they reside in, runners, cyclists, and endurance sports athletes everywhere all know a familiar sound. The delightfully gratifying chirp of a fitness tracker as you complete your next mile, achieve a new PR (personal record), or record a new VO2max.
Ever since I entered the world of endurance sports training eight years ago, I’ve heard athletes talking about their VO2 max, how to improve it, and how accurate (or not?) fitness trackers are at actually measuring these values.
I decided to explore the technology of Garmin fitness watches to understand how VO2max is calculated and do a baseline comparison of how these wearable technologies VO2max predictions compare to laboratory testing.
Firstbeat Technology’s Fitness Test is used by Garmin and other fitness companies to calculate VO2max for a variety of different activities. Described in patent US20110040193A1, this Fitness Test calculates users’ VO2 in the following steps:
1) The personal background info (at least age) is logged
2) The person starts to exercise with a device that measures heart rate and speed
3) The activity collected data is segmented to different heart rate ranges based off the persons background info and the reliability of different data segments is calculated(reliability is measured based off how continuous the activity is- uninterrupted segments are better than those where the user has to stop)
4) The most reliable data segments are used for estimating the person’s aerobic fitness level (VO2max) by utilizing the person’s heart rate and speed data
Speed data from reliable segments are used to calculate a VO2, oxygen consumption, during that segment. 20-30s bouts are used to calculate VO2 across segments using one of the following theoretical VO2 calculations:
Walking and Pole Walking: Theoretical VO2 (ml/kg/min)=1.78*speed*16.67[tan(inclination)+0.073]
Running on a Level Ground: Theoretical VO2 (ml/kg/min)=3.5 speed
Running in a Hilly Terrain: Theoretical VO2 (ml/kg/min)=3.33*speed+15*tan(inclination)*speed+3.5
Cycling: Theoretical VO2 (ml/kg/min)=(12.35*Power+300)/person’s weight
Rowing (Indoor): Theoretical VO2 (ml/kg/min)=(14.72*Power+250.39)/person’s weight Unit of speed=kilometers per hour (km/h)
Unit of inclination=degrees)(°)
Unit of power=watts (W)
Unit of weight=kilograms (kg)
From these calculated theoretical VO2 values, heart rate information is used to determine effort of segments. Heart rate zones based on user information are utilized to evaluate effort, and then effort is used to determine that VO2 as a %VO2max. VO2max estimates are made for each segment using %VO2max. These segment VO2max can be weighted based off heart beat derived parameters and performance parameters, and then used to calculate VO2max.
An affordable mode of tracking your VO2max through measuring heart rate and speed data – pretty neat, right? But how accurate is this technology and how does it match up to laboratory testing?
Firstbeat conducted their own study to validate the technology and its effectiveness at estimating VO2max. They found that “[t]he accuracy of the method when applied for running is 95% (Mean absolute percentageerror, 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”(4). Error in estimated VO2max was less 3.5ml/kg/min in most cases, which is fairly accurate considering most submaximal testing has an error of 10-15%. Method accuracy varied with respect to estimated maximum heart rate(HRmax). ” 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”(5). This study suggests a high degree of accuracy for Firstbeat’s fitness test technology in predicting VO2max.
A group of scientists at Southern Illinois University Edwardsville evaluated the wearable technology’s accuracy by conducting a laboratory VO2max test on male and female runners, then allowing participants to use the wearable technology to calculate VO2max in a 10 minute self guided run. They found that the Garmin Forerunner 230MAX and 235MAX measured VO2max within -0.3 ± 3.4 ml/kg/min, p=0.02 for the 230MAX and -1.1 ± 4.0 ml/kg/min, p=0.026 for the 235MAX for female runners, and -1.1 ± 3.4 ml/kg/min, p=0.149 for the 230MAX and -3.2 ± 4.2 ml/kg/min, p=0.002 for the 235MAX for male runners. There is a greater amount of variability in the male group; however, this could be due to miscalculations in HRmax and potential variations in levels of effort in participant during the 10 minute self guided run. Although there is greater variability within the male group, the devices still appear fairly accurate at predicting VO2max.
Wearable conducted an evaluation of their own putting fitness watches to the test – assessing the accuracy of Garmin, Fitbit, and Jabra devices in measuring VO2max. They found that Garmin technology provided a VO2max estimation within 0.3 ml/kg/min of their study participant, which was the most accurate of all devices tested. The high degree of accuracy found in their study remains consistent with other larger scientific studies.
Across the board, there appears to be a high degree of accuracy with Firstbeat’s Fitness Test in estimating VO2max. For endurance athletes everywhere, this is a huge sigh of relief. Rather than partaking in expensive, strenuous VO2max testing, we can monitor our progress utilizing the technology in the watches we wear everyday. In addition to watching our paces, heart rates, and overall progress, we can also monitor our cardiovascular health and athletic progress as we continue to train and push ourselves everyday.
Seppanen, M., Pulkkenin, A., Kurunmaki, V., Saalasti, S., & Kettunen, J. (2016). U.S. Patent No. US20110040193A1. Washington, DC: U.S. Patent and Trademark Office
 Firstbeat Technology(2014). Automated Fitness Level (VO2max) Estimation with Heart Rate and Speed Data.
Snyder, N. C. , Willoughby, C. A. & Smith, B. K. (2017). Accuracy of Garmin and Polar Smart Watches to Predict VO2max. Medicine & Science in Sports & Exercise, 49(5S), 761. doi: 10.1249/01.mss.0000519024.10358.0b.
Stables, J., & Stables, J. (2016, December 21). The big VO2 Max test: Fitbit, Garmin and Jabra go head-to-head. Retrieved from https://www.wareable.com/running/best-vo2-max-devices-tested-9129
I did not know Garmin watches even gave VO2 max estimates, let alone accurate ones! The level of accuracy compared to laboratory testing is pretty impressive, especially considering these are relatively cheap, simple wearable technologies. I was interested in participating in the VO2 max testing we saw at STAR Campus, but I think wearing a Garmin watch during a less strenuous run sounds a bit more appealing. Plus, being able to track VO2 max, along with various other exercise-related physiological measurements over time wouldn’t be too bad either. Might have to invest in one.
Wow that is really impressive to see that Garmin has such accurate estimates of VO2max. I bought a few knock off Fitbits for some experiments in my lab, then did my own pilot testing to compare them to Fitbit. They were mostly accurate in steps, but really struggled with heart rate (luckily we want to track daily activity with steps). If I extend my work to have an exercise group, I might have to have my lab invest in a Garmin. Plus as Ben said, I’d like to get some idea of my VO2max, so maybe I’ll borrow it from the lab for a few days.
In a controlled environment I imagine that the device is accurate however when adding variables like weather and added weight during my walking I have noticed while my strength and endurance are increasing my vo2 max has been decreasing at an alarming “ you should go to the doctor” rate.
My resting heart rate has decreased over time which shows how my cardiovascular system is becoming more and more healthy over time but my vo2max calculation says that I am becoming less and less healthy with each subsequent journey of more than 3 miles a day
I am eager to see if, as the weather starts to cool, my readings get more accurate. Otherwise I’m going to the cardiologist because the healthier my body becomes the worse my heart is. In my humble opinion until you can add additional variables such as weather (when exercising outdoors) and added weight (I walk with 2 4lb D weights ) the vo2 max feature on the Garmin vivosmart 4 is flawed during the summer for sure and quite possibly the winter in extreme cold conditions. This means I can only trust it in spring and fall which makes it of no use more than 1/2 the year
I have experienced exactly the same during a period of doing more cardio training. Reduction from 47 to 38 during good cardio excersize and experiencing a better condition. I can see if I run in a more hilly terrain that the vo2max plummets rapidly in the stats. I guess the system is sensitive to variation in mode of excersize an possibly also temp and other conditions
that is an interesting observation and I wonder where you are now with it, because since I started running again (due to COVID), my resting heart rate has decreased and my vo2max has increased quite a bit. Have you gotten any more information?
I noticed that most of my VO2 max last year was 58 and it plummeted 1 point when I push hard my training run (with anaerobic capacity training effect per Garmin) my ave HR goes up as well. Recently I did an easy run with a lower ave HR and surprisingly my VO2 max increased by 1 point to 59. I’m currently using Garmin Fēnix 6 Pro.
Could be a rounding error or just expected variation. Claims to be accurate within 5%, 1 point out of 59 is less than 2%.
Have you updated the weight and HR personal data in your watch. If your HRR has changed may be that’s he only issue.
I have exactly the same experience: gained some weight, went off training for a while, and restarted running – and now, even though my perceived fitness is up since I restarted, my VO2 max has fallen off a cliff.
I would love to see some research on wearable devices accuracy on calories expended!
For some time now I’ve thought that a good measure of exertion per unit time would be the count of heart beats during that interval, rather than the number of steps. Like carrying a 25-pound load over a given distance compared to carrying nothing: Same step count, but a different level of exertion.
The app already can count heart pulses, so this seems an easy thing to program, I would think.
Pingback: Can you trust the data on your fitbit? Here is the latest on Fitbit accuracy
Will recording a walk (obviously slower than running) as a run lower the estimated VO2max?
I have several different Garmin devices, wear a small activity tracker everyday, 635 with HR strap for running and use a cycling computer on my mountain bike. I get that VO2 max running vs cycling will be different. but for running both activity tracker and the 635 differ running V02 max 41 vs 48. Really not sure which number to trust. Like the other comments doing intervals and training harder is actually decreasing V02 Max
pasted from Garmin’s site “If you are an active cyclist and runner, you will notice that you have one VO2 max for cycling and a similar but slightly different VO2 max value for running. This activity type-based variation is normal in both direct and indirect VO2 max measurements and is simply the result of different muscle groups being used and sport-specific training adaptations.” You can find the full info here https://www.garmin.com/en-US/blog/fitness/v02-max-and-aerobic-performance/
I run and cycle but sadly I get no VO2max from my ride activities just because my bike does not have a power meter. It’s not good enough to have a Garmin Enduro on your wrist, calculating VO2max requires another expensive computer mounted on the bike.
My vivosmart 4 shows a VO2Max of 42 while my Forerunner 235 shows VO2Max=50. I wonder if this is due to the vivosmart using the formula for flat terrain (even when connected to my phone’s GPS) while the Forerunner 235 might use the formula for hilly terrain. I did not expect the difference to be so big though.
My Garmin Vivoactive tells me I have a VO2 Max of 31. Which = 75 years old. I am only 65 but in the last eight days I have done 3 x 5K runs, 64 lengths in a pool freestyle and a bit of walking. Including calibrating the Vor max, which they say takes 15 minutes but mine did in 5min.
Do I need see a doctor?
Need to do more training with more load….
Does the Garmin calculation include correction for the altitude that the exercise is performed? I have seen a paper that found VO2 max decreasing 6.3% per 1000 meters elevation, linearly over the range 0 – 2800 M. Since I live at 7400 ft (2255 M), My VO2 max of 34 would increase significantly to 39 (age 70), if an elevation correction needed to be applied.
Thanks for this. I was wondering why the cycling v02 estimate is total shit on my watch. But considering it uses power to calculate cycling v02 and my watch rarely connects to the power meter that would explain it.
Good article, really clears it up for me, particularly providing the formulas.
One interesting personal observation is that a couple of rest weeks (lower volume not no volume) have increased my vo2max. This after a 3w training block at increased volume (40-45 km/wk which is new for me).
I was actually disappointed that my vo2max didnt increase from all that training, but perhaps fatigue was to blame, or my body hadnt had a chance to adapt. Two weeks on I can feel my easy loops feel easier and lo and behold, Garmin agrees with an extra point of vo2ma: brilliant!
Interestingly this is at odds with Strava’s fitness measure (which I think is an exp wted average of HR-adjusted training effort). Mind you, if I combine fitness and freshness in Strava I would see that I should be at my fittest 🙂
Wrist HR monitoring can be problematic if your activity involves using your arm. For example, road cycling is fine, mountain biking (especially on rough trails) can give false readings. That’s because the contact with the wrist gets lost. I have noticed this with rowing vs running.
I have also found that the highest (accurate) HR readings come from the activities that you do the most and are best at.
Hi, I’m having similar issues with VO2 on my vivoactive 3 watch.
I’ve had it for 13 months and initially my resting HR was 58, occassionally going up to 62 and V02 gave me an age range of 64…which lowered to 63 (I’m 70) .
Just lately, past 10 days or so my resting HR has dropped to 53 and my age range is given as 73.
I’m wondering if there could be a glitch in the watch.
Has anyone else had results like these?
Garmin don’t use wind or climbing to calculate Vo2max !
So it’s not true !
For example, i’m 56.
If i run flat around 130 bpm, without wind, garmin says my Vo2max is 47.
If i run flat around 130 bpm with some wind, garmin says my Vo2max is 45.
And if i run in hill, still 130 bpm, garmin says my Vo2max is 43…
Garmin Vo2max calculator is a bulls…..
I am 76 years old, 72 kilos, 1.88 m tall and enjoy my unspectacular running (circa 31 minutes for 5k). My Garmin 245 shows my VO2max as 45 and my fitness age as 20!!! Go figure 😏
I’m 73, and my Garmin Fenix 5X shows my VO2 Max as 32, in the top 45% of my age group. It goes on the add that my fitness age is the same as an excellent 20 year-old. I suspect these two are in conflict.
MAYBE it is time to throw away the watches and just enjoy running?
My forerunner stated a vo2 of 42 while walking[ gym shut] now back at the gym and working harder the watches says 40 when running and on line it says 39. I am 65, run at 8mph
The life battery well ran 80 minutes at 5, not sure this helpful. also the other day watch had gained 6 minutes