Smartwatch VO2 Max Accuracy | Science Explains The Truth

[@jamjam-1418]

Hi there! Coming from someone who administers these VO2 max tests on the side and who loves getting into the granular details, I’ve found that current Garmin watches are usually within 5-10% of an accurate reading (given you’ve been training with it for 3-6 months). Which, as you state in the video, is quite extraordinary for a wrist wearable. The important metric here is accurate weight at the time of testing. If you train with your garmin for a full year, that 5-10% shrinks to 0-3%. Quite impressive! Note that Garmin updates their software regularly, so VO2 algorithm accuracy is contingent on a solid software performance - shots fired at you, Enduro 3. PS: Hats off on the 63.5! Well done, sir!

[@jjndz3782]

Would you know why my VO2 max sometimes goes down after a run? E.g. before a run I am on 61 after the run it goes down to 58.

[@dbarn8448]

My results using Garmin Venu & Fenix 6 w/ Polar h10 versus professional VO2Max test ( run about every 6 mo) has them both w/in a couple ml of each other. That is accurate enough to measure my progress via Garmin Connect.

[@oxey_]

yeah i think that's the way you should be using them anyway. The point of stuff like having running power, hrv, vo2 max is to compare with your usual/past values anyway to track trends. Bragging rights over a number like this means nothing

[@iggalan]

On July I got a Coros Pace 3 that I've used for two months straight, running, swimming, cycling. My estimated VO2Max was 47. My Fenix 7 estimates my VO2Max at 53, so quite a difference. Of course I've used the Fenix for years but still. I have done two fitness tests with the Coros, plus high intensity trainings so the Coros should have enough data. Heart rate zones were set exactly the same in the two watches plus everything else, like weight, age, etc. Interestingly enough race predictions on both watches do not differ that much for 5 K, but for a Marathon Coros believes I can do 3h25 minutes and Garmin 3h51, but my best so far is 3h 46m 😅 .

[@marc5279]

i've been using coros for 4 months now and it also seems to me that coros overestimates the running pace for the marathon. I feel like it ignores how differently peoples marathon pace can drift after trying to maintain it for more than 3h, and it just assumes you're going to show almost none instead of calculating each persons individual drift. I never run more than 15k on flat and i can confirm the pace prediction for the 5k and 10k is on point, but i do have almost the same doubts about the 21k prediction

[@boredbootneck81]

Would love to see things like the 12minute run stacked up against the lab. I use it not for accuracy but for a consistent metric thats cheap and easy, but would be curious.

[@teplak4805]

Did a lab test for both running and cycling. Both were exactly the same as my Garmin results.

[@harryholmes1971]

My garmin bike computer told me my vo2 max was around 56, peaked at 59, when I'd entered my max hr as 201, which my hr had gone to on several hard sessions. Then a year or so later I do a DIY max hr test and got up to 199, so entered that as my new max hr. Since then my vo2 has week on week tanked and is now hovering at around 51, taking me out of "superior" for my age group, and almost down to "good". Goes to show how important having the correct info is when using smart devices to calculate these things!

[@GTE_Channel]

Finally someone who explains the different protocols. Personally I do not care about my absolute number, its the trend that interests me.

[@NoelSlevin]

I got lab tested last year - my Garmin measured 58, lab test was 56. The same watch has measured me at 60 for the last couple of months, but my lab test a few weeks ago came out at 63.42. So the watch is within a reasonable margin of error, but obviously the lab test measures the data directly so is going to be much more accurate. On the watch side, it also depends whether you are using a wrist sensor, where the accuracy between watches can vary a lot, of a chest/arm strap which tend to be much better at high intensity. But for VO2 Max, I'd say the exact number doesn't matter a lot, more the trend - whereas with lactate testing the idea is to find your LT1/2 to target them more precisely.

[@durachoks]

I have seen couple vids where people speculate that if you do mostly high intensity than watch will overestimate vo2 max and underestimate it if majority is done at low intensity

[@ross88mac]

Thanks! I've always wondered how accurate it was. Pretty happy with my score!

[@erictravis5995]

One data point doesn’t prove accuracy unfortunately. I would love to see a range of athletes of varying fitness levels all perform this test to see if the results are similar, if the watch tends to overestimate for the less fit and underestimate for highly fit, etc… importance of max HR estimates in the watch’s algorithm, etc… would also be good.

[@rustyheyman214]

My thoughts too. Would be nice if Heather and James did it too. Even if the results and not tests themselves are shown. N3 > N1

[@ChiTownBrownie89]

Same, I did a lab test and got 72, my watch says 60-65

[@Dandegas1-BRFC]

I did a lab test which was 51 and my garmin watch was 49. My watch is now 51 so I guess mines got up a bit.

[@catinapottedplant2324]

They have done an actual study on this and it lines up with this video more or less, around 4% discrepancy (PMID: 35094376). the couple studies I found both said that this accuracy falls off with people who are extremely fit or extremely out of shape. Also, at least a dozen other channels on YT have done the same test (I think GCN even did it too) and got basically the same conclusion as this one. These watches are pretty accurate, accurate enough for the average amateur athlete.

[@brum293]

You also have to factor in that the lab result can also be slightly different depending on lab equipment, environment, test protocol, test day fitness/recovery/mental state etc.

[@cmarnold78]

This is in keeping with friends who have also done lab tests. A variation of 2-3 between lab and watch. Only thing I find a bit suspect about the watch (garmin) is its variability. I have seen it vary from 2-3 from one day to the next. Normally going down, often very dependent on weather conditions which it doesn't seem to take into account eg it will drop you back if you go out in hot/humid conditions then take a couple of weeks to bump you back up. However for long term trends over months/year it is a pretty useful indicator of where you are.

[@andrewmcalister3462]

Mark, one thing I wish you had mentioned was that the accuracy of the watch estimate depends significantly on having a good estimate of your HRMax. Those just using the default 220 minus age are likely to get a less accurate VO2Max estimate. The FirstBeat White Paper on the topic goes into more detail on this. Love the contrast between your suffering on the treadmill, and the scenery where you did the field test. 😅

[@markthrelfall3577]

Good point. They actually all differ in how to calculate it. I can’t talk for other brands, but Coros doesn’t need an estimated HrMax

[@2WheeledMohawk]

My last lab base tests at my uni (Teesside) my lad result and watch result (Garmin Fenix 6) have been within 2 ‘points’. Very impressive considering the watch is estimating

[@angelicdanielle]

Great example of all this Mark. 61 or more accurately 63 is actually very good. Goes to show you can still go deep (and do) from time to time. How about Head units vs watches for cycling

[@walter5334]

Nice video as always Mark! BUT, I am sorry to disappont you: it's actually possible to detect your LT thresholds from your watch data. You just need a chest strap for very accurate HRM. Or a custom external sensor to track additional data. It's just that Garmin and Co. haven't realized it yet 🤫 Don't believe me? Share the watch data from a step test with me and I'll find your thresholds with a reasonable degree of accuracy ;)

[@dragonfractal6361]

Care to share a formula?

[@walter5334]

@dragonfractal6361, thanks for asking! I'll spill the beans, in the hope that athletes may find a straightforward way to track their Thresholds, without having to prick their fingers and buy expensive lactate measuring devices. Assuming that you can accurately track your Respiration Frequency (Rf a.k.a. BrPM) and Heart Rate (HR) during vigorous exercise, here is the short answer: Perform a carefully designed step test, either outdoor or on a treadmill, and plot Rf (Y-axis) against HR (X-axis). You should be able to see two ventilatory thresholds on the plot (VT1 and VT2), which correlate strongly with LT1 and LT2, respectively. I hope this helps. I remain available to help anyone who would like to perform the Respiratory Threshold analysis from their watch data. Keep in mind that although VTs and LTs are strongly correlated, this procedure will need to be validated in the lab by recording both Rf and blood lactate from a cohort of athletes. Some notes: 1. In addition to your watch, you'll need a chest strap HR monitor capable of recording Respiration rates. 2. To extract the data for the Threshold Analysis, log into your Garmin connect, select a running activity (step test), and export data in .csv or .gpx. I'd be happy to provide help with parsing the .gpx data. 3. If you cannot record Rf, I can help you build a custom wearable that accurately record Rf during running. It will cost approx. 30 euros. I will also share the algorithm I developed to call the respiratory peaks, calculate the Rf, and perform the Threshold Analysis. Cheers!

[@jonathanzappala]

A lower guess by the watch is definitely preferable to overestimate. I took three weeks off last December, then injured myself getting back in, my watch number fell away from 65. Now it decides to creep up slowly one point every few months. It’s back up to 64 despite running as good or better. I think maybe it’s an old watch problem that they haven’t pushed an update in 12 months.

[@realstradman]

I'm 62 and having worn it continously for the last 9 months my Garmin Epix Pro 2 has a pretty good record of my runs HR etc. It shows my V02 Max at 52. Did a lab test a week ago on a treadmill after a couple of rest days and that showed a V02 max of 44. Go figure....

[@winklertribe5268]

I would LOVE to see a comparison of multi sport smartwatches! What is the best triathlon watch?

[@usr-bin-gcc]

Performance based VO2 max estimates are bound to have a lot of uncertainty in them because of factors like running efficiency (you will need more oxygen to run a given speed with poor efficiency, so you are likely get an under-estimate of your actual VO2 max). The Cooper test is one of the few where the method is published, so I re-implemented it from the information given in the paper (as closely as could be done), and the 95% prediction interval for my VO2 max is 42.42 - 55.43 mL/min/kg, which is quite a big range! It would be nice if smartwatch manufacturers would give a range as well as a most likely value, but I suspect they might not want to for commercial reasons. However (IMHO) what matters is not the accuracy of the absolute value of the estimate, but the accuracy of *changes* in the estimate over time as that ought to be a better guide to training.

[@richardblom4358]

I am wondering, how accurate is a vo2 max test in the lab? Compared to a watch? (Garmin claims 5%) When you do a lab test, you need to take into account the precision of measurement of the gas analyser and daily fluctuations of your vo2 max (how well rested are you?)

[@molochz]

I've seen some of them say +/- 1.5%

[@PolarBear9733]

At the end of the day, as I understand it, the lab test is a direct measurement and will be your actual VO2 max subject to instrument error. The watch is an indirect estimate using the body parameters you give it and then measuring your HR at given running speeds. These numbers are plugged into and empirical formula to estimate VO2 max. If the watch gets within +/- 5% of the lab, I would view that as excellent!

[@v6i838]

My Garmin currently estimates mine at 62. I think mine is actually anywhere from 59-63. Long aerobic runs seem to helped mine go up

[@robertcioata4034]

My watch says 65, and my lab tested score was 66, so not too far off, but I seem to generally fit the population average for estimating things like max hr. Lactate threshold and lt1 hr were perfect between the lab and watch which is what I think is most important since I base my heart rate zones around them.

[@ChrisCapoccia]

Assuming a basic VO2Max test is $100, it's not something regular people would do more than once a year. How does knowing the number with greater precision change anything about the training plan?

[@Nehpets1701G]

Pair it with blood lactate testing and you can get yourself a very precise idea of your training zones. Loads of people in cycling are training with an ftp and training zone number which are set too high, and fatiguing themselves too much during VO2 Max intervals. Phil @ bike racing without mercy, here on KZbin did a few videos about it and his subsequent performance increased a fair bit because of it.

[@ironman140.6]

EXCELLENT MARK!

[@fiddleronthebike]

as many others have found that too (most data I know are from Garmin watches, but in most cases they were only +-1 off) I think for us amateurs these watches are accurate enough

[@HolisticHealthEducation]

I would be interested to see the difference the wearable indicate in VO2 max, if for six weeks you only do slow long runs, and for the next six weeks you only do short max effort fast runs. My wearable, which I won’t name as to not embarrass the manufacturer, shows significant changes in results.

[@DJ_Gallow]

Very nice intro

[@Adgum1]

eyyyyy. straight into it

[@99cya]

What about the polar fitness test where you dont even run to calculate vo2max? How accurate is that?

[@Allride_]

Hm, let's see... 🤔 who's the current GTN sponsor? Coros. You're asking for testing of a Polar watch... 🤔 So the correct answer is: I guess we'll never know

[@99cya]

@@Allride_ thats a pity. Would have loved to know about polar. Polar is giving me 60 vo2max. Dont want to spend the money for a lab test on this.

[@wemakerobotsai]

A test / analysis I would love to see is MHR training zones versus HRR training zones. I have used both and find HRR zones to be a far better calculation. I like most things about the Whoop but the fact that it only offers MHR calculated zones is really frustrating… and somewhat deceptive considering how much they push the performance measurement personalisation message. Hence I’m cancelling my subscription. I don’t need to spend £35 pm on a random number generator.

[@abernardes2]

Well, I did recently The Great North Run with a carbon plate shoe and was overtaken by a guy wearing clock shoes! He finished in 1:47 min and I did in 2:02. Compare then both would make a great video 😂😂😂

[@questgivercyradis8462]

Very nice test!! I am curious what the error margins are, and what direction they skew. Seems like watches are good for ball-park estimates, and tracking self improvement trends, which is excellent! But scientist in me has more questions. Are watches usually under-estimating, or over-estimating? And, since I am a cyclist and noticed on my Garmin that there is a Running VO2max and a Cycling VO2max, what contributes to differences in these? Seems like something the triathletes may have an idea on. I'm up to Garmin-measured 50 running, 52 cycling, after being a super out of breath weakling 5 years ago. I run on rare occasion, and am going to add a few more ~5k runs to my fitness schedule for the sake of variety, and comfort at higher heart rates (right now a 5k run on flat is harder for my legs than a 100mi ride through mountains). I'm curious if the VO2 max difference is a matter of familiarity with the sport, or a regular thing regarding ability to breathe effectively depending on sport.

[@darrenc.3263]

Would have loved this when more with even just a couple more people and data sets.

[@jonasvieth]

Forget the Crocs, I want to see Mark with his Jesus sandals in the lab 😄

[@jamesfussey9651]

Don’t know if you’ve already done a video on it or not but the difference between a disc wheel cover compared to a disc wheel. A full disc wheel is way out my budget but a cover is only £200 just wondering the difference between them

[@him050]

I have a gen 2 epix pro and mine thinks my VO2 max is 53 and predicts that I can run a 5K in 20:30. My Parkrun PB (which admittedly has a bit of a hill) is 22:53 and it damn near killed me. So I have no idea where it’s pulling these numbers from. I’ve only been running a year with my only exercise before that being weightlifting, yet it says I have a superior VO2 max?! How can that be possible?

[@realstradman]

Yup I'm 62 y.o and got identical watch and my Vo2Max is 52 and it tells me I should be able to run 5k 21:23 but I highly doubt this without requiring a defibrillator at the end of the race....

[@francoisrouviere]

Wondering if the cycling VO2 max calculations would be considered to be similarly as accurate. My cycling VO2 max has been declining in the last couple of months yet my running VO2 max has been increasing. Have not made any changes to my training.

[@amyennion6196]

Maybe one for coaches corner, but does Bath's lab have any view on the applicability of a watches VO2 max reading for swimming? I'm an ultra marathon swimmer, and I've always noticed that my heartrate zones are CONSIDERABLY lower in the water than they are for running (and also very different OW/cold water swimming vs pool swimming) so my question is: is there also a predicted difference in VO2 max? Or should the physiological basis be the same, and so the heart rate readings I can take (with a chest monitor) while swimming should also give me a ballpark-accurate VO2 max, specifically for swimming? Thanks in advance!

[@PeterFairhurst-dl3hj]

How do the predicted race times stack up against real performances ?

[@bikeanddogtripsvirtualcycling]

looking at historical data on my garmin account I can see that my cycling vo2 max has gone from 51 back in Feb to 64 as of right now (I am 51 years old). What I have noticed over the past few months is that if I over train for a week or so (I get the "overreaching" message from Garmin) then I will lose a point or 2 from the vo2, then after a few easy sessions and say a serious effort hill climb or a race I will gain those points back plus an extra point or so. strangely though, my general vo2 max for running and general fitness shows currently as 61 . I'm certainly no athlete .

[@overcookit1433]

How about the VO2 max-calculations of the bike computers, e.g. the Garmin Edge 530? Are they also pretty closed to the results of lab tests? In the last weeks, my Egde 530 commonly saw my training rides being good for VO2 max, with a recommended resting period of 72 hours, only once or twice it said I should rest for 4 days.

[@otifaf]

My VO2 max for cycling is much higher than for running. I'm also curious about the accuracy of the Edge 530. Those recommended rest times after workouts are hilarious-I get them too and completely ignore them because I need to keep up with the training program.

[@overcookit1433]

@@otifaf I also ignore them, otherwise I could only ride every three days. I suppose the Edge takes into account my age of 51. It were great if one could add Weightlifting to the activities, now the device thinks I did nothing when I did not ride, but in the days between the rides (monday, wednesday and friday) I train with a legpress. The older you get, the more important becomes the power training.

[@jerryyoung6494]

Any ideas if newer watches are better than older ones? I’m hoping to justify a new purchase maybe

[@dri1811ya]

Algorithm is probably identical - but HR sensor may have improved, giving better data & approximation. If you use a chest strap for every run (who does, seriously) then it's moot. I upgraded from FR245 to 955 over Christmas, and my VO2max estimate dropped from 54 to 49 😅

[@kayy-03]

Likely not by a lot

[@squngy0]

Depends on how old your watch is. If it has power data already, then it is probably not going to be very different. Also, though you might not like to hear this, but it doesn't really matter how close to the lab the watch VO2Max is. It just matters that it shows your progress (or lack there of)

[@LiberaTeTutemetExInferis]

I wish you had been wearing the watch during the test...same effort

[@SueWilson-i6w]

Brown Causeway

[@deanoboxing83]

My garmin says 59 but my lab test last week was 69

[@Tritiuminducedfusion]

Why is my Apple VO2 Max 4.7 points higher than Garmin?

[@otifaf]

Because the Apple Watch sucks as a sports watch.

[@sebastianfischer2520]

Mine was 10 lower in the lab :yikes:

[@oxey_]

keep in mind that might also be due to your fatigue/comfort and performance on the day, you might have simply not been at your best

[@realstradman]

Yup. Similar experience to me. it was 8 lower in the lab

[@nickmacaluso6585]

You must have insane running economy, is what that tells me.

[@nickjacobsss]

My apple watch says im far below average, but still bike 200 miles a week? makes sense.

[@sigurdbr08]

My official vo2max is 67.5, but my Garmin says 57

[@stevecollier21]

Correct me if I'm wrong, but you didn't disclose in the verbal commentary or the video that you are sponsored by Coros

[@philipeick-vocalmusic]

My watch estimated 55 while I got 63 in the lab 🤷‍♂

[@patchworks4102]

I got the same! 55 to 62.7. Garmin?

[@philipeick-vocalmusic]

@@patchworks4102 Yes 😅

[@MrMars121]

Same here. 56 Lab, 48 Garmin

[@oa1021]

my apple watch gives me 58 (running) and garmin 68 (cycling) anyone have any ideas

[@zackbelcher3896]

Watch says 50, lab tested at 60 for me I’m an underperformer

[@jasonmccullum9424]

my watches are both useless for VO2 Max.. My apple watch says 30.1 my garmin says 41.... what gives

[@pal_lokomotivet2679]

It’s all statistical average. It has no chance to measure it. I’m a heavy cyclist that measure 60 in lab when the watch gave me 52. Because when I run my efficiency is very low. I’m not getting good speed from the legs despite a ok vo2. So Garmin assumes that I have a much lower due to my running speed. Both measured running. If it’s right for you it’s only because you fall into the database averages Garmin use.

[@brum293]

Why not measure it on a bike if you’re a cyclist?

[@ianclunie9753]

Watch might be right. Your VO2 will be different for different sports because you might be an efficient cyclist but not a great runner. If you use a power meter, most Garmin devices will be able to estimate your cycling VO2, which should be considerably higher than your running VO2 max.

[@julekogurek]

Got tested at 59 when my garmin was giving me 57

[@BrowneNorton-h5u]

Moore Edward Thomas Jose Harris Eric

[@MMichiganSalveRegina]

Relative VO2 Max (which is what this is) is definitely comparable across people. That's a stupid comment

[@IanButterworthyyc]

Sponsored by Corus - you forgot to mention. I bet if the lab results didn’t come close, this video wouldn’t have been posted? N=1

[@IanButterworthyyc]

Sponsored by Corus - you forgot to mention. I bet if the lab results didn’t come close, this video wouldn’t have been posted?

[@keepleft99]

do a sweat test!

[@JaegerDives]

My Garmin is out of its mind. It says my VO2max is 50 and that I have the fitness level of a 20 year old. I'm 60 and am only running 40km (25mi) a week. In actuality I have the fitness level of a 60 year old 😢

[@stennan]

But Garmins estimation is meant to give you happiness, and make sure you buy a Garmin watch next time 😉

[@playbandcflat9829]

Ah, so you were that result of the 1964 experiment that actually produced the supreme human. I was the other that well, failed.

[@therealsvensson]

@jaegerdives his is a well-known Garmin bug that puts you at 20 years fitness age. Mostly when you don’t have enough data for the algo to work on. Your fitness age can never be more than 10 lower than your actual age. Part of their algo and well-described all over the typical forums. You’re still fit tho, owning a Garmin already makes you fitter than 90% of the western world. 😂

[@christopherbornman6643]

@@therealsvenssoncant they fix this bug. It’s so irritating. No matter how fit or unfit I get I remain an excellent 20 year old.

[@usr-bin-gcc]

@@therealsvensson " Your fitness age can never be more than 10 lower than your actual age. " I think that is only for newer Garmin watches. My forerunner 245 gives me the fitness age of 20. It isn't a bug, it is just that my VO2 max is about the average for someone of age 20. AIUI the reason they changed it is that VO2 max is only one aspect of "fitness age". @JaegerDives can still say he has the VO2 max of a 20 year old, because it is true - I hope I can still say that when I am 60!

[@stuarthammond2030]

Watches are a waste of money. Calories burned and sleep are way out!!!

[@brum293]

@@stuarthammond2030 If you think it’s not worth it then just don’t buy a watch. lol 😂 For me it’s very valuable because I can track my training.