58 thoughts on “Garmin Running Power – An Explainer [2022]”

1. Will a watch that is worn fairly loose affect the accuracy/consistency of power calculated from a watch in a significant way?

   1. quite possibly yes.

      1. That is my thought as well, and is one of the reasons why I use Stryd (apart from better distance and speed in general) even though I’ve a Polar Vantage V. For similar reasons I use Polar H10 and not the built-in optical, and my Vantage V is generally only worn during training.

         But Stryd also has some inconsistencies (from run to run) due to use of different shoes, so nothing is perfect.

         That said, I can see why many would opt for a single solution in a watch.

2. Good summary and refresher, thanks! I used RP on my FR935 out of curiosity and found it rather useless. Moved on to Stryd and was impressed by the ecosystem, the CP curve especially, but gave up on it after seeing how variable the results could be from one run to the other and from one shoe to the other, +/- 2%. Doesn’t sound like much but it’s huge when you’re pacing a marathon! Back to pace and HR!

   1. 2% variation in distance or power?

      1. As you point out “running power” cannot be related to anything “real”, distance can and in my testing (a 1km officially measure straight line) I found that distance could vary by +/- 2% even with the same shoes. Stryd deny that it’s possible of course and go after anyone who finds any variation of more than 1% (pushed to their limit they will finally admit that their “can” be 1% variations which is “still much better than a bike power meter costing up to Yx more”).

         In my experience (and despite what they claim too) “Stryd power” IS dependent on pace/distance, so distance variations will have an impact on power and there could be other factors of course for unexplained variations but that’s probably the main one.

         In spite of all the aggravation and insulting attitude by “Gus from Stryd” I tried to make it work for a year, until a 10k when the distance was 4% short and I tried to hold a power (based on my CP) that was way too high. The last Km was…ugly obviously ! Stryd never acknowledged that there could be something wrong with my device (“it’s impossible”, “user error”, “course not measured properly”, blah, blah, blah) and never even opened the door to sending a replacement device. Pretty much says all you need to know about them.

      2. if you auto-calibrate with a garmin it will use GPS (well, garmin use GPS for the stryd autocalibration factor…stryd themselves don’t use it)
         if you manually calibrate it, that’s it.
         the distance you run will differ from the measured distance eg if you run on the right-side line of lane 1 around a 400m track you run about 1% further.
         saying that, a 2-4% difference sounds too much.

         please (Will !) correct the above if wrong. i think it’s correct

      3. Yes manual calibration of course, many, many adjustments done over the course of a year, on a track, on a measured road, etc…

         It seems I was not clear but THE issue is that you can nail down the calibration perfectly one day and then one day out of the blue you can have a +/- 2% variation, even with the same shoes.

         So yes 99% of users will take Stryd’s word for it but when I’m told something is accurate I’m going to be checking that claim especially after repeatedly noticing that autolap points varied on my usual runs.

         The main Stryd FB group has few to no complaints because Gus just moderates them (you get a PM telling you to contact support because you’re likely doing something wrong). There’s an independent FB group where a guy who’s even more OCD keeps running accuracy tests with multiple units and and he’s also finding differences, as one would expect, he’s around +/- 1.6% I think.

         If they sold it with a 1.5% accuracy claim, or even 1%, it would be fine, it’s just dishonest from them. But you can see why they’re doing that, if you’re trying to pace a 10k at 320 watts based on their tool but your Stryd is telling you you’re at 307 watts (4% too low) you can imagine what’s going to happen !

      4. I’ve manually calibrated my Stryd and on a common run I make the distance varies from 11.20 km to 11.13 km, so about 0.6% difference. Difference is likely due to how I cut corners, which side of the road I run and so on. I also use two different pairs of shoes, so that can cause difference as well. On top of that even Stryd itself has errors in it’s own measurements.

         I find Stryd pretty accurate and consistent when not using the GPS auto calibration of my Polar Vantage V.

      5. I think you’ve got it. I’ve never done a track circuit calibration. Falls in to the category of “life too short” for me 🙂

      6. Pity you didn’t get the tech support you were after.

         The power target for a 10k or 10.4k race is only 1W difference (weight dependant). So if you tried to hold target power and faded badly in the final 1km then it was more likely that the target was wrong, rather than distance calibration being the over riding factor.

         Stryd’s race power targets rely on an accurate Auto CP. That can be quite a minefield to establish. For a 10k race I’d cross check Stryd’s target with doing a 5k test a few weeks before. Then take which ever gives the lowest estimate. Come race day, it’s nicer to start off conservatively then progressively build on the final stages, than the other way around.

         The other method is to look at your Power Duration Curve, turn on Modelled Ability. Look what power you have for the estimated race time. Ask yourself; was that target a previous race, a tempo training session etc. Then for the race you can hopefully go a little bit higher than in training.

         I race quite often, so maybe it’s easier for me, since a past race is a decent predictor of a future race. If your 10k was you main goal for the year and it doesn’t work out – that hurts.

      7. No one reporting accuracy issues is getting honest tech support from Stryd, either you’re happy with their product or you’re doing something wrong. To their credit some people are likely doing something wrong, like auto-calibration, not running properly on a track, think GPS is perfectly accurate, etc…

         As for race strategies, yes I’ve probably run 100 races over the past 10 years including many marathons so I know what works and what doesn’t 😉 What Stryd sell you is that they’ll help you optimize your training and race strategies and while their ecosystem is certainly well designed and the power curve concept well executed it’s stil the same old “garbage in, garbage out” problem.

         It may be that there’s a 1W difference in the target watts for a 10k and a 10.4k (or 9.6k) race but that’s not the info you’re getting on race day if you find yourself facing a 4% variation, the watch/Stryd don’t know they’re 4% off so you’re not going be told to adjust your goal by 1w, you’re going to be told you’re running at say 307 watts instead of your 320 watts objective so you’re going to be pushing way too hard to reach that goal. Hopefully you’ll realize after 1k when you auto-lapped too late (or that you’re running too fast if you lap manually) that something’s wrong and you’ll ease off but a lot of damage will have been done by then and it won’t make for a fun race 😉

      8. I’ve never seen a 4% variation in power with my various shoes, so can’t help on that front.

         I agree, a 4% power change is large as that’s in the region of 5k target vs 10k target.

         I have read of people not reporting air power correctly with Nike shoes that have offset laces. 4% airpower is very common at race pace. Is that relevant to you?

         Another issue I’ve read is people forgetting to switch the Workout app to outdoor mode after indoor use. Indoor mode turns off air power. Just a thought.

      9. I find it strange that anyone would ever take one data point (Power in this case) as their only reference during an important race. If you have Stryd, and a FR935, then you have Power, HR Zones and pace zones all to check at the same time, along with cadence and running feel.

         If you know one of them isn’t working, then why wouldn’t you just switch in the first few hundred meters to the next best?

         During a race I would know on the flat what my pace should roughly be compared with my power so if one was wrong, (which it was on occasion with VaporFly) I just switched to use one that wasn’t as wrong.
         Seems odd that an experienced racer would let technology rule an important race or any race.

         Of course, S customer service has no excuse.

         That said, I never had any problems and ran all of my PB’s with Stryd (and a great third party running plan) so what do I know!

      10. Right and it seems I wasn’t clear but that’s exactly what I did after 1k at 5k pace (rush of the start/flow power target) but I could only really find out for sure when I hit the 1k marker.
          That’s the problem with Stryd, you’re told it’s perfectly accurate so use it for distance/pace, calibrate it perfectly and then it shifts for no reason.
          So yes I could have not used it for distance as a safety net but if you can’t trust it what’s the point? And it would have taken 1k too. Like I wrote earlier garbage in/garbage out.
          Again, maybe it’s my Stryd that shifted for no good reason but they never gave me a chance to find out as they refused to swap it (why wouldn’t they since they’re dealing with someone’s who’s spent hours collecting and sending feedback? Odd…) and there’s a lot of similar feedback in the non Stryd moderated FB Stryd test page. And the “crazy” tester on that page found some +/- 1.6% variations, so pretty close to an overall 4%.

      11. I think Stephen Cousins who you’re referring to, only talks about distance variation with Stryd, not power.

          With time, I think Stryd as good as admit they consider +/-2% OK for distance accuracy.

          But this doesn’t explain your power variation. Have you checked air power across your runs?

      12. Yes SC found some distance variations in his videos (and was aggressively attacked by Stryd in comments) but some far more detailed testing is being done in that FB Stryd testing group. Stryd has (and based on my experience) will never publicly admit to anything else than perfect accuracy and zero variation. Again the only thing they’ll admit to when faced with unexplained variations is that even 5x priced power meters are rated for a percentage of error. Strangely no one at Stryd has deemed it appropriate to put that info in their manuals/box 😉

          As for power variations, we’ve got nothing to measure them against so it’s impossible to prove but I’ve seen them with the same shoe from one race to the other run in the same conditions and it’s common sense that they’re using some forum of pace input to calculate it and if it’s off by 4‰ well…

      13. Yeah – I’m not a fan of how Stryd challenge users when they give negative feedback. Interestingly, Kun (co-founder) when he does reply is very much more pragmatic. So I think it’s just style, more than substance.

          Stryd are quite coy about how accurate it is and just say “accurate out of the box”. But without quantifying how accurate, it kind of means nothing. I’ve read quite a few polls and many users find it 1-2% accurate.

          Stryd comparing % accuracy to power meters is comparing apples to oranges. Stryd is discussing distance accuracy whilst power meters are power/force accuracy, for which there are agreed laws of physics, unlike run power. But any rate, you can now buy a single sided crank power meter for a similar price to a Stryd pod. It’s a senseless debate – one I’m sure I’ve been guilty of in the past.

          Can you post links to two activities that have 4% power variation?

      14. Kun had been quite OK as a replier on their club webpage until users started to make him remember on a constraint of their product: https://club.stryd.com/t/uphill-power-model-regressed/15806

          To be honest I have no FB account and thus cant confirm whether they are suppressive there or not, but on club.stryd.com they are not. And they do not do censoring.

          They just stay in silence sometimes as a tool of evading debates about something they cannot solve.

      15. It’s very obvious that the Stryd device have errors due both to the hardware itself and the algorithms used. This is true for all devices of whatever type, so the relevant point is: Are the numbers reported consistent enough for training and racing guidance?

          Yes, they are, for guidance, but you seems to want rely in it 100% at racing, and that quite simply is a misunderstanding of the limitations of the technology.

          That said, decide on a test protocol with Stryd Company and send your device back for testing.

          I, for one, would love to know how that went.

      16. Webvan, you may be remember my nickname both from forums.garmin.com and club.stryd.com, and if so then you must remember that I am not a user who accepts statements simply nodding, yes, yes, yes.

          Even I think your unit must be defective or your shoes setup may be special, because apart from very low cadence (walking up on steep climbs) I can confirm that the distance measured by Stryd is pretty accurate. My pace range does not cover everything, but I tested my Stryd pacing between 4:50 and 7:30 min/km and it was really accurate, or just really consistent. I cannot give you exact number of deviation, but both on my testing loops it was well within +/- 0.5%.

          I never calibrated it on a standard 400 m track, but on a non-standard track being said to be 525 m. Even if it is not 525 m, just 521 m or 530 m, my backtests showed that the total distance varies less than 2 m/lap. The shorter the test course/track is the more important is how to start (after waking up your Stryd you have to walk to the start, wait for 3-4 seconds) and how to stop (exactly on the start/finish line and you have to wait for 3-4 seconds before you push the stop buttons.

          As a sidenote I must add that even with this protocol if you later during a run on a track/course/loop you always push the lap button exactly when you at the start line it may happen that there is a deviation more than a 1%, but I guess it may be simply due to the constraints of the processor of my watch. Even a Fenix 6x can be too busy just when it should record a lap. Or maybe Stryd is too busy and there is some time lag sometimes. But this phenomenon does not have any impact on the total distance.

          I may be simply lucky, though.

3. it works on Epix 2 now as on FR955 or a future firmware?

   1. soon

      1. It will just appear, soon, on Epix 2 (and offshoots like Quatix 7) with an automatic update? No need for the CIQ app, right?

4. Garmin has some description how running power is calculated.
   https://www.garmin.com/en-US/garmin-technology/running-science/running-dynamics/running-power/
   What is interesting, Wind/Air Power component comes not only from a weather forecast, but also barometer.

   1. It’s quite non-specific what the barometer sensor is doing.

      “ The Running Power app uses your heading from GPS and reported wind conditions for your area, augmented by data from the barometer on your watch, to determine how much wind you’re likely experiencing.”

      1. yes, i’m reasonably sure that wind does not come from the barometer.
         however that article says “Barometric data to detect local conditions”, so perhaps the barometer is used to check the actual conditions agree with the forecast ones. if not then turn off the auto-wind part of the calculation.

         anyway, to cut a long story short. forecast wind is inaccurate.

      2. I’m sceptical to the general usefulness of weather forecasts for wind power calculations, and there should be an option to disable that. Even Stryd can be iffy at times.

5. Reference your comment on weight: I imagine many new Garmin and Apple run power uses who have smart scales will turn on auto sync by default. It’ll make intuitive sense to them.

   1. indeed so.
      i’ve changed the wording, above. i think we agee but i am perhaps not being clear.
      basically..smartscales are bad for running power! 😉
      (I use Garmin smartscales most morning but don’t update my weight in stryd, it seems to be updated in Garmin running power as my w/kg is heading South for the Summer)

      1. thread drift – which garmin scales do you have?
         I had the 1st generation from 5yrs ago. They did a firmware update which basically “killed” the fat % part. Many users complained of this on the fourum.
         I wonder if the Mk2 is better now?

      2. https://the5krunner.com/2020/10/23/garmin-index-s2-review-smart-scale-2-wifi-2020/

         2 years on and 2kg on I’m not sure that i belive the 2kg extra muscle I’ve supposedly put on!

6. Hi,

   Maybe I missed it, is there a built-in running power guided test to determine your power zone like the lactate threshold guided test?

   1. IDK, sorry.
      I get the impression that CP mysteriously comes from the identification of LTHR (ie power at LTHR when lthr is determined, that’s a half-educated guess). so if you did the lactate test then the CP should be calculated at the same time.

   2. I don’t know if there is a built in test, but running coach Steve Palladino has a suggested protocol in this paper he wrote about CP:

      https://docs.google.com/document/u/1/d/e/2PACX-1vSz0ryPMD4vVJLi2HGjkqoUKRzRPhN78TUNlCmuYcrLTjeLPx-ZaiROk9NG96iatJ4zYKAQcbjEEot0/pub

      1. yes, a 30-minute or 10k maximal effort should be enough to trigger whatever Garmin is doing behind the scenes. Steve’s protocols would likely also trigger something similar but it’s a bit complex to read all of that for runners new to power.

   3. this is quite interesting. it’s for cyclists power zones but running is similar https://www.cyclingapps.net/blog/new-take-on-the-classic-power-zone-calculator/
      (The TTE part wont be similar). often it’s just easier for you to put in whatever maximal value you already have than to do a test.

7. Maybe I’m missing something, but it seems the only advantage of running power over heart rate is it’s more instantaneous, right? But to me, when I hit a hill my heart rate climbs pretty darn quickly. What am I missing just staying with heart rate?

   1. Lag and drift are known issues with HR. Power will respond quick than HR. HR takes a mile or so to get up to steady state, then will drift upwards during the run, for the same intensity. HR also has day to day variation.

      It’s not that power is without its own issues, but generally offers a better metric for gauging intensity

   2. there is drift over long workouts
      there is the impact of starting fatigue
      there is the impact of caffeine and other factors that elevate hr
      initial delay of 10-15 minutes at the start then the 30 second or so lag as you shift efforts during the workout

      as Will has said elsewhere you should still look at hr and pace. it’s just an extra tool and maybe it can be your primary one. perhaps the one issuette i have with stryd is that their ethos comes from the ‘only use power’ viewpoint, although that’s changing

8. I’m sorry to question the facts written above, but the HRM-Dual isn’t just a typo?
   I’m quite sure it doesn’t have running metrics and it wasn’t supported in the iq app before.
   I would be happy to see it added though.

   1. yes you are right, ty

9. Are the Power Workouts only available if you have a native Run Power watch? I can’t seem to make one, but I have a Fenix 6.

   1. i havent got an f6 to test that on but it should work?
      have you got the garmin running power ciq data field installed?
      i assume so.
      then how are you trying to create the power workout?

      1. I’ve been trying on the Garmin Connect web app, but there doesn’t seem to be a spot for me to add a power step?

      2. https://the5krunner.com/wp-content/uploads/2022/08/Capture-1.png

10. Great article! But I’m still don’t know how to configure my watch to work with Power Running zones. I have Garmin 945 and HRM Pro. Since I bought HRM Pro I downloaded the app Combo Running Power from GarminLabs. It shows my power running (the graphic and the Avg Running Power), but I couldn’t set my zone power. I downloaded Running Power also from GarminLabs and setted the zones power on ConnectIQ app. I can create a workout on Connect app (iOS) with intensity target setted to power zones (I can’t create this workout with power zones on Garmin website) but the zones on the Connect app aren’t the same as ConnectIQ. When I do the workout It shows the target range on the watch but the power isn’t show on the watch (as in the first image of this article where it shows too low 0, mine shows –).

    If I change my Data Screen to show ConnectIQ Field it seems to show the zones I setted on ConnecIQ. So I can conclude it is taking the power, it just can’t use the power on the screen that shows the power range (to show if I’m in the target, to low or to high).

    Also When I try to set the power zones on Connect app (more > Garmin Devices > Forerunner 945 > User Settings > Power Zones) it just shows Cycling and Cross Country Skiing.

    Do you know if this feature just works in 955? Or am I doing something wrong?

    1. It doesn’t sound like you are doing anything wrong.
       The article is focused on native power on the 955.
       the ciq power should function similarly (identically!) but some vagaries, as you describe, wouldn’t surprise me.

11. Hi, can I disable Garmin Run Power ?
    I use Stryd and numbers from Garmin are getting in my way (they’re prioritized in Garmin, Strava and other ecosystems)

    1. other than not wearing a garmin hrm strap, i would love to know the answer as well.

12. very good answer, thank you
    It seems only HRM-Pro and Pro-Plus produce Power data (calculated from their running dynamics).
    The odd thing is they are not connected as PWR sensors (something that could enable 3rd party PWR sensors, like Stryd)

    1. hi
       no the HRMs (or the RD POD) provide 3-d force data. that’s one of the inputs for the power calculation that takes place in the watch
       the likes of stryd is a power sensor i n the sense that it calculates power internally and transmits it

13. How is the wind impact on power disabled? I can’t seem to find the setting to disable wind. FYI, I have the Quatix 7, and the new software with power etc was released a few days ago.

    1. see above where the article briefly talks about a connection to garmin connect within the last hour

14. At least with the last firmware of Epix2, running power values on indoor runs(treadmill) are not shown in Trainingpeaks. It shows there rTSS or hrTSS instead of TSS.

    1. i dont really use training peaks but it sounds like an issue for them to sort out.

15. I have an epix 2. I also have stryd. Can you set the watch to take it’s native power reading from the stryd footpod? Garmin’s numbers are quite a bit different than stryd’s. I would love to build a power based workout… but my stryd data is what I’m accustomed to interpreting. However I’m not paying a stryd membership, so i can’t build a power workout through the stryd app (i think it’s ridiculous that after spending $200 on the footpod that you don’t even get to access a workout builder😡).

    1. hi
       you can build power workouts elsewhere without stryd membership (I don’t have membership) – try the limited free version of training peaks or the free version of Final Surge
       you have to use the stryd ciq data field as garmin has implemented a PROPRIETARY form of running power on its watches

Comments are closed.