Thanks @nasatt. I used Gotoes and made an excel. Average including zero’s was the same as intervals.icu.
Unlucky still no explanation why the laps averages of the Garmin Connect are higher.
1 Like
When you exclude the zeros, does it align with Garmin Connect?
Connect: Average power = 279 W,
Intervals: Average power = 270 W
Excel: Average power without zeros = 317 W
Excel: Average power = 271 W
How does your head unit record averaging power? Including zeros or excluding?
The default is excluding, so you have to change it to include.
Cadence averaging: Do not include zeros and Power Averaging: include zeros
Hi David,
Did you already had a closer look?
I noticed that all my ride activities on intervals give somehow lower power values than Garmin does.
For example yesterday activity (lap button pressed once the race started, total lap time: 1h12m): 273w average and 295w normalized. While Garmin gives 288w and 309w. The Garmin values are also the same values that I saw on my bike computer during the race, therefore I tend to think the Garmin values are ok and that somehow Intervals is going funky town on the power averages.
For me Connect and INtervals always have a difference , even with Distance , but i only ride indoors with Power , so i prefer use BreakAway to record my rides, and Garmin just for “in case” ; i used before because of Moxy , but now Breakaway add Smo2 on his calculations - I still dont know what to do when i update my Garmin to use “alphaHRV” , i tried use Gotoes or fix-fitfiles to add Smo2 rides with FitFiles , w/o sucess , so when i will add DFA-1 , will be a new issue with Garmin
I have just had a look at that activity on fitfileviewer.com and using Google sheets and the streams CSV. According to the power data in the file the average for your race lap is 273w. I will DM you a link to my sheet. There aren’t any “extra” power fields. I don’t understand why such different data is being recorded for the laps. Maybe Garmin is using the power set on the indoor trainer for that?
You might want to contact their support about this.
Go to Settings and under Power section what is your Power Spikes setting? As David posted earlier, Intervals “fixes” your power data and then calculates average and normalized power. That can be one big difference between Intervals and other analytics. I’ve set my Power Spikes to 999% to effectively turn off “fixing” data that is actually good data.
Same here, Garmin connect and Intervals do not line up:
Intervals: 283, 331(NP), 290(20min)
Garmin: 286, 334(NP), 291(20min)
This may be the reason for the difference between power shown Garmin and Intervals:
''So this gets a little complicated, but the difference is due to a couple of things going on. One is that there are radio outages that are causing some of the records to be back filled. The second is that the tools you are using to process the data are using the power values assigned to each record and are not using the accumulated power values. Using the accumulated power values is the better approach. Because of this the tools are subject to how the dropped messages are being backed filled.
The accumulated power record at the start of that #6 30 second window is 199200. The end value is 208219. That is a difference of 9019 kj. Average that over 30 seconds gives 300.63 or 301 watts. This is what the Edge reports.
The last 2 seconds on that lap and the 1st second of the next lap have the same accumulated power value because of an outage.
Using the 1 second power values that have been computed using the backfilled data because of the radio drops gives 9330 kj over 30 seconds or 311 as you get.
For the next 15 second lap the accumulated power starts and ends with 208219, 210262 or 2042 kj over 15 seconds 136.2 watts.
The sum of the individual power records gives 1720 over 15 seconds or 115’’
1 Like
It seems Garmin calculates average’s using the accumulated power / time.
Here an excel file and the original fit files. The excel file excludes the auto pauze moments. I have put 3 different fit files in the sheet and the average’s calculated using the accumulated power / time are the same as Garmin.
Averages calculated by averaging the power values out of the fit file are the same as Intervals.icu.
Undoubtedly Garmin has reasons for calculating the averages this way. Curios what this reason is.
I can’t answer this question, but I would ask a question that might answer it…
Is it easier for their device software to process the accumulated power divided by duration to be able to show it on the device while you are out riding?
If I’m following the explanation, here is one very good reason:
- only calculate average power from data actually collected
In other words, don’t “create” data that doesn’t exist.
Which also means don’t change the power data actually collected. Intervals “fixing” power spikes was my biggest issue when first using this really nice tool, because I can generate large spikes in power over 1-3 seconds. I’ve seen these “spikes” with all power meters I’ve used - Stages gen2 single side, Stages gen3 dual sided, SRAM Force, SRAM Red (2), Garmin 200 series pedals, and Stages stationary bikes at the gym.
I don’t know if this was figured out. Thought this might help. Big discrepancies might be easier to see?
Difference in Garmin Avg Pwr 181 vs. Intervals Avg Pwr 102.
I came looking for this thread because of this similar experience. Edge 540. Garmin Average power is 181. (calculated based on the time the timer is running per Garmin on Edge 540 excluding zeros by default.). Garmin Max Average 103 timer running. (Garmin help states Average excludes zeros while the Max Average includes zeros.). Intervals Normalized is 147 and average is 102. Garmin Moving Time 1:12.49 (time user was actively moving). Total Time 1:13.29. ( appears to be as long as the timer is on. In my case with auto pause on it does not include time the timer was paused. Should theoretically be the same as Moving Time but they explain Moving Time as stopped moving and Time/Timer as time when timer is stopped from Auto Pause. On their online documentation this is called Time/Timer). Elapsed Time 1:55:41 (the moment one hits start on the device and until it is manually stopped). Intervals :31 Coasting.
So for reference, this ride was a mountain bike ride. Intervals created 33 intervals of varying length rating my Max Power at 187% to 337% of FTP. So lots of intervals. Coasting :31 so lots of zeros. Intervals showed 33 Intervals and they were Zone 7. How many more or others, certainly, but I don’t know.
I noticed a few things. First, when I start pedaling from zero, the power and cadence are not immediately displayed on the Edge. Second, the Average Power numbers displayed at the bottom of the Intervals section in Intervals have a range(?) displayed. My Average Power column ±40. Max Power column -+56. Lastly, there was a significant number of zero cadence and zero power periods. Intervals shows .31 minutes coasting. So was moving 1:12 avg power Intervals 102, Garmin Avg 103, Intervals coasting :31, and Garmin 181 watts Average Pwr timer is moving, and Intervals 147 Normalized. Mean anything?
Anyway, I thought the large difference might help illuminate the reason for any discrepancies. Huge variation of Garmin’s Average Power but under unusual conditions. This particular mountain bike system had a lot of ridges, gulches, switchbacks, and steep drops and ascents. Lots of time power was off the pedals and then followed by power on. Possibly lots of filled in zeros? Errors in length of efforts? Differences in timing and reporting of power by the power devices which is exaggerated by the number of pedal starts and stops experienced during the ride? Breaks in communication between the power pedals and the computer while the timer is still running because didn’t auto pause, the bike is still moving per GPS and altimeter.
I’m not going to lose sleep over this. I feel like I got a version of a Threshold HIIT ride 
and that’s close enough for my purposes. I can see though how this kind of thing can drive someone crazy.
Sorry for long marginally constructed post.
1 Like
Do you have „power average without zero“ activated on Garmin? Because intervals will consider the zeros (during coasting or pause times) into the average calculation.
Yes, I do. Well, so that’s that.
Interestingly Garmin reports to both Strava and Intervals. Strava was showing 102 Average power, but they have a new Beta AI which I asked this morning for more information. It divulged when prompted for more information beyond a basic report that I was, “pushing your limits with a power of 181.” If Strava has that information so would Intervals.
So, the information is all there, just a matter of what the software displays. I think I like having a number without the zeros. To me that seems to be useful additional detail for the actual effort of a ride beyond the load number.
Is there a way to see that number in Intervals?
I don’t think you can do it in icu. You could create intervals for your active times and see the averages of that parts.
And there are a lot of posts in the forum, that show calculating with zeros make more sense.
Just see it that way. Do 5m at 100%FTP. Stop/Pause and rest for 10m. Repeat it another 3 times. Average without zeros would be your FTP value for 20m.
Average with zeros: a third of your FTP for 1hour.
You can’t compare the average without zero, because you didn’t do these watts for the whole 20m, which would be far more challenging.
2 Likes
Hi @david
My new racing season just kicked off and, once again, I’m seeing that the average and normalized power numbers from Garmin don’t align with what Intervals.icu reports. I’m still hopeful this issue can eventually be addressed.
Over the winter, I spoke with a developer at Garmin, who confirmed that Garmin calculates average and normalized power based on accumulated power values. I also contacted a software developer at Power2Max, and they indicated that from a power meter perspective, accumulated power values are indeed the correct basis for both average and normalized power calculations.
To summarize the difference in how the data is processed:
- Garmin uses accumulated power values, which helps smooth out short-term dropouts or gaps in the data.
- Intervals.icu uses power data interpolated to 1-second intervals, which can introduce discrepancies—especially during short, high-intensity efforts like in criterium racing.
This difference seems to lead to a significant error in those short, punchy intervals. It would be great to see some alignment between the two platforms—or perhaps an option in Intervals.icu to better match Garmin’s method.
Thanks again for all your work on the platform!