Enduro Elevation Gain Accuracy (Again)

I've been noticing the same thing recently. On a recent trail run (16miles/3000ft), the Enduro elevation was 15% higher than another Garmin model on the same route and the value calculated by the Strava for the same route. I'm not sure what the issue is, but something seems a bit off.

But if it is the case with your friends’ Enduro watches, and you do activities together and all of you have phones at you with constant BT connection (so  the chance of having a difference in connected-DEM availability is more or less excluded) then your watch is overly sensitive to pressure changes. I know you tried 2 Enduro watches, but I have no other idea. Maybe your Enduro is confused by sweat and the other are not, but to be honest I saw this phenomenon mainly with Fenix 3 HR, rarely with F5+, and almost never with F6X. And I deem the latter is the same construction as Enduro as regards barometer hole.

As I posted in the other thread, I am about to start to use an Enduro and will report my findings about any altitude difference coming from the fact that Enduro can use the DEM of our phones, but lack the DEM map placed in the watch.

Btw I am afraid that this connected-DEM is used only at start and not continously, but the support articles are not clear. If so then it is critical what sort of GPS+barometer algorithm is used by Garmin in these years. 

it had been my favourite algorithm before watches with maps (and thus with DEM data) appeared:

patents.google.com/.../en

https://patents.google.com/patent/US7142152B2/en

After I burnt almost a full working day to set and finetune everything on my new Enduro involving a hard reset as a needed step after I had tried to save some time and had copied some predefined sport apps from my Fenix 6X, I went for a run.

The first run was not fully flat, but close to it. My F6X ended with 62/72 and my Enduro with 66/72 as regards total ascents/total descents.

I was continuously observing the (absolute) elevation, that is altitude, but Garmin has been using elevation as a datafield name for decades, and I realized that the elevation of F6X was fluctuating around that of Enduro. So I bet the connected-DEM  data coming from the phone is used only at start. While Fenix 6X was trying to substitute the calculated elevation which comes from pressure differences, with the DEM data of the map. But certainly this small fluctuations did not create additional fake ascents and descents.

To be honest I like the method of Enduro, I always preach that a barometer based elevation finetuned by the averages of gps based elevation using big enough dataset is better than a barometer based elevation partly overwritten by DEM data. Especially if the activitiy does not last for hours, and your path includes wildly uneven terrain. OK, if there is strong wind DEM data help to stabilize the “final” elevation as a  displayed data.

On the other let me repeat that I observed that DEM data is used not in a brutal and stupid way, and it is why the total ascent/descent data are acceptable.

My method now was first to write my post, and just after it to check the recorded elevation chart, and I am happy to announce that my observation and long term belief is confirmed by the charts. The very small drift can be seen in Enduro’s chart while it is eliminated from F6X’s chart, but in the section which was out and back, the former chart is more symmetric.

Yesterday I had a run+walk activity of 49 km on almost full flat terrain except for some very small bumps like going up levees.

My Fenix 6X and Enduro recorded almost the same calories, training effect and so on, but while my F6X calculated 247 m as total ascent, my Enduro was cleverer to measure it as only 132 m.

https://connect.garmin.com/modern/activity/11667754349

https://connect.garmin.com/modern/activity/11667770297

Sidenote on physio:

I have never used at the same time two "Performance conditiion capable" watches running on the same software, but it is quite interesting that how F6X and Enduro work.

While their total calories burnt  numbers are close to each other from the first activity of the newly acquired Enduro, the training effect numbers are differing even if all the physio-related syncs are switched on, although it seems the the difference is decreasing. My conclusion: not all the previous physio data were shared with Enduro and/or HRV data are detected differently due to different CPU operation (note: Enduro is slower to have bigger battery life).

Finally I could not decode how the performance condition calculation works, because on the first run my Enduro gave 2-3 points lower number from the first detection and the difference persisted for the whole run, while yesterday my F6X did just the opposite. Until now I saw no sign of convergence,

Sidenote on sleep detection

Sorry to say, but I lost the remaining (almost low) confidence in sleep phase detection and consequently the accuracy of scoring. These two watches say totally different things as if they were not just two different generation of Garmin outdoors watches, but like watches of totally two different manufacturers.

The good news is that the sleep data of the primary device is kept, while the other is disappearing day after day, and each day it seems it was just a bad dream to have totally different sleep data. :-)