Invalid Total Descend

My friend, before start writing many words, read carefully my post.

"The thing is, your watch doesn't know for sure your loop ends at the same elevation as it started" -

Im not comparing garmin reported elevation with actual elevation. IM COMPARING GARMIN REPORTED ELEVATION VS GARMIN REPORTED ASCENT-DESCENT.

Total ascent 327m (correct by your words).

If you look at the values listed in Garmin Connect link:

So if your start elevation was 315.6m and your end elevation was 317.4, then your total descent is 325m. All the maths check out, there is only some rounding involved to display the values as integers. If you are seeing some larger drift, link the activity so I can see what you mean? And keep in mind start elevation is not necessarily equal to end elevation :D

Here same route https://connect.garmin.com/modern/activity/11410494666

Now difference 74 meters.

As i mentioned before - there is another example. But if you dont want read all post - just first sentence, here another example, with same route https://connect.garmin.com/modern/activity/11410494666

Now, also rounding error ? laugh now ?

This one definitively looks very odd, something is going wrong with the calculation. Garmin-Sierra can you take a look at this?

Route starts at 314m asl, ends at 314m asl (according to the elevation plot in Garmin Connect), but ascent is listed as 399m while descent is listed as 322m.

As a side note, this linked activity also shows how Garmin algorithms are severely underestimating the effort of uphill, as I mention in https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-965/335446/pacepro-how-broken-is-pace-pro-with-elevation-e-g-trailrunning - here in the form of "grade adjusted pace": just look at this plot of heart-rate and grade adjusted pace. Heart-rate is indicating the downhill part is a bit easier than the uphill part, but grade adjusted pace is claiming the uphill part was easier than the dowhill part.

I looked at the GAP vs normal pace on this activity and on the uphill the GAP is faster than actual and on the downhill GAP is slower than actual. Thats seems logical? 

On the graph you pasted the GAP on the uphill is slower than the GAP on the downhill, I don't see how that indicates the uphill is easier. The closer the pace is to the top of the chart the faster it is.

Elevation looks strange for sure.

S.M.R. Fast is higher up, the blue line (grade adjusted pace) is clearly lower (=slower) for the uphill and higher (=faster) on average for the downhill? And the heart rate vice versa, clearly higher for the uphill (especially if you ignore the first few minutes where it ramps up from resting), and clearly lower for the downhill.

So this grad adjusted pace algorithm is claiming downhill was harder work than uphill, while the heart-rate tells us the true story is likely quite the opposite.

If I would have to guess numbers for averages, they are around:

Uphill HR (ignoring first 1/3): 162 bpm

Dowhill HR (ignoring first 1/3): 149 bpm

Uphill grade adjusted pace: ~5min/km

Downhill grade adjusted pace: ~4:30min/km

Grade adjusted pace - is useless metric. This is only tell how Garmin "think" how you "should" work on uphill-downhill. But unfortunately, my body don't care what "Garmin thinks" about it

I got another fresh example for you:

new one https://connect.garmin.com/modern/activity/11511317392

Now, descend values missed.

Same route, same start-end points.

Yes, Garmin grade adjusted pace is useless, that is my point! This is something I wish Garmin would address, update their algorithm on how to calculate grade adjusted pace, as it stands it is useless.