Avg GAP higher then avg pace

Could you scroll the table from GC to the far right, and post also the values of the Moving Pace?

sure

btw. that's also interesting or rather wrong as I never stop or walk and according to what I see regarding moving time, every time I click lap watch lose 1s, moving pace should be the same as pace here.

To me the data looks all right. The ascents and descents are identical in all but five laps, and even in the five laps the difference is minimal. Without knowing the exact elevation profile and the algorithms they are using, it is hard to guess which method is better, but TBH, I find strange that Strava shows faster GAP at laps with negative elevation gain (lap 2 and 10). In these cases I'd believe more Garmin (showing slower GAP on descents) than Strava.

You can also check the data with some online GAP calculators. For example when I check the 9th lap with the pace of 5:24 and elevation gain of 3m, the Grade-Adjusted Pace (GAP) Calculator for Uphills and Downhills - from Running Writings shows the GAP of 5:15, which is pretty close to the GAP of 5:16 in GC. Strava tells 5:21.

At the 10th lap with pace 4:14 and elevation loss of 1m, the calculator shows GAP 4:15, exactly the same as GC. Strava claims 4:10 for some reason.

It would be better comparing the GAP on runs with a less accidental elevation profile, with long steady ascents, and long descents, always keeping each lap purely ascending, or purely descending.

Not sure how you can say that data looks ok. First of all as I mentioned before Time should be
the same as Moving time and Avg Pace should be the same as moving Pace and almost for every lap
there is difference (4:29.3 even after rounding it's not 4:28). I uploaded this activity to
TrainigPeaks and it's exactly the same. Also you mentioned above that if elevation gain is close to zero Gap
should be the same as Pace. I provided example to prove that's not the case. For lap 4 is the same
and then for lap 6,8,12 is not. Then you have recovery laps when again even when elevation gain is zero Gap is
not the same and in most cases is faster then Strava. Regarding calculator you just selected one lap when there is elevation
gain and gap is faster, for other laps it's not the case. You also mentioned that is strange as Strava shows
faster GAP even when there is negative elevation gain and it's not true because on Strava there is no negative gain,
elevation gains are different for some reason then in Garmin Connect and what is surprising they are more accurate.
My whole point here was to get answer what the *** this metric is and how is calculated and you are trying to
convince me that is not wrong where there is no explanation for those numbers :) I will stop with this as
it's not so important for me, but maybe someone else have other idea/thoughts on that ?

I agree that this looks wrong. For any given lap / period of time, if...

- ascent equals descent

- ascent and descent are not 0

- and even better both, both ascent and descent correspond to going up and down the same hill (or two hills with the same "profile" / length / grade) (I realize this part doesn't apply to your situation though)

...then I would always expect grade adjusted pace to be faster than actual pace.

Why? Because it's known that the energy you gain from going down a hill is less than the energy you lose from going up the same hill.

If I run a downhill 10K, followed by running up the same 10K route, I expect the effort to be harder than running a flat 20k route.

Similarly, for a given segment, if ascent equals descent and both are not equal to 0, and especially if the uphill part of the segment is somewhat symmetrical to the downhill part of the segment, then I would assume that GAP should be faster than actual pace.

At worst I would expect GAP to maybe be the same as actual pace, when ascent and descent are equal.

What I wouldn't expect is GAP to be slower than actual pace, unless there was some crazy difference in elevation profiles for the downhill part vs the uphill part. Like maybe you descended 5 m on a steep 10% grade, but you ascended 5 m on a gradual 1% grade? Something crazy like that.

That's just my intuition though. All I can say is runalyze and strava both produce intuitive GAP values for me, while Garmin's GAP values often seem counter-intuitive (and are qualitatively the opposite of runalyze/strava in many cases - i.e. where runalyze/strava show me a faster GAP than actual pace for a given lap or activity, Garmin will often show a slower GAP than actual pace.)

Maybe the best way to definitely test this would be to run up and down the *same hill* for a single lap and compare the results for Garmin, Strava and Runalyze. idk maybe that wouldn't

Others have made similar comments that Garmin's GAP seems wrong:

https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-955-series/313929/grade-adjusted-pace-comparison-garmin-vs-strava-vs-runalyze-bug

I'm not an expert on GAP, so I'm not really qualified to talk about anything except my own intuition on the subject.