Hike Activity Highly Inaccurate.

Jesus_Arango said:

Erratic GPS or location have nothing to do with the problem

It has everything to do with the problem. If the step counter indicates that you've taken two steps, but your GPS location has moved by tens of meters, then the algorithm should be discarding the distance calculation as it's physically impossible to move in such a way.This logic applies whether youre shuffling around and making some lunch at your camp site, or belting out a 4min/km on the trail.. 

If you have a look at the GPS plot of those hikes where you've measured >10% over the planned distance (whether you have auto-pause enabled or not), you'll notice that the plot appears to be a haphazardly drawn scrawl as a result of reflected GPS signals when close to large objects (buildings, rock walls etc) or in valleys.

I've also noticed when checking data in the Garmin Connect app on my phone after a hike and after syncing the hike data into Connect, that it will give the average speed of the hike, say 2.5 mph (which is reasonable), and the max speed, often around 25-30 mph, a physical impossibility since no one walks that fast.  I think there is probably a tie-in here with the distance inaccuracy issue.

It would indeed be nice for someone from Garmin to weigh in here, but I have the feeling that no one from Garmin pays any attention to these threads.

hi everybody, do you found solution for this issue or not, it is happen with me in Fenix 5 Plus. 

When I said that GPS accuracy is not the problem I meant that any decent movement detection algorithm can work even in the presence of GPS errors. No algorithm would decide if the subject is moving or stationary based on a distance threshold between two points. If the threshold is too small than GPS errors can be greater than the threshold. And if you make the threshold larger than most errors then the subject would have to run to keep the device in "moving" state.

What algorithms do is to keep a set of points over a moving/rolling window of time (i.e 10 seconds). If the distance between any two points in the current time window (i.e last 10 secs) is greater than a threshold then the subject is moving, otherwise it is stationary.

On Fenix watches, this algorithm is not correctly implemented.

I find the accuracy of the clock outstanding. I have already covered thousands of kilometres in the Alps, in New Zealand, on the islands in the Mediterranean. Either jogging or hiking!
I never use automatic start and stop measurements. When I stop, I stop the watch. If I keep running, I start the clock manually again.
Automatic functions are problematic because rest is almost impossible to define. For example, if I sit in a restaurant and get up for a moment to get something at the bar, is it still rest or is it already moving again?

Translated with www.DeepL.com/Translator (free version)

I have an instinct which tracks my hiking mileage accurately my friend has the fenix 5s and her mileage is consistently higher. We swapped watches I wore hers and she wore mine to test it out. When we did mileages were both correct and almost exactly the same. What is going on?

My instinct is very accurate when tracking hiking mileage. My friends Fenix 5s mileage is consistanly higher. We swapped watches I wore hers and she wore mine to see if the Fenix was still higher. When we swapped both watches were accurate and almost identical on mileage.
what’s going on here? 

6527400 said:

When we did mileages were both correct and almost exactly the same. What is going on?

Distance isn't 100% GPS acquisition acquired. When your watch loses signal, which you will not be aware of, your watch's accelerometer does its best to use your programmed user profile settings for height etc. to capture steps and predict the distance as accurately as possible. Bottom line, it is highly complex and not as straight forward as it can seem to be.