Page 1 of 1

Speed to high - Skiing FIT import The data

Posted: Mon 7. Jan 2019, 22:11
by froodproton
I have imported .FIT file from Garmin and see the Speed for Skiing activity being way too high.
The watch was showing realistic highest speed at about 72 km/h but Runalyze shows 96,8 km/h peeks in the area "pace data".
It was imported today.

Re: Speed to high - Skiing FIT import The data

Posted: Tue 8. Jan 2019, 10:18
by laufhannes
There's nothing we can do about.

Calculating some 'max speed' based on gps data is extremely inaccurate (that's why we don't show a field for max speed at all). Gps coordinates are recorded every second and pace can only be calculated based on these differences (based on coordinates or recorded distance) every second. In addition, plotting accuracy is limited (in general we can't show a data point for every second) and you can be sure: you don't want to see that data as speed data is extremely noisy.

Feel free to adjust the precision of your plots. In your configuration, you can choose from one data point every 50m up to every 500m or a total number of data points (50 to 1000). The precision will highly affect the visual max speed. But still, you need to accept: There is (based on the recorded data) no 'truth' for your max speed.

Re: Speed to high - Skiing FIT import The data

Posted: Tue 8. Jan 2019, 13:20
by froodproton
I do not understand the difficulty.
There are two datapoints - position and time - distance and time --> Speed
The Garmin app calculates the speed - most probably - in the same way. Maybe with even some more smoothing algorithms - I don't know.

I have attached the FIT file - maybe it helps.

Re: Speed to high - Skiing FIT import The data

Posted: Tue 8. Jan 2019, 13:42
by laufhannes
froodproton wrote:I do not understand the difficulty.
There are two datapoints - position and time - distance and time --> Speed

That's what we do and that's what you are not happy with. The difficulty is to explain everyone that there is no one truth and that things like Garmin has value X, you have something else --> You are wrong, please fix. are not correct. (Most frequently on the subject of elevation)

froodproton wrote:
Tue 8. Jan 2019, 13:20
Maybe with even some more smoothing algorithms - I don't know.

Yes, of course we can think about the smoothing algorithm and try to adjust. But it's not the right way to do this based on one single activity by assuming that Garmin's value of 72 km/h is the one truth.

As I already said, we have a configuration option for the plot precision which is one kind of a moving average - one that is easy to understand for everyone.