Cycling Data Mismatch

LeetleGreyCells

Un rouleur infatigable
My wife got me a Wahoo Elemnt Bolt for my birthday and yesterday I used it for the first time when we went for a slow family pootle. She bought the bundle, so I had fitted the sensors (cadence, speed and heart strap) before we started the ride as I wanted to test it was all working together and correctly. It was and did.

After the ride, the data uploaded to Strava and Komoot as I had set it. But none of it matches even though it was all recorded on the same device. Here's what I mean below.

Which one is accurate? I realise there's not much variation (other than the max. speed on the Wahoo Companion App), but I like accuracy. I may be hoping for too much. Any thoughts?

From Wahoo Companion App:
fullsizeoutput_176c.jpeg
fullsizeoutput_176b.jpeg

From Strava:
Screenshot 2019-07-08 at 06.42.32.png


From Komoot:
Screenshot 2019-07-08 at 06.43.03.png
 

gaijintendo

Veteran
Location
Scotchland
I think Strava uses a lot of science to get things right. I dropped a few quote marks there. "Science". "Right".

I think they fix gps blips and calculate power based on everything from incline cadence, you/your bile weight and wind direction.
 

Dogtrousers

Kilometre nibbler
They seem to agree pretty well. Apart from the bonkers 140kmh speed, which you can discard as an outlier.

The elevation readings have a wider spread (364 / 364 / 190) I'd guess because the 190 is inferred from map elevation data and the 364 from GPS /altimeter data. Does the GPS unit have a barometric altimeter?

Just pick the set that is most convenient for you to use and stick with that. Pretty much all of your use of this data will be comparative anyway (was this ride faster/further than last week's?) so the absolute values don't matter so much.
 
Last edited:
OP
LeetleGreyCells

LeetleGreyCells

Un rouleur infatigable
Each of them recalculate their own 'moving time', so if you were going really slowly you can see a disparity. You have a huge amount of 'paused time' so I can see how it would give weird figures.

"A man with a watch knows what time it is. A man with two watches is never sure."
There was a lot of waiting about at the top of 'hills' while little legs caught up.

Here's today's solo ride data which should be (and looks) more accurate. Still not perfect, but then I'm asking too much methinks!

From Wahoo Companion App:
fullsizeoutput_176e.jpeg
fullsizeoutput_1770.jpeg


From Strava:
Screenshot 2019-07-08 at 13.24.17.png


From Komoot:
Screenshot 2019-07-08 at 13.24.39.png
 

Dogtrousers

Kilometre nibbler
There was a lot of waiting about at the top of 'hills' while little legs caught up.

Here's today's solo ride data which should be (and looks) more accurate. Still not perfect, but then I'm asking too much methinks!
I'm not sure what you're expecting. They seem really consistent to me, apart from the elevation, and the reason for that is probably because Komoot uses a different estimation method.

"Moving average" speed is never going to be "accurate" because to be accurate you'd need some kind of gold standard you could measure against. And such a thing doesn't exist. Various systems do a bit of guesswork to correct for stopped time but it's all guesswork/estimation.

"Distance climbed" has the same problem. How would you measure that "accurately"? From a map or using surveyor's equipment? No, that doesn't take into account minor ups and downs in the road. Do you include every stone and pothole that you ride over? It's all guesswork/estimation.

Distance - yes that can theoretically be measured accurately, but remember your front wheel will be going further than your back wheel as it wiggles slightly more so there's a limit to that too.

Time. Yes time can be measured with great accuracy.

If it's accuracy (closeness to the "true" value) that you want then you must be able to define what that "true" value is. You can't do that for moving speed, or for elevation. If it's precision (confidence that measuring the same thing multiple times will give the same result) that you want, then just choose one method and stick with it. It should be precise and repeatable even if it doesn't give the same value as some other method.

Comparing multiple methods will just lead to disappointment and confusion. Like @Milkfloat 's man with two watches.
 
OP
LeetleGreyCells

LeetleGreyCells

Un rouleur infatigable
I'm not sure what you're expecting. They seem really consistent to me, apart from the elevation, and the reason for that is probably because Komoot uses a different estimation method.

"Moving average" speed is never going to be "accurate" because to be accurate you'd need some kind of gold standard you could measure against. And such a thing doesn't exist. Various systems do a bit of guesswork to correct for stopped time but it's all guesswork/estimation.

"Distance climbed" has the same problem. How would you measure that "accurately"? From a map or using surveyor's equipment? No, that doesn't take into account minor ups and downs in the road. Do you include every stone and pothole that you ride over? It's all guesswork/estimation.

Distance - yes that can theoretically be measured accurately, but remember your front wheel will be going further than your back wheel as it wiggles slightly more so there's a limit to that too.

Time. Yes time can be measured with great accuracy.

If it's accuracy (closeness to the "true" value) that you want then you must be able to define what that "true" value is. You can't do that for moving speed, or for elevation. If it's precision (confidence that measuring the same thing multiple times will give the same result) that you want, then just choose one method and stick with it. It should be precise and repeatable even if it doesn't give the same value as some other method.

Comparing multiple methods will just lead to disappointment and confusion. Like @Milkfloat 's man with two watches.
I agree :thumbsup:. The part I don't understand is that the journey was recorded with one device, yet when the one GPX (or whichever format it uses) uploads to Strava and Komoot, you get different readings. It is what it is I suppose :scratch:. I shall go with the Wahoo Companion App readings I think (unless there's a blip like the 146 km/h reading earlier) as they are direct from the device.

In terms of the sensors, I placed the Speed sensor on the back wheel hub as I had the same thought about the front wheel travelling further due to course corrections.

Oh, and I checked - apparently the Bolt has a barometric altimeter. You learn something new every day! :laugh:
 

Dogtrousers

Kilometre nibbler
I'm a RideWithGPS user. I occasionally upload rides to Strava and I use Garmin Connect, as a go-between, but my go-to source of info is RideWithGPS partly because all my rides are there but mainly because RWGPS is a bit more generous with its speed and elevation calculations, which in my view makes it righter! ;)
 
Top Bottom