Garmin/Strava accuracy

surrey_commuter

Just been looking through approx 20 Ride london results and whilst distances are within 1% of each other the height variance is between 529 metres and 918 metres (both Garmin).

Was this caused by the weather or are height stats to be taken with a pinch of salt?

rower63

the expected GPS error in elevation is necessarily much higher than for ground-level positioning, enough to often make the elevation results worthless while the ground-level ones will still be fine.

leeefm

You can change the polling frequency on the device, but it probably won't give you much more (and will probably drain the battery more quickly).

meanredspider

Depends upon a lot of things:
Barometric or GPS elevation
If barometric, were there any weather fronts passing through
If GPS, how was line of sight to horizon

I've found GPS elevation is good where line of sight to horizon is good (top of a high hill) but not great otherwise.

I don't think Strava has elevation correction.

elbowloh

meanredspider wrote:

Depends upon a lot of things:
Barometric or GPS elevation
If barometric, were there any weather fronts passing through
If GPS, how was line of sight to horizon

I've found GPS elevation is good where line of sight to horizon is good (top of a high hill) but not great otherwise.

I don't think Strava has elevation correction.

Strava has a "elevation correction?" button under the total elevation figure for each ride/activity (on the full website).

I think it looks at everyone else who have rode similar segments and takes some sort of average.

meanredspider

elbowloh wrote:

meanredspider wrote:

Depends upon a lot of things:
Barometric or GPS elevation
If barometric, were there any weather fronts passing through
If GPS, how was line of sight to horizon

I've found GPS elevation is good where line of sight to horizon is good (top of a high hill) but not great otherwise.

I don't think Strava has elevation correction.

Strava has a "elevation correction?" button under the total elevation figure for each ride/activity (on the full website).

I think it looks at everyone else who have rode similar segments and takes some sort of average.

Ok - because there's a flat piece of road in Inverness that Strava has called Cat4 because of what I guess is iPhone data.

rower63

elbowloh wrote:

I think it looks at everyone else who have rode similar segments and takes some sort of average.

it's what they should do, they've got sufficient data for most segments, but I seem to recall reading that they don't. It's how satellites measure ocean surface level to within centimetres/millimetres from a grillion metres up - the power of averaging over many many readings.

bigmonka

I thought that Strava can pull the elevation data from other sources (Google maps data?) as I did a charity ride with a friend where we only had tracking on the horizontal GPS position but when we uploaded it to Strava it gave us elevation too.

davmaggs

Not all GPS units are identical. Differences in antenna for example will affect how easily the signal is picked up. Some people have cycled around using two devices, and found the differences.

I've found a dedicated GPS device like the old Garmin HCX outperforms a phone for example. Even between Garmin devices you'll find some that work under cover (trees, roofs etc) whilst others won't.

Ian.B

Strava's elevation data (via my android phone) has always been hopelessly inconsistent. On my dead flat commute, starting and finishing at river level, it shows an elevation gain of anything between 30m and 250m.

Am I alone in also thinking that the horizontal data is showing an increasingly wiggly, drunken and erratic route? It is not unusual these days to find that I have been riding down the middle of the river, or even recording a segment that I have not even ridden!

rower63

Ian.B wrote:

Am I alone in also thinking that the horizontal data is showing an increasingly wiggly, drunken and erratic route? It is not unusual these days to find that I have been riding down the middle of the river, or even recording a segment that I have not even ridden!

I think that's your phone's issue, not Strava's - Strava just accepts whatever GPS-recorded data the unit uploads. I have noticed, though, that whenever I see a really wiggly route, or a spurious KOM, more often than not it seems to be from an Android phone. (Arsey, you reading this? )

Ian.B

rower63 wrote:

Ian.B wrote:

Am I alone in also thinking that the horizontal data is showing an increasingly wiggly, drunken and erratic route? It is not unusual these days to find that I have been riding down the middle of the river, or even recording a segment that I have not even ridden!

I think that's your phone's issue, not Strava's - Strava just accepts whatever GPS-recorded data the unit uploads. I have noticed, though, that whenever I see a really wiggly route, or a spurious KOM, more often than not it seems to be from an Android phone. (Arsey, you reading this? )

OK, that's interesting to know.

Do you know if a phone's GPS recording might deteriorate over time for any reason? I may be imagining it, but I have the definite impression my commutes are more wiggly than they used to be.

rower63

Ian.B wrote:

rower63 wrote:

Ian.B wrote:

Am I alone in also thinking that the horizontal data is showing an increasingly wiggly, drunken and erratic route? It is not unusual these days to find that I have been riding down the middle of the river, or even recording a segment that I have not even ridden!

I think that's your phone's issue, not Strava's - Strava just accepts whatever GPS-recorded data the unit uploads. I have noticed, though, that whenever I see a really wiggly route, or a spurious KOM, more often than not it seems to be from an Android phone. (Arsey, you reading this? )

Do you know if a phone's GPS recording might deteriorate over time for any reason? I may be imagining it, but I have the definite impression my commutes are more wiggly than they used to be.

I don't know, I'll ask my electronics mate.
But, I do know that the quality of a GPS-determined position depends hugely on how unobstructed its "view of the sky" is, and how much of the sky it can "see", as the more satellites in clear line of sight, the better-quality positioning. That's why trees, tall buildings etc will cause corrupted signals. Have you changed your phone-carrying habits lately? For example, do you now also take, say, a laptop and always put the phone under it in your rucksack (from the pov of the sky)? Or carry the phone in a chest pocket which will cause your body to obstruct satellite signals?

Mikey23

I error corrected my 500 as it was doing a lot of stoopid auto pausing and it gave me 2401ft, which i believe was only about 20 out

Applespider

It's the max speeds on Strava that always make me smile.

For RL, apparently my max speed was 59.5mph

In terms of my commute, it's generally fairly good at giving the same elevation each day...regardless of where it is (on me or in my pannier)

surrey_commuter

Mikey23 wrote:

I error corrected my 500 as it was doing a lot of stoopid auto pausing and it gave me 2401ft, which i believe was only about 20 out

so the correct answer is 720 metres?