Elevation profile discrepancies

[livpoksoc]

Just been out on a ride I plotted out using Ridewithgps, loaded to my edge 200 as a course.

It's a simple 29 mile ride on country lanes. No huge climbs, some undulation but nothing major, yet when I review the ride there are some discrepancies between the devices/sites. Given that all the sites use Google Maps, I am unsure how this happens.

Ridewithgps = 1462ft
Strava = 1215ft
Garmin Connect = 924ft
Garmin Edge 200 device = 3658.1ft

Over a 30 mile route that's some big differences. Does anyone have an answer to this as I am not sure I can trust the Edge device, but even still Garmin's connect app hasn't taken the device at face value but still makes it 66% of both RWGPS and Strava.

Anyone got a clue why this may be?

 

[Milkfloat]

You certainly cannot trust the Edge 200 as that uses GPS altitude which is really poor quality. For the other 3, just because they overlay map data with google maps, it does not mean they use them for the DTM (digital terrain model). Without a true altimeter in your device you are reliant on third party data, which varies in quality considerably.

 

[Dave 123]

In April 8 of us went to Mallorca. 6 of us were using Garmins, either touring, 810,800 or 500. We were all doing the same rides, all were uploaded to Strava.
There were discrepancies there too. The best one was my brother Keith, on our ride to Formentor he apparently rode out several miles across the sea!

 

[vickster]

This gets asked quite often, some threads here might give an idea

https://www.cyclechat.net/search/7542592/?q=Elevation&o=relevance&c[title_only]=1&c[node]=55

 

[livpoksoc]

You certainly cannot trust the Edge 200 as that uses GPS altitude which is really poor quality. For the other 3, just because they overlay map data with google maps, it does not mean they use them for the DTM (digital terrain model). Without a true altimeter in your device you are reliant on third party data, which varies in quality considerably.

In April 8 of us went to Mallorca. 6 of us were using Garmins, either touring, 810,800 or 500. We were all doing the same rides, all were uploaded to Strava.
There were discrepancies there too. The best one was my brother Keith, on our ride to Formentor he apparently rode out several miles across the sea!

This gets asked quite often, some threads here might give an idea

https://www.cyclechat.net/search/7542592/?q=Elevation&o=relevance&c[title_only]=1&c[node]=55

Thanks for the replies. Looks like Connect's correction will be the most accurate then.

Didn't realise it was such a known issue but interesting to read why. That might be the push I need to go for the 510/520.

 

[vickster]

Is it such an issue for you? The rest of the data is probably less than 100% accurate too. My phone and garmin 800 show speed and distance as being different too, who knows (or really cares) which is right?

That said, if you want a new gadget, go for it

 

[livpoksoc]

Is it such an issue for you? The rest of the data is probably less than 100% accurate too. My phone and garmin 800 show speed and distance as being different too, who knows (or really cares) which is right?

That said, if you want a new gadget, go for it

Not massively fussed, just concerned that my device said I was climbing twice what the other apps say & my biggest issue at the moment is getting over the hills, so I want a clearer idea of how much I am doing.

don't mind a 10% discrepancy but 100% is ridiculous.

 

[Bollo]

Thanks for the replies. Looks like Connect's correction will be the most accurate then.

Didn't realise it was such a known issue but interesting to read why. That might be the push I need to go for the 510/520.

Not a bad idea. Elevation is never completely without error but having a 'puter with a barometric altimeter makes it much more accurate.

I'm a big fan of the 520 - there's lots in a very small package.

 

[Dogtrousers]

A few options.
Get a unit with a barometric altimiter, but bear in mind these can go a bit bonkers if not properly calibrated. This will probably improve accuracy, but it won't help to make it consistent with other means of measurement.
Always use mapping data. RideWithGPS allows you to replace your own elevation data with map-derived data. This will be consistent.
Always compare like-with-like.

As long as you get consistency the absolute accuracy may not really matter, it depends why you're gathering it.

Have a think about why you want a measure of metres* climbed. Personally I like to see it because it enables me to say to myself "that was hillier than my ride to X last year" and so forth. I don't really care how many metres it was, as long as I can make relative comparisons. YMMV

*Or feet, or rods, poles or chains, or whatever you use.

 

[livpoksoc]

A few options.
Get a unit with a barometric altimiter, but bear in mind these can go a bit bonkers if not properly calibrated. This will probably improve accuracy, but it won't help to make it consistent with other means of measurement.
Always use mapping data. RideWithGPS allows you to replace your own elevation data with map-derived data. This will be consistent.
Always compare like-with-like.

As long as you get consistency the absolute accuracy may not really matter, it depends why you're gathering it.

Have a think about why you want a measure of metres* climbed. Personally I like to see it because it enables me to say to myself "that was hillier than my ride to X last year" and so forth. I don't really care how many metres it was, as long as I can make relative comparisons. YMMV

*Or feet, or rods, poles or chains, or whatever you use.

Thanks...how do I replace it with my own data on rwgps?

My need is based on wanting to get a training plan similar to the ration of the pru. IIRC it is c. 3500ft climbs, so I want to get 10 mile splits with 300ft climbs on average. It's probably a mental thing, but that's me.

 

[Dogtrousers]

Thanks...how do I replace it with my own data on rwgps?

RWGPS allows you to replace user-supplied elevations with elevations from maps (click edit, then click on "Is the elevation data bad? Check here to re-fetch elevations for this ride.") IIRC this typically ends up with less total metres climbed, and therefore something I don't do (give away those hard won metres? Never!) unless things have gone badly awry with my GPS (like the time it reported I was deeply below sea level for a while - probably a calibration issue).

 

[PK99]

My need is based on wanting to get a training plan similar to the ration of the pru. IIRC it is c. 3500ft climbs, so I want to get 10 mile splits with 300ft climbs on average. It's probably a mental thing, but that's me.

That pattern does not replicate the pru. The first 50 is very flat, then the hills kick in, 3 big one over 30 miles then a long flat 20 back to London with a small nasty kick at Wimbledon.

The mental thing you want is a long steep climb on tired legs. 10 mile splits @300ft is gently rolling.

 

[BSRU]

A silicone case on my Garmin made a difference.
Before buying one the same route would result in large variations in elevation information. After putting the case on they are consistently within a few metres of each other.

 

[livpoksoc]

That pattern does not replicate the pru. The first 50 is very flat, then the hills kick in, 3 big one over 30 miles then a long flat 20 back to London with a small nasty kick at Wimbledon.

The mental thing you want is a long steep climb on tired legs. 10 mile splits @300ft is gently rolling.

Fair point...I need to find something appropriate. Either way double the real elevation on my Edge is doing my bonce.

 

[Smokin Joe]

I.ve just ordered an I-GOTU GT-820 Pro Bluetooth Wireless GPS from an ebay seller, £24.99 posted. No idea how good it will be but it's cheap enough and gives me a bike computer without faffing around with sensors and wheel magnets.