I suspect this has been covered before but my search skills have failed me. Why is there such a big difference between estimated and recorded ascent on different apps/devices? Are any of them anywhere near accurate and if so which is likely to be more so?
Example. I planned a route today using my OS Maps account that suggested the elevation gain would be 885.82m. When I exported it to Garmin Connect to upload to my device it said it would be 750m. When I actually did the ride, my Edge Explore 2 recorded 1,106m of elevation gain which it sent to Connect and Strava. I'm guessing the Edge is the most accurate as it uses (I believe) a barometric altimeter and the others estimate based on map contours? Not that it matters particularly, but I'm curious why it varies by such a large amount.
I understood that the apps in particular really struggled when a route crossed contour lines at certain angles, although I find it hard to believe the issue hasn't been resolved in the 20 years since Memory Map came out.
I remember plotting a lap of Mull on the road bike and it came to something mad like 3500m elevation gain over 150km or something similar.
I only pay a small amount of attention to elevation figures when I plot routes on Strava, I assume they're all a bit off but hope they are 'consistently' off 🙄
My Garmin Instinct (my preferred method of ride recording) has always given a much inflated elevation figure following rides compared to the OS app or Strava routes. My Garmin 530 tends to be closer to OS & Strava. However, once uploaded to Strava I tend to just do the 'Correct Elevation' (desktop only) and it brings the elevation back in line with expected
I don't use OS myself, but someone I ride with does. We concluded OS in particular was massively out on elevation.
Lezyne GPS computers with barometers don't give as wacky elevation numbers as they used to, but I always hit "correct elevation" in the Strava app for Android, once uploaded.
I then out of habit click "correct distance" on the desktop webpage.
Think about it this way...
If the contours are at 10m intervals, it's relatively easy to keep tabs on how many times a route crosses the contour lines and sum all the changes.
However, if you are on a route with lots of small undulations (less than 10m) you aren't crossing the contour lines but a barometric device will capture this changes where's a course calculation won't.
So, I'd expect a route which consisted of a big, consistent up followed by a big consistent down to be more accurately calculated then an undulating route
If the contours are at 10m intervals, it's relatively easy to keep tabs on how many times a route crosses the contour lines and sum all the changes.
However, if you are on a route with lots of small undulations (less than 10m) you aren't crossing the contour lines but a barometric device will capture this changes where's a course calculation won't.
Makes sense and what I suspected. Ta 👍
I used to log elevation data on rides using GPS on both a phone and a barometric altimeter watch. I then uploaded the GPS data to Bikehike and looked at both the raw elevation stats and the “resampled” stats using the map data. The altimeter watch and resampled stats were generally consistent with an average error between the two of less than 2% across a large number of ride logs. The raw GPS data was always a long way out, typically overstating the elevation gain and loss by at least 3 times.
I’m pretty sure that the overstatement in the raw GPS data is basically a hardware problem not a software problem. The logging software is recording what it’s told very accurately but the GPS sensors are very poor at doing elevation. They are much better horizontally doing latitude and longitude. I believe this is because GPS works off angles and the angles for elevation are very obtuse so a small error in angle measured translates into a much bigger position error than lat and long for which the angles are more acute. Hence better navigation devices have barometers to measure elevation. Or something like that anyway. Maths!
Find a friend with a Coros.
Ask DC Rainmaker
My take is non of the devices give the correct answer
I always use the strava elevation data after upload. That is then consistent with a route plotted in strava. Os seems to be a touch less, komoot alot less
At a deep level the question “what was my elevation gain isn’t simple”. I think most of us wouldn’t count bumping over rocks as lots of short climbs. What about 1 metre high rollers?
@susepic thanks, that's quite a good explanation of how Strava does it and the factors they consider. Interesting that they claim to use data from the Strava community to refine their base elevation map and value data from barometric altimeters over GPS and other sources.
We generally consider the data from barometric altimeters to be of higher quality than the data derived from GPS signals and prefer that when processing activities.
Still, interesting that pre-ride their map thought my route yesterday was 750m and afterwards 1016m. It might be interesting to plot it again in a while to see if they've used my data to improve their map estimate? Or perhaps they need data from more sources before they change it? As I said in the OP, none of this is that important, I just find it interesting 🙂
At a deep level the question “what was my elevation gain isn’t simple”. I think most of us wouldn’t count bumping over rocks as lots of short climbs. What about 1 metre high rollers?
See also; how long is a coastline. 😂
My Fenix 7x can over read if I wear it on my left wrist without gloves, my theory is the sensor hole gets sweat in it and certain wrist positions blocks it or pushes sweat in, if you zoom in on elevation via Garmin connect, you can see split second dips down to 0 elevation, so if you are riding along at 300ft you get a sudden 300ft addition of gained elevation.
Wear it on my right wrist, or upside down on my left wrist it’s fine, I also calibrate the altimeter at the start of each ride, only takes a second, but it’s normally a bit out, or a lot out if I’ve been on a flight beforehand.
As Creaky alluded to, it's hard measuring change in height using GPS alone. Assuming GPS satellites are in a low maintenance orbit, that's 22,000 miles above the surface of the earth. 🙂
Not done a real comparison test for ages, but iirc it was something like a factor of 2 between the lowest and highest reported total ascent for the same ride recorded on the same barometric device and shared to multiple services. Some apply smoothing algorithms to get rid of all the accrued +1m / -1m stuff while riding along a flat road or canal towpath. One took the raw file and the total was exactly the somethingion of all the + and - from 1 data point to the next. Website and Android app or the same ride would show different values. And one made stuff up (I think Strava), and gave a total higher than the plain data.
Therefore Strava is the correct one 😉
Never did get round to leaving my Garmin in fixed spot for 6 hours, recording that "ride" to see how far it went and how much ascent/descent it did. I really should try with both the eTrex (with barometric) and Edge and compare.
Glad it's not just RidewithGPS then:
Surely anything using a barometric system is totally weather dependent.
If the air pressure stays pretty constant like the last few weeks then all good, but if a more typical weather front drives through then the air pressure can be all over the place. Then surely this will affect the perceived measurement of altitude
Have you calibrated the barometer on the device? That can also make a difference...however, I don't care if it is a larger number as it makes me feel good about my attempt at an effort.
- Maps account that suggested the elevation gain would be 885.82m
Surely if you ride a loop your elevation gain is zero?
Which one told you you enjoyed your ride the most?
Surely anything using a barometric system is totally weather dependent.
If the air pressure stays pretty constant like the last few weeks then all good, but if a more typical weather front drives through then the air pressure can be all over the place. Then surely this will affect the perceived measurement of altitude
In my experience, the error from air pressure fluctuations is much less than the error from other sources. You can always gauge it by checking the measured altitude at the beginning and end of a ride and I don't recall the difference ever being more than about 20m.
The signal from gps is weak. Less than the noise in the circuits . There are 2 problems with altitude. The first is 10m on the level isn’t much but it’s quite alot verticaly, well in the minds of cyclists. The other is it’s based on the difference in the time of arrival of signals. So it’s best with satellites when the difference is big. That’s easier horizontally than it is vertically. More of the satellites are low in the sky than high in the sky
