Having done Poppy Scotland long route sportive yesterday, which was brutal in that wind but very good route, I was amazed at the difference in height quoted on komoot route 5775 feet, and garmin connect planner at circa 6169ft versus what my Garmin fenix 6 watch recorded which was 7800 feet.
That is a very big difference in actual versus map planner - I have to say that I expected variation but that is a big difference and judging by how cream crackered everyone was in riders close to me I think its closer to the latter.
Any experience/ thoughts?
Could be that the Fenix is calculating ascent based on air pressure using its barometer. Was there a low pressure front moving in yesterday?
I've used the "correct elevation " option in Strava with my three Lezyne GPS computer recordings over the last ~6 years. Back in 2017 the downwards adjustment was sometimes massive from my Y10 Enhanced Super GPS, this became less variable with my Mega XL from '20 and so far my Super Pro seems to have even less adjustment (sometimes upwards, sometimes downwards).
I've yet to hear of a GPS unit with a barometer that doesn't get dodgy elevation numbers.
From my limited experience with Garmin Connect this year, their route planner gave me slightly higher elevation numbers compared to Strava.
As above ^^^, my Garmin Instinct watches (I have both the standard & solar ones) are normally way over and I use the Strava 'Correct Elevation' to sort it. I know its the watches as on a standard route like the SDW I know I cant be doing 3k ft more ascent than everybody else!
I've very much found that online planners give elevation about 1/4 to 1/3rd greater than the total elevation when I upload to Strava - don't think I've ever seen it go the other way. Barometric sounds like a possibility
It's a total mess, I've fed the same GPX file from my bike computer into four different tools (Strava, Polar, Garmin Connect and one other) and they all disagreed massively on the total ascent - like 50% difference!
I would assume that when you upload or transfer to a different app, the elevation data are recalculated based on location because recorded elevation data are often inaccurate (+/- 50m). The value you get will then depend on the sampling frequency and the elevation dataset that is used. I plan most of my rides in OSMaps - the OS quote a vertical accuracy of +/- 5m (for the OS locate app, which should be the same accuracy as OSMaps).
I’d say that Strava corrected to base map is reasonably consistent. So if you plan in Strava and up load to Strava then you generally get what you planned for.
Anything measured in the field is basically wrong. gps is less good elevation at a fundamental level.
Elevation data base map is quite coarse and a small route deviation of a few metres can be the difference between the top of a cliff, and the bottom (or a less extreme example, 20 metres down the hill). And contouring round a hill, you could be moving up and down within the 10 metre contour lines quite a lot.
Strava has an article on the challenges of this, I think they do stuff like take historic elevation data from devices that have barometers to improve the height data for popular routes and apply it to routes recorded by units without barometers. However it's still limited by the accuracy of your unit and how often it records it's position.
Very interesting set of replies, I was assuming the watch measured against GPS vs barometer but see that I am wrong. It is eye opening on the variance of various platforms