Total Ascent Difference

It seems a certainty that Garmin knows more about measuring this than you do.

That you think it would be easy doesn't mean much.

Nice of you to assume something incorrect just because you don't like the answer.

You are not answering the question that i've made you. Bro science?

More wild assumptions. 

How do you know your device isn't accurate?  It could be bang on and your friend's device could be a bit off.  Strava's and Garmin's elevation corrections cannot be taken as that accurate either due to the way the elevation models they use were created.

GPS is very accurate in the horizontal plane (latitude and longitude) but it is not very accurate in the vertical plane due to how it works.  Because of this consumer GPS device manufacturers started adding pressure sensors to provide more accurate elevation data.  Before this you just had to rely on GPS elevation.

These barometric pressure sensors certainly can have issues.  They are affected by the weather changing.  Calibration problems will cause issues.  General manufacturing and measurement tolerances will have an affect.

How much a particular device is out elevation data wise no-one can really say unless you have a known accurate reference.  You certainly could get a very accurate elevation model of where you ride to compare against but that would be very expensive.

I think most people are paying for devices that are good enough , not ones that are perfectly accurate.

You can easily get more accurate GPS receivers.  They are used in surveying but they are very expensive and you get the accuracy by post processing the recorded data to remove the errors recorded by nearby GPS base stations at known locations.

Consumer grade devices have limitations.  In terms of track recording Garmin devices are perfectly good enough for what people use them for.

Omg... for your first question, read de op.

Totally agree, the barometric pressure sensor is not a direct measurement of elevation to start with. There're quite a few factors which could affect the readings. I have done roughly 400 rides on my 820, and just about 200 on the 830. The elevation values between the two for the same rides is roughly 5%, with the 830 recording slightly higher numbers. Most importantly, neither the 820, nor the 830 had reported identical values for the same rides, the closest I've seen is about 2%.

Your numbers are most accurate.

"10% is not very close. And it's not 10% it's 13%. A LOT."

It's not 13%.

You can't include the "corrected" value. It's using a completely different methodology.

(1979 + 1836)/2 = 1908 (average)
1979 - 1836 = 143

143/1908 = 7.5% (total difference / average).

Each number is +/- 3.7% from the average.

====================

"when i correct elevation, total ascent now shows 2112 mt."

If this is the "correction" Garmin provides, Garmin provides it for units that don't have barometers.

That is, Garmin deems it as less accurate than the number using a barometer.