Garmin 800 or Strava elevation

littledove44

My garmin 800 calculates total elevation barometrically. It says I did about 1500m climbing.

Strava elevation correction changes this to 2500m.

Which one is the least inaccurate?

ai_1

littledove44 wrote:

My garmin 800 calculates total elevation barometrically. It says I did about 1500m climbing.

Strava elevation correction changes this to 2500m.

Which one is the least inaccurate?

It depends....

No easy answer on this. Technically there is no correct answer. The result will depend on the resolution of the measurement. Infinitely fine resolution would give infinite elevation gain and loss.
This one you may have heard before in relation to coastlines. If you look up geographical info for a country you may find the length of the coastline stated. But this is really just a made up figure. If you measure in straight lines between headlands and add these up then you'll get one figure. If you mark points every 1km along the coast and measure between these you'll get a different, larger figure. If you reduce that to say 1m intervals you'll get a larger figure again and if you were to reduce the intervals to sub millimeter precision and measure around every grain of sand you'll get a different vastly larger figure. The smaller you go the bigger the distance.
So really if different systems use different measuring criteria they will give different but equally correct lengths.

I have no experience with Strava so can't comment on how it compares with the Garmin data. But it may not be a case of better or worse accuracy.

rolf_f

Ai_1 wrote:

This one you may have heard before in relation to coastlines. If you look up geographical info for a country you may find the length of the coastline stated. But this is really just a made up figure. If you measure in straight lines between headlands and add these up then you'll get one figure. If you mark points every 1km along the coast and measure between these you'll get a different, larger figure. If you reduce that to say 1m intervals you'll get a larger figure again and if you were to reduce the intervals to sub millimeter precision and measure around every grain of sand you'll get a different vastly larger figure. The smaller you go the bigger the distance.
So really if different systems use different measuring criteria they will give different but equally correct lengths.

You should really only see any significant difference if increased resolution of data picks up changes in gradient direction that are otherwise not picked up - eg a rise, followed by a short fall and then another rise. Effectively, increased data resolution does a better job of peaks and troughs and makes little difference.

Strava have some sort of fancy correction tool for computer GPS data and it is clearly complete nonsense. I use a Bryton computer. Bryton reckons a particular route had 5544ft of climbing, Strava converted that to 5315 feet (which doesn't seem to bad a discrepancy given the potential for barometric errors) and gave a friend from the same village, using a Garmin, 6030 feet. The same route plotted on the Bryton site (ie just using the digital terrain map with no barometric data) comes in at 5308 feet.

Of course, none of these figures is necessarily accurate. Strava say:

"Strava detects devices with barometric altimeters and recognizes the data from that source. Since elevation data derived from a GPS signal is fairly inaccurate, Strava automatically corrects elevation derived from a GPS source by consulting elevation databases to determine the elevation at each point in the activity".

This is clearly guff. The barometric data ought to pick up the fine detail as I said above - the detail can then be applied to the digital terrain info to create the final elevation. However, it obviously doesn't do this as there is no way that you can lose 700 feet in 6000 just down to that sort of fine detail. And it also seems hard to believe that the DTMs are so wildly different as to account for that sort of difference either. Except that plotting the same route on Bikeroutetoaster gives 4650 feet for the same route and Bikely 7340 feet. :roll:

But Strava should be using one digital terrain map - whatever that is, and however accurate or inaccurate it is, it should produce broadly similar elevations for the same route irrespective of what computer is used. For now, I'd say that the Strava elevation is fantasy but probably so is everyone elses. The problem with Strava is that it claims accuracy way beyond what it actually achieves. In my case, it just seems to mildly tweak what the computer calculated which isn't really achieving anything at all.

The best conclusion you can make is that you will never have the slightest clue how much you have climbed in absolute terms - you can only really judge it relative to other routes that you and your computer have done in the past.....

Additionally, Strava generously randomly gave a friend an extra ten miles on a route yesterday that they didn't ride. The cross section had the correct mileage on it, the ride time was correct but the total was 10 miles higher. Did wonders for their average speed which was significantly higher than everyone elses despite them being a slower climber who was occasionally waited for!

littledove44

Clearly science will not help, so I have decided to potentially delude myself by going with whichever gives me the biggest number.

Actually, my ride did feel a lot more like 2500m than 1500m, so that's what it is.

If there is a rule that says "if it isn't on Strava it didn't happen" then surely there should also be a rule that says "if it is on Strava it did happen".

rolf_f

littledove44 wrote:

Clearly science will not help, so I have decided to potentially delude myself by going with whichever gives me the biggest number.

There's no arguing with this sort of logic! If you are going to do something, do it properly.