Accuracy of Web based mapping tool climbing data

[Louis]

Does anyone know how accurately the various web mapping tools (say Map My Ride because that's the one I know best) calculate total altitude climbed for a given route? We have lots of short, steep rollers around here and I'm wondering how accurate those calculations are for that sort of ride profile.

Thanks

[rice rocket]

They all rely on Google's terrain mapping (dunno about Garmin). If their elevation profile isn't accurate, then everyone's data is skewed.


There's certain portions of Google maps where you can see "stitching errors", where two separate datasets don't line up. That said, the rolling hills are probably fine. If you're riding at the edge of cliffs, then it's probably not as accurate. Also dependent on your GPS data accuracy, of course.

[Peter B]

Not very accurate in my experience. Tends to be quite optimistic. Around here it treats a bridge crossing a canyon as having ridden down to the bottom and back up, even if I crossed up high.

Take it with a grain of salt unless you can validate by experience.

[rice rocket]

Ah, I never thought of that, or even checked for that.

[palincss]

They all rely on the google map interface, but I don't believe they all come up with the same numeric values. I've not experienced this personally, but I've heard from others that mapmyride and ridewithgps can have as much as 50% difference in total elevation for the same ride.

On-bike GPS data can vary wildly from ridewithgps as well. I led a ride a few weeks ago that, according to ridewithgps, had a 12% maximum grade. One of the riders had an on-bike GPS that recorded 19% for that hill.

[AngryScientist]

i've found them to be pretty inaccurate with the type of terrain you're talking about. over long distances and gradual elevation changes they are pretty good, but roads with lots of little "bumps", often these elevation changes are missed.

not sure of the nature of your question, but i just picked up a nice sepialized computer with altimeter based elevation tracking for days i dont want to take my garmin out, it works great and was only $90. if you're trying to track elevation, altimeter based (with temp compensation) is the only way to go.

[Onno]

Interesting question. I recently downloaded the Mapmyride app for my android phone, and tested in on a very hilly ride. It recorded 42 miles and almost 3000 feet of climbing, which didn't seem enough. The elevation profile of the ride said we never got above 9%, which was certainly wrong. Several of the hills we tackled on the ride are above 15% for stretches of several hundred yards. I entered the route on Ridewithgps manually. The exact same route on the latter site said we'd climbed over 4000 feet, and the elevation profile showed that the grade was over 15% at the points we knew this to be true.

So it certainly seems to me that Ridewithgps is a lot more accurate than mapmyride. My guess is that the latter site uses larger intervals between elevation points, thus averaging down grades, and losing detail in terms of total elevation. The difference is very striking, though--25%!

My conclusion: Use Ridewithgps!!!!

[Birddog]

I vaguely recall that one of the RidewithGPS dudes did indeed state that their
system used the most frequent sampling of alt available. He also said that it is still imperfect.

[palincss]

That's [i.e., to use ridewithgps rather than mapmyride] good advice regardless of the elevation discrepancy. Ridewithgps is a great site, and the developer is a member of this forum. He's been very responsive to user input, too. And of course, no annoying ads that obscure your view and make it impossible to work with the site, like mapmyride.

[rnhood]

I like the Ridewithgps user interface and the guys running it try to keep it up to date however, the elevation profiles are often optimistic, and at times considerably overstated. Bikemap.net provides more realistic elevation gain data however, the user interface is not as simple. But Bikemap.net probably provides the most accurate profile estimations of any of these mapping websites in my opinion.

[Onno]

I like the Ridewithgps user interface and the guys running it try to keep it up to date however, the elevation profiles are often optimistic, and at times considerably overstated. Bikemap.net provides more realistic elevation gain data however, the user interface is not as simple. But Bikemap.net probably provides the most accurate profile estimations of any of these mapping websites in my opinion.

I don't really understand this. If these route programs are sampling elevation points from map data (and what else can they do?), how can they overestimate elevation? The finer grained the sampling, the more accurate it will be, but it will still be an underestimate. The only way it can be "optimistic" is if the map data itself is wrong. Or am I missing something here?

[Louis]

Thanks guys.

not sure of the nature of your question, but i just picked up a nice sepialized computer with altimeter based elevation tracking for days i dont want to take my garmin out, it works great and was only $90. if you're trying to track elevation, altimeter based (with temp compensation) is the only way to go.

This is where I'm heading. I've recently been wondering what the elevation gain is on some of my rides and I'm not interested in using a fancy Garmin-type system, so I'm debating whether or not the map-based info is good enough or if I want/need an on-bike measurement using an altimeter-based device. A while back I tried out one of the early Garmins and because of all the trees (I think) it was essentially useless.

[fiamme red]

I was delighted when I found that Bike Route Toaster could give me elevation gain immediately after drawing a route. I was later disappointed when I found that it didn't always give the same number for the exact same route, and there were occasionally some obvious blips (like abrupt peaks or troughs where the road is flat).

[dave thompson]

Louis, you can climb? I'm stunned. :D

[Louis]

Louis, you can climb? I'm stunned. :D

Hey Old Man Thompson, any time you want to come out here for a ride just let me know... :p

[BillG]

The grades on Lincoln Gap are a lot worse than this seems to suggest:

http://ridewithgps.com/routes/173577

[Louis]

The grades on Lincoln Gap are a lot worse than this seems to suggest:

http://ridewithgps.com/routes/173577

That's one of the reasons I don't trust those map elevation calculations...

[Ti Designs]

If you're gonna talk accuracy of mapping and grades, you also want to calibrate your legs too. I have a test hill, it's called Prospect hill in Waltham. Some days I climb it in the 44/17 in the saddle, other days I struggle with the 44/19. I'm pretty sure the hill is the same...

[Louis]

Some days I climb it in the 44/17 in the saddle, other days I struggle with the 44/19. I'm pretty sure the hill is the same...

Depends on whether the moon is pulling you up or down the hill.

[jemoryl]

I've been wondering about the same thing. On a recent ride in the Catskills, my computer, a Mavic Alti (barometric altimeter) showed a total altitude gain of about 3000 ft. The route we took was from ridewithgps (a site I love) and there it showed a gain of 4300 ft. So there can be a big difference.

Eyballing the altitude profile makes me think the Mavic is closer to the mark.