hill gradient/percentage on strava

gubber12345

in the analysis of a ride how accurate would you say strava is when it comes to the percentage of hills/gradients...reason i ask is that around here there are some short but quite steep hills,but when i check them on my ride analysis on strava they only say its about half of what they are.i say this cos i've checked the hills/gradients on google earth and its double if not more on some of them.not that it really matters but just would like to know what your thots are.

got a garmin 500 as well how accurate is it as regards to the percentages of hills as well. :?:

i'm a bit of a geek for things like this as i just like to know :oops:

supermurph09

Not sure if I understood properly but Strava will give the average % of the climb not the highest %. User veloviewer to see the various % points on the climbs. Much better IMO.

Strith

Ave gradients seem roughly ok over a few hundred meters or more, more finite and it seems to be massively over, or all over the shop.

I've no idea how they calculate the elevation gain, but I did one ride in Asia last year which my garmin clocked as 3000m climbing, but when I clicked the 'recalculate elevation' or whatever it is, it came up with 9000m, which I know is miles over, literally.

I'd take it all with a pinch of salt to be honest.

AchillesLeftKnee

For longer climbs at a consistent gradient, a barometric altimeter such as the 500's got, will be very accurate indeed*. Where it falls apart a little is on climbs with small, steep bits. Garmin's "smart averaging" means it reduces the data churned out to a nice lil' manageable file, cutting out the detail. Consequently, if you're riding up, say, a 10% gradient for 50 seconds with 10 seconds of 20% in the middle of it, you're probably going to see that represented as a gradient of about 12% rather than the actual fine detail. If you turn off smart averaging (which can be done on the 800, no idea about the 500) then you'll get a lot more gory detail on the gradients, but the resultant files will be one hell of a lot bigger.

With respect to Strava, always look at your nice, barometric altitude data for the segment, not the segment's data. That segment may have been created using GPS-only altitude data, which is, without putting to fine a point on it, absolute horse feces.

* Assuming you climb it sufficiently fast for there not to be a dramatic weather change mid-climb.

Pross

All Strava can do is take whatever level information is available on their mapping data and calculate a gradient from that.

Mikey23

It's fascinating to watch the elevation percentage change on my 500 as I go up a hill. It's not always logical and sometimes its still going up when I'm going down.

Recalculate elevation is a bit weird too. When I'm on a flat course it gives me more and when on a hilly course it gives me less. And this is always different to what the same data does in garmin connect. I've given up worrying about it

Grill

Google Earth is less accurate than your Garmin. Found this:

The google earth elevation data uses a 90m SRTM grid. In between the grid points, it interpolates i.e guesses. It's very inaccurate around sudden vertical changes - like cliff faces and roads where there is cut-and-fill construction. So it's gradient calculation can be way off the real figure.

gubber12345

thanks for your thoughts chaps....basically all in all thats what i was thinking.