How is maximum gradient determined?
 

On the surface this seems like a simple question. With percent gradient being a simple calculation of rise/run *100% (or more precisely tangent of the slope angle *100%). However, how do popular cycling web applications such as mapmyride ridewithGPS and strava determine the maximum gradient of a ride? I recently did a 54 mile ride where the maximum gradient on ridewithGPS was listed as 7.2%. When I actually rode the route the steepest hill had me struggling to turn the pedals in 34-28. A fellow cyclist said the maximum grade of the hill was 16%. Looking back at a route that only includes the 1 mile long hill ridewithGPS lists the maximum grade as 13.2% mapmyride has the steepest section of the hill at 18% and the steepest part of the strava segment for that hill is 16.2%. So how do these websites calculate maximum grade and why does it vary so much between sites and why does the length of the ride change the maximum grade?

Tech answer for your GPS ? IDK .. Internet search: https://en.wikipedia.org/wiki/Grade_(slope),

Maximum would be using the "Lift"/ Elevator ; Straight Up.

The variation likely happens because of differences in the sample length. Imagine a long hill with an average 5° grade, but with a short section of 100 feet with a sudden rise and 10% grade.

If the grade is measured for only that short section the maximum grade reported will be that 10°, however depending how long each sample is and how that section was blended with others you'll get a lower number based on the blended 5 and 10° sections and how much of each type was included.

You realize that n * 100% = n right?

That being said...Strava operates on many different data points including topographical data and data sets supplied by various riders with an altimeter. The altimeter data, I would assume, is manually inspected and only on an area of concern. Say a very popular segment that is literally ridden by hundreds of riders and say 10-20 people flag the elevation data there -- a person may look at the data gathered for the segment and alter the elevation correction data. Let's be honest, topographical maps are not very accurate for a newly paved bike path, for example, but may be accurate enough when determining how much higher above sea-level a town is than the nearby river and/or drainage system.

That also being said, it is a very simple question.

Divide the rise by the run. Multiply this number by 100 (not 100%) and you have the percentage slope.

You can't change that calculation ... your question is related to where Strava obtains their data points -- which is a question for them specifically...because no one is going to answer that but them for a specific location.

The GPS sites aren't very accurate. They just use satellite data, which may not even be measuring the road. Overall grade is fairly easy, but the best way to measure max would be with a grade indicator on the bike. That way you're measuring across the immediate 40 inches, which is what you feel when riding.

The folks I know who talk about maximum gradient seem to do it by feel...

Very good call.

It's when I get off and Push the Bike up the Hill .

It is supposed to the the tangent of the angle * 100. So rise over run on the horizontal not the hypotenuse.
45 degree is 100% grade.
But...not everyone does it the way they are supposed to.
30% grade is arctan .3 = 16.7 degrees.

On a slight tangent, how would I find the grade for a particular road if I don't have a GPS or Strava? There's a hill at the end of my daily commute, and while I don't have to get off and push, I do have to stop a couple of times, 'cause that sucker is steep, and I'm curious about just how steep. Thanks.

Do you stop every 40 inches to write down your Rating of Percieved Gradient, in a little notebook?

You're right. Although at common grades, measuring the rise over the hypotenuse is 'close enough for government work.' The induced error is only 0.3% at 20 percent grade.

Yes, but when you are calculating the MAXIMUM grade over a multi mile climb, you have to pick a interval for intermediate measurements since you can't take the instantaneous grade at every point on the road. And most of the road I've climbed with > 10% grades usually are in crappy shape, so you can probably find different spots on the same parallel that have different grades.

Civil engineers probably figure the grade for some shorter segment like every 100 yards and then pick the steepest of those.

See if you can find the section map on the USGS site and then find the elevation contour where your street is located.

There's a lot of ways to do this.

By an inexpensive contractors sight level. Find your eye elevation and know your pace length. Find the start of the steepest pitch and sight ahead to a point on the road. That's run and your eye elevation is rise. If your eye elevation is 5' and you paced 100', that's 5%. 50' and that's 10%... , 25'..20%. I've found local roads (old subdivisions, military fort access roads, lighthouse access roads, logging roads etc.) that have short (say 200') heart pounding pitches of 15-18%.

https://www.amazon.com/Johnson-Level.../dp/B000KKWQ9S

Good point. I'd bet they just look up the US G Surveys. That is a guess, but no point in using the million devices to report what they think it is when your software - and the web APIs exist (I recall they exist) in the cloud.

There is some natural variability and errors with the GPS, especially with vertical recording. Heck, I've had flat segments show odd jumps in speed and location.

I have had some odd recordings when crossing a bridge, or riding along a steep slope to the side of the road/path, where the recording will seem to jump to another location on the slope, or drop down to river level below the bridge, so I wonder if the data is being augmented with map data.

Personally I think the Cycle/Run GPS sites should take all the heat maps, and regularly recalculate slopes and routes. Ride an area 100 or 1000 times, and they could even out the errors. However the systems seem to be stuck on a single ride. And in the case of a segment definition, the actual original recording of the segment.

I've noticed on a few segments that Strava also can't seem to recognise gross elevation changes.

Here is a local segment where Strava confuses people riding an overpass vs underpass. Also note the dip in the middle that doesn't exist. The climb is a steady climb.
https://www.strava.com/segments/1062935

Here is another segment with a switchback that someone else found that Strava confuses ascending and descending riders.
https://www.strava.com/segments/3267169

Here is a segment I periodically hit. A few rolling hills, nothing major. The map shows a maxim of a 61.3% grade at MP 2.8. I can guarantee I'm not hitting a 61.3% grade, unless perhaps it counts a pothole somewhere.
https://www.strava.com/segments/1450894

Go to the Strava site, do a segment explore and see if there is a segment defined for that street. Strava will show the gradient info.

If you want an accurate calculation of how steep a hill is. Take a level and tape measure with you.

Set the level to level, and measure straight down to the road at 90° from the level. Calculate rise/run.

(All numbers approximate.) So the run is 1250 ft., the rise is 130 ft. Dividing the rise by the run makes it a 10.4% grade??? That doesn't sound right.

Yes, your math is right.
Pretty steep, but doable. You said it was steep, right?

What isn't right about it?

Regardless of how it may feel or seem, if your input info is right, hat's the grade. Of course if either the rise or run numbers are off it's a case of GIGO.

Very steep. 10% just doesn't sound that bad. It's steeper than trail 3 in the community forest, and that sucker's steep, as well as 3/4 mi. long.