Martin, I hope you don't mind me commenting.

I'm curious to know how you know where the maximum gradient is (between VPs 14 & 15)? If you're seeing red and assuming that's where it is I think you may be mistaken. As with most things gradient related it can be a little more complicated.

Here are a couple of previous discussions that might help:

https://cycle.travel/post/4595 & https://cycle.travel/post/5728

In essence, the shorter the route the more accurate the elevation calculations will be (steep mountain passes, bridges & tunnels excluded) and the colouring of gradients is the display of an average over a set distance, more of an indication than a declaration -  a short section with a gradient of 20% may not be as obvious as a longer section at 7% (although the max gradient should still show up at 20%)

Your examples are quite interesting, especially the section from the bridge at Shaldon heading to VP 15/12 at Buckland.

In your original route (Luton) there is lots of red and it's spiky! In Luton 1 there's no colour at all and seems much more benign. If I isolate that section I get something between the two, the three 'spikes' of Luton reduced to one and two very, very gentle ascents. 

The takeaway, I suppose, is to treat gradient and climbing information as a guideline.

Good morning, Hobbes. I don't mind you commenting at all. I value your opinions.

Firstly, I find my maximum gradient by clicking where it says max 14%. It then draws a vertical line on the elevation profile. I think this is a feature that Richard introduced quite recently. However, I have just noticed that if I click it again the line moves.

I agree with you about the need to treat gradient calculations cautiously and the effect that the route length has on them. However, in my examples, removing the three via points only decreases a 34 mile route by 5 miles.

As you point out, it is interesting the way that much of the spiky red disappears. I can't think of a reason why it should do this.

Hi Martin,

Firstly, I find my maximum gradient by clicking where it says max 14%. It then draws a vertical line on the elevation profile. I think this is a feature that Richard introduced quite recently

I didn't know this, so thanks for sharing. 

Multiple clicks don't change the location for me, unless I'm inaccurate with my mouse placement. Then the line moves with the mouse.   

Just focusing on the max gradient (something that I never do) it's interesting how it varies across all three versions of the route. 

In 'Luton' it is 14% at approx the 40 km mark (in the middle of my shortened version)

In 'Luton 1' it is 8.7% at approx Viapoint 3 (despite the previous max gradient location being on this route too).

In my 'cutout' route it is 10.3% and not in the same place as the 'Luton' route says. 

Having written all that, maximum gradient is very low down on my priorities.

I think that my interest in maximum gradient has been directly proportional to my age. When I went past 65 I began to realise that anything over 15% was more trouble than it was worth. Legs are not what they used to be.

There’s a bug at the moment where it’s over-smoothing the elevation in some circumstances. I’ve just reverted to the previous version of the code for now, and will see if I can work out what was going on!

@ Richard, that's looking much more consistent now. It frazzles my mind if I think about gradient too deeply so kudos to you. 

@ Martin, I certainly didn't mean to be dismissive of gradient, just that in my experience, 'Max gradient' is just one snapshot out of millions on a route and can be misleading. A useful app that may be helpful to you is Osmand. It's a real PITA to master but it does give elevation data in a similar way to CT's surface info: X km @ 0-4%, Y km @ 4-8%, Z km @ 8-12% etc. which I find much more useful in determining the suitability of an unknown route (I tend to travel heavily loaded). That 1 km @ 16-20% is actually 100m here, 20m there, etc. However, it's routing is nowhere near as good as CT's. 

For most uses though, the colour coding on CT is great for a general understanding of the climbing on a route. 

It’s smoothing out all the tiny little bumps in the raw elevation data that’s the hard bit – without that, the total climb and max gradient would be vastly inaccurate. You can get a bit of a glimpse of what c.t does by dragging a via point and watching what the elevation graph does before you release it – the smoothing only happens when you let go.

@ Richard, as Hobbes says it is definitely more consistent. I saved the route, Luton, again and rode it this morning. Gradient seemed spot on.

@ Hobbes, I didn't think you were being dismissive. We have different priorities with regard to routes, GPS devices, and I suspect, lots of other things. It would be a drab world if we were all the same. Thanks for the suggestion of Osmand. I will take a look.

That’s great – pleased it’s working better.

Here’s the most extreme example of the need for elevation smoothing – following the Meuse valley downriver. Without smoothing the total climb is 1200m. With smoothing it’s 106m!