Firstly, let me thank you again for a fantastic site. I'm back from a trip to Burgundy and again, CT proved to be very valuable to route me neatly thorugh the beautiful countryside, in very picturesque landscapes.
I have found a case, though, where CT made a choice which proved wrong (in my view).
In order to always improve your routing engine, let me share it with you.
I created a test route here.
The chosen small road goes up a very steep section (10% or even more on about10 meters) where another option would have averaged out the climb and avoided that effort. I reckon the avoidance of D158 Route St Honoré was based on assuming higher traffic. You probably do not have the traffic data for the Nièvre Département, but still, without knowing, I would think that the trade-off traffic vs steep climb would hardly seem worth it for such a small distance. If my assumption is right, the parameter to consider might be how much of the "busier" road is it worth avoiding when steepness (or distance) are concerned. In my humble opinion, I think not much. And that's assuming the road is busy, which it isn't (but that, CT may not be able to know).
Again, this is very minor. I still thought I'd share it because it might give you a real-life test case on a matter which is probably less trivial than it looks. In those situations, nothing replaces real examples.
Comments
It's very close between the two routes – just 2% difference in c.t’s score!
Yes, I think there’s probably some opportunity to adjust the elevation penalty slightly. Unfortunately, as you guessed, there’s no machine-readable traffic data for Nièvre (and no traffic data at all for the D158), so it defaults to an assumed value for the OpenStreetMap road types. I do have a long-term plan to address that but not yet…!
Thanks for your reply. Nothing urgent. Again, CT is already very good right now so I can defenitely wait for what you have coming LT.