Route GuidesRoutes City GuidesCities Map Log in

route planner deviates from main road to return to it a few hundred meters later

18 Sep 2022
Find a better bike route. Try our map & route-planner »

Become a supporter

Hi all, Richard

I cycle a lot in Belgium and abroad with all kind of bikes 'm really a big fan of the website & routing!

However, there is one point that I find annoying enough to write out here :)

I even have a colleague who stopped using cycle.travel for his long distance routing because of it.

Sometimes the routing algorithm takes a main road, but then whenever there is a even a very short road alongside, it takes that one to return to the main road just a few hundred meters further. This complicates the ride unnecessarily, slows you down and you loose your right of way. Also, quite often the smaller road is in bad shape. It is especially annoying when you are going downhill, or when you have to cross the main road twice. In a hilly environment, the main road is also generally the least steep. 

I have this on almost all my longer routes that I plan, and you really have to zoom in to see those deviations, most often I only discover them when doing the tour. 

Example of today is this fragment, you need to cross the road twice, see picture,
(I'm not talking about the avoidance of the Kerkstraat which is in cobble stones so OK):.

https://cycle.travel/map?from=Tombeek&to=Terlanen&fromLL=50.7539068,4.5674104&toLL=50.7702123,4.6029216 

What I would suggest is to set some kind of minimum length to deviate from main route, and if possible, if going downhill or if you have to cross twice set that length even longer.

I hope something is possible here!

And many thx for all your efforts!

Dominique

Comments

Tue 27 Sep 2022, 12:27

Thanks!

I do agree it’s annoying and I’d like to fix it: there’s a similar example near where I live. Unfortunately it’s a hard one to tackle because the fast routing algorithm used by cycle.travel doesn’t really have any memory of where it’s been so far on the route, so it can’t think “oh, I was on that road earlier, I should stay on it”.

What cycle.travel does do is penalise turns to try and minimise unnecessary weaving. I’ve just done a bit of work on revising these penalties, and I think this might make a difference in this example. I’ll be running a map update in the next week or so and it’ll be interesting to see how it changes!

Wed 5 Oct 2022, 19:35

I do agree 100 pct on what is has been said by Dominique De Munck. When touring it is easy to face this problem at least a couple ot times everyday. Appart of being annoying, it creates two unsafe situations when the deviation is on the non-driving side of the main road.

On the adjoined example the deviation is less than 200 m.

For me this issue is the only drawback of c.t. For the rest, congratulations!

Mon 17 Oct 2022, 12:35

Hi Richard

I read on the newsletter that you modified some settings to lessen the 'weaving' issue. 

Are they already in place, because my example and your example seem to deliver the same detours as before on the website?

Another approach might be to run a "cleanup task" after the main routing has been done. Mimic the behavior I now do manually, follow the route, and if you return to the original road, skip the deviation. That way you keep the fast routing, but avoid things as this. 

Probably some work implementing this, but personally I would prefer this before Android app.

Thx

Dominique

Mon 17 Oct 2022, 12:51

They are but there’s still some fine-tuning to do. In particular the example you gave is almost there but not quite – the difference in cycle.travel’s scoring between the two routes is now only 0.1%!

I’m not sure if a cleanup task would be possible – after the route is generated, it doesn’t have access to the geometry that wasn’t chosen. So in this case although it might be possible to ‘clean up’ the detour onto Oude Abstraat, it would only be able to do so by drawing a straight line between the two points. It also wouldn’t know whether the road was avoided because of a genuine issue, e.g. a blockage or bad surface. But I’m continuing to give it some thought.

Tue 18 Oct 2022, 09:10

This is a very intersting topic . I also agree with previous comments.

What this tells me is that we, cyclists, must also refrain from blindly follow our devices' instructions without putting attention into the global picture: a real map.

Let me explain.

If I just let C.T. plan a route, then upload to my Garmin and follow turn by turn instructions, I will suffer from this annoyance.

If I use C.T. as a mobile Web App, I can display the map and suddenly these oddities become obvious and I can make my own decision to stay on the main road.

We have all heard (or experienced) these stories about folks following GPS instructions into stupid detours by car, simply because an error was made in setting the destination. Errors that can be avoided if one looks at the proposed route rather than just setting off thoughtlessly.

My strategy is to examin the proposed route before I start to memorize what comes ahead and most of the time I avoid these problems. No earlier than last sunday my Garmin took me onto a detour without any reason and because I had my phone tucked away I stubborny followed it. Had I used my common sense, and observed the situation on the ground, the Garmin error was obvious and I should have avoided it or at least checked on the map to see it before waisting time detouring.