Won't route there unless you're going to an address on the street. Also, once I drove it and got Waze to recalculate, it gave me a time of 2 minutes to get to the end of the block, so I think it IS adding time to the route due to all of the junctions. Actual time should be about 45 seconds.

I drove it again last night with the map zoomed out a bit and the car chugged along pausing at each junction. The junctions aren't visible, but once I got the routing to recalculate, I noticed that the purple line had a zig-zag look to it.

I guess I can go ahead and remove that now.

Haven't had a chance to get rid of it yet. I'll get a screen shot though.

andrewkjs wrote:This is great investigative work. Quick question though. We tend to segment our long highways into 1km sections or shorter in order to have traffic congestion appear earlier and more accurately show the highway section that's affected.
Would the advantages of this editing practice outweigh the disadvantages (i.e the effect on routing ETA)?

The general consensus is that a traffic jam between an entrance and an exit on a highway is a traffic jam, regardless of where on that segment it is and how long it is. Once you've gotten on that onramp, you're trapped. Navigation should avoid it either way if there's an alternate route, but I'd much rather see the road red at end of an onramp (navigation off), and decide to take another route than to get on the freeway and discover a red segment a mile (or 5 miles) later and be stuck in traffic.

Now that I've driven over the route several times, the routing doesn't seem to be an issue anymore. I guess it has gathered more speed data.

However, it still has a bad effect on the display on the client. The little car chugs along pausing at each junction. And with routing, still the zigzag display.

foxitrot wrote:The conclusion: a penalty was clearly there. But ETA was not increased by a second.

No, no increase in ETA. I had a weird result the first time I drove it, and it gave me several minutes to travel a couple of blocks, but I didn't see that again. May have been that the nodes killed all the speed data, and once I drove it, it had no problem calculating the correct ETA.

No. I posted that after trying to reproduce my earlier results in Livemap. See my later post. The Livemap goes back and forth between routing on it or not, but the client won't route on it at all.

EDIT: it seems the post with my final results are in another thread that someone started, but I'm not on the computer right now, so I can't find it. I'll cross post it here later so everyone can see it.

bgodette wrote:Did you start off with a entirely new drawn segment that you added nodes to, or did you use an existing segment?

Existing segment.

Either way, there's nothing good and a whole lot of bad that could potentially happen if you don't remove them. In any complicated system a good mantra is "anything that can go wrong will eventually go wrong".

iainhouse wrote:There's another effect of extra nodes mid- road.

If you decide to turn off the Waze route - for example because you want to go a different way - Waze will direct you to do a u-turn at every node until you've gone far enough along the new route that it agrees with you.

I didn't think that Waze directed a u-turn at a node. I thought that it only directed two left turns at the intersections of separated highways.