theclem54 wrote: "A service road is a road running parallel to a significant road for the purpose of serving access to properties. Both ends of the service road are connected to the more significant road, it is not intended for through traffic. Such a road should have the Parking Lot Road type when needed to discourage the Waze routing servers from using it to bypass a section of the adjoining road. Service Roads which also service access to other roads, should not be Parking Lot Road types."

What should we do with this then? Logan Road Service Road Permalink

Another report has come in for Logan Road Service Road. DS previously made Logan Rd Service Rd a street, but it seems a reporter has been sent down the service road again. Should we change LRSR to parking lot road as discussed here?

Can I have some more replies to this topic? I have an outstanding report that I am trying to get back to and would like to know what to do with Logan Road Service Road.

From what I have seen from the wiki, and from the previous attempt to prevent this issue, which was a street, I am in favour of changing Logan Road Service Road to Private Road.

There are no other branches or other streets that come off Logan Road Service Road and LRSR returns directly back to the main Logan Road.

iainhouse wrote:Having butted in once, I guess I can again.

The solution is definitely to change all the segments of Logan Service Road to PLR. As I detailed earlier in this thread, Waze is directing users onto LSR. When they ignore this and continue on SR95, the app cannot tell they have done so - their (presumably good) speed on SR95 therefore gets added to the historical data for LSR, making it appear faster than before and therefore even more likely to be offered as the best route.

Feel free to butt in. I am not concerned about what State or Country the respondent is from, as long as they are aware of the issue being discussed, which is the title of the thread.

This is in the Australia section, so comments from any experienced editors across Australia are certainly welcome. This issue of Logan Road Service Road and what to do with it might need some collective input from experienced editors about how to change it.

When I first added Logan Road Service Road to this thread of Don't trust the purple line, it was because drivers were being sent down the service road and then being put back on the main road, whilst speed data was being accumulated on the service road.

And that this issue was only appearing every now and again.

kaiglynn wrote:Garvin, is there any way to tell how close a segment has to be in relation to the speed data trace for it to be linked to that same segment?

This knowledge probably is buried somewhere in one of the wiki's, or is regarded as proprietary knowledge of waze and so is only known to the owners of waze. I am not sure. I will admit, just pure speculation on my part. Need a much more knowledgeable editor in this area than myself to answer the specifics.

If there is, would it be possible to spread two parallel roads far enough apart so that the data doesn't get linked to the wrong segment?

This is not always possible. Quite often, the lay of the land, or the road structure does not allow it. I am thinking of North Quay and the Riverside Expressway here as one example. But there are quite a lot of others where it would not be possible to move the 'lesser' road further away from the more major road because it would then put it on top of other roads/streets, or other buildings etc.

Would it then be one of those situations that the app doesn't necessarily have to completely mirror what things are in reality?

The waze map editor and the app never completely mirrors what 'things' are in reality. The purpose of the waze map editor and the app is to provide the correct TTS instructions to the driver. As editors, we adjust the map editor to get the correct TTS.

Seems like all feedback for Logan Road Service Road has come in, or at least all those wanting have their say have had enough time and there has been plenty of time for back and forth.

To let everyone know. I am going to change LRSR to parking road and see if this assists with sorting out this issue. Hopefully we will not see another report from this site.

https://www.waze.com/en-GB/editor/?env= ... ,244117227

We have this report, which is another service. Again called Logan Road Service Road, but this time at Underwood. The two segments are a little bit different in that they have another street branching off them, and also in this case Logan Road Service Road does continue further North along side Logan Road 'proper'.

From the discussion on here previous, seems like we should keep Logan Road Service Road as a street. Please have a look the purple line, the discussion in the report and my summary here and add your thoughts.

And ianhouse, feel free to barge in

Is there any commentary on this latest issue?

The reason for this is simple enough: 2 segments with the same endpoints. It's a very long-standing bug/feature in the Waze works. Validator and Toolbox will highlight it and it's in the Wiki here.

The problem is that the routing servers return a route that is basically a list of junction nodes. If you have 2 junction nodes connected by 2 (or more) single segments, then there are 2 (or more) possible routes between the 2 nodes and no way to determine which will be taken.

This leads to situations like you see here - where the purple line appears to have taken one route, but the driving instructions (which were presumably given in the client) indicate the other route.

The solution is simple enough: split the service road with an extra junction node. JNF will do this automatically if you use it to disable/enable turns on either end of the loop.

I would also advise you to change the segment type of the service road to PLR. If Waze directs a user to go via the service road, GPS resolution is nowhere near good enough to spot that the driver continued on the main road. As a consequence, the good speed on the main road will be recorded for the service road. That makes it appear faster and therefore more likely to be chosen next time. You get a vicious circle where the service road keeps getting offered and keeps getting improved speeds.

This won't affect the middle of longer routes (where higher road types are strongly preferred), but can affect short journeys or the start/end of longer routes.

GarvinGray wrote:Can I have some more replies to this topic? I have an outstanding report that I am trying to get back to and would like to know what to do with Logan Road Service Road.

Having butted in once, I guess I can again.

The solution is definitely to change all the segments of Logan Service Road to PLR. As I detailed earlier in this thread, Waze is directing users onto LSR. When they ignore this and continue on SR95, the app cannot tell they have done so - their (presumably good) speed on SR95 therefore gets added to the historical data for LSR, making it appear faster than before and therefore even more likely to be offered as the best route.

The reason for PLR is my general assumption that drivers use service roads to access and stop at destinations. PLR will suppress any traffic jams that might be caused by drivers stopping but leaving Waze running. Apart from that consideration, Private Road would work in exactly the same way.

As also noted earlier in this topic, this solution is fine here, since LSR does not provide access to any other streets. If it did, then the PLR change would affect routing to those streets.

kaiglynn wrote:if we made the service road a private road the routing would see a penalty of 1 once it tries to route onto the service road.

That's not quite correct. The routing engine adds a large penalty when leaving PLR/PR segments onto another road type. This means that a destination on a PLR will be routed to without penalty. Routing through PLR and back onto another road type will attract the penalty, so that type of route would be avoided.

On the original subject of this thread ("Don't trust the purple line!"), I notice that there is an extra junction node in the middle of LSR. I don't know if that's been added to avoid a 2-segment loop, but it's not needed here, since SR95 has plenty of junction nodes between one end of LSR and the other.

Hope that helps.