We might have 200' in some of the target area. I'll take a look later today.

At the narrowest point, the roads are now 62' apart. Both roads are at the right edge of the satellite picture. So we could probably get 100'. But not 150'.

Does Waze have any threshold for "oh, you're not really on the roadway I suggested!"

In other words, occasionally I'll take the 'other' roadway on the Tpke. Waze will put my car icon on the purple line, but the yellow GPS debug arrow will show my actual as where I really am.

Are the parallel roadways in Fort Lee set up such that spreading our segments won't overcome whatever Waze is looking for to decide it's not GPS error? Or was the situation I've described a fluke?

***
Confidence is the feeling you have before you fully understand the situation.

(Sent via mobile.)

If we're ok with the segments not being exactly over the middle of the roadway they represent, we can try dragging the express/local roadways as far apart from each other as we can (regarless of where the middle of the roadway is in the aerial) so that it does not cause map errors and/or unexpected re-routing, right?

It's also a process that would probably require ongoing monitoring of feedback, mostly through URs (or significant reduction thereof) and tuning as needed.

If you're looking for a scalar number of some sort, it doesn't exist. Or is not published. Rather, the spec is that Waze will recalculate your route once you've gone "reasonably far enough off course." My guess is that it's difficult to separate legitimate close-range reroutes from typically-experienced GPS noise, and that it's much more important to not get probably-unneeded, data-hogging recalculations due to random minor/typical GPS noise.

So, the official recommendation is to try pulling the parallel roads going in the same direction as far apart from each other as possible, such that traffic in the furthest lane in the same roadway does not get recalculated but traffic in the nearest lane of the other roadway would get recalculated.

So yeah, some fine tuning required. And yeah, if any of those unpublished parameters change, it may require fine re-tuning.

EDIT: Also a potential downside to this method is that by moving the segment away from the center of the roadway, you're also increasing the chance that random GPS noise/less-than-optimal reception will impact someone's trip by causing a reroute when one is not expected - in other words, the opposite extreme.

If it hasn't produced a meaningful effect on 287 in Bridgewater, it's likely not going to produce a meaningful effect here, or most other places along the turnpike, either.

I would just stick with the plan of educating reporters to always try to use the roadway they're directed to, or if they can't, then manually stop your trip in Waze and ask for a reroute - not much more difficult to do in Waze while driving down the freeway than submit a UR, considering your destination should be at the top of your recents list - and, just among us, it's much more productive than whining about bad directions.

Or, if they just plain don't like Waze's route selection and think they know better, then I would just ask why are they asking Waze for directions, then?

PhantomSoul wrote:If we're ok with the segments not being exactly over the middle of the roadway they represent, we can try dragging the express/local roadways as far apart from each other as we can (regarless of where the middle of the roadway is in the aerial) so that it does not cause map errors and/or unexpected re-routing, right?

It's also a process that would probably require ongoing monitoring of feedback, mostly through URs (or significant reduction thereof) and tuning as needed.

I am hesitant to try do this. We ran a brief trial on I-287 Bridgewater where the interstate forks into an inner and outer configuration back in June. Despite having dragged both parallel segments to the extreme edges of the roadways, multiple tests between skbenard1 and me did not yield any positive results. Every time we tried to force a GPS snap by taking the opposite side that Waze recommended, we were unable to force a recalculation.

My assumptions seem to be confirmed that any GPS snapping needs to be >150' as the maximum distance between the roadways was around 148'.

Unless we're heard directly from Waze or we have ≥200' that we can pull the segments apart, I would recommend not doing anything.

mrsmith66 wrote:At the narrowest point, the roads are now 62' apart. Both roads are at the right edge of the satellite picture. So we could probably get 100'. But not 150'.

Thanks for checking. I would recommend not doing this given the 20 tests skbenard and I ran on I-287. I would still love to hear what PS has to say given he had the direct conversation with a Waze Champ about the approach to pull segments apart.