I’m looking at a UR (permalink) raised today in London at 17:14, for which the suggested route fragment shows Waze routing via a right turn which is marked in the map data as being restricted for all vehicles between 0700-1900…
Livemap routing also seems to be ignoring TBRs right now, not just at this junction but also at a few others around London which I know for certain used to behave correctly in Livemap.
In the back of my mind I have a vague recollection about a suggestion that Livemap routing may not be working properly anyway, so I’m not clear on whether or not the Livemap testing I’ve done here is necessarily proof of anything beyond Livemap itself playing up. However, there’s clearly no denying that in the specific UR noted above, the in-app routing was ignoring the TBR at that one particular junction, even though I can’t see anything amiss with how the TBR is set up.
Also, over the last month or so I’ve been responding to a hell of a lot of URs raised by people who’ve been routed through taxi-only turns or along taxi-only segments. At first it just seemed like people were incorrectly setting the vehicle type in their app, but the consistent number of new URs being raised for this same issue is starting to make me think it can’t just be down to so many Wazers deliberately/accidentally altering that setting from its default value…
So is there a problem with how the routing server handles TBRs specifically, or is there also a problem with its handling of vehicle type restrictions as well?
There also seems to be some issue with the London CCZ restrictions today - planning several random journeys in the app (latest Android release) is inconsistent in showing whether or not a route into/through the CCZ is flagged as a toll route, and therefore whether or not the avoid tolls option in the app makes the slightest difference to the routes being offered…
Bank experimental traffic scheme affected as well… seems like every other entry in my feed today is a UR complaining about restrictions being ignored :shock:
My understanding has always been that in-app routing has no information about TBRs or closures and these are only available when the app is using the routing server. I’ve taken it for granted that, when a user has responded that they’ve found themselves not to be in taxi mode, their app has taken over routing due to poor data connection and hence been happy to route through an active TBR.
Fortuitously, the CCZ is not operational this week, but I understand that Bank Junction is. As Bank Junction is not available for routing any Waze-recognised vehicle Mon-Fri 0700-1900, my immediate conclusion has been that they’ve been subjected to in-app routing.
However, the number of users reporting routing through taxi-only turns has been, for some months, worryingly higher than might normally be expected from wrong vehicle-type selection.
Sorry, the terminology I used may be slightly misleading here - my references to “in-app” routing were just to differentiate from the route tests I was doing with Livemap, where I know there can be some differences of opinion, and weren’t in any way meant to imply the use of the offline router built into the app…
I’m aware of the limitations of this offline router, but presumably if that many users were having problems with connectivity back to the servers to get online routing data, wouldn’t they also be having problems posting the URs to tell us about it?
And in any case, the app-based testing I did myself with a solid network connection still showed restrictions being ignored, so the finger of suspicion has to point to it being server related somehow…
I’ve now come across a user apparently being routed through a private road to be able to make a turn which is otherwise not possible. It’s not apparent from the traces in the UR but the user has confirmed that’s what happened. Although it’s currently pointless doing so, I’ve added a buses-only restriction to the segment.
If this goes on for a long time, a lot of people are going to give up on Waze.
I’ve been routed over the Humber bridge today, which is also a toll road, even though I have the avoid toll roads option selected. Would it do that if it’s the only feasible option? I didn’t mind too much, the view was nice, but avoid toll does mean avoid toll…
If it genuinely is the only way to reach your destination, then yes. As I understand it, all (?) restrictions in Waze are just implemented as exceeding heavy routing penalties as opposed to an absolute block on routing no matter what, so under the right (or should that be wrong) circumstances, the restriction will be ignored…
It looks as though the problem may have gone away.
I’ve been doing some testing using the Routes Script in WME for the reported URs that highlighted the problem. I’ve also done a test of one route in the app which makes me think that restrictions appear to be respected again.
So far, yes, though I think I’d prefer to wait for some sort of official indication to filter down to us from HQ that the problem has indeed been addressed, before I go clearing off all the URs I’d been leaving open “for further feedback” whilst this problem was occurring - bad enough so many users had to suffer (in some cases repeatedly) crap routing over the past week or so, the last thing I’d then want to do is tell them everything’s back to normal again until we’re certain it actually is…
We had another response on the Direct Access thread:
I’m not at all happy with that response and I have said so. Sure, connection issues easily explain this type of problem. But that doesn’t go any way towards explaining why we’ve had such a massive increase in such problems.
I see speculation above that the problem might be resolved. I hope that is true. Next week, everyone is back to school and back to work. If the stats run like they did last year, I expect to see a big spike in Waze usage taking us beyond 40,000 users in London. I don’t want that to be accompanied by a big spike in reports of restrictions being ignored.
Please, if you have more examples of this type of report, particularly if they have happened very recently, post them here. I hope HQ will continue to monitor this topic and I will update again when I hear more.
It’s stuff like this which makes me wish HQ would stick to their oft repeated promises to be more open and communicative with us, because if all of their feedback on this issue had been out in the open for us all to see then
a) it’d have given us a clue that they were struggling to replicate it, in which case we could have tried capturing more data for them which would have proved it wasn’t just an offline routing issue in the app
and
b) if nothing else, it’d have least let us all know earlier on that they were even aware of the problem so that we weren’t left having to deal with all the URs without any idea what was going on…
I had one here reported on the 2nd in Leeds. This one wasn’t reported en-route, but at another location, so it may have been another time.
Another here in Bradford on the 3rd.
In both cases I have referred the reporter to this thread.
In the Direct Access Topic there was some to-ing & fro-ing about whether there was a problem. I’m not sure they are looking directly at this topic, so I posted a long list of permalinks directly there. This covered the large clusters of URs around Oxford Street and Bank Junction, and a few others I encountered whilst I was checking those.
This is the latest response:
Basically, to improve Waze routing when traffic is anomalous, Waze has a “holiday” mode. During the period of an expected holiday, Waze uses weekend traffic patterns as it’s base model during the week. With London having so many users, I would expect traffic estimates to be pretty good regardless of the base model being used. But it looks like, whilst using weekend traffic patterns, Waze also uses weekend restrictions! :o
Quite how this would work if there was a restriction for only Saturday or only Sunday, I don’t know. However this does explain all of the URs around Bank.
It doesn’t explain the URs around Oxford Street (some of which may be caused by users having taxi mode on), or the others I quoted: so I’ve asked again for those to be looked at.
It’s also been suggested that this behaviour with TBRs during holidays can be disabled for the UK - or that it’s a design flaw that can be rectified. It’s not quite clear which, but I’ve said it needs disabling and we’ll chase that.
In the meantime, with no holiday periods in action or upcoming, I would definitely be interested to see any new URs with routing through restrictions - anything from Monday 8th or later.
This tickles a memory of something similar happening a few years ago on a bank holiday Monday or Friday, where restrictions were mysteriously being ignored only to start working again the next day… So aside from Christmas week, do we now also know if this holiday mode applies in the UK on other public holidays as well?
As you say, London traffic levels are so high at all times, that it’s unlikely there’d be any real need to guesstimate what the holiday traffic might look like just based on weekend traffic (and given how the CCZ affects weekday traffic in London, I wonder just how accurate the predicted holiday week traffic model was anyway compared with the actual data that would have ordinarily been used…).
And since most of the restrictions we have in London aren’t AFAIK relaxed by default during holiday periods anyway, it’d be preferable for the servers not to be messing with the restrictions like this - better for the few which are lifted on certain holidays to remain in place and generate a handful of URs, than for the entire country to be hit with a sudden (and hitherto unknown) overriding of all restrictions to their weekend forms without us having any control over it.
Sure, connection issues easily explain this 