The October tracking data from this route reveals just how bad the problem was, and shows that this is part of a constant problem on the route.
Buses run in convoys on 35 Jane on Saturdays, and to a lesser extent on Sundays, for hours on end producing extraordinarily irregular service. This would be bad enough in pre-covid times, but with crowding being such an issue in 2020, the TTC’s inattention to reliable service bears added responsibility.
This article reviews the route’s behaviour on October 17 in detail, and then turns to other weekends to see how common the situation might be.
A challenge both for someone like me who wants to analyze service “as operated” as well as for riders, and even for the TTC, is to track the hundreds of “Run as Directed” (or “RAD”) vehicles (mainly buses) in the system.
During the Covid pandemic, the TTC opted to cut back on scheduled service on many routes and run unscheduled extras to be used when and where required depending on demand. Their name for this is “demand responsive” service.
The idea is good as far as it goes, but it has produced many problems:
Internally, most of the vehicles are associated with a “route 600”. More accurately, that is where the drivers’ crews are. However, when one of these buses goes into service there are problems.
The driver does not “sign on” to the route they are serving, and so a RAD bus running on Steeles West does not show up under the route 60 tracking data.
Because the bus is not associated with a scheduled run, NextBus does not know what to do with it, and the bus does not appear in projections of vehicle arrivals. A RAD bus might be just around the corner, but riders will not see it on transit service apps.
From a service tracking point of view, the vehicle hours are not charged against the route where the bus operates (which could be several in the course of a shift).
There is a further problem with the tracking data in that a bus “dead heading” to or from service (or even between RAD assignments) looks exactly the same as one that is serving passengers on a route.
The TTC is aware of these problems, and hopes to enhance its tracking system, VISION, to compensate. Current plans are for more regularly scheduled service to return by February 2021, subject to budget issues, and the number of RAD vehicles will decline.
I have been working with TTC IT folks to figure out a data extract from their VISION system that will allow visualization and reporting on the RAD buses. The charts in this article are a first step and are provided for those who are interested.
In many articles over several years, I have written about the quality of transit service in Toronto and the degree to which it varies from the sometimes sunny presentations by TTC management. Since the onset of the Covid-19 pandemic and heading into an extremely difficult budget year for 2021, understanding service from a rider’s perspective has become more important than ever to retain and rebuild demand on the transit system.
On the budget side, there are already harbingers of cuts to come. The TTC proposes to remove poor performing routes from the network and to trim hours of service on some routes. This includes the 14x Downtown Express services with their notoriously high cost/passenger and a few routes’ late evening operations. This is really small-scale stuff especially considering that the saving from cancelled express routes is already in place since Spring 2020.
The larger problem Toronto will face will be to decide what deeper trimming might look like, how candidates for cuts might be chosen, and how to evaluate the operation of what remains. There are already problems with erratic service that accentuates crowding problems coupled with an underutilized fleet of transit vehicles. Conversely, advocates for service retention and impovement, including, one hopes, TTC management, need solid ground to support calls for specific improvements and to measure them when they occur.
Management reports monthly on service quality and vehicle performance, but the metrics used fall far short of telling the whole story. Recently, CEO Rick Leary mentioned to the TTC Board that these metrics will be updated. This is worthwhile to the extent that new information is actually revealed, not simply a rehash of what we have already.
This article reviews the metrics now in the CEO Report and proposes updates both to the metrics and to the standards against which they report.
Broadly the areas covered here are:
Ridership and Trip Counts
Budget, Scheduled and Actual Service
On Time Performance and Service Reliability
Vehicle Reliability and Utilization
This is a long article because it covers many topics and I wanted to put the arguments together so that the way factors interact is clear. If you want to skip all the details, at least for your first read, there are consolidated recommendations at the end of the article.
Technical note: Many of the illustrations were taken from the October 2020 CEO’s Report. Although I have enlarged them for readability, their resolution is limited by the quality of the source document.
The Tyranny of Averages
Almost all TTC performance metrics consolidate data into monthly average values and, sometimes, into annual moving averages. While this approach simplifies presentation and shows long term trends, it hides a great deal of variation that is at least as important to quality measurement as the long term view.
As I have written many times:
Passengers do not ride average buses.
Telling riders that on average buses are not full and that their arrival is within standards is meaningless to someone who waits twice or more the scheduled headway (the time between vehicles) and finds a crowded bus when one shows up. This problem existed long before the pandemic, but crowding and the effect of service cuts combine to make it a greater concern than before.
Averaging in the performance of off-peak services such as evenings and weekends with overall route behaviour masks poor quality service. Conditions during busy periods are diluted by data from trips when demand on a route is lower.
Averaging performance across the network dilutes the behaviour on busy routes even further by including vehicles running with less crowding and better reliability.
In recent articles, I have reviewed a pervasive problem on the TTC’s network with uneven headways (the time between vehicles). This is annoying enough in “normal” times, but with the concerns about crowding during the pandemic, anything that contributes to crowding is more than just an annoyance.
The CEO’s Report presents only a few measures of service quality:
the proportion of service that is “on time” leaving terminals,
the number of short turns, and
the amount of service fielded compared to what is scheduled.
These metrics are utterly inadequate to showing how the system is behaving at the level seen day-by-day, hour-by-hour by riders for several reasons:
Most riders do not board vehicles at or near terminals, and measures that look only at terminal performance do not reflect most of each route.
“On time” is a meaningless concept for most routes because service is (or should be) frequent enough that riders do not time their arrivals at stops to meet their bus.
A count of short turns only indicates that most vehicles reached their terminals, not whether the absence of short-turns contributed to poor service quality. Moreover, there is good reason to believe that the number of short turns is under-reported.
Statistics are averaged on an all-day, all month basis and do not reflect route-by-route or hour-by-hour variations.
Riders to not ride “average” buses and streetcars – they take what shows up when they try to use the system. Indeed, if the TTC actually provided “average” service, a lot of problems would be solved because the service would be much more reliable.
At the recent TTC Board meeting, CEO Rick Leary noted that the metrics in his monthly report will be changed in the near future. This is long overdue, but it remains to be seen just how informative these will be.
This article includes displays of headway reliability in a new type of chart I have been working on recently. Reader feedback on this is welcome.
In a recent article, I reviewed the scheduled service for several routes and the potential problems associated with uneven headways and crowding. As noted in that article, several routes had their schedules modified in recent months to reduce or eliminate these problems.
In my review, I only looked at a subset of routes across the system. The TTC has provided a full list of the routes with new schedules and their plans for coming months.
We have been actively working to reduce uneven headways created by the cancellation of crews in the May 2020 board period.
In most cases, we utilized the built-in recovery time at the end of the route, and by shifting departure times, we were able to minimize significant gaps in the schedule. This approach allowed us to improve on most of the last-minute changes implemented in the spring while continuing to develop multiple service level scenarios through the summer and fall. Unfortunately, due to time constraints (working well past normal deadlines), full schedule re-writes were not possible and we continued to provide suboptimal headways on some routes across the system.
The shifting of schedules was implemented on the following routes in the September board: 7, 11, 25, 29, 32, 34, 51, 61, 63, 68 (gaps remain), 72 (southbound only), 74, 75, 79, 86, 88, 89 , 90, 98, 102 (northbound only), 105, 116, 124, 160 and 163.
We are currently working on the January boards. Most routes will return to pre-pandemic levels and scheduling anomalies resulting from cancelled crews and extra trippers will be eliminated.
Mark Mis Head, Service Planning & Scheduling
When the January updates are announced, I will publish the details as usual. Stay tuned.
This post arises from a Twitter thread in the early evening on Saturday, October 17. Photos of crowded TTC buses are not rare these days, and one appeared from the 35 Jane bus. Intrigued, I took a look at NextBus to see whether bunched service might be the culprit.
The first snapshot at 5:09 pm was what I found. All of the buses on the route were clustered in three groups with wide gaps between them. Oh well, I thought, maybe the route supervisor will sort this out at the terminals by spacing the service, or possibly by short turning a few buses. This was not to be.
Half an hour later, at 5:41, the only change was that the three groups of vehicles had rotated positions on the route, and this continued for the next hour. It was not until 8:15 pm that I came back to see what was happening, and found that almost all of the service was running southbound with an enormous gap northbound.
Regular readers here will remember that I wrote about route 70 O’Connor which had a chronic problem with buses bunching late on Saturday afternoons.
At the TTC Board meeting of December 12, a motion by Commissioner Brad Bradford asked staff to report back in the first quarter of 2020 on my findings. The pandemic emergency intervened, and there has never been a report.
The situation on 35 Jane on October 17 prompted me to look at vehicle tracking data for past months. Because this is a route with a proposed reserved lane similar to the BRT red lanes recently installed in the Eglinton-Kingston-Morningside corridor, I have been collecting data for it for several months.
In a series of articles, I reviewed the quality of service on many bus routes during a period, the lull in traffic and demand during the pandemic, when it should have been relatively easy for the TTC to operate reliable service.
A consistent factor on almost every route was that buses are running in bunches with wide gaps between them. Those gaps translate to crowded buses followed by lightly-used ones, and riders rightly complain about long waits and an uncertain arrival of the next group of vehicles.
The TTC argues that service is not really that bad because they have a large number of unscheduled extras (aka “RAD” or “Run As Directed”) buses that do not show up in vehicle tracking records. Leaving aside the obvious need to track all service, not just the scheduled buses, this does not explain why buses run so close together so much of the time. These are tracked vehicles that have a schedule that should keep them apart.
Or so one might think.
TTC Service Standards include provisions for headway quality (the reliability of spacing between vehicles), but this is fairly generous, and it is never reported on as an official metric of service quality.
However, another problem is that on some routes, the service is actually scheduled to come at uneven headways. This arises from three issues:
Some routes with more than one branch have different frequencies on each branch. This makes it impossible to “blend” service with, for example, alternating “A” and “B” destinations.
In response to the pandemic, the TTC quickly adapted schedules by cancelling all express buses, and selectively cancelling individual runs as a “quick fix” to avoid complete schedule rewrites across the system. Where local trips were cancelled, this created gaps in the scheduled service.
On many routes, notably those that formerly had express service, the TTC scheduled “trippers” to supplement the basic service. However, these trippers were generally not scheduled on a blended basis leaving riders with scheduled, but erratic service.
In some cases, the September and October schedules corrected some of these problems, but many persist. This article looks at a number of routes where the summer (August) schedules had uneven headways to see what, if anything, has changed by mid-October. (The most recent set of schedules went into effect on October 11, 2020.)
All of the data presented here were taken from the TTC’s schedules as they are published in GTFS (General Transit File Specification) format for use by travel planning apps. This almost exactly matches information on the TTC’s online schedule pages.
Updated October 12, 2020 at 6:20 am: Travel time charts originally published with this article included data for April 2018 for Saturdays as well as weekdays causing a dip in values at the end of that month. The charts have been corrected.
With the gradual appearance of red pavement, a reserved lane for buses is appearing on Eglinton Avenue East, along Kingston Road to Morningside, and up Morningside to Ellesmere. This lane has no physical constraints on use by other vehicles and therefore is not really a full BRT implementation, but rather “BRT-Lite”.
Like so much recently in our town, it has acquired a moniker “RapidTO”. I will leave it to readers to concoct a name for the really rapid service provided by the subway.
For reference, here is a portion of the TTC route map showing the affected area (click to enlarge).
Four routes leave Kennedy Station:
86 Scarborough local buses
116 Morningside local buses
905 Eglinton East express buses
986 Scarborough express buses
The reserved lane begins at Brimley and then continues along Kingston Road. At Guildwood, the 116 Morningside route splits off, but the other services continue on reserved lanes.
At Morningside, the reserved lanes turn north along with the 905 Eglinton East Express buses and the 116 Morningside locals which have come up from Guildwood. The various 86 and 986 Scarborough service continue east and north to their destinations without a reserved lane.
The reserved lanes end at Ellesmere near the University of Toronto Scarborough Campus (UTSC).
The lanes on Kingston Road and Morningside will also be used by the 12D service from Victoria Park Station to UTSC, but this is a peak-only operation that I have omitted from this discussion because it is so infrequent (roughly two buses per hour).
This has been announced with much fanfare, but it will be important to track what actually happens with service in this corridor. The TTC has an uncertain attitude to the benefits of reserved lanes, and this goes back years to (at least) the St. Clair right-of-way project.
Reserved lanes can reduce travel times by keeping motorists out of the way assuming, of course, that someone makes the effort to keep the bus lanes clear. The TTC uses this saving in two different ways:
Running the same number of buses on a shorter scheduled trip time means that more frequent service and capacity can be provided at no extra cost. This is the approach taken with route 116 Morningside.
Keeping headways (the time between buses) the same but reducing the trip time allows a route to be served with fewer vehicles while maintaining but not improving service. This is the tactic used for 86 Scarborough, offset by express service on the 986.
Over the years, there have been many service reductions on the TTC through the reverse of the second point above. Traffic congestion might become a problem, and TTC management wants to guarantee that fewer or no buses are short-turned. To achieve this, travel times are increased so almost all buses will arrive on time or early at terminals with a generous provision of recovery time before their next trip.
This has usually been implemented by running buses less often so that the trip time can be longer. The downside, however, is that scheduled service is reduced and this becomes a “new norm”. Service increases to deal with capacity problems are often held hostage to budget limitations.
Two other changes will happen on the Eglinton-Kingston-Morningside corridor concurrently with the new bus lanes (effective Sunday, October 11):
The 905 Eglinton East and 986 Scarborough express buses will resume to operation
Some stops will be moved or consolidated.
The service to be provided on the express routes is comparable to that provided in March 2020 before much of the express network shut down in response to the pandemic. This will substantially increase service for those who travel between express stops.
Express service stops:
Eastbound 905 buses stop only at Kennedy Station, Midland Avenue, Brimley Road, McCowan Road, Bellamy Road, Markham Road, Kingston Road and Eglinton Avenue East, Guildwood GO Station, Galloway Road, Lawrence Avenue, Morningside Avenue and Kingston Road, Ellesmere Road, Ellesmere Road and Military Trail, University of Toronto at Scarborough Loop.
Westbound 905 buses stop only at University of Toronto at Scarborough Loop, Military Trail and Pan Am Drive, Morningside Avenue and Ellesmere Road, Kingston Road, Kingston Road and Lawrence Avenue, Galloway Road, Guildwood GO Station, Eglinton Avenue, Eglinton Avenue East and Markham Road, Bellamy Road, McCowan Road, Brimley Road, Midland Avenue, Kennedy Station.
Eastbound 986 buses operate express from Kennedy Station to Celeste Drive (Guildwood GO Station), stopping only at Markham Road and Celeste Drive; 986 buses then operate local from Celeste Drive to Meadowvale Loop.
Westbound 986 buses operate local from Meadowvale Loop to Celeste Drive (Guildwood GO Station); 986 buses then operate express from Celeste Drive to Kennedy Station, stopping only at Markham Road and Kennedy Station.