Analysis of Bloor-Danforth and Yonge Night Buses: May 2021

From time to time, there are reports and photos in social media of crowding conditions on the two principal night bus services, 300 Bloor-Danforth and 320 Yonge. The TTC responds in its usual way saying that they monitor crowding and assign extra buses as needed, but they do not address a fundamental problem: buses on these routes run in packs with gaps that cause overloading. The situation is at least as bad, if not worse, than on daytime routes.

In past years, I have not been able to review night route performance because the old CIS tracking system fairly routinely went offline for a few hours most nights at about 3 am leaving a big gap in the data. The Vision system has far fewer outages, and gives a full view of how the service behaves.

The TTC now provides crowding data to some of its online service apps such as Rocketman, but this information is not yet available on an historical basis for review alongside the vehicle tracking info. Correlation of gaps and crowding must be done in real time, something that is not practical for month-long retrospectives. There is no announced date for crowding data to be available for research outside of the TTC.

As I have shown in other articles, headways might be within “standards” at the terminal, but they deteriorate as vehicles move along their routes. Moreover, the TTC averages data from all routes and time periods. The modest contribution of the night buses to overall “on time” performance is quite small. Nothing in the TTC’s methodology identifies problem routes, locations and time periods.

As with the daytime service, the cheapest form of additional capacity is well-managed service with vehicles arriving on regular headways spreading the load evenly.

Background Material

For a guide to reading these charts, please see Understanding TTC Service Analysis Charts: A Primer.

300 Bloor Night Bus

The 300 Bloor Night Bus is an extremely long route running from Kennedy Station to the Airport with two overlapping services:

  • 300A operates from Warden & Danforth to the Airport.
  • 300B operates from Kennedy Station to the West Mall.

Service operates between about 2 and 6 am, but on Sundays is extended to 8 am due to the later opening hours for the subway. Although the early hours mainly handle late night home-bound journeys, the route serves many work-bound employees whose jobs require travel before the subway opens. It is typically from these trips that photos of crowding have appeared.

As a starting point, here are charts of headways westbound and eastbound at Yonge Street in the middle week of May (the week before Victoria Day).

As the charts above show, there is a lot of bunching from just before 4 am onward. Of particular note is the purple line for Monday, May 17. This pattern shows up in each week’s data, and it arises from the fact that Monday morning is really Sunday evening from a service planning point of view, at least for the period up to 4 am

Plotted as quartiles in block-and-whisker format, the data look like this. There is a wide range of headways and especially the outlying upper and lower quartiles.

Yes, that 40 minute gap on May 17 at 4am really was there, and it can be directly traced to service management, not to any delay. Here is the service chart for May 17.

At about 2:30 there is already a problem with bunching of two eastbound buses (“lime green” and “dark blue”) which are joined by a “baby night bus” (“grey”) that runs from Bathurst to Coxwell passing “lime green” at Yonge Street. But the real problem comes after 3 am on the next page.

The pair of buses that came east together (“lime green” and “dark blue”) have a siesta at Kennedy Station. After about 10 minutes, “lime green” leaves only to be joined by another bus westbound on Danforth, and the two buses travel across the city together. The chart shows many other examples of buses traveling in packs sometimes because they leave terminals together, sometimes because two buses meet up where the 300A and 300B services merge.

For comparison, here is Monday, May 10, one week earlier. There is no big gap in service, although some bunching does occur.

Here is Monday, May 24 (which is a holiday). Here there is a wide gap because the 300A and 300B services did not mesh properly. There is also a lot of bunching, and some of this began right at the beginning of service, 2 am, and carried onto this chart at 3 am.

Finally, here is Monday May 31. By comparing these charts, one can easily see that service can vary a lot, and that gaps followed by two or three buses are not uncommon. This translates directly into poor service for riders.

Saturday morning service shows fewer large gaps, but headways are still spread over a range from 0 to 15 minutes.

Sundays show much more scattered headways as well as two days that are special cases.

  • Sunday May 9 there was a Vision outage from midnight until around 5 am as part of the changeover to new schedules system wide effective that day.
  • Monday May 24 is plotted with Sundays because it is a comparable day from a traffic point of view. However, subway service starts at 6 am on statutory holidays.
  • The lines swooping upward a product of Excel attempting to project the lines based on the area where there is data. They do not represent real data values.

Looking at the service inbound is comparable to what we see at Yonge Street.

Here is the service westbound at Danforth Road where the 300A and 300B services merge. From about 3:30 am onward, buses are running on short headways indicating bunched service.

The situation eastbound at The East Mall is similar with short headways common throughout the week’s data.

Looking at individual day’s service charts shows that the situations described here are common. Many buses get generous terminal layovers (although not all) suggesting that simply adding “recovery time” will not fix the problem. There is a need for the overlapped services to be managed so that they truly blend, but there is no evidence of any attempt to do this. This is a general problem on branching routes, but on overnight services it is critical.

320 Yonge Night Bus

The 320 Yonge Night Bus operates as a single continuous route from Steeles Loop (east of Yonge Street) to Queens Quay. Like the 300 Bloor-Danforth Night Bus, it runs between 2 and 6 am except on Sundays when service ends after 8 am because of the late subway opening.

On weekdays, service operates frequently until about 4 am at which point it drops back to every 15 minutes. Service builds up again in varying degrees depending on the day of the week, and can be quite frequent just before the subway opens.

As on Bloor-Danforth, service on Mondays (following Sunday) is less frequent than on other weekdays. Here are the range of headways northbound at Bloor Street.

Note that the headway lies at the 4-minute mark until just before 4 am and is subject to the usual problems of routes with shorter headways that service bunches and headways are spread over a range from 0 to 10 minutes. After 4 am, the scheduled service becomes less frequent, but there is still some spread in values with quite irregular headways.

Saturday morning service runs off the weekday schedules, and so the service pattern for Saturdays is comparable.

Sunday service is quite frequent late at night and after 6 am with a lot of bunching and gaps. As on weekdays, the period of less frequent scheduled service between 4 and 6 am has a wide range of headways.

Southbound service behaves similarly. The charts here show service at Eglinton in the middle of the route. Weekday and Saturday service is typical of routes with short headways.

As with the northbound service, there are some wide gaps during the 4-6 am period with buses as much as half an hour apart on Sunday mornings.

TTC measures schedule adherence at its terminals. The service southbound from Steeles is fairly well-behaved, although there are some wide gaps and bunching at about 5 am.

At Wellington northbound (just north of the south-end loop via Yonge, Queens Quay, Bay and Front), the headways are also in a narrow band until after 4 am.

As we have seen on other routes, headways might be within “standards” at the terminal, but they deteriorate as vehicles move along their routes. Moreover, the TTC averages data from all routes and time periods, and so the modest contribution of the night buses to overall “on time” performance is quite small. Nothing in the TTC’s methodology identifies problem routes, locations and time periods.

The service charts for 320 Yonge are necessarily different from those on 300 Bloor-Danforth because there is moe service on Yonge Street. Here is a weekday.

What is quite striking here is the abrupt transition to wider headways at 4 am. Accompanied by bunching, this can create some nasty gaps. The chart below is typical.

Sunday service shows the same pattern with wider headways and gaps after 4 am, but an extended period of frequent service from 6 am onward. As with other charts here, note that there is little sign of the effect of congestion or longer dwell times affecting travel time. The bunching is entirely due to how the line is operated. This is particularly troubling when a bunch originates at the terminals Steeles or Queens Quay.

9 thoughts on “Analysis of Bloor-Danforth and Yonge Night Buses: May 2021

  1. Just as with the day services, the TTC seems incapable of blending the two 300 branches evenly. This has been a problem for years.


  2. Some may ask why doesn’t the TTC operate the subway 24/7, like New York City?

    From Wikipedia:

    “Though the subway system operates on a 24-hour basis, during late night hours some of the designated routes do not run, run as a shorter route (often referred to as the ‘shuttle train’ version of its full-length counterpart) or run with a different stopping pattern. These are usually indicated by smaller, secondary route signage on station platforms. Because there is no nightly system shutdown for maintenance, tracks and stations must be maintained while the system is operating. This work sometimes necessitates service changes during midday, overnight hours, and weekends.

    When parts of lines are temporarily shut down for construction purposes, the transit authority can substitute free shuttle buses (using MTA Regional Bus Operations bus fleet) to replace the routes that would normally run on these lines. The Transit Authority announces planned service changes through its website, via placards that are posted on station and interior subway-car walls, and through its Twitter page.”

    Also, some of the subway routes in New York City have headways of up to 10 to 20+ minutes, or more (IE. 40 minutes?).


  3. I think the last five charts tracing the 320 Yonge on May 19 and May 23 demonstrate how operators affect the consistency of bus service. The performance of TTC service is documented and route supervisors are responsible for the service.

    We have all experienced waiting forever for a bus and had no way to prove to the TTC that service was bad. These charts can prove you waited too long and the delay was not because of traffic conditions.

    At first glance these charts are difficult to read. The chart shows the activity of every bus on the route during the shift. Each coloured line represents a different bus and the TTC can even identify its operator. One easy condition to spot is two buses running too close to each other, that is where two lines are close to each other for a long time. At the top and bottom of the chart you can see that some buses leave the terminal too close to each other, again the lines are too close together. The situation I’m most interested in is finding the long wait. That condition can be spotted where there is a big horizontal gap between two lines. That’s where you waited forever for the next bus to arrive. Ideally the lines should have even spacing among them. This not the case and the bunching affects service.

    The TTC has all this information available real time and can spot when drivers are tagging each other and surely can control the spacing between buses leaving the terminal. Route supervisors are responsible for making the buses run evenly spaced but there are no reports showing how well they do.

    How well the TTC executes these manageable activities should be the basis of TTC performance reports. We all know that bunched buses means poor service somewhere else on the line. If the TTC improved these manageable activities, service levels could be better and there would be less need for special lanes or reserve buses and other more expensive measures.

    Steve’s efforts have dealt with producing management performance reports to reflect our experience of service. We have all been late because we waited forever for a bus. The charts can prove your wait and for some cases the delay was caused by operator behaviour. The TTC will not have any part of this.

    Steve: Real time charts are available at Transsee. They are free for streetcar lines, but it is a paid service for the rest. As an example, here is a link to display the current status on 512 St. Clair.

    As to chart complexity, yes, the service charts can be messy, but they give an overview showing how all of the vehicles behave relative to each other. There are other ways to display the data that I have used in previous articles, and which are described in the linked article. This type of analysis is hardly new, but the TTC persists in using a simplistic set of metrics of “quality” that hide more information than they show.

    Finally, I make no claim to “inventing” this type of chart. The idea of plotting train movements in a distance-time chart dates to the 1840s and Charles Ibry. With the modern ability to track vehicles in real time, it is a small step to produce regular reports and analyses.

    I visited Vancouver shortly after SkyTrain opened in 1988, and they were producing charts of actual operations from the day they opened as a way to review how well line management strategies performed. Although SkyTrain is automated, designs of operating plans and decisions about dealing with unusual events are basic parts of line management.

    In 1984, the Streetcars for Toronto Committee undertook a review of service reliability on streetcar lines with volunteers recording locations of cars as they passed various points. This included charts in the familiar graphic format, although not as much fine detail because of the low number of observation points.

    In the early days of this blog, I wrote about the SFTC study. Service quality has been central to my activism for a very long time.


  4. Thank you for doing this Steve. Anyone who uses these night routes over the years as I have will encounter the bunching and gaps sooner than later. Even if your bus happens to arrive in a good amount of time, with the advent of tracking apps like Nextbus you can often see some bunching and gaps somewhere on the route.

    What always struck me is how the drivers don’t seem to care. They would often be the cause of the problem, or at least would do nothing to try and fix it. I appreciate the service they provide, especially on night routes, but their culture and attitude seems far more focused on the brotherhood and sisterhood of drivers than on the riders. It’s these drivers that become the route managers right? In years past I’ve seen supervisors during the day just standing there doing nothing as clear bunching roles through.

    My question to Steve and others is how much do you think this is purely a result of lack of management and planning, i.e. they don’t even know it’s a problem and so don’t do anything to correct it, or if the power and culture of the transit union plays a role in insulating operators from criticism and corrective action, making them somewhat unmanageable?

    To be clear I’m not an anti union person. I worked for one years ago. It’s all about balance and it seems to me like the culture TTC operators is often more focused on their fellow workers than the people they serve. It’s a tough job and they deserve to be properly compensated and protected, but riders also deserve a well managed service.

    Steve: There is a long history to this situation going back decades. Part of it arises from the transition from on-street to centralized management through computer consoles on a system (the original CIS tracking) that did not do a particularly good job of actually “tracking” vehicles, especially if they wandered off route or short turned. That wasn’t fixed for over a decade until GPS was retrofitted to the system. What had once been on-street route supervisors with a good sense of how their routes behaved and a working relationship with operators (including knowing the good and bad apples) morphed into a hands off system, compounded by the lack of an actual strategy for how service should be managed.

    There was a period when a major labour issue was the amount of unscheduled overtime because vehicles were habitually late due to poor scheduling. The contract settlement made it punitively expensive to keep an operator past their shift as an incentive to ensure they stayed on time. For a while, this produced a lot of short turning just to ensure that crew reliefs would work as scheduled.

    Then came the “no short turns” era of Rick Leary who, reacting to political pressure, attempted to drive the short turn count to zero by having schedules so padded that it was almost impossible to run late. This wastes vehicles most of the time and results in a lot of queuing at terminals. There are still short turns (I see them regularly in tracking data and occasionally encounter them in my travels), but they simply are not reported. Leary gets his gold star.

    The metric he counts on is “on time departure” from terminals which as I have shown in other articles is (a) not well observed and (b) does nothing to preserve service quality along a route. In particular, branching routes have no enforced time points to ensure blended services inward from where each service merges in. These points are not “terminals” and so they are not monitored.

    When schedules have a lot of padding, an operator can take a long siesta at a terminal, including leaving late, then speed across a route to repeat the process. This is more easily done on bus routes than streetcars. In some cases, it is impossible to avoid running early because there simply is too much running time. TTC has finally started to prune schedules with the May and June 2021 updates.

    Management is rewarded for achieving a meaningless goal (and they don’t even do that particularly well), while operating staff shrug their shoulders and gripe legitimately about contradictory expectations from management. A good example, now inherited by the fare inspectors, was a desire to make sure the TTC collected every fare due to it, but at the same time to be kind and gentle to riders even if they were known scofflaws. This inconsistency flowed all the way from the top with the hawks on the TTC board winning the day, except when a child was denied passage for want of a fare, at which point the operator should have “used their discretion”.

    Over many years, “the way things are done” becomes accepted practice, and it is even difficult to impose new rules and discipline. That’s where we sit today.

    Having said all of this, there appears to be no interest in senior management to address the problem beyond figuring out how to make the stats look good. Actually providing more reliable service is not a priority. If the TTC Board can pull its attention away from Covid recovery, they might ask key questions about the difference between service quality reports from managament and complaints from riders.

    Liked by 1 person

  5. Thanks for the detailed reply. I’m definitely with you on the bad stats the TTC board is using and the emphasis being on “juking the stats” (to steal a term from The Wire) rather than the actual service experience.

    I think the TTC board may use the fact that our bus network is probably one of the best in Canada and the US due to our high service levels and coverage to ignore complaints and criticisms towards how the system is actually managed. I saw you mentioned the SkyTrain using operations charts like you posted here. I wonder how the bus management compares in Vancouver, or Montreal.


  6. On the B-D time chart, what is supposed to happen where the lime-green bus meets up with the other bus at Victoria Park?

    I assume the bus at V.P. is leaving on time but the longer bus is late.

    Steve: The bus leaving from VP has just had a long siesta, and if it left earlier, would not be running as a pair with “lime green” coming down from Kennedy Station. The two of them ran across the city together in a huge gap.

    Liked by 1 person

  7. What is the point of route supervisors? what the hell is that guy with a clipboard doing? why in 2021 does he even have a clip board and not a tablet? why is it necessary for route supervisors to drive around in full sized SUV’s? Why is management so tone deaf to day to day real world conditions of their ineptatude? And why does the Board not take some seious action and do some top down house cleaning. starting with that useless twit of a CEO.

    Liked by 1 person

  8. Surely in this age of apps and tracking technology and GPS and AI, the TTC ought to be able to control vehicle movement to the precise minute, on a STOP-BY-STOP basis in order to provide predictable service?

    I shudder when I think about the cold and windy Tuesday afternoon in late November when I stood there with dozens of other shivering potential passengers at the northbound streetcar stop at Spadina and Harbord. After literally half an hour, six streetcars arrived, all going southbound!!! And the Wellesley bus? Not one in either direction the whole time!!! I decided to walk up to Spadina Station. And guess what? The Presto gates were not working at that entrance and I could not get in. Not until some one came along to exit. Escalators were off, too. What a fiasco.

    Liked by 1 person

  9. I can’t recall hearing a peep from Leary during the pandemic period. I doubt anyone would notice him gone if he were suddenly relieved of his duties, whatever they may be.


Comments are closed.