With the shift to the larger low-floor streetcars, the TTC will begin schedule changes on Sunday October 13 (Thanksgiving weekend) for route 506 Carlton. In many periods of operation, the headways will widen in recognition of the capacity of the new cars, although the change will also bring a capacity increase.
The schedules are designed on the basis that most service will be provided by Flexitys, especially in the off-peak period. This could lead to problems such as those seen on other routes where CLRVs operate on headways designed for larger cars and are badly overloaded as a result. A lot depends on there being enough new cars to fully populate the route.
On the schedules, the only CLRVs remaining in operation will be the 506 Carlton trippers and 511 Bathurst (which is planned to start conversion late in 2019).
The TTC Service Standards set design capacities for vehicles which in theory dictate the level of service a route received. The standard is more generous for off-peak than for peak service.
If service were replaced purely on the basis of scheduled capacity, then there would be about 60% as many Flexitys/hour as on a schedule for CLRVs.
The new service design (click to enlarge, or retrieve PDF version) is shown below.
The change in AM peak and midday headways is from 5’40” (340 seconds, 10.6 cars/hour) to 7’50” (470 seconds, 7.7 cars/hour), a decrease in cars per hour to 72% of current service. This will be partly offset by the larger capacity of the new cars plus the benefit of trippers scheduled during the height of the peak period.
The change in PM peak headways is from 6’00” (360 seconds, 10 cars/hour) to 8’30” (510 seconds, 7.1 cars/hour), a decrease in cars/hour of to 71% of current service.
This change comes on top of wider headways introduced in August in an attempt to improve route reliability by giving streetcars more running time and longer terminal layovers in a bid to cut down on short turns.
The capacity goes up, but the service frequency worsens. This will almost certainly be compounded by the TTC’s inability to maintain even spacing of vehicles as they move across the route. Even though they might leave their terminals somewhat “on time”, this situation quickly deteriorates into bunches and gaps.
There is always a challenge when the TTC substitutes larger vehicles on a route and widens the headway. Laissez-faire route management that might work tolerably on shorter headways falls apart when headways are wider. If two vehicles nominally ten minutes apart actually operate as a pair, then there is a twenty minute gap and the service might as well not exist at all. Equally, the short-turn stats might look wonderful because both cars have time to reach their terminal, but riders see the same gap they would have if one of the cars short-turned.
The TTC has a huge challenge in this regard, but shows little sign of trying to fix the problem preferring to assume that more running time and longer layovers will do the trick.
The charts below show the evolution of mid-August 2019 service at various points along the route. (The AM peak bus trippers are not included in these charts.)
Here is the service leaving Main Station. The headway values are clustered around the trend lines, although there are a number of gaps over 10 minutes.
By the time service reaches Coxwell, it looks like this. Even though this is only a few kilometres from the terminal, the headway values have started to spread out.
By Yonge Street, the situation is even worse with more cars in the 10-15 minute band, and correspondingly more on headways of only a few minutes.
When the service gets to Dufferin, the spread of headways continues with several at 1 minute or less.
Finally, we arrive at Parkside (High Park Loop), and the headway chart does not look good at all.
For the eastbound trip, the service leaving High Park is somewhat better organized than what arrived there westbound (the previous chart). This is due in part to the recovery time built into schedules so that some of the gapping/bunching seen above can be smoothed out. Note, however, there there are still many cars on very short and long headways.
When this service reaches Dufferin eastbound, the evolution of bunching is clear with more headways in the 0-3 minute range compared to what left High Park Loop.
By Yonge Street, the situation is roughly the same.
When the service gets to Coxwell, there is a wide range of headways, many well away from the trend line (which roughly matches the scheduled headway).
Finally, the service arriving at Main Station. The range of headways bears no relationship to the schedule and, as is typical for the outer ends of routes all over the system, wide gaps and bunching are common.
It’s at that slippery slope point for a lack of service to turn off riders. When you have a 20% chance of having to wait 12-15 minutes for a streetcar to show up in morning rush hour, that averages out to once per week. That essentially means you always have to plan on waiting 15 minutes for one to show up. That’s not the kind of service one wants to have in late January.
My impression is that implementation of the VISION system has improved on-time terminal dispatch. And by on-time, I mean vehicles at terminals are departing when prompted by VISION.
I wonder if VISION provides the TTC with better schedule adherence data by operator, route & run – – and also wonder if the TTC will bother to look at this data assuming it is available.
Steve: I have been doing some preliminary work with VISION data and the results vary a lot from line to line. One technical issue I have to deal with is that the VISION data are reported by stop, and do not include the ;point by point tracking of GPS location every 20 seconds the way that CIS data extracts do. This makes the resolution of position somewhat less precise. The TTC is working on moving to a continuous GPS extract for analytical purposes, but it’s not available yet. You can see some examples in this article. Note that it is a technical backgrounder written for people already familiar with the tracking technology and the types of analysis published over the years on this site.
I am working on an article about 41 Keele because there is a motion coming to the TTC Board about service quality on that route.
PM peak frequency will be higher on Sundays than weekdays?
Good grief …
Steve: This shows how the weekday service is constrained by the available fleet. The handful of CLRV trippers will not make up the difference.
Since I cannot make a comment about the VISION data on [that] post, I’ll make it here. Steve, you mention that the terminal behaviour for the 941 Keele Express isn’t as bad as other routes. The northern terminal stop times should be correct as the first/final stop is on street at Keele/Finch while the layover location is 200m north inside Finch West Station bus terminal. As buses always have to go beyond the final stop to finish the trip, I would assume this is why the VISION behaviour is much better for the 941 at the northern end.
Steve: Thanks for that detail. I also think the problem is related to the point where operators routinely take their buses out of service. I am hoping that when the full tracking data are available, we will see the vehicle behaviour even when it is “resting”.
I see the flexitys and CLRVs daily. It appears none of the streetcars show up on schedule and seemingly leave when the terminal has 1 or 2 cars filling up space. The flexity’s on the other hand seem to leave at a more “fixed schedule”. They usually grab up passengers with 1-5 minutes and are off. On a Sunday I saw 5 flexity 506 Streetcars bunched up at main street and not a single had left for 7 minutes. It was crazy and people who are not used to seeing flexity streetcars on 506 Carlton would just not stop looking.
I currently seem to find an issue with the Main Street streetcar terminal for flexitys. It can only hold 2 full flexity streetcars and the rest are sometimes left outside of the terminal inside main street waiting for the first 1 or 2 to leave.
Any plans of ttc to fix the main street streetcar terminal – my best guess is to create some tiny terminal on the turning point in front of 62 Mortimer’s bus bay and then add a designated walkway and buses should stop in front of it when a streetcar is dropping off passengers. Of course 62 Mortimers bus bay could be combined with the 23 Dawes bus bay. See my point?
Steve: If you want to see backlogs of Flexitys, go to Dundas West or Broadview Station where the fact that cars have too much running time is evident daily in the queues out on the street.
The subway uses stop-and-stay signals to gap the trains. Could this work for other vehicles? With GPS tracking the location of vehicles, periodic signals along the route could be programmed to break up clusters of vehicles by gapping the following vehicles. With some additional programming they should also be able to bring the route back into operation faster following a major holdup.
Steve: Actually, headway management has been turned off on Line 1 as you may have noticed from the irregular headways. It is still active, I think, on Line 2.
Gapping vehicles following a delay does not require high technology, only a route supervisor keeping an eye on the route and spacing service. Of course there are gaps all over the place with no “major holdup” simply because some operators drive to whatever “schedule” they please.
Steve, I appreciate that high tech isn’t needed for even vehicle spacing/proper route management; but surely Vision allows this to be done?
Real-time data should be used to instruct drivers, and/or notify Transit Control if needs be. There will never be enough on route supervisors to keep a direct eye on every route, on every shift.
Surely we should be making best use of technology in lieu of same. Using field supervisors and/or transit control to deal w/things AI can’t do, at least not well, not yet. (ie managing emergencies, diversions, etc.)
Steve: There are two issues here. First is the distinction between “on time” and headway-based management. TTC stubbornly looks at “on time performance” with a six-minute window, an utterly meaningless metric for routes with headways under 10 minutes, and especially for headways in the 6 minute or better range. Bunched vehicles can be “on time” and the KPI looks just fine, thanks, with a gold star for management. Second, TTC does not considered irregular service from a rider’s perspective and the implication that bunched service is just as bad as short turns because the gap is 2x the advertised headway. Having two or three buses show up every 20 minutes is not “service”, and this actually underutilizes vehicle capacity. However, again the average vehicle loads look just fine even if every other bus is half-empty.
Before we expect technology to “fix” things, we need to determine what we should be measuring and what our goal for “good service” really is.
Steve, you mentioned that headway management has been turned off on the Yonge Line, is there a specific reason for this? I’m finding it a bit absurd that they would do this on the busiest subway line.
Steve: This was done as part of the ATC conversion because trains are making better time under ATC (the effect of slow operators is reduced). If ATC were still on, there would be backlogs from the headway management points. That said, the TTC has still not modified the schedules to reflect the shorter running times and this causes terminal queuing especially at Finch which is not yet ATC.
In regards to the declining frequency of streetcar service. I’ve been noticing that operators are now warning passengers of the impending delays for getting into Broadview to which they suggest getting off the last stop before station and catching the 505 bus. To even the extreme case of an operator dumping passengers off in front of Broadview via the front doors and where the operator has stop traffic to let passengers get off (passengers becoming irate).
Steve: I live near Broadview Station and see this all the time. It is part of the problem with giving too much running time in the schedules so that short turn stats will go down. We are wasting a lot of vehicles and service on layovers, but management gets to show huge improvements in the KPI they are judged on. I plan to write an article looking at this problem on a system-wide basis later this fall because a lot of routes have received new, generously padded schedules.
I can’t help but sense that all this declining frequency and bustitions is all the doing of the current CEO. His disdain for streetcars and his legacy at MBTA seems to flowing into the TTC. I hope we don’t come to the point of having 15min frequencies on the subway like it occurs on Boston’s T. It’s an ominous period for streetcar service, the continual decline of frequency in the name of reliability, the bustitution and the lack of new streetcars for capacity are seemingly turning into the perfect reason for replacing the entire system for buses. Steve, do you ever feel that there might be a need to start rallying up support to strengthen the streetcar system with the current trends that are happening?
Steve: Yes, Leary was initially not pro-streetcar although he was rumoured to be grudging coming around simply because of the level of demand. The situation with Bombardier has not helped one bit. As for the system as a whole, the big problems are funding for large capital projects, the focus on subway expansion, and the provincial meddling in responsibilities and priorities. I don’t hold much hope until there is a change at Queen’s Park. and I dread transit’s future if the Tories win federally.
I think it’s interesting to compare these proposed frequencies, particularly with the offpeak numbers, to what was put forth in the first new streetcar implementation report from 2013 here. Check out pages 7 and 8 for the numbers.
Back in those days the idea was to run offpeak service with the new cars at the same frequencies as the old cars (1:1 replacement) but turns out that was a lie. We should have known it was too good to be true because we’re getting cuts after all just like when we moved from CLRV’s to ALRV’s.
Sorry I had to do a double take and in most periods it is actually 1:1 replacement but the pernicious effect of running time increases eats away at frequency. A few periods have actual reductions and the effect is even more pronounced.
Maybe an alternate solution would be to chop up the route into shorter, overlapping routes? With shorter routes, the difference between slow and fast drivers will be lessened since the speed differences won’t accumulate so much. You can have different vehicle frequencies on different parts of the route too, so you get some of the benefits of short-turning. Almost no one rides the full route anyway since it’s so long and slow that it’s usually faster to take the subway if you need to go long distances. I admit that it’s more confusing for riders if there are different streetcars plying the same route.
I just realized that one amusing side effect of splitting up a long streetcar route into shorter ones is that you can make use of loss aversion techniques. I’ve heard these psychological techniques are becoming increasingly used in the gaming industry and in marketing. Apparently, people really don’t like it when things are taken away from them. But if you reframe things as a bonus, they’re much happier with the result. So if 50% of buses are short-turned, riders on these short-turned become really angry because they feel like they’re losing something. But if all buses short-turn by default, but 50% of these buses are upgraded to run the full route, then riders feel like they’ve received a nice bonus. The end-result is (mostly) the same, but overall satisfaction is higher.
Steve: There are several problems with your scheme. The first is the assumption that the problem is “slow and fast drivers” when the real issue is a system-wide tendency for buses and streetcars to run off schedule in pairs or worse. This is as much a question of drivers having a flexible relationship with departure times not to mention the comparatively easy life of being the second or third bus in a pack.
Next there is the issue of overlapping travel demands and the fact that, for streetcars at least, turnback locations do not necessarily coincide with natural breaks in the route. There will inevitably be the need to overlap some services, and some riders who now have a one-seat journey will be forced to transfer. There is absolutely no guarantee that service in the overlap areas will be reliable as one can see on any existing route. 504 King is a good example with the A and B branches scheduled on the same headway but routinely not providing a properly blended service. Another example is the challenge of designing stopping patterns for express and local bus services to maximize the benefit of the split service. Those buses have set routes and do not short turn (as per your proposal), but riders whose trips cannot use express buses because they start or end at a local one are frustrated by buses they cannot use passing them by.
Probably the best counter-example I can think of to your scheme is the 54 Lawrence East bus which, despite stopping at Lawrence East Station where regular headways could be restored, continues on its way with bunches and gaps.
I agree that very long routes are troublesome, but they need better management and a corporate ethic that values properly spaced service along the route, not just at terminals.
I noted the Flexity cars weren’t running on the 506 this weekend much like how the ALRV/Flexity runs on the 511 were weekday only runs. I continue to question why they keep scheduling and running the old unreliable cars when they don’t have to if they are really as bad as TTC brass say they are. I counted 91 Flexity’s in service Sunday afternoon which leaves more than enough to run all Sunday service with new cars.
Steve: Yes it was quite strange to see all CLRVs on Carlton when there are tons of spare Flexitys.
How are the headways for ss early morning and ss afternoon on Sunday the same and yet there’s one-two more streetcars?
Steve: Different running times. Takes more cars to maintain the same headway on a longer trip time. 120 minutes in the morning, 140 minutes in the early evening, 130 minutes in the late evening.
I’m noticing the number of flexities in use on Carlton has recently been reduced although there are plenty of flexities at Leslie Barns. Is there an issue with operators not being flexity trained? Are we saving them for ???
Take the 506 every day. Capacity never a problem. Problem is frequency and unpredictability thereof. Have stood at 7 pm at Queen’s Park Station waiting 24 minutes for a car. Once I walked from there to Ossington with a streetcar never overtaking me. I already very often will take the 63 north to the subway and subway across to Yonge, then subway down to College. More dependable and often much faster than depending on the 506. With less headway, I (and I’m sure others) will depend even less on the 506, meaning ridership will decline and then the TTC will say, well no riders means we need less service. This has been the history of the 506 for decades.