The Queen Car is Fine as it Is, Thank You (Updated)

The TTC agenda for May 6 includes a report on the proposed restructuring of route 501 Queen in two optional ways:

  • Return to the old setup with a separate 507 Long Branch service west of Humber Loop, or
  • Operate a 507 service to Dundas West Station overlapping 501 Queen service between Humber and Roncesvalles.

Needless to say, TTC staff recommend against any changes.

The Humber option never made sense because:

  • The connection at Humber Loop is in an isolated location where passengers do not feel secure, especially at off hours, and
  • Short turns of 501 Queen cars at Sunnyside leave a gap in the route and cause highly unpredictable travel times for passengers trying to transfer between the 501 and 507 in either direction.

Since the construction of much new development on The Queensway, the provision of reliable service is more important than ever.

The Dundas West option is designed to address several problems, none of which are addressed by the TTC report:

  • Move the transfer point east from Humber Loop to Roncesvalles.
  • Provide additional service on The Queensway that is not affected by short-turns further east.
  • Provide additional service on Roncesvalles to partly compensate for the frequent short-turning of 504 King cars (once that route returns to its normal configuration).

The TTC report lists many favourable aspects of the 507 Dundas West option:

  • The restored 507 route would operate with CLRVs freeing up five ALRVs for use on the 504 King route.  This is the equivalent of adding 2.5 CLRVs worth of capacity to the 504 in the peak period, but no value is attached to this by the TTC.
  • Service would be improved between Humber and Roncesvalles, and between Queen and Bloor on Ronces.  This is stated relative to the scheduled service, but a good deal of this service never actually reaches the terminals especially when there are disruptions downtown.  Nearly 10,000 trips per day would benefit from these improvements.
  • A small number of trips between Lake Shore Blvd. and the Bloor/Dundas area would have one less transfer.
  • The TTC does not mention the benefit of reliable service on Lake Shore, the value this would have in reduced wait times for riders there, and increased riding that could result.

On the down side, the TTC claims:

  • About 2,500 trips per day that travel from points east of Roncesvalles to west of Humber Loop would be inconvenienced by this arrangement, an that about 300 rides per day would be lost due to this change.  When I first proposed the 507 Dundas West service, the scheme included improved service on the 508 specifically to preserve through trip options.  This was not included in the scheme the TTC reviewed.
  • There would be little benefit for the Neville-Humber service because the west end of the line is not a source of congestion and service disruption.  This is exactly the reason for splitting off the 507 service so that it can benefit from the relatively trouble-free environment.  Also, with the 501 service having shorter trips, operators would not be faced with interminable runs across the city before they get a break. 
  • The TTC claims that sharing the platform at Dundas West between the 504 and 507 services could cause delays similar to what happened when the Dundas and King routes shared a common track.  I beg to differ.  First, the 507 service is less frequent, and there is a runaround track available if a 507 is so early arriving at Dundas West that it would hold a King car from leaving.

After all of this, the report concludes that “overall, the change would make service better for customers”.

However, the TTC rejects the scheme because it is estimated to have a marginal cost of about $825k annually.  As mentioned above, they give no credit for the value of the additional capacity provided on the King line by reassignment of equipment nor for the ridership effect of more reliable service.  Indeed, the only ridership change the TTC cites is the potential loss of 300 rides per day from the loss of a transfer-free trip through Humber Loop.

In discussing the route’s history, the TTC states that the amalgamation in 1995 was intended to eliminate the transfer at Humber.  What they omit is that this change was made primarily to reduce operating costs.  The decline in service quality west of Humber has long been a complaint from the community, and it is a direct result of the restructuring coupled with general service cutbacks on streetcar routes through the 1990s.

The report includes a chart showing ridership declines on the route, and the TTC argues that the fall occurred before the routes were amalgamated due to declining employment.  The TTC neglects to mention a large cut in service on the Long Branch car in the early 1990s that drove away riding even before the route’s amalgamation with the Queen car.

An important note about the Long Branch route, something evident in a previous TTC report on the subject and to anyone who rides the line, is that there is considerable local demand that never gets east of Humber Loop.  Service that is managed (and short turned) on the basis of somewhat empty vehicles at Roncesvalles will short-change riders who don’t board until west of Humber.  This is particularly so during off-peak periods.

The TTC report is self-serving with selective analysis intended to put their preferred option, do nothing, in the best possible light.

At a minimum, the TTC needs to carry out a trial operation of the 507 Dundas West option following restorarion of streetcar service on Roncesvalles late in 2010.  This trial needs to run long enough to allow meaningful analysis.  A related service change should be improvement of the 508 Lake Shore route to provide more through peak period trips to and from downtown via King Street.

49 thoughts on “The Queen Car is Fine as it Is, Thank You (Updated)

  1. @James Bow:

    (I assume) When people are saying the Flexitys are too big, they are referring to the size of them: the Flexitys will be 28m, whereas the CLRVs are 15m, and the ALRVs are 23m. 1 Flexity is more than twice as big as a CLRV. In theory, the TTC is replacing about 223 CLRV equivalents with about 408 CLRV equivalents.

    Also, I read the full report after I posted my first comment. Overall, it appears restoring the original 507 (Humber – Long Branch) would bring few benefits, if we look at the number of passengers impacted. As for the Long Branch – Dundas West Stn via Roncesvalles branch, the benefits are clear:

    “Service would be improved for customers who now use the stops between Humber Loop and Roncesvalles, as service would be more frequent. Approximately 1,800 customer-trips each day would benefit from this service change.

    Service would be improved for customers who now use the stops on Roncesvalles/Dundas between Dundas West Station and Queen Street, as service would be more frequent. Approximately 8,000 customer-trips each day would benefit from this service change.

    Service would be improved for customers who travel between Lake Shore Boulevard, The Queensway, and Roncesvalles Avenue/Dundas Street, and destinations on Bloor Street West east of Dundas Street, as they would have a new direct trip where they currently have to transfer. It is estimated that up to 350 such customer-trips each day might be made with one fewer transfer.” (The TTC claims analysis of Transportation Tomorrow Survey data indicates that very few people make such trips. Considering how many people use routes 123, 44, 110, 73, 66, 77, and 80, I beg to differ.)

    The downside is this:

    “Service would be made worse for customers who travel between stops east of Roncesvalles Avenue and stops west of Humber Loop, as they would have to transfer between streetcars where they now have a through trip. Approximately 2,500 customer-trips each day would be made with an additional transfer. Of this total, approximately 300 customer-trips each day are projected to be lost to the TTC, as these customers would stop using the service because of the additional transfer.”

    My thoughts? If those (supposed – this is a hypothetical analysis) 300 customers care more about a one-seat trip downtown on subpar service, than they do about more reliable service with a transfer, then something is wrong here. As much as the TTC’s points system can be useful in hypothetical scenarios, in practical application, it’s rather useless, especially if the scenarios don’t play out the way the TTC and riders would like it to work. For someone like me, who lives in northern Etobicoke, I would much rather have a shorten waiting time (weight of 1.5), even if it means an additional transfer (weight of 10).

    Even the TTC concludes that: “the change in weighted travel time shows that the benefits of a shorter wait and saved transfer is greater than the inconvenience of an additional transfer. Overall, the change would make service better for customers.”

    But they insist on finding a reason not to run the service. This reason is:

    “The Neville Park – Humber and Dundas West Station – Long Branch services would overlap on The Queensway, and the Dundas West Station – Long Branch service and the 504 KING route would overlap on Roncesvalles Avenue. These duplications of service add operating costs and reduce efficiency, and use up resources that could be better employed elsewhere on the TTC system where they would benefit more customers.”

    Except you (The TTC) aren’t using those resources elsewhere. In fact, you’re beinging to cut back on some of the improvements you made post-RGS by cutting services in tiny bits – 1 bus here, 3 buses there. (See the TTC’s service changes page for details.) To say you’d use them elsewhere when you don’t is plain lying. So much for the better way.


  2. Grzegorz’s comment above refers to the TTC’s “points system” (i.e. weighted travel time analysis), and that is the problem with this analysis. The weighted travel time calculation only considers four components:

    In-vehicle travel time (1:1 ratio)
    Waiting time (1.5:1 ratio)
    Walking time (2:1 ratio)
    Transfers (10-minute penalty)

    These variables are then fed into a ridership model based on service “as scheduled” (or with adjustments to account for planned modifications, but still “as scheduled”). They don’t account for things like service reliability, short turns, schedule adherence or headway fluctuation. If the primary goal of the Lake Shore changes is to improve reliability, it’s no wonder that the analysis didn’t find any benefit, because the analysis method doesn’t account for service reliability. (It does weight waiting time more heavily, which in theory could account for some level of uncertainty, but it doesn’t differentiate between waiting time for a route that runs like clockwork, and for the frustration that results from a route that is supposed to run every 5 minutes but hasn’t shown in 15.)

    Also, because the analysis doesn’t account for variation in headways, it likely understates the benefit in waiting times. The analysis is based on average headway, which is different from average waiting time. Here’s a simple example:

    Consider a bus route that operates at a 6-minute scheduled headway (10 buses/hour). Passengers arrive at a particular stop regularly, at a rate of one passenger every minute. If the route operates at even headways, each trip will board 6 passengers, and the average waiting time will be half the average headway (3 minutes).

    Now let’s look at the same route and ridership patterns, except that buses are bunching up so that two buses arrive every 12 minutes — specifically, the first bus arrives after a gap of 11 minutes, followed by another bus arriving one minute later, then repeated. The average headway doesn’t change (still 6 minutes). The first bus will board 11 passengers who have waited an average of 5.5 minutes each. The second bus will board 1 passenger who has waited an average of 0.5 minutes. The average waiting time here is 5.1 minutes ((11×5.5)+(1×0.5)/12). Or, looked at another way, the perceived average headway is 10’10” (double the average waiting time) — and the perceived average headway is 69% longer than the actual average headway.

    This is a simple example for analysis purposes, but you could easily come up with a “perceived” average headway using the GPS data (accounting for gaps/bunching and for short turns). The challenge would be in determining to what extent a revised service pattern would improve the distribution of headways, to come up with a perceived average headway under future conditions.


  3. The points system is a very simplified route weighting method, but deeply flawed, from a network topology perspective.

    Transfer penalties, in reality, are not a simple uniform value…they are fundamentally dependent on the frequency of the route you’re on and the one you are transferring to.

    For example, out in the suburbs, transferring from one infrequent 15-min headway bus route to another is a large transfer penalty.

    But transferring from a 5-min headway streetcar service to another 5-min headway streetcar service is a negligible penalty, as anyone who makes transfers in the downtown area on a subway or streetcar will tell you.

    I transfer between streetcar routes and downtown bus routes all the time and I don’t think twice, because I know the next streetcar or bus is coming within 5 minutes (except the 501 Queen, of course!).


  4. You know what the ironic thing is? I recall the TTC’s report that launched the combined 501-507 service, and they DIDN’T recommend a full amalgamation.

    What they considered and ultimately recommended was extending the 507 streetcar downtown via Queen, looping via Church, Richmond and Victoria, which is a restoration of a service the TTC operated until the early 1970s.

    The idea was that frequency on 501 Queen (Humber-Neville) would be reduced, and some reliability within the Queen service would be restored by relying on overlapping services through the downtown (507 Long Branch-Church; 501 Humber-Neville, 502 Victoria Park-McCaul).

    This didn’t happen. I have no idea why the TTC took this route, when their own report made no mention of it. This thing sort of came out of nowhere.

    Steve: The original scheme required more cars to operate compared to the unified route, and in the mid-90s, cut cut cut was the order of the day.


  5. Glen asks:

    “That has always been my major concern with transit expansion in Toronto. If the city cannot grow its employment base, why increase public transit? Increasingly the commuting flows are outwards and cannot be completed via PT.”

    A lot of south Etobicoke was developed in the 1950s. There are streets in Alderwood where it’s bulgalow, bungalow, bungalow, bungalow. Many of these bungalows still have their original owners. Obviously, they’ve retired now.

    Here’s the thing. As Toronto’s population ages, employment goes down. But older people do have places to go, and they often have problems driving. (My father is fine living by himself, but the government pulled his licence a few months ago.) It’s also quite possible that the elderly can’t really afford a car. (Wait ’till the post-defined-benefit workers “retire”. For me, it’s going to be cat food, and not the deluxe kind either.)

    In fact, there are a lot of older and mobility-impaired people along Lake Shore; more than on any other streetcar route I am familiar with. I think it makes sense to have good transit for these people. For physical reasons, they may not be able to walk long distances to a stop, nor to tolerate long waits in hot or cold weather. For financial reasons, a private vehicle or taxi could be out of the question. What does this leave them, other than public transit? Automatically going into retirement homes?


  6. Steve,

    Here is another thought how Queen and King cars can fixed:

    * Convert Queen and King to one way streets. Queen will run eastbound from Ronces to Bayview, King will run westbound. Two left lanes on both to be used for autos, the inner track as a transit ROW and the right curb lane as a bike route with bays for taxis and deliveries.

    * Next, split King and Queen into 5 FS routes:

    1) 501W will run all day from Long Branch along Queen and turn back at Parliament along King.
    2) 501E will run from Neville along Queen and proceed on King westbound to Shaw and turn back along Queen.
    3) 504W will run from Dundas West station to short turn at Parliament loop.
    4) 504E from Broadview along King and return along Queen after passing through the Shaw loop.
    5) During peak hours more cars (preferably ALRVs) will run on the Queen East – Parliament – King West – Shaw loop.

    Note that some traffic lights on King and Queen can be replaced with twin pedestrian crossovers and full priority signals can be given to all traffic (transit and auto in the downtown loop).

    You get reliable service on all sections of the routes, higher transit capacity in the core, better flow of traffic despite reduction in the number of lanes and a safe bike route.

    Steve: There are several issues with your proposal. First off, you have ignored parking completely. If removing parking is an option in general, we also need to think about how the streets would work today as two-way operations with no or with much less parking. From my experience with traffic proposals at City Hall, I would say that this aspect alone is at least as contentious as making the streets one-way. You also make assumptions about the availability of room for both a bike lane and taxi/delivery space. This would require narrowing of the sidewalk in almost all parts of Queen and King Streets, something that is not a good idea considering the volumes of pedestrians. Such a change would also require relocation of all hydro poles on the affected side of the street.

    There is also an accessibility issue in that this would put the entire area from Queen to Lake Shore more distant from eastbound service. People living south of King would have to walk up to Queen to get an eastbound streetcar. That’s 750m from the furthest point south (using Jameson which is densely populated), considerably longer than the (occasionally observed) walking distance criteria for transit routes. Also, the rail corridor east of Dufferin prevents a straight-line path north to Queen for people living in the developing Liberty Village area. There are similar issues further east on the route.

    New track would be required at Queen and Parliament for the east to south turns. Note that there is no “Parliament Loop”, and the site at King has been redeveloped. I presume you mean that the 504W would run east via Queen to Parliament and return west via King.

    The absence of a layover point for the downtown “loops” will not exactly endear you to the operators. This was a big issue with the split 501 operation.

    Any operation like this, given the considerable overlap of routes between Shaw and Parliament, will almost certainly require more cars than are now on the street. Although I agree that there should be more service, this has to be costed into any proposal.

    Finally, converting the streets to one-way operation will severely hinder the ability of service to divert around accidents, parades and other events. Queen and King now are used as a pair to spell off each other depending on where the problem lies. We have only recently seen streetcar service return to King west of Shaw following road and track construction. Under your scheme, the diversion via Queen would have been impossible. Similarly, an upcoming diversion of Queen cars around bridge construction at Dufferin would be impossible with a westbound only King Street.


  7. Good god … it’s bad enough having to walk all the way north to King to get a streetcar. And someone suggests that to go east, we’d have to walk all the way to Queen?

    I take it the proponent neither lives nor works downtown!


  8. Hi Steve,

    First of all, thanks for such a detailed reply. Agreed, there are issues and they need to be addressed. While I believe that some limited on-street parking can be preserved in Parkdale, Liberty and east of Parliament I believe that parked cars in the downtown core are a waste of limited space as there are alternatives for drivers and residents. First of all, there is no shortage of short term (and free) parking on residential streets west of Spadina which can be used for running errands downtown. Regarding the core itself, there are plenty of commercial parking facilities available for those who choose to drive. It should be noted that while lots of parking lots were removed due to intensification lately there were plenty of underground and structural parking lots built. A loss of a of a few hundred spaces that are not cheap anyway and can’t even be used during the rush hour won’t adversely affect availability since most parking lots are not full even at midday.

    Steve: While I agree that the use of arterial curb lanes for parking is a waste of limited space, I am not sure that your claim of alternatives is viable. Your one-way scheme extends well beyond downtown into areas where parking is not banned in both directions during peak periods. You assume that this is already the case. You also assume that space will be free on residential streets outside the core, but in fact many of these streets are already well-used by residents who do not have off-street parking available.

    While I did not elaborate on the curb lane usage due to lack of space, it never meant to be used as a continuous taxi/delivery space. Instead, one bay designed for six to eight vehicles per block would do to meet the needs of couriers or people that drove their car because they need to load a few bags of rice or a box of potatoes. Most of the curb lane will actually be narrowed to about 1.5 meters which will give enough space for bikes and the sidewalk can be widened for pedestrians, vendors, street performers, artwork, fountains, street furniture and everything else that makes the street vibrant. While the delivery bays might require a marginal widening of the roadway, the overall space on the sidewalks will increase and the streetcar will travel closer to the sidewalk which will improve both safety and accessibility. Finally, there will be a turning curb lane at the end of each block so right turning autos won’t block the through car lane and/or the streetcar lane.

    Steve: You still seem to be unaware of just how wide (or not) the streets are in various locations. Narrowing the sidewalk is not an option. You can’t just wave this away by implying that the average amount of sidewalk space will stay the same or grow if, in the process, you create pinch points. We have already seen the sort of problems this brings with the City’s street furniture program where the amount of walking space is severely constrained.

    Cycling lanes are tricky if they are not continuous, and designs such as on Roncesvalles are a compromise. Ronces was lucky to have a fairly wide existing sidewalk for much (not all) of its length and this gave some wiggle room.

    The only space for a turning lane at intersections is the existing curb lane, and that only works if there is an area clear of parking (or of a transit stop) far enough back from an intersection to provide a turning bay.

    While I agree that 750 meters is well above the accessibility target and will result in more than a ten minute walk for some people living in South Parkdale, Liberty village or East Bayfront this can be addressed by improving the feeder bus service. Lansdowne bus can be easily be extended further south and turned back along Jameson – and a large number of people living in apartments will benefit from service at their doorstep. Similarly, Ossington and Sherbourne services can be extended further south and the number of extra buses will not be significant.

    Steve: You omit the requirement for access to the north-south services and the extra waits inherent in the transfer. Also in some critical locations (notably Liberty Village) the rail corridor makes north-south travel on foot or by a feeder bus impossible. Extending Lansdowne south would require a loop via, say, Close southbound and Jameson northbound because Lansdowne ends at Queen. These are narrow streets, and Jameson is quite congested at times. The Ossington bus already runs south of King as does Sherbourne. You seem to be unaware of the existing route and street layout.

    As you mentioned, there is currently no Parliament Loop. It will need to be built. Additional connectors will need to be added in the core possibly along Simcoe and in Parkdale west of Dufferin. This will make diversions possible for construction, accidents and special events. For instance, on a Marathon Sunday morning streetcars that arrive from the west will be short turned along Simcoe. The vehicles arriving from the east will short turn along Church and a small number of shuttle buses will fill the gap. Finally, the outer track that is used at present for westbound service on Queen and eastbound service on King won’t be dismantled. Instead, crossover tracks will be built and streetcars will be able to bypass a disable vehicle in mixed traffic. Needless to say, the streetcar lane will be available to emergency vehicles at all times and police officers might even use it to divert auto traffic if necessary.

    Steve: Parliament Loop cannot be built because the property is no longer owned by the TTC and has a building on it. The Google Street View of the intersection is old enough that it shows this only as a vacant lot with piles of earth. As for turnback tracks on north-south streets, that’s easier said than done in some locations, particularly residential streets.

    There are no layover facilities downtown and it is not feasible to build one. Instead, drivers assigned to the downtown loop can be relieved every few hours. On street port-o-potties accessible only to TTC staff should take care of the rest and several can be installed along the loop.

    Steve: This won’t fly at all. Operators need to have a break more often than “every few hours”, and the idea of TTC loos scattered around the loop is really stretching things.

    While I agree that the total number of vehicles will rise if additional feeder buses are taken into consideration extra streetcars won’t probably be needed. First, the combined length of 501/504 will decrease due to the elimination of the two way service on each street. Second, vehicle speed will significantly increase as impediments to speed including mixed traffic, schedules, left turning cars, hostile traffic lights, single door loading will be removed. While theoretically speeds of 20km/h and above (similar to city centre trams in Europe) are not out of the question keeping distance between vehicles is much more important to provide ultra frequent and reliable service. Assuming speeds of 17km/h it would take between 30 to 35 minutes to traverse the loop between Shaw and Parliament therefore not more than twelve extra vehicle will be needed at the rush hour. I assumed fixed headways on all branches to simplify the headway-based route management (which can be done centrally using the CIS data).

    If implemented, the route will serve as an alternative to the western part of the proposed Queen Street subway or Waterfront LRT. It will provide capacity, frequency and reliability at a low cost. No megaprojects will be needed to serve this part of the city and money (if raised) can be used to build the Eastern part of the DRL from Yonge to Pape and possibly, even Eglinton and Don Mills.

    Steve: What you have ignored is that the combined existing demand of the King and Queen corridors must be carried in the peak direction, but because these routes are loops via the one-way pair, you have to provide the same level of service on both streets. In fact, the demand today on King is higher than on Queen, but you would have to provide the same capacity on both sides of the loops. Your assumptions about improved traffic speed are amusing when we consider that the Spadina car, with a reserved lane, has a much lower scheduled speed than what you expect to achieve through the dense central part of the route.

    As for this being an offset to a supposed megaproject to serve the area, no such project is on the books, and so you are not “saving” anything. Some comments on this site have advocated a Queen West subway and/or a “DRL West”, but I don’t agree that this is an appropriate response to the demand pattern that is evolving along the Queen and King corridors.

    Please note that this is the last extended reply I will make to this proposal.


Comments are closed.