Note: The charts in this article include a methodology problem. Short turn counts for vehicles crossing two screenlines (such as eastbound on Queen at Coxwell and at Woodbine) are distorted when these events do not occur in the same hour. This article has been replaced with a revised version, but I am leaving the original here for reference.

One of the many annoyances of trying to use transit service is to discover that your bus or streetcar has been “short turned”, that is to say, will not reach the destination advertised. This might happen before you board so that an arrival prediction turns out to be for a streetcar you can’t use, or as a “surprise” when the operator gets on the PA to announce that Transit Control wants to short turn the car.

This has been a problem for as long as I have been involved in advocacy for better transit service.

TTC Board members and Councillors hear about this problem a lot, and they in turn beat on management to eliminate the practice. This can produce unwanted side effects, notably the padding of schedules so that it is almost impossible, at least in theory, for a car to be late and, therefore, short turns should not be required.

Alas it is not quite that simple. Short turns occur for various reasons including schedule issues, crew change timing, major delays/diversions and “operational problems”, that catch-all phrase covering everything from a stuck door to a plague of locusts. (Some explanations for transit service problems have been with us so long they have taken on an almost Biblical character.)

Meanwhile, the CEO’s Report happily tells us every month that short turns are a thing of the past, that they are so rare that it might not even be worth tracking them as a service metric.

The big drop in the metric in spring 2019 coincides with a point where a “no short turn” edict was issued by the CEO. This is not really practical as there are many bona fide reasons for short turning vehicles, but the numbers obediently went down and have stayed down.

Regular riders, however, might choose to differ in their day-to-day observations.

Since 2019, we have come through the pandemic era when a great deal of traffic congestion and ridership disappeared. For a time, the type of event that would disrupt service was comparatively rare. However, with “normal” conditions returning, service is no longer insulated by these effects.

In my own travels, I routinely encounter streetcars that are not going to their scheduled destinations. Let me be the first to say that I understand the need for short turns, but am rather amazed that the reported counts stay very close to zero. This simply does not match actual experience. A short turn is a short turn, regardless of why it is required.

The question, then, is how to count these events reasonably easily without standing on street corners clipboard in hand. Vehicle tracking data that I already receive from the TTC’s Vision system (and which drives the many arrival prediction apps) provides a simple mechanism.

In this article I will review several common streetcar short turn locations to see what is actually happening.

If readers have specific bus routes and locations they would like to see, please leave your request in the comments.

501 Queen Eastbound to Neville

A short turn long-feared by 501 Queen riders is the discovery that a streetcar supposedly bound for Neville Loop is actually short turning at Kingston Road into Woodbine Loop.

To determine how many times this happens, I compared the counts of streetcars crossing a screenline east of Coxwell to the counts at Woodbine. If a car is seen at the first screenline, but not at the second, there is only one place it could have gone.

A few explanations are in order here:

• For each hourly group of columns, the first the values are for all weekdays, Saturdays and Sundays respectively.
• The other five columns break out data for weekdays on a weekly basis to show the variation through the month.
• It is possible for there to be values over 100% when, due to bunching, a group of cars crosses the first screenline within one hour and the second screenline in the following hour. This evens itself over time.

On an all-day basis, fewer cars reach Woodbine than are seen at Coxwell, and these are the short turns. Note that there are more of these on weekends (orange and green bars for Saturday and Sunday) than on weekdays (red) in the “All Day” values. The proportion of service that gets through to Woodbine (and hence to Neville) varies by time of day and is sometimes below 80 per cent.

Considering that on a weekly basis, there are almost 4,000 scheduled trips to Neville Loop, that is more than a handful of short turns.

Total trips (short turns):

Weekdays: 2811 (143), Saturdays: 501 (49), Sundays: 459 (56)

Dundas Westbound to High Park

A common short turn for Dundas cars is at Lansdowne using “College Loop”, an around-the-block loop formed by the Lansdowne-College-Dundas triangle. The screenlines here are westbound at Lansdowne and at Roncesvalles. (Note that all 505 Dundas service is diverted to High Park Loop due to construction at Dundas West Station.)

In this chart there is an obvious question about how the Sunday value for 6:00am can be 200 per cent. In fact, there was one day, June 26, on which several cars were held westbound at Ossington between 6 and 7am. Other cars diverted via Ossington and College. The diverted car were not counted at Lansdowne because they were off route. As a result, there were twice as many cars counted at Roncesvalles as at Lansdowne.

On a similar note, for the weekend of June 5-6, all data have been omitted because the service diverted via College Street for a street festival on Dundas.

There will inevitably be cases like this that require investigation and some data cleansing, or at least an explanatory footnote, but the overall premise remains.

As the chart shows, there are a lot of short turns at Lansdowne especially in the evening. One could well argue that the full level of service is not required beyond this point and the short turns concentrate cars where they are most needed. However, this does not make them vanish from the counts.

Total trips (short turns):

Weekdays: 2405 (189), Saturdays: 316 (19), Sundays: 273 (26)

The Dundas route also has a common east end short turn via Parliament, Gerrard and Broadview, but until June 19 Broadview was the eastern terminus due to construction. I will return to this when data for July are available.

Carlton Westbound to High Park

The premise for 506 Carlton cars is the same as for Dundas with a screenline at Lansdowne counting the “before” values, and another at Roncesvalles counting how many cars went through to High Park.

There are values much above 100 per cent in the 6am data, but the reason for this is different from the 505 Dundas chart above.

Many streetcars enter service from Roncesvalles Carhouse by travelling east to College Loop (Lansdowne) and then heading west to High Park. This gives a higher count for cars at Roncesvalles than at Lansdowne.

As on other routes, values well below 100 per cent are not uncommon here, especially just after the AM peak (9-10am) when many cars short turn at Lansdowne.

Total trips (short turns):

Weekdays: 2772 (358), Saturdays: 465 (87), Sundays: 459 (44)

Carlton Eastbound to Main Street Station

A common short turn for Carlton cars in the east end is to run to Coxwell-Queen Loop rather than going the full distance to Main Street Station. Note that all service on 506 Carlton originates at Roncesvalles Carhouse and so no trips end at Coxwell as running-in trips to Russell or Leslie.

Total trips (short turns):

Weekdays: 2498 (192), Saturdays: 397 (40), Sundays: 386 (20)

King Westbound at Spadina and at Bathurst

King cars have two common short turns westbound at Spadina (looping via Adelaide and Charlotte) and at Bathurst (looping south to Fleet Loop or north to Wolseley Loop).

For these charts, the screenlines are westbound at Peter (two blocks east of Spadina), at Portland (one block east of Bathurst), and at Tecumseth (one block west of Bathurst).

Short turns at Spadina are not uncommon.

Short turns at Bathurst compound those that have already occurred at Spadina.

When the effects of the two short turns are measured comparing service east of Spadina to west of Bathurst, here is the result. This affects service to Liberty Village, and does not bode well for the eventual restoration of streetcar service to Roncesvalles Avenue late in 2022.

Total trips (short turns):

At Spadina Weekdays: 5107 (207), Saturdays: 689 (23), Sundays: 701 (15)

At Bathurst Weekdays: 4900 (432), Saturdays: 666 (38), Sundays: 686 (35)

With the eastern part of 504 King under construction until June 19, I will leave analysis of short turns there until the July data come in.