Getting From “A” to “B” — Is There More Congestion?

Those who have come to this site in the past year to read, among other things, detailed analyses of route operations on King and Queen Streets may not be aware that this has been done before.  Back in May 1984, the Streetcars for Toronto Committee organized manual observations of the major streetcar routes for three days.  A detailed post on the subject appeared here in April 2006 and it makes interesting reading for any who think that service problems are new to the system.

At that time, we documented a very high proportion of cars short-turning in the afternoon peak and a systemic problem that the actual times required to make trips across the system was higher than the scheduled time.  Short-turns were poorly managed and contributed to the general chaos in service.  Not much has changed, although the headways are a lot wider now than they were in 1984, and the reliability of service much lower.

Considering how much store the TTC puts in “congestion” as the explanation for all its woes, it is worth looking back two decades to see what changes have been made in the schedules.

506 Carlton

When we surveyed the line in 1984, the scheduled pm peak round trip time (RTT) was 114 minutes for the Main Station to High Park trip, and the actual observed average was 121.  Today, the schedule (based on resumption of service to High Park) is 126 minutes including 3 minutes recovery time.

The headway in 1984 was 3’00”, and today it is 4’20”.

505 Dundas

In 1984, the scheduled pm peak RTT was 96 minutes for the Broadview Station to Dundas West Station trip, and the actual observed average was 101.  Today, the schedule (like Carlton’s, based on resumption of service west of Lansdowne to Dundas West Station) is 110 minutes including 7 minutes recovery time.

The headway in 1984 was 2’24” westbound from Yonge, and 4’00” eastbound.  Today it is 5’15” both ways.

501 Queen

In 1984, the scheduled pm peak RTT was 120 minutes for the Neville to Humber trip, and the actual observed average was 128.  Today, the scheduled time for the Humber branch of the 501 is 143 minutes, including 15 minutes (!!) terminal time to recover from delays. 

The headway in 1984 was 2’33” westbound from Yonge, and 1’32” eastbound from Yonge including cars on the Downtowner route.  Today, the Queen car is scheduled every 5’30” with ALRVs (equivalent to 3’40” with CLRVs), and the Downtowners, on a 14-minute headway, make almost no contribution to the eastbound service.

Back in 1984, the maximum headway observed at Neville Loop was only 17 minutes, and the standard deviation was 3’49” minutes.  This meant that a majority of the headways were below 6’22” (one headway plus one standard deviation).  That is only slightly higher than today’s scheduled service, much of which never reaches Neville during the pm peak.  What does arrive there does so on a very wide and unpredictable range of headways.

504 King

In 1984, the scheduled pm peak RTT was 100 minutes for the Broadview Station to Dundas West Station trip, and the actual observed average was 110.  Today, the scheduled time is 127 minutes including 11 minutes terminal time to recover from delays.

The headway in 1984 was 2’33”, and today it is 3’45”.

Congestion Then and Now

As you can see, the running times today are almost identical to those over 20 years ago, but the change lies mainly in the addition of “recovery time” at terminals.  At the Queen car forum, the TTC claimed that operators get 3 minutes at the end of the line, but in many cases there is much more time allowed in the schedules.  On Queen, the amount of recovery time seems to have much more to do with making the service merge properly at Humber than with the actual time needed.

This is borne out in my analyses which show quite long layovers for cars at both Humber and Long Branch.  This is time (and vehicles) that could be reclaimed and used for productive service if the line were split up.

As for traffic congestion, our comments in 1984 mirror those in my recent posts.

The TTC’s usual explanation for short-turns is that unpredictable accidents and traffic congestion require them. We observed only one accident which blocked service, and traffic congestion occurred at predictable locations and times. Moreover, two of our three surveys occurred on fair-weather days when the traffic congestion associated with rain or snow could not have been present.

The only real difference in over two decades is that the TTC recognized that scheduled running times in 1984 were inadequate, and these were lengthened.  The only other change, one which came after months of reviewing our report, was that the size of fleet numbers was roughly doubled so that route inspectors could identify a car at a greater distance.   (These larger numbers survive on the bus fleet.)

The big change since 1984, of course, is that there is far less service.  The TTC slashed streetcar headways disproportionately to the rest of the system in the 1990s, and has only grudgingly given back some of the service.  Ridership fell, but it fell most dramatically on Bathurst and Queen, the lines whose headways were further widened for ALRV operation. 

(For a longer-term view of service levels, please refer to my post on this subject from January 2007.)

Queen had the additional insult of the amalgamation with the Long Branch route, a disaster for which there has never been a formal review and which has just about destroyed this line as a solid, reliable local service on Lake Shore Boulevard.  This is the same place, by the way, that the TTC hopes to lure riders with the Waterfront West LRT line.  They might make a start by running credible service on the existing streetcar line.

Wider headways and crowding lead to longer times at stops, not to mention increasingly frustrated riders.  Every study of rider behaviour has shown that waiting time is perceived as a far greater inconvenience than travel time, and yet the TTC persists in offering “service” where wait times are unpredictable and long.  Transit priority can yield some improvements, but the biggest single change needed is for service to be reliable.

My opinion of the TTC’s inability to understand and acknowledge its own operational problems has not changed one bit since 1984, but many readers here may not know just how long-running and deeply-rooted this situation actually is.  We have a long way to go to recover the quality of service for which the TTC was once renowned.

3 thoughts on “Getting From “A” to “B” — Is There More Congestion?

  1. Downtowners, on a 14-minute headway,

    I think the TTC is outsourcing the Downtowner to GO Transit, and it is being scheduled like a Lakeshore GO Train 😛


  2. Do you have the comparison data for the Lakeshore line?

    Steve: We did not survey that line in 1984 because it was an isolated route on which reliability and short turning were not issues. However, the Long Branch car operated for years with 25 to 27 minute one-way trip times, and these are comparable to the values I see in present-day CIS information. There is no congestion on Lake Shore, only erratic service, even inbound when the “recovery time” should have smoothed everything out.


  3. Steve writes:

    There is no congestion on Lake Shore, only erratic service, even inbound when the “recovery time” should have smoothed everything out.

    The streetcar I rode last night was showing +11 on CIS approaching the loop — he was flying as soon as we passed Roncesvalles westbound. I wonder if they have a TV in the little operatior blockhouse and catch their favourite show there? There was probably a good 25 minutes before the car had to leave the loop eastbound; leave 5 minutes late (not unusual) and you’d have a full half-hour layover.

    Steve responds: How could he possibly have been that early with all of the “traffic congestion” downtown?


Comments are closed.