In two previous articles, I have reviewed the St. Clair car during its first month of operation on the new right-of-way over the complete route from Yonge to Keele. Running times during busy periods are down compared with April 2007, when the only right-of-way was between Bathurst and Yonge Streets. However, the situation with headways, an important factor in how riders perceive service quality, is quite another matter.
For the entire period of construction, the idea of regular, scheduled service was something of a fairy tale on St. Clair, and both the streetcars and buses made their way such as they could along the route. One would commonly see vehicles taking long terminal layovers, and headways were not a big priority.
In analyses of other routes, there is a common factor that is independent of the route’s length, the time of the year, the weather, eclipses or any other phenomena: vehicles do not leave terminals on a regular spacing. They leave when they get around to it, a practice abetted by the TTC’s standard that ±3 minutes is considered to be “on time”. Pairs of vehicles can travel together on routes with short headways and remain within this standard.
Here are a few samples of headway data for the month of July 2010.
These charts are in the same format as the link times presented in the previous article. Data for each weekday are grouped onto four pages in each set, followed by a page with all of the weekdays. The purpose of this is to show the individual data with trendlines for the days, as well as the “cloud” of data points for the entire month.
Weekend days are shown separately because they run on different schedules. Note that streetcar service did not run to the west end of the line on all weekend days due to street festivals.
The times shown for St. Clair Station are after any layover the cars may have taken. These charts show that although the headways trend at roughly the scheduled levels, there is a considerable variation from car to car. Although this display shows headways rather than schedule variation, it is self evident that it would be impossible for service that was actually on time to exhibit such a spread in headways.
When the service starts off from the terminal running ragged headways, it is inevitable that this will persist across the route unless there is active intervention by route management. Things have improved slightly by the time the service reaches Bathurst westbound, but the weekday “cloud” still lies mainly in a band between 0 and 10 minutes.
Eastbound, service crosses Keele Street on headways as scattered as at Yonge, and on Saturdays with particular irregular spacing. Things have improved somewhat by the time we reach Dufferin, possibly due to having passed both the crew relief point and the Route Supervisor at Lansdowne. Again, however, “better” is only a matter of degree and there is quite a mixture of short and long headways. By the time cars reach Tweedsmuir (east of St. Clair Station), the headway cloud has spread out again in the typical pattern that cars carrying gaps fall further behind as they cross their routes.
I have not bothered to include comparative charts for April 2007 here as they are much the same, and you can review them in the earlier article on that period’s operation. If anything, the problem of irregular headways appears to be slightly worse in 2010 than it was in 2007. This could be a question of less rigourous operating procedures, or that comfortable running and little traffic interference encourage leisurely breaks that can easily be made up within the scheduled trip times.
Another way of looking at the headway data is to view the spread of values.
These charts are not intended to be used to read individual data points, but to give an overall impression of what is happening.
Each cluster of vertical lines represents one day’s data broken into individual hours. These come from the same data that generates the daily headway graphs. The small dots (seen more easily if you magnify the PDF) are the average values for the hour, and the bars show the maximum and minimum values within that hour. The longer the bars, the more spread out are the data values within one hour. The many lines extending above the 10 minute scale show that waits considerably longer than the advertised headway are not uncommon, and the same periods also see some very closely spaced cars.
I am sure that someone will ask about Standard Deviations, and, yes, I did produce a bunch of plots to see how those numbers behaved. I have not included the charts because they are a bit messy thanks to the comparatively low value of “n” for each hourly period on each day. What is quite consistent, however, is that the values lie mainly in a band from 2 to 4 minutes. The 2007 data are actually better behaved than the 2010 data, and this is consistent with my other observations.
If there is a lesson to be learned about managing service on a line with little interference from traffic and generous provisions for recovery time at terminals, it is that headway management is even more important here than on routes where the rigours of traffic congestion and quick turnarounds are the norm.
In the final article of this series, I will examine a few days in detail to show examples both of ordinary operation and of the upheavals caused by external events.