Route 501 Queen in December/January 2007/08 (Part 3: Headways)

The 501 Queen car is a mess — that’s no surprise to anyone.  In the past two posts, I have looked at where the cars go.  More, to the point, I have looked at where they don’t go thanks mainly to short- turns.

Riders are frustrated not just by the unpredictable destinations of Queen cars, but by the lack of reliable headways.  The schedule may say there’s a car every 6 minutes, but this is at best a general average.  TTC “on time” goals say that an error or plus-or-minus 3 minutes is an allowable margin, and this means that gaps of up to 9 minutes, followed by only 3 minutes , is “on time” performance. 

This ignores the fact that the gap car will carry 3 times the headway, be much more heavily loaded, and the average rider’s perception of the service will be much worse than the average implied by the schedule and the loading standards.

If the service actually stayed within 3 minutes of the advertised schedule, life would be annoying, but tolerable.  There might even be a car in sight much of the time, if not the “always” of the TTC’s slogan for good service from the 1920s.  We should be so lucky.

This article reviews the actual headways at selected points during the past December and January.  Without question, “weather happened” in various ways over these two months, but not every day, and this cannot explain all the data we see here.

Each linked PDF contains seven pages.  The first four of these show the headways over weekdays for each of the four weeks in the month with each day plotted in a separate colour.  The trend line follows the broad pattern of the values, but is not disturbed by short-range changes.  When service is running properly, the trend line will roughly follow the scheduled headway.  When service is poorly spaced, but otherwise all present, the trend line is quite similar to that of properly spaced service because the average remains consistent.  However, the values will be scattered around the trend.

The fifth page contains the data points for all weekdays.  No trendline is plotted (it would be more or less meaningless and just get in the way), but this chart shows the overall range of values seen on weekdays through the month.  The more tightly grouped the points are, the more reliable the headways.

The sixth and seventh pages contain Saturdays and Sundays.  For the purpose of these charts, Christmas and New Year’s appear on the Sunday chart, and Boxing Day appears with the Saturdays.

Westbound From Yonge Street

Westbound at Yonge Street — DecemberJanuary

At Yonge Street, the first three weeks of December show a gradual rise in the scatter of weekday headways as Christmas approaches.  Week three, as we have seen earlier in this series, also suffers from the carry-over effects of the storm on Sunday, December 16.  In week 4, the scatter of headways is less than early on the month because there is much less traffic on the road.  This builds up, of course, on New Year’s Eve.

In January, the weekday headways are more scattered in week 1 reflecting the weather, but they settle down for the remainder of the month.

The charts on page 5 of each set contain clouds of points showing the overall distribution of headways on weekdays in each month.  The December values have greater spread with noticeably more points above the 20-minute mark than in January.  These “clouds” are nothing to be proud of because they show the degree to which many trips exceed the three-minute rule for “on time” performance on a headway basis (never mind where the cars are actually supposed to be).

The scheduled headways were:

  • AM Peak:  4’52”
  • Midday and PM Peak:  5’30”
  • Early Evening:  7’22”
  • Late Evening:  10’00”

On a 10 minute headway, the commonly-seen values above 15 minutes are more than tolerable nuisances, and they discourage transit riding.  On a 6-minute headway, such gaps are far to common to be acceptable.

Saturdays and Sundays are almost indistinguishable from weekdays except that the outliers above 20 minutes are much less common.  December 16 is a special case due to a snowstorm.

Westbound at Parkside Drive — DecemberJanuary

By the time we reach Parkside Drive (the start of the right-of-way at the east side of High Park), cars short turning at Roncesvalles or further east have dropped out of the service, and the average headways are wider.

The December pattern is comparable to what we saw at Yonge Street, but with longer headways, some of which are quite wide.  The trend line for weekday headways has moved up to around 8 minutes during the daytime, and longer gaps are more common.  Friday, December 21 (effectively Christmas eve from a traffic point of view) was particularly bad with its trend line pulling above 10 minutes during mid-afternoon.

In January, the headways are better-behaved, although far from ideal, and the trend lines are lower indicating that more service operated west of Roncesvalles.  The cloud of headway points for all weekdays is tighter in January than in December.

Saturday headways are badly scattered showing the effect of shopping traffic in December, and t oa lesser extent in January.  December 1 is an oddball day because daytime service west of Roncesvalles operated with buses on a noticeably less reliable and frequent basis than the streetcars on other weekends.

Sunday headways are also scattered, again because Queen is busy seven days a week.  The December 16 snowstorm caused some very wide gaps in service.

Westbound at Royal York Road DecemberJanuary

Once we get beyond Humber Loop, things really fall apart.  Whether in December or January, headways are all over the place, pairs of cars and gaps of over 30 minutes are common.  The “cloud” of weekday headways is worse in December, but January is nothing for the TTC to crow about.  During periods when the scheduled headway is 11 minutes, many data points lie at 20 minutes and above.

Weekend service is a bit better in January, but there are still far too  many cases of cars running on headways wildly different from the schedule.

Eastbound to Yonge Street

Eastbound at Royal York Road — DecemberJanuary

After cars go west to Long Branch, most trips have a scheduled recovery time at the loop.  This is supposed to be taken as a layover only if the car would be early leaving inbound, but many operators take a considerable break here regardless of the schedule following the 90 minute trip from the east end.  Inbound service does not operate on a reliable headway, although the spread tends to be a bit better for the outbound service.  This is more noticeable in January than in December.

Throughout both months, cars running in pairs are commonly seen throughout the day.  This shows a complete absence of headway management by TTC and a major problem with service quality for Lake Shore residents.

Note that these charts include cars that were short-turned at Kipling, and the service inbound from Long Branch to Kipling would suffer relative to the charted values because of these short-turns.

Weekends in January brought service that managed to stay better aligned with the scheduled headway compared to service on weekdays and in December.

Eastbound at Parkside Drive — DecemberJanuary

Service eastbound at Parkside is goverened to some extent by cars leaving from Humber, and the spread of headways is tighter than we saw outbound.  January weekday headways lie within a band below about 15 minutes.  For a 5’30” headway inbound on a reserved right-of-way, this is less than ideal, but it is better than December where the band spreads out to the 20-minute line.  This situation arises both from short-turning at Roncesvalles and from irregular inbound headways from the two terminals, Humber and Long Branch.

Eastbound From Yonge Street

Eastbound at Yonge Street — DecemberJanuary

Headways at Yonge Street eastbound behave quite similarly to the westbound data except that they are slightly more spread out.  This reflects variation in travel times inbound through the west side of downtown which has no counterpart for cars arriving from the east.

As with the westbound data, the January headways are more tightly bunched around the trend lines than in December.

Eastbound at Woodbine Avenue — DecemberJanuary

At Woodbine, the effect of short-turning is clear with the trend lines sitting well above the scheduled headways (5’30” for midday and pm peak service).

In December, the headway cloud is spread out with many data points well above the 10-minute and 20-minute lines for weekdays.  Saturdays are slightly better with most data below 20 minutes.  Sunday December 16th shows the effect of the snowstorm with the trendline rising to about 25 minutes.

In January, the headway cloud is not as bad as in December, but it still has many points well above the 10-minute line.  Weekend service shows some scatter up to the 20-minute line, especially on Sunday.

Westbound To Yonge Street

Westbound at Woodbine Avenue — DecemberJanuary

Weekday headways at Woodbine are all over the map especially in December when gaps over 20 minutes were common, and quite a lot were above 30 minutes.  Management of inbound service from Neville is clearly non-existent.

The situation on weekends is slightly better, but still very poor.  Once again, the December 16 snowstorm shows up on the Sunday chart where the trend line is tracking values that are off of the page above the one hour mark.

Conclusion

These charts show the combined effects of short-turns and unregulated headways along with effects from weather and seasonal traffic (road and passenger).  The quality of service on the inner part of the line (between the two carhouses) is erratic, but beyond these points it is really atrocious.  The scheduled headways are meaningless, as would be any riding counts or projections that assume something like reliable service.

A major flaw in TTC’s service planning is that it takes no account of the riders lost through poor service.  For years, light ridership was used to justify further service cuts.  The long-standing operational practice of concentrating on service downtown discourages those riders for whom the private car has a strong competitive advantage.

The TTC needs to do much, much better.

In the next installment, I will turn to link times (the time taken to travel from point to point on the route) and discuss the issues of traffic congestion and transit priority.

2 thoughts on “Route 501 Queen in December/January 2007/08 (Part 3: Headways)

  1. It’s hard to see in the headway graphs the gaps in service, followed by a pair of cars. That’s because there are multiple days and the horizontal timescale is compressed. It can be impossible to pick out each sequential point in a day, because the points are up and down so much, and any down by the y=0 axis get mixed up with all the other car pairs. On other headway charts, you had lines connecting the headways, like you have with the destination-ratio charts.

    Steve: The problem here is that if I put in a line to actually follow individual points, you wind up with a complete mess. I have done that before on headway charts, but only when one day’s data was displayed. On the weekly charts, I am trying to show the overall patterns — how the headways vary over time, how much spread there is around the trend value.

    The cloud of points where all weekdays are combined is intended to show just how badly spread out the headways are on an all-day basis. The thickness of the cloud stays more or less the same through the day, although weekend clouds can be better than weekdays, and January’s can be better than December’s. The underlying problems, when we get that far, affect all hours of service, and any strategem that is aimed only at a few hours per day will fail.

    When we get to the link times (now in progress), you will see that congestion delays are almost exclusively on the portion of the route from Yonge to Roncesvalles (no surprise to anyone who rides the line), and that they don’t just happen in the peak period.

    Every one of those points that is close to the y=0 axis is a close headway. Considering that the target is at most schedule minus three (2’30” east of Humber, 8’00” west of Humber for much of the day), it’s obvious that a huge number of pairs or very close headways are operating, and the TTC does little or nothing to fix this problem.

    I could change the page size to legal to extend the horizontal axis, but this would prevent many people from printing them.

    On the other hand, the destination-percentage and destination-ratio charts may work better with trendlines (as you used for the headway charts).

    Steve: I have created a version of the chart with trendlines, and have included only weekdays as it would be meaningless to trend through Saturdays, Sundays and Holidays.

    You can access it here. I will roll this into the main post when I get a chance to make detailed comments.

    Also, the destination-ratio charts should have a y-scale that shows better detail. Anything falling below 0.50 is bad for service in the west end.

    Also, the destination-ratio charts are a bit misleading: what throws off the casual glance is that the same distance above and below 0.50 means something different. In the LB:Neville chart, if absolutely no cars get to Long Branch, then the ratio is zero: a mere 0.5 below the mean line. But 0.5 above the mean line doesn’t mean that no cars get to Neville, it just means that the same number of cars get to Neville as to Long Branch. No cars to Neville will be at infinity.

    If only one out of every ten cars made it to Long Branch, the data point would be at 0.1, but if only one out of every ten cars made it to Neville, the point would be at 10.

    This argues for the scale to be either logarithmic, or measured against all active runs (in which case the scale for both Neville and Long Brach can range only from 0 to 1.0, and means the same thing for both).

    Eyeballing the percentage of service reaching Long Branch, it looks like the 1800 to 2100 time period is especially bad, with 40% or fewer runs getting to Long Branch, and it gets worse towards the end of January. This fits with my observations, trying to get home, waiting for a Long Branch car around 1830 at Spadina.

    Steve: I wrestled with the problem of the scales for Neville versus Humber and Long Branch services. Simply multiplying the Long Branch numbers by two would restore the balance presuming that twice as much service is intended for Humber/Neville. However, there are times when more than half the service actually goes to Long Branch, and values over 1 will occur. As it is, I have adjusted the Y-axis in the trendline version to have a maximum value of .7 which holds all of the data values.

    Your observations at Spadina also tell me that the Long Branch runs are changed to Humber cars (or worse) somewhere east of you. This is an interesting conundrum of line management and shows how a one-size-fits-all rule doesn’t necessarily work.

    A big problem people complain of is “surprise” short turns where a car leaves Yonge signed Neville but only goes to Woodbine Loop. To avoid this, the strategy is to decide on the short turn before the car reaches Yonge eastbound. Given the distance from Yonge to Woodbine, a reasonable guess can be made of whether the car will need to short-turn or not, assuming you are managing to the schedule, not to a headway.

    However, the westbound trip to Long Branch is huge, and there is no way that a route supervisor or CIS can know when a car is, say, westbound at Jarvis, that it won’t be able to make it to Long Branch. Changing the run to a Kipling or a Humber that far ahead potentially wastes time a vehicle might well have by the time it gets further west. Moreover, it puts big gaps in the service to Long Branch as you well know.

    My gut feeling is that Long Branch runs should NEVER be short turned unless they are already west of Humber, there is a dire emergency and another car is available close behind. This should be a basic rule for any service with wide scheduled headways for obvious reasons.

    Of course, managing to a headway rather than to a schedule would make all of this a lot easier. When traffic or weather gets bad, the headway would widen out, but the intent would be to keep it as regular as possible, not to short turn everything with the hope of preserving a schedule.

    Like

  2. Steve,

    Two questions.

    1) If the TTC started managing to a headway instead of schedule tomorrow, assuming they did it competently, what would you expect to see? I would assume that this should eliminate short-turns except in the circumstance that a track blockage (disabled car, badly parked auto, snowstorm..) puts service behind to an unrecoverable point. We’d therefore see a drastic difference in service, yeah?

    Steve: I would hope to see a considerable improvement. “Drastic” is a word I would use only for a near-miracle of reliable service almost all of the time. Right now, I’m concerned that the TTC will concentrate a lot of effort on periods where they think they have problems while forgetting off peak periods and locations that are not right downtown.

    2) Any advice on what we, as ordinary riders, can do to affect change? The TTC doesn’t make itself easy to talk to, but surely there is something that the thousands (?) of 501 Queen riders can do to make it clear to the TTC that we’re peeved and aren’t going to take it anymore.

    Steve: A public meeting is planned for June as a follow-up to the one held last December at Metro Hall. Details are not yet finalized.

    Like

Comments are closed.