Updated May 20, 2019 at 4:40 pm:
Another factor in the travel time conundrum may be related to a “no short turns” edict that went into effect for the second half of April 2019. One effect of this can be that there are often large gaps and associated loading delays which drive up travel time. This will be the subject of a separate article.
Updated May 21, 2019 at 8:45 am:
Charts for detailed changes in eastbound travel from Humber to Neville have been added to this article.
With the June 23, 2019, service changes, the TTC will officially make route 501 Queen a low-floor streetcar route. Concurrent with this change, the number of cars on the route will be reduced during all operating periods. The degree of reduction should be cause for celebration because reduction in cars is nowhere near the ratio of low-floor to CLRV capacity. However, at the same time, the TTC will substantially increase running times during many periods. The combined effect of fewer cars and longer scheduled trips will be much wider headways than are now provided, especially on weekdays.
The number of cars/hour will be cut by about one third on weekdays and one quarter on weekends. This is very much the sort of change Queen Street riders saw when the two-section ALRVs replaced the shorter CLRVs, and the end result, thanks to the lackluster line management, was a loss of 1/3 of the route’s riding. This is precisely the opposite direction the TTC should be moving at a time when there is known demand on the streetcar lines as shown by the King Street Pilot.
When the fact that 501 Queen is already partly served by the larger Flexitys, the capacity on the route will actually be reduced on June 23 compared to what is operating today.
In the peak periods roughly half an hour will be added to the scheduled round trip time from Neville to Humber (from 149 to 176 minutes AM, and from 169 to 205 minutes PM) resulting in much wider headways that would be provided if the Flexitys ran on the same operating plan as the existing service.
TTC management seek to reduce or eliminate short turns with extended running times, but in the process produce a service plan that will be demonstrably worse for all riders. The wider scheduled headways will be compounded by the uneven spacing of cars that is commonly seen on Queen and other routes, and it would not be surprising to see pairs of cars travelling on headways over 10 minutes almost all of the time.
This would be a far cry from the type of service that clearly warrants transit priority on King Street. Any discussion of priority for Queen would first have to ask why the streetcars that are so infrequent would deserve so much road space to be devoted to them.
Such a large change in scheduled travel time raises the question “why”. Is this simply a padding of schedules to minimize short turns and make life easier for operators, or is it a shift to recognize a change on the route? If the latter, what has changed and how have travel times evolved over time.
Looking into the details, what appears is that there has been a build-up in travel times across the route over the past year, and particularly in recent months. Further exploration will be needed to determine what is happening, but the situation is cause for concern when there can be such a large change in travel times and, in turn, a degradation in service frequency.
501 Queen Trippers
The schedule in effect in May-June 2019 includes five tripper cars operating from Russell Carhouse (at Greenwood/Connaught) in both the AM and PM peaks. Three of these provide through Long Branch/Downtown service. In the AM peak the other two provide two round trips between Greenwood and Sunnyside Loop (runs 83 and 84 below). In the PM peak , they make only one trip from the carhouse west to Humber and back to Neville (runs 88 and 89).
The choice of timepoints below may look a bit odd, but these are the ones used in schedules exported by the TTC to NextBus. Glendale is the St. Joseph’s Hospital stop west of Sunnyside Loop. Triller is the first stop east of Roncesvalles.
In the June-July schedule, buses will provide three trips each way from Kingston Road to Sunnyside in both peak periods. These are separate vehicles, six in all for both peaks. These are one-way trips. With the peak design load of a bus being about 40% of a Flexity, this does not represent a substantial amount of service beyond what the streetcars will provide.
Eastbound from Sunnyside:
- AM: 8:15, 8:20 and 8:25
- PM: 4:30, 4:40 and 4:50
Westbound from Kingston Road:
- AM: 8:15, 8:20 and 8:25
- PM: 4:35, 4:45 and 4:55
The 501L Long Branch service is not affected by the schedule change according to the service memo. I will verify with the TTC that the Long Branch/Downtown tripper streetcars remain in the new schedule.
Trip Times on 501 Queen
Travel times across the Queen route can vary substantially by day of the week, within weeks of the month and of course seasonally. No “one size fits all” schedule is possible especially if the goal is to completely eliminate short turns. If running times are so padded that even under the worst conditions cars will never need to turn back, they will have extremely long layovers at terminals, assuming there is even space for the vehicles there. This is a waste of streetcars that could be providing service rather than sitting near the end of the line.
On a full month average basis, the travel time between Neville and Humber Loops follows a similar pattern in each direction. (The measurement is from the first stop west of Neville at Silver Birch to a point on The Queensway just east of Humber Loop to avoid including terminal layover times in these values.) The one way time peaks at about 90 minutes for cars leaving the terminals in he hour between 5 and 6 pm. Equally important, however, is the standard deviation value which is typically close to 10 minutes. This means that a substantial proportion of the trip times lie in a band ±10 minutes of the average. This is far to large an amount to be absorbed into scheduled recovery time.
Breaking down the values by week reveals that the last two months of April (green and blue below) saw longer trip times than the first two (red and orange). This begs the question of whether some change on Queen Street affected service from that point onward. Later below I review the route segment-by-segment and this shows that the changes vary by location and time of day. The effect is widespread enough that one must ask whether there has been a change in (or removal of) transit signal priority at some times and locations on the route.
Even more variation is revealed on a day-by-day basis. The charts below are for Week 4 (including the two days of the short Week 5) which had some of the highest travel times. Note that these charts are clipped at 120 minutes, but there are some values above that line. (Each dot represents one car’s trip, and the collection of dots lies generally within a 20-minute wide band as the 10-minute SD value leads is to expect.)
Full chart sets:
The new schedule’s PM peak round trip time is 205 minutes (192 minutes driving time plus 13 minutes of recovery). With the high end of average trip times sitting at about 90 minutes each way, and allowing 8 minutes for each terminal, a long round trip should take less than 200 minutes. The actual scheduled value is generous and will minimize, but not eliminate, the need for short turns.
Also important will be headway management so that streetcars, including those that are short turned, actually operate on a regular headway. The TTC “standard” allowing a six minute “on time” window would allow peak period streetcars on Queen to run in pairs every 13 minutes or so while being “on time”. That is not acceptable service on what is supposed to be a major route.
CLRV vs Flexity Trip Times
Among the reasons cited by the TTC for a jump in short turns on this route during March 2019 was the claim that operating characteristics of the new Flexity and legacy CLRV fleets made for uneven service requiring more short turns.
The actual average travel times for the two vehicle types are slightly different, with the new cars being a bit slower, but not by a wide margin.
The Evolution of Trip Times on 501 Queen
Updated May 21, 2019 with charts for eastbound data
Current travel times across Queen are now higher than scheduled times, and that is part of the reason for the schedule change. This prompts the questions of when the change happened, and where along the route. This section delves into those questions for travel between Neville to Humber. Westbound and eastbound data are similar in character although some of the details vary because demand and congestion effects vary with direction and time of day.
The chart below shows the average travel time from Silver Birch to The Queensway just east of Humber Loop in (mostly) April of each year from 2012 to 2019. Each group of bars shows the progression of values over the 8 years for each hour of the day. (The time refers to the point where a car left Silver Birch, and so the 6:00 bars show data for cars westbound from the Beach between 6:00 and 7:00 am.)
Small annual increases occur, but there is a large jump in most periods in 2019.
Eastbound travel times show the same general pattern.
Of equal concern is the evolution of the standard deviation values. Although there are some spikes in earlier years, 2019 stands out for very large SD values indication a wide variation in the range of travel times particularly during the AM peak and immediately following.
Looking at the past year on a more detailed level, month-by-month, reveals that the jump is quite recent. This chart shows a slight growth over the past year for most time periods, but a big jump in April particularly in the mid-day and PM peak. The change is more pronounced for westbound travel than for eastbound.
Looking at the data on a geographic basis with the route broken into segments shows that there are increased travel times in April 2019 compared to April 2018 on most segments of the route except west of Roncesvalles:
- Neville to Jarvis (red): Travel times on the eastern part of the route are up particularly in the midday period.
- Jarvis to Bathurst (yellow/orange): Increases for this segment are greater in the PM peak.
- Bathurst to Roncesvalles (green): This sement shows increases through the midday and into the evening.
- Roncesvalles to Humber (blue): There is little difference in travel times over this, the only segment which is largely on private right-of-way.
This appears to be a route-wide problem and cannot be explained by the possible diversion of traffic from King to Queen Street which, in any event, would already have been well underway by April 2018.
Looking at an even more detailed level reveals which parts of the route have seen the most change. This is broken down in five charts because there is too much information to fit on one page.
- Neville (Silver Birch) to Woodbine (red/brown): Travel times in this segment are up from 9:00 am onward. This suggests that there may have been a change in traffic or parking restrictions, or possibly in signal timing after the AM peak.
- Woodbine to Coxwell (green): Like the section east of Woodbine, the area to the west to Coxwell shows a jump in travel times after the AM peak. Again this suggests a change in traffic regulations or signals.
- Coxwell to Greenwood (blue): The data here are similar to the other segments.
The effect is stronger in the westbound than the eastbound direction.
- Greenwood to Broadview (red/brown): There is a slight increase in average travel times during many periods of the day for this segment.
- Broadview to Parliament (green): Travel times are up slightly across this segment.
- Parliament to Jarvis (blue): Travel times are up slightly here only for the period from 8:00 to 11:00 am.
- Jarvis to Yonge (red): Travel times are up here from 8:00 am until the early evening.
- Yonge to University (yellow/orange): There are large changes in travel times here especially from mid-afternoon through early evening.
- University to John (green): There is a small increase here through much of the day.
- John to Spadina (blue): Travel times in this short segment have not changed much, and in some cases are slightly lower.
There is a pronounced difference in the eastbound data compared to westbound for this section of the route with less change, especially from University to Yonge, eastbound.
- Spadina to Bathurst (red): Travel times here have not changed much at all westbound, but have risen eastbound.
- Bathurst to Ossington (green): Travel times are slightly increased in both directions.
- Ossington to Dufferin (blue): Travel times here have changed slightly in some periods both up and down.
- Dufferin to Lansdowne (red): Travel times here jump from 11:00 am onward right through the late evening westbound, but are almost unchanged eastbound.
- Lansdowne to Roncesvalles (green): Travel times here are up a bit with the largest changes in the off-peak hours westbound, but for eastbound travel there is a major change throughout the day.
- Roncesvalles to Humber (blue): There is a small increase in travel time mainly in the evening westbound, but through much of the day eastbound.
Full chart set:
Increased headways, sigh: “When will they ever learn“ …
Thanks for the great analysis Steve, and I have a few points.
Firstly, when the headways are this wide, is the TTC just going to kick off a whole packed Flexity full of passengers to short turn and fill the likely enormous gap going the other direction? Then oh no, the next streetcar is 10 minutes behind and everyone has to wait for that packed vehicle. I don’t buy it for a second that this will significantly reduce or eliminate short turns, not a chance. And with that I have to say replacing the streetcar trippers for bus trippers is also another downside. Putting 3 buses doesn’t really help very much.
Secondly, over the past week, there has been an effort to reactivate the ALRVs as at least 2 have seen service. Are ALRVs scheduled anywhere in the June-July schedule or would they just sit there and rust again or run as unscheduled extras?
Steve: In theory, there are five ALRV trippers on the current schedule, of which three run to Long Branch. The actual vehicle assignments do not match the original plan. In the June schedule, they might show up on Bathurst, or as unscheduled extras.
Thirdly, from observation, it seems like Woodbine loop has been used a lot recently. I’m not sure if this is a factor or not, but does the time it takes for track switches and turning at Queen/Kingston contribute to the changes in travel times from Woodbine to Coxwell ? And to that as well, does the replacement of the 502/503 streetcars with buses contribute to a change in times there as well? And between Broadview and Parliament, does the increased use of low floors on the 504 coming off Broadview then having to turn down King contribute as well?
Steve: The arrangement at Kingston Road and Queen has been in place for several years and cannot account for the change seen so recently in travel times. Moreover, those changes are seen on many segments of the route. Their sudden onset, and the fact that some of them appear immediately after the AM peak, tells me that (a) this is not due to Flexitys replacing some CLRVs and (b) there is a good chance this is a change in transit signal priority. Very fine grained analysis would be needed to delve into that, and it’s easier to simply ask the city “what did you change”.
I remember a specific instance where a 504 was deploying the wheelchair ramp at Queen/Carrol and in that time, around four 501 cars, 3 of which were Flexities were bunched up behind it. Then they got moving and the 504 took a long time to turn down King since there was a lot of eastbound traffic. It just became such a bottleneck, worst of all, this wasn’t even on a weekday.
I can’t speak about the other parts of the route, but a couple of visits downtown suggest the Yonge to University travel time changes could be due to southbound car travel changes. This includes both blocking intersections and generally reducing capacity including for cars on east-west streets trying to make turns. As I understand this is due to the York Street construction project reducing car capacity there. On evenings when there is a sports event at either of the arenas it can be particularly bad.
Last week I was surprised to actually see an officer enforcing not blocking the intersection at Queen and Bay.
Parkdale also has delays/queuing around Jameson/Lansdowne misalignment but I am surprised to learn it used to be better. There is also that weird light cycle eastbound at Lansdowne that goes from green to green with a 3 second red in between…
Steve: Yes, I have seen some of those effects on King as well as on Queen. However, the change across the entire route in mid-April suggests that something is afoot on a wider scale. The year-by-year and month-by-month averages don’t look too odd, but the recent spike is definitely a whole different class of change.
I doubt much should be expected of the ALRVs – 4207 and 4221 were in service in the early part of last week, but 4221 apparently had a failed compressor near Glen Manor. Meanwhile 4230’s tracker went offline at Exhibition Loop and nothing has been seen since. 4207 has been in Russell since then.
I have been on some CLRVs that have been following extremely slow Flexities across town. This suggests that the main reason for the increase in travel times is slow Flexities, at which point there are a few CLRVs hanging out right behind. This would then increase the travel time for CLRVs as well, because they can’t pass the Flexities.
The problem with Flexities is that they can hang around for a long time at a stop. The doors start to close, someone runs out and presses the open button, and that’s another 30 second pause. It is not unusual to see a Flexity sit at a minor-ish stop such as Brock St. for a full traffic light cycle, with the doors opening and closing a few times before it manages to move off.
It may be already in the data you did on headways, but it could be worthwhile to see whether Flexities are overrepresented as cars carrying a big gap, while CLRVs are running on short headways right behind the Flexities.
Steve: I will try to tease that information out of the data, but there are a few things worth noting.
First, the change is quite marked beginning in week 3 of April although the Flexitys were deployed at the start of the month with the new board period. Second, there are periods and locations where the change in travel time is substantial, and others where it is not. This does not line up with a “Flexity effect” through massively slower operation. I expect some effect from the long dwell times you describe, but not on the scale that we see on Queen.
But I will have a look to see what there is to see in the headway data.
Updated: I looked at some of the data, and there are multiple factors at work, not the least is the uneven headways departing from terminals and from merge points where short turns or carhouse pull-outs join the stream. These are erratic at the best of times, and a following car will always catch up with its leader. This is probably easier for a CLRV to do as the operators drive them in a more “sprightly” fashion. However, as ops become familiar with the new cars, I have noticed that when they get a chance, they move quickly to the point of throwing riders off balance with acceleration.
The whole zone with the shared 501 and 504 track is still a huge bottleneck now even if the 504B is theoretically less frequent. Because of the size of the Flexity cars, bunched cars will only load right at the stop unlike the old ones which would do so together in pairs. Getting more than 1 car through the Broadview intersection during 1 cycle is a miracle.
And also the crew changes at the corner of Broadview take much longer on the Flexitys compared to the old cars and often cause a streetcar to waste the transit phase or miss cycles entirely.
There was a period earlier in the year at Broadview when the transit phase was being triggered on every cycle even if there was no streetcar waiting to turn which stole a lot of time from the other movements in the intersection causing localized congestion throughout the day.
506 cars needing diversions east of Parliament also need to clog the Broadview intersection to make its turns but those probably don’t show up as more than blips for an hour here and there.
Steve: I direct your attention to the chart which shows that while travel times Broadview-Parliament are up a bit, this is only a small increase compared to the changes seen at other locations. As for the eastbound transit priority signal at Broadview, that has no effect on the Queen service westbound beyond whatever green time it stole when it was tripping on every cycle.
It is important to look at effects which exist on multiple days, times and locations, although I of course can look at even finer details than I have published (knowing that even avid readers have their limits).
Am I right in assuming that this is Doug Ford’s fault?
Steve: No. This change was concocted by the TTC completely separate from the Ford Follies at Queen’s Park.
As Citynews reported, TTC is tearing apart a Leslieville intersection for the third time in five years and this is not even the last time and all this for TTC streetcar track work. The project initially expected to cost $14 million has already costed almost ten times as much. Some years more than half of the entire streetcar network is ripped apart for reconstruction and some routes like Queen are ripped apart for 9 months every year for several consecutive years. Give me one good reason why Toronto should keep it’s streetcar network. I am not a car owner and I don’t bike but I am tired of never ending streetcar track work. Subways are much better and faster.
Steve: I was looking at the work on Leslie and it appears to be only the concrete outside of the rail area between the curb lane and the tracks, not the entire trackbed. As the news reports already indicated, the problem lies with faulty original construction. Track is rebuilt all over the city routinely that does not have these issues which came about from a new and apparently incompetent contractor.
A big chunk of the original cost increase was thanks to a change in scope. What was originally thought to be a job where an existing old watermain could be protected turned into a complete replacement, frankly something that should have been foreseen at the outset. There were a lot of screwups with Leslie barns right from the beginning of the project.
As for “half of the streetcar network” being ripped up every year, that is flat out not true.
I suspect that you’re also for lower taxes, but if we’re going to replace streetcars with subways, get ready to pay a lot more.
The “side strips” of concrete on Leslie south of Lake Shore are also very bad in several places; several sections were fixed (permanently?) last fall but still more to do. Clearly a poor job by the original contractor (Pommerleau).
I am guessing that the increased headways due to the 1:1 substitution is necessary for the TTC to have enough streetcars to supply the other major routes (Carlton, Dundas, Bathurst and Queen West (Humber to Long Branch) and maybe a few for Kingston Rd. I fear that Kingston Road service will remain bustituted permanently. Only a quick purchase of additional streetcars will save this service.
Steve: It’s not even 1:1 substitution. The peak service has fewer cars in it with Flexitys than the scheduled service today. Same for all other periods. Allowing for the fact that there has already been a capacity boost with partial Flexity service, this is actually a cut in capacity.
So what will the TTC do with the CLRVs it currently uses on Queen?
Steve: Some of them will go to 511 Bathurst and some will be retired.
What exactly will service on Bathurst be like? Will it be all CLRVs or a mixture of CLRVs and Flexitys?
Steve: As I have said before, it is scheduled for CLRVs, but Flexitys may run as part of the service depending on availability. That’s all the info the TTC has put out, and all you’re likely to see until the day streetcars return when, I suspect, the mix will vary from day to day.