Over at torontoist, Adam Giambrone responds to a question about bunched service outbound from Mimico on the Lake Shore with a collection of the usual lame TTC excuses about irregular service.
This is getting tiresome, and it is distressing to see the TTC Chair spouting so much of the party line from TTC management. The reasons for irregular service, according to Giambrone, are:
- Bunching caused by minor variations in the time spent at stops and traffic lights. This applies to frequent routes where the headway (as cited by Giambrone) is fairly close to the variation in delay times. The last time I looked, the best scheduled service on Lake Shore is every 8’40” on Saturday afternoons, rather worse at other times including peak periods. Minor delays at stops do not account for bunching.
- Traffic congestion. Yes we have heard this before. The point, as we have seen in many of my analyses of route operations, is that congestion is manageable, and bunching should only occur when there is an actual blockage of service. See below.
- Traffic signal delays. Yes, signals are being changed to give priority to transit vehicles, but this has already been done on much of the Queen route. Major intersections, where traffic engineers feel that transit priority could be counter-productive by its effect on cross-street traffic, run on their regular cycle.
- Surge loads. Yes, they happen, but they don’t explain routine bunching. Moreover, on Queen, the line uses all-door loading at major stops.
- A shortage of supervision. See “traffic congestion” above. The TTC feels that if it can just put a small army of route supervisors in the field with better technology to let them know where the cars actually are (see Next Bus display at Spadina Station when it works), they can manage the service better. As some comments in the Torontoist thread point out, there is a big problem with operators leaving terminals more or less when they feel like it causing ragged service, and little seems to be done to manage the gaps and bunches out of the service. This happens on many routes.
- Short turns, larger vehicles, more service. This bullet in Giambrone’s presentation is, to say the least, unclear. We know that busy routes have delays and need short turns, although changes in the management style for the 501 eliminate most of the need for this tactic (a point completely missed in the article). Larger vehicles will help provided that the total capacity of the route is also increased. Queen has suffered for decades with the effect of a reduced number of cars providing allegedly equivalent capacity. Between cases where short cars are running in place of long ones (and they get late because they can’t handle the demand) and the larger impact on waiting times of missing cars, the change to wider headways has been a disaster for riding on the line. There is no indication that the TTC understands this problem.
There are three fundamental problems with service on Queen and on Lake Shore (where the original reader comment arose):
- The Long Branch 507 should never have been amalgamated with 501 Queen. The route west of the Humber River has a large amount of local demand, but the decline in service quantity and reliability of the merged route has never been acknowledged. The TTC just does not understand that service is important on the outer parts of lines, not just at Yonge Street.
- The amount of service on Queen is insufficient to provide a reliable headway and capacity for demand. The TTC may point to declining ridership over the years, but this is not reflected on other parallel routes like King or Dundas. The irregular service drives away riders.
- Route supervision leaves a great deal to be desired. For clarity I don’t just mean the guys standing on the street corners making notes on their clipboards, but the whole strategy of how a line and its operators are scheduled and managed. The TTC is working on this, but changes are slow to come.
I have begun detailed examination of Queen route operating records for December 2008 and January 2009, and will be publishing results from this work here soon.
Steve Munro 
That’s what I thought when I read that. It seemed so much like PR spin. And I’ve seen a lot of this sort of PR spin in school, and it’s so frustrating. Just for once I would like to see the head of any organization come out and tell it like it is, instead of hiding behind his/her media training.
LikeLike
Giambrone or Leach: which one’s worse? I say Leach, but only by a hair.
Steve: There’s no comparison, but given the litigious nature of one of the parties, I won’t go into the details.
LikeLike
Traffic monitors shouldn’t be required…we should be able to do it…
twitter #badttc if you see two trains leaving the station at the same time…times/train numbers and pics if you can get them…if it’s a valid reason for them leaving at the same time, then someone from the TTC should respond…
LikeLike
In all fairness, what do you expect him to say? … “yes, our streetcars bunch up because our operators do what they want and we simply can’t be bothered to micro-manage them and make headway adjustments when delays do occur”.
Steve: All those route supervisors need to do a better job of making headway adjustments, not to mention the central CIS staff. Even if they don’t do their jobs properly, it’s sad to see Adam just parroting the standard lines. He could have chosen to address a different question.
LikeLike
This is hardly a problem unique to Toronto. Any major city I’ve visted where a bus route has headways of ten mins or less experiences bunching during rush hour. The streetcars operate like buses, so it’s hardly surprising. The underlying cause is that the streetcar system operates at capacity during rush hour.
Steve: If it were only during the peak and on routes with frequent service, this would not be as much of an issue; however, off peak service can also be erratic even at times when the standard excuses such as “traffic congestion” simply do not apply.
LikeLike
Which is why you should support someone else, like Bombardier, running the TC lines — they couldn’t do a worse job.
Steve: I beg your pardon. “Not doing a worse job” is no qualification for anything. The issue will be whether there is a performance contract, what penalties it contains, and whether they are enforced. At least in the public realm, we have the recourse of angry voters.
LikeLike
Sometimes the fine admiral is quite candid about problems, other times, not so much. I think he probably believes a lot of the old TTC canards about congestion, etc.
LikeLike
Really looking forward to your route analysis! My impression has been that overall service has improved somewhat, but there are still major problems at certain times of day (6 p.m. streetcar die-off, I’m looking at you!).
Not to mention that if two streetcars are running in tandem and they decide to short-turn one, it will almost always be the second one. The first one waits if you’re really lucky. I now assume this will happen, and try to get on the first (packed) one if I can. I’m sure this exacerbates the bunching problems.
LikeLike
Bunching isn’t really a big problem…unless it creates a massive gap in service afterwards!
The problem is that there are simply not enough streetcars on the routes … so unless they are spaced perfectly, it causes massive gaps.
Look back at the old Toronto archive photos of streetcar service along Bloor or along Yonge street. There was always bunching of streetcars … but there was also always another car coming in the distance, too.
Bunching only causes a problem when you’re operating the absolute minimum number of streetcars on the route … if they had the additional cars to add to each route, then this problem would probably be fixed.
Part of the problem is that they’re just stretched so thin along the route that it takes more effort than they’ve had to ever use before to manage the route. When there were more streetcars on the route, they could be lazy in route management because the extra streetcars would mask the problem.
LikeLike
It comes down in part to logistics. You have to plan how to run the streetcars effectively. If you are going ot have some cars turn at McCaul or Church, and head back east again – simply plan on it. Don’t leave it up to changing each day.
Also, if they can place devices on fire trucks to keep traffic lights stay green (or turn green if need be), then the same goes for streetcars. Traffic lights have to set up to allow a streetcar to continue through on a green light (I have been on 501 streetcars that stop at a location while the traffic light is green, and by the time everyone has boarded the streetcar, we are now stopped by a red light.)
Things can be changed – the problem is we need a city council that actually is pro-transit, and not pro car. (It means minor changes sometimes, for example I once said that streetcars on a revived 507 run could turn at Queen and York. While traffic runs north on York today, city council would merely have to change its direction to solve that part of the problem.)
Steve: Part of the problem is that Council generally won’t take on the traffic engineers when they predict doom and gloom if transit gets too much priority. This is changing very, very slowly, but it isn’t helped when the TTC asks for 7×24 reserved lanes on King through downtown while running service that does not exactly dominate the street at all hours. Spadina at least has frequent service almost all of the time including weekends.
As for the 507 to downtown, I think it should be on King Street (as the 508 is), but that’s another thread.
LikeLike
re: Dave
contrast the situation with the subway, where they often run way more trains than are really necessary, say on a Sunday afternoon. Lovely for riders of course. Imagine that on all the streetcars.
Steve: If the ratio of trains to passengers on parts of the subway system were duplicated on the surface network (of whatever mode), there would be creams of outrage that we were wasting a fortune on needless transit service. Only recently, a member of Council was proposing that the recent off-peak service adds should be rescinded as a budgetary measure. The subway represents a huge investment in physical plant, plus the cost of operating the stations and having emergency crews to handle various subsystems like power and signals, that the cost of the trains is comparatively small.
This is part of the LRT vs subway debate we never hear — the huge operating cost of running so much more service relative to demand on the subway lines.
LikeLike
Really? Can you vote out the management and workers of the TTC?
Steve: If you don’t like the way the TTC is being run that much, then change Council and have them change management. The worst problem is that those who would cut costs usually don’t care about the quality of service and regard services like transit as something “other people” use.
Meanwhile if a contract with a private operator isn’t properly written and enforced, then we can have problems with service quality and costs that we cannot control. I doubt that the folks who bought Highway 407 would have done so without having a free hand to charge whatever they felt like. This eliminates any sense of control, any incentive regarding quality and provides instead a reward for greed. Wait until the 407 infrastructure starts to age and we see just how much the owners plan to spend on it without some source of additional revenue.
LikeLike
I drive a Queen streetcar and have never jumped an end. The schedule sucks — during the week (Mon-Thurs) one must drive ridiculously slow to stick to it due to Fri/Sat downtown traffic, it doesn’t change. it is the most frustrating thing, and yes we do get in trouble for getting too far ahead of schedule. They are managing?? it all the time.
LikeLike
Funny, no one ever mentions the attendance of the employees. How one late operator, can screw everything up. Living in an area full of TTC operators who work at Birchmount division, which is by my house, a lot of the new hires has been throwing wrenches into the carefully planned timetables for the operators. I drink at a bar by my house, and some of the guys come by after work for a “few.” They are griping about some of the new hires messing thing up. I am not saying it’s the problem, but it is playing a factor with the routes by my house.
LikeLike
I think Adam Giambrone should look at blaming his fellow employees at the TTC rather than traffic congestion than anything else.
I remember when I was taking the streetcar about a month ago and there was one operator driving really slowly westbound to Long Branch at 5th steet. That same driver decided to switch with another driver going eastbound (the streetcar I was in). I was sitting close to the front and as we passed by the other 5 streetcars that were bunched up going westbound every single driver we passed gave my driver the expression as to why she was driving so slow. My driver just shook her head. So as you can see traffic isn’t the only problem here, it’s operators taking their time as well.
Steve: That swap of operators is an on street crew change, one of those little peculiarities found only on streetcar routes. I don’t know why local 113 hasn’t demanded that they be abolished years ago for safety reasons.
LikeLike
It’s easy to blame the operators, and in some cases, accurate. However, THERE IS NO REASON FOR TWO STREETCARS TO BE BACK TO BACK IN LONG BRANCH, as was the case when I was out there a couple of Saturday afternoons ago. In that situation, incompetence of route supervisors would seem to win the day.
Long Branch is basically an LRT, of sorts. It definitely was that day. While I appreciate what Torontoist is doing with this, that it gives Mr. Giambrone an opportunity to spew the same old crap that is the TTC company line is distressing.
Oh, and Steve, while I appreciate the spirit of the line of voting out council, and telling them to fire the TTC Board and Management, considering this city votes out incumbent city councillors as often as an eclipse, forgive me for not holding my breath. Mr. Giambrone seems to have pissed off a fair number of people in his constituency, so we can hope.
Steve: My distinction about voting out Council may offer faint hope, but that’s better than a private company answerable to nobody with an ironclad contract.
LikeLike
Actually, the 407 is managed quite well. They’ve invested millions in lane expansions to address congestion (expanding it from 3 to 5 lanes) –snow removal is always A-1, and new exits and entrance ramps were added where they were lacking. Yes, the tolls are higher, but they’ve re-invested a lot of that money back into the highway. I don’t know why you’re always against private operation of any sort. With respect to the TTC, in terms of line management it might work if routes were outsourced on a trial/temporary basis. Let’s put it this way, it can’t get much worse.
Voting the “TTC management” out by changing council is a bit of a stretch, don’t you think? You would know more than anyone else that their culture hasn’t changed in over 50 years. When they used to win those “best transit system” awards in the 70s and 80s, they were proud of the system they had and tried to run it well, but now, well, they really couldn’t care less. Why you keep banging your head against the wall in the hope that they’ll change is beyond me. That “culture” is firmly embedded in the organization.
Vic said …
“I think Adam Giambrone should look at blaming his fellow employees at the TTC rather than traffic congestion than anything else.”
You’ve got to be kidding! — you expect the chair of the TTC to publicly criticize his own organization and employees? I’m sorry, but that remark is truly laughable. If he did something like that, he’d be “de-chaired” the very next day!
LikeLike
I think that Wotan’s comment on transit priority, and your response, may be two different items. I think Wotan is talking about the example where a streetcar arrives at a stop at the point where the light would normally turn yellow, but the transit priority algorithm kicks in and the light stays green. Passenger activity at the stop takes so long that the maximum green extension period is reached and the light turns to red just as the doors are closing. In this case the transit priority technically worked as designed but actually made things worse. It would have been better if the signal had gone to red as the streetcar arrived, the streetcar took the 20 or so seconds to load on the red, and was ready to leave once the light turned green again.
A similar problem exists with left turns. A streetcar arrives at a signal and has a car or two ahead waiting to turn left. If oncoming traffic is heavy, cutting off the green at the normal time would allow the left turn queue (and streetcar) to clear, but if the streetcar triggers a green extension, it simply increases the wait for the streetcar. A similar problem occurs where a streetcar is waiting to turn left and is facing a steady stream of oncoming traffic (say, at Main and Gerrard).
Steve: Yes, both of these are problems. The underlying issue is that operators have no way to signal to the traffic controller whether they actually want extended green time “now”, or will be stopping to pick up a substantial crowd in which case it would be better to let the cross-street have the green first.
LikeLike
Bunching on its own is not inherently a streetcar issue (as Steve’s CIS analyses have shown and as bus riders can vouch), or a TTC thing (Google “bus bunching” and you will see rider complaints from plenty of other cities). So Giambrone’s answer isn’t necessarily out of line. Uneven loading does result in bunching, for example. The question is whether there is anything that the TTC does specifically that exacerbates it, in addition to the usual causes that are experienced by any large transit agency.
The one that comes to mind for me is making sure that operators are spaced out evenly (or according to schedule, whichever approach is appropriate) when they leave the terminal. There are enough factors outside the TTC’s control that contribute to bunching — at least there is one thing they can and should be controlling, but all too often we see three buses leaving the terminal within the span of a couple minutes.
Actually, there is one characteristic unique to streetcar bunching — the inability to pass other vehicles. When buses are bunched up, the trailing bus can go around buses at stops, which speeds up service (recently I was in a pack of 4 buses on Vic Park, and because the buses were essentially operating skip-stop it cut off at least five minutes off the trip). When streetcars are bunched up, all streetcars are operating according to the lowest common denominator, the plodding and overloaded lead streetcar that is having to stop at every stop. The answer would be for the lead streetcar to stop to drop off passengers only and let trailing streetcars pick up waiting passengers, but I don’t think the TTC or riders are ready for that idea.
LikeLike
You and Dave mention the difference with the subway, where trains are run at policy headways even if not always supported by ridership. The other difference is that subway headways are micromanaged. Transit Control has a big display showing all train locations and all stations; each station has a countdown timer showing the time since last train departure, and there are significant efforts to ensure that average headways are not exceeded by more than a certain threshold (I think it was 2 minutes). The result of these two measures is that service is frequent (the policy headways) and is reliable (the micromanaging of headways). Both of these factors contribute to the public’s perception of the subway as offering preferable service. The frequency is the most apparent to average riders (especially in rush hour), but the reliability is arguably more important (the Queen car is scheduled every 5 minutes or so apart on paper, but what we see on the ground may be totally different).
Steve: Before the TTC implemented a maximum policy headway on the subway, there were times when the scheduled service on BD was 7 minutes at times. The variations of a few minutes showed up in some very long (by subway standards) gaps. One big problem with the way the subway train dispatching works is that is is relative to scheduled departures. If the whole line is late, then it pushes out trains as fast as possible with no heed to actual spacing. To compensate for this, the controllers at the tower must adjust for the offset between scheduled and real time and this seems to be the prevailing way the line is run today.
For minor delays, they can just let the system push trains out quickly until everyone catches up by using the padding built into the schedule. For major delays, the controllers must intervene.
LikeLike
The government operates the TTC, yet there is no sense of control, there is a lack of quality and greed (cost-cutting) is definitely rewarded.
If drivers feel the tolls are too high, then they have a free hand to not travel on the 407. Lack of quality? You mean like ADDING more lanes and more on-ramps/off-ramps? You mean like providing free roadside assistance? How evil of 407 ETR respond to demand. Your beloved TTC is government operated and can’t minimize bunching.
So what if the owners raise tolls to replace infrastructure? What is wrong with the users of the road paying for its upkeep?
LikeLike
I hope planners and engineers address these problem we see on our current streetcar routes before they build Transit City.
LikeLike
I witnessed a positive example of terminal departure management recently, so there seems to be some hope. At Spadina Station there was obviously a prior line blockage as indicated by the GPS locator screen and by the number of people waiting. The first car in (signed for King) let out its passengers, turned off the interior/signage lighting to indicate not in service, and promptly left the station to make up time and help clear the passengers further down the line. The next car in (signed for Union) dropped off and picked up as normal as did the other three behind it. I know the first car remained in service because I saw it twice later on.
I realise this was a response to a specific incident, but it was a step in the right direction. There’s lots of bunching on this line, but it seems to have more to do with the service density and intersection signalling than anything else. This level of service is probably as near to a subway as we’ll ever see on the surface lines. It’s hard to say whether or not the apparently abandoned plan to couple runs in pairs would have made the situation better or worse.
LikeLike
I was just woundering the Spadina street car always seems to be busy. Does it have enough ridership to make it a subway?
Steve: The peak service on the Spadina route is about 30 cars per hour. The TTC uses a design standard of 74 passengers per car, although a crush load is considerably more. In practice, the crush load cannot be sustained for long because it makes for severe loading delays, and space at the rear of the car tends to be underused as passengers get off at the rear doors, but are not replaced by new riders from the front. This gives a service capacity of about 2,200 passengers per hour. More could be carried if the TTC ran larger cars on Spadina, but this would still be nowhere near the level of 10,000 usually considered a minimum for subways.
There are two additional problems. First, Spadina works as a route because the stops are close together and it can serve many destinations. As a subway line, this would not be as easy unless it were built like the Yonge route downtown with many stops. That adds to both construction and operating cost.
Second, there are physical limits to the size of car or train that can operate on the route due to platform lengths. Spadina Station platform will be challenging when the new cars arrive and two 30-metre cars attempt to be on the platform at once, one unloading and one loading.
LikeLike
One of the reasons for bunching is the problem of traffic light cycle time versus headway. A streetcar on a six minute headway can get bunched up by traffic lights that are on a slightly different frequency. The Spadina car has this problem with the lights at Front and at Lakeshore. These lights are on a much longer cycle time than the rest of the downtown and if one car just misses both while the trailing car makes both this can reduce the six minutes to three between these two cars.
Steve: This is a huge problem on Spadina/Harbourfront which has to deal with some very long cycle times at various intersections with Front and Lake Shore. I keep hearing about how there will be a new era of enlightened (pardon the pun) traffic signal priority for transit, but nothing concrete has occured.
In the 60’s and 70’s the old city had an 80 to 90 second cycle time at major intersections and double this at some in between side streets but the green phase for the side street was still short. A vehicle that just missed a light only had to wait 40 seconds for the next one. With the longer cycle times they can now wait upward of 90 seconds at some multi phase intersections. The far side stops on Spadina and St Clair seemed designed to ensure that a vehicle just arrives at the next intersection as it goes red and it has to sit for the entire cycle and then go through and load. The St Clair cars that I have ridden on seem to over come this problem by driving at 15 km/h. This gets them to the next light as it turns green but the passengers die of boredom.
Steve: Yet another example of how the “transit priority” signals may be active in theory, but in practice work to thwart fast transit operations. The whole idea of farside stops is that cars get a green as they approach the intersection, and then load with the red phase “behind” them. The road designers love farside stops because the “shadow” of the loading island creates space for a nearside left turn lane. However, the signals are not arranged to suit the stop arrangement. Another problem we have is the proliferation of closely spaced traffic signals at locations where there is no car stop. It is claimed that call-ons for green time can only be triggered when a streetcar is in the block facing the traffic signal, and that as these blocks shorten, the advance time to cycle the cross-street in transit’s favour becomes longer than the time needed for a car to reach the intersection.
This is hogwash, and especially with GPS-based systems where we don’t have to dig up the road and install new detector loops every time we want to adjust the control scheme, the City and TTC should be doing much more experimenting with transit priority that reaches more than a few hundred feet upstream of an intersection.
A related problem will no doubt be found with the antique CIS system on the vehicles. This system works by polling each vehicle every 20 seconds, but that is too infrequent to deal with closely spaced signals and makes no provision for a proactive “I want it now” type of signalling from a vehicle.
The last two times I was on Queen’s Quay the operators either had learned how to time the lights or the cycles have been adjusted to give more priority to transit. The north south green times are still too long and they are longer than what is needed for pedestrians but they said they understood this at the Queen’s Quay meeting and were going to play with them for this summer. They also said they were going to enforce the no parking and tag and tow illegally parked vehicles and move the tour busses to off site parking.
I am going to go into the downtown with a stop watch and check cycle times in May when it is warmer and compare them to what I got when I did a project in a traffic engineering course I took in the early 70’s. I don’t know for sure but I believe that the cycle times have lengthened at most intersections.
LikeLike
Hi Steve and Brent:-
From Brent’s comment on the 10th re the VP bus leap-frogging and his comment at the end about the idea of the lead streetcar stopping only to exiting passengers I am thrilled to hear someone else encouraging this kind of operation on TTC routes. There was a period when TTC operators were not allowed to leap-frog those ever so flexible diesel buses for fear of receiving management’s ire in so doing, but with your note on VP and my own recent observation on the Ossington route I’m happy to see that it appears that this heavy handedness appears to be easing, thus allowing right thinking drivers the opportunity to exercise some seat of the pants customer service.
Too, I commented a few months ago about the same idea as yours Brent, that the lead streetcar, if not needing to stop for alighting passengers, shouldn’t, if there is another car on it’s anti-climber. The car immediately behind can look after the passenger at the stop. My example related a day when I was in Hamilton on an extremely miserable heavy snow day to ride and photograph the trolley coaches and witnessed this kind of operator co-operation. The weather played havoc with the service on this particular day and all of the factors that can lead to a vehicle being delayed contributed to this one example I was privy to on the Barton route. Since trolley coaches have a similar characteristic to streetcars in being tied to their overhead, unless the lead coach has its poles pulled to allow another to slip past, then the skip-stop method works well and indeed on this particular day the skip stop surprised and delighted me in learning that in Hamilton, unlike the big ole ‘Hogtown’, the operators were allowed to think for themselves and their riders and thus supply a better service than that dictated by the powers that be in TO. Our TC arrived downtown more quickly than it would have, coach spacing was partially restored and no one was upset with the operator’s choices.
This said though, in TO the hew and cry from the ridership here would be loud and long, for as sure as shootin’ here the leading Carlton streetcar would go through to Main Station while the following car(s) would be short turned at Coxwell to re-establish a downtown headway. So much for what works in the real world and what can’t possibly in Torawna! Brent I believe the riders can handle service improvements here; you may be right about the other factor though!
Dennis
Steve: I was reminded this morning of the problem of that “downtown headway” on which so much effort is concentrated. The TTC does not seem to understand that many routes don’t just have demand at Yonge Street, and that they perform local service functions elsewhere. A particularly striking example is on my local route, 504 King, which is in effect also a Broadview car and a Parkdale/Roncesvalles car. Each of these areas has its own demand and boardings that go nowhere near Yonge Street. All the short turns play havoc with service reliability. The folks on Lake Shore suffer a similar fate, although those in the Beach benefit (for the time being) with intensive efforts to maintain good service there.
LikeLike
There could be a mathematical reason (i.e. control systems and chaos theory) for groupings of streetcars.
http://en.wikipedia.org/wiki/Bus_bunching
Steve: The article cited here in fact states that mathematical theory does not explain or predict the actual behaviour of buses, and gives a number of alternative reasons that closely parallel observations here in Toronto as well as my own findings from review of TTC vehicle monitoring data. Although the citations are few, I must caution about circular references as the date of this article follows the publication of my own work.
LikeLike
One of the problems with transit priority signals on wide streets like Spadina, St. Clair, Queen’s Quay and Lakeshore is the presence of pedestrians. These pesky devils require 1 second for each m or roadway that they must cross; not all are members of the local track team. If you have a 5 second steady walk signal begins then you need another 25 seconds to clear the intersection which means that you minimum green time foe each cross street on Spadina is 30 seconds. (Spadina has about 18 to 19 m of traffic lanes and 6 to 7 m of transit lanes so say 25 m total). This is the length of most of the cross street green phase now. You cannot shorten it with a transit priority signal without wiping out a lot of little old ladies and men and possibly me.
University Avenue uses shorter cross street times because the people have a centre median where they can wait for the next green signal. In order to cross University safely in one light you have to facing the on coming traffic on the cross street, start as soon as the light turns green, walk quickly and you will finish the last two lanes well into the amber or red, but the cars are stopped at the far side of the street you are crossing. If you go the other way you will still have a lane or two of traffic to cross when the light goes green and the cars are there with you. About the only thing the traffic priority signals can do is put up the lunar white street car turn phase. If they put the Queen’s Quay right of way on the south aside of the street then there will only be a 12 to 13 m wide street to cross and they can shorten the cross street green phase accordingly.
The intersections of Spadina and Front and Spadina and Lakeshore are another matter. These are designed to maximize the numbers of cars that get through the intersection and the rest of you be damned. With the number of left turns to get onto or off of the Gardiner there is not an efficient way to shorten the light times. Perhaps they could put in a quick transit phase at Lakeshore before the south bound auto traffic gets its long left turn phase. This would speed up the transit. They would also get their existing phase as well so they would have two cross cycles for each complete road cycle.
I don’t think the northbound left turn phase at Front is long enough for this to make much of a difference. Most of the cycle time is for the west to south left turns. The one thing that the Front Street extension had going for it was that it would get rid of south to west turn at Front and Spadina. Perhaps the province can build a super ramp from Front St. just west of Spadina to the westbound Gardiner. If this turn is not eliminated or replaced by something else then it will always be a nightmare. Since there are no westbound left hand turns from any major east west transit route onto Spadina all the autos go south to Front and use it.
The mathematical concept that predicts bunching of street cars or any other group is called queuing theory. That is why the checkout counters at the super market are empty when you enter and while you are shopping but are lined up three deep when you want to leave.
Steve: This is related to a behavioural phenomenon in which everyone queues for a popular brunch spot just before it opens because they assume it won’t be busy. The really canny gourmets try to arrive as the first wave is all finishing, and the late risers haven’t yet found the need for their bowl of cafe au lait. Just as the Queen car can be fouled up by unexpected events, the presence of a distracted waiter can alter the expected service times and behaviours of the diners.
If you choose your brunch spot carefully, you can observe all of these phenomena at work both for your meal and the passing Queen cars at the same time!
LikeLike
Further to my comment on queuing (or queueing) theory. If you really love all those nice mathematical symbols at the start of the show “Numbers”, if you are suffering from insomnia or have sado masochistic tendencies then read the first web site below. If you want a more humorous look check out the second site especially the section on queuing humour. A google of queuing theory or queuing models will give you a ream of information, most totally incomprehensible unless you are a math geek.
Click to access qt.pdf
http://web2.uwindsor.ca/math/hlynka/queue.html
Steve: Only a masochist would read it. A sadist would make someone else read it and explain it at a car stop swarming with would-be Queen car riders.
I used queuing theory to model traffic flow through intersection in grad school but when I looked these up again it was an instant headache. Queuing theory basically says that if you have 60 randomly travelling cars arriving at an intersection per hour you will not have one each minute but groups of different sizes with large gaps in between. Transit vehicles are not supposed to be random but rather they are supposed to be well behaved. Unfortunately the rest of the traffic is random and not well behaved. Also the number of passengers for each vehicle is not well behaved.
All these facts plus the problem of having a headway that is not an integer multiple of the cycle time for the traffic lights causes the transit vehicles to lose their even spacing. The longer the line is or the more cars on the line the more pronounced the problem becomes. This is what is happening on Queen and Spadina. The subway tends to be a well behaved system because the trains are dispatched regularly, held at certain stations and have predictable dwell times in stations. When a car breaks down or a passenger becomes ill then the system is no longer well behaved.
Queen is too long to run in mixed traffic without suffering from a series of random events like a car making an illegal left hand turn or truck parking and blocking the track. These problems are easier to control on a shorter line because they happen less often and the problem does not keep building.
I have driven in Toronto recently on Lansdowne, St. George and Harbord which have been reduced from two lanes in each direction to one with parking, bike lanes, bump outs to mark the end of parking and the start of intersections with turn lanes. This seems to force the traffic to be better if not actually well behaved and the traffic seems to flow more smoothly. I first noticed this on Spadina when the number of lanes in each direction was reduced from four to two with designated parking and turning lanes. The traffic flows much more smoothly now than before except when a truck parks illegally at a bump out to make a delivery. At least it is not mucking up the streetcars. There needs to be more ridged parking enforcement with harsher penalties for vehicles that park illegally and disrupt traffic and transit. I am waiting to see what happens on Roncesvalles if the proposal for reducing it to two lanes like Lansdowne goes ahead. I think it will work wonders.
Proper management of the line is important, especially when a disruption occurs. I was in London when one branch of the Green Line had a power failure. The system did not miss a beat. The transit controllers just sent all the cars to another branch and turned them on it so that they came back in their proper place. The fault was cleared in about ten minutes and service resumed on that branch had no impact on the rest of the line. Toronto needs more pocket tracks where a train can be turned and/or held to clear a disruption and get the cars properly spaced, especially in the return direction.
Steve: By analogy to subway operations, the TTC does have the option of spacing cars, if not precisely, at least reasonably at dispatch points such as termini and at points along the route where traffic is not so dense that a car stopped to kill a light (not an unknown phenomenon) won’t cause complete chaos. There are two major problems, however.
Cars do not leave the termini properly spaced, and the gaps just get bigger as cars move along the route. Pairs form because cars catch up to each other. This happens under every weather condition including ideal days when traffic (both passenger and auto) is low.
The other problem is that the attempt to regulate this through CIS is botched in two ways. First, the system only tells an operators that they are late or early relative to the schedule, not to the actual prevailing situation on the route. Second, the running times assumed for calculating intermediate schedule points don’t always match actual conditions on the street. Operators are flagged for running “early” when they are just trying to get a buffer for a known congested area further down the line.
We will never get rid of the random events, but there are many ways to change line management that would improve service overall.
LikeLike
In terms of crosswalks, at the transit priority intersections such as St. Clair and Spadina where streetcars operate in a ROW and left turn movements are controlled via separate signals, it would be very useful to install crosswalks on either side of the streetcar ROW, so that for example at Bathurst and St. Clair, I want to catch a WB streetcar (assuming they ever come back…), but I am standing at the northeast corner and Bathurst has the right of way. I could then cross 1/3 of the way and wait for the next cycle to cross Bathurst and reach my platform as opposed to using three light cycles and likely missing my car.
LikeLike
Bunching happens on any system. The question becomes how to manage it so that customer experience will remain positive. The Spadina trams bunched up, but it is still acceptable. I take it all the time and I usually do not wait more than 2 minutes before a tram shows up. Even on a computer controlled line like the Scarborough ICTS, bunching still happens. Just go there at around 8 AM in the morning and see what happens at Kennedy where trains line up waiting to enter the station.
Of course computer controls are far better than CIS or GPS based tracking. CIS and GPS systems cannot calculate the vehicle load capacity and speed. GPS is a one way communication link. Spread spectrum radio signals used in the Bombardier CityFlo systems are two way. If vehicle speed is known, it can increase the accuracy estimated arrival times. GPS signals are also not as accurate in poor weather.
The best way to combat bunching is not shot turning vehicles, but to build exclusive right of way structures for transit. Sometimes, the best way to fix bunching is to hold the vehicles at a station. On Queen St, this is not possible. It is somewhat possible on a line like Spadina, but the traffic lights makes it much harder to run the vehicles at a higher speed to catch up. However, if run trams on Eglinton on a dedicated guideway, it is much easier for a computer to hold a tram at a station or speed up another one.
LikeLike
Jonathon Says:
April 11th, 2009 at 6:24 pm
“In terms of crosswalks, at the transit priority intersections such as St. Clair and Spadina where streetcars operate in a ROW and left turn movements are controlled via separate signals, it would be very useful to install crosswalks on either side of the streetcar ROW, so that for example at Bathurst and St. Clair, I want to catch a WB streetcar (assuming they ever come back…), but I am standing at the northeast corner and Bathurst has the right of way. I could then cross 1/3 of the way and wait for the next cycle to cross Bathurst and reach my platform as opposed to using three light cycles and likely missing my car.”
Unfortunately there is no room on the north side of the westbound streetcar lanes for you to stand. If there were the road would need to be 3 m wider and this is not going to happen. You could stand on the end of the eastbound platform and then cross Bathurst on the south side of the right of way and get to the westbound platform. You would not have to worry about any turning vehicles just watch out for the trolley.
LikeLike
What drives me really nuts is when two routes meet where one uses near-side stops and the other uses far-side stops. Almost always my connection requires me to cross twice. When both routes use the same stop placement I only have to cross once for either transfer direction.
I realise that when near-side/far-side combine there are two transfers that don’t require crossing the street at all. However, when one route has a ROW, I still have to wait for a walk signal to cross half the street to reach the island. Intersect TWO ROWs with far-side stops, as in some future designs, and that signal-less transfer now requires you to wait for two walk signals anyway. Witness the amount of jaywalking on Spadina to guage the level of frustration caused by the loss of the vehicles’ ability to stop traffic while passengers are exchanged. If the public adopts this suicidal behaviour on some of the larger arterials in the Transit City plans, there will be a significant increase in accidents.
LikeLike
I have not read every letter on this subject here, so hopefully I am not repeating information.
I operated streetcars from 2006-2008. Although I never supervised a route, its my understanding that Supervisors have been informed to avoid overtime hours for streetcar operators. This means ensuring that each operator ends their shift early or on-time. This often leads to short-turning streetcars. Allowing a late streetcar to make it to the end of the line would result in paying overtime to the operator.
Short- turning a streetcar can often result in disasterous bunching. Here is why; most short-turns are done on tracks that are off the main routes and these turns involve manual switches. Even the most seasoned operator can run into a myriad of problems with manual switches. These problems can result in serious time wasted allowing numerous streetcars behind to catch up in bunches. Problems with manual switches are numerous; since they are not used frequently they are often caked with dirt and debris which needs to be swept out with a broom to ensure a switch point makes proper contact. Even if the switch is clean, some can be quite heavy or stiff and can require several attempts to move. Remember that the operator is doing this task in the middle of a downtown street. Staying alert to the dangers of ignorant and inattentive drivers is not a recipe for a speedy undertaking.
Before the switch is moved, the operator has to off-load the passengers from the streetcar. Again, this can often take a lot longer than you would think. Telling a streetcar with 40-100+ passengers that they all have to get off at the next stop usually requires repeat anouncements. Following this comes the explanations demanded by certain passengers, and there are usually a handfull of confused passengers who need special attention and directions for how to complete their journey. Then there are the people who do not understand English. There are always a few. Not to mention the people sleeping or who are drunk or generally confused.
So, you finally get the people off your streetcar, deal with a stubborn switch, then as you get the streetcar moving around the corner, as often happens, your pole falls off the power line.
Sometimes an operator can execute a short turn without missing a beat, sometimes, real life gets in the way. This is definitely one of the reasons that bunching happens.
Steve: Yes, definitely, any routes that are used for short turns should have electrified switches that are kept clean by track crews. The TTC seems quite selective in where it installs these switches, and several that are in place don’t actually work. There has been a project in the capital budget for years to replace the faulty systems that control these switches, but nothing has actually happened on the street. Meanwhile, we have a stop-and-proceed rule at all facing point switches, even the manual ones, to correct for the unreliability of these switches. Imagine if subway trains had to do this!
LikeLike
I’m wondering if the Termini is the place for a supervisor, the idea being that if the headway is 2 minutes, and 3 cars arrive at the same time, dispatch one car, wait 2 minutes, dispatch the next car, wait 2 minutes and dispatch the next. Rather then having a set time for each run to start, the idea being if at each end of the line you can re-establish the headway, that it would never get very far off. The difficulty would of course be where there are cars from different routes using the same platform, where you really want the third car arriving to go first. Of course at corners like King and Spadina where multiple cars from different lines join together, you would need the ability to hold a car until the other car goes by.
LikeLike
I think one of the non-functioning switches is at King and Church as I often see a TTC worker stationed there just to operate it. The trackwork here was rebuilt about 3 years ago so this switch is, presumably, quite new. Because Church is not only a common short-turn point but also on the route of the 508 heading along Wellington I find it strange that the TTC finds it better to hire additional staff to operate the switches rather than simply fixing them. Sometimes it is hard not to think the TTC might be better run by others, such as Metrolinx!
Steve: Ah but Metrolinx would object that all of those streetcars stop far too frequently, and that if only they ran express from Broadview Station to Dundas West stopping only at Yonge and (maybe) University, people would be able to zip across the city so much faster. This all seems to be part of a general attitude that making the surface system work, especially the streetcars, is something to be done “later”. I suspect part of this developed during the dark years of the 1990s, and the organization permanently learned a bunch of very bad habits. I would be interested in hearing from any who were “on the inside” to comment on this.
LikeLike
I’d like to thank Aaron for his insights. It’s refreshing to hear once in a while from the people actually in the driver’s seat.
His point about short-turning cars to avoid paying OT is well made. While I’m sure a lot could be done to make the physical turn faster, let’s pause and consider whether this should happen at all. Why not schedule the drivers at 15 minutes less per shift, but expect them to continue for 30 minutes or so on the bad days when we need to keep their car running. The net OT would be small (unless every day is a bad day) and the 40-100 passengers on the car that isn’t short-turned would be very happy.
Steve: The TTC is working on various schemes to avoid linking a car (and its passengers) with the specific needs of an operator, but it’s taking a long time just to try it out on Queen, let alone expand it to other parts of the system.
LikeLike
When I read the TTC’s report about their experiences with the alternate crewing arrangements on Queen, it became immediately clear that the bulk of the problems on the route were directly caused by turning operators to keep them on schedule. The statistics speak for themselves, and they contradict most of the tired old lines about traffic congestion and loading delays.
I think it is more than fair to ensure that the employees get home on time, and unnecessary overtime would be wildly costly over the year. But that doesn’t mean there can’t be substantial changes made to the crewing practices to ensure that meeting those goals doesn’t introduce problems into the reliability of the system. It would be much easier to tackle the other every-day delays and stabilize the service without these additional crewing-related issues churning things up.
LikeLike
Wouldn’t things be better if they had switchbacks at endpoints. In other words, when a streetcar arrives at its destination, there is an operator there to take it out again immediately without the need of waiting for the operator’s break to be over. The operator would have his/her break while waiting for the next car to come in, and then take that streetcar out right away. It would seem to very well for the 501.
LikeLike
People have mentioned short turns and King & Church and that got me thinking about things I’ve seen. I don’t know if they’ve fixed things or not but one thing that I remember is poles frequently coming off the wire when going from eastbound King to northbound Church.
I’ve seen countless short turned King cars and Queen diversions lose power on that curve. I’ve even seen it happen twice (!) to the same streetcar as it was making its way through that turn.
In each case the operator would spend several minutes stopped in the middle of the intersection while he or she tried to get the shoe back on to the wire.
I even remember a few years ago (around 2003) when a part of the overhead at that intersection was wrecked (by a loose pole maybe?) which backed up the line for hours while a wire crew patched it up.
Could poor overhead like this be a factor in delays? It certainly makes me wonder how many frequently used spots in the system are like this.
Steve: I don’t think it’s a major factor in delays, but it certainly complicates short turns. There lesser-used track and overhead doesn’t get attention as often as the main lines, and this can cause problems for short turns and diversions.
LikeLike