basics: dead running

P1010471 The British/Australian term “dead running” means “running out of service, unavailable for passengers.”  I like the term because it could be the title of a zombie movie.  I look forward to seeing if it attracts hits.

The North American term is “deadheading,” which may remind you of rock fans of a certain era.  In both cases, the word “dead” is apt.  It’s a waste, a loss.  It’s unavoidable up to the point, but transit agencies are always trying to turn dead time into live, useful time.

The Brisbane Times has a piece by Marissa Calligeros announcing that ’28 per cent of bus services are “dead running”, where passengers are unable to board.’  That’s a confusing way of putting it.  The reporter must mean that Brisbane buses are dead running 28% of the time.

That sounds like a lot, but “dead running” is the result of two different issues that need to be kept separate.

  • All transit vehicles must travel between their operating bases, where they are stored and maintained, and the beginning and endpoints of service.  Rail services usually have bases directly on the rail line, but may still have to dead-run through the rail system to reach their trip’s starting point.  Bus bases can be anywhere.  The location of bases (called depots in Britain/Australia, often divisions in the US) is a major issue.  It’s often worth spending capital money to save operating money, and careful investments in bases, reducing dead running, can do that.
  • Transit agencies that run extensive one-way express incur massive amounts of dead running.  Brisbane, for example, is a very, very centralized city, with a downtown far out of scale to anything else.  That means a huge demand for one-way trips into downtown in the morning and out in the evening.  All those services that are needed in only one direction usually have to get back in the other direction so that the driver’s shift can end where it began.  (The other alternative is to pay the driver to hang around downtown all day, which is even more expensive.)

There’s no way to understand Brisbane’s “dead running” issue, or that of any other city, without separating these two causes.  The Brisbane Times article talks only about the first, but the Brisbane bus system has a massive one-way peak due to its single dominant downtown and relatively lack of direct rail paths for much of the city.  Can dead running be better addressed by a rigorous review of whether these one-way peak services can be combined, replaced by links to rail, or otherwise made more efficient?  GIven the higher cost of dead running for one-way peaked service, could some of it be converted into two-way, all-day service at less expense than it would first appear?

That’s an important question for any city.  Details of driver costs vary, but in general, dead running is one of the main reasons that one-way express service (bus or rail) can be more expensive than it looks.

48 Responses to basics: dead running

  1. Daniel June 9, 2011 at 10:25 pm #

    “so that the driver’s shift can end where it began”
    And presumably so the bus can be parked somewhere. It might not be unreasonable to ask the driver to catch another bus to where they need to be, but large numbers of buses can’t be parked in the CBD until the evening peak.
    PS. The BT article’s author was Marissa Calligeros, not Matthew.

  2. ant6n June 10, 2011 at 12:00 am #

    I’ve always wondered whether the extra cost of turning dead heading into reverse commute service would pay for itself in the sense that the extra cost due to the delays from stopping and not going directly is less than the new riders that could be attracted. Both numbers are probably small, and it could probably go either way, depending on the situation.

  3. GD June 10, 2011 at 12:57 am #

    Vienna’s streetcars carry passengers while deadheading at the end of the day, which can be confusing, but is popular with people living along the route (then again, is there real deadheading when the depot is in a rather densely populated area?)
    The same thing applies with some subeway lines that have their yards in a semi-central location. After a certain time, most trains only go half the way, though the penultimate train usually runs the entire route.

  4. Nicholas Barnard June 10, 2011 at 2:01 am #

    I have the same question that ant6n did.
    The other side of the question is what other cargo can be carried? I used to work in the trucking business and we HATED deadheading, although it was necessary.
    If say deadhead runs were filled with mail (don’t laugh, put the mail in a mail bag and put it on a seat.) would this be doable and make sense?
    The basic problem is imbalance in demand which is often made up for by charging more for peak services.
    This also shows up in the trucking business. I used to work for a banana company when we had to put bananas into Denver we paid through the nose on a per mile basis. Why? Denver makes very few physical products, but requires physical products (such as bananas!) so truckers charge more because they’ll be deadheading out of Denver to pick up their next load..
    I’d argue the push for transit agencies to charge one fare at any given hour for a given distance helps contribute to deadheading. An real world example using the my commute from home to church in Seattle:
    I live downtown, the church is near the intersection of 35th Ave NE and 65th Ave NE. The route that runs all the time between these locations is the 71. The 71 is part of the 71/72/73/74 series that connects downtown with the University District. During the day and peak hours the 71 and its fellow route mates make no stops between downtown and the University District, running “express” either on the Interstate or local roads, depending on the exact details of the trip. But, in addition to the 71, there is the 76 that runs at peak hours only and in only peak directions. This route bypasses the University District only serving the tail end of the 71’s service area. (The routing is identical after both buses reach 65th Ave NE.
    The 76 by definition requires more deadheading than the 71, but when the 76 is running the fare for a ride on it is the same price as a ride on the 71. The 76 could, and should have a higher fare, but the operating agency King County Metro (KCM) doesn’t do this. (Not to say they don’t have good reasons, the one I can think of off off the top of my head is rasing the price for the 76 would push some trips to the 71 which is already near or at capacity during the hours they run at the same time.)
    Sure KCM could just replace the runs of the 76 with runs of 71, but this’d likely result in a higher cost, because a 71 requires more time, and you don’t need the peak hour trips on the way back, so you’ll still need to deadhead them.
    Deadheading can be reduced, but I strongly doubt it can be eliminated to the levels needed.

  5. Max Wyss June 10, 2011 at 2:22 am #

    There can be good reasons for deadheading over running with passengers for reverse commute. And this reason is when the deadheading bus can run another extra/express service. In such a situation, the time to get back to the originating point is critical, and must be as short as possible.
    A possible solution for the “type 1 deadheading” when the bus line is relatively far away from the operation base, and the driver change shift is that the operator has a dedicated passenger car, which is normally at the operation base. The driver starting the new shift takes this car and drives to the point of the bus route where he will take over the bus, and the driver finishing the shift will drive that car back to the operation base. With such an approach, the bus itself will remain on the line for more than one shift, and deadheading in a small passenger car is way cheaper than deadheading the bus (and helps keeping the number of vehicles in the fleet lower).

  6. Max Wyss June 10, 2011 at 2:27 am #

    With rail systems (light rail, streetcars, to some extent also trolleybuses), it is pretty common to have the operation base close to an active line. And it does make sense to let passengers in from/until the closest stop to where the access to the operation base happens. In some cases, this may lead to a few interesting direct connections, appreciated by the regular, and not hurting anyone.

  7. Zoltán June 10, 2011 at 3:10 am #

    “could some of it be converted into two-way, all-day service at less expense than it would first appear?”
    That’s precisely what the trend has been in Leeds – where, because buses are privatised, the commercial imperative defines everything. There used to be a number of one way express bus routes, nearly all of which have been sacrificed to provide more frequent local service on arterials (as stops are rigidly kept about 400m apart, that’s not excessively slow).
    The only express services that remain run all day to and from outlying towns, mostly every 20 minutes or better.

  8. Zoltán June 10, 2011 at 3:19 am #

    @Max Wyss
    A few points in response:
    – Where a second run must be made shortly after, it still might be possible to accept passengers on the non-stop trip to the terminus, and then allow those passengers to alight as the bus traveled back past their destinations – essentially turning the route into one large circuit.
    – The passenger car-based solution to deadheading is precisely what Go North East, a private bus operator in North-East England, does. They have a number of what they call “staff van” duties, where a driver takes a small vehicle to the bus terminus where they need to take over from another driver.
    – You’re right about rail carrying passengers to and from the closest stop to the terminus not hurting anyone. Hence in Glasgow, though Garscadden and Scotstounhill stations on the way to the EMU depot are only reckoned to merit a 15-minute frequency during the day, it does no one any harm to provide them with a 2-5-minute frequency early in the morning, just because that’s how many trains need to leave the depot that way, and that is precisely what is done.

  9. John June 10, 2011 at 4:50 am #

    I have heard that Metra used to (maybe still does) have some lines where the conductors would operate the train inbound in the morning, go to work at an office job just like the passengers, and then work on the train going back out in the evening. Assuming you can find somewhere near the CBD to put a bus for the daytime hours, I wonder if using more part-time/peak-only employees to drive buses would help with this problem. Even if you have to put the bus a couple of miles from downtown, you could shuttle the part-time drivers into the CBD with a van or something. It’d be cheaper than hiring full time drivers and sending them dead-heading back to the suburbs.

  10. Paul K McGregor June 10, 2011 at 4:57 am #

    In Austin, it was the operating rule that every bus is in service even if it is pulling out to its route. Each route has an established deadhead route that it follows to get to the start of its route so anyone wanting to go where that bus travels would usually have the chance to ride. The only downside would be that if you pick up too many passengers, you might not get to the start of the route at the designated time. But usually we were talking about a few people.
    For light rail, the ideal case would be to have a yard at the end of the line. Most times that is not the case. Our yard here at UTA is towards the southern end of the line so as soon as the train hits the mainline, it goes into service. Same with the pull in. We let people ride to the last station before the yard track and them get them all off. We provide alternative transportation if they want to go further.

  11. Matt Johnson June 10, 2011 at 5:46 am #

    I don’t know that this is still the case, but a few years ago, I came across what I believe to be one of the worst examples of deadheading possible.
    Atlanta’s newest transit operator is GRTA (pronounced “Greta”) Xpress, which operates motorcoach express buses from the suburbs. The county I grew up in was one of the last to join the system, so we got buses late.
    By that point I was living in the city, but I could use the PM outbound buses on Fridays to go visit my parents. Once I got to chatting with the operator at the end of her shift, and discovered that the buses for most of the routes are based in Forest Park, south of the city.
    So this route, #490, starts in the wee hours of the morning at a bus garage in Forest Park. It then drives north, through the city, to Canton, a distance of over 50 miles – without passengers. The bus picks up passengers in Canton, drives an express route to Midtown/Downtown Atlanta and disgorges passengers. It then drives 10 miles or so back to Forest Park to spend the day. In the PM, it repeats this in reverse.
    In the case of express buses that are peak direction only, it would seem to make the most sense to base the buses near their suburban terminal or perhaps less ideally at their downtown terminal.
    But #490 had (not sure of the current schedule) only 3 daily departures, so a base doesn’t make sense just for that one line. Especially since it’s 15-20 miles to the next nearest suburban terminal (circumferentially).
    Anyway, basing the buses on the far side of the city seems a bit ridiculous. But GRTA doesn’t run enough buses to have too many bus garages.

  12. Aidan Stanger June 10, 2011 at 7:15 am #

    The Brisbane Times article talks only about the first, but the Brisbane bus system has a massive one-way peak due to its single dominant downtown and relatively lack of direct rail paths for much of the city. Can dead running be better addressed by a rigorous review of whether these one-way peak services can be combined, replaced by links to rail, or otherwise made more efficient?

    Brisbane CBD may appear dominant, but when I tried to check what its primacy was, the best figure I could find was a quarter of the Brisbane LGA. There’s no reason to give up on the other three quarters.
    Looking at the route map of their BUZ network (where the deadheading reportedly is), a couple of things stand out: firstly they do serve significant non-CBD destinations such as universities and hospitals. Secondly, even with buses running in the reverse direction, the lack of cross city services may discourage passengers going to other destinations.

  13. Aidan Stanger June 10, 2011 at 7:34 am #

    @ Paul MacGregor
    I’m impressed. I thought London was good at avoiding dead running (with all contracts tendered out by individual route, usually to the company with the depot nearest the end of the route as they have lower costs and the contract specified service requirement is the same in both directions on nearly all routes). London even has some longer distance buses (from Essex) parked in Westminster all day – or at least they did ten years ago and I presume they still do.
    Does anyone know what proportion of bus milage in Hong Kong is in service?

  14. Alan Robinson June 10, 2011 at 7:53 am #

    A few years ago when they were based at Oakridge, Vancouver’s trolleybuses almost all ran in service straight from the depot instead of deadheading to their route. I believe they still do this from the new depot. For trolleybuses, deadheading doesn’t make sense unless there is a dedicated overhead power for out of service vehicles to pass other busses on their way to and from the depot. 41st/Oak anyone?

  15. D Matyas June 10, 2011 at 9:15 am #

    I work for a small bus transit agency in Massachusetts, USA (UMass Transit in Amherst and we generally have very little need for deadheading due to the way our routes and garage are set up. Our situation is pretty unusual in that we are in small college city in the middle of nowhere, with lots of driving through farm fields and forrest (our service drops to 1/3 full service during the summer and winter breaks). Because of this our practices are probably less feasible in many other cities, but I think it’s still a useful comparison.
    Most of our routes radiate from the campus with the campus in the middle of the route and ends in apartment complexes or nearby town centers. Generally routes start and end in the middle of that loop at campus because it’s close to the garage and one of the major sources of traffic. Because of this most of our deadheading is less than half a mile and even then we’ll still sometimes transport people along the way on campus back to the garage to supplement the campus shuttles.
    While I can only think of one deadhead longer than a mile or two the ones that we do have are usually important to avoiding any delays to keep on schedule (like sending overflow busses that don’t appear on a public schedule during the morning rush). When you’re picking up passengers on the off-peak direction the 5 or 15 minutes (or more with wheelchairs) of delay from passengers and traffic mess up timing and slow service to the bulk of the passengers. And there’s usually another bus servicing that stretch nearby anyways.
    We do use crew change cars and vans frequently for our couple crew change points during the day and they work quite well. We also couldn’t function without them because we simply don’t have enough busses to begin and end every shift at the garage (all drivers are students so shifts are usually in the 2.5-5 hours range to fit around class schedules). Even though the garage is less than a mile from the major crew change point, that is in the middle of the routes so passengers would have to change busses every time. Also there’s the 5 minutes to and from the garage during heavy traffic and the need for the driver to inspect the bus before heading out and so forth which all would add a few minutes and slow things down pretty fast while also clogging up our small garage. (Also, many drivers can set up their schedule so when their class ends they’ll walk over to the crew change spot, get on the bus and drive a couple loops before another class. Sometimes hard to fill shifts will be partialed out in single loop or half-loop segments as short as 40 minutes to fit class schedules.)
    The few times that we do have to deadhead are really useful to keeping on schedule and avoiding traffic however. If there’s already other buses servicing the off-peak direction in close proximity to the dead bus then it can be more efficient to sent the dead bus express around the traffic and get back into the peak direction faster. And I can imagine that for many other agencies that have similarly-tight bus supplies those few minutes can make the difference between having enough busses or not for full service if there’s too many in maintenance.
    If I had to guess I’d say that our routes have less than 5, maybe 3% deadheading, not counting training or special hired “field trips.” I’m curious to see how our ratios and practices compare to the rest of our sister networks that we connect with which are all more conventional agencies with full time drivers (collectively the PVTA

  16. Nicholas Barnard June 10, 2011 at 10:46 am #

    @John I’m not sure about Metra, but Kitsap Transit has this. Kitsap has a major transit customer of a military base, plus a set of connections to ferries. For more info:

  17. Nicholas Barnard June 10, 2011 at 10:53 am #

    I’m in solution mode this morning. Another way to tackle the deadheading problem is to copy the airlines. They position airplanes at out stations in the evening to fly out in the morning. (its called RON – Remain OverNight) The plane gets cycled through to a major hub for heavy maintenance, but light maintenance is done at outstations.
    Given fuel prices this might become more feasible for transit agencies. That being said you bring on issues such as security of buses during the evening. Although, when KCM drivers abandoned buses all over the system during our snowpocalypse in 2008 I didn’t hear of any major reports of vandalism.

  18. EngineerScotty June 10, 2011 at 11:20 am #

    TriMet (in Portland) has three bus depots (Merlo Garage in Beaverton, City Center Depot in inner SE Portland, also the location of HQ, and a garage at 92nd and Powell), and two rail depots for MAX (one at Elmonica in Beaverton, essentially next door to Merlo Garage; and one at Ruby Junction in Gresham). Portland Streetcar and WES have their own depots as well.
    I’ve often argued that this isn’t enough. My home is not far from Merlo Garage, and it seems that a significant fraction of the busses I see are deadheading to and/or from the depot. TriMet has probably enough “full-service” depots, with the capability for major repairs and the like; but it probably could benefit from a few more “minor” depots, particularly in the southern parts of the metro area, where busses can be refueled, washed, and have minor maintenance on things like tires and oil, and drivers can clock in and out.
    TriMet takes a different approach for MAX. Both rail facilities are presently located on the Blue Line; Red, Yellow, and Green trains have to travel along the Blue Line until they reach a point where the can switch to the normal service route. But rather than deadheading, TriMet runs these trains (both to and from the depot) as revenue trains, so in the mornings or evenings you can occasionally catch a one-seat ride from Gresham to North Portland.

  19. Steve Lax June 10, 2011 at 5:11 pm #

    In my experience, we always tried to convert non-revenue trips into revenue trips (both pull-ins/pull-outs from the depot and deadheading from one end of the line to the other) where reasonably practical; but, in many cases we needed to exercise caution in doing so.
    Reason 1: We would develop a very small/small ridership on some non-revenue trips and then no longer need the primary reason for the trip and be faced with discontinuing service for this small number of riders. (Example: Primary reason for an afternoon pull-out trip was to serve an industry with a second shift beginning at 3:00 PM. Industry discontinued second shift. Primary reason for pull-out no longer existed.)
    Reason 2: Converting non-revenue miles to revenue miles for the (potential) benefit of a very small number of passengers makes a bus route’s operating statistics worse (passengers per revenue trip, passengers per revenue mile, revenue per revenue mile actually decrease if a large number of trips are converted and ridership does not develop.) Then the route appears less successful than its peer routes. This puts the route in danger when measuring it against the agency’s performance standards.
    Reason 3: We would lose some flexibility. Often, we could convert a deadheading trip into a revenue trip on a nearby route experiencing some overcrowding (with some additional scheduling shuffling) if we had a deadheading vehicle available at less cost than adding another vehicle to the route experiencing overcrowding. If the deadheading fleet had been converted to live (revenue) trip operation in these very marginal cases, we would not be able to respond as effectively to true demand situations.
    Reason 4: On pull-ins/pull-outs to/from the depot, the small additional time required to convert a non-revenue trip to a revenue trip might cause us to incur operator overtime costs in excess of potential revenue.

  20. Daniel June 10, 2011 at 8:59 pm #

    Oh, a couple of other things to throw into the mix:
    In Melbourne we have some services which are “non-PSR” (Passenger Service Requirement) trips. That is, they would be dead-running, but have been made available for passengers to travel on… but the operator does not get penalised if they don’t run, and they don’t appear in the timetable.
    On the suburban train network, these are most often runs out of the CBD in the morning (eg counter-peak) to suburban depots. They became known as ghost trains following this article in The Age:
    On the tram network, similar services are said to account for about 10% of service kilometres. This article specifically highlights the mystery route numbers they have been given:
    See also
    Recently there has been a proposal from the operator to alter the tram route numbers to something more meaningful.

  21. EngineerScotty June 10, 2011 at 10:09 pm #

    Reason #2 (on Steve’s list) doesn’t strike me as a particularly good one. If the standards for recordkeeping are imposed somewhere else, and any revenue service counts but non-revenue service doesn’t, you may be stuck–but it seems to me that deadheading ought to be figured into a line’s efficiency metrics.
    After all, the driver’s getting paid and the bus is accumulating miles whether or not anyone is is riding.

  22. Moaz Ahmad June 10, 2011 at 10:28 pm #

    Dead running in Kuala Lumpur is pretty simple … the buses come into the city in the morning, then sit at the roadside until they fill up, then move again. At the suburban terminals, they do the same thing.
    Regards, Moaz

  23. buscommuter June 10, 2011 at 11:35 pm #

    Some dead heading is obviously required.
    But Metro deadheading in Seattle is particularly annoying.
    Seattle and cities on the east side of Lake Washington are connected by a bridge (SR520) which is a major choke point, and has massive rush hour congestion in both direction in the morning and the evening.
    Metro has one way commuter routes in both directions, so you see empty dead heading buses going in opposite directions across the bridge.
    Most annoying are those that head back to base, which is about 1 mile away from the transit center in Bellevue, so it frequently happens that 15-20 people, waiting at a stop on the Seattle side, are being passed by an empty bus going to within a mile of where they want to go. Sounds like and opportunity to either add more service for little more expense, or cut expenses with little impact to service.

  24. Max Wyss June 11, 2011 at 10:49 am #

    @Paul McGregor: When the operation center is close to an end point, it is pretty common practice to run the train through to the end point, and from there to the operation center. This provides maximum service of the line, and reduces the need to provide “alternative transportation”. When reducing the frequency, for example after the evening rush hour, there will be the regular train leaving the end point, and the one turning in follows immediately; again best possible service for the customer.

  25. Max Wyss June 11, 2011 at 11:03 am #

    @Steve Lax: Taking passengers on deadheading runs should nevertheless not count as revenue run. Because it is an operational necessity. So, instead of counting the deadheading as additional “revenue run” don’t count it, but count the passengers as “revenue” for the line the bus operates. This actually improves the revenue of that particular line by a marginal amount.
    Assuming that the deadheading with revenue passengers is along a regular line, the deadheading vehicle would be a non-scheduled “extra” service, so, if you have to help out elsewhere, there would be not much harm done.

  26. Max Wyss June 11, 2011 at 11:19 am #

    @Daniel: Assignng a special number for the “ghost service” strikes me as, hmmm… unusual. Assuming that the vehicle will go to an end point of the target line, why not just sign it as part of that line. For some stretch, it will be away from the standard route, but it will eventually hit it, and then appear as a normal vehicle of that route.
    If for whatever reason, it has to be part of another line internally, so be it, but don’t show that to the customer.
    And if it really does not work out, I assume that Melbourne has “Extras”, jumping in wherever needed, so, show it as “Extra” to the end point of its route…

  27. EngineerScotty June 11, 2011 at 2:34 pm #

    Is there a preferred term for nonstandard revenue runs of a transit vehicle in lieu of deadheading? I nominate “liveheading”, if not. 🙂
    (And a minor linguistic note: The term “deadheading” long predates the Grateful Dead or their famously devoted fanbase…)

  28. Zoltán June 11, 2011 at 3:39 pm #

    @buscommuter This seems like an argument for converting one-way expresses into two-way routes with limited stop sections, that presumably would provide frequent service on common sections.

  29. Steve Lax June 11, 2011 at 4:28 pm #

    @Max Wyss – A trip may operate off the public timetable; but, if for no other reason than operator accountability, it has to be either a revenue trip or a non-revenue trip.
    If the trip is filling a gap in the headway (however one wishes to define gap), it should appear in the public timetable in the interest of showing the public the frequency of service at a particular time. However, if the availability of the deadhead trip is not going to be constant (for example, the vehicle operates along this route in deadhead mode only on days school is not in session and operates along a different route on days school is in session), then the public is not well served by the trip appearing in revenue mode in a hit or miss fashion.

  30. Steve Lax June 11, 2011 at 4:40 pm #

    @Engineer Scotty – Concerning my Reason #2 – I noted some measures where converting non-revenue trips to revenue trips could make it appear that a route was performing more poorly than its peers. There are, of course, many other measures which do take non-revenue mileage and associated costs into account.
    Sadly, some times only some of the statistics relating to a route’s performance are easily comprehended by those not in the transit industry. And these statistics can be misused. One example: Passengers per revenue trip. If a route is primarily a peak direction route, operating buses in revenue service in the non-peak direction can significantly impact the passengers per trip count. And this is a statistic that politicians and newspaper reporters like to use, even if they do not understand fully the meaning of the statistic they are using.

  31. Max Wyss June 12, 2011 at 12:54 am #

    @Steve Lax: There is one thing I kind of assumed, which may not be the case in many cities: The “regular” schedules have to be dense enough that people are not looking at the timetables, but simply go to their stop, and take the next vehicle coming (and that behavior begins at intervals shorter than maybe 12 to 10 minutes). At these intervals, a public timetable can only be conclusive if the vehicles operate within a minute or two according to the timetable; if that is not assured, people will have to go to the stop and see what comes. That said, these extra services do not need to be published, because they may appear as “early” or “late” regular services.
    About accounting: particularly with zone fares, it is not possible to get exact numbers of use. In such cases, vehicles equipped with a passenger counting system are used, creating sample data which can be used reliably. Now, with routes being seriously imbalanced, there would be the other trick, to assign another route number (such as “extra services” to those extra services.
    Also, as soon as you have a network, accounting for a single route becomes questionable, because it is no longer for sure that traffic for a certain “corridor” can be properly assigned.

  32. Steve Lax June 12, 2011 at 4:17 pm #

    @Max Wyss – You are correct that if service is frequent, having an off-timetable (or off frequency) trip appear along the route is of no consequence. However, even in systems with large numbers of frequent service routes (including NYC), there are routes that do not operate with frequent service, even in peak periods. Some of these routes may be near a depot where there can be a valid debate as to whether pull-ins and pull-outs, though random in terms of a headway, should operate in revenue service.
    Concerning accounting: There are additional ways to count passenger activity; but that is somewhat immaterial as long as one has a reliable base count. Assigning another route number can be done to extra services; but one must be careful. In the U.S.A., the FTA has this thing about running trips not known to the general public. It is a no-no. While extra sections that are off-timetable are tolerated; anything that looks like a “special trip”, not designed to accommodate the general public can lead to a reprimand or worse. Thus, any route number operation needs an accompanying timetable. My system received grief for coming up with a route number with the generic name “School Services” for trips that operated in revenue service only on school days even though the trips serviced the general public.
    I agree that accounting for a single route in a network is a questionable practice. But every time the budget was tight at the system at which I worked, someone had to crank out route by route performance numbers.

  33. Brian McCann June 12, 2011 at 5:58 pm #

    When publicly questioned (criticized) about buses not in service, Glenna Watson at COTA Central Ohio Transit Authority responded. “You know, they’re all empty when they go back to the garage.”

  34. Chris, Public Transport June 12, 2011 at 8:55 pm #

    Some thoughts to consider:
    1) When pulling into or out from a garage and deadheading, picking up passengers could potentially make the bus leave late on its first trip, affecting all the passengers on the route, or arrive at the garage late on its last trip, which could cost the transit system money in overtime costs.
    2) Passengers generally do not like to be ordered off the bus in the middle of the route, which is likely to happen if a deadheading trip carries passengers to the garage. This kind of thing certainly has caused a lot of transit agencies goodwill, including Toronto’s dreaded random short-turn streetcar trips.
    3) As mentioned, the running time savings from deadheading usually enables additional trips to be made with the same bus. If there were no savings, then the bus would not (hopefully) be deadheading.
    4) Overall, a lot of deadheading is caused by poor facilities planning and poor scheduling practices. A suffering transit agency in the San Francisco Bay area has chosen to locate its two major garages far from the start of any bus route. In addition, poor scheduling causes buses to deadhead continuously to operate different routes. The practice of some operators to extensively use pull reliefs instead of street reliefs also creates unnecessary deadheading. Operators with multiple garages also have the opportunity to adjust blocking to ensure blocks both start and near the same garage.
    Comment about leaving the buses overnight at the terminus point of the route, or at a different garage: even if we don’t care about inconveniencing the driver, including potentially leaving them stranded, necessary night activities such as 1) fare removal; 2) bus servicing; and 3) bus maintenance would be problematic. Fleet management would be a nightmare, and miss-outs would likely increase as drivers (and supervisors) would forget where they were supposed to report on any given day.

  35. Dave June 13, 2011 at 5:01 am #

    @Chris; I have to disagree with your 2nd point. Sure, people don’t like being kicked off a vehicle mid-route without reason. However, if it’s marked as a non-PSR route – eg the typical route is a 100, and the vehicle is ‘100d’ (for deadhead, depot, whatever makes sense in your city) then pax are informed and will make decisions accordingly. Not all passengers using the depot run will go all the way to depot either; if they only want to stay on for a few stops it may be irrelevant if it’s a depot run or a regular run.
    It’s when you get unexpected / ‘random’ short runs that people get annoyed – and fair enough. But I don’t think this should happen if there’s sufficient information on deadhead/depot runs for passengers to know that’s what they’re on.

  36. Max Wyss June 13, 2011 at 11:12 am #

    Adding to Dave, by running the deadheading vehicle kind of immediately after a regular one, there won’t be many passengers anyway (well, that does apply more for rail-based systems, as overtaking is not really possible.
    In short, it all depends on the individual circumstances. But minimizing deadheading is one thing where schedule planners really can show how good they are.

  37. EngineerScotty June 13, 2011 at 12:41 pm #

    One other thing that can help, particularly for busses: BRT street improvements (signal priority, exclusive ROW) that deadheading busses can use to get back to the garage faster. One advantage of the “green bridge” being built for the Milwaukie MAX project (construction on the bridge is scheduled to start in about two 1/2 weeks), is that it will provide busses with a faster crossing of the Willamette river than does the existing bridges. Close to the west end of the bridge is the southern terminus of the transit mall–and close to the east end is the City Center depot. Busses finishing up routes downtown will have, hopefully, a slightly shorter route back to base–and even a few minutes saved will add up, in terms of more reliable schedules and the like.

  38. Jarrett at June 13, 2011 at 12:42 pm #

    Scotty.  Is it still the case that deadheads from downtown to Powell division run as Line 9 shortline trips?

  39. david vartanoff June 13, 2011 at 2:21 pm #

    SF Muni rulebook (aspirational???)
    “4.17.1 Operators on Cable Cars, coaches, and
    electric cars, in revenue
    service, or on pull-out and pull-in trips,
    are to stop for intending
    passengers, when there is room enough to
    board, except as follows:
    A. When a number of coaches or electric cars
    having the same destination
    are bunched due to an unusual delay. In such
    instances, the operator on
    the first coach or electric car may pass up
    every other stopping point
    where persons are waiting to board;
    B. When coach or electric car is loaded to
    C. When coach or electric car is disabled;
    D. When transferring motor coaches between
    E. When motor coaches are operating during
    out of service trips;
    F. When operating Limited Stop and Express
    trips in areas where stops are not
    G. When directed by authorized offical.”
    The reality is that drivers evict passengers for pull ins and very rarely pick up passengers either “in” or “out”.

  40. Steve Lax June 13, 2011 at 5:13 pm #

    @Chris, Public Transport, part of 4) – The use of street reliefs, especially if mid-trip, has some disadvantages, also. The system at which I used to work actually has reduced street reliefs and increased pull reliefs for a number of reasons:
    1. Complaints from on-board passengers about delays to their trips.
    2. On frequent service routes, the probability that bunching would occur due to the time the relief takes [operator being relieved has to sign off bus systems (radio, farebox, etc.); relief operator has to sign on, adjust run number on window display, adjust operator seat and mirrors]. Relief time for a mid-run street relief is four-six minutes.
    3. If the operator being relieved is late, the relief operator is “in the hole” and, depending on recovery times at the end of trips, may not be able to make up the time for the entire run.
    4. Where the relief operator was expected to use public transit to reach the relief point, occasionally, the relief operator was late because of a transit delay. Where a company car was provided; occasionally, the car would not be available being late returning from a previous relief.
    So, my system traded off and increased non-revenue miles/hours to alleviate these issues on many routes.

  41. EngineerScotty June 13, 2011 at 5:19 pm #

    I don’t know the answer to your specific question… what I do know is that about half of the 9 busses heading east from downtown only make it as far as the Powell Green Line station (which is right next to the Powell Division garage); the other half go out to Gresham. So while the line is billed as frequent service, it’s really only frequent (using TriMet’s definition) as far as SW 92nd or so; east of there it runs at half-hour headways, at best.
    But how many of those busses go into the garage, vs turning around and heading back downtown, I don’t know.

  42. Pete June 14, 2011 at 2:24 pm #

    On my local route – Service 231/232 Bath – Chippenham dead running is confined to the start and finish of the operating day. Otherwise buses run in service providing a 30 minute frequency throughout the majority of the day. The bus depot is in Bath, so when the frequency steps down in the late afternoon/early evening buses leave service on arrival at Bath bus station and make a short dead run to the depot. Here is a link to the timetable:
    Here is a link to the route map:
    Route length is approx 14 miles, with an end to end running time of 1hr 20 mins.

  43. Max Wyss June 14, 2011 at 3:30 pm #

    @Steve Lax: With all due respect, but I am shocked about the relief practice in your system. There must be serieous organizational issues at that system, as with the systems I am a bit familiar, the operator change takes 2 minutes at most (and that includes a chat between the operators).
    Even on the bus system where the bus driver has the farebox unit (I don’t know the correct term of that in English, sorry), he starts closing the farebox during the drive suring the last segment, takes the receipt, pulls out the memory unit, and takes out his money drawer. So, when he arrives, he is checked out, and the new operator can install himself immediately (insert money drawer, insert memory unit, maybe print out a login receipt. Adjusting mirrors may take a few moments, as well as adjusting the seat (does not take much time with the state of the art Grammer seats…)
    The case where the relief driver is not there… just not allowed to happen. Period. A call to the control room will be the first thing to do, and in many cases, the driver may just have to go on, and the relief will be organized, or has to catch up using a taxi or whatever.
    The system I am most familiar with is mostly rail-based, and the routes are in relative proximity of the operation bases. And in these cases, the operator change is very fast; adjusting seats is non-critical, and is usually done on the move; there is no farebox handling, and mirrors are big enough that they don’t need to be adjusted (except maybe the one for the inside, but that too is non-critical, and can be done at a further stop.
    It boils down to operation practices, which may or may not be efficient.

  44. D Matyas June 14, 2011 at 5:07 pm #

    To add another perspective on the in-route crew changes: For us it is generally closer to a couple minutes with experienced drivers who rarely bother to adjust the mirrors too much because they rely on them less. For less experienced drivers such as myself (which is a high percentage of our drivers given that we’re all university students) I generally take the extra minute or two to make sure the mirrors are properly adjusted. On the older busses it’s usually very quick to adjust the mirrors to be good enough (and then fine-tune at our convenience later), but the newer busses have very finicky mirrors that take a solid minute or two to adjust properly. So in practice it usually takes us between 2 and 5 minutes depending on the bus and experience levels of the drivers.

  45. Steve Lax June 15, 2011 at 6:05 pm #

    @Max Wyss – I did not believe the 4-6 minute time for a relief mid-trip (and I emphasize mid-trip where the operator being releived cannot sign off the radio and farebox safely until reaching the relief location in contrast to an end-of-route relief where it might be possible to do so at a red traffic signal a short distance before the route end) until I rode “undercover” and timed it myself. The time it took to change from a petite female operator to a tall and heavy set male operator actually exceeded six minutes with both operators hustling.
    Obviously, the specifc seats, mirrors, bus type, and electronic equipment vary by bus type. My system strongly discourages operators signing off/on the electronic equipment while the bus is in motion. And, as I have stated in previous threads, each system is somewhat unique. While we can learn from others’ experiences, we also have to consider the hand we are dealt locally.

  46. Eric June 15, 2011 at 9:41 pm #

    "All those services that are needed in only one direction usually have to get back in the other direction so that the driver’s shift can end where it began."
    Why? Because the only way for the bus driver get home or go anywhere during his break time is drive the bus back to the base in order to pick up his car?
    Sounds to me like an admission that the transit system is a failure? With a functional system, after completing a shift, especially in downtown, drivers should be able to board a regular bus or train as regular passengers and travel to wherever they need to go.
    If the bus is slow enough to make this cost prohibitive, it is probably too slow for the riding public to consider it as a serious mode of transportation, and the mode share of such a system is probably very low.

  47. Nicholas Barnard June 15, 2011 at 10:02 pm #

    @Steve – This is a pretty good example of systems not really being designed to meet the users needs. Whoever came up with the design requirements for the system didn’t do a good job describing the requirements as far as time. I’d be curious if getting rid of all those deadheads would save enough to facilitate the redesign/reworking of the system..
    In Seattle, they can usually switch over in about 3 minutes. It doesn’t happen that often, and its at stops where they’d be likely to be backed up behind other buses anyway.

  48. D Matyas June 18, 2011 at 2:59 pm #

    @Eric Don’t forget that in even cities like Boston the system shuts down at night, let alone much smaller systems, so the only way for the first or last shift drivers to get to work is to drive.
    @Nicholas Most systems I’m familiar with were never designed in whole as they are today but have evolved and usually started on very tight budgets where many compromises had to be made at the expense of small efficiencies simply for the system to be able to afford to exist at all. And that problem only compounds over time and expansions where agencies never have the sort of money they’d need for a proper design overhaul beyond very incremental improvements taking decades.
    Also, @Max remember that things like which busses (or trains) an agency acquires is often largely out of their control due to the bureaucracies they’re stuck under. Things like “Buy American” programs and the many strings attached to the federal/state money often leave very few options of manufacturers/designs that meet their needs and the affordable customizations can only do so much.