the dangers of travel time comparisons

Revised in response to early comments. 

Are you sure you know which of your transportation options is fastest?  It depends on how you think about travel time.

A recent Boston Magazine article about the private bus service Bridj featured  typical "race" between two transit modes: the MBTA subway and Bridj, which provices luxury buses on fixed routes and schedules running only at times of peak commute demand.   The newspaper sent someone by each path at the same time.   The outcome of the race is supposed to be decisive:



Why is this not a fair race?  Well, it depends on when you start.  From the article:

The MBTA passenger arrived last, [sic] even though she had a head start and boarded the train six minutes prior to Bridj’s departure.

Why a six minute headstart?  Why not 10 or 20?  What headstart would be appropriate?  The headstart is your cue that there's something wrong with this methodology.  

What's really happening here is that a service that is available all the time — the subway — is being compared to one that's only available at a few special times — Bridj's specialized commuter buses.  Any "travel time race," with any headstart, is going to miss the real point of this comparison.

The notion of travel time seems so self-explanatory that most people miss how deeply misleading it is in discussing transit.   The imagined user is someone who happens to be going at the ideal moment for the preferred mode to succeed.  We talk about travel time this way because it's how motorists, cyclists, and pedestrians experience it: as something that begins at the moment you want to go.  

But that's not how transit travel time works, so the comparision implied by the term "transit travel time" is often a false one.  

When I teach transit planning and rhetoric, I encourage people to think of a weighted sense of travel time that includes average wait time, or more generally the difference between when you wanted to go and when you went.  A bus that's 10 minutes faster is of no use if gets you somewhere 30 minutes before you needed to be there [an 8:00 AM class or meeting, for example] because that's the only time it ran.  

Purveyors of low-frequency transit services, such as classical North American commuter rail, do this as well, bragging about how fast you can get from A to B without mentioning that this travel time is available only once a day.  

Unless you are sure that you will absolutely always travel at the same time each day, transit travel time figures have to be viewed with skepticism.  Whenever you hear about travel time, ask about frequency! 

15 Responses to the dangers of travel time comparisons

  1. Ben June 30, 2014 at 5:50 pm #

    Hmm, I just read the article and I’m not coming to the same conclusion.
    “The MBTA passenger arrived last, even though she had a head start and boarded the train six minutes prior to Bridj’s departure.”
    That quote makes it sound like the exact opposite of what you’re claiming, unless I’m reading it wrong. Care to clarify?

  2. Mo June 30, 2014 at 5:52 pm #

    And often not factored in to car travel are such time-bandits as retrieving the car (from where it is parked), topping up fuel, finding somewhere to park it, and getting from the parking to the actual destination. I have been known to decline offers of a lift because I know it will be quicker to go for a bus & walk home from the bus-stop.
    I believe all public transport system decision-makers should be required to travel by the public transport they manage, if not daily, then at least one working day each month – including to attend meetings during the day (and not re-scheduling meetings to avoid their ‘public transport’ day). That would be a reality check for them.
    When I worked in Public Transport I had fellow-managers asking me very basic questions about travelling on the system – they had no idea of the layout of vehicles, facilities, impacts on travellers, actual needs of the public-transport travelling population, etc, etc.
    I usually travelled by train and worked on the train – something that would have been dangerous if I was driving. Of course it was also safer than road travel, and more consistent in arrival times.
    I choose to rely on public transport and bicycle, occasionally borrowing a car for special trips on 3 or 4 occasions each year. I find it generally easy to schedule my tasks around the transport timetable. If I am going to be an hour early for an appointment, I schedule a coffee-break, take something to be working on, take a walk, smell the roses and think. Admitted, there are some times when this feels like an imposition and I would prefer to just get there and get it over with. Still, you can’t tell me busy people don’t have something they can be getting on with.

  3. Brent June 30, 2014 at 8:06 pm #

    @ Ben: although I read the article the same way as you, I think it is still a fair point, at least at the infrequency of the current Bridj service. They only run three routes; two offer three morning trips roughly every hour, and one offers five morning trips at half-hourly frequencies.
    However, the biggest point is that in order to achieve these savings, you need to have an origin AND destination within a short walking distance of a Bridj stop. If you need to walk more than 10 minutes on either end, you lose any travel time advantage, regardless of frequency / schedule.
    Add those two factors together, and MBTA comes out ahead in almost all cases. (And this is basically what Jarrett preaches — along the same lines of coming out ahead with a grid network that requires a transfer as opposed to systems designed for one-seat rides that result in convoluted networks, inconvenient schedules and inefficient operations.)

  4. Rob June 30, 2014 at 8:38 pm #

    When I read about Bridj, I had a different reaction. I remember when I was doing rail planning almost 30 years ago (seems like yesterday) and I wondered to myself whether someday we might do a better job serving commute trips by massively customizing service based on survey data rather than developing specialized peak routes and overlays. It was an idle question then (since massive survey data would have required unaffordable expense and sophisticated analysis), but in the context of big data – rich datasets on commute patterns and new methods to synthesize them – it’s no longer unfathomable that the peak overlay could be replaced with service designed for more precise markets.
    For a subset of people – those who really do want to travel to and from the same places at the same time of day each day – customized services could be faster and more convenient. In theory, buses scheduled this way could be more consistently full. Again in theory, if such service existed, you would not compare the travel times in traditional transit terms; you’d have to look at the whole package to compare the value of customized routes against the flexibility of travel time offered by frequent transit.

  5. Ben Smith June 30, 2014 at 9:17 pm #

    There is also the fact that the bus is likely running in mixed traffic, exposing itself to unreliability. While subways are hardly immune to delays (suicide jumper, rail congestion, limited abilities to pass obstacles, etc), the fact that it is in a controlled environment helps to improve its reliability.

  6. Ira July 1, 2014 at 5:16 am #

    Bridj also picks on the most dysfunctional of the MBTA’s services. The Green Line is old (some of it dating to 1897), low capacity (most trains are two-car LRVs with a crush capacity of 350; a third of the Red Line) and hobbled by very outdated infrastructure. It’s also the only MBTA service without real-time departures (coming next year, apparently) and transit signal priority (coming, oh, god, it should have been implemented years ago). And Bridj also cherry picks the morning rush hour, when thanks to high transit use, Boston doesn’t have much traffic (the streets downtown at 8 a.m. are empty, except for pedestrians crowding the sidewalk; by 5:00 it’s often gridlocked on the surface.
    And yes, it’s a specific trip, and it’s not particularly scalable (a large number of 12-25 passenger minibuses clogging popular destinations would not sit well with local authorities, and enough pedestrians and cyclists are injured by errant taxicabs that putting more vehicles on the road is not really a positive development). The claim that many of these passengers would otherwise drive is spurious, and most will probably gravitate back towards the T when prices go up to $5 or $6 per ride. It’s mainly catering to the 1%—and using blacked-out-window limo-style vehicles to boot—and even then might not fly because, well, has anyone ever made money providing urban scheduled transit?

  7. Patrick July 1, 2014 at 7:02 am #

    I was passing through Coolidge Corner this morning and noticed the Bridj bus loading/unloading passengers. Looked to be a fairly large Minibus (20 – 25 passenger). It looked like a fairly large crowd surrounding the bus, around 30 to 40. It seems like they’re doing ok ridership wise but then again they’re offering free ridership during their “BETA” period.
    This reminds me of a lot of the Rural Transit service in my college town where Minibuses offer deviated fixed route style operations.

  8. Jim D. July 1, 2014 at 9:47 am #

    Is anyone shocked that a bus which takes a more direct route between two self-selected points along roadways with available capacity is faster than a rail system which requires riders to travel all the way to downtown on one line and transfer to a second line? If Bridj offered service on routes like Quincy-South Station or Malden-North Station that required significant travel on frequently congested highways, the MBTA service would suddenly look a lot more favorable.
    The more interesting comparison would be if Bridj does start up a Brookline to Copley Square route as mentioned, which will allow Green Line trains to compete directly with the buses. Of course, a valid comparison would feature multiple riders leaving their homes at various times during the morning peak and factoring in waiting time for whichever mode they are taking, but these sorts of media ‘races’ aren’t known for their statistical vigor.

  9. Jarrett July 1, 2014 at 9:57 am #

    NOTE: The original version of this post incorrectly assumed that the Bridj and MBTA trips started at the same time. Comments up to this point are in response to that previous version of the post. The post was revised 1 July ’14, around 9:55 AM Seattle time to correct this error, but the point of the post remains. Thanks for the great comments, and keep them coming.

  10. davistrain July 2, 2014 at 1:12 am #

    One wonders how the drivers of these buses are scheduled and paid. What do the buses do after their rush-hour runs are completed?
    I second the motion of Mo, who recommended that “transit system decision makers be required to travel by public transport they manage”. When I worked in telecommunications, the “prime directive” was “The job’s not done until you’ve tested the system the way the user will use it.” When the managers have to use a ticketing system that would give an MIT graduate fits, maybe they’ll come up with something user-friendly.

  11. Jim D. July 2, 2014 at 10:05 am #

    The buses are provided by Academy Bus, so the drivers are Academy employees. Presumably Bridj charters the buses for a set period of time.

  12. Al Dimond July 3, 2014 at 10:03 am #

    Rather than seeing the comparison as an indictment of Boston’s mass transit, I see it as a compliment. Most people can make their commutes faster driving than taking transit, and Bridj appears to be a point-to-point service that can achieve similar travel times to driving when it’s moving. But it’s only an attractive service in a city where people live and work close enough together than common pick-ups and drop-offs work, where walking and transit are useful enough that they don’t regularly need a vehicle at work, and where the rest of the transit system is good enough to get them home at the end of the day if they stay late or go out, or to provide a reasonable option if they are leaving between Bridj trips. And it’s only financially viable if it’s attractive enough to fill enough vehicles to cover overhead. This is only true in a city where transit works already.
    Since the 4th of July is coming up, it’s worth noting that, though some public transit systems do a great job on the 4th, most don’t, and often fireworks celebrations aren’t at permanent transit nodes. A system designed specifically to be dynamic might be able to handle irregular nighttime event exits better than a public agency that’s schedule-oriented. But, of course, this only works in a city where transit and walking are already the best ways to get around.

  13. Brent July 3, 2014 at 8:09 pm #

    If anything, one could argue that mass high-demand special events, such as fireworks displays, need basic, brute-force service. Up in Toronto we just had our Canada Day fireworks display in the eastern beaches. Beforehand, the TTC cranks up the service levels on the regular bus route running from the Danforth subway down to the lake (a little over 2 km) so that there are buses running as frequently as they can be loaded — easily every 2 minutes or better. Then afterward, there is a queue of buses picking up crowds two or three buses at a time, again as fast as they can be loaded, to take them back to the subway. There is no room for nuance in situations like this; in extreme cases drivers don’t even bother taking fares — I suspect they’re directed to just get things done as quickly as possible, and besides, the majority of riders probably have Metropasses anyway. Relying on dozens (or hundreds?) of custom shuttle routes at a major event such as that would be a logistical nightmare.

  14. el_slapper July 15, 2014 at 6:40 am #

    Total time is from door to door, from when you want. I think you nailed it.
    For example, my daughter has dance lessons wednesday at 2pm. Walking takes 35 minutes. Bus takes 6 minutes. Plus 5 minutes to go to the stop, plus 2 minutes from the stop to the dancing room.
    Add to that the fact that buses are at 1pm or 2pm, and you’ll get why she walks as soon as it’s not raining. It’s not for love of sport. It gains her time.
    There is something else : when in car, your activity is driving. When in public transit, it may be standing clogged against other people(not better)… or seating sleeping, reading, or playing the smartphone, which is much better. In Paris, when given the choice between B & D line from Paris Nord to Chatelet, I prefer the slower-but-less-crowded D line. As I can have useful time inside. Travel time is not everything.

  15. James Turner December 2, 2014 at 3:28 pm #

    An interesting read with some good points 🙂