London: the Circle Reaches an End

550px-Circle_Line.svgOne of the world’s most famous continuous loops is finally reaching an endpoint.  The London Underground’s Circle Line, which has long attracted tourists with its simplicity but bedeviled its operators with lack of rest, is to be broken apart into a “tadpole” shape, with trains leaving the circle at Paddington, via a spur of the current Hammersmith & City Line, to end at Hammersmith.

As discussed here, loops touch deep things in the human psyche.  Nobody wants to reach the end of the line, really, except for the transit operator.  In fact, the “end of the line” shows up fairly often as a metaphor for death.  So at many scales, from the London Underground to tiny downtown circulators, there’s a constant temptation to offer services that never end, services running in continuous circles.  Few people actually want to travel in circles, but the embracing quality of circles, and perhaps the subtext of forgiveness in the notion that life will eventually bring you back to anything that you missed, keeps causing them to appear, not just in plans but in operations.  Montréal is the latest to propose a giant continuous loop in its subway system.

Why did the Circle Line come apart?  Because it was hard to operate reliably.  Reliability depends on regular breaks that give a late service the opportunity to catch up to its schedule.  Human drivers obviously need breaks too.  Actually, the transit industry usually distinguishes between two kinds of break:

  • Layover is the time a human driver needs at the end of a line, to do personal business, meditate, snack, whatever.
  • Recovery is spare time in the schedule whose purpose is to enable a late vehicle to get back on time.

Many labor contracts and operations policies have complex formulas governing these two things, but in broad terms, a bus or tram in mixed traffic needs at least 10% of its time to be layover and recovery, and often closer to 15%, while exclusive right-of-way services may need a bit less.

Since people rarely want to be waiting on the vehicle while these breaks occur, efficient transit lines have endpoints where the vehicle will be empty.  Most transit lines that are presented as continuous circles have some provision for this, but it’s often awkward.  On the Circle Line, drivers took their breaks by leaving the vehicle and taking over one a few minutes later, a method called “operator fallback,” which provided layover but didn’t solve the problem of recovery.  Now that they have both, they should see better operations, and perhaps even happier drivers.

8 Responses to London: the Circle Reaches an End

  1. Jeffrey Bridgman November 22, 2009 at 9:19 pm #

    I wonder how the Yamanote line in Tokyo works… perhaps recovery isn’t an issue since trains run so frequently, people rarely look at the schedule (except to see when service starts in the morning, or ends at night), so who knows, maybe they don’t follow the schedule at all during the day. I think they run like 300 trains a day, who’s gonna know if it’s actually 299! 😉

  2. Alon Levy November 23, 2009 at 1:02 pm #

    Jeffrey, my understanding of Japanese scheduling is that all trains follow a schedule, even subways, and if a train is more than 5 minutes late, the operator produces delay certificates for the riders to be able to explain why they were late for work.

  3. anonymouse November 23, 2009 at 7:49 pm #

    The Yamanote Line (and Moscow’s Circle Line) have the benefit of dedicated tracks, without the many merge conflicts and flat junctions that London’s Circle has. In Moscow, it’s also standard practice to not run trains at full speed, so that when a train falls behind schedule, there’s some slack that can be picked up by going faster.
    Anyhow, it’s not obvious that the new arrangement in London is necessarily an improvement: you end up with Edgware Road as a chokepoint, with Circle and H&C trains running through the outer platforms, and District and Circle trains terminating on the inner platforms, plus there are more conflicts at Praed Street junction. It’ll be interesting to see if this is in fact any better than the original arrangement.

  4. jfruh November 25, 2009 at 11:35 am #

    Another problem with circle lines is that trains have to be periodically reversed, or their wheels will wear asymmetrically!

  5. Nathanael November 29, 2009 at 1:04 am #

    Recovery wasn’t necessary in London’s Circle Line, because from a passenger POV it ran on headway operation — “renumbering” was used to get the trains “back on time” to the extent it was needed.
    What mattered — the actual operational issue — was being able to fit the trains in around the District, Metropolitan, and H&C trains.
    I suspect the “teacup” will turn out to have severe operational problems. The problem comes, as noted, at Praed Street Junction, and when reversing at Edgware. If there were a four-track ROW from Praed Street to Edgware, or if the route from the west side of the circle stopped at Paddington, it would relieve operational problems. As it is it makes them worse.
    It doesn’t really provide space to get out of the way of the other lines, and the line crosses itself.
    I wonder if they’re going to end up making people change trains at Paddington when going west-to-north. This would be annoying as it’s a long walk, but it *would* eliminate the operational problems by largely eliminating the use of Praed Street Junction.

  6. Paul February 28, 2010 at 2:44 pm #

    I wonder then if an automated train system would be more suitable. It wouldn’t need to take a break or a layover and so it can continuously keep running.

  7. Andrew July 3, 2010 at 3:04 am #

    The Yamanote Line in Tokyo uses an automatic train system that keeps the service running relatively consistently so that headways between services don’t change significantly from one service to the next. The other thing they also do is they have two dedicated stations which are used for starting and terminating services (Osaki and Ikebukuro) as they have extra platforms.

  8. Joel N. Weber II November 27, 2011 at 5:12 pm #

    For a transit system that runs in mixed traffic, can traffic signal priority that gives transit priority only when it’s behind schedule help with the recovery time problem with loops? (Or, alternatively, that denies priority to vehicles that have gotten too close to the one ahead?) Also, can a system that instructs a vehicle to hold at a stop for an extra 30 seconds to help with keeping headways help out with this to some extent?
    (In Boston in the United States, the SL1 bus route has a loop at the airport, and then runs in a line between a few other stations and South Station. There’s a proposal to extend the bus tunnel to Chinatown / Boylston and then a portal, and I also wonder if a new portal right next to the I-90 westbound exit it uses into the bus tunnel could speed things up; in such a case, a one way loop of airport stops then I-90, the new portal, Courthouse, then South Station, then Boylston / Chinatown then out the Silver Line Phase III portal, to Hearld Street, to a new bridge into an existing tunnel under Fort Point Channel back to the airport seems like it might be more efficient than having to double back through Boylston / Chinatown, South Station, Courthouse, and probably World Trade Center by enough to outweigh the inconvience of waiting for a minute or two occasionally for schedule recovery. SL1 runs on very consistent 10 minute headways all day long, so it should probably run to maintain headway rather than on a constant schedule.)