04 October 2014

The Top Ten Problems Facing Caltrain

Caltrain is searching for a new General Manager.  This person will need a briefing on the key issues now facing Caltrain.  Here is a list of the top ten problems that are clouding the "blended" vision of Caltrain and high-speed rail:

PROBLEM #10: Bloated Staffing.  There are well over 100 Caltrain conductors, among an operating staff of close to 500.  The traditional job description of a conductor (ensuring safe movement of the train, acknowledging signal aspects called out by the engineer, announcing stops, operating doors and lifts, checking fares, giving out information, etc.) is slowly being made obsolete by technology.  The new signal system will ensure safe train movements.  Modern trains will automatically announce the correct station stops, and allow remote door operation from the driving cab.  Smart phones give everyone up-to-the-minute information.  We are quickly reaching a point where union-mandated crew size is no longer justifiable, and an employee-to-locomotive ratio of 20 is an unsustainable excess.  Solutions:
  • Do more with less.  Make modernization pay off by taking a hard look at staffing levels
PROBLEM #9: An Overcomplicated Timetable.  Caltrain's timetable is impossible to memorize and confusing to figure out unless you ride the same train every day.  There are large and irregular gaps between trains.  Irregular stopping patterns make it impossible to plan timely and reliable connections to buses and employee shuttles.  Solutions:
PROBLEM #8: Too Much Deference to "Tenant" Railroads.  In a "blended" system, Caltrain's interaction with high-speed rail will be far more intense (as measured in train-miles) than with other "tenant" railroads that use Caltrain tracks, such as Amtrak, ACE or Union Pacific.  And yet, Caltrain's efforts on interoperability are focused on a handful of trains that use only a fraction of the corridor, whether in the planning of the new CBOSS signaling system or the discussions about platform height.  There is even serious talk of allowing Amtrak to revive the Coast Daylight service into San Francisco.  While interoperability is a worthy goal, this is the wrong kind of interoperability.  What matters most is interoperability with high-speed rail.  Solutions:
PROBLEM #7: Skipping the Wrong Stops.  When the Baby Bullet was launched in 2004, fast service and shiny red trains stole the headlines.  The under-reported back story is that speed came at the price of cutting service to a number of previously healthy station stops.  Caltrain is primarily a commuter service, and it should stop where people and jobs are located.  Throughout jobs-rich Silicon Valley (Palo Alto southwards), almost every stop has a similar density of nearby jobs.  Stops like California Ave, San Antonio, Lawrence and Santa Clara are severely under-served.  Solutions:
PROBLEM #6: Russian Roulette Station Dwell Times.  Caltrain is inherently unable to stay on time, but not for lack of trying.  Without level boarding, the occasional wheelchair customer can randomly inject an unanticipated delay of three to five minutes.  The unpredictable nature of these delays forces Caltrain to build a very conservative timetable with generous padding to absorb whatever might happen on any particular day.  This conservatism, which translates directly into lower utilization of the tracks, reduces the amount of traffic the rail corridor can reliably carry.  It also prevents Caltrain from operating reliable timed transfers or overtakes.  Solutions:
PROBLEM #5: Slow Average Train Speeds.  Caltrain's electrification project will help in this regard, but it will not sufficiently increase average train speeds to blend successfully with high-speed rail.  While electrification improves end-to-end run times by up to 12 minutes for an all-stops local, the excessively long station dwell times (on average 45 seconds) blunt the benefits of modernization.  The greater the average speed difference between high-speed rail and Caltrain, the fewer trains the line can carry.  To preserve enough slots in the timetable for Caltrain to grow, run times must be reduced beyond the basic step of electrification.  Solutions:
PROBLEM #4: Limited Corridor Capacity.  Under any "blended" scenario, high-speed rail will consume a large share of the traffic capacity of the peninsula rail corridor, leaving a limited number of rush hour slots for commuter service and severely constraining Caltrain's future growth.  In today's rush hour, 5 trains per hour per direction x 650 seats = 3250 seats/hour/direction are often standing room only.  Ridership is sure to zoom again when service to Transbay is inaugurated.  Unfortunately, the "blended" system will limit Caltrain to no more than six trains per hour per direction unless the corridor is widened to four tracks.  Solutions:
  • Pack 'em in with comfort, using five-abreast seating in extra-wide trains 
  • Buy the new EMUs at least 8 cars long, not just 6 as currently planned
PROBLEM #3: Traffic Jams into SF Transbay.  Despite marketing as a "Grand Central of the West," just six platform tracks for both HSR and Caltrain will require a tightly choreographed ballet of arriving and departing trains that will degenerate into cascading delays at the slightest disturbance.  As it stands, Caltrain is shut out of 2/3rds of the platforms in a station that could generate nearly 50% of its future ridership.  The solution to this problem is most assuredly not to terminate trains at 4th and King (see Problem #2).  Solutions:
  • Adopt a common platform height so any arriving train can be assigned to any platform, giving much-needed flexibility to relieve congestion in the station's approach tracks and to recover from a service disturbance
  • Optimize the track layout of the station approach for routing flexibility and higher throughput
PROBLEM #2: Incomplete Service to SF Transbay.  Today, the Transbay location has more jobs within a half-mile radius than exist within a half-mile radius of all the other Caltrain stops from 4th and King to Gilroy combined, and the intense development around the site will push its potential even further into the stratosphere.  Meanwhile, every preliminary or conceptual timetable published by Caltrain planners in the last five years shows the majority of rush hour trains terminating at the existing 4th and King terminal.  When asked about this, Caltrain invokes baseball game service and completely misses the point.  Solutions:
PROBLEM #1: An Agency Culture That Doesn't Put Service First.  At Caltrain, infrastructure projects are dreamed up over a timeline of many years, planned by layers of consultants until a funding package is cobbled together, and then pursued doggedly and almost for their own sake.  Caltrain is still an agency that thinks of itself as running a railroad, rather than providing a transportation service to its customers.  The capital projects that result aren't the outcome of a focused planning process that puts service first--case in point, the San Bruno grade separation, a $155 million piece of infrastructure that has achieved exactly nothing for the average Caltrain rider.  Solutions:
  • Use a capital planning approach driven by service quality metrics
  • Make the timetable, and its future possibilities, the focal point of all capital planning
  • Consider waiting time and first-mile / last-mile connections in all planning decisions
  • Live and breathe the mantra: service drives planning drives infrastructure
The Baseline of Mediocrity

The top ten problems and solutions, by contrast, give us a detailed picture of the most mediocre outcome that Caltrain could achieve by 2030:
  • Infrastructure first, planning optional, service an afterthought
  • Sub-optimized service where ridership already is, not where jobs and people really are
  • Rush hour trains that don't serve San Francisco Transbay
  • Caltrain squeezed into 2 Transbay platform tracks due to incompatible HSR platforms
  • A slow and inefficient train traffic jam in the tunnel approach to Transbay
  • Narrow European-size EMUs that don't take full advantage of the available clearances
  • EMUs that are just as short as today's diesel consists, limiting passenger capacity
  • Russian-roulette dwell times, with wasteful timetable padding to absorb them
  • Unreliable timed transfers and delay-prone overtakes
  • No level boarding, resulting in only modest improvements in trip times
  • Fast acceleration, but not enough giddy-up to blend efficiently with HSR traffic
  • Overstaffed train crews that increase cost and drag down fare box ratio
Caltrain has always had a lot of potential.  With such juicy ridership demand as this corridor enjoys, it will be hard to call Caltrain's "modernization" a failure, but so far it doesn't promise anything more than this baseline of mediocrity.  To be sure, electrification is a huge step forward compared to most other commuter rail systems in the U.S.  But that's the wrong frame of reference.  This is San Francisco, this is Silicon Valley, where we never stop short of making things better, where "good enough" is quickly left behind, where innovation is prized, and where ambition fuels a relentless drive to change people's lives for the better.  There is no other place like it in America.

Will Caltrain's new General Manager have the vision it will take?


  1. Problem #11: Lack of weekend, midday and night train.
    New management need to change "Commuter Rail" mentality.

    1. 11 = 10 + 1!

      Problem #11 = "PROBLEM #1: An Agency Culture That Doesn't Put Service First" + "PROBLEM #10: Bloated Staffing".

      In fact, everything = "PROBLEM #1: An Agency Culture That Doesn't Put Service First" + (something or other which is just a special case of Problem #1)

    2. Certainly they have a major tunnel vision (no pun intended) on special events - e.g., regardless of what you think of it, the S.F. Pride Parade is one of the biggest events in the city, drawing literally several hundred thousand people. But whereas they'll run special trains for concerts at Shoreline or the several thousands coming from a Giants game, on Pride day it's the (already deficient) standard Sunday schedule - standing room only before it even gets up to San Mateo...

  2. Dwell times on popular trains at popular stations are rarely as low as 45 seconds anymore, so *average* dwell times are surely higher.

    Level boarding cannot come soon enough. With the alarmingly long dwell times I've been seeing on jam-packed trains, I'm beginning to think level-boarding might be *more* important than electrification.

    Caltrain also needs to dump the horribly inequitable fare zone system, which, for example, has many short one or two-stop trips crossing a fare zone boundary costing as much as long trips from Redwood City to SF or Menlo Park to Tamien. This has the effect of discouraging many shorter trips/commutes which Caltrain would otherwise be perfect for.

    The fare zone system is a holdover from the days the crews sold tickets onboard and used paper "hat checks" with holes punched to indicate that they'd checked a rider's ticket and how far they were riding. Since manually-issued tickets are long-gone and all ticketing is either via Clipper or TVMs, it is high time for Caltrain to adopt more equitable distance-based fares similar to BART's.

    1. Reality Check,

      .. except Caltrain, in their infinite wisdom, bought ticket-vending machines set up for zones. They'd have to replace all the TVMs. "Oh, woe, waste!"

    2. @Kiwi, zones or stations (doesn't matter) ... it's just a software change ... the TVMs run Java. I worked out the button mappings and various use-cases long ago ... staff said they'd take a look at it maybe next time they revisit the tariff. They were just shining me on, however ... since that was a good 10 years or so ago and, at the time, they had already made the decision to go with whatever zones their (stupid) high-priced consultant study recommended.

  3. Adding Cal Ave and San Antonio to the current bullet trains would only serve to lower the ridership to Mountain View. I agree that Cal Ave needs more service... but the other stations I question. AMD runs their shuttle to Mountain View instead of Lawrence because the employees value the time savings of the bullet trains. Turn the trains into locals and you'll be "able" to take the train anywhere but won't actually do it.

    1. It's important to separate the peak and reverse peak schedules from each other.

      In the peak direction, it's pretty obvious that a majority of riders want to go to SF. Secondary destinations include Palo Alto and Redwood City, maybe Millbrae. So, baby bullet service skipping through SM County is justified, considering trains are packed by the time they get to PA. The current schedule is okay, though I think having that last train stop at PA would do a really good job redistributing train loads.

      In the reverse peak direction, most riders are currently headed to Mountain View and PA, not SJ. Stopping at least the non-bullet trains at Sunnyvale, Santa Clara, etc. probably wouldn't hurt existing ridership while attracting more potential customers. Also, the timed transfer system is not very useful. A nonstop stretch from PA to Santa Clara without even stopping at California Ave or Mtn View is pretty baffling. In any case, the reverse peak schedule needs to be blown up.

    2. Current timetable is created in 2005, when Caltrain have only 30~35K weekday ridership and many hold-out stations. Current 55~65K ridership and limited number of hold-out station, Caltrain should change the timetable dynamically.

      In the peak direction, stop pattern should be simplified just express and local train. All express should stop Sunnyvale, Mountain View, Palo Alto, Redwood City and Milbrae. Some of express stop California Ave, Menlo Park, Hilsdale and/or San Mateo, depending on ridership.
      Local train become connector of express train, allowing by-passed at Lawrence,Redwood junction, Milbrae 4S platform and/or Bayshore. With multiple by-pass location, Caltrain can operate express train every 15 min while providing local train every 30-minutes.
      If this model works, build more 2-platform/4-track station for local-express transfer.

    3. "Adding Cal Ave and San Antonio to the current bullet trains would only serve to lower the ridership to Mountain View"
      Adding Transbay will lower ridership to SF Fourth Street.

      Is that bad?

      Maybe too many trains might stop at San Antonio (but NOT California Avenue!) given ridership. Just like too many BART trains stop at West Oakland. But what's the alternative? Every third train stops at San Antonio, but only before 5:45pm, and not on days containing an "r"?

      Given good (not Caltrain!) performing trains, given normal (not Caltrain!) level boarding, and stacking the cards against stopping by lowering (not Caltrain!) timetable padding, it looks like skipping San Antonio saves less then two minutes.

      So very little cost that the benefits (mostly being super simple regular clockface schedule every 15 or even every 10 minutes) seem to predominate.

      "It's important to separate the peak and reverse peak schedules from each other."


      "In any case, the [present day] reverse peak schedule needs to be blown up."

      Indeed. People have been saying that for about a decade now. Customer focused Caltrain response is "ridership is up, therefore we are doing everything perfectly. Suck it."

  4. Problem #10 A CALTRAIN = SAMTRANS. I remember once being lectured by a Caltrain HQ staff member on the "benefits" of staff sharing between the two agencies. Let's give up on that idea and have an independent, accountable agency staffed by service oriented individuals instead.

  5. Hey Clem, I'm not sure if you've seen this yet, but it looks like there's been some progress with level boarding and platform compatibility! They're at least talking about it now, which is definitely an improvement!


    1. Hey AGTMADCAT, unfortunately this troubling news came only days after the Green Caltrain blog posting:

      CHSRA and Caltrain Still Not Cooperating on Platform Compatibility

      "In a spec published just yesterday, the HSR trains are to have a floor height of 1295 mm (51 inches). This spec will serve as the basis for a train procurement process that has now begun. 1295mm is incompatible with Caltrain requirements. It really makes no sense — other than being compatible with the NEC. Why are they trying to maintain backwards compatibility with a rail line thousands of miles away?"

    2. Good news: they're talking.

      Bad news: talk is cheap. They've had half a decade to solve this problem and there has been no demonstrated action. As of this week, HSR isn't budging by even one inch from their prescribed 51" floor height.

      Thursday's TJPA agenda features a joint HSR Caltrain presentation on the issue. We shall see if they can even agree on a compromise PowerPoint slide template... gotta start somewhere, right?

  6. Problem #0: the passenger information system at the SF terminal consists of wooden paddles put into place by pitchfork operators. Solution #0: fire the pitchfork operators and replace them with one or more screens that show the next departures. Maybe even have a "big board" like at any other major train station anywhere else in the world. And maybe install these systems at San Jose too, now that there are 9 tracks and 3 operators and lots of potential for confusion.

    1. We will provide those affected employee for new assignment: Ticket inspection, and/or platform safety.

      With increase ridership, station like Palo Alto and Mountain View may necessary to be staffed in the platform.

  7. One platform (2 tracks) of Transbay should be good enough for Caltrain if they run only 6 train/hr. Fremont BART and Richmond BART handles 8 train/h with 2 tracks. Why electrified Caltrain cannot do the same?

    1. The problem isn't so much platform capacity as it is the traffic jam in the station's throat. Separate platforms, as planned, cause inbound Caltrain movements to conflict with outbound HSR movements and vice versa. In an off-nominal delay situation, this will make a mess of both systems' timetable and will severely limit the throughput of the station approach tracks. Fremont and Richmond BART easily handle this traffic because there are no conflicts.

    2. Richmond BART has tail tracks and a railyard to move trains out of the way.

      At Fremont, BART has difficulties adhering to the schedule. Trains coming into Fremont will often be stuck (for as long as 5 minutes) waiting for a platform to clear.

    3. Consist length of Caltrain (4~8 cars?) will not be long as HSR (12~16 cars?), then it can be build tail track for Caltrain.
      By the way, JR Chuo-line in Tokyo terminal handles 24~27 train/h with 2 tracks with no tail track. Your mentioned Fremont BART issue caused by poor on-time discipline of BART management.

      Another option is couple/decouple local and express train. For example, 12 car south bound express from Transbay decouple at Redwood City (or Palo Alto). Front 8 car remain express to San Jose but rear 4 car changed to local. Decoupling takes 1~2 minutes and coupling takes 2~3 minutes with automatic coupling system and additional signals.
      This is very common in Japan because of difficulty in platform extension and/or limited track capacity. However, capital expense of platform extension can be concentrated into express stop where those spending are jusified.

    4. Keikyu's operating practice is too high level and no other Japanese operator can do same level of couple/decouple within one minutes. Unless Keikyu operate Caltrain, It will takes 100 year for Caltrain to catch-up this level of operation.

    5. JR East couples and decouples Shonan-Shinjuku trains at Koganei and Kagohara. How long does that take?

    6. Ignoring for a moment that you need to have the right sort of coupling system and regulations (no FRA-mandated brake test, etc.), why couldn't this lightning-fast uncoupling technique be taught or learned?

    7. Caltrain already have waiver for using non-FRA EMU. They should request all the waiver which related FRA nonsense.

      Alon, JR East's coupling used to takes 3~4 min without automatic coupling and no leading signals. Engineer attach several jumper cable between the train manally. Odakyu is about 2 min with automatic coupling and leading signal.
      Only Keikyu do 1 min coupling because express train arrive the station with full speed (green signal) and coupling car are comes from siding which does not affected by main track signals.

  8. Clem wrote: "[...] the San Bruno grade separation, a $155 million piece of infrastructure that has achieved exactly nothing for the average Caltrain rider."

    Sadly true. And yet, amazingly, there's this:

    HNTB-designed San Bruno Grade Separation Project earns Transportation Project of the Year award from ASCE
    Awards honor outstanding civil engineering projects; new grade separation opened in April 2014

    OAKLAND – HNTB Corporation has earned the Transportation Project of the Year award from the American Society of Civil Engineers San Francisco Section for the San Bruno Grade Separation Project in San Bruno, California. HNTB Corporation was lead designer on the project for Caltrain.

    The ASCE San Francisco Section Awards recognize outstanding civil engineering projects in the San Francisco region. HNTB received the award for advancing the limits of engineering to achieve visionary architecture with the San Bruno Grade Separation Project. This year’s award winners were announced at the annual ASCE San Francisco Section meeting September 25.


    1. "American's Finest Transportation Engineers"

    2. That's just the perfect illustration, the cherry on the cake: a project run by civil engineers, of civil engineers, for civil engineers. Passengers? What are those? Huh?

  9. Clem writes: “Caltrain is searching for a new General Manager. This person will need a briefing on the key issues now facing Caltrain. “

    How can the new GM be briefed if you don’t come to meetings and make a presentation?

    You have made a good case with “The Path To Level Boarding”

    But how is Caltrain going to see this unless it is brought to their attention in a public meeting/forum?

    I would much rather see these kinds of issues discussed rather than DBE/WBE and other irrelevant bullcrap.

    Caltrain does respond to issues that are often brought up before them or in frequent written correspondence. Issues such as loud horns and bike capacity have been the subject of much criticism and they responded to the frequent complaints.

    1. Sounds like you're already going to the Caltrain board meetings, Jeff. Tell them what you think they need to know or hear and let us know how it goes. Problem solved, right?

    2. Yeah, I know 10am on Thursday sucks. It means losing a day of work but what can we do?

      There is an evening forum, the JPB/Caltrain Citizens Advisory Committee, 5:40pm on the third Wednesday each month.

      They seem to be more interested and more informed than in past (pre-Scanlon) years. Yes they are impatient, and always approve staff recommendations. They are disengaged from the customers and the service… When is the last time a board member rode on Caltrain and talked with riders?

      What solutions do you propose for a change to a more receptive agency?

      Unfortunately, we are mired in circa 1900, FRA and CPUC regulations and union work rules. Regrettably, change doesn’t come that easily. USA railroading is a century behind the rest of the world.

    3. @ Reality Check…

      Yes I do go to the meetings each month been doing it for like 30 years… I guess I am a glutton for punishment…

      I do tell them what I think they need to know and hear but as pointed out there are only two minutes to speak and I can only say/do so much. Sometimes I do get under their skin which means that I am making valid points.

      There needs to be more than just us ‘usual suspects’ speaking up each month. The more people that bring up legitimate points, the more apt they are to listen and make changes.

      The anti loud horn folks wrote to Caltrain often and also spoke at board meetings, so Caltrain made changes to the horns.

      The bicycle coalitions are well organized; they speak and write often to Caltrain so they have increased bike capacity. Unfortunately they will not add a third bike car and guess what the excuse is?

      It relates to point # 10... The conductors insist that there must be a third crewmember to police/handle the additional bike car.

  10. Have to disagree with #10 - I've been on plenty of trains with drunk & disorderly conduct, or other "ASBO" behavior, where a conductor was needed to step in and stop, or avoid, problems.

    Also disagree re: Baby Bullets - making all S.V. stops destroys the point. I lived in San Mateo and worked at San Antonio for 2 years and didn't find the schedule that problematic (other than midday, when there are no express trains at all).

    1. Reverse commute Baby Bullet stopping all the S.V. completely makes sense because of scattered office location. Increased traveling time can be offset by frequency convenience (= shorter waiting time, more options). At this point, we cannot expect run more train because of Caltrain's management.
      Traditional commute direction is still question but simplifed stop pattern is necessary. 5 different stop pattern should be consolidated into 2 or 3 patterns (Means, Baby Bullet, limited and local)

    2. Based on statistics, the baby bullet is unnecessary past Mountain View. SJ's ridership is not high enough to warrant express service at the cost of Sunnyvale, Lawrence, and Santa Clara. Below is a classification of stops south of RWC wrt reverse peak potential. I used census data and 2002 ridership information to fill in some gaps:

      Tier 1: RWC, Palo Alto, Mountain View (1-2k riders on/off)

      Tier 2: California Ave, SJ Diridon, Menlo Park, Sunnyvale, Lawrence, Santa Clara (most trains should stop - at least 3, preferably 4 per hour)

      Tier 3: San Antonio (~250 riders, 2 or 3 trains per hour should be enough)

      Tier 4: Tamien (Non-existent reverse peak destination, but some trains should run there so they can pick up peak direction riders headed north)

    3. "Based on statistics, the baby bullet is unnecessary past Mountain View ..."
      That's true, in the sense that limited-stop service is unnecessary, and in fact counter-productive.
      It's also somewhat true, in that today's poor Caltrain "reverse" peak ridership drops off a cliff south of Mountain View.

      But against that, consider that there isn't a huge operating or equipment savings from turning back trains at Mountain View. With better (non-Caltrain!) equipment and level (non-Caltrain!) boarding Mountain View to SJ Cahill is about 16 minutes, meaning that turning back every second train (for 2tph Mountain View—SJ vs 4tph or better SF—Mountain View) only ends up saves one train and crew out of 20 or so, and 36 train-km/hour which is under 4% of the total. Oh, and Sunnyvale is clearly under-served today, but if the turnback is moved further south there's even less saved.

      A valid technical argument can be made for doing that, but given the "Capital of Silicon Valley" small-endowment-compensation political nonsense, it's north worth making it. Just suck it up and run a couple extra nearly empty trains back and forth to try to make the bleaters shut up for a little while.

      Basically it's hard to beat a peak service plan of 4tph SF—Redwood—SJ (limited north of Redwood) with 2tph (maybe 4tph) local shuttles SF—Redwood to cover the limited-stop gaps.

  11. Turning back at Mountain View or Sunnyvale is making sense with current FRA operation.
    There is space for one more track in NB platform at Sunnyvale station. Let's ask VTA to pay for this additional track as the justification of Levis train turn around facility. (They paid for light rail)
    Caltrain can operate 3 train/h between Tamien to Sunnyvale and 6 train/h between Sunnyvale to SF under same cost/equipment/crew size.
    3 train/h (every 20 min) depart from SJ diridon should be good enough for current demand. Tamien, Santa Clara and Lawrence will get more convenient every 20 min service.

    1. Even with bad slow FRA equipment and long dwell times and long turnback times, turning back at Sunnyvale instead of SJ Cahill only saves one or perhaps two trains.

      In fact, the less efficient the trains (worse performance, longer dwells, longer turnbacks) the less the savings, there being only three additional stops.

      I agree that this ought to be worth analyzing especially given Caltrain's constant bleating about poverty and incessant demands for extra public money to throw away and burn. But given the infantile politics around The Capital of Silicon Valley, a 5% operating cost saving (and note that Caltrain never pays attention to any objective metric anyway) isn't going to be enough to considered.

  12. SF and SJ Diridon have so many tracks/platform. That's why Caltrain have no motivation to reduce the turn around time in each terminal. In this point, built additional tracks/platform at SJ Diridon was completely waste of money.
    Turn around at Sunnyvale half the rush-hour train need ~15 min of turn around time. 15 min is archievable number as weekend baby bullet turn around at SF only 20 min but with many riders.
    As Richard mentioned, saving from Sunnyvale to SJ is only few stop. However, turn around at area wtih limited truck capacity makes Caltrain no excuse.

    1. Northeastern commuter lines routinely turn in 5-10 minutes at their outer ends, even push-pull diesel operations like the MBTA.