average weekday schedule for 3-car Link trainsReader Greg Briggs wonders “if there is a way to just get the full low level schedule” for 3-car Link trains, because they’re much more comfortable to ride than the 2-car trains during rushhour.

There is, although it’s important to remember that Sound Transit will change the schedules on days of big events. Furthermore, if there’s a disruption, operations will try to maintain 6-minute headways rather than adhere to the schedule.

Regardless, at right are the 35 trips from UW and 31 trips from Angle Lake that have 3-cars in them on a typical weekday (thanks to Oran) [Correction: the original schedule had the directions swapped. It has been fixed.].

45 Replies to “The 3-Car Train Schedule”

  1. Great data! Now we just need to work out the schedule at each station?

    The other thing worth mentioning is that these runs seem to avoid an operator change that you otherwise get in the later side of peak.

  2. What’s up with the late night 3-car trains? Trippers? Or some of the PM peak ones stay in rotation while the 2-car ones are pulled?

    1. I was wondering that myself. I thought there was a baseline of two-car trains that formed 10 minute service until 10pm ish, then additional three-car trains were inserted (and departure times adjusted) during peak only for 6-minute service.

    2. It was reversed in the last announcement a month or so ago. Now three-car trains are the baseline, and two-car trains are the peak extras. That accommodates the frequent spikes on weekends and midday and evening ballgames. The peak extras are two-car because ST doesn’t have enough cars for tree-car trains until the North Link order arrives in a few years.

      1. The reversal is only for major weeknight events, such as the Seahawks’ Monday Night Football game.

        AFAICT, 3-car trains are still running all day on weekends, at least when there is some sort of major event, which seems to occur more often than not.

      2. We’re only running the “inverted” schedule with the base of three-car trains and two-car supplements during peak hours on weekdays with special events like MNF. Weekends are all three-car trains with or without special events.

  3. What would be even nicer is if OneBusAway showed the type of train. This would also be useful for peak express routes, so you could tell if Metro is running an articulated bus on that particular run.

    1. OBA for Android does. Find the arrival time in the list, tap it, and select “Show trip status” from the popup list. The vehicle number is displayed at the top which you can cross-reference with this list. The number is shown as x_yyyy where yyyy is the vehicle number and x is the agency ID: 1 for Metro, 40 for ST, etc.

    2. I can’t believe anyone would wait for a three-car train when a two-car train is coming in a few minutes. The wonderful thing about Link is you never have to wait more than five or ten minutes for a train unless something’s broken or it’s after 10pm. Waiting 10-20 minutes for a 3-car train reminds me of the experience of scheduling your life around an infrequent bus and the forced idleness of waiting for it. I’d rather just take the next train that comes, even if I might have to stand.

      1. If they’re running on schedule, sure. My experience has been that if two are coming close together the first one will be a packed 2-car train while the second will be a spacious 3-car. Without real-time data (seriously, we still don’t have that?) it’s a crapshoot unless you hear the “next train in 2 minutes” announcement twice before any train arrives.

      2. Anyone who has difficulty standing for an extended period in a moving vehicle, or perhaps parents with more than one small child in tow, is not likely to be willing (or able) to board a packed train, especially at University Street and Pioneer Square stations southbound, where Westlake crowds have already filled the train.

        If you are able to plan ahead, or have some idea whether waiting for the next train is a good bet, it can be a good improvement. Some of these demographics have already learned that the train is not a usable option during peak.

        Or if you work downtown daily, you might “aim” for the 3 car train as a choice of preference, if that was an option.

      3. I hadn’t thought about the issue of the 3-car trains having shorter dwell time (well, I thought about that half, and am on the losing side of a debate as to whether they should all be three cars, all the time) and the 3-car trains bunching with the 2-car trains. But now that you mention it, I have seen 3-car trains come through at SODO Station 2-3 minutes after a packed (by American I-don’t-want-to-stand-in-the-raised-area standards) 2-car trains. The 3-car trains were reasonably full, too, except for the third car.

        The obvious stop-gap solution is to always have the 3-car trains wait an extra minute before starting their runs, to even out the load factor.

        If Sound Transit were willing to run 3-car trains all the time (and it isn’t), riders would get used to waiting over the full 3-car length, dwell time would drop, DSTT congestion could be reduced, and running only 18 peak trains might work, and manage to maintain 6-minute headway much better than now. If there were only 18 peak trains, all 3-car, ST would meet its 10% LRV spare ratio.

        More even headway, better distribution to all three cars, and the realization of consistently more capacity would induce more peak ridership, I would hypothesize.

        The downside is running up mileage by having extra relatively empty LRVs running all day, thereby ballooning maintenance costs, and shortening the useful life of the current fleet. For those of us who can’t wait for ST to get into the open-gangway business, that last point may be a feature, rather than a bug. Alas, the new LRV order is clearly not open gangways.

  4. Does that represent a new schedule? In my experience, PM peak trains leave Husky Stadium at 4:57, 5:03, 5:09, 5:15, 5:21 etc.

      1. Yes. But I am saying that there is no train leaving Husky Stadium at the times (at least in the 5PM hour) that this schedule says it does.

    1. Pretend the headings are reversed, and look again. I think we editors just didn’t catch a simple transcription error.

  5. Maintaining consistent headways is far more important than holding to a rigid timetable. Link light rail is not a bus route.

    1. Except it’s very common to transfer from Link to a bus route, and I always look at the schedule and find which train gets me closest to the bus departure time. If it’s really close, I’ll take the train before, but if there’s something like a 7 minute transfer window, that’s pretty much ideal.

      1. Alex, as long as either vehicle has to contend with street traffic anywhere on the route, schedules times can be sometimes workably close, but never precise.

        In one capacity or another, I’ve been struggling since drafting board days in 1983 to get buses to rail efficiency. Same with trains, and combination of the two. Even if it’s the last day of joint ops, it’d all be worth it to see the Tunnel work anywhere close to its potential.

        Which unfortunately has always required a degree of training and teamwork that maybe three people on Earth besides me ever considered worth a dismissive thought. “Relax. They all come out the other end!” Don’t they, Roger?

        But one of the DSTT’s worst rush hour obstacles is a driver who insists on running the Tunnel to schedule. Backing up several hundred passengers behind him all the way back up the tube. See above paragraph about acceptable level of training. And supervision.

        Tunnel was also designed for Proof of Payment. Fare-boxes in the DSTT tells me that the dissolution of the Municipality of Metropolitan Seattle was the worst mistake our transit system ever made. Those boxes might as well be iron wheel-blocks.

        So I think you’ve already discovered that at this stage, best way is to board LINK early enough to be on the Tunnel platform ahead of scheduled departure from Staging for your bus. Erring on the generous side for both modes. At least in the Tunnel you’re out of the rain.

        But also, my every comment on joint operations has to be judged in context that if some ST-3 or other problem keeps joint ops alive for twenty years, I’d still be bugging the system to get it right!

        Mark

    2. Is maintaining headway important enough to invest in running 3-car trains all the time, to avoid the bunching caused by longer dwell time trying to squeeze a few more people onto the 2-car trains, which involves getting some standers already on the train to get out of the way?

      1. Link drivers should not be extending dwell times to allow every last straggler onto the train. But Link drivers are all former bus drivers, who are taught to do just that. Riders need to understand that the next train will be along in 6 – 10 minutes, and just wait for it. Trains need to operate like trains; they are not just big buses.

      2. To be honest, the same goes for the tunnel bus routes that are very frequent (41, 550, etc.) Have often seem people running for a 41 NB at Westlake, which then holds to wait for them, delaying a following train–not always, but often. I just want to tell them “there’s another one in 4 minutes!” (they’re not just running to the Lake City one) The platform attendant that deals with back door boarding should tell them.

        Of course, this would be reduced, probably dramatically, if there were real-time arrival info for all trains/buses provided on platforms and mezzanine. It still staggers me that we don’t have that at our busiest station–they could even have somebody at each portal manually enter it if necessary, since once vehicles enter the tunnel the order remains the same and time shouldn’t vary too much.

  6. If you are boarding at UW Station or Angle Lake Station, board the rear train. If you are waiting at any other station, look across to the marker on the other platform for the front of the trains going the other direction. Wait in line with that. If your next train has a third car, board it. It will be less full than the front two cars. If your next train has only two cars, the rear one will still be less full than the front one.

  7. Maybe it’s Chicago and San Francisco experience, but I think that on any transit line with enough passengers to justify tunneling, a train with a single empty seat at rush hour would be canceled for lack of ridership.

    Because exactly like with lane space on freeways, more cars don’t mean more empty seats. Just more railcars full of people. STB rightly considers Density in land use and building occupancy a good thing for transit. So therefore…..

    https://en.wikipedia.org/wiki/R160_(New_York_City_Subway_car)

    No conspiracy of silence here. Planners and engineers just assumed that everybody who loves transit also loves big cities, which people depopulate whole rural regions to flee to. Leaving behind the voting demographic that elects State legislators who hate transit. And therefore have already given hearty consent to close rolling quarters with other cross Cascades refugees.

    Benaroya Hall Rhapsody in Blue” audiences sadly have an unspoken (if they know what’s good for them) understanding that George Gershwin would never have been inspired to write “Rhapsody in Blue” if noise suppression measures had demanded a muffled bass drum instead of the horn section for the elevated line’s signature melody.

    So maybe we can re-name LINK “The Alki Line”. After the promoter’s promise, as a suicide control measure as his fellow passengers stood on the ship’s railing tying ballast to their legs because it was swifter than drowning in the rain, that Seattle would shortly- Chief Seattle’s people said “Alki”-become New York City.

    Mark Dublin

  8. BTW- difference with freeway is that the standing passengers on trains will get home a lot sooner than the sitting passengers in the lanes. Giving standing train riders a psychedelic view of lit-up red and white slowly-moving lines a lot more “Oh, Wow!” than Rainier in the moonlight.

    Mark

  9. It’s too bad they don’t have enough cars to just run 3-car trains full-time. Having varying train lenghts is silly.

    1. My math says they do have enough if they can run the peak algorithm with just 18 trains. My sense is that there is padding going on within the trip to enable 19 trains to fit, thereby making the trip a couple to a few minutes longer than necessary. That padding is annoying when it happens all day to keep consistent travel time. Move that padding into longer breaks for the operators at each end, please. And have off-peak travel times reflect how long it really takes off-peak.

      Running all 3-car trains consistently, with markings along the platform to get riders to spread out, should cut dwell time by about 30%. Removing the padding will drop it by another minute or two or three.

      Yes, they can. Yes, it would also substantially raise maintenance costs. But they have to gradually ramp up maintenance staff anyway to be ready for the next extensions.

  10. In the peak periods, all trains could have three cars, if the scheduled headway was seven minutes rather than six. More capacity would be provided. No intending rider would face a two-car train. This would have the benefit of both Link and buses going faster in the DSTT, as the friction between modes is reciprocal. The average wait time would increase to 3.5 minutes from three minutes. Until new LRV arrive and are tested in 2019 or later, there are only 62 LRV.

    1. It wouldn’t be a clockface schedule with 7 minute headways. Six min. = ten trains per hour; seven min. = 8 4/7 trains per hour.

      1. For a while, TriMet was doing 7.5 minute headways (half the 15 minute clockface), which solves that particular issue. However, once they get down to that frequent stuff like surface disruptions play much more of a role in screwing with the exact headway than being able to tell the difference between 6 minute headways and 7 minute headways.

    2. I’m not convinced headway would have to be increased to run all 3-car trains. Has anyone else noticed much longer dwell times than necessary, including during off-peak?

      1. On the other hand, do you really want to sacrifice reliability to shove in frequency? A 7.5 minute schedule with all three car trains might be better than an unreliable 6. Below about 10 minute headways the passenger perception is headway decrease is not as significant as making changes from, say, 12 to 10 or 20 to 15.

  11. Is a clock face headway as important as the other factors if the access and egress time at stations varies so much among passengers?

    1. One of the ideas of a clock face timetable is that you can reliably leave a certain number of minutes before the scheduled departure and get to the train. If that is 5 minutes that you need then it is 5 minutes. If that is 10 minutes then it is 10 minutes.

      If you are dealing with service every 12 minutes, you can’t reliably leave at a certain number of minutes past the hour every time. You have to know the timetable.

      I’ve been told by someone who should know that operators in Europe that go with a clock face schedule system can double their ridership just with that change.

      One bus route I use from time to time is now every 35 minutes. It’s terrible to try to plan around.

      1. Glenn, 12 minute headways do give you a clockface schedule. That’s five trains per hour. As long as it’s an integral number of TPH, it works. So, 1, 2, 3 4, 5, 6, 7.5, 10, 12, 15, 20, and 30 minute headways all work.

        Maybe you meant to say the original 7 minutes.

      2. BTW, I feel like a dummy for not picking up on the possibility that maybe “7 minute headways” was really intended to mean 7.5 minute headways.

      3. Apparently I wasn’t thinking that through too well.

        Even 7 can be made to work if it is 7.5 instead. That’s a double frequency of a 15 minute service.

        12 isn’t quite as intuitive as 10 or 15 to figure when the next arrival will happen. However, at least it doesn’t drift over time.

        It’s the stuff that drifts over the course of the day that is much harder to work with.

Comments are closed.