Link riders at UW Station

From now on, board the third car if you want to avoid crowds like this.
Photo by SounderBruce

Sound Transit will be rolling out increased capacity on Link Light Rail beginning Monday, featuring twelve 3-car trains from opening of service to at least end of PM peak, seven AM peak-only 2-car trains, and seven 2-car trains in the PM peak that will stay out through close of service. On event nights, all 3-car trains will stay out, and the 2-car trains will come back to base at the end of PM peak.

What that means for those of you waiting for a train on a weekday, delineated by time of day, is as follows:

Before 6 am: Expect a 3-car train. Stand to the left of the last between-car-barrier, and board the third car, for maximum spaciousness.

6 am to 9 am: You have a 12-in-19 chance of your train being a 3-car train. Stand near the left-most between-car-barrier between the second and third car, and be prepared for either length.

BH Link Station, southbound platform.

Between-car-barriers / photo by Matthew Bates

9 am to 3 pm: Expect a 3-car train. Stand to the left of the last between-car-barrier, and board the third car, for maximum spaciousness.

3 pm to 6 pm: You have a 12-in-19 chance of your train being a 3-car train. Stand near the left-most between-car-barrier between the second and third car, and be prepared for either length.

6 pm to close on Mariners game nights and other major event nights: Expect a 3-car train. Stand to the left of the last between-car-barrier, and board the third car, for maximum spaciousness.

6 pm to close on non-event weeknights: You have a 5-in-12 chance of your train being a 3-car train. Stand near the left-most between-car-barrier between the second and third car, and be prepared for either length.

Weekends and holidays have been all 3-car trains for most of the past year. However, most riders have not caught on. So stand to the left of the last between-car-barrier, and board the third car, for maximum spaciousness.

If you are boarding in the transit tunnel, there are no between-car-barriers yet, so just stand toward the left end of the platform any time besides weekday peak and weekday evenings.

In Westlake Station, there are floor markers showing where the doors of the first two cars are, so use them to judge where the break would be between the second and third car. Please don’t stand on the arrows pointing away from the train. They are there to provide a clear path for riders to disembark.

Westlake Station door markings

Westlake Station door marker / photo by SounderBruce

55 Replies to “Midday 2-Car Link Trains End Today”

  1. The suggestions for where to stand relative to between car barriers should be the opposite (“right-most” and “right of”) for all center platform stations (i.e. Beacon Hill).

    1. +1 I was confused until I saw this. I mostly use Capitol Hill and UW stations, so it seemed backwards to me. I forgot it would be opposite for the stations with non-central (side?) platforms.

  2. I know ST says it’s a gigantic technical hurdle that would require overhauling the communication system, but is it really so hard to have two alternate messages play based on car length to tell us the next 3 car or 2 car train is arriving in two minutes?

    At a bare minimum, they could simply show the number of cars show on the front display of the train so at least you could see and react from a distance instead of waiting until the back gets out of the tunnel tube to see the length.

    1. Even easier – have a big card with a “2” or “3” displayed in the driver’s cab. Keep both number cards in the cab and the driver can display whichever sign indicates the car count on that particular train that day. Surely ST can get to Staples and pick something up…

      Washington DC (no paragon of transit-agency capability) has had electronic train-length displays since at least the 1990s. Of course, I forget that Seattle is super special and nothing works here (open gangway trains, train length displays, not having an inane announcement every other minute, enough fare machines and card readers, working elevators and escalators, etc. etc.).

      Sorry for the rant, but once you experience other cities/countries doing transit you have to bang your head into the wall about the issues we’re still dealing with here.

      1. I’m not sure what’s worse: these basic design mistakes apparently from naïveté or the refusal to acknowledge and act on them.

      2. Personal Opinion, ST as an institution is slow to change its mindset from a construction agency to an operations agency. They also tie everything to a plan (ST 1, 2, 3, etc.) and heaven forbid they have the agility to make a move outside of the boundaries set in one of those plans. It has happened, but not as often as it should. When ST originally spec’d out the next train system it was probably typical of the technology of the era, in this case those decisions were being made in the late 1990s. Time has progressed on, the system is now obsolete. They probably did not want to mess with the budget and risk more than they had to for the Capitol Hill extension, which this should have been added to it as they were upgrading the SCADA system anyway but they did not. I think its coming with Bellevue or Northgate though. Why they cant spend that money early and get those features installed within the next couple of years is beyond me, but it probably either had to do with scheduling the funding or how it lines up in the construction contracts. In any event, it just goes to show that they are unsure of who they are and what their real function is since they have so many different competing interests.

      3. With what money? ST 1, 2, and 3 had a budget ceiling based on what they foresaw that they would need. To do more would require going back to the voters for more money, and people are feeling pretty maxed out right now with the magnitude of ST3 as the reaction against the car tabs shows. The money doesn’t all come at once, it comes in a trickle every month, so you can’t just advance expensive things. If they could do that, then they could get Ballard done sooner. Maybe ST should have thought more long-term when it had a window of opportunity to change things in 2008 and 2016. But it would take somebody with a lot more knowledge of ST’s budget details and the details of the engineering challenges than most of us have to say credibly how much more it could have done.

        As for “They probably did not want to mess with the budget and risk more than they had to for the Capitol Hill extension”, isn’t that a good thing? ST has been risk-adverse ever since it was too confident in the 1990s, and as a result we actually have a line running between UW and Angle Lake, and ST has a good credit rating, and its critics can’t point to anything substantial it has done wrong beyond having a big party for U-Link opening and maybe not shouting to legislators “This is the old expensive car-tab formula!”

        “it probably either had to do with scheduling the funding or how it lines up in the construction contracts”. And if it was, what could ST have done about that? If things can only fit in a certain order, and take so much time to complete, and can only be done when their chunk of money comes in, then that’s the way it is.

        The only persuasive critique I can see in all this is that the board could have set the projects’ priorities better in the beginning; e.g., told the staff to favor newer state-of-the-art alternatives whenever they had a choice in designing. But regarding the details of the projects and their execution timelines, the critics have offered no evidence things could realistically have been better.

      4. BART also manages to do train length announcements. I can’t remember if they do verbal announcements, but definitely the LED displays show it.

    2. In 2023, there will be two lines anyway. The system must change eventually!

      ST should be on this right now! They can simply input a two-car line as one line and a three-car line as another in their system. Then, all the 70k riders each weekday can know!

    3. They can’t even get real-time-arrival right. How can they possibly get this right?

    4. ST’s current plan is to have all 4-car trains, all the time, commencing with the opening of Northgate Link in 2021. So, any additional train-length announcement technology would have only a few years of useful life.

      Moreover, ST should be able to move to all, or nearly all, 3-car trains when the buses get kicked out of the tunnel in September 2018. If it still has to run 1-4 2-car peak trains, the first few new LRVs arriving in 2019 will bring it up to fleet capacity to run all 3-car trains by sometime in 2019.

      However, turn-back trains of shorter length remain a possibility between UW Station and the SODO base for 2019-2021, once enough new fleet arrives.

  3. What that means for those of you waiting for a train on a weekday, delineated by time of day, is as follows:

    Before 6 am: Expect a 3-car train …

    6 am to 9 am: You have a 12-in-19 chance …

    6 pm to close on Mariners game nights and other major event nights: …

    Is there an app for that? Seriously. I have never coded up phone apps, but it doesn’t look that complicated. Maybe a little chart would make sense.

    But then again, I think Patrick is spot on. Knowing (instead of guessing) how long the train will be would go a long way towards getting people to spread out. Or maybe, as you put it, stand near the left-most between-car-barrier, and be prepared for either length.

    1. I’ve actually built a OneBusAway app for Windows 10 PC and Phone (OneAppAway if you want to download it, but it’s very buggy at the moment). It is actually quite complicated. I’ve been working on a bug update, but it’s taking a very long time with how little time I have for it. One of the features I’m working on is being able to see whether a Link train is 2 or 3 cars, but it’s based on today’s schedule, which is obsolete.

      I got the schedule of trains from STB’s post on the 3-car train schedule. Does anyone know where to get the schedule of 3-car trains from Sound Transit so if/when I release this thing, it will have the correct 3-car schedule?

  4. I will ask the big elephant in the room question: Why, after months of avoidance or excuses responding to outcries, did ST change things with only a four-day notice? If it was this easy, why wasn’t this done last year — when the problem was first apparent?

    1. My guess is with ridership hitting 70,000+ weekday for the first time this past April (and the summer is the busiest season for ridership), Sound Transit decided rolling out more 3-car trains is finally necessary. They’ll be stretching their fleet a little thin until the new LRVs arrive in 2019, and probably wanted to hold off on more 3-car trains as long as possible.

    2. Didn’t they just take delivery of a bunch of new rolling stock? Seems to me it was just this month, so having a few weeks to figure out which trains to lengthen doesn’t seem that out of line.

      1. I believe that was an order for more train cars. Sound Transit was using a clause that allowed them to purchase more cars at the same rate as the first order. They weren’t able to buy all the cars at once due to the amount there are aloud to be bonded out and capital in hand required. The first of the new cars orders will be arriving in 2019 at which point every thing will be three car trains. so in approx. 18 months this discussion about 2 vs 3 car trains wont matter. and by 2023 they will all be 4 car trains.

      2. The powerful and unaccountable ST Board gave emergency powers to Emporer Constantine, who decreed a Grand Fleet of the Republic. The Grand Fleet made an immediate and sudden appearance, when the Emporer announced Order 101, sparking terror in the Outer Counties.

        Rumors of a secret Order 66 have neither been confirmed nor denied.

  5. Still disappointing that peak hours have an inconsistent train length. I’d prefer a slightly longer headway and consistent 3 car trains.

    1. Or if they could at least announce the train length ahead of time or as it’s arriving to the platform. TBH I prefer the shorter headways – it makes for smoother bus-Link transfers.

      1. Because having very crowded trains (two car trains) and somewhat less crowded trains (three car trains) isn’t as good as slightly less frequent less crowded trains. It sucks to be left on the platform with two car trains. At least with the new pattern maybe I’ll be pretty much assured that if I don’t squeeze on a two-car train the very next train will almost certainly be a three car train. It’s really annoying to not get on a two car train hoping that the next train is three car only to find out it’s another packed two car. (I usually board southbound at Westlake during afternoon peak when I’m not biking)

      2. Some people want space, others want to wait less. There’s no universal answer that pleases everybody. But there is the general principle that frequency is among the top two or three concerns of passengers. People would rather be on a train or bus moving perhaps less quickly than they’d like, than standing on a platform waiting for it. Especially when there’s no real-time clock to give more certainty and make the time go faster.

        As for low-cost solutions, as I’ve said before Moscow and St Petersburg have count-up clocks that show how many minutes since the last train. That at least tells when the next train should come, and how likely it is that one will come in the next couple minutes. Link has two of those, a southbound one at Mt Baker and a northbound one I haven’t seen, but they’re for the drivers not the passengers — to avoid train bunching — so they aren’t at every station.

      3. The mix of 2- and 3-car trains causes train bunching. The 3-car trains get everyone on and off using less dwell time.

        The shorter wait time for more frequent mixed-length trains is at least partially made up for by longer times getting through the DSTT than if all the trains were 3 cars, especially while bus platoons still have to transit the tunnel.

        ST has opted to prioritize clock-face schedules and minimizing wait time. I disagree, and favor reliability and minimizing travel time, but it is a debate that will likely become moot within a couple years.

      4. There may be uneven headways but I’ve never seen catastrophic “train bunching” with one train immediately after the other. Although that may explain why Friday afternoon from UW southbound there was one train two minutes after another train. (I arrived for the second one.)

  6. It ain’t that hard. Just stand near the middle of the platform.

    If it is a 2-car train walk towards the 2nd car.

    If it is a 3-car train walk towards the 3rd car.

    If you can’t figure out the difference between a 2-car and a 3-car train take the bus.

    1. Additionally, The worst place to wait is where a train never stops (until four car trains): the first 100 feet of platform. Especially cumbersome in the DSTT where all the passengers waiting for busses are (except the 550).

      You can never go wrong and will probably find a seat to sit the farther down the platform you stand. But don’t tell anyone – it’s our secret!

      1. Jeff, suddenly dawned on me that subconsciously, especially since present loads are so new, many passengers gravitate to the front of the train because this is where they’ve been boarding buses all these years.

        Possibly proximity to the driver. Just in case maybe they might need to ask something. Matter of habit, not intellect. Would be interesting to take polls to find out how many passengers every day have barely started riding LINK at all.

        Good chance to also survey passengers as to why they gravitate to the front cars. Also, might be good to add an advisory animal to the illustrations of bad transit habits.

        Maybe a graphic of two whole cars stuffed with Seat-hogs and their luggage in the “Don’t” frame, the “Good” Seat-hogs and their luggage evenly spaced, or front two cars packed with Seat-hogs who’ve put their luggage in the third car.

        Honestly, I think we’re only a year into a transit system that could last a hundred. And is unlike anything transitwise Seattle had ever done before. Not only is it too much to ask that everything be worked out to the letter right now. Anything rigid now is going to break.


    2. Good grief. Thank you, Lazarus.

      And don’t stand on the arrows?!?

      1. Yes, people are haplessly standing on the arrows pointing away from the train. Maybe there needs to be printed instructions with the arrows.

  7. Why would they take the 3-car trains out of service at the end of the PM peak instead of the 2-car trains? That makes no sense. Is it really that much more expensive to run a 3-car train than a 2-car train?

    1. I’m not in maintenance but I would assume that it is b/c the 3-car trains will require much more work. Not only are they 50% longer, but they will have been out 2x as long (and so need deeper cleaning). Someone who is better at math can work out how much more work (time) is needed.

    2. My guess would be the consists are semi-fixed sets. While cars can be switched out they probably remain in their 2 or 3 car configurations, until there’s an absolute need to pull a car for maintenance or inspection. It would not surprise me if the inspection times are the same on all the cars in the consist so they don’t have to break the sets up nightly for work. So theoretically, the same 2 or 3 cars could be coupled into a trainset for upwards of 30 days or even longer depending on the inspection schedule. By bringing the 3 car trains in earlier, and since they are the first cars to go out the staff have more time to clean and maintain them every night, whereas when the 2 car sets start to come in, most of the 3 car sets will already be done being serviced and the 2 car sets can than be worked on for the rest of the night since they are departing later the next morning.

  8. Longer trains require more electricity to operate but I think the real reason is security. Fewer cars in a train at night means more people in each car, which increases security. In the old days a fair number of agencies operating rail lines (including I think BART) cut trains in half during off-peak periods but I think it got to be too much of a hassle. Not sure if CTA still does it, that’s another one I’ve seen before cutting train sets in half at the Howard yard.

  9. Seattle has only had anything like modern rapid transit since March 19 last year when UW Station opened. To me, fact that the public gravitates to the first two cars is a good sign.

    Because when we really get up and running, however many cars we’re running will always look like that. Longer headways with more seats? Same as conveyor built with larger scoops, but more space in between.

    The longer time between trains, more people will accumulate on the platforms. To pack the cars same amount when they arrive. Remember, a rush-hour load is like sand or gravel pouring down a chute. Slow the scoops and load will begin to spill off the platform.

    Slide might work better than stairs and less repair than elevators an escalators.. Smooth, no moving parts. Only problem is that grown-up passengers will be shoving kids out of the way to scramble back up the stairs and slide down again. Turning Security into hall monitors.

    Being headquartered in Sweden, Securitas could have a problem. Geneva Conventions forbid fifth grade girls in plaid skirts and highland knee socks who twist violators’ ears ’til they squeal. Evelyn. First time I understood that WWII wasn’t over yet and we had to keep fighting ’til she went to junior high.

    But really, snowballing ridership will take care of spacing problem. Because in a couple of years max, sheer number of passengers will squeeze everybody like Tom’s organic toothpaste (Well this IS Seattle) up and down a 360′ tube.

    Saw chilling fore-taste this aftenroon: An actual New York City subway passenger confronted with a broken elevator and a guard yelling at him (always losing contest for us vs. NYC) not to run across the tracks.Trust me. One empty car out of three, let alone out of a thousand, is a fleeting problem.


  10. today, with Angle Lake to UW, maximum capacity would be provided with seven-minute headway and three-car trains. All trains would have three cars. DSTT would flow better for both modes. Why is midday service at 10-minute headway? Let’s offer less waiting.

    1. If only 11 of the 61 LRVs are available for maintenance the whole day, maintenance will fall behind quickly.

  11. Some more PR gold for ST’s wonder child which is excellent for getting to the next neighborhood and not much further.

    Meanwhile a simple cost effective improvement to Sounder that is > connector buses < goes ignored :-)

    1. Sounder has some buses that feed stations but the network is hardly complete.

  12. How much are crowded trains a problem? I travel mostly between Capitol Hill and UW and any car I board has plenty of room. I mostly choose the car closest to which exit I intend to take. I have seen cases where there’s room for only one or two more people, but it has been a long time and mostly just at ballgames.

    1. The bottleneck I’ve observed is southbound PM peak, with the trains fullest between Pioneer Square and International District Station. That’s without adding game traffic.

      There may be other bottlenecks I haven’t observed.

      Our expectations of personal space on a train may be less than others, for whom the appearance of full traincars may be a deterrent.

    2. Busiest stretch is between UW and IDS Stations with the worst of worst typically being near the ends of that stretch.

      But I’m with you, Although crowding occurs, I don’t think it is a huge issue – YET. Hopefully this tweet by ST will do an even better job matching capacity to demand.

      And don’t forget, once the buses come out of the DSTT ST can add an interlined turnback line and add both frequency and capacity to the core where it is needed most. But the buses need to come out first.

      1. There is, at best, barely enough fleet to run all 3-car trains once the buses come out, and no additional LRVs available until sometime in 2019. Reducing any of those trains to 2-car trains is likely to result in some passengers occasionally having to wait for the next train to head to the south end during PM peak, especially on game nights.

        Doubling the frequency between SODO and UW is unlikely to be viable until late 2020 or so. Nor is the SR 520 reorg to maximize transfers at UW Station likely to blow northbound PM peak capacity once all the trains are 3 cars. As Mike Orr points out, the northbound PM peak trains are already pretty spacious, even with a dominant 2-car train deployment.

        I feel the pain of SR 520 bus commuters whose commutes are going to get worse while waiting for East Link to open. But doubling the peak frequency on part of Link will only reduce their average trip time by 1.5 minutes, or really a little bit less than that as headway reliability decreases.

      2. @Brent,

        A turn back line alleviates the LRV availability crunch by only putting the additional resources where they are needed. Basically only Stadium to UW would get the extra resources.

        Essentially ST would be “rightsizing” the service levels in the core and periphery of the system. Essentially you get more service and frequency in the core with fewer LRV’s because you aren’t wasting those LRV hours at Angle Lake.

        It is a path to more and better service with fewer LRV’s.

        But the buses need to come out of the DSTT first.

      3. @Lazarus,

        Do you have data suggesting south end peak ridership can be handled by 2-car trains, and that 3-car trains running every 6 minutes will not be able to handle north end ridership?

      4. I often see large masses getting both on and off at Capitol Hill. But I’d call it “surprisingly robust” rather than “overcrowded”. In other words there’s plenty of room for another person even if it’s standing room only. (And it’s rarely standing room only for me; only at ballgames.)

    3. In the morning rush, the two car trains heading north (from Angle Lake to Downtown) are usually very heavily loaded by the time they get to Beacon Hill. I would imagine dwell time at Beacon Hill is affected by folks trying to pile in. I wouldn’t call them crush loaded, but a minute or two delay might get them there.

      This morning was a good example, although it might be a result of people planning on catching a 3-car train and finding the schedule reversed. Our train that left CC at 7:40 was filled to the brim. Again, I wouldn’t call it crush loaded, but it was getting uncomfortable and loading and unloading were starting to take a while.

Comments are closed.