It’s the seventh day of the restructure, and Link had power outages this week. How are things today?

Mini news roundup:

What happened to RapidRide G this week and how is Metro fixing it? (Ryan Packer, The Urbanist)

How to deal with cost overruns in light rail extensions? (Stephen Fesler, The Urbanist)

Sound Transit executive commitee grapples with West Seattle Link and its costs

Open threads, the Friday edition, commentary that tells the agencies and governments where to go.

This is an open thread. (The tagline in the middle paragraph is a spoof on the As It Happens intro.)

55 Replies to “First Week, Part 3”

    1. Reading the interview, it sounds like someone well-versed in politico-speak. I see her strongly advocating for not much.

  2. I read recently, I think in Capitol Hill Seattle blog, that much of G line delays could be attributed to buses having to wait for up to 5 light cycles to get a green light on Eastbound Madison/Boren… (If I recall correctly)

    Basically SDOT fucked up the signalling to make anti-transit priority or something…

    1. At Boren the buses would be in the middle lanes going straight, so that wouldn’t be it. They might have to wait for a long cycle, but only one.

      It could be the turn from 9th to Madison (eastbound) but from what I’ve seen, the problems are all over the place. Just looking at One Bus Away (as well as reading anecdotal stories) there were various delays throughout much of the route. These problems compounded themselves because drivers would arrive back at the base way too late (and bunched together). The drivers need a break, which means there would be a big gap (e. g. ten minutes) with the buses leaving Madison Park. This would make that run even worse, as it would get a lot more riders which means that riders farther down the line (e. g. First Hill) get delayed even more. This is classic bus bunching, but it was happening early in the run (not late).

      As Ryan Packer mentioned, having more drivers helps alleviate the big gaps of buses leave Madison Valley. But it doesn’t address the cause of the delay in the first place. Based on the article, it appears it is a bunch of things that they are dealing with (and addressing). It may take a while before it is has the kind of service we expected.

    2. I misremembered the intersection: CHS says 5th/Madison
      https://www.capitolhillseattle.com/2024/09/rapidride-g-arrives-with-ambitious-public-transit-goals-and-plenty-of-hiccups-over-madisons-new-buses-stops-signs-and-signals/
      Also unclear but this might have just been an opening day issue/ might have been fixed since:
      ‘The launch brought some major signalling screw-ups with the city. At 5th and Madison, the bus hit a light right before stopping and others cars received priority. The RapidRide G bus waited through five rounds of lights.
      “This is why all the buses are late,” the driver said.’

      1. The downtown section has BAT lanes, not bus lanes. I could see that being a problem. For what it is worth I could see buses backed up going up the street (on Spring). Same issue. It is legal for cars to share the lane with buses, as long as they turn right at the next intersection. But if they are turning right at the next intersection, it means the bus could easily back up behind it, since it is likely the car will have to wait for pedestrians.

        The other thing that can happen is that the lane backs up with cars and the bus can’t enter the intersection. I could see this being an issue at 5th. Cars can’t turn right on 5th. So no cars should be in the BAT lane approaching 5th. But cars can turn right on 4th. So I could see how cars could fill up Madison between 4th and 5th. There are two ways in which cars can get there. The first is if they are going on Madison and change lanes. The other is if they are turning right from Fifth onto Madison. These drivers are either using the BAT lane as a turn lane (and not getting out of it in time) or there are too many people making U-Turns. The latter seems unlikely. So it is basically too many cars on Madison heading south and turning right on Fourth and/or too many cars turning onto Madison and stuck trying to get into the middle lane.

        My guess is most of the problem comes from people on Madison. I could easily see this. A car uses a lot less space. Furthermore, they could use the middle lane (which may not be crowded) and then change lanes later (as they get past the intersection). This means that just when it looks like there will be a gap it fills up again, with people changing lanes.

        One remedy is for the bus drivers to switch lanes, or at least claim that middle lane (right before the intersection) to prevent that. That is hard to do, since the bus stop is right there.

        There are other things SDOT could do. For example, ban right turns from Madison to Fourth. You would have to turn left and loop around (https://maps.app.goo.gl/dMtkpnENyv28Le126). That is probably a bit much for SDOT.
        Another alternative is extend the Madison signal at 4th, but not 5th. That way 4th clears up well before cars (from Madison) are allowed there. My guess is that is what SDOT has done (if this was fixed already).

        In my opinion they should not have gone downtown this way. I think they should have run on Madison and Marion, not Madison and Spring. That way the buses would be in their own lanes the whole time (contraflow). We have that already, on Fifth (https://maps.app.goo.gl/NsYnAiLiUTL9bTCL7). Notice that there are no cars at all in that lane. They aren’t allowed to be. Various signs say “Do Not Enter” and all that. These are bus lanes, not BAT lanes. There are drawbacks to this approach. It is only one lane, which means that buses can’t pass buses. But in this case that isn’t an issue, since the buses don’t pass each other — they are supposed to be spread apart. (In contrast it wouldn’t work if there were a bunch of different routes converging onto the street.) The other issue is traffic lights. Sometimes they are timed towards one lane of traffic. Again, that isn’t an issue with this.

        But there are advantages to going the way they did. They are closer to Link. It makes for better line spacing (assuming they get rid of the 2). Of course the irony is until then, the spacing would actually have been better on Madison/Marion. Of course if the 3/4 ever gets moved to Yesler, then using Marion would actually provide better spacing. I think the strongest argument for the routing they chose is that it is closer to Link. Too bad they didn’t build a stop at Madison with the bus tunnel (as they originally planned) all those many years ago. Another (more radical) approach would be to change the streets, not the route. Pine is westbound, and this is fairly arbitrary, especially since Stewart is westbound as well. Flip Pine around so it is eastbound and then flip all the buses south of there. That would mean that Madison and Spring would be reversed (for cars) which would allow the buses to run contraflow. You wouldn’t have to move a stop for the G (but you would have to move a lot of other signs).

      2. To the RossB comment about Madison-Marion v Madison-Spring. SDOT wanted to reduce the transfer distance with Link at USS, now Symphony station. SDOT rejected the 49M alignment; their consultant argued against the ship canal crossing. Route 49M would have served Capitol Hill station and would have made up the deleted First Hill station better.

      3. Yeah, they wanted to move it closer to Link. That is a reasonable choice. They still could have made it run contraflow though. From Capitol Hill turn right on 9th, then left onto Spring. Turn left on First and then left again on Madison. The left turns would actually be a little easier on the buses (dealing with the slope is easier if you begin your turn midway through the intersection). Both left turns would require left-turn arrows that would be triggered by the bus (the only vehicles in those lanes). The stop on first would have to be in the middle of the street (which was the plan with the streetcar anyway).

        As far heading to the UW goes, I think they made the right choice. I don’t think you can justify 6 minute headways on Broadway, north of Aloha. You also have plenty of other transit on Broadway. There is the streetcar and 60. That means three fairly frequent routes and yet they all go willy-nilly, with no real coordination along the corridor. That doesn’t count the 43 and 9.

        The thing to do is get rid of the 9 and 43. Straighten out the 60 and combine it with the 49 (at least from Beacon Hill to the UW). You can’t justify 6 minute headways along that corridor, but you can justify 12 minutes. Run the streetcar opposite it. That way from Capitol Hill Station to the south end of Broadway you would have 6 minute effective headway. Trust the grid.

        Then make Broadway faster. Move the bike lanes to 12th. Add transit lanes along Broadway.

    3. I reported on the slowdowns I saw all along the route Tuesday afternoon, that raised the 10-13 minute travel time to 29 minutes. It was mostly congestion, cars turning right into a full lane of traffic, or cars illegally jutting into the bus lane after turning while waiting to get into a full lane of cars in the next lane. The middle one was at 8th, not 5th. I’m sure Packer’s article is right and Metro identified other issues, and I’m glad it has a supervisor and standby driver stationed at the eastern end now, that sounds pro-active.

      It will take some time before I’m willing to ride the G again or take other people on it for a route tour. My feeling trying to ride it Saturday and actually riding it Tuesday was, “Riding the G is an ordeal.” Not necessarily worse than the 12, 2, 3, and 4 have been for decades, but not the dramatic improvement we were expecting. I’ll wait until there’s more dramatic improvement, or at least I can be reassured of not waiting more than 10-12 minutes for it (when it’s supposed to be 6 minutes).

    4. If there was ever a place for AI, its traffic lights. Understand the vehicles at the intersection and adjust the lights accordingly, it’s almost like having a traffic officer at every light without the high labor cost. Bus approaching? Time the lights for its arrival, but AI would learn whether the bus is making a stop (and how long it takes) and let the cross traffic go while the bus is loading/unloading and give it green when it’s ready to go.

  3. Link still passing veeeery slowly into and out of UW Station northbound. Any knowledge or speculation as to what the problem that borked that train earlier this week is?

  4. Last night, on the 1 Line, Sodo Station southbound, we were delayed over 20 minutes because two people tried to open doors as we were leaving the station. It caused the train to stop. As we had started pulling out of the Station, the operator could not open the doors. They managed to somehow force the track side doors to open and they left that way. They damaged the doors and we had to wait for maintenance to fix them.

    1. looks like we’ve got our historical photo for next week’s midweek roundup!

      edit: looks like it’s “all rights reserved”, so permission to repost is limited. neat photo though!

    2. This is how we need to be building new elevated rail… fast, cheap, simple and not over engineered. No wonder West Seattle Link is blowing through $5 billion for an over engineered complex line and finding now its billions under a more likely cost.

  6. The problems with Rapid Ride G are more systemic and because of this the reliability for the G is unlikely to improve. there are many factors that could cause a bus to fall behind on its schedule. But if it falls behind at any point in the day it will be very difficult for it to ever catch up because there isn’t enough recovery time built into the schedule. With the G, Metro has implemented Advanced Service Management. There is a limited amount of space at the MLK terminal so the run cards are designed to have Operators switch buses at the terminal and keep the buses in service all day.

    “Advanced Service Management aims to test delivering high frequency bus service using headway management (buses arrive at a set frequency rather than a schedule) and fall back operations (buses continually in service while operators take breaks) to increase the reliability of public transportation and reduce bus bunching”
    https://kingcounty.gov/en/legacy/depts/transportation/metro/about/long-game/enterprise-initiatives

    Here’s an example of a run card for the G – this is the schedule for a particular coach that is switching operators throughout the day. You can see there is 11 minutes from the time the bus pulls into the terminal until it leaves again. If the bus is on time the new Operator can take over and make the adjustments to the driver compartment and leave on time.

    6:10 – 6:49
    7:00 – 7:37
    7:48 – 8:25
    8:36 – 9:13
    9:24 – 10:01
    10:12 – 10:49
    11:00 – 11:37
    11:48 – 12:25
    12:36 – 13:13
    13:24 – 14:01
    14:12 – 14:49
    15:00 – 15:37
    15:48 – 16:25
    16:36 – 17:13
    17:24 – 18:01
    18:12 – 19:12

    1. Sounds to me like “Advanced Service Management” is neither advanced, nor service, nor management.

    2. Based on Packer’s article the problem isn’t the number of buses or the space for them, but the number of drivers. This confirms what I could see on One Bus Away (on Saturday). You could see a bunch of buses heading northbound to the base (at Madison Park). Southbound, you could tell it had been a while since a bus had left base. But then even after the buses got to the base it took a long time before another bus left. Clearly the buses were just sitting there. The problem is, the drivers need a break. It is part of their contract.

      So, as Packer explained, they have an extra driver at the base, in case this happens. The buses will leave the base more consistently.

      As far as ASM is concerned, it would actually help with this. ASM is designed to prevent bunching. What it means though is that the buses could be very slow, especially if there are multiple delays. So instead of catching up to the bus ahead of you (when they get delayed) you hold back and wait too. It is essentially the same idea with a bus delaying half way through the route. You set an expected time to reach a particular point. The time is basically a worse case scenario. If the bus is running faster than that, it waits there. You end up with slower buses, but better overall frequency (less bunching). With ASM is it similar but a lot more dynamic. But if this happens a lot, then a lot of buses get delayed.

      This brings up another issue. It may be that when the dust settles it simply takes longer than expected to complete the route. They need to build in a lot more float (or it just takes longer). That would mean you need more buses and more drivers unless they simply ran less often. That is basically the worse case scenario. My guess is they have enough buses, it just depends on how many drivers they want to use for the route.

      I also think it is quite likely that they fix these problems bit by bit and it gets a lot closer to what they promised in the first place. As Packer wrote, there are three categories of problems (or four if you count the lack of drivers). One is mechanical. Another is driver training. I expect both of these to go away fairly soon. The third is signal timing. These will be mitigated, but it may take weeks.

      It looks to me like this was basically rushed into service. SDOT was still finishing up the street days before launch, which means that Metro didn’t have much time to train the drivers. The agencies were overly optimistic, given the new equipment and what is essentially a new route. Buses have served Madison/Spring for a long time, but not in this way.

      1. “As far as ASM is concerned, it would actually help with this. ASM is designed to prevent bunching.”

        I was suggesting the same thing before I heard about ASM. When RapidRide launched Metro tried to make it headway-managed during 15-minute periods. There was a huge backlash saying headway management is only appropriate when headways are 10 minutes are less.

        With scheduled routes, if there’s a bottleneck and buses bunch, when it ends, all buses go at once to try to get back to their first scheduled run they can. With headway management, when the bottleneck ends, only one bus goes and the rest are held back until another headway interval. This is fine if the interval is 10 minutes or less, but when it’s longer than that, the impact on passengers who have waited through the gap is, the first bus may be overcrowded and they can’t get on, or if they just miss the first bus they have to wait a whole interval for the second bus. Whereas if the buses just all came together, they could immediately get on the second bus. And it messes up transfers to other 15- or 30-minute routes if you never know when the RapidRide bus will come. So Metro relented and published full schedules for the RapidRide routes.

        However, it’s different with the G situation. The G is nominally 6 minutes. If there’s a 10-25 minute gap, it doesn’t help passengers to have four buses coming a minute apart. Passengers still have to wait 10-25 minutes between pulses; the second bus can pick up any overcrowding; and the third and fourth buses will be practically empty. So send just one bus, or maybe two if crowding is likely, and hold back the third and fourth buses until 6 and 12 minutes have passed.

      2. @ross you seem to think that the Operators are waiting in a queue and once they finish their scheduled break they take the next bus from the terminal. Is this actually how the ASM works? I was under the impression that the based on the run cards they are assigned to relieve a specific run so its not a random shuffle of buses and Operators, the Operator is taking over a specific bus based on their run card.

      3. Long bus routes certainly have issues with delays but short transit routes I think are worse for delays, and plague lines like the SLU Streetcar and now the G. They don’t leave their terminal on time (often recovering from a delay of their previous run), so they aren’t on schedule, nor are they able to be tracked with real time arrival since they haven’t departed yet. You end up waiting for it to come/leave and have no clue of its departure. At least longer routes can be tracked with real time arrival as the buses are underway (so long as you aren’t near the very begining of the line).

      4. You seem to think that the Operators are waiting in a queue and once they finish their scheduled break they take the next bus from the terminal. Is this actually how the ASM works?

        It has nothing to do with ASM.

        I was under the impression that the based on the run cards they are assigned to relieve a specific run

        Of course. If things are running well. When they aren’t, you have the problem that was happening. Go back to that previous comment: https://seattletransitblog.com/2024/09/20/first-week-part-3/#comment-941467. Look at the table. So the driver is supposed to drive the bus from 6:10 to 6:49, then take a ten minute break and start driving again at 7:00. But what happens if they arrive back at the base at 7:02? They need a break. So they wait.

        But as Ryan Parker explained, what if there is another driver there, waiting to take over the duties of the 7:00 driver (or any driver)? Great! Now the 7:00 leaves as planned. Now the driver gets at least a six minute break. That is the basic idea. It is a little more dynamic than just allowing more float, but it is the same idea.

        the first bus may be overcrowded and they can’t get on, or if they just miss the first bus they have to wait a whole interval for the second bus.

        For what it is worth, ASM is designed to prevent that problem. The main reason the first bus is so overcrowded is because it is running so much slower than the bus in front of it. But if the bus in front is slowed down then it doesn’t get that crowded. It is just another bus. It is as if all the buses are connected together with a cord so that are all the same distance apart. Except instead of distance it is time. Of course there are potential problems that can make the thing fall apart. There could be a surge of riders on a bus for some reason (e. g. the ferry landed). Now that first bus is crowded even though it is running properly. Another is more fundamental: Do you want every bus to slow down just because one bus is running slow? The system is designed with wiggle room — it isn’t meant to be rigid, necessarily.

        It is worth noting that trains typically work the same way. You don’t want a train catching up to another one. Quite often you can’t (for safety reasons). But rather than wait until the trail is clearly too close, it slows down a bit at each station. You don’t even notice it. For automated systems (like in Vancouver) this is how they operate, but even with manually operated systems they do the same thing.

      5. Long bus routes certainly have issues with delays but short transit routes I think are worse for delays, and plague lines like the SLU Streetcar and now the G.

        It is more about how much time you want to allocate for waiting and float. There is probably a correlation, but that has more to do with delays that are more common for shorter routes. In the case of the G they were just overly optimistic. In theory there is plenty of time to complete the route, rest and go back out again. But now they know better. But the same thing is true of buses like the E. If they don’t allow enough time to rest, then the drivers can’t turn around in time when there is a delay.

        It may be that they “stretch the envelope” with some routes more than others. Infrequent routes might have a long wait time, just because there is no point in leaving early.

  7. Rode the Rapidride G downtown and back yesterday evening for an event, and it was great – buses came on time and weren’t bunching (according to Transit app, at least). We are halfway between Husky Stadium light rail stop and the G, so the G gives an alternative to get downtown.

    Now, if the 48 didn’t drop to every 30-minute frequency after 9:20pm, I’d be living the dream, but a Lime bike was an acceptable substitute for me.

    1. The 48 is a core route. It is part of our Night Owl service. It shouldn’t drop to every 30 minutes so soon. Maybe after midnight, but not after 9:00 pm.

    2. I’m hopeful that Route 48 gets more riders when Judkins Park Station opens on the 2 Line. It will be possible to get from Pioneer Square Station to the Route 48 bus stop on 23rd at the station entrance in about 6 minutes, for example.

      If Route 48 is run at a high frequency, it could be an attractive option to get to Downtown Seattle from areas as far north as Union Street. But to have that draw the frequency needs to be higher.

      It’s a classic transit dilemma. What comes first? more frequent buses or more riders?

      1. When the 48 and 45 were split, it was assumed the 48 would have the most riders. It was positioned as a future RapidRide. Seattle’s transit levy raised the 48’s and 45’s frequency to 10 minutes daytime, 15 minutes evenings. But when people voted with their feet, it was the 45 that got the most riders while the 49 lagged behind. I think that’s because the 45 was chosen to be the University Way route, and was interlined with the 75 so that it was a one-seat ride for all the overlapping trips between U Village and 65th & Roosevelt. After that lagging performance of the 48, it got less priority. If it’s at 30-minute evenings now, that’s due to the driver shortage or Seattle lowering its levy rate.

    3. How do you find the 48/G transfer at 23rd? I’ve been concerned about the RapidRide stations being a block away and up/down a hill from 23rd. (The 48/11 transfers were also not that close.)

      2. I think the 48/G transfers are the biggest missed opportunity. Everything about it is wrong. Just to back up there, there are multiple potential transfers, but some are less likely than others. Madison runs at roughly a 45 degree angle compared to 23rd. The G doesn’t go to Madison Park (there may be issues with the 48/11, but I’ll ignore that for now). That basically leaves two common transfers:

        1) Southbound 48 to westbound* G: The stop for the 48 (on 23rd) is just south of John. If the stop for the G was just west of 23rd it would be fine. It isn’t. It is on the other side of the street, so riders have to cross 23rd: https://maps.app.goo.gl/G6mNdhBn9kgSMMS37.

        2) Eastbound G to northbound 48: In this case the closest stop is way back by 22nd. To make matters worse, none of the stops along 23rd are close to Madison. The best option for riders is to walk south on 22nd, then across on Olive (https://maps.app.goo.gl/uYorz9YYbA18gn8CA).

        I’m not sure why it is so bad. My guess is one thing led to another. There is a stop close to 17th, so they wanted to put another stop somewhere around Denny (for good stop spacing). But they didn’t want to put the stop on the right side of the street, and block people turning right (I guess). Or they simply didn’t want to add another stop so close to that one, next to 23rd. Fair enough, but why isn’t the stop on 20th (instead of Denny)? There may be other considerations. A stop on 23rd would block those making a hairpin turn from westbound Madison to southbound 23rd. Tough luck. Same goes the other direction. Maybe they were concerned about cars being unable to turn onto Madison. Again, tough luck.

        In short, I would do the following:

        1) Move the westbound G stop closer to 23rd.
        2) Add a stop for the 48 just south of 23rd (just south of the hairpin turn).
        3) Move the eastbound G stop on the other side of 23rd (next to Aegis Living).
        4) Move the Denny stops to 20th (for better stop spacing).

        * When I write “westbound” and “eastbound” for the G it is basically inbound (towards downtown) and outbound (away from downtown). Because the street runs at an angle travel is southwest and northeast.

      3. The RRG – FHSC transfers are pretty terrible too.

        The big problem there is the streetcar stops. The streetcar doesn’t stop close to Madison. It stops south of Marion. The next stop (to the north) is all the way at Pike/Pine. This means the quickest way to transfer is to walk along Boylston. They did about as well they could, really (https://maps.app.goo.gl/FH54zRGETanboAa46). The stop is between Summit and Boylston (it is a bidirectional center-stop that takes up the whole street). Ideally you would try and squeeze it into the gap between Madison and Broadway, but they probably felt like there isn’t enough room there (especially with everything else going on at that intersection). Ultimately it wouldn’t be much closer — it might even be farther (https://maps.app.goo.gl/FH54zRGETanboAa46). It would involve a bit more of a slope as well.

        Ideally they would move the streetcar stop close to Broadway and squeeze in stop for the G somewhere in there. I could maybe see something working better, but it wouldn’t be trivial.

      4. “There is a stop close to 17th, so they wanted to put another stop somewhere around Denny”

        The stop should be at the TRANSFER! More people are transferring north-south to east-west than are going to Denny Way.

      5. “There is a stop close to 17th, so they wanted to put another stop somewhere around Denny”

        The stop should be at the TRANSFER! More people are transferring north-south to east-west than are going to Denny Way.

        I agree. There is nothing fundamentally wrong with the stop at Denny. But there would be nothing fundamentally wrong with a stop at 20th either. The difference is that the latter would allow you to put the stops on 23rd in an ideal location while still retaining go stop spacing.

        But as I noted, it may be that they didn’t put the stop on that side of the street for some other reason (such as traffic concerns). Maybe they didn’t want to block the right turn onto Madison (from southbound 23rd). I also don’t know why there is no eastbound stop for the G on the other side of 23rd. The next stop is instead close to John. That does allow a same direction transfer with the 11 (which is good) but there is another stop at 27th that does the same thing.

        The stop for the eastbound G should be just east of 23rd and the stop for the northbound 48 should be just south of Madison, so that riders can go north (e. g. downtown to Montlake) without crossing the street (https://maps.app.goo.gl/QNJzyQCUTPM7wK1D7). (Technically you have to cross Denny, but that is easy, since there is no traffic light. https://maps.app.goo.gl/UMbT3E2WSPG13UDA6).

      6. “The RRG – FHSC transfers are pretty terrible too.

        The big problem there is the streetcar stops. ”

        Note that both of these projects were led by SDOT. So blaming FHSC is still blaming SDOT.

      7. I am not blaming anyone. I’m saying the streetcar stop is nowhere near Madison. If you put the RapidRide G stop close to Broadway it actually makes it worse. You would have to move the streetcar stop, which is a lot more difficult than moving the bus stop (as I explained).

        The same problem exists with the 48 and G. I’m not blaming Metro for the bus stop location. That isn’t the point. You would have to move the bus stop for the 48. But moving a bus stop is a lot easier than moving the streetcar stop. You don’t have to deal with the rail.

      8. The SDOT designs for transfers at both 23rd avenue and with the streetcar have long walks. It is their pattern; see their future long walks with north end of J line and with route 40 in Fremont.

      9. see their future long walks with north end of J line and with route 40 in Fremont.

        Moving the streetcar stop would be quite expensive. The 40 was because they wanted more bike lanes. The north end of the J line is temporary — eventually the bus will (and should) go up Roosevelt past 65th.

        In contrast, there really is no reason why Metro/SDOT messed up the transfer at 23rd. It would mean moving the bus stops for both the G and the 48, but they should have done that.

  8. The G line is serving a difficult market. There are frequent signalized intersections. There is no chance for buses to pass buses, so they tend to bunch. I was at the terminal Tuesday afternoon and we saw the supervisor allow three coaches to leave atop one another; they began their inbound trips bunched. We alighted at 12th avenue to snoop around; there was a 30 minute gap in service. The inbound bus stop on Madison street far side MLK Jr. Way is long enough for one bus; it will be jammed by routes 8 and 11 and the G line. ( my suggestion is to consolidate routes 8 and 48 and restore route 43). The Metro execution of the G line will improve; the strategic errors of SDOT will be permanent. It appears that the inbound route 2 will be stuck in congestion east of 12th avenue.

    1. The G line wasn’t designed for any rider to have an easy time using it. It’s more about vanity, photo ops and interest groups.

      The mere fact that bus bunching is immediately a recurring problem when so much of the route has exclusive bus lanes — and especially exclusive bus lanes in the median to not conflict with right turns yet the operation is a problem — is truly a testament to how badly executed the operation is.

      Please let’s get SDOT out of the bus technology business once and for all! That’s two operational fiascos — RRG and FHSC — in the last decade.

      1. The G line wasn’t designed for any rider to have an easy time using it. It’s more about vanity, photo ops and interest groups.

        Oh, bullshit. It was intended to be the first (and only) Bus Rapid Transit in the state. Center running is a good thing. It means the buses can run faster and more consistently. If they run faster and more consistently they can run more frequently (at no additional cost and without bunching).

        There are definite weaknesses with the implementation but the routing is solid. The only major problem with the routing is that Metro hasn’t adjusted their routes to it. Just look at this map: https://kingcounty.gov/en/dept/metro/programs-and-projects/madison-street#gallery. Show anyone in the world what this city looks like and then tell them that the G route runs in its own lane much of the way, every six minutes. There will be a collective “Makes sense”. It is a major diagonal street — the only street that operates that way.

        But they will immediately question every other route. Why is the 2 running so close to the G? Is it running as frequently? (No, not even close.) Why are the 10/12 branching there (or branching at all)? How come the 60 doesn’t run on Broadway? Why does the 49 go downtown given it runs right by Link and crosses a series of frequent buses that go downtown? Why do the 9 and 43 even exist? Even the 3/4 (which is better than most) makes an inexplicable little dogleg in the middle of the route, only to work itself back to the main corridor at the end. It could just stay on James/Cherry the whole way but it doesn’t.

        You can blame SDOT all you want for not providing enough right-of-way and signal priority to make the buses faster, but the G is one of the best routes in the area. It literally goes straight from the moment it leaves downtown to when it finishes! If this is indicative of SDOT as a route planner, well done. The problem is SDOT hasn’t done the part that is really their responsibility: making the buses run really fast.

      2. “Oh, bullshit.”

        My comment wasn’t about providing a BRT.

        My comment was about the ridiculous decision to put in median stops that require left-hand doors under the theory that it’s better than a curb bus lane (supposedly slightly faster and laughably with more reliable frequency) and that it’s better to have articulated buses that aren’t on trolley wire — driven by a promise to allow bicycles onboard with no consideration of an 18 percent slope. That’s the vanity part.

      4. The center lanes are to avoid cars turning right. That’s already a proven problem in the side-lane areas. Left-side doors could be avoided if the buses ran contaflow as in the Bellevue Transit Center. You can make the center stand out more visually so that it looks like two ordinary streets rather than one four-direction street. Then pedestrians’ and drivers’ instincts would be correct.

        SDOT didn’t want contraflow, so we ended up with island stops with left-side doors, separate right-side stops for parallel routes, and no way to put ordinary trolleybuses on the G without losing the anti-congestion advantage of the center lanes and island stops.

      5. It’s going to be interesting to see how the G Line route handles very mild winter conditions. Not even a layer of snow, but what happens when there is a little bit of slush on Spring street between 1st and 2nd, for example. Or there’s a few splotches of ice on Madison east of 23rd. Something tells me the entire route will come to a standstill if even one or two blocks are a tiny bit slippery. Most other routes have more options under similar circumstances.

      6. “The left side door was to allow the bus to be on the left and avoid right turning traffic. I’d have to look through the documents again but I remember for instance Madison and at Boren there would be a high number of right turning traffic.”

        SDOT has no problems banning turning movements at major intersections. For example, Columbian Way at Beacon Ave has a relatively new right turn only lane separated from through traffic with left turn prohibitions. It operates as a different phase from through traffic. Southbound Westlake has a right turn prohibition at Mercer so the streetcars can move through that busy intersection.

        Or SDOT could have installed right door median islands and staggered the stop directions by an block. Eastbound cars are to the right of the median island for the eastbound FHSC at Occidental.

        There are plenty of design options available before summarily requiring a fleet of bespoke articulated buses. Design is always a compromise so nothing is perfect. But the whole thing could have been done differently without needing special buses. It’s just some people thought it was a novel idea.

        Let’s not forget too that a grid system depends on good transfers, and all of the major crossing bus and streetcar routes to RRG are not easy transfers.

        And it’s worth mentioning that SDOT didn’t give a crap about vehicle conflicts and delays on First Hill when laying out the FHSC.

    2. I’ve always felt that Route 2 and Seneca was where RapidRide G belonged. I would have left the occasional ETB on Madison (with its steepness) and 19th alone and otherwise let the street be for cars. Seneca across First Hill could have worked so much better and been faster!

      But it’s done and we now have to eat this questionable service operationally and financially for several years.

      1. SDOT was told that three lines had overhead directly connected with the Madison Street overhead: routes 2, 12, and 49. (I favored Route 49M). Your suggestion of Route 2M is solid. During the reductions period, 2012-14, folks campaigned to preserve Route 2 as is. The irony will be clear when SDOT causes the inbound trips to slow after they improved the outbound trips west of I-5. SDOT really wanted articulated buses to load bikes on board. Otherwise, they could have had standard ETB that would have climbed the Madison Street grades. See: https://www.seattletimes.com/seattle-news/transportation/metro-and-seattle-promised-electric-trolleybuses-on-a-new-madison-rapidride-turns-out-that-wont-work/
        Transit agreed to all the SDOT choices. Sad.

      2. The 49M (which I assume was the Broadway-Madison concept) seems to replicate the problem with the 49, and for Madison doesn’t go far enough east. So you want to continue south on Broadway but the 49 turns west. And now instead of the retail district, you end up in an office ghetto and then have to go north-south to wherever you really want to go. The 12 turned north on 19th, making it useless for going to 23rd and transferring to the 8 or going further to Madison Valley, the Arboretum, or Madison Park. The 49 would double down on this because it wouldn’t serve 12th, 15th, or 17th either. With the 49M, would the 11 and 12 both have to continue to serve the rest of Madison, so you’d end up with three routes serving different parts of Madison?

      3. Seneca basically ends at Capitol Hill and then you have to do the weird little zig-zag where Union intersects Madison. It would likely be slower.

        It also makes for poor stop spacing. You are practically overlapping with Pike/Pine buses. The 3/4 runs on James, which means there is a big gap there. The gap would be even worse if they move the 3/4 to Yesler (and I think they should). No, it makes way more sense to just run straight on Madison. It is a straight shot from the Central Area to downtown. The route they took is fine — they should just run it as contraflow (to avoid traffic downtown). That would mean entirely exclusive lanes west of 15th (where the buses transition from center-running to curbside running).

        During the reductions period, 2012-14, folks campaigned to preserve Route 2 as is.

        Right, which is one of the big problems. The line spacing is terrible downtown. Buses are way too close to each other. It runs against official Metro policy and just common-sense routing. We need to consolidate on various corridors. The 2 should run on Pike/Pine.

        It probably played a part in the lack of contraflow. When they were discussing the advantages and disadvantages of routing proposals, adding BAT lanes on Spring was seen as beneficial since it would make the 2 faster. The problem is that if you added contraflow lanes for the G on Spring, the only way for the 2 to use them is to also add contraflow lanes on Seneca. This greatly expands the scope of the project. So instead they went with BAT lanes.

        The problem is, BAT lanes are second-rate. They are definitely better than nothing. There are areas of the city (e. g. Denny) where BAT lanes would be a godsend. But in a very urban area they are problematic. The only cars allowed in them (legally) are cars that are turning at the next block. This is terrible, as it means that they are delayed by pedestrians (and there are a lot of pedestrians downtown). It is also much harder to enforce. To even be in a contraflow lane means you ignored a “Do Not Enter” sign. Sure, it happens, but only the most brazen driver does it on purpose. In contrast it is quite common for drivers to use a BAT lane before they should. This happens all the time in Ballard. They are approaching Market on 15th (having just crossed the bridge). They enter the right lane several blocks before turning on Market. This just doesn’t happen with contraflow lanes.

        This is why they have center running buses in the first place! It is kind of crazy when you think about it. On the one hand they spend all this time, money and effort building bus platforms and buses with doors on both sides so they can run in the middle of the street. Great! But then for downtown they put in BAT lanes, instead of contraflow.

        The thing is, they could still fix it, but it would require a lot more work. You would have to move all the curbside bus stops downtown. So that is six stops. The five on Madison and Spring are straightforward — just move them to the other side of the street. The stop on First would be in the middle of the street (which was the original plan). You would need to change the traffic signals, but that shouldn’t be too hard. Otherwise it looks pretty good. This slide shows it pretty well. If a bus was going outbound (towards Capitol Hill) on Madison is in its own (right) lane it would simply go right into the middle lane at sixth. That is actually a much better transition than exists now. Going the other way, the bus would turn right off of Madison on 8th or 9th. Then it would turn left onto Spring (in its own lane) and go down to First. You would have to move the bike lanes.

        This wouldn’t be trivial. The most expensive and disruptive part would be adding concrete (assuming the slides are correct and it is just asphalt). But you would end up with a much faster bus route. It is easy to chide SDOT for not doing this in the first place, but it is better to fix a problem then pretend it doesn’t exist. I’m not saying this is high priority — there are a bunch of things they should do first. Maybe this turns out OK, as people avoid these streets (over time). But it just doesn’t make sense to have the buses run exclusively in their own lanes (with all the advantages and disadvantages of that) and then run in BAT lanes downtown.

      4. The 49M (which I assume was the Broadway-Madison concept) seems to replicate the problem with the 49, and for Madison doesn’t go far enough east.

        Exactly. The biggest problem with our transit system — by far — is that the buses don’t run frequent enough. One of the big reasons for this is waste. A 49M would make it worse. Not only is it a service disconnect (do we really want buses running every six minutes from Capitol Hill to the UW?) but it also creates additional, useless overlap. I write “useless” because overlap can create a “spine” — a section where multiple buses combine to provide really good service along a corridor. In this case they don’t. Just consider how you would serve Madison between Broadway and 23rd:

        1) You don’t. You just depend on the buses that run north/south/east/west. The problem is, you’ve introduced awkward three-seat rides for short, common trips.

        2) You just run a bus like the 11 or 12 on Madison. That works, but now you are overlapping downtown (and thus creating waste).

        3) A bus like the 43 follows Madison from the east and then heads towards Jackson. OK, now you’ve created waste on Broadway. You’ve got the streetcar, the 60 and now this new bus.

        It just doesn’t work. In contrast, the G creates the possibility of a vastly more efficient network by creating three major corridors going to the Central Area. Madison (the G), James (3/4) and Pine/Pine (every other bus). The first two are central corridors — the buses don’t branch along it. But with Pike/Pine, there are several branches, everything from Union (the 2), Summit (the 3) and 15th (the 10). Thus Pike/Pine is a “spine”, and the other two are major transit corridor with effectively only one bus.

        Meanwhile, you have two major north-south corridors: Broadway and 23rd. Broadway is served by the combination of the streetcar and a bus going from the UW to Beacon Hill. Match the frequencies and you provide good combined service along Broadway from CHS to Jackson. It worse. You’ve made the system a lot more efficient which means that effective headways are much better and you don’t have to wait so long for a bus.

      5. “Not only is it a service disconnect (do we really want buses running every six minutes from Capitol Hill to the UW?)”

        Better too frequent than not frequent enough. I’d like to see 5-minute service on all core routes, which includes whatever runs between Capitol Hill and the U-District. The problem is the zoning in between, not that buses are too frequent. If you take one problem (can’t walk to retail) and then add another problem (have to wait longer for a bus), that makes things worse, not better.

Comments are closed.