How should we value an individual transit station?  Proximity to jobs and housing is of course important, but even important is how the addition of that station increases the value of the network as a whole.

When University Link opens in 2016, it will be great for people who live and work in Capitol Hill and the UW.  At the same time, neighborhoods along Central Link (Downtown, SoDo, Beacon Hill, Rainier Valley), will instantly have access to two significant new destinations.  Property values surrounding Central Link stations will likely increase substantially, even though actually nothing’s “changed” in those neighborhoods.  It’s as if you woke up one day and Comcast had doubled the speed of your internet connection.

This is roughly the principle behind Metcalfe’s Law.  Metcalfe proposed that the value of a network was roughly proportional to the square of the number of nodes. If five people have telephones, then the network has a certain value. If 10 people have phones, the value of the network quadruples, rather than doubles.  It’s why Facebook and Twitter — with their hundreds of millions of users — are  so valuable.*

To take the analogy back to transit; if you have a station near your neighborhood, then the value of that station increases exponentially as stations are added elsewhere throughout the city.  When light rail goes from the initial 13 stops to the planned 33, the value of the network will be roughly 644% greater than it is now due to all the new origin-destination pairs that will open up.

Of course, Metcalfe’s law assumes the value of each station is equal.  In reality, that’s rarely true.  On social networks, some users contribute more than others.  The same will be true of Link stations.  A freeway station in North Seattle may be attractive to the people who live nearby, but probably not a frequent destination for people who live in, say, Overlake.  By contrast, a station right in the heart of Capitol Hill or Bellevue or the Airport has the potential to be valuable to everyone on the network.

This is why it’s so important for every station to maximize the development opportunity around it, and why even people who don’t live in the neighborhood ought to be able to weigh in on the development of the area around a planned station.  A botched walkshed doesn’t just affect the neighborhood that surrounds it, rather it degrades — substantially, quadratically — the value of the network as a whole.

* Some have challenged Metcalfe’s equation with respect to very large networks, but the man himself is sticking with the law for smaller numbers like our transit example.

74 Replies to “Metcalfe’s Law in Transit Planning”

  1. There should be two geometric terms to the usefulness equation. A well placed stop that is one hour or one minute from another point in the network is not as useful as one 15 minutes away. Discounting overlapping walk sheds, the marginal usefulness of a transit stop will go as:
    Sum over all stops(exp(-{total_travel_time}/{1 hour})
    and the total network value goes as
    n^2exp(-{mean trip travel time}/1 hour)

    Metcalf’s law is a powerful effect, but it’s not quite n^2 in a large transit network.

    1. I’m not sure I agree that’s an improvement. Now you’re just discounting the value of a transit stop based on its distance to the other transit stops. In other words, time between stop is now the only value. I don’t think that’s true at all. Just because it takes longer to get to Johannesburg, South Africa than it takes to get to Hoonah, Alaska, doesn’t mean that regular flights to Hoonah are more valuable to than regular flights to Johannesburg. Or, when I lived in the Bronx, just because the East Tremont Ave stop was closer to me than the Central Park stop, didn’t make the East Tremont stop more useful.

      1. @Jeffrey

        You are correct, the value of the location a stop serves is more important than the time required to access that stop.
        Metcaffe’s law assumes that each new node is equally useful to all existing nodes in the network when it is more useful for those that are within a reasonable travel distance. Both this law and my previous comment discount the quality of the access provided by a stop and just assumes more nodes are better.

        My formula earlier for the aggregate value of the network is incorrect. At the network gets larger, the value of an added stop should go assymtotically proportional to n.

  2. For the purpose of this equation, should the streetcars be counted as part of the network or not? How about connecting buses?

    How does the speed and frequency of each segment figure in?

    There seem to be more relevant variables here than just the mere number and quality of nodes…

  3. It is disappointing how long it takes to build transit infrastructure. WS-DOT builds highways and tunnels much much faster than Sound Transit. The tunnel to UW/Husky is finished, the station boxes are basically finished. Why does it take 3 more years to get it operating? Open it sooner!

    1. The tunnels and boxes are done, but not the stations themselves. The track is not yet complete either.

      Capitol hill also is going to have a west entrance on the other end of Broadway that they have not yet dug the tunnel for (I assume cut and cover *through* Broadway):

      http://www.soundtransit.org/Documents/pdf/projects/link/north/ULink/CHStationSitePlan(0).pdf

      I agree I would love to see these things done more quickly, but it seems that these underground station just take many years to build.

      In comparison, the Angle Lake station construction began recently and will be done in the same year, 2016.

      Its something to think about I suppose when advocating for underground stations. In the right places they make a lot of sense, but they do take a lot longer to build than the surface ones.

      1. And you can always build faster, but it takes a lot more money. Money ST is short of due to the recession.

  4. Sorry, but the law is not transportable to Link stations. You’re comparing apples and wheelbarrows.
    Social networks on the internet start with the premise that all the users have roughly equal access via broadband, although they choose to participate at various levels of activity.
    Transit stations or stops vary by night and day, or desert and highrise if you choose. Continueing E.Link past Redmond North Bend, with stops every mile would double your station count and do virtually nothing for increasing the network effect and ridership.
    And how would you measure the 644% increase anyway? The logical answer would be in ridership, but you don’t mention that. Property values throughout the network? Maybe, but what’s the use of having a vastly better network, if nobody is using it?
    And the network HAS to consider all the journeys being taken, and by what mode. Isn’t that the whole point of building mass transit? So that the masses will use transit instead of the paltry 5% of all trips taken in the region. We’ve made some progress over the last 2 decades, but at a snails pace and kings ransom to do it.

    1. I’m not sure how you’d measure the percentage. I use the number just because I think it’s arresting to stop and think about how much value is being created.

      Social networks aren’t the greatest example of the law, ad Metcalfe himself states. He used telephones and fax machines as examples, which I think are more relevant.

      1. In you favor, applying the 644% to Central/Airport Links 2020 estimate of 47,000 daily riders yields 302,998 riders when ST2 is built out. That’s darn close to the 296,000 estimate for all the lines in 2030, so maybe the damn formula works.
        OTOH, PSRC pegs the ridership much lower, so who knows.
        More reason to stick around to 2030 to see how all this plays out!

    2. The law is an abstraction. Even in telephone networks, some nodes are more valuable than others and get used more than others. What the abstraction represents is the ability to get from point A to point B. E.g., I don’t ride bus 48 every day, but the fact that it runs every 15-30 minutes means I can use it whenever I have reason to. That makes me more willing to downsize my cars because I know the 48 is there. The characteristics of the transportation itself may also contribute more to the value than the particular station does. For instance, if the network is frequent and grade-separated, then station B is more valuable than if the line weren’t so.

  5. So this inviolable law says that value = nodes ^ 2, except you then fall on yourself why only super special perfect places in Seattle can have these stations. Obviously you can’t allow this law to exist or a station on Kent East Hill wold add exponentially to the value.

    1. John,

      A station on Kent East Hill would add a lot of value for you and other Kent East Hillers, but little to others. There’s no “there” there so it would be LRT used as commuter rail, a costly and poor investment.

      I would be like mic’s mooted extension to North Bend. Well, a little better, but not much.

      1. The “rule of squares” part of the law works, but it isn’t about the nodes, it is about the number of people who you add to the system. Kent East Hill just won’t add that many. I go into what a better analogy is down below. Whether by pure people standards, or weighted by destination, Kent East Hill is just not that good of a spot. Weighted by class, it gains a bit of benefit (as it brings in more diversity). It might also be fairly cheap to add the stop, which is always a factor.

      2. But that’s not the point.

        Value can be measured in usability…it’s like adding movies to Netflix. Even if only a few people watch the movie, the knowledge that the library is comprehensive increases its value.

        Or think, for example, of the old days of the NYC subways where there were three private lines. Each station they added made that line seem more universal. Even if the stations were redundant.

        So, the knowledge that if I walk into a LINK station I can get to anywhere is the value, and hence this law applies even to Kent East Hill.

    2. That’s where the suburb’s responsibility comes in. It can make a super perfect place around the station. A super perfect place means it has several destinations within walking distance, both commercial and residential. And it fosters one-of-a-kind companies that attract people from outside the neighborhood.

      For instance, Roosevelt station has a Whole Foods a few blocks away, and a row of specialty audio-equipment shops. Other Whole Foods aren’t near rapid-transit stations, and no other concentration of audio-equipment shops exists. So those are two things people will go to Roosevelt Station for, and that makes the station more valuable beyond just the immediate residents. Roosevelt station also has Greenlake, and bus transfers in four directions, and the high school. And it will have lots of apartments. You may not like apartments, but it means more people taking transit.

      Capitol Hill station has several hundred things you can walk to. Beacon Hill station has a library branch. Etc. What will Kent East Hill station have?

      1. But within what walking distance of a train station? Or how much money would you need to spend to build a parking garage for people outside that radius?

    3. Auburn Station works well as a Sounder stop; but, according to WSDOT research, it wouldn’t work well as a stop for AmtrakCascades because it would add very little value to the network (and might actually reduce the value of the network). On the Link extension to Lynnwood, ST is studying building stations at 130th, 145th/155th, 185th, etc. Why not build all those stations and create a super-valuable line?

      1. Of course, some of us (like me) object to that study because it only looked at having all trains stop at Auburn in addition to Tukwila, or having some trains stop at Auburn and others at Tukwila. The thing we’ve been proposing is having all trains stop at Auburn instead of Tukwila; they didn’t look at that.

      2. Tukwila is completely unreachable by bus transit in any amount of time.

        Auburn however, is supremely reachable (more so than Kent Station unfortunately) because the ST Express buses stop there.

    4. That depends on how the land around the Kent East Hill station is developed. As-is, the parking-oriented strip malls it would give access to are already duplicated all over the line.

      Might be better than the Mountlake Terrace station, though.

      1. Maybe we should offer a station to the first town willing to tear up a bunch of parking lots and turn them into 5 story+ apartment complexes?

    5. So Mr Bailo, where exactly would you want an East Hill Link station to be? What would you expect to see around it? Where would the line go to?

      1. I foresee a rail line that runs along the Kent East Hill ridge starting in Renton (which would be connected to the system by then) and travelling up Benson. Then it would go along 108th/104th until it reached Kent East Hill. From there it would turn up Kent-Kangley Road and head towards Covington and then Black Diamond.

      2. I would put the Kent East Station smack dab in the center of it all — the confluence of Canyon Drive, 104th Avenue and Kent-Kangley.

        This is a rich shopping area, and school district, with importance as a transit area, automobile area and a nascent commercial hub.

        So right here:

        http://goo.gl/maps/MHhbu

      3. So in other words, you wouldn’t change any of the buildings or anything in the station area, but just leave them as they are.

        Would this line go in the middle of the street like MLK, or would you insist on elevated?

      4. I think that the strip mall to the right…where Applebee’s is located, is tremendously under utllized. Not only would it make a fantastic transit station (even BRT) but also a good place for some TOD.

        108th and Kent-Kangley are troublesome for me. They are really highways that were never developed into highways and handle very large volumes of traffic all day long (sometimes more than I-5…believe it or not, but that’s what Google told me)!

        So, the transit in me says, Road Diet. I would put something in on one side of the street, probably the east side where there are more malls and reduce the turning lane. There is so much traffic and yet so much pedestrian and some bike activity that it all becomes really dangerous and unlivable…and a road diet would make it pleasant.

        BUT. As I said, it’s also used as a highway. And I don’t know what you would replace that highway with if you did do a road diet. I wish that they had built more highways back when they were developing the area…but that is water under the bridge.

        Maybe, as crazy as it sounds…we would want a tunnel….a Kent East Hill tunnel!

  6. Just to play Sam’s advocate: That trip from Columbia City Station to the future Capitol Hill Station takes about 23 minutes on the Metro 9. Link will improve it to … 21 minutes! Jump for joy, kids! Columbia City is going to be overrun by skyscrapers! Sign me up.

    That super-fast clear ROW to Montlake is certainly going to improve my ability to get to UW fast as well, right? Well, the 48 takes 25 minutes to get to UW from Mt Baker Transit Center. Link will do that in 22 minutes. Hey, why did you take away my one-seat ride on the 48 to UW? It is a lot more valuable of a line because of all those stops, er, “nodes” as you call them. That stop diet they did on the 48 cut the value of my bus line by 3/4 … or was it exponentially?

    1. The 48 taking 25 minutes to get from Mt. Baker Station to the UW is a joke, which is based on extremely optimistic paper schedules. Much more likely, the bus will actually take 35 minutes, not 25, and simply show up on OBA as being late.

      Also, don’t forget that the scheduled time on the 48 does not include wait time. The 48 is notoriously unreliable. Even when it is supposedly running every 15 minutes, it is not at all uncommon to have to sit and wait 20 minutes or more for the bus to show up (even if you are very close to the terminal, it doesn’t matter – you still have to wait for the driver to finish his coffee break after his previous run pulled into Mt. Baker 20 minutes late)

      The figure of 21 minutes with Link should be much more reliable than this, and the headways and span of service are significantly better too. Also, when buses eventually get kicked out of the tunnel, hopefully, the time will be faster too.

      1. ST’s long-range plan recommends 24-hour service, but that’s a long way off and there’s no timeline for it. The issue of track maintenance would have to be worked out.

        The relevant night owls are of course the 83 and 7, not the 48. If Metro reorganizes the night owls may change, but I doubt the 48 will be a first-phase candidate. It’s also likely that the night owls will be eliminated entirely in next year’s cut, so it would depend on whether they are reinstated later.

      2. My wife takes the 48 to UW every day. It is not uncommon for it to take her 50 minutes to get home.

        The breaks that drivers take after they are 10-20 minutes late are also infuriating, and they are often quite surly about telling people who count on public transit that they will now be late for work because of a bus driver’s coffee break.

      3. Breaks are a legal requirement, of course. You want to eliminate the driver’s break so he’s more tired and surly and less safe driving? The lateness is not caused by the driver’s break but by Montlake/520 congestion, cash fumblers, wheelchairs, and the number of stops (so that two people get out at adjacent stops rather than the same stop, and at peak hours the bus ends up stopping at every single stop).

      4. It should also be noted that when the bus is sitting at the terminal, OBA has no information on what’s actually going on. If you’re getting on half a mile later, you literally have no idea when it’s going to show up.

        In the past, I had experimented with making Issaquah->U-district trips by getting off the 554 at Ranier and taking the 48, rather than going through downtown. The theory was that the 48 runs every 15 minutes on Saturday, and that avoiding downtown would compensate for the 48’s slow travel speeds.

        The first couple of times, it went ok, but the next time I tried it, OBA said the bus was 5 minutes away when I got off the 554. 3 minutes later, when I reached the bus stop, it was still 5 minutes away. And, after 20 more minutes of standing at the bus stop, it was still 5 minutes away according to OBA. At that point, I gave up and walked back to Ranier Freeway Station to take the 550 into downtown and find another way home. I don’t know how long it was before the 48 finally went by. Needless to say, since then, I never attempted to take the 48 from the Mt. Baker area ever again.

        Experiences like this is why no sane person would choose the 48 over Link just because it drops you off slightly closer in the U-district.

      5. The breaks are find. It’s just impossible to know if the bus that OBA says is on its way will contain a break-needing driver, thus adding to delays on the 48.

    2. That stop diet they did on the 48 cut the value of my bus line by 3/4 … or was it exponentially

      Overlapping walksheds don’t increase value, or did you miss that part.

      Also, the 48 is overcrowded and has an on-time rating somewhere around 10%.

    3. Three points:

      I think that Frank is generally right in that the benefit effect is around all stations, and not just individual station pairs. For example, Beacon Hill to UW will be a direct and desirable trip using Link, and will likely reap the property value benefits more than Mt Baker will.

      I also think that rail offers a level of ride comfort that buses don’t have. I’ve known many people who tell me that they will ride a train but not a bus because they get motion sickness on a bus but not a train. Not everyone has the same value of time.

      Finally, reliability is just as important as travel time. If two routes are expected take about the same amount of time but one is prone to unanticipated delays on a routine basis, I suspect that people will gravitate to the other one. In choosing routes, I’ve found that many people make trip planning decisions based on the expected “worst case” transit trip time, not the “average case” or “best case”.

  7. Regardless of the specifics of the math involved, I’m concerned about the effect it’s having on the people who presently live in the areas directly the stations. There is no doubt whatsoever that the value of the areas surrounding Link stations (present and future) is greatly invested in the success of the Link system as a whole. We’re already seeing rises in land values, putting pressure on the populations currently there. There’s been a lot of bluster in the news lately about region-wide spikes in rent and home values, and while the numbers themselves are up for debate, the concept is anecdotally plan to see. What’s going to happen to the affordability of Capitol Hill – an area already experiencing rapid (perhaps even unsustainable) growth rates – once the Link station opens up in a few years? I realize Capitol Hill is an extreme example, but other Link station neighborhoods are seeing it in the present (Columbia City and Othello).

    1. Seattle has an extremely high quality of life. You could blame Link. You could blame the Seattle Center and all its mega-events. You could blame the Seahawks, Sounders, Mariners, and Huskies. You could blame Capitol Hill for being Capitol Hill. You could blame the low unemployment rate and abundance of six-figure jobs. You could blame the relative tolerance for the homeless, that draws the homeless from all over the country to live here, and then compete for the low-end rental space. You could blame our silly law against sleeping in cars for pushing up rent. You could blame the hills and the valley and the mountains and the 4 months of pure blue sky. Just don’t blame all the single-family neighborhoods for bringing the world, as that is *not* what draws people to move to Seattle. Rather, that is where people end up, once drawn by all the other stuff that contributes collectively to quality of life. But do blame all the acreage used up on single-family houses for helping push rents up.

      1. @Brent: I think you are underestimating the importance of that SFH. Of course it doesn’t distinguish Seattle from anywhere else in America, but that doesn’t mean it isn’t important. People think about there future when they move across the country, and f raising a family in a SFH is still a pretty important part of American culture. Think of it like oxygen: it isn’t why people move to Seattle, but a lot fewer would move here if it wasn’t there.

        Also, I think you underestimate the drawing power of some of the streetcar suburbs. The concentration of Craftsmen style houses, for example, is a draw. [I suspect that the dream of a lakefront SFH is also a significant (albeit improbable) draw.]

      2. You could blame the low unemployment rate and abundance of six-figure jobs.

        Yes, that’s it.

        Remember the “Will the last person to leave Seattle please turn out the lights” billboard in the early 70’s? We had the same nice single family homes then (more of them and they were a lot more affordable). Relatively speaking, transportation (both private and public) was better then (it took less time to get where you wanted to go). As much as I think a lot of Seattle’s single family neighborhoods are really nice (and much better than a lot of suburban neighborhoods) I can’t see it playing a very big part in drawing people out here. If anything, it is a negative overall. For every person who falls in love with a Seattle neighborhood and comes out here to live in one, there are probably two who think “Seattle is really nice and I can get a job there, but can’t afford to live there”. People used to say the same thing about San Fransisco all the time. Like San Fransisco, attempts to preserve some of the those homes only increases the value of them.

        That doesn’t mean that I think we should ignore the historic (and aesthetic) value of the houses. It means we should acknowledge that there is a price being paid when we preserve them, and the price is being paid by everyone who doesn’t own property in the city. Balancing those principles suggests that other values (such as parking, or shadows) should be pushed way down on the list (in my opinion). If we build a new building, we shouldn’t require parking and we should allow as many people to live in it as possible. Doing so will, all other things being equal, keep the price of houses and apartments lower. Of course “all other things being equal” is the key phrase. The main reason housing is really expensive right now is because we have such a low unemployment rate.

      3. “Think of [SFHousing] like oxygen: it isn’t why people move to Seattle, but a lot fewer would move here if it wasn’t there.”

        That makes even less sense than applying Metcalfe’s Law to a milk run bus route.

        Demand for whatever form of housing is available in Seattle appears to be insatiable. Maybe there are people who wouldn’t move here without the single-family neighborhoods, but there are plenty of others to replace them who would move into town if there were more apartment vacancies. I’m not trying to say one is morally superior to the other, just that demand for potential housing in Seattle far exceeds supply, hence the sky-high purchase costs.

    2. This is why we need an extensive rapid transit system with many good station areas, both in Seattle and the suburbs. Three stations will make rents rise exponentially around those stations, but six stations will diffuse the benefit and allow more reasonable rents at the less-prestigous stations. Rents will inevitably spike on Capitol Hill, but not as much in Othello or Northgate or Roosevelt because there’s less demand for housing there. The main cost of housing is not the building itself (around $20,000 for a single-family house anywhere in Pugetopolis) but the land under it. The Station at Othello Park can’t charge as high rents as Capitol Hill, and a similar building at Angle Lake station or Kent-Des Moines Station would charge even less. Of course, the latter two don’t have anything like walkable station areas yet, but they could in the future.

      1. The $20,000 figure you quote is obviously a typo, for a 1000 square foot house that would be only $20 per square foot, which is about a third of national average. So what did you mean to say?

      2. There are a lot of ways of making houses affordable without building rail to the suburbs and hoping that people build affordable housing out there. You are absolutely right in saying that if we make parts of Seattle really pretty, and put all of the jobs here, then housing will continue to be really expensive. But we can add density both here and in the suburbs, and that will (all other things being equal) bring housing costs down. This doesn’t mean just more apartments, but that is definitely a big part of it. The apartment “tax” that is paid for new construction is paid for by everyone who doesn’t own property (to a certain degree). The “tax” involves things like restrictions on height, or even worse, the number of tenants. It involves parking spaces (parking spaces, for crying out loud!).

        Then there are lot sizes. Generally speaking, most of the houses in Seattle are on relatively small lots. If the suburbs (and that includes areas that used to be outside Seattle) had the same size lots you could increase density, while still producing plenty of single family houses. If folks had built lots of small houses on small lots instead of huge houses on huge lots, then a lot of homes would be a lot more affordable right now. In other words, it is all about zoning.

        Oh, and don’t get me started about Mother-in-law apartments…

      3. I’ll assume you meant single-family house buildings are worth $200,000 each, and the rest is land. It is true that for SFH, the land value is a significant part of the total value.

        The reverse is true about urban buildings. A 6-story apartment building will typically cost 10x as much to build as the land value, i.e., land value is less than 10% of the total value. Since construction costs are virtually the same between Capitol Hill and Othello, even a wide swing in land value doesn’t change the overall project cost much. Hence, an Othello apartment building would need to charge basically the same rent as the one on Capitol Hill.

      4. I have known people who have built their own house in Seattle or costed out a premanufactured house for a lot they expected to own (e.g., by inheritance), and they say it’s around $20K or $50K tops. OK, you can add another $20K for the construction company’s profit if you’re having it done, but the point is that the vast majority of housing costs is not for the building but for the land. People want to live in convenient locations, and drive the price of those locations up. If we can spread convenient transit to all parts of the city, then it would break one part of the price premium that the most pedestrian locations currently command.

      5. @Mike:

        According to the Manufactured Housing Institute, the average cost per square foot of a new manufactured home was $41.24 nationwide in 2009, The average MH being sold is around 1500 sq ft. My research suggests that 20K will just about buy you a lightly used 800 sq ft single wide. .

      6. t the vast majority of housing costs is not for the building but for the land.

        Absolutely not true for new construction. ROT for banks is 2/3rds house 1/3 land value for a construction loan. And yes the vast majority of people have to borrow not build with cash in pocket on land they inherited. $20k gets you a bathroom. $60k gets you a new kitchen. Can it be done for much less; absolutely. But again the vast majority of people don’t have the know how, time, tools or patience to live in a construction zone while working two jobs (i.e. your day job and then come home and play amateur contractor). On older homes yes, the value of the land can exceed the dwelling by a considerable amount. That’s because in many cases those homes if sold will be bulldozed to build a McMansion or multifamily.

  8. Great post and great analogy, but I think I can get a little closer to matching it. Rather than “stations are like nodes, except the analogy breaks down a bit because some nodes are more important than others”, I would say “human transportation is all about how many people can meet other people in a given amount of time. Link should be measured by how it improves the network by that measure. There are no concrete nodes, since transportation is too fluid and complicated for that, but the rule of squares still applies.”

    Generally speaking, this is what drives modern cities. It isn’t about the factories, or the proximity of coal, iron or other goods to those factories. It is about people. More and more, businesses are dependent on getting people together to share ideas and expertise. Likewise, for recreational purposes, people want to get with other people.

    I can’t emphasize the “transportation is fluid and complicated” part of my statement enough. New York works extremely well on my measure not only because it has a decent subway system, but because it has great elevators and people are willing to walk a few blocks. Every day, millions of people ride elevators and walk to their destination. It might include a cab, subway or bus ride, but walking and riding elevators are a key part of it.

    Of course, the analogy breaks down in a couple ways. There are some locations that are simply more suited for large numbers of people. On any given weekday morning, there are probably a lot of people that are making these random, point to point connections (going to grandma’s house) but the bulk of them are going to large employment centers. That being said, the trend seems to be away from large employment centers, and more towards dispersed employment — forty years ago, Bellevue was just a bedroom community while Fremont and South Lake Union had houses and a bit of industry. Despite these changes, the bulk of people choose to get together in certain spots, and those spots tend to be close to other spots.

    Another way it breaks down is that to a lot of businesses, some people are more important than others. If you operate a high tech office, you may not care about how the cleaning folks get to work, but bend over backwards to make sure the tech worker has a nice commute. Unfortunately, as neighborhoods have evolved classes, this seems to further exacerbate and increase class stratification.

    All of this explains why University Link is such a big deal. Whether weighted by the extra two measures (quality of location or value of the individual) or not, it is a great system. It will serve (to quote Sound Transit) “the three largest urban centers in the state of Washington”. A lot of those people will walk to Link. A lot of them will first ride another bus, or ride a bus after riding Link. As a result, the number of people who will be able to get together with another person in, say, twenty minutes, will go up a huge amount when it opens.

    All the more reason that complimenting the system with good bus service and other facilities is vital.

    1. All of your points are well-taken, but I again must rush to remind non-New Yorkers to avoid overestimating the percentage of the city that is high-rise. The high-rises are not responsible for nearly as much of the density as you may think, and the number of daily elevator rides per capita is orders of magnitude lower than the number of subway rides and pedestrian journeys.

      And that’s to say nothing of the primary gathering places for commerce and recreation, the portions of the city with the highest quality of life and greatest diversity of experiences-per-block, that are in mid-rise rather than high-rise areas. Yes, including the vast majority of Manhattan.

      1. Yes, but mid-rises have elevators, too. I’m surprised, though, that subway rides per capita are exceed by elevator rides. Then again, just about everyone who rides an elevator to work first rides a subway, so I guess that makes sense. I just assumed that there were plenty of people that walked or took a bus then rode the elevator. Obviously there are, but that number must be exceeded by folks who take a subway then don’t take an elevator. As someone who doesn’t take an elevator unless I’m going more than three stories up, it may be that a lot of New Yorkers walk those stairs in the same way.

        I think we both agree that walking is a very underrated form of urban transportation. So too are elevators (although I may have overrated them).

      2. When I worked in a high rise office building, I’d ride the elevator probably 3-6 round trips per day, to get out for lunch and coffee breaks and just “stretch the legs” during a sedentary office desk job day. But, I only commuted office-home once per day.

      3. Good point, Steve. All the more reason I would like to see d.p.’s study comparing per capita elevator use versus subway use.

      4. Again, I think you both are envisioning Manhattan as defined by its skyscrapers, and grossly overestimating its elevator dependency. The majority of even that tallest of boroughs exists in walk-up range.

        New York City as a whole has roughy one million structures, and only 60,000 registered elevators. The majority of New Yorkers rarely set foot in an elevator. The vision of the city as full of Gordon Geckos zipping up and down all day from their apartments-in-the-sky to their offices-in-the-sky and martini-lunches-in-the-sky and dentist-appointments-in-the-sky is not accurate.

        I’m only harping on this because the your line of “vertical city” thinking eventually begets brain-dead Futurists advocating
        stuff like this, which is both deeply unsustainable and very, very bad for city life.

      5. A fun test: Pull up a Google Map of Manhattan and drop the Street View cursor onto it completely at random. Then do it a dozen times.

        Nearly everywhere the cursor lands, you will be see buildings without elevators: with the exception of some institutional buildings, nearly everything 6 stories or under, and with less width than height, will be walk-up.

        These will pepper your Street View landscape even in the heart of Midtown. These will appear between the uniformed-elevator-operator swank of the Upper East Side. And from Little Italy to Murray Hill to Harlem to Inwood, these will dominate.

      6. OK, d.p., at this point we’ve entered the “silly argument” part of the discussion. It is like two guys at a bar arguing whether a player is great, or just really good.

        I never said that New York is dominated by Skyscrapers. I would guess that very few people live in them. I would also guess that most people work in buildings too short to be considered skyscrapers. Because New York is an old city, I would think that elevators play a smaller part in the lives of people than one might expect. If I see a six story building in Seattle, chances are it has an elevator — I would guess that isn’t true in New York. As you pointed out, it is more untrue than I originally guessed.

        From my own experience, I can tell you that I’ve worked in 7 office buildings. One in the U District, two in Bellevue, two downtown and two in other parts of Seattle. Oddly enough, the only building that didn’t have an elevator is one of the buildings in downtown Seattle. Like the buildings you mentioned, it is old. It is also slated to be torn down and be replaced by a bigger building (with elevators).

        In Seattle, I would guess that the old building is an outlier. Seattle is young enough that most of the buildings (even the short ones) have elevators. I would love to see the studies detailing the types of growth that are occurring in both housing and office space in the city. While much of it is in town houses and duplexes, a lot of it appears to be new six story buildings (with elevators). Likewise, I think the tide has turned with regards to office space. Although some companies prefer the Microsoft model of sprawling short buildings, companies like Amazon are obviously proving that big buildings in an urban setting aren’t out of fashion after all.

        By the way, if I wanted to describe a city where it is common for people to live and work in a skyscraper, I wouldn’t have mentioned New York; I would have mentioned Toronto. All in all, it is probably a better city to mention as a transportation model anyway (since their subway system is so good). Even saying that, the number of people who live in high rise condos are still way outnumbered by the folks living in houses. Although, if the trend continues, that might change in a few years.

      7. Fair enough.

        Pretty much any publicly-accessible multi-story building will have elevators these days, for completely valid ADA reasons. Including the 2-story sprawl at Microsoft. Which is why elevators-as-public-transport is a silly discussion even when trying to quantity verticality — and much more so when trying to quantify real, functional density.

        Toronto is one of the few major North American cities I’ve never been to. Opinions seem to be divided on whether their subway is one of the best, most efficient things to appear in post-war North America, or whether its emphasis on reaching single-use towers-in-parks on the outskirts, at the expense of adequate capacity and coverage through the older-style density of the central city neighborhoods, has rendered it a crush-loaded nightmare on its central segments.

  9. I live close to 145th/155th and am looking forward to having LINK as my (main!) option to get to downtown. I look forward to the quick travel to Northgate, U-District, Capitol Hill and downtown as well as points south. Really looking forward to the winters that shut down the city EXCEPT for LINK. And yes, I think my SFH value will definitely rise when the line is complete. I just wish there was a way to speed up the process/construction because I am really impatient and want to use it NOW, instead of waiting another 10 years or so.

      1. Actually for my job, I cannot take the bus from home due to my hours. I have to drive to #5 or #358 since #347/#348 end too early for me.

      2. … and if those routes end too early, the 511 (which drops at 145th post-evening peak) probably ends too early, too. Sort of a shame, for such a trunk service. I don’t think any of ST’s bus routes have outbound trips after midnight.

        (When I worked in SnoHoCo I worked later than the 511 was running a couple times. The thing is, it’s only a little worse than being stuck in downtown Seattle late at night the way our bus service works.)

      3. I mean, as regards the idea of value to the network, 145th/155th isn’t exactly UW. Or even Lynnwood, yet.

        As for late-night stuff… Central Link’s last outbound trip leaves downtown around 12:30. I sure hope that gets improved with North Link, though I’m not sure my fingers should be crossed. Aurora, too, is shortchanged for late-night trips, though Metro helpfully runs night-owl trips right up to the Seattle city limits (circa 1940).

      4. ” I don’t think any of ST’s bus routes have outbound trips after midnight.”

        You’re almost right.

        The 512 has a trip on Sundays that leaves 4th & Jackson at 00:08.
        The 545 on weekdays has a trip that leaves 4th & University at 00:01. It leaves the terminal before midnight.
        The 550 has a trip that leaves ID station at 00:06. It leaves Convention Place Station before midnight.

  10. “To take the analogy back to transit; if you have a station near your neighborhood, then the value of that station increases exponentially as stations are added elsewhere throughout the city.”

    Hmmmm… Since the value of each of the X stations is approximately X^2, isn’t that relationship quadratic and not exponential?

    1. I love trying to apply semi-complicated math to socioeconomic concepts. I remember how my college economics textbook tried to explain how a minimum wage is bad for low-income workers.

      The value-increase to the whole system may approach a quadratic limit, but the value increase to a single station is linear. Moreover, the farther off the new stations (travel-time-wise), the less value they add to that particular station.

      Simply put, the land values are going to rise, but not quadratically.

      1. I second Brent’s statement. Metcalfe’s Law assumes, critically, that each new node is directly connected to every other node. That’s why phones are a great example as any phone can call all other phones directly. However, since link largely follows the hub and spoke model, the connections between all lot of these nodes won’t be even close to direct, but rather a 45+ minute ride. For that reason having an urban core of crossing rail lines (Seattle Subway) gets much closer to this quadratic growth of value than what we are currently building.

        In other words, a Westlake to UW line via Ballard would be more helpful (even before considering density) than say, an extension to Tacoma with an equal number of stops, because the average distance between nodes (in terms of stops) would be far less in the Ballard case. Considering Metcalfe’s law is valuable in so far as it prescribes network choices that reduce the average number of stops between nodes, without sacrificing too much coverage (i.e. valuable nodes) of course.

  11. Another little mathematical rule is that mobility (defined as the number of people and places you can see with a given travel time budget) increases as the square of travel speed. That is why expanding Link is so important – it allows you to get around faster than the dismal 6 mph (or so) speed that urban buses achieve between many points in Seattle.

    Increasing density also increases mobility (as defined above), but the increase is only linear. Thus speed trumps density as far as mobility is concerned, and that is why spread out suburbs connected with high speed car travel works so well in many ways.

  12. The law has a third dimension, which is the primary land uses at the station. If every station only has residential development it doesn’t multiply as the line doesn’t connect to destinations that a resident would want to visit, for example. On the other hand, if a station with mostly residential around it connects to jobs, event centers, shopping and hospitals, it value does indeed multiply.

    While that concept is often understood about rail stations, it could also apply to buses. For example, there are several KC Metro routes that run primarily through residential areas but not quickly link directly to jobs, medical officers and shopping. It’s not surprising that those routes have low productivity.

Comments are closed.