To me the new bus network presents Auckland Transport with many opportunities to improve transport in Auckland. These include:

  • It simplifies the bus network and removes unnecessary duplication and inefficiency.
  • It allows for greatly improved frequencies across much of the city.
  • The simpler network structure means that it will be much easier to market the PT network as a whole rather than only being able to focus on individual parts.
  • The network structure encourages AT to put effort into improving bus stops and interchanges around the network further improving its quality – an example is the proposed Otahuhu Interchange.
  • The network provides a ready-made plan for what roads will need better bus priority built (e.g. bus lanes, bus priority at intersections etc.)
Frequent Bus Network 2016
The 2016 frequent network (although missing the changes made to the southern area)

With this post I want to look at another opportunity that AT will hopefully make as part of the roll out of the new network – changes to bus stop spacing. The problem is that along many routes the spacing of bus stops is extremely close, sometimes less than 200m apart. I only checked a few routes but on them stops appear to only average about 300m apart.  While that might be good for those living nearby meaning they don’t need to walk far to get a bus, it can make the bus slow by stopping frequently.

What kind of impact could changing the stop spacing have?

For the purposes of this exercise I’m going to assume that on average a bus loses about one minute of journey time per bus stop. That includes time to brake, dwell at the stop while passengers get on and off and then accelerate again. As a comparison, over a distance of 12 km (many bus routes will be longer than this) a bus that stops on average every 300m will make 40 stops, a bus stopping every 400m will stop 30 times and a bus stopping every 500m will stop just 24 times. Based on that one minute average stop time above then stopping every 400m could save 10 minutes per run while it could be 16 minutes saved with a 500m stop spacing. It’s also worth pointing out that in the situation above all stops are treated equal as in reality a larger stop where lots of people get on will likely be more efficient on a per passenger basis during the dwell time phase.

What are the benefits to increasing stop spacing?

The first one is that passengers get a faster trip whilst they are on the bus but that would have be partly offset by a longer walk to the bus stop. From an operations point of view, depending on the route that time saved might just be enough to allow an extra bus to be run without needing to buy another bus and hire more drivers. In effect it means that can either reduce how much we spend to get the same level of service or get more service (and therefore patronage) without having to spend more. Both options are positive.

The risk with increasing station spacing is that it the service becomes less attractive due to being harder to access stops. Bringing this back to the new network, the question is whether the increased frequencies on the new frequent bus network could be enough to offset the extra walking that would be required due to wider stop spacing. A research paper from Australia last year has attempted to answer exactly this question by surveying people in the major Australian cities. The abstract from the paper is below

Network planning of bus services requires addressing the trade off between frequency and coverage. Planning for good coverage of bus services using the rule of thumb that people will walk four hundred meters to access bus based public transport services means sharing the available budget between many services. For the same budget, the alternative approach of concentrating frequency on core corridors implies lower coverage and that some travellers would need to walk further to access bus based services. An understanding of to which extent people are willing to walk to a bus stop with higher frequency would provide empirical information for bus network planning.

The research question addressed by this paper is whether travellers are willing to walk further to a more frequent bus service in the context of Australian cities. A Stated Choice Experiment approach is used to elicit the trade off between walking further to access more frequent bus services. In doing so the paper investigates the potential success of reorientating a coverage approach to network planning, prevalent in many Australian cities to one predicated on concentrating frequency in corridors. The results show travellers in Australian capital cities are willing to walk around 206m to 327m further for a ten-minute reduction in bus headways. These research outcomes provide valuable Australian evidence confirming travellers are prepared to walk further to a more frequent bus service.

And just expanding on the conclusions they say

The major contribution of this paper is the quantification of the trade-off between walk distance and bus frequency as identified by the MRS. The results suggest that the travellers are willing to walk further to a more frequent bus service in all Australian capital cities. Travellers in Australian capital cities are prepared to walk further by between 206m and 327m for a ten-minute reduction in bus headways. The policy implications for network planning are that increasing frequency, even if it means travellers have to walk further to bus stops, will attract higher patronage. If budgets are fixed, this suggests that moving from a policy of coverage to the ‘European’ approach of concentrating frequency in corridors is likely to be a good policy if increasing public transport patronage is desired. Of course, concentrating frequency in corridors will require some travellers to walk further to access bus based public transport and will require policy-makers to consider and implement complementary policies to ensure accessibility is not reduced for those travellers unable to walk the additional distance. This could take the form of lower frequency access services or more flexible services to provide on-demand access to high frequency corridors.

In other words if we assume that we are at least a little bit similar to our cousins across the ditch then we are likely to see similar results too. It suggests that not only is the new bus network the right thing for Auckland Transport to be doing but that people will be prepared to walk further for higher frequency services without it compromising patronage (which would be boosted by the higher frequencies being provided). Implementing changes to bus stops spacing at the same time as the new network rolls out is something AT really should be doing.

Share this


  1. I’d be cautious about removing increasing the spacing at the same time as increasing the frequencies as it may turn out that buses will stop less often anyhow.

    For instance, if buses run just a few minutes apart in the am peak, then it is highly likely that buses will overtake each other at each stop. No need to stop if a bus is loading passengers already. The concept of “leap-frogging” buses could potentially halve the number of stops.

    Going home is a bit different in that typically the bus starts off full and it’s likely the bus will need to stop at every bus stop on the route.

    1. On the contrary. I think changing spacing at the same time as frequencies is the perfect time to do it. If freqs are increased but stops not changed then people get an expectation that it is how things have to be and will complain even louder if you go to change them if some at the same time then people are able to much more clearly see the trade off being made and understand the reasoning for it.

      1. I agree. It is basic PR that ideally one combines as much pain (change = pain) into one to get it over with, if one wants to do it anyway.

    2. Buses overtaking other buses need quite a bit of a gap in the left lane. Not only does it slow down the line, but it forces the buses to lurch even more than usual with multiple lane changes. This makes the service very uncomfortable for users who will be tossed this way and that. And it’s not the same as spacing further stops. You also risk ending up with bus convoys instead, empty buses being stuck for some time behind full buses that are very slow.

      I think it is much better to implement limited services on very used lines. They do it in New York, where some routes have both “local” and “limited” buses, the limited buses stop only at each kilometer or so, the local buses stop at every 200 meter or so. They also recently started having the limited buses have payment before boarding, to speed up boarding. It’s some sort of BRT-lite, they call it “Select Bus Service”.

      From my calculations, on heavily used lines, lengthening stop spacing from 200 meters to 500 meters can increase speed by nearly 30%, lengthening spacing to 1000 meters can increase it by nearly 45%. But that’s on very heavily used lines (around 6 to 8 boardings per kilometer). Lengthening stop spacing starts making sense at about 4 boardings per kilometer if my calculations are correct (in which case a 500 m stop spacing results in 10% higher speed and 20% with 1 000 m spacing). Which ironically means that on very long routes in the suburbs, you can have a stop spacing of 200 meters or even 150 meters without problem, because there are so few passengers by kilometer that it doesn’t matter at all, the bus will bypass most stops anyway.

      And even though 4 boardings per kilometer seems high, many lines do not have balanced boardings on their entire route. So an average of 3 boardings per kilometer may mean 1 boarding per kilometer on half the route and 5 boardings per kilometer on the other half. So even if overall the average seems to indicate that lengthening spacing isn’t useful, it may actual be pretty useful because of that higher section.

      1. not sure how AT would jump at limited stops ? …. they certainly stopped the limited stop express trains
        and also trying to remove or limit the express buses which follow a similar philosophy (although less stops
        than a limited stop service as you describe would use) …….
        not that I am against the idea personally……..
        one thing to point out though – not a week goes by where we don’t have some poor soul jump on our
        bus and then realise once it hits the motorway that they wanted to go to Newmarket/Otahuhu/Greenlane
        etc etc …… and its a bit tough luck now …………and its always cleary stated on the sign its an express – I would iimagine a limited
        stop system might be prone to that too

        but now come to think of it the ability to catch an express bus at 40-50 mins into town v a great south rd bus at 60-80 mins
        is probably well worth an extra walk !

    1. No need to assume, they also looked at that in the research paper quoted above (buried in the data table). If I am not mistaken the result for that aspect (in the kind of terms they used for the frequency reduction) was people were willing to walk an extra 179-343m for a 10 min reduction in journey time. Also people in that study were willing to walk further for a less crowded bus.

      However, while I think this paper provides good support for AT’s planned changes to a “frequent corridors” approach, I’m less convinced if it says much about the benefits of increasing stop spacing. The study seems to have be done with the view of looking at the benefits of a frequent corridors approach and I would expect the questions asked in the study would be different if you were aiming to find out whether stop spacing is a good idea. With this kind of study even relatively subtle changes in the way the questions asked can make a big difference to the answers you get.

      In any case, I would say that making bus stops further apart would need to be done very carefully rather than as some kind of blanket policy. The kind of trade-offs that are likely to be needed to make it work would not be easy to achieve on all routes. For example on routes where busses already run every 10 mins, I can’t see how reducing bus stop distances would have enough improvement on journey time/frequency to compensate for the increased walk distance. There is a danger that any changes like this may reduce patronage significantly if not done well.

  2. The effect is even better if we were to have some stops with good cycle parking facilities.

    There may only be a few of these on each route but they would allow people to cycle up to 3kms to that stop in less than 10mins. Perhaps they could be placed at the local shops where more frequent foot traffic will make the bikes less likely to be at risk (although bike theft is not a big issue in NZ relative to a lot of other countries).

    You often see in the Netherlands fairly isolated looking cycle parking at bus stops even in quite rural areas.

    1. Ideally, if you’re already on a bike most people should be able to cycle far enough to get to rail or the busway, which is going to provide better service. It’ll also be easier to provide security at a few larger locations, where cycle cages/CCTV etc. are practical. There’s a few town centres that might make sense for bike-and-ride parking facilities, though, especially at major bus interchanges (Mangere, Pakuranga, Mount Roskill etc?)

      On the flip side, the fact that some people might be cycling to a stop is not a good reason to reduce stop spacing for the people who are walking there. Regardless of the merits of bike parking, it doesn’t really have any relevance to the design of the network.

      1. I agree 100% and I wasnt trying to argue against less stops – quite the contrary I think it is a great idea. I just see cycling as a great way to reduce the effect of the larger distances even more. I took a bus the other day (which I dont do very often as I use the ferries) from Takapuna to Browns Bay and the major delay (as compared to the same trip by car) was the closely spaced stops – as well as the lack of bus lanes. It was a surprisingly busy bus going both ways so we stopped frequently.

        I also agree that on the Isthmus for example, most people are within 3kms of a train station and on the North Shore a busway station/ferry terminal. I was thinking more of the areas that are devoid of PT with its own right of way like the east of the city and areas in the south.

        Without a doubt, interchanges need to have very good cycle parking as well as cycle infrastructure from nearby residential areas.

        Cycling and PT just work so well together that it is crazy not to have it front and centre in our mind. The vast majority of Aucklanders do not want to cycle more than 5kms to work. However, I think many would cycle 1-3kms to good quality PT and that distance (even now, let alone if the CFN was implemented) would capture most people in the city.

        1. My second-to-last sentence was a little ambiguous: I meant that even if some people are cycling to bus stops, we should still be designing stop spacing around people who are walking to the stops. So more like 400m than 1km. I am arguing that we still need lots of stops, although not necessarily quite as many as we have now.

          Not that I’m opposed to having a bit of cycle parking at the local shops. But I suspect they’ll be more used by people actually going to the shops.

          All of this assumes that the streets are actually hospitable to cycling, of course. The absolute best case is like the Netherlands – crash programme, ten years or so. If we have the political will for that, then the rapid transit network should look at least like this: – which makes a good dent in the “PT sucks in East Auckland” problem. Hopefully almost everyone in urban Auckland will have a busway or rail station by then.

        2. Goodsoid, those East Coast Bays buses are indeed very popular and I’ll bet that you found that a relatively large proportion of passengers were gold card users or were paying with cash (and/or having a “does this bus go to…?” conversation with the driver), both of which obviously cause delays. The answer of course is more aggressive promotion of Hop.
          Not much opportunity for bus lanes on Beach Road.

          1. And I’ll bet he didn’t. As a Shore Boy and PT user myself, I see very few gold carders, casuals or ad initio users. References please and drop the patronising tone.

          2. Johnski What time do you travel ?? I am not a northshorite so I could not say for sure but my experience in my two hats
            is a worker-commuter (or prev student) to the city at peakish hours I see mostly the same sorts of people
            mostly under 50s a few high school students……..however in my other hat with days off with the kids I see
            a shed load of gold card users and Mums with prams etc using PT to go to shops, doctors, outings, school/plunked/playgroups, social gatherings etc ………..Perhaps it very much depends on the time that you travel compared to DavidT ?????

          3. Actually David is bang on as it was off peak. Gold Card users and cash users were another major delaying factor.

            We really need to move to the driver having no responsibility for tickets and he/she just drives. Tickets are enforced by random inspections by plain clothes inspectors. Then you can open both doors, everyone floods on and it is their problem whether they have a ticket. Tickets are sold at vending machines or at dairies, service stations etc (if no HOP card).

            This works really well in most of Europe and increases the reliability of services no end. There may initially be some increase in fare evasion (as there was in Melbourne when they got rid of inspectors on trams) but onmce people are pinged a few times for no ticket, that will change. I suggets it will also be off set by increased usage as the service becomes more convenient and easier.

          4. most our off peak are cash fare too I note …….big education campaign by a couple of writers
            to our local paper to help older people with gold cards get around the big issues with the train ticket machines
            by getting gold cards ……… my comment would be if you want them for the bus then make the machines
            a bit more robust and plentiful please …….
            Same too goes for having dairys etc sell them – would have to be every other dairy rather than just the few that do/did the
            train concessions and bus monthly passes

            but yes most regular commuters I know would love to have the cash fares stop ………..
            whereas I used them despite the extra cost due to the pain of the bus tag on and off system for me …
            contributing to the issue of course ! but trying to have right change as often as possible

            so fine go cashless – but with competence, working machines, lots of places to buy – don’t make it a pain in the butt

          5. goodsoid, yes that would make a huge difference but AT don’t seem to be in a hurry to increase the number of Hop retailers so it will be a long time coming. Don’t see why it couldn’t be done on the NEX though once Ritchies have implemented Hop.
            jjay, are you still having Hop card issues – or do you mean it’s just hard to use with kids, shopping etc?

          6. DavidT – my main issue was with the snapperHOP on the bus …..not used the ATHOP on the bus yet
            have used it on the trains with just standard issues with the machines and queues etc for that……..
            snapperHOP issues were yes primarily tricky to juggle the stroller full of child, daycare bags, work stuff etc
            and tag on and off without impeding fellow passengers trying to disembark/embark – it was not always successful
            tag on and off………and sometimes I’d forget to tagoff (esp if we had a chaotic small person ride and I just wanted to get home ASAP!) … if you added up the penalties for
            those failure to tag offs or things not tagged off properly it worked out just as economical to pay cash…….

            I did actually talk to the snapper people – saying was there a way of putting 2 sensors on each door and seeing if they
            might pick the card up better ……..but they said they’d already had issues with it being too sensitive and picking up cards when it should ……..

            when I return to work this time I imagine we’ll have ATHOP on our buses …….I am going to see how that compares ……

          7. Jonski, your tone was equaly patronising and you were incorrect….

            I catch that bus fairly often living on the 858, and 839, and near the 879/ 875 and off peak less than 15% wold be using hop,despite the company having been on it for months now.

            On the Supergold/hop issue, I had to tell George Wood in December that you couldput supergold concessions on a HOP card, hopeless.

        3. yes mini cycle interchanges at intersections with major roads very useful, especially where this road goes to an isolated spot that is naturally going to have less bus service. Lake Road at intersection with Bayswater Ave a really obvious example. Less useful on Isthmus where have parallel bus routes, but very useful where lack grid structure.

  3. On the other hand, when there’s no-one getting on or off at a stop, the bus isn’t delayed at all. Seems like a different tradeoff in the peak, with very frequent buses, and lots of passengers, versus mid-day or evening where there are fewer buses and most stops will be empty anyway.

    The obvious compromise is having some stops that are OFF-peak only: but is that just being too complicated?

    1. Hi Steve – having stops that work only part of the time would in my opinion be a huge disservice. It really confuses occasional or new users, leading to frustration and uncertainty. It also means that we need to spend more headspace planning our trip when one of the advantages of PT is that I can use it half-asleep, or while reading a book / playing a game on my smartphone.

      Express buses have some uses, but should be handled very carefully, otherwise, if you get on a route, you should be able to depend on being able to get on/off at any stop on said route.

  4. I’m also all for increasing stop spacing in order to increase speed and frequenct. It would also be great if the fewer remaining stops could be upgraded at the same time to increase shelter, provide bike parking, real time signs, timetable information etc etc. I also agree that consolidating the network should go hand in hand with investment in bus lanes on the bus corridors to improve the speed, frequency and quality of the service on those corridors. This way, although people will have to walk a greater distance to catch their bus, they will be rewarded with a much better experience when they get there.

  5. State Choice experiment??? I’m giggling.

    You can’t trust what people “say” they would do.
    We lie in order to make ourselves look good.
    And we don’t even recognise we’re doing it.

    Why didn’t they actually do some comparisons between routes with different stop / frequency spacing and survey how far people ACTUALLY walked to get there?

    I would say that trading 100m for 10 minutes (actually on average 1/2 that over the route length) would be a pretty good bargain though.

    1. I am with you on this one Roger. Peter Winder showed me data from the introduction of the Perth Diesel units in the early 1990’s. The stated preference results showed twice as many people said they would swap to the train from their car compared to the revealed preference surveys carried out after they were introduced. A better way would be to model the perceived cost to a passenger of a longer walk versus a shorter bus ride. ART3 can do that as can the old APT model.

  6. To me, the obvious answer is “yes”, and that’s what I do sometimes.

    The goal of commuting is to arrive at some destination, with different modes having trade-offs between comfort, speed (time taken), costs, level of attentiveness required, etc. Other things being equal, the more accurately you know how long the journey will take, the more likely you are to use that mode. Having fewer bus-stops will make your assessment of how long your commute will take more accurate, as will having more frequent buses.

    Having fewer bus-stops with more space between them is better than having more stops because you can estimate how long you’ll need to walk to arrive at the further stop and incorporate that into your commuting plan, but you’ll never know how long a bus will be delayed at each stop.

    While walking more usually is good for most of us, there may be a problem for senior citizens and young children.

    1. The travel time trade off also needs to account for different perceived or behavioural costs. Most people would prefer 1 minute sitting on a bus to 1 minute walking. (You are dry and you are not worried about whether the bus will arrive. The EEM1 manual suggests for work trips $6.60 per hour for walking, cycling and standing passengers and $4.70 for a seated passenger. So for each additional minute of walk you need to provide a reduction in bus travel time of 1.4 minutes to break even.

      1. Though there may effectively be a preference to being on a bus as opposed to waiting on a bus, I think that overall, the advantages are greater than the disadvantages.

        1- Higher speed: the gains are relative to how heavily used the lines are, there would be no speed gain on feeder bus lines in suburbs with 1 people hopping on board per kilometer, but a simple halving of the number of bus stops can increase average speed by up to 30%.

        2- Higher reliability: if you plot average speed by number of boardings per kilometer, what you can notice is that at first speed will fall quickly as the number of passengers increase, but at one point, you reach a plateau, speed still decreases but much slower. The plateau is reached when the bus has to stop at all its stops. As this point is reached faster if the bus has fewer stops, it can guarantee a more reliable travel speed on heavily used lines. Meaning buses will be more likely to stick to their schedule. So not only can reducing the number of bus stops reduce the time spent in the bus, it can also reduce the time waiting for the bus. (and personally, I prefer 1 minute walking to 1 minute waiting for a bus that is late, at least I am doing something).

        3- Higher frequency: the higher speed of buses if there are less stops means that it takes less time for each bus to do its route. This allows the transit operator to increase frequency and capacity. If we ignore idle time, if a route takes 30 minutes to run through, a bus can do it twice in an hour, but if it takes 20 minutes instead, the same bus can do it thrice. For a marginal increase in cost (gas and maintenance), you can increase capacity and frequency on all lines by removing bus stops. This higher frequency can reduce waiting times for people who do not plan ahead their trips but instead just go to frequent lines and hop on the next bus that comes.

  7. How about we let market forces decide? What do you mean I hear you say?

    Well imagine if most stops remain as they are, but the every 500m or so ones are designated A stops, rest are B stops or if B stops are too close, adjacent ones are merged into one stop.

    B stops charge users more to use them (easy for HOP to manage I’m sure) than the users at the A stop.

    To counter folks using cash – cash fares on buses are always at the B stop price for that zone regardless (rounded up to the nearest $).

    This means if going to your nearest stop is a B stop you will pay more for the convenience – you may soon wise up and decide to walk to the A stop where the fare is lower even if the walk is slightly further or a less convenient stop.

    How much lower?

    Well how about 50 cents different – not enough to really impact people who still need to use B stops but enough to encourage people to notice and change their behaviour.

    Furthermore, if a bus stops to let someone on at a B stop, why not have the HOP system then credit all the other HOP users on that bus (via their tagged on HOP cards) a few cents per stop as a small token. That way, those who bear the most inconvenience of having buses stop at B stops – i.e. the existing folks on the bus – get some reward for the disruption it causes.

    Of course, if a bus never stops at a B stop then theres no cost to anyone. Alighting at a B stop could either not cost you anything or you could also pay the higher fare – not sure on that one.

    So even if you trip is slow that morning (thanks to all those B stop passengers), you know that AT will give each of you a small credit for that trip, paid for by the folks who use the B stops.

    And if the bus stops a lot at B stops, you may get easily 50 cents or more off your fare for that trip.

    Over time that would naturally drive people to A stops who use PT a lot who want to save money, and still not discourage the occasional PT user who uses their local B stop as they don’t know better. Once they get wise to the difference and become a regular PT user, then they can decide if the extra walk is worth the saving.

    My gut feel is people will vote with their feet and use A stops if you give them an incentive or lower fares to do so.

    1. That would provide close to none of the benefits of actually closing the stops, if there were still a few people willing to pay to use them, AND be incredibly difficult to implement with the HOP card, AND not be compatible with integrated zonal fares, AND be (accurately) perceived as incredibly unfair. Think of how you launch this – you’re just up and announcing that some random bus stops will now cost extra for a difficult-to-explain reason.

      1. A major drinks company tried out a drinks vending machine that charged more on hot days (when demand was higher) as a way to increase profit and be economically efficient. The problem was people thought it was unfair and simply didn’t buy a drink as they didn’t want to be ripped off. They gave up and went back to a fixed price. A free market is efficient but as you point out fairness matters too particularly with public transport where there is a positive externality and you are trying to encourage greater use. In terms of measuring how far people will walk to avoid a cost that information is easily available at fare boundaries such as Mt Eden shops. No one pays the additional to use the first stop in the higher fare stage.

        1. A better way is the drop the price on cooler days, not put it up on the hot days – that is seen as gouging.

          Plenty of retailers have Cheap Mondays or Cheap Tuesdays (e.g. Movies), or Taco Tuesday 9at the local Mexican), The local Indian has $10 “Big Wednesday” lunches instead of $12.50 on other days – whatever it is you name it, people discount sometimes to bring in customers.

          Same with bus stops – A stops = “Cheap Tuesday”, B stops = “normal/the rest of the week” prices.
          How hard is that?

          Do you see people not going to the movies as they feel ripped off the rest of the week (well actually I do, but thats I reckon because movie tickets are outrageous no matter how cheap – given the crap they put on, but thats a different discussion for elsewhere).

      2. not too fussed about the issue – but what about the fairness of closing bus stops and making some people walk further while others get to retain their stop?

        1. It’s still unfair, except that people accept it as a necessary part of a tradeoff to speed up the buses. Fewer stops = faster buses is logic that people can sympathise with.

        2. I hate this concept of individual ownership of a public good. Public goods shoud be designed to maximise the public good.

      3. Well the alternative is simply closing the stops – so why not ask people what they would like then?

        I think some people would accept paying a little more to keep their stop – if the alternative was no stop there at all.
        And if no one uses it as they don’t like the “price” – well the bus never stops there, no harm done.

        And lets be clear – people have to accept a trade off between convenience of bus trips and speed of bus trips. The more of stops you have the slower the trip.
        They’re called Omnibuses for this reason. Making it an economic decision does help people grasp the concept.

        As for HOP not handling it, it knows what stop you’re at now when you tag on/off, how else does it calculate the fare you have to pay otherwise?
        So, all it needs to do, is deduct the “full rate” fare when its a B stop, and a discounted fare (by 50 cents or whatever it is) when its an A stop.
        Hardly rocket science. And if you enter and exit at A stops, you get the full discount, use a B stop for entry or exit then you pay a higher fee.
        How hard is that? If the credit the other HOP users on the bus is too hard, then don’t do it. As I said its an optional step.

        As for making it clear – the “A” stops get a discount, B stops don’t, easy to inform users with a “A” or “B” letter at the top of the Bus stop sign to indicate which it is (and/or change the colour of the background on the sign too to make it even more obvious), this is the sign that shows the bus stop number and/or shows the bus routes that stop there.

        These all have to be redone during the PTOM rollout anyway as the bus route numbers are all changing so its not hassle if you do it at the same time.

        Once a stop is an A or B stop I doubt it will change, just like bus stops don’t move around much. But, if the patterns change, then I’m sure some adjustment can be made.

        1. Well the alternative is simply closing the stops – so why not ask people what they would like then?

          That’s not how anything works in local government. The status quo of keeping the stops open is always the alternative by which any policy is going to be judged.

          I think some people would accept paying a little more to keep their stop – if the alternative was no stop there at all.

          And lets be clear – people have to accept a trade off between convenience of bus trips and speed of bus trips. The more of stops you have the slower the trip.

          Closing the stops completely has a clear and easy-to-explain benefit: the bus won’t stop there, and the journey will be faster. Your proposal doesn’t close the stops, and is simply the worst of both worlds. In reality, most people will walk to the cheaper “A” stops, grumbling the whole way about the bizarre fare rules, and a few people will still use the more expensive “B” stops, thus still slowing down the bus. The worst case is that the bus still stops at the “B” closest stop to you, but you walk farther to get to the cheaper “A” stop. If the “A” stop is earlier in the run, you actually get to watch someone else board the bus at the “B” stop, knowing that the council just made you walk a couple of hundred extra metres for no reason.

          It’s possible that some stops will go from having a few passengers to no passengers on many runs, but then why not just close the stop completely?

          As for HOP not handling it, it knows what stop you’re at now when you tag on/off, how else does it calculate the fare you have to pay otherwise?

          It knows what general geographic area you’re in, and often is accurate enough to get the exact stop, but not always. People also can and do tag off before they get to their stop, which would lead to both people being overcharged when they happen to tag at the wrong time, and the savvy can tag off at the last cheap stop before where they really want to get off.

          It also doesn’t work with zonal fares, which AT are intending to introduce this year, or daily/weekly/monthly passes. You pay one fare, no matter how many trips you take, within a specific set of zones across the region. That can’t be combined with charging based on the exact stop you use for a specific trip. Part of the reason for zonal fares is that they are simpler to understand, and that will encourage new people to start riding: charging extra for some stops is a huge step in the opposite direction.

          There’s also Gold Cards – if you get free travel anyway, the 50c fee is irrelevant to you.

  8. I think one important nuance of this discussion is that the *average* stop spacing is what determines the route speed. It’s easy to get hung up on one stop being only 300m w along when the ideal spacing is 400m, or whatever, even if bye average spacing along the route is 500m.
    More important in my opinion is simply locating stops well. Say there is a cluster of apartments on one side of a motorway gully and a town centre on the other. I wouldn’t worry if they were only 300m apart if they were ideally located for each , or in other words you wouldn’t push the second stop 100m down the road out of the town centre just to get ‘ideal’ stop spacing.

    1. Its getting right stop locations, some areas may have more stops than others, most importantly stops should be located at major intersections. The type of route will determine stop spacing, busy routes should would have less stops than quieter feeder or community routes.

  9. If people are only willing to walk 179-343m longer for a 10 minute reduction in journey time, well that’s hardly anything for such a huge reduction in journey time.

    1. I’m not sure the interpretation is that simple, due to the squared relationship that exists between walk-distance and catchment size, i.e.: Catchment = pi*(walk_distance)^2.

      For example, a stop with a walk-distance of 500m has a catchment of 785,000 sqm. Increase that walk-distance by 179m to 679m and the catchment increases to 1,448,000 sqm – an increase of 84%. Increase that walk-distance by 343m to 843m and the catchment increases to 2,232,000 sqm – an increase of 184%.

      The non-linearity between catchment and walk-distance makes it difficult to assess the relative effectiveness of reductions in stop spacing (and the associated increase in in-vehicle speed) and/or increases in frequency. Which is one reason why I agree with other comments that have suggested revealed preference studies would be more useful, albeit perhaps difficult to find.

      Perhaps AT could look to use the implementation of the New Network to understand the question in more detail, by mapping stop catchments before/after the changes?

  10. – There’s a good discussion of the maths of stop spacing at

    – The paper doesn’t discuss survey responses broken down by age group. It’s to be expected that the elderly will be happy with the less walk/ less frequent status quo. To gain political acceptance for change, their needs need to be acknowledged – for example, by setting up a clearer two level network with less frequent ‘coverage’ routes interspersed between the more frequent ‘patronage’ routes.

    – The survey did not mention punctuality as a variable. But let’s remember that frequency and punctuality are intimately related. Frequency and reliability (that is, random variations in frequency) go together to create the actual average wait for a randomly arriving passenger. Extra wait caused by unreliable service has a much greater disutility than ordinary wait time.

    Compare a bus that nominally runs every 30 minutes, but has a common lateness of up to 10 minutes, with a bus that nominally runs every 10 minutes, but has a common lateness of up to 10 minutes. In the second case, the lateness clearly has much worse effects in undermining the usefulness of the high frequency, since – it varies the actual headway much more, as a percentage; – as a result, it will cause more overcrowding.

    So the ‘frequent network’ should always be considered as the ‘frequent reliable network’, and promoting it should give high profile to the things that are needed to improve reliability.

  11. For low frequency semi rural areas i like the sound of the “Transport on demand schemes” being used in the uk, scotland europe. I think for the outlying areas these less frequent areas could also be combined with parcel delivery services. But basically book it in like coundown online and the bus shows up at your house.super helpful for elederly, people with disabilitie and young.

  12. I generally agree with the trade-off of a larger distance between stops for a faster service. However, I think Auckland’s topography shouldn’t be ignored in all of this. The 400m circle is often used as the walking distance that people will use but that depends on whether you’re on the flat or on a hill. Walking a further 100m up a hill is enough to put people off using PT. I used to live 600m from my nearest bus stop but most of the walk was up a steep hill – enough to put me off. If that 600m had been flat I’m sure I wouldn’t have minded. I realise that if you go uphill to the stop in the morning you get to go downhill at night – but just one of the trips uphill is enough to convince you to drive instead.

    I think it would be an interesting GIS exercise to map the walkabilty of PT stops in Auckland using a generally agreed walking distance on the flat, and a walking distance on steeper terrain. Including in the analysis busy roads that take a few minutes at the lights to cross etc would also give a truer illustration of what areas in Auckland can viably walk to a PT stop.

    Also, this may be a dumb question about the new network, but how would someone travel from Highbury to Takapuna? The network doesn’t show any buses heading this way (without taking a circuitous route)?

    1. The map only shows the “frequent network” of buses that at least every 15 minutes, 7am-7pm, 7 days a week. There’s also other less-frequent services: a connector network that runs at least every 30 minutes, and various sundry local and peak-only services.

      Adding all those other services to the map would make it far too complex to show the whole region in a single map, and the exact services haven’t been decided for most of the region yet, anyway. There’ll be individual maps like this one, that show absolutely every service, when the routes for the North Shore are consulted on (sometime this year, I think?)

  13. is not only the distance but the consistency also. In Melbourne Bus and Tram stops are often located where you expect them to be, even if you don’t know the area, just walk 300m to the intersection and sure enough there will be a bus stop there. And another one across the rintersection for the service going the other way. Quite different in Akl where there are isolated bus stops like this very lonely one

    1. The position of that bus stop would make a lot of sense if there were a practical way to cross the five lanes of traffic between it and Tapora Street.

      1. I know. The only “legal” way from the south side is to cross the Sliplane then the light at Tangihua, then cross Quay, then cross Tangihua again the port side trying to dodge the always red light running trucks and walk to the bus stop crossing two slip lanes with very heavy truck movement. Funny thing is that it’s quite a busy stop but most of the people just runs across quay. I cringe everyday when I see it.
        Seems like the new beach road cycleway is coming that way too with the plan to install a signalized pedestrian crossing just close to Tepora st. Not holding my breath

        1. It’s “legal” to cross wherever you like, but frankly it’s stressful and dangerous playing live-action Frogger for maximum stakes.

    2. Its actually a reasonably well used stop, surprisingly (I use it and the one opposite twice a day…).

      Mostly with students but with some nearby office workers, too.

  14. Can I ask a question, not related to Bus Stops.

    I have got a new monitor on my PC and I notice some comment threads have a grey background.
    I note that the original poster always gets their comments in the grey background but it seems some other folks do too.
    Is this background because the Blog has made a change or is it that my monitor was really crap and didn’t show that grey background so I never noticed it before?

    1. The blog’s regular authors all have grey backgrounds. It’s so you can tell which comments are the “voice” of the blog, and which are random unrelated yahoos like me.

    2. No it was just your monitor, hasn’t been any changes. Light grey backgrounds appear on every second comment thread i.e. first comment is white, second is grey, third is white etc. Replies to a comment stay in the same colour as the original comment. If the post author makes a comment it go darker still and has a think dotted red line around it instead of a grey one.

  15. Maybe take out every second stop for the frequent network, and keep spacing the same as it now for the local network.

    1. Agreed in principle. Connector services are going to be slow and infrequent with low patronageanway,keepstopspacingtighter.

  16. those that can will probably walk further …those that can’t will miss out – just as for some
    there is already impacts as this article highlights

    though this does not highlight spacing of stops you can see that for those described in the article
    stop spacing beyond a certain distance would have a negative impact too
    (though I guess since they won’t have a bus soon it won’t)…….

    So the point here is that
    1. Some people can’t easily walk longer distances – despite wanting to use and indeed even relying on public transport
    2. lack of close scrutiny in choosing stops to close will also affect people (by that I mean for e.g. have you removed
    the stop that now means Granny/SchoolChild/VIsion impaired person etc now needs to
    cross very busy road with no signals to get to/from bus ? or …. have you closed the stop that is more accessible somehow/
    what about the footpath and traffic conditions to pedestrians to get to the remaining stops
    is it in good order ? is it still safe in the evening/morning hours (e.g. walking past unlit
    area when you previously did not might deter some commuters esp in the darker winter months).
    Lots and lots of questions like that spring to mind…….
    There would need to be people out their walking these routes to see that does not happen.

    Of course it would not affect all users but a significant portion of users could be greatly affected by it if not done smartly ….

    From a personal perspective myself I am not opposed to an additional walk in the right conditions to save time on a bus
    and shorten my OVERALL commute ……..
    but if its hard to get to the bus, or too far and I have the pram/bags etc and its piddling with rain, or I no longer feel safe etc that would affect
    my usage …….

    1. In the long term the new network will provide most advantages to groups which don’t own or use a car. More frequency will give them more freedom travelling to more places at any time in Auckland. Yes, some transfers will be needed so integrated ticketing should be highest priority. My granny (in Munich) was a master in figuring out where to exchange that you don’t get wet, don’t have to walk too much, where to get on a train to find an empty seat, etc. I have confidence that the people mentioned in the article will be some of the first to use the new network to the full extend.

      1. good – you can approach them and give them a helping hand ………..
        since no-one at AT seems to be offering and they seemed to have realised too
        late and missed out on making a submission too

    2. Note though the second half of the article where the Housing NZ residents were initially very concerned at the removal of stops until the full plan was explained to them at a meeting. Then they were really happy as they realised they would have more frequent services, 7 days a week.

      So it seems at least some people are happy to make that trade off.

      I sympathise with the 487 bus users and they are obviously people for whom even travelling a little further is an isue. However, I do wonder whether their fears on transfers will be allayed when they see the frequency of the new buses. It may offer them a better over all service with more frequent and faster options despite the transfer. There does seem to be an assumption in AKL that transfers are always a problem but that is a legacy of Auckland’s crap PT – it doesnt have to be that way and transfers can be a positive experience.

  17. Bus stop distance decisions should also take into account Aucklands volatile weather. I wouldnt use a bus stop further away in wet weather. The australian data shouldnt be relied upon for this very reason . Why not have more bus stops and more buses not fewer bus stops and have buses stopping in turn at each 4th bus stop. That way you get faster times and more convenient bus stops then we already have. Lets have a bus stop every 150m which would mean every 4th stop would be 600m. The extra cost of buses would be justified by the greater patronage as no one would have an excuse NOT to catch the bus.

      1. Yes the heat can be so bad that the close schools and people don’t go to work. And indeed in winter Melbourne gets far colder and can get far wetter than Auckland (the recent long drought notwithstanding). I’ve seen trams float off their rails down Elizabeth St in flash flooding. Brisbane gets to 35 degrees with 100% humidity, Darwin is worse.

        Auckland has mild and comfortable weather compared to Australia, in fact compared to just about any place we are lacking in weather extremes.

        1. Don’t we have horizontal rain though (combination of rain and wind) that make most umbrellas pretty useless?

          I thought our “unique” climate conditions were a documented reason for why all those mono-clad houses leaked in NZ (and not so much in Aus where the mono-clad stuff came from).

          Sure we don’t have extreme heat or freezing cold/snow, but it can sure be a nasty turn off to using PT when Auckland has a horizontal rain day
          – you don’t feel like walking to the gate, let alone the next PT stop whether it be 200m away or worse 500m away.

          1. Yes – been my experience ! The rain thing ! Having even from the bus stops walked probably 0.5 km from my current
            and previous house home been caught in that a few times – with a stroller (which don’t come with rain covers which is why I always
            carry their jackets and waterproof pants along with all the other paraphernalia)
            even been stopped multiple times by concerned people to say its too wet to walk hop into my car!
            The trains are > 1 km walk – so that’s worse – hence in a rainy day (while I have the option) I would choose the bus
            plus of course the commute time is comparable – trying to think how much commute time saved would make me choose
            the longer walk on a wet day ? Would depend on if it was home (where there was a change of clothes and a shower) or the to work trip I think …………see so from just one persons random musings you can see there is heaps of factors in play in making those choices ………
            would love to see the survey done here – and broken down into age groups etc ………..

          2. They have horizontal rain in Melbourne too, all the main cities in fact. Anywhere you have a port city on the edge of the water you’ll get wind and rain.

          3. all this talk of weather comparisons may be null and void anyway warming and the changing weather patterns
            and the higher incidents of extreme weather events we are having will probably mean going forward weather as we know
            it will change somewhat

        2. having lived and commuted here and Sydney would have to say Aussie hotter yes
          but in terms of variation in one day (over the whole year) Auckland would still win
          on the rain front …….

  18. I live on West Tamaki Road Glen Innes. I have two stops close with two different services. Both hourly at non peak. If further away and frequent I’d walk (Cnr Line Road and West Tamaki Road – Red on map) Having the ability to not run on a timetable to get into the CBD / beach would suit me perfectly. Any idea what frequencies these would run at?

  19. stopping frequency could be a major problem if reduced for the elderly and disabled. I thought part of the appeal of PT was regular stops. Imagine the outcry in London if Boris announced he was going to get rid of half the tube stations as people could ride his bikes between stations.

    1. Stop spacing on the tube ranges from 600 to 3000m. People are happy to walk over 600m to get to a tube station because the service is frequent and dependable.
      We should do the same.

  20. I notice there is no peak bus services between the upper east coast bays and westgate along the motorway.
    Given how much traffic it sees at peak times, and (I assume) how many folks commute from one location to the other, why is there no bus service from say, Constellation bus station to westgate? Is that on the cards, or have I not read the map correctly?

    1. That map only shows the premier Frequent Network routes, i.e. only those routes that run at least every fifteen minutes all day, seven days a week. All the normal routes aren’t on that map, including ones that only run at peak times. Yes there is a bus from Constellation to Westgate. I believe it will depart once every thirty minutes across the day with more frequency at peak times.

  21. Discussion of the same topic on The Atlantic Cities

    The researchers believe that at a walking threshold of 800 meters (roughly half a mile), the Fairfax bus system could eliminate 53 of its 121 bus stops and still provide coverage to 82 percent of the area. Removing the stops would reduce both travel time and operating expenses by about 23 percent; plus, faster service would attract more riders and generate more revenue. And since only 10 percent of the population would lose coverage from the cuts, the researchers see stop removal as a win on balance.

  22. 1 minute saved per stop is a huge overestimate. It probably loses no more than 5 seconds for a bus to come to a complete stop and open the doors, and another 10-15 seconds to close the doors, pull off, and accelerate back to 50 km/h. It’d be a whole minute if you count the time for people to pay and get on, but they’d be paying, and getting on, no matter whether they board at a few main stops or lots of little ones.

    There are a few extra seconds while people walk from the bus shelter to the bus’s front door, but we could solve that by making sure the shelters, real-time signs etc. are actually at the front of the stop. At the moment, many aren’t.

    There’s logic to closing stops, but you’ll get to diminishing returns (or start putting people off) pretty fast. You’re also getting fairly limited returns in the off-peak, where most trips will be skipping the majority of stops anyway.

    1. Let me calculate that for you.

      First, it takes more than 5 seconds to brake for a bus. They can brake that fast, but that is a uncomfortable deceleration, going from 50 to 0 km/h in 5 seconds, assuming a constant deceleration, would be about 3 m/s2, one third of gravity. In general, in public transit, acceleration and deceleration are about 1 m/s2, that’s a comfortable acceleration/deceleration, and it is quite required when you have people standing in the bus, holding on to straps. So slowing down from 50 to 0 and accelerating from 0 to 50 should take about 14 seconds, and it takes 100 meters to do it. As to time spent stopped, let’s say 2 seconds to open the doors, then 2 seconds for each passenger to get onboard, and 2 seconds for the doors to close. Assuming 2 passengers, that would be 8 seconds stopped. All in all, a bus stopping to pick up 2 people will take 32 seconds to travel 200 meters, from the moment it starts decelerating to the moment it starts accelerating.

      In contrast, a bus that doesn’t stop and remains at 50 km/h will travel 200 meters in 14,5 seconds. So each stop delays the bus by 17,5 seconds, but if you assume that the number of riders will be the same, then just counting acceleration/deceleration and opening/closing doors, each stop delays a bus route by 14 seconds. So each stop you can avoid having the bus do will save the route 14 seconds. If you have a 10 km route with 50 stops, 35 on average being used, and you remove 30 (the 15 who are least used and 15 others), you spare the 10-km route 15 stops, so 210 seconds.

      You may question my figures for acceleration/deceleration rates, but I found a source that shows that deceleration is perhaps a bit higher (like 1,1 instead of 1), but acceleration is poorer (around 1,1 up to 16 km/h, then it goes down until it’s 0,7 until 48 km/h) so I’m in the clear. See table 3 (in mph unfortunately, divide by 2,25 for m/s2):

      You are right that the advantage of closing stops is dependent on how used the line is. Which is why an interesting solution is simply having a local line and a limited line following the same route. The limited line has fewer stops and may only be around say from 06:00 to 20:00 during weekdays, whereas the local line is always around. Of course, you need to have a street or road with 2 lanes by direction or more to pull this off, otherwise the limited line is likely to just end up pushing the local in the rear.

      However, I’ll point out that I believe that in dense areas, the high number of stops is a self-fulfilling prophecy. Meaning that the high number of stops implies that when a lot of people use the line, it becomes slow and unreliable. Since the line is slow and unreliable, you annoy people and convince them to take a car instead, reducing the number of riders and so reducing the number of stops where there will actually be people getting on or off. At one point, you will achieve balance, but this balance is one with less riders than there would have been if the line had been faster and more reliable. The local bus without pre-payment, in mixed traffic and with a lot of stops is thus pre-destined by its very characteristics to be a low ridership transit option, if ridership is high, its service becomes crap, pushing away people. It can offer decent speed and reliability, only as long as few people actually ride it.

      1. I applaud the fact you figured that out.

        But you might want to redo the calculation with the average traffic speed in Auckland.

        32 km/h, which you can find somewhere else on this blog.

        1. As a traffic engineer, I better know how to figure it out, that’s the kind of thing I do for a living.

          I get what you mean by referring to the average speed of traffic, but I don’t think it is relevant in the comparison I made. The average speed is the result of traffic lights, stops, and needs to slow down for whatever reason. The maximum speed is probably around 50 km/h on most major arterials I figure, which is exactly where the most heavily used buses are likely to be. The 32 km/h average is the result of vehicles traveling at 50-60 km/h, then having to stop for a bit, then accelerating back to 50-60 for a few hundred meters, then slowing down or stopping, and accelerating again, etc…

          So when I want to compare the time lost stopping and accelerating for a bus, I think it is appropriate to compare with the maximum speed the bus will reach, and not the average speed of traffic.

          That being said, if you want to see a graph of what the average speed of a bus line on a 50 km/h arterial road looks like depending on stop spacing and boardings per kilometer, I have already made that graph.Okay, it’s in French, the vertical axis is average speed in km/h (uh, made a typo, I wrote k/h…), the horizontal axis is the boardings per km, the three lines represent the lines whether stops are spaced every 200 meters, 500 meters or 1 000 meters. You’ll notice that if there is no passenger, the speed of all three lines is about 30 km/h, pretty close to the average of 32 km/h you refer to, which is a result of myself including stops for traffic lights and stops in my calculations, before adding in bus stops. The lines are based on an assumption that the average number of passengers per stop is 2 at least (if I had started at 1, the 500m and 1000m line would be even more advantageous than the 200m line).

          Anyway, enough explanations, here is the graph:

          1. Definite advantage there then, I had to google the formulas 🙂
            I can understand your reasoning for that.
            I would have figured that bus stops being near pedestrian crossing and lights, etc would bias it the other way.

            That graph is a excellent illustration of the asymmetries involved in all this kind of thing.

            Which I think what a lot of the discussions here are about.
            Different people weight them differentyl.

          2. My car shows the average speed (time / distance) on each trip meter (one tripmeter is reset automatically each time I refuel the car, the other keeps going til I manually reset it).

            Typically, the first tripmeter, I use to measure my commute each day which is to Newmarket on arterial roads, the speed it gives me is about 28-30kph at best
            – even when averaged over say 20-30 trips/ over 10+ hours of driving its still pretty consistent month in month out.

            And thats with me not stopping to pick up passengers like a bus, but I am using the same roads as the bus, getting in the same traffic and getting stuck at the same sets of traffic lights.

            So a bus getting 30-32kph average speed shows how little impact that stopping to pick up people actually has when compared to a car driver who doesn’t – in effect its a mere 2-4km of distance “lost” by the bus for each hour of bus operation – when compared a bus and car doing the same trip.

          3. Greg N, I think you’re misreading the graph. The X-axis (on the bottom) is about the number of passengers the bus has to stop and pick up per kilometer it travels. So at 0, of course the bus line has an average speed of 30 km/h, because the bus would then go just as fast as a car traveling on the same road. But the more people the bus needs to pick up, the more its average speed declines. For the standard bus line with stops every 200 meters (the light blue line) if 2 people board the bus every kilometer (that means 40 people on a 20 km route), then the speed is down to about 24 km/h. If 4 people board every kilometer (80 people on a 20 km route), the speed is down to 19 km/h. So stopping and picking up riders has a big impact on the speed of a bus.

            In practice, bus lines with 30 km/h average speed are very rare, they are essentially suburban buses with very low ridership. That’s assuming that they travel on the street and serve all the area they travel on, if they use highways for a section of their route, the speed can be higher, but the calculations are still valid on the section of the route where they travel on streets and have frequent stops.

      2. I said that you lost 5 seconds braking to a stop, not that it took 5 seconds. I was making a similar estimate to you: simply stopping at a stop loses about 15-20 seconds.

        Limited stop services can fix the problem, but you’re effectively halving the frequency by doing so. You need to be running buses every couple of minutes before you’ll save individual passengers enough time from either faster trips (if you go for fewer stops) or shorter walks (if you go for more stops) before having that split is better than either limited-stop only or all-stops only. The only route in Auckland I can think of that has both the frequency and the overtaking opportunity for that to work would be Great North Road from New Lynn to the city.

        1. Oh I see, my bad.

          As to the time advantage, first of all, it’s not necessarily halving frequency. Well, on half the stops, or a third of the stops, the frequency is the same, if you add the two services. Plus, you must consider the advantage for the transit operator. If a bus takes 20% or 30% less time to make its route thanks to having fewer stops, it means it can be reused on the same route or on another quicker. Let’s take a circular route, if a local bus takes 30 minutes to run it (and has no idle time), it can make the route twice in an hour. If the bus is limited and it takes only 20 minutes to make the route (unlikely but possible time gain), then it can make the route thrice in an hour, thus adding 50% of capacity and frequency to it.

          You seem to consider the possibility of people going to transit at random times. I get where that comes from, the importance of spontaneity and whatnot. But for those who actually look at schedules, the limited buses have a huge advantage: they are far more likely to stick to their schedule than frequent, heavily used lines who frequently end up in convoys, one bus 2-3 minutes too late, the other 2-3 minutes too early. Limited buses do not have that problem. And that’s because they probably stop at all their stops anyway, so additional passengers just delay the line by 2 seconds, whereas additional passengers on frequent local lines can delay them by up to 20 seconds if they take the bus on a stop where they are alone.

          That means that schedules for limited buses, even at peak times, are much more likely to be respected than schedules for frequent lines. Which means less waiting time and less chances to miss the bus for people who actually know the schedule of the limited bus. This can be significant, I’m next to a frequent line and I frequently either miss the bus as it was early or wait 2-3 minutes more than I should as it is late (and full). I’ve seen cases of buses that were supposed to be 7 minutes apart running one after the other, the first one 4 minutes late, the second 3 minutes early (and empty). Part of it is the traffic lights, but a large part of it is also the stops.

          1. The specific numbers are hugely important in deciding if it’s worth it. If you go from 200m spacing to 333m spacing over a 10km route, you’ve eliminated 20 stops. Suppose that route takes 30 minutes, one-way. It’s now 25 minutes, if you saved a (fairly massive) 15s x 20 stops = 300 seconds each way. That’s 10 minutes per round-trip, so how much does it increase headways? Imagine that you’re running six buses on the route, which do the round-trip, including layovers, in 1 hour 20 instead of 1 hour 30. Headways were 15 minutes, they’re now… 13 minutes 20 seconds. Good, but an average person now has to walk another 66 metres, at 1m/s walk speed you’ve just lost two-thirds of the benefit. Total time savings per trip = 33 seconds. Does that make up for the inconvenience of the extra walk for people who have difficultly walking, or have luggage or small children etc.?

            Plenty of people do check schedules, but in terms of the actual time you spend, it’s irrelevant. People either naturally want to leave at a specific time, or have appointments at a specific time. Either way, you’re just moving the delay from standing at a bus stop to sitting around at home or work or wherever, waiting until it’s time to leave for a scheduled bus. Modelling waiting times based on random arrivals is the fairest way to model the inconvenience of longer headways.

          2. I disagree that you just “move waiting time around”. Waiting in your home or spending a bit more time in a shop is vastly more pleasant than waiting at a bus stop for a bus that is late. When I take the bus, it makes a massive difference for me. 1 m/s for walking speed is very slow, the average human being walks at around 5 km/h, so about 1,4 m/s. And walking, to my mind, is more pleasant than waiting. I also notice you have omitted the time savings of the greater speed from your calculations, you only compared the headway gain versus the greater walking distance. So the fact is that in your example, the walking/waiting time is in favor of greater stop spacing, and so is greater speed. You shave off a few minutes easily, and that’s just one bus, if you need to transfer to another, double the savings. The result is that the area you can reach in 30 minutes or 45 minutes of travel has become much greater, you have significantly increased mobility and the attractiveness of the bus service.

            Okay, you have increased inconvenience to a few mobility-impaired people. And I do mean a few, mobility-impaired people tend to be gathered in specific places: retirement homes, hospitals, etc… People who struggle having to walk 100 meters more and take transit are unlikely to be living alone independently. You just need to make sure you keep stops near these places, to preserve these stops and you have not made things worse for most of that clientèle. For the rest… Now there’s a difference of vision.

            Some consider transit as essentially a welfare program, for those too poor to own a car or too ill to drive one. In that case, what do they care that the service is slow? It still beats walking, and that’s all that matters. So they just have stops as close as possible to each other and do not care about the average speed of the resulting bus lines. In that vision, yes, inconveniencing a few people is unacceptable.

            Others have a different vision, it is to create a competitive and good quality transit system to act as a separate and decent mobility network apart from cars and roads. This serves not only to increase ridership even amongst the healthy and the car owners, but to orient urban developments, to allow them to be denser by making living without a car easier. That’s my vision. So if I have the choice to increase speeds by 20 or 30% in heavily used corridors and increase the ridership and capacity of lines at the cost of inconveniencing a few people, I would do it without hesitation.

            Of course, my personal preference is to have transit networks made up of a backbone of subway and LRT lines, and have buses serve essentially as feeder lines for the high-quality transit or as ways to fill gaps in areas where population is too low to merit any kind of better transit. In which case, I wouldn’t care much for longer stop spacing on the bus network. But if I’m stuck with fragmentary rail networks and buses everywhere else, then longer stop spacing or limited/local lines running on the same routes are essential to make sure that the network can take higher ridership without crawling like a snail and becoming useless.

          3. Waiting in your home or spending a bit more time in a shop is vastly more pleasant than waiting at a bus stop for a bus that is late.

            Yes, but if you drove, walked, or cycled, you wouldn’t be waiting at all, and if you caught a cab or got a lift, your wait time has just become better in comparison.

            1 m/s for walking speed is very slow, the average human being walks at around 5 km/h, so about 1,4 m/s

            That figure takes into account waiting time at intersections, and is an average for all sorts of pedestrians, in all sorts of temperatures, not just fit adults on clear spring mornings. 3.5km/h = 1m/s is the standard I’ve had recommended to me (I’ve developed walking/cycling/PT journey planner software for my day job).

            I also notice you have omitted the time savings of the greater speed from your calculations

            Yes, sorry, that was very unclear. That was only intended as a response to your comment that “[a bus] can be reused on the same route or on another quicker”. You do get the time savings when you’re on the bus itself.

            And I do mean a few, mobility-impaired people tend to be gathered in specific places: retirement homes, hospitals, etc…

            Not at all. There are loads more people who can walk a bit, but not indefinitely, and live in the community. It’s not about a 100 metre walk, it’s about the effect of an extra 100 metres, and how off-putting that is. Most people already have to walk some distance to get to the road the bus runs on in the first place. It’s also not just one walk – you’ll be making similar walks several times during the day if you’re making a round-trip and walking at each end.

            Some consider transit as essentially a welfare program, for those too poor to own a car or too ill to drive one. In that case, what do they care that the service is slow? It still beats walking, and that’s all that matters.

            Transit competes with cars, but as you say, it also competes with walking (and cycling). No matter what the actual travel speed, once you have to allow time for the bus to arrive, add extra time in case of delays, maybe make a transfer, or if the bus just doesn’t go close enough to your source or destination… the sort of person who’s happy walking any distance to PT is often happy walking all the way to their destination, or will walk further still to a train or a proper, high-quality busway.

            Public transport in Auckland is a million miles from being able to compete with cars for overall travel time. It’s just not even in the same league for long-distance trips. The sort of cities where PT is even comparable (London, New York, Tokyo…) haven’t gotten there by speeding up PT, it’s because they’ve made the smart decision to give up trying to build their way out of traffic congestion, and so average car speeds have become low enough that PT can compete. Even then, it’s because grade-separated trains are competing with cars that have to stop at every intersection. The way PT-heavy cities reduce travel times is by taming sprawl and reducing trip distances, not speeding up the trips themselves.

            We can’t think that we’ll build public transport and have time savings “lure” people from their cars, in a city designed around driving. PT needs to provide an overall value package – being more pleasant, making better use of the time you do spend travelling, saving you money (including by not needing to own a car, or as many cars, and needing less parking), saving you from having to drive, and making it practical to live in a denser neighbourhood that’s walkable in its own right. In that sense, it is a welfare programme – but a welfare programme for everyone, not just those with no options.

            … a backbone of subway and LRT lines, and have buses serve essentially as feeder lines for the high-quality transit or as ways to fill gaps in areas where population is too low to merit any kind of better transit.

            I definitely agree with you there.

            My point isn’t that Auckland’s current stop spacing is perfect. The optimum probably is higher spacing than what we’ve got at present. But we need to take into account that we get diminishing returns pretty quickly, especially if you’re starting to lose passengers put off by the walk, and that for the forseeable future, PT passengers in Auckland are not going to choose to catch buses because they’re strictly faster than cars along the same route.

          4. Wait times becoming better is not to be neglected. But I think that when you know the schedule, it’s not really a waiting time unless you are yourself on a schedule. If you want to go walk downtown after dinner, you go by bus, you check the schedule and see that the first bus is at 13:15, okay, you go then. If you went by bike or by car, you wouldn’t “wait”, but you would probably set an arbitrary time, you can even set one after 13:15.

            “Not at all. There are loads more people who can walk a bit, but not indefinitely, and live in the community. It’s not about a 100 metre walk, it’s about the effect of an extra 100 metres, and how off-putting that is. Most people already have to walk some distance to get to the road the bus runs on in the first place. It’s also not just one walk – you’ll be making similar walks several times during the day if you’re making a round-trip and walking at each end.”

            I do believe most of the mobility-impaired who take transits are gathered in a few places, those in the community who still can take transit are a minority of a minority. These people likely live at some distance from stops, so they already have to walk some distance. If they can manage to walk there, they can probably manage to walk a bit further. If they couldn’t… they wouldn’t be taking transit in the first place.

            “Transit competes with cars, but as you say, it also competes with walking (and cycling).”

            I don’t think transit really competes with walking. It is COMPLEMENTARY to walking. Transit is like a shortcut in a walking trip, you still end up walking before and after. Some places are considered too far to get there on foot, that’s the utility of transit, to bring pedestrians to those places that they wouldn’t go to by foot, not to be an alternative to trips that people would willingly make on foot.

            In my mind, a good transit system is divided like a road network:

            Equivalent to highways (very fast travel between cities, very few access points), there are intercity trains.
            Equivalent to boulevards (fast travel within metropolitan areas, few access points), there is rapid transit like subways, LRTs and even BRTs.
            Equivalent to arterial streets or collector streets (relatively slow travel with many access points), there are frequent bus lines.
            Equivalent to local streets (very low speed, extremely high access), there are feeder bus lines… and walking.

            From my point of view, walking is part of the “transit continuum”.

            ” PT needs to provide an overall value package – being more pleasant, making better use of the time you do spend travelling, saving you money (including by not needing to own a car, or as many cars, and needing less parking), saving you from having to drive, and making it practical to live in a denser neighbourhood that’s walkable in its own right.”

            Except that travel times are an extremely important part of that value package. No matter how cheap you make transit or how dense a neighborhood, having decent travel times in transit is essential for people to rely on it. If going by car takes 30 minutes, but going by bus takes 2 hours, forget it, people are not going to take transit no matter what you do, and if they need to go there, then they will either buy a car (ruining attempts at building walkable communities) or move away from the neighborhood, weakening it. For transit to be useful enough that it makes dense walkable neighborhoods livable, it needs to have sufficient mobility, and that mobility is directly dependent on travel times. We cannot ignore speed. The faster transit goes, the larger the area it will bring together and make walkable.

            Cities with only slow transit may have dense downtown areas, but this area is extremely tiny, it’s hitting an invisible barrier, and that barrier is the slow speed of transit. The core area cannot spread because it is too far by transit from the rest of the dense urban area, even if zoning gets out of the way. Development follows transport axes. People make the choice of where to live depending on what is within acceptable travel distance from them, and that distance is not measured in kilometers or miles but in minutes. If the only way to reach what they need to reach (jobs, businesses, services, etc…) is by car, then they will ask for car-friendly neighborhoods, because they need cars to get where they need to go. If they can reach all the things they want by walking and transit, then they will demand a walkable area that makes their trips comfortable. If you want to spread walkable areas where you live, you need faster transit to tie walkable areas together. It’s not avoidable. And if we can’t expect rapid transit anytime soon, then we need to speed up buses.

            And if you want to convince people to take buses, you need to reduce the travel time penalty of taking transit over cars.

            “The sort of cities where PT is even comparable (London, New York, Tokyo…) haven’t gotten there by speeding up PT, it’s because they’ve made the smart decision to give up trying to build their way out of traffic congestion”

            Actually, they have gotten there by speeding transit. All of the cities you named have rapid transit with average speeds of 30 to 40 km/h, which is faster than what cars can achieve in their street networks. They also have relatively few parking, and if parking is hard to find, that is a major impediment to car travel, plus it makes travel times higher. What they have not done is sped up cars excessively. Transit cannot come close to competing with the speed of freeways, so whenever cities build a network of urban freeways, it’s all over but the crying for transit. Freeways are the worst sprawl inducement that exist. It’s not necessarily about capacity, it’s about speed, when cars acquire a huge speed advantage over transit, transit declines and so does walkability. When cities convert highways to urban boulevards, capacity doesn’t decline, but speed does, and that is sufficient to bring an urban renewal in the area. Capacity only matters insofar as it reduces car speed.

            In New York, local opposition prevented the building of freeways on Manhattan, leaving a very complete freeway network in the area with a huge gaping hole in the center. That’s what saved transit and walkability in NY. If Robert Moses had his way and built the Manhattan freeways, I think New York transit mode share would have collapsed like Chicago’s, which also has a very developed transit system, but at the same time has a lot of urban expressways, which results in transit mode share being significantly lower than New York’s.

            At least, it’s about RELATIVE speed. Speed for speed’s sake leads to sprawl, because then distances increase but travel times do not go down because everything is further away.

            I’ll stop there, I am feeling like I’m coming up with a “Unified theory of transport, development and urbanism” and it’s still in gestation and complicated even in my head. Continuing on this way would have me writing a train of thought essay right here on this page.

      3. Time to decelerate then accelerate from 32km.h at 1 m/s = 2 x

        8.9m/s at 1 m/s2 = 9 seconds to stop

        d = u.u / 2.a

        u = 8.9, a = 1

        d = 79.21 / 2 = 39.5m

        = 18 seconds
        = 80 metres

        Time for boarding can be ignored as constant if you assume patronage is constant (3 people at each of 2 stops vs. 2 people at each of 3 stops for example)

        80 metres at 32k/h takes 9 seconds

        so each additional bus stop would cost the trip 9 seconds + door open/close time, you say 2 seconds each

        that would be a grand total of

        13 seconds.

        13.9 m/s at 1 m/s2 = 14 seconds to stop

        d = 193 /2 = 96.5

        = 28 seconds and 193 metres

        less the 14.5 seconds @ 50k/h + 4 seconds for the doors gives

        17.5 seconds 🙂 Good they agree

        over 15 stops thats

        195 seconds vs 262.5 seconds

        3.25 minutes
        4.4 minutes saved by more than halving the number of stops, which probably increases the inconvenience factor by say 25%?

    2. Funny thing. I actually have a relative (by several different convoluted friend / marriage relationships, got to love small towns) whose job it is to design and install bus stops. Quite surprising the number of elements they have to consider.

Leave a Reply