Almost a year ago, I told the story of Los Angeles line 305, a diagonal line that ran every 40 minutes zigzagging across the city's high frequency grid, between Watts and Beverly Hills. The line was so infrequent, and the surrounding grid service so frequent, that if you just missed the 305 it was faster to take the frequent grid routes and transfer than to wait for the next 305.
The 305 comes up often in my presentations because it's such a useful example of symbolic transit. The purpose of the line was not to be useful to very many people, but rather to announce, as a matter of symbolism, that "we run from Watts to Beverly Hills"! That's certainly how Jennifer Medina of the New York times described it, when it was first proposed to be cut in 2011. The Times's graphic:
The NYT headline of the time even claimed that cutting this line would "make a long bus commute longer," which was factually untrue if you count waiting time as part of travel time.
Well, a year later, LA Metro is finally cutting the 305. Very few people will experience any loss of travel time as a result, but the system will be simpler, more frequent, and ultimately, more liberating for anyone who wants to get where they're going. Of course, like such worthily deleted lines as San Francisco's 26-Valencia, the 305 will still be useful as a parable!
This was an unfair comparison then and it still is now.
1. You assert that waiting time for the 305 was a big problem. Presumably because the bus might sometimes leave a stop early (so you would have ‘just missed it’). Well, that’s going to happen with the transfer too – except it’s far more likely, because you have two unreliable buses joining together rather than one, and there’s nothing you can do to improve your odds (like show up 10 minutes early to the old 305).
2. You assert that frequency builds ridership better than directness. This is nonsense in most cities where modeshare is this low – you are not competing against other buses or against walking; you are competing against a car the prospective passenger owns and currently drives to work.
Frequent but slow is far worse than infrequent but fast – as long as the reliability is roughly the same in both cases – and this is from direct experience. Years ago here in Austin, I used to use a bus as a boost for a bike commute – and the same stop was served by the 982 – express – and the 3 – frequent local. About 5% of the time, I’d arrive too late to catch the 982 – and it had a 30 minute wait for the next one. I’d take the #3 in those cases because I had no other choice – but it was so slow that were I not using it as an assist to allow me to bike all the way home in the afternoon, I would literally never have taken it. The 982 was close enough in speed to my car, though, that I actually took it once in a while even when not biking.
If I remember correctly, travel times were something like 15 minutes versus 40 – so even if I waited an extra half hour for the next 982, it’d still be only 5 minutes slower on the times I actually missed it (and on the 95% of the time I didn’t miss it, 25 minutes faster).
M1EK–what was the stopping frequency on the 3? Was it one of those busses that stopped every other block? How does that compare with LA’s Metro Rapid busses?
@M1EK, this isn’t really a direct comparison. In the case of the 982 in Austin, the bus took the freeway for a considerable part of the distance, so I assume the choice was between a slow but frequent service and fast and infrequent service, both direct to your destination. The 305 in LA runs on surface streets, at the same speed as a rapid bus (well, slightly slower since it makes so many turns), so the choice is between a slow and infrequent service direct to your destination, or two slow but frequent services that require a connection. In this case, I think the travel time comes out to very nearly the same, unless you’re going from one end to the other, in which case taking the Blue Line saves you 20 minutes even with a transfer (to the 720, which runs every 2 minutes, unlike the 3 in Austin, which seems to run every 20 minutes). But in general, with the 305 or a pair of grid buses, your travel time is likely to be much slower than a car, because cars can use freeways, which make things appreciably faster. Grade separation can have this effect on transit too, as you can see with the Blue Line example above.
I don’t think Walker’s point is that this service wasn’t useful to some people for whom it managed to fit their exact travel needs. But you can find that for almost any route. The question is what is the best way to offer service given limited budgets, etc. etc.
This route is redundant in that there are already services that get one from one end to the other in pretty similar time. And it is better to have the larger grid network be fast and frequent than to spend resources on odd routes like this that do not expand overall mobility as much.
Of course, this route had riders and it probably worked well for many of them. But I think the point is that it’s better to invest in a grid based fast & frequent service than routes that serve somewhat obscure point to point destinations.
One critical issue in a system like Los Angeles:
There are no free or low-cost transfers. Thus, if you are a low income rider who cannot put aside even enough money for a day pass (for example, a day laborer or domestic worker who is paid at the end of each work day), it is quite a bit cheaper to have a single seat ride than to have to pay for two rides coming and going.
This cost cutting measure might contribute to the strength of the overall network. In other contexts, transit planners also need to know when and where network bypass (ie direct service) is justified or desirable in the efficient provision of network services. Slavish adherence to network structures that don’t produce ridership or revenue results is not a great way to ensure the ongoing success of transit systems.
I don’t see the development as positive as Jarrett does, that L.A. will be better off without the route, but I do think Metro was right to cut its losses with Line 305.
First off, Line 305 would have never been an issue if riders would just have used the bus. It ran for more than a decade, and it initally had 30 minute service, then had its frequency and span reduced because too few riders took it.
Second, even the revealed ridership data undercut the Liberal Morality Play angle of the New York Times story, as well as the L.A. Times carbon copy of the same story a few months later. The journalists had fully invested in the narrative of “underclass bus riders losing their only connection to a better life,” but the anecdotes don’t really pan out.
In another thread on this site, another poster found the Stop Dots diagram of Line 305 and linked to it. Line 305 didn’t even serve its purpose of bringing South L.A. residents to the Westside. If it did, the dots would be larger on the segment north of I-10 and smaller but dispersed in South L.A.
Instead, the dots are very small north of I-10 and large in South L.A. This would imply that riders were just using the bus as a local circulator, or simply got on it because it was the first one to show up. (That stair-step pattern in South L.A. duplicates several high-frequency lines, so riders are not losing any transit access).
The line got 25 passengers an hour, averaged over all trips, including the low ridership night and weekend ones. The constituency was not vocal or politically connected. Although every trip on the 305 can be made with two buses some of the routes, like the Sunset and San Vicente buses, are 30-40 minute service routes and not much better than a 40-60 minute Line 305.
Look at the 442, another route that completely duplicates other MTA service, which has been on the cut list by staff for almost a decade now, but saved due to political meddling. It doesn’t serve affluent communities – connecting Hawthorne, Inglewood, and South Central to Downtown LA – but it does serve a politically potent demographic of African American government workers who know how to work the system.
I would generally agree with Jarrett, I prefer more frequent connecting service over an infrequent direct line. Where I live there is a direct line to a popular shopping center that runs every hour and two lines that connect to the same shopping center that run roughly every 15 minutes. I can be there in the same travel time without having to wait for another direct bus.
anonymouse, both the buses in Austin I referred to were slower than my car – but the infrequent-but-fast route was only a little bit slower (travel time only); while the frequent-but-slow route was much slower.
Perhaps in all of this people should identify a target demographic they’re trying to ‘serve’ here. I can see the logic in frequent but slow with connections if you don’t have and will never be able to get a car; but that’s kind of limiting your market (especially in cities like LA!).
The target demographic of the 305 was to connect low income communities in South Central LA with job on the westside and UCLA, while avoiding Downtown. There was a whole host of these lines: the 576 (Vernon Avenue – Pacific Palisades), the 422 (USC to Ventura County reverse commute), etc.
The replacement line 30/330 on San Vicente is not that frequent – 30 minute peak service, 40 minute off peak service to replace 3 buses an hour during the base (305 + 550).
The problem with LA is that it’s very difficult for a bus on surface streets in mixed traffic to compete with cars on surface streets, and nearly impossible to compete with cars on freeways. You could run buses on freeways, of course, but freeways tend to not actually be where people want to go, so you end up with effectively point-to-point routes connected by a freeway segment, which limits ridership, and doesn’t work as well without a strong center. The other possibilities require some amount of infrastructure: either various priority measures for the buses, up to separate lanes, which could be politically and practically difficult, since most LA boulevards have four lanes plus parking, so you’d end up cutting significant parking or traffic capacity, which people obviously won’t be happy about. And even then, the bus is likely to be competitive with driving on surface streets, but not with freeways. The other alternative is to build new grade-separated lines, whether subways or light rail with absolute preemption: these can actually be faster than driving on surface streets, and competitive with freeways.
Cutting direct routes and forcing people to change is unhelpful to travellers, except where all of the following conditions are met:
a) There are cheap system-wide tickets/passes valid on all operators’ services;
b) Services on each segment are frequent (at least every 15 minutes) and reliable;
c) The interchange point is user-friendly.
In the UK, these criteria are in general only fulfilled in inner London and partially in Manchester and certain other PTE areas. Elsewhere, population densities are typically too low to support frequent services, and tickets are rarely interavailable between operators/transport modes.
I believe they intend to cut the 550 also, which is a similar diagonal route from San Pedro (I think?) by USC to West Hollywood. It seems that they ought to run some sort of relatively frequent bus that takes advantage of the unique diagonal of San Vicente, but I don’t know what is best.
“either various priority measures for the buses, up to separate lanes, which could be politically and practically difficult, since most LA boulevards have four lanes plus parking, so you’d end up cutting significant parking or traffic capacity, which people obviously won’t be happy about.”
Most LA boulevards have four lanes plus two parking lanes, as you say. This means there is *plenty of room* for bus lanes.
It’s now well-documented that the “four lane plus” boulevards are the most dangerous and unsafe type of street there is; and they don’t add much throughput over one-lane-each-way streets because of weaving.
A pair of bus lanes would still leave a two-way street with parking, with the same speeds, close enough to the same car throughput, but a lot safer. (Alternatively, you could reinstate the light rail lines which used to run in the middle of some of these streets… sigh.)
Right, the 550 was eliminated north of USC during peak hours, and north of Harbor Gateway Transit Center on middays and weekends. San Vicente goes from four buses an hour during the peak period to two. And just ten years ago, both the 305 and 550 had 20 minute peak and 30 minute midday service. So in the span of ten years, MTA has gone to six buses an hour during the peak and four buses an hour during the midday to two buses an hour peak and 1.5 buses an hour off peak.
Incidentally, all of these were Bus Riders Union routes which were implemented after the Consent Decree. Most of the BRU routes have now been eliminated – the 577, 603, and 605 are the only ones left.
@Nathanael, that’s pretty interesting about the capacity of boulevards. Do you have any links to the research on this? Having something to cite would make advocating for bus lanes that much easier.
@calwatch I kind of wonder whether that’s the MTA’s resistance to change imposed from outside, or just because service planning by court order isn’t the most effective kind.
“Incidentally, all of these were Bus Riders Union routes which were implemented after the Consent Decree. Most of the BRU routes have now been eliminated – the 577, 603, and 605 are the only ones left.”
So from the recalcitrant staff POV mission almost accomplished.
Sigh. Unfortunately, I wasn’t keeping references when I ran across the boulevard vs. everything else accident rate study.
If it helps you find it, I believe I was looking at *pedestrian* injury rates and only *within urban areas*.
The high rate of intersections combined with the high volume of cars was speculated to be the cause.
There are probably confounding factors, of course.
Oh, wait, you asked about capacity, not the accident rate. It’s actually easier to find capacity references.
I was kind of cheating when I said “nearly the same rate”, because two-lane-each-way boulevards *do* have a significantly higher car-moving capacity than one-lane-each-way roads.
It’s just a lot less than twice the capacity. I’ve read numbers from 1.5 times to 1.8 times, I believe? Sources disagree. Weaving motions are the key to this.
The difference in *people-moving capacity* from replacing a car lane with a bus lane is more than made up by the speed improvement in the buses moving people in bus lanes, at least when the buses are full (as they are in LA).
Ummm…..how is the system going to be “more frequent” now than with this cut line? Unless they’re putting the buses on this line that they cut onto the frequent lines you mention (which you don’t say they are).
Chris. Travel paths by most existing grid lines are vastly more frequent than the 305. Jarrett
Whether transit planners like it or not, users often show a preference for a single seat ride as opposed to potentially faster trips involving a transfer because every transfer increases the chance of a long wait for the connecting route. This preference is aggravated by transfer surcharges.
>>Chris. Travel paths by most existing grid lines are vastly more frequent than the 305. Jarrett<< I understand that. But those existing grid lines are no faster after this line was cut than they were before this line was cut. Riders already had the option of taking those grid routes if they wanted. As far as I can tell, all they've done is remove a route. That doesn't make the remaining routes any faster or more frequent than they were previously. So unless the buses from this cut route are being redirected to the grid routes to provide more frequent service, how are riders better off now than they were before? It seems they just have one less option.
I think the key is that the productivity of the 305 was only 25 passengers per hour, which is less than half of the Metro average. I suspect if it averaged 50-60 passengers per hour than it would have been kept, or if the line served an area that had no other bus service.
An interesting point brought up in the comments was the idea that one reason the 305 cut harmed low income riders was because since LA Metro does not issue transfers their day transportation cost went up from $3 (1.50 each way) to $6. Personally, as a transit planner I don’t think we should be planning transit from the perspective of how much in fares people are paying, since taking into account the frequently nonsensical fare payment structures would likely result in a non-efficient network, but I’d like to see Jarrett comment on that because I don’t remember reading about it in his book or on this blog.
C. MacKechnie.
Agree completely that fare structure shouldn't distort network design, and I like flat and zone systems because if they're well designed, they don't.
Don't agree that productivity was the sole cause of 305's demise. It wasn't just unproductive, it was redundant, which is a different thing.
LA’s fare structure is … whatever. I don’t know why they do it that way. It is something LA has been getting wrong for a long time.
“Cutting direct routes and forcing people to change is unhelpful to travellers, except where all of the following conditions are met:
a) There are cheap system-wide tickets/passes valid on all operators’ services;
b) Services on each segment are frequent (at least every 15 minutes) and reliable;
c) The interchange point is user-friendly.”
Actually, all of these were true along the Line 305 route (though not in areas to the west of the terminus).
One additional shortcoming of Line 305 that hindered its ability to provide a more compelling service was all the left turns along the route. Each left turn could cost 1-2 signal cycles, significantly adding to the total travel time. Even if a Rapid and a Line 305 bus showed up at the same time, taking two Rapids to your destination might not take that much longer.