Comment by scottfr | Hacker News Reader

Comment by scottfr | Hacker News Reader

Parent Full thread

Quick back of the envelope calculations. For AC Transit (the primary Bay Area bus system):

  The Annual number of riders is: 54,987,132 [0]
  The average trip is: 3.1 miles [1]
  The budget is: $398.4 Million [2]

This means the cost per passenger mile on AC Transit is: $2.33 (398,400,000/(54,987,132*3.1))

According to [3] the cost of an Uber trip in San Francisco approximately: "$2.20 plus $0.26 per minute plus $1.30 per mile" (I don't use Uber so don't know how accurate this is.

If these numbers are accurate and generalize to other US transportation networks, public bus systems in the US are massively inefficient and are ripe for disruption/better solutions (note that Taxi's which aren't massively VC subsidized, are also ball-park cost competitive with the price of AC Transit on a per mile basis);

(Of course an Uber/Taxi has huge advantages for the user over buses in general, not the least of which include for most points of origin/destination pairs the large time savings compared to a bus.)

[0] http://www.actransit.org/about-us/facts-and-figures/ridershi... [1] http://www.actransit.org/wp-content/uploads/designing_with_t... [2] http://www.actransit.org/about-us/facts-and-figures/budget/ [3] https://www.quora.com/How-much-does-Uber-charge-passengers-f...

(sorry that these numbers may come from different years, don't have a unified source)

Edit: Also when calculating the cost effectiveness of the Uber/Taxi you should divide by the average number of riders. It's probably close to 1, but even something like 1.1 on average would result in a 9% savings on a per passenger mile.

Ok now lets imagine peak hour downtown in any reasonable sized city with 60 private vehicles in place of every bus.

Uber and similar definitely have a role to play in transit but really the system is built around the peak loads which uber alone would deal with poorly for a large place

Or a single subway train being replaced by 750 private vehicles. And through central Stockholm we have more than one such train pass through per minute.

Why 60?

A city bus has around 30 seats. To get as many passenger seats from sedans, you need 10 sedans. Standing is, of course, an option, but personally, if public transit has standees for a nontrivial distance it is woefully underfunded, overcrowded, and failing.

You're essentially saying that most of Tokyo's transit system during commute times is failing?

They literally hire people to shove riders into cars. That has to slow down boarding, which becomes a vicious cycle after you already needed more throughput than you have.

London too, apparently.

Based on personal experience and media reports (but no hard data), I would say yes, it is failing at peak times. In particular, the trains on the commuter lines to and from London.

If a train is so full there are many people forced to stand, is that a failure because customers are poorly served, or a success because demand for the product is high?

It's the same kind of failure as highway congestion: too many people are using something relative to its capacity, because it's really compelling, and as a result it gets worse for all of them.

30 seats, no standees? Yeah...if public transit were designed to waste resources as much as possible - that's what I'd call failure.

It's a waste of resources to avoid having people stand on moving vehicles for extended periods of time?

What else shall we cut? Heating and air conditioning? Running water?

I know how to solve the housing crisis! Just build Army barracks in tents on all the park land. You don't need parks and you don't need more than a cot and a backpack. By your standards, apartment buildings are a waste of resources.

A New Routemaster has a capacity of 87, including 25 standing. This sort of capacity is common in double-decker buses which can run on busy routes.

You cite the average trip cost as being ~$7.2/trip for busses, and for an equally lengthy uber trip, ~$6.2+0.26/min/trip. Unless your Uber drivers drive 3 miles in under 4 minutes reliably (which is probably illegal), how is the bus system massively uncompetitive ?

He is probably saying there should be more economies of scale for the bus. For example, if a diesel bus can get 6 mpg [1] and a honda civic gets 32 mpg [2] then the bus only needs 6 passengers to be more efficient on fuel per passenger-mile.

[1] https://en.wikipedia.org/wiki/Energy_efficiency_in_transport...

[2] http://www.fuelly.com/car/honda/civic

If only fuel cost was the major part of the cost of car transport, or bus transport... Or even significant. As it is, it seems to be between 5 and 10% of total cost.

For example, taking RATP as example (serving Paris Metro Area), in 2015 energy costs where 191 M€(http://rapportannuel2015.ratp.fr/assets/pdf/fr/rapport-finan... page 126) while personnel cost where 2 587 M€...

Yeah to be honest US transit systems are generally appalling on these sort of metrics. They manage to hit everything that makes them super expensive:

1) Often very heavily unionised, driving prices for staff up to eye watering levels, especially if you include retirement benefits

2) Low population density meaning it's hard to get enough fares

3) Paradoxically, often underfunded badly and lurch from one crisis to a next, which is actually really expensive as you have capital assets under-utilised and maintenance costs soar as simple repairs don't get done, resulting in much more expensive repairs later down the line.

This wikipedia article has a good table where you can see some farebox recovery ratios (not really the same stat, but quite interesting): https://en.wikipedia.org/wiki/Farebox_recovery_ratio

Hah I knew there'd be someone blaming unions somewhere in this thread.

US public transport problems have nothing to do with unions.

AC Transit is not the primary Bay Area bus system. The public transit system in the bay area is a huge patchwork of systems, and AC Transit is not considered an especially well run system.

I think Muni and Bart do far better in cost effectiveness calculations.

Muni isn't much better. It's about $6.45 per boarding, which has got to be at least $8-9 per trip, accounting for people who transfer. [1] Average trip distance is 3.4 miles [2], so $2.50 per passenger mile. This is more expensive than UberPOOL. The problem is that if you subsidized Uber like transit is subsidized, ridership would skyrocket since it's so much faster and more comfortable, and the budget would be blown. This could be solved with a big subsidy for lower income people and a small subsidy for higher income people. You could also run pools with bigger vehicles, say 8 passenger van/suvs that would make more stops and take longer but be significantly cheaper. I'm not even sure it would make congestion worse since busses are themselves significant sources of congestion with their large size and slow speeds. The biggest issue would be congestion caused by Ubers constantly stopping on the street. This might be fixable by turning street parking into loading/unloading zones, since you wouldn't need as much parking with everyone taking the cheap Uber to work instead of driving. Obviously when the cars are self-driving, the economics get dramatically better yet.

[1] https://www.sfmta.com/sites/default/files/agendaitems/2016/4...

[2] https://www.sfmta.com/sites/default/files/pdfs/2015/San%20Fr...

You didn't factor in the economic advantage of taking hundreds of thousands of cars off the road during the rush hour.

Compare it to chicken buses in Guatemala and El Salvador, etc, that serve the most remote places in the country and cost pennies for a ride, and the companies make a profit. The chicken bus system is really an economic marvel.

Just a nit, but the SF Muni us the larger of the bus (and general transit) systems in the bay area[1]. Its busses more then double AC ridership and including LRV it more than triples ACT's ridership.

[1]https://www.sfmta.com/news/press-releases/muni-ridership-inc...

Uber is without a doubt the single greatest thing that has ever happened for people who don't have a driving license.

Nope, that's still going to be the bicycle... but check back in again in 200 years and see if uber has caught up.

The horse was fairly popular with people who lacked drivers licenses. Maybe check back in a few millennia.

Reply on news.ycombinator.com