Hacker News Reader

Comment by tptacek | Hacker News Reader

From this data: almost half of Amtrak's ridership (north of 45%!) occurs on roughly 1000 miles of track, or just 3% of Amtrak's almost 28,000 total miles --- all in the Northeastern US, DC to Boston through NYC and Philly.

Another gigantic bite of Amtrak's ridership comes from short-hop California routes --- particularly LA to San Diego.

The sharp increase in Amtrak ridership that Brookings notice includes the zero riders to 3.3MM riders from the introduction of the Acela service, plus another 2+MM from other Northeast routes, presumably knock-on benefit from the Acela.

Most of Amtrak's route map seems, with this data, to be a pointless waste of money. The "renaissance" is the Acela, and not much more.

View on HN

I think Amtrak's fans are as big of a problem as its detractors. The romanticism of idyllic long haul routes where nobody is in a hurry is not what Amtrak needs as an organization. The real demand is shuttling people along the northeast corridor, and those travelers don't want idyllic. They want on-time. The money saved by shutting down the long haul routes could go a long way toward the maintenance and repair issues that hold back on-time performance in the northeast corridor.

... or fund a Chicago-MKE-Madison-MSP route, which would actually make sense, unlike Wichita-ABQ and Omaha-Denver.

Or you shut down the long haul routes and what support Amtrak has vanishes.

Similar to what tptacek said below, why should a taxpayer in Arizona support a DC-NYC shuttle?

They probably shouldn't, but the DC-NYC route is Amtrak's most lucrative; it is the Amtrak route that requires the least amount of support from the rest of the country.

They shouldn't. The DC-NY route generates an operational surplus, and the states along the route should shoulder the capital costs by issuing bonds, they way they do for any other infrastructure project.

So all the federal dollars should come out of supporting it? I guess then we'd learn what the subsidies really are.

The northeast corridor has a major capital expenditure backlog due to years of poor financing, about $21 billion. On an ongoing basis, the required capex is probably about a billion a year, about half of which can be paid by operational surpluses. The rest should be borne by a group of states, the same way NY/NJ/CT maintain the commuter and airport infrastructure in those states.

I think Amtrak could see the most success by building out to support the cash cow that is the Northeast corridor. That would mean improving service on the Northeast corridor, to improve speeds for Acela, and adding additional higher speed service to connecting cities.

Two good targets for this would be Upstate NY, where you could reduce travel time to Buffalo from 8 hours to 5 hours, and make travel to Chicago and Toronto much less painful. Improving the Philadelphia -- Pittsburgh route would also make a lot of sense too.

Philly to Pittsburgh is longer than DC to NYC, or most of the ridership of the Acela put together, wouldn't serve all that many people, and would be at least a 4.5 hour trip even with the unlikely assumption that you'd get Acela's average speed for the whole trip.

I'm also not sure what building out NE service does to make Chicago more accessible. Chicago is prohibitively far from all of the east coast cities to make NE rail meaningful.

I think you may have misread, improving service along the NYC-Buffalo route, which is part of the route to Chicago, would speed up the trip to Chicago. Shaving off 3 hours, means that someone could get on a train in New York at 6PM, and arrive in Chicago at 9AM, which while for most people isn't competitive with flying, it is a reasonable and comfortable way to make the trip.

No, I understood the subtext; I just don't think a 10+ figure capital expenditure that leaves us with a best-case 15 hour trip from NYC to Chicago is a good use of funds. The person who suggested a high-speed post-security rail route from Penn Station to LaGuardia and JFK had a better idea. The person who called Amtrak SF-LA 12 hour service a "tourist attraction" had the right idea, too: under any imaginable scenario, that's what NYC-Chicago rail would be.

Would that it were otherwise! I live in Chicago and have lots of friends and business associates in NYC. :|

The money saved by shutting down the long haul routes could go a long way toward the maintenance and repair issues that hold back on-time performance in the northeast corridor.

Amtrak agrees and they've been trying to shut down a lot of those routes for some time now. The problem is that Congress won't allow them to.

This tells me that speed increase made a large difference. Most people drive when it's faster than the train. So, when we get our first high-speed rail in this country, in a densely populated region, it should do well.

Cost is a factor too. I'd love to take Amtrak from Los Angeles to the Bay Area, a trip I make a couple times a year.

But the current Amtrak service is incredibly slow and inconvenient in both timing and destination. I just can't justify it, as much as I'd like to.

So I usually drive, occasionally fly. High speed rail (in 2099 or whenever they actually finish it) would be a nice choice. But if it's several times the cost of driving, not including the car I'd have to rent at the other end anyway, it will be a lot less appealing. The drive isn't that long, and even HSR won't be that much shorter.

I'd be a lot more likely to take HSR to Las Vegas, weirdly. There I don't need a car at the other end.

There's a bigger gap between the 75MPH that normal Amtrak trains operate with at peak speed and the 130MPH the Acela does than there is between the Acela and a European high-speed train. Also: high-speed rail is going to go where the Acela already is, and the Acela is already fast enough to be viable and competitive. So, I kind of doubt train technology is going to make a huge difference.

This is incorrect on all accounts.

The Acela doesn't run in California and it doesn't run in Texas.

http://texascentral.com/

Trains in Eurpoe are going to 220mph.

http://www.bloomberg.com/video/take-a-look-at-europe-s-faste...

The average speed between NYC and Boston is something like 79mph. Saying you have a top speed 130mph, and doing it only for a small portion of the trip is quite misleading. At any rate, peak has little meaning. I've been on commuter trains that peaked at 90mph for 5 minutes.

Huh? Of course it doesn't: neither Texas nor California have routes that would be economical for an Acela-like service. LA-SD is a single point-to-point route; neither LA nor SD connect economically with any other California city.

I also don't understand your "peak versus average" speed thing. You're comparing apples-oranges, not me: I'm comparing peak-to-peak. Normal Amtrak trains also don't average 75MPH!

That's the point. Peak speed is really irrelevant on the northeast corridor. The bottleneck is track conditions and tunnel capacity.

Prior to today I had seen this anecdotally on the Keystone Corridor. I live outside Harrisburg, PA and have seen traffic increase dramatically once they upgraded the track to Philadelphia to dedicated "high speed" rail (110mph max) with free WiFi. Looking at the map both Harrisburg and Lancaster have seen half a million new riders in the past decade, all bound in the direction of Philadelphia. Unfortunately the Pittsburgh leg is shared with freight traffic and turns a 3.5 hour journey by car into an impractical and relatively expensive 6 hour train ride. There's a plaque in Pittsburgh commemorating the first rail line from Philly in the 1860s which shaved travel time from 7 days to 8 hours and all I thought upon seeing it was what a shame it was that so little has changed since then.

Not quite true. I'd probably be a bit faster door-to-door from where I live in Massachusetts to Manhattan Midtown. But that's, to my way of thinking, an awful drive. I think that's generally the case for a lot of the Northeast corridor routes that many people take. Train doesn't take much longer and they're not remotely pleasant drives.

What higher speed rail could do in the Northeast is make the New England to the Pennsylvania/DC routes more attractive which don't really work for most cases today. But that's sort of a marginal and it doesn't address the more interesting cases where the economics and travel time math don't really work vs. air (e.g. NYC to Chicago).

You really have to look at the sum of Acela and Northeast Corridor but it was still a substantial increase. The speed increase certainly helped ridership; it's probably worth noting that electrifying north of New Haven (required for Acela) also made the regular regional trains faster because they no longer needed to change engines midway from Boston to New York.

What these speed increases did was they made train travel roughly equal (well, probably within an hour or so) in time to plane for popular routes-BOS to NYP and NYP to points south. And given post-9/11 security, that made the train pretty attractive especially going downtown to downtown. I'd argue though that "faster" isn't an automatic win though. It has to be competitive with air.

The introduction of Acela also created a huge amount of publicity for taking the train rather than the plane

And yet the Pacific Surfliner loses as much money as the Texas Eagle. I suspect that's because of multiple trains per day on the Surfliner.

I'm not sure, though, that the fact that a route loses money necessarily means that route is a pointless waste of money.

Is it not possible that running a government funded service has value even if it does not produce a profit?

What I'm getting at is that profit probably isn't the best way to measure the value of these routes.

As far as the renaissance, ridership is up considerably on most routes. Even if they aren't profitable, they're certainly serving a lot more people than they were.

And, about that: why does Brookings list a 611MM deficit "on routes traveling over 400 miles"? That conveniently excludes a number of non-profitable routes, including the Surfliner.

I think that threshold was chosen because every route longer than 400 miles is losing money. The longest route with an operating profit is 381 miles.

Great point. Does maintenance of roads have value if it does not produce a profit?

The roads serve vastly more people than Amtrak does.

They could also be completely user-funded, if politicians weren't so squeamish about the gas tax and other user fee-like schemes. (Albeit somewhat unfairly, because folks using "profitable" Massachusetts interstate highways would be subsidizing vastly less-traveled roads in Montana.)

Amtrak, outside the NE corridor[1], could not possibly be purely user-funded. Increasing fares would decrease ridership significantly enough to offset any added revenue. Whatever fare rates would maximize revenue couldn't pay for the system.

[1] And I'm skeptical about this, too: Gas taxes pay for the lion's share of capital expenditures to build and repair our interstate highways. Reports like this about Amtrak's "profitability" are limited to operating costs, which are all you need so long as your tracks and trains last forever.

That's because relatively no one takes trains because they are too slow. There are clearly major population areas where fast trains would be quite popular.

Sure. It's just that very few of them are between the 10 largest US cities:

https://news.ycombinator.com/item?id=8713324

Very few train riders are large-city-to-large-city passengers. The most successful train lines may terminate in a large city, but it is the density of the route that matters, not the termini. There are plenty of very successful routes in the world that are between two small cities, or only with 1 large terminus. The LGV Mediternnanee, from a demographic standpoint, isn't much different than a line from Indianapolis to Chicago line, and the Nagano Shinkansen line not much different from a Seattle to Portland line. And yet both of those lines are extremely profitable.

This is mostly due to simple competition...the existence of passenger airliners, regardless of how competitive the train or airline offerings are, reduces market share. I guarantee you that the most profitable ticket sales on California's HSR fairy tale will be between (LA|SF) and (Fresno|Bakersfield|Modesto). Not between LA and SF.

> I suspect that's because of multiple trains per day on the Surfliner.

This is a surprisingly common misconception in the United States. For public transport, the top three determinants of ridership are, in order, frequency, frequency and frequency.

As a simple example, here in the equally long-distance-transport benighted state of Australia, there's a shiny train line that goes from central Sydney to the nation's capital Canberra in a bit over 4 hours, which is quite competitive with driving (~3.5 hours pedal to the metal, realistically 5 with Sydney traffic and breaks etc) and flying (45 min in the air, but several hours of hassle to get to the airport, security, waiting around yadda yadda), and the cost of a return ticket is on the same order as the gas to drive back and forth, and around 3-4x cheaper than flying.

Yet I almost never take the train, opting to drive or fly instead. Why? Because there are three (3) trains a day, leaving at 6 AM (nope), 12 noon (meaning you spend the whole workday on the train) and 6 PM (well, this one's semi-decent, although it's past 10 PM by the time it arrives). And should you miss your train for any reason, you're completely screwed.

If, on the other hand, the trains ran hourly, none of this would be a problem: I could pick a time that suits me, instead of having to fit myself to the trains, and missing one would be an inconvenience, not a catastrophe.

Much of the route map outside of the Northeast has been designed as much for congressional favor as for actual ridership.

Because it's Amtrak's only proper line.

I mean what's the point of running a train between San Francisco (which is actually on the OTHER side of the Bay) to Los Angeles only once a day, taking over 12 Hours!? That's not a real train service. That's a tourist attraction.

It's mostly because, as I noted, taking a train for short trips has become a better option due to the hoopla and bullshit surrounding flying these days.

For long trips though, you're looking at a day of travel vs days of travel. For most people that's not hard to choose between.

My mental model for flight schedules is:

* Add an hour on each side (getting to/through airport, getting from the airport).

* Assume a 2-hour minimum flight no matter what the route.

* Round up to nearest hour.

For routes with a flying time less than 2 hours, I prefer to drive. Amtrak, though, is never an option. My mental model for Amtrak would need to build in the time to complete the trip on a bus.

Reply on news.ycombinator.com