You're probably right. The California Zephyr or the Empire Builder are on my "should do once when the stars align" list but they probably won't be an option forever. I should probably consider doing one of them at some point when the stars get at reasonably in the vicinity.

Google say it takes 31h to drive between the two -- any idea why the Amtrak takes twice that time?

If you're driving alone -- a reasonable assumption if you'd be in the market for a train -- a 31-hour "Google Maps" roadtrip takes 3 days, assuming ~10 hours of "map-time" driving, followed by 6+ hours of civilized sleep.

The biggest constraint is "civilized" sleep, which is best had in a real bed at a real hotel, whose locations are constrained by settlements and whose typical availability hours begin at 15:00 local time and end at 11:00 local time. When you're driving west and passing through timezones, you can gain an extra hour that you can use for rest, while if you're driving east, you're racing doubly against the clock, as you're losing an extra hour you could be sleeping in a hotel at every timezone.

My longest "map-time" drive sprint (not including stops) between sleeps is 19 hours, which was filled with delirium towards the end, and I would never recommend. I've done a 16-hour sprint to similar results, and have done multiple ~12-hour sprints with no issues whatsoever. In my experience, 12 hours is around the upper limit of how much "map-time" distance one can safely drive between sleeps, especially if one is to perform a comparable drive for three consecutive days.

Over the road truckers are limited to 11 hours per day for good reason, your brain thinks falsely it can safely do more, but studies show otherwise.

I have found on my cross country trips that there are plenty of small towns scatters across the US to make finding a hotel within an hour easy (though I'm sure you can find exceptions). I have also found a lot of interesting "small town" parks and culture because I forced myself to stop and find something (anything!) to do after 11 hours on the road.

There are very few parts of the U.S. in which you'd travel more than an hourr or two between hotels. Virtually all towns have at least some services.

Nevada, Wyoming, Montana, possibly the Dakotas, are principle exceptions in the lower 48, though this means an upper bound on this interval, at least along principle rotes. An example is the 75 miles along I-94 between Buffalo and Cheyenne, Wyoming, with no highway exits or onramps at all. There are, however, services, at either end.

Alaska is a domain unto itself.

Yes, but trains don't need to sleep. I'm not saying I'd recommend driving 31h at a time, I'm saying I don't understand why that would be so much faster.

Plenty of stops and built-in delays with overlaid industrial trains. Additionally, you aren't going generally as fast as a car can, especially on long stretches of rural highways.

Driving you don't need to pull over to the side of the road for 2h and let semis pass you.

On the train, you will often have to pull over to let freight trains pass you.

American rail is built for cargo, so there are no sections as fast as a freeway.

Also there are several sections that are particularly slow. The track through the Rockies is the original 1869 route, with original curvy single track and original tunnels. Speed there is around 20mph. Sometimes you come close to I-70, where cars are going 60-80mph. The Sierras also have slow sections. That said, those were my favorite parts of the trip.

This is mostly true, but in fact the route currently used by the California Zephyr through the Rockies (Salt Lake City - Soldier Summit - Moffat Tunnel - Denver) was built by the Denver and Rio Grande Western Railroad or its predecessors [1]; the section in question was completed no earlier than 1883.

The more famous railroad finished in 1869 takes a more northerly route, through Ogden instead of Provo, and through Rawlins, Laramie, and Cheyenne in Wyoming instead of adjacent to I-70 through Grand Junction and Denver.

The mountainous sections of the Zephyr, specifically the crossings of the Rockies and the Sierras, are indeed the least competitive vs. driving. Based off of the official timetable [2], the Zephyr takes:

~9 hours between Chicago and Omaha vs. 7.5 hours driving

~8.5 hours between Omaha and Denver vs. 7.5 hours driving

~15 hours between Denver and Salt Lake City vs. 8 hours driving

9 hours between Salt Lake City and Reno vs. 7.5 hours driving

7.5 hours between Reno and Emeryville vs. hours 3.5 driving

Lengthening the driving time to include breaks, the train is viable between Chicago and Denver, and between Salt Lake and Reno, but is nearly twice as slow as driving on other parts of the route.

[1] https://en.wikipedia.org/wiki/Denver_and_Rio_Grande_Western_... [2] https://www.amtrak.com/content/dam/projects/dotcom/english/p...

Stating the obvious here, I know, but Chicago-Denver has so many flights it's not even funny, United, American, Spirit, Southwest, Frontier, well over twenty flights any day.

The nominal max speed of Amtrak trains is about 79mph [1]. When I was traveling up I-57 in IL (speed limit 70mph), the Amtrak train was going almost identically in speed.

Most of the class I railroad track was built to handle high speed (by early 20th century standards) traffic, so trains can generally reach 70-90mph (although not so much in suburban or mountainous areas) with little modification. The difficulty is that there's no economic reason to push freight trains to high speed, and mixing higher-speed passenger and low-speed freight creates scheduling conflicts, so passenger speed tends not to hit its top speeds.

[1] On freight track, that is. The NEC, which is all passenger traffic and all-electric, has higher nominal speed: even the "slow" NE regionals will hit 100mph on the empty track in MD.

To add to this, the FRA regulates track speed based upon condition of the track and some other requirements. Track condition is typically quite good (rail companies really want to avoid derailments), but one of the biggest impediments is grade crossings. The faster the track, the more extensive protection needed to keep trains and cars separate. Table 42 [1] here gives the requirements. Most of the mainline track is Class 4. You'll notice to hit 125 mph every grade crossing needs a full barrier between the car and train. And faster that that, complete grade separation (ie, bridge). US rail was not designed with this in mind and while they close a grade crossing any chance they get, they're still a major hindrance to high speed operation.

[1] https://safety.fhwa.dot.gov/hsip/xings/com_roaduser/07010/se...

There aren't no sections, but there are certainly very few that beat 70mph...for example, IIRC the Amtrak Downeaster hits 70-80 at at least one point (the speed limit of the train is 79mph throughout the route)...for like a mile or two as it crosses the Scarborough marsh ;)

Driving in shifts for 31h is pretty awful. It’s really an apples to oranges comparison. That’s really a 3 day trip, especially if you’re targeting hotel check in times.

The train is generally pleasant and relaxing. I used to do Albany->NYC a few times a month. It takes less time in that case, but I always looked forward to that trip to work, reflect or just observe.

The problem with the trains is inconsistent times. They share track with freight in many routes and there is little rhyme or reason for many trips. The error bar for arrival for a Boston->Buffalo trip is 3-4 hours, easy.

I'm not asking about the feasibility of driving nonstop, I'm asking why the average speed of the train is significantly slower.

3 days is a relatively aggressive schedule for the road trip. The train makes up for its low average speed by not needing to sleep overnight.

The real problem is that the cost is greater than or equal to a 4-hour flight.

Why is train travel so much more expensive? Fewer tickets being sold comparatively?

Also because it's so slow. You're tying up a lot of personnel on the train, as well as capital equipment, for multiple days, rather than hours for a plane. With the same personnel (which cost about as much as pilots/cabin crew), you could run ~10 equivalent flights. Capital equipment is cheaper for rail than in the air, but the portion of rail capex/opex (either directly, or usually via fees to freight lines) is also high.

This video is a good analysis. https://www.youtube.com/watch?v=fwjwePe-HmA

Also, maintaining tracks is expensive, especially in a large country like USA, as compared to just maintaining two airports.

Kinda makes me wonder how much money is spent on road maintenance that we are not thinking as a cost while driving.

> Kinda makes me wonder how much money is spent on road maintenance that we are not thinking as a cost while driving

Here in Japan all the highways are privatized and are supposed to be self-funded by tolls. Where I live pretty much any road trip over 45 minutes you're likely to hit a toll.

All tallied up (gas+tolls), driving 3-4 hours costs about as much as taking the 1hr 40minute bullet train (which isn't cheap - about $90 one-way). Driving only becomes economical when you're 3+ people in the car.

I imagine there would be riots in the US if the interstates all started charging tolls.

Amtrak doesn't own most of the track they use. They pay track usage payments to the freight railroads that actually own the track. In FY 2009 that was $115.4 million (3.3% of total operation cost). Cannot find the current numbers. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&c...

$115.4M seems really cheap. You couldn't lay too many miles of track for that, especially if you're buying land rights along the way.

Did the interstate highway system have to buy land rights?

The government used it in Eminent domain power to acquire the land (which if owned was paid for). The interstate system was built by the government, so it’s a totally different situation as the railroads built and maintain the track Amtrack is using.

I guess that then the solution is obvious.

It doesn’t seem like a good use of eminent domain to build a slower and less capable transportation network when airline service is already ubiquitous. “I prefer to spend more for a slower journey because I like trains” is a weak justification for the use of such an invasive power.

I'm not sure what is obvious. The government pays rent on something it didn't build. Cargo is vital to so many sectors of the economy, and passenger trains are not.

For shorter routes, it's the inverse; I go from STL to CHI (and back) often, and it's $30-40 for a pretty enjoyable short trip; 5-6 hours (compared to driving (4-5 hours), or flying (1 hour, but 2.5+ with airport time and stress included). Flights are usually $90+.

I understand the root issues is Amtrack unions. Amtrack staffing is one employee to four passengers. I dont know what the ratio is for airlines but it must be lower. This video does a pretty good explonation i think: https://m.youtube.com/watch?v=fwjwePe-HmA

It's certainly a labor issue, but can you explain how it's the union's fault? Staffing a train for 30+ hours is 10x more expensive than staffing an airplane for 3. It's not the number of workers, but the time cost of workers. Don't forget to account for the ground crew when discussing planes.

> Amtrack staffing is one employee to four passengers

I've been on Amtrak trains and have seen nothing like that number of employees. My impression is that 1 in 40 or 80 might be closer.

Wendover Productions has a good video that explains this and many other reasons why trains suck in the US

https://www.youtube.com/watch?v=mbEfzuCLoAQ

Spoiler: freight companies own most of the track

Track work, waiting on local traffic (especially commuter), train police pulling someone off the train, etc.

You can tell the people who've never taken the train before when they start asking why it's stopped.

I say this out of love, to set expectations. Cross country, if you want to get somewhere ASAFP take a plane. If you want to enjoy a leisurely week+, drive. If you're willing to spend a few days for the scenery of a road trip without having to do the logistics, take the train.

The timetable is set partly so that you sleep during the boring bits and have daylight in the mountains.

At least it was when I did Chicago -> SF a few years ago. It was great, after the prairies you woke up in Denver & lots more people got on, spent a day going through the Rockies, sleep, and then a day in the Sierras. (I'm pretty sure we spent some of the night parked in a siding.)

Probably having to yield for other trains.

I see you have never taken Amtrak before :)

I’ve not taken the train before, but I’ve heard that the experience is quite different in a good way. No traffic. No hustle and bustle.

The rights-of-way are shared with, and subservient to, freight rail, and the roads themselves are owned by the freight lines.

Top speed is limited to 79 mph, and is lower in many areas.

As with software development, delays tend to be additive, and not made up.

Any service disruption (track, equipment, weather, accidents, crews) means hours, occasionally (though fairly rarely) days of delays.

33 stops.

I do live in a bad spot for trains (Oklahoma) however they keep trying to build more rails (which make no sense in our region).. to determine how much sense they made to me personally I studied a trip to chicago from oklahoma city.

Turns out I couldn't find any destination where trains beat driving or flying in the metrics of time or cost even.

I do get that trains work in some regions in the US ~ but we should define those regions or what makes a region good for rail and focus there. Amtrak to me is one of the largest scams the government does - it's a constant money losing operation. It takes in so much money - and loses money on every thing it sells from the tickets to the drinks and food.

Building roads is also a money losing operation. Complain about something else. There's plenty here.

Reconstructing an existing road is even worse. Want to repair a failing street? OK, but only if you make every aspect of the supporting infrastructure ADA compliant with the most current laws. We are legislating ourselves out of modern infrastructure.

Or, just maybe, it’s good to make infrastructure accessible?

He's got an unpopular opinion (and calling out the ADA specifically is probably the worst of the examples that can be chosen) but its worth a thought. When the choice is forced to be "perfect, or none at all", a possible outcome is none at all.

On the balance, I find most of these rules to be acceptable most of the time, but it should always be considered that it can be taken too far.

Regulation is a commitment that we prefer "none at all" to the status quo. No more, no less. We talk about regulation creating desirable outcomes, but that's an oversimplification - the ADA does not hire or conscript anyone into the business of creating accessible infrastructure. It threatens that if they won't make their infrastructure accessible, they can't have any. The threat of "none at all" is the whole mechanism.

Which is just fine so long as we remember that none at all is inaccessible to everyone.

Is opposing accessibility the hill you want to die on?

Do you have a ballpark for the financial and human cost of making our basic infrastructure not-accessible?

A larger percentage of people who pay taxes use roads than the percentage that pay for trains and don't use them.

Only place trains win are very specific regions and for cargo

Umm, the purpose of government is not to make money, it is to serve the public. While you can make arguments whether or not trains are where we should put public money, but you can't use the argument that 'it loses money'.

Discontinuing Amtrak's long-distance services -- which are clearly not competitive in terms of ticket price or travel time with alternatives -- comes up time and time again, by members of Congress and commentators alike.

However, successive directors of Amtrak have always opposed the discontinuation of long-distance routes. Congress could likely force the issue, but it's interesting that Amtrak leadership always leans in favor of them.

There are some routes in the northeast that Amtrak turns a profit on [1][2][3], and some others are subsidized by states enough to still make them worthwhile to operate (CA, IL, NC). But if the long-distance routes were defunded, they'd likely never come back under the Amtrak banner, and significant layoffs would ensue.

[1] https://news.ycombinator.com/item?id=15968720#15976146 [2] http://reasonrail.blogspot.com/2012/12/amtrak-routes-by-2012... [3] http://www.riarp.org/blog/2014/11/25/september-amtrak-report

A lot of the issue, I believe, is that those long-distance services cater to a lot of small towns that have no airport or good bus service. So Senators don't want to lose that service in their state. And they fight to keep the service, even if they are conservative.

I'm still surprised that Oklahoma City to Fort Worth route exists. It's 2-3 cars, and stops in such exciting places as Gainesville, TX. (sarcasm definitely intended.)

Yet every day, there it goes. Bypassing a city like Denton, TX where a lot of people live and commute to Dallas or Fort Worth. The opportunity for commuter rail is there... it's just not happening, and I wish that it would.

It often runs a bit late in the morning as well.

A few reasons: the train has to slow down and then get back up to speed for each stop; there are stops in a few cities for refueling / shift change / restocking onboard supplies. These stops take a lot longer than just filling up your gas tank on a road trip, anywhere from 1 to 3 hours. And the article points out that Amtrak is lower priority than freight when track conflicts occur, so you will spend some time sitting on the prairie waiting for a freight train to go by.

Realistically, nobody's making that drive in 31 hours anyway without a lot of caffeine and breaking the speed limit.

With that much caffeine you'd be making a lot of pit stops.

I did do it in 40 hours and some change, but with one speeding ticket in Utah, and by the time I got home I knew I ain't doing that on "google schedule" again.