[nickbauman]

No it doesn't have a world class freight rail network. The North American rail infra is incredibly primitive. Most of it is "dark territory" (no track sensors), unlike Europe. I used to write rail automation software for a German firm. They were appalled at the state of affairs here. One of the most lucrative rail systems in the US had an average speed of their trains in the single digits MPH!

[twawaaay]

I think the term "world class" is unfortunate with connection to freight rail networks.

Freight does not need to travel super fast or super high tech. What it needs is to be able to travel everywhere at high throughput and cheaply. US is doing quite well in that regard.

[bobthepanda]

High throughput comes with a caveat, since it’s high throughput given the existing poor conditions.

The US used to have much more tracks, but the private railroads stripped a lot of them as far as they could get away with. There are lines that were four-tracked or were electrified that have now been reduced to unelectrified single track, so you now have a much more sluggish, polluting and congested railroad, and on top of that much is poorly maintained to save money.

—-

Also a lot of the freight is bulk freight like coal. This has led to some interesting dynamics where freight railroads oppose coal plant closures, because they will lose a major source of tonnage.

[panick21_]

> Freight does not need to travel super fast or super high tech.

That depends on what you want to ship. In the US the train just gave up on many other class of freight. Yes, large scale slow bulk transport doesn't need speed, other things might.

[seanmcdirmid]

> Most of it is "dark territory" (no track sensors), unlike Europe.

Have you seen the USA? The places where they lack track sensors are basically out in the middle of nowhere with no one around for miles.

> One of the most lucrative rail systems in the US had an average speed of their trains in the single digits MPH!

That really isn't bad for freight. They optimize freight for throughput, not latency (something passenger rail is more concerned with).

[sudosysgen]

Track sensors are especially useful in they middle of nowhere.

And there are plenty of latency sensitive applications for freight rail which are developed in other places. They don't make sense in the US because the capability isn't there, not because there's no market for it.

[seanmcdirmid]

> They don't make sense in the US because the capability isn't there, not because there's no market for it.

There really isn't. Freight companies are responsible for maintaining investing in the rail, and if it doesn't make them money, they aren't going to put it there. Heck, a lot of places are single rail (meaning, no two way traffic at the same time), because it doesn't really make sense to dump more money into an extra set of tracks in those places.

[sudosysgen]

Again, you're conflating things. There is a market for low latency rail freight. The rail companies find that it's better to keep the rail as is and invest the profits somewhere else. That doesn't mean that the market doesn't exist.

The correct approach is for low latency rail freight to operate on passenger rail systems which already have the necessary speeds and flexibilities. This is structurally unfeasible in the US but it's still definitely a market that better rail systems can service at no extra cost.

[seanmcdirmid]

The freight companies own the railway, they optimize the rails for freight, which is why we move much more freight by train than Europe. Passenger service is something they do for the federal government subsidy and nothing more.

And actually, sharing tracks between passenger and freight service is something that they don't really do in Europe. Because they share tracks, American passenger trains have to build at a weight on part with freight trains. Most lines in Europe separate out passenger and freight service lines so they can run lighter trains for passenger service.

[sudosysgen]

A big part of why the US moves more by rail is simply because it moves more goods overland than in Europe.

Low latency freight for smaller, high value items is often done on passenger lines (or even passenger trains) because it doesn't put any scheduling pressure on passenger service.

As far as use of freight in Europe, the elephant in the room is that the EU uses a lot more sea freight than the US. Indeed, while the modal split for EU trucking is around 50%, it's around 70% in the US, and it thus seems clear that the real reason that there is less rail shipping within Europe is because there is much more competition from maritime shipping.

[nickbauman]

This is a perfect example of path-dependent policy. Because the decision was to select for freight, other kinds of rail are "unfeasible". I'm using scare quotes to suggest that it's not really unfeasible, it's a choice.

And since trucking capacity has been maxed out in the US for quite a while (Amazon is the big mover for this situation, even before covid), you can bet this path dependence is biting us now.

https://en.wikipedia.org/wiki/Path_dependence

[panick21_]

> And actually, sharing tracks between passenger and freight service is something that they don't really do in Europe.

Happens a lot in Switzerland.

[cptcobalt]

I'm surprised by the general opposition to your comment. I agree. US transit infrastructure, including rail, is anything but world class. Sure, we move tons of freight, but is that the standard alone?

Just because it works doesn't mean it can't be improved better. It's always ok to reject the "don't fix it if it's not broken" mentality.

[SllX]

In the context of freight, it is the lone standard because as another commenter pointed out, throughput is more important than latency in bulk goods transport whereas latency is a much more important variable when passengers are involved.

US rail owners and operators know what the variables are that they care about and their customers care about are, and also what insurance companies care about and as a result, they are adept at moving goods coast to Great Lakes to coast, across the Appalachians, Missouri-Mississippi river system, the Great Plains, the Rockies, the Great Basin, the Sierra Nevadas, the Cascades and the California Coastal Range.

If they’re not using some software package or have complete sensor coverage on their tracks, they probably judged that they don’t need it. If a competitor actually finds advantage with these things tomorrow, then they will all adopt it.

[skellera]

I think it’s easy to call something out as primitive but what changes could be made and how much impact would it have? It doesn’t seem like our freight trains are the bottleneck when moving goods around the country.

[ant6n]

They kind of are if you consider how many goods are still shipped by trucks.

The US is perfect for rail - lots of long trips, with lots of goods. It could probably have more market share if goods could move more quickly and flexibly.

[bluGill]

While there is a lot US rail could do to get more freight, the fact is we send a lot more freight by rail than Europe.

[_delirium]

Just to add some numbers, the freight modal splits as of 2018 (most recent year with complete data), measured in tonne-kilometers, for a few countries [1] and the EU taken as a whole [2]:

US: 45% road, 38% rail, 17% other (water/pipeline)

France: 75% road, 15% rail, 10% other

Germany: 62% road, 25% rail, 18% other

Spain: 92% road, 4% rail, 4% other

EU: 76% road, 19% rail, 5% other

[1] https://stats.oecd.org/BrandedView.aspx?oecd_bv_id=trsprt-da...

[2] https://ec.europa.eu/eurostat/statistics-explained/index.php...

[ant6n]

Europe is not perfect for rail. Shorter trips, complicated geography, etc. etc. there are some legacy technology constraints that the us doesn’t have or solved a hundred years ago (loading gauge, max train length, couplers).

The fact is the us could do better, pointing to Europe which could also do better is moot.

[nickbauman]

I believe this as well. And I happen to know of one of the largest rail systems in North America changes its topology weekly. I'm not sure "throughput over latency" is as much of a mindset as those here who say it is.

[throwaway0a5e]

And how much tonnage moves on your "smart" rails? I'll let you pick the metric.

[cptcobalt]

Imo, your choice of scare quotes around "smart" telegraph your unwillingness to consider even a well-founded data informed argument, for what it's worth.

[jakear]

> Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith.

GP asked for a well founded data-based argument, assuming they don't actually want that is in bad faith.

I agree with GP that the connection between sensors and high speed to better freight rail is tenuous, whereas large amounts of tonnage moved more clearly indicates good freight rail.