An exercise:

Imagine for a moment that you’re tasked with providing commuter rail service between Providence and Boston. Much of the track supports some of the fastest rail speeds in the country, and it’s fully electrified.

In what reality would the train technology look like this?

Running diesel push-pull trains under wires on mostly straight track? Unserious.

The Providence Line is the busiest commuter rail line on the MBTA’s 12-line network, with nearly 20,000 daily riders, and with Providence ranking as the 13th-busiest Amtrak station in the country, the line clearly deserves better.

At present, the trip from Providence to Boston takes 70 minutes (varies based on trip). The Acela does it in 48 (with far fewer stops). So how can a local line like the Providence Line close the gap if 48 mins is the ceiling? Quite easily, actually.

How to Improve the Line

The basic improvements the Providence Line needs could be made anywhere, but the incredible work of the Momentum Report and the TransitMatters report highlights the changes we all know would make the line far faster (please read and support both organizations).

The current runtime of MBTA #836 is 78 minutes, of which only 36 minutes is actual running. That means 54% of the trip is spent accelerating, decelerating, and stopping at stations.

Subscribe now

High Level Platforms:

Of the 12 stations between Providence and Boston's South Station, 6 are currently low-platform, which means everyone boarding has to climb stairs to get on and off the train, slower than simply walking through a wide door that's flush with the platform.

At the extreme, Amtrak can dwell for more than 5 minutes at a low platform with narrow doors; in places like Providence, the Northeast Regional can take 2 minutes thanks to those narrow doors and passenger luggage. The fastest dwells belong to the LIRR, where wide doors and flush platforms let people board and alight in as little as 30 seconds. Compound that across many stops and the delay adds up fast.

Time savings: Converting every platform on the line to high-level alone would save 9 minutes.

New Vehicles with better Acceleration

It goes without saying that heavy diesel trains take longer to reach cruising speed than a lightweight electric one, same as with cars. Caltrain's electrification is the famous example: what takes a diesel push-pull train (MBTA's technology) nearly 90 seconds to reach cruising speed takes an electric EMU a fraction of that.

And even if the trains stayed diesel but ditched the push-pull setup (a locomotive hauling cars), you’d still see savings, if MBTA’s Providence Line trains looked like the TexRail equipment below.

In a perfect simulation, the acceleration savings could approach 60 seconds at every single stop.

Time savings: Converting the current fleet to electric EMUs would save 16.5 minutes on its own.

More with the Same

I almost always run my own calculations on work like this, but given the great work of the Momentum Report and the TransitMatters modernization report, I’ll use their numbers, I’d rather not promote something outside the bounds of their advocacy. So: high-level platforms at every station, and electric EMU trains.

Per my very rough math, at max peak service there are 5 trains running between Wickford Junction/Providence and Boston, producing 20 one-way trips each way to Providence (8 of which continue to Wickford).

If you assume all five trains can run a 47-minute runtime to Providence and a 73-minute runtime to Wickford, with each train needing 10 minutes to turn around at each terminal, the time savings let you run an additional 12 trips in each direction every day.

Why Don’t We Just Do This?

What is happening on the Providence Line is a microcosm of American transit. We look at METRA, MARC, VRE, and Metrolink, and we accept sluggish, low-platform diesel operations as the natural order of things. But Caltrain’s 57% ridership jump post-electrification proved what happens when an agency stops accepting the status quo. Fixing this isn’t a question of technological limits, it’s a test of political will. We don’t always need to build new multibillion-dollar systems from scratch. Sometimes, we just need to stop choosing to run the systems we already have so poorly.

Sources for this work: Transit Matters Report & Momentum Report

Author’s note: I was formerly contracted with the MBTA. All information in this article comes from fully public sources; none of it was gained from my time at the agency, its employees, or anything not public-facing.

Subscribe now

Share

But what about the trains?

The results are, frankly, insane. Lancaster has the second-highest rail ridership of the 14 Amtrak stations in Pennsylvania, beating out much larger cities like Pittsburgh and Harrisburg. It sees 1,255 riders a day, averaging 44 riders per train served. In fact, of the top 25 Amtrak stations in the country, Lancaster ranks fifth per capita.

Defining the “Lancaster Model”

What’s happening in Lancaster isn’t a happy accident. It’s a repeatable formula for regional success I hypothesize as the “Lancaster Model”. For a city to replicate these numbers, it must meet four specific criteria:

The Four Criteria

1. Satellite Geography: Located 50–100 miles from a Tier-1 metro (the “Sun”).

2. Frequency Floor: Must be frequent enough to be useful (>10 round trips daily). No “peak-only” service.

3. The Connected Urban Anchor: The station must be in a walkable, gridded core, not a highway island.

4. The Speed Rule: The train must beat or be equal to the car. This requires electrification and high-level platforms to minimize “speed sinks” like slow acceleration and long dwell times.

The Model in Action: Lancaster, PA

Satellite Geography: Lancaster is 65 miles from Philadelphia and 35 to Harrisburg, this is within the hypothetical model outlined.

Frequency Floor: The station sees 13 trains a day in both directions between Philly and Harrisburg (The Keystone Service), plus another (The Pennsylvanian) that takes you through to Pittsburgh. While that pales in comparison to the S-Bahn or regional rail networks of Asia and Europe, it provides enough “turn-up-and-go” convenience for a commute or a leisurely trip to Philly or the airport.

• Good: Lancaster (14 round trips)

• Bad: Cincinnati (3 round trips... per week)

The Connected Urban Anchor: While Lancaster County’s bus network is modest, it is strategically anchored. All routes start or end in Downtown Lancaster, and four of them directly serve the station. Is the bus network helpful for Lancaster County, not if you’re not going to Downtown Lancaster, but for purposes of train station access, it works great. More importantly, the station is located right at the edge of a well-gridded, dense urban core, not isolated in a parking lot.

• Good: Lancaster (Walkable, gridded urban core)

• Bad: Syracuse (The station is essentially an island off the highway)

The Speed Rule: Electrification is the secret sauce. Diesel trains take 10+ minutes longer than the electrified Keystone service to reach the same speeds, and in transit, 10 minutes is the difference between taking the train and driving. The train averages 60 mph, reaching max speeds of 110 mph. It is 12 minutes faster than driving to Philadelphia in perfect conditions, and significantly faster when you factor in the I-76 nightmare.

Share

The "What If" Scenario: Janesville, Wisconsin

Until 1971, Janesville was a rail town, served by the Varsity and other daily lines. It briefly saw the Lake Country Limited Amtrak service in 2000, which failed spectacularly (a post-mortem for a later article).

What if Janesville saw 14 round trips a day to Chicago, Madison, Milwaukee, and Rockford?

The Lancaster Model proves this isn’t a pipe dream. A 40-minute trip to Madison, 30 minutes to Rockford, and 80 minutes to Milwaukee would instantly transform Janesville. It becomes a viable bedroom community for Madison and makes catching a Cubs game on a summer evening a reality without the I-90 traffic. You cannot fix a local economy without tapping into its neighbors.

Subscribe now

Applying the Model to the "Anti-Rail" State: New Hampshire

New Hampshire is notoriously one of the most anti-rail states in the Union. But what if they embraced this model by extending the MBTA Lowell Line to Nashua and Manchester?

Satellite Geography: Manchester and Nashua are cities of 100k people within 70 miles of the Boston CBD.

The Frequency Floor: The proposed service calls for 16 round trips per day, more than Lancaster.

The Connected Urban Anchor: Both cities are gridded, walkable, and have existing transit "bones." Looking at the rail Right-of-Way (ROW) in Manchester, the integration into the built environment is already there.

The Speed: With high-level platforms and electrification, Manchester could be 70 minutes from North Station (average speed of 47.6 mph). That beats driving during rush hour every single time.

Because the model is so efficient, the latest DOT study shows nearly 80% of operational expenses would be covered by fares. While I don’t believe transit needs to make a profit, in the context of New Hampshire lawmakers, that is a massive selling point.

Where else could the Lancaster Model work?

When I look at the map, I see “Lancaster bones” all over the country:

• Rockford, IL

• Reading, PA

• Allentown/Bethlehem, PA

• Asheville, NC

• Scranton, PA

• Fort Collins, CO

All of these are currently being studied by the FRA Corridor ID program. In the case of Rockford, METRA is going to be there by 2027.

I want to hear from you: What do you think of the Lancaster Model? What other medium-sized cities within 75 miles of a major metro are screaming for fast, frequent, electrified rail?

Leave a comment and let’s talk shop.

(Author’s note: I should mention that ridership is definitely driven by the Amish who do use the trains, so while it’s a cheeky article, I definitely think it’s still a model for how we should treat small US sized cities with rail)

Subscribe now

Leave a comment

But while I was wrestling with that, the 2026 CAHSR Business Plan just dropped. The new total cost estimate? $231.3 Billion. Frankly, that number is insane.

While the full deep dive is coming, we need to talk about the biggest headline in this new plan: the agency is looking at using Metrolink’s Antelope Valley Line as an interim connection instead of immediately boring a 26-mile tunnel under the Angeles National Forest.

It makes sense to be critical of the Palmdale to Burbank segment. It is the most expensive piece of the entire puzzle, estimated at $15.94 Billion, or roughly $419 Million per mile.

Here is the breakdown of what that segment actually requires (and remember, these are just estimates since construction hasn’t even started):

The project section includes approximately 27.95 miles of bored or mined tunnels (per the Final EIS):

• 1.52 miles of cut-and-cover tunnels

• 0.47 miles of at-grade cover tunnels

• 13-mile tunnel between the Palmdale area and Agua Dulce

• 1-mile tunnel near the SR 14 freeway

• 12-mile tunnel underneath the Angeles National Forest and the San Gabriel Mountains National Monument

To put that in perspective: while it isn’t one single continuous bore, the total tunneling required for this one segment would make it the fifth-longest rail tunnel in the world if it were uninterrupted. It’s a massive engineering lift, and it’s currently the only thing standing between the Central Valley and Los Angeles.

You need the tunnels.

Before I go any further, I want to be crystal clear: these tunnels are absolutely essential to the project. Why? Because they turn the trek from Palmdale, the gateway to the Central Valley, to Burbank into a THIRTEEN MINUTE trip. 13. I cannot overstate how transformative that is for regional mobility.

There is simply no other way to achieve those kinds of time savings without punching through this mountainous terrain with concrete and steel. We need these tunnels, and if you’re reading this, you likely already understand the benefit. That’s why the agency’s pivot toward “abandoning” them in favor of the existing Antelope Valley Line in this new business plan is so incredibly eye-raising.

The Pivot to the Antelope Valley Line

The new CAHSR draft proposes a major shift: using the existing Antelope Valley Metrolink line, which already serves Palmdale, Burbank, and LA, as an “interim” solution. The goal is to connect the Central Valley to Los Angeles faster than the currently anticipated 2039 timeframe.

Honestly? It’s a smart move. In my opinion, an isolated segment between Bakersfield and Merced will never net the ridership or operational profit to make the service truly productive. You need the big anchors, San Francisco and Los Angeles, to make the math work. That isn’t a controversial take; it’s just reality.

Subscribe now

Decoding the “Interim” Plan(s)

Let’s be real: I have no idea what is actually going on here, and I’m not sure the agency does either. CAHSR is notoriously vague. They tend to bifurcate information so nothing is ever set in stone, presumably to stay “nimble,” but a little clarity would go a long way.

Reading between the lines of page 21 of the business plan, the agency suggests that once San Francisco and Palmdale are connected by 2039:

“At Palmdale, a connecting Metrolink/Surfliner express service to Los Angeles and San Diego could give passengers a preview of a truly statewide service.”

They also mention that early investments in the LA Basin, like upgrading shared corridors, could happen concurrently with the Central Valley-to-Bay Area work. They point to the 2024 Caltrain electrification as a model for how these “early investments” prepare the tracks for future HSR through-service.

Based on this, I’ve deduced a three-stage roadmap (or options) that the agency is proposing to lead us to the full tunnel buildout:

1. The Transfer: Passengers take HSR to Palmdale, then physically transfer to an Amtrak or Metrolink train to get to LA/San Diego.

2. The “Blended” Upgrade: Electrifying the existing Antelope Valley Line to allow high-speed trains to run directly through to LA (similar to the Caltrain model).

3. The Original Plan: Building the original 28 miles of tunnels and running full-speed HSR directly into Burbank and LA.

The travel times of each are as follows:

The financial logic is that the SF-to-Palmdale service is projected to be a massive cash cow, generating nearly $1.2 Billion in profit (a 236% recovery ratio). That profit would then be used to help pay for the $16 billion tunnel between Palmdale and Burbank, while options 1 and 2 keep riders moving in the meantime.

Subscribe now

How bad is it, doc?

Let’s look at the numbers, because this is where the “interim” plan starts to hit the reality of physics.

Currently, the Metrolink Antelope Valley Line (AVL) from Palmdale to Los Angeles takes 1 hour and 59 minutes. That’s with stops, but it’s the baseline. Now, compare that to the law: Proposition 1A, which voters passed in 2008, legally mandates a nonstop travel time between LA and San Francisco of 2 hours and 40 minutes (or 2:42 depending on how you read the technicalities).

Adding nearly two hours of “interim” travel time is a total non-starter for the integrity of the system. It’s not just a delay; it’s a legal and operational crisis.

Per my calculations, even if we run high-speed trains on the existing Metrolink corridor, we are looking at adding roughly 44 minutes of travel time compared to the tunnel plan. This is because the existing terrain is famously “curvy”, to put it mildly. It was built for 19th-century freight, not 21st-century high-speed rail. You can electrify the tracks, but you can’t “electrify” away a tight curve that forces a train to crawl at 30 mph.

In the end, though, we have to face the music: electrifying an existing commuter line to allow for “through-service” is a far more viable interim step than waiting another decade for a connection to Los Angeles with zero revenue flowing in. It’s a bitter pill, but it’s better than no medicine at all.

Downsides to the interim plan

As much as I want to see this system connected, we have to be honest about the downsides. An interim plan is a compromise, and in transit, compromises usually come with a heavy price tag and a side of “permanent” temporary solutions.

1. We shouldn’t even need an interim plan. In principle, the idea that we need a “bridge” solution is an admission of failure. As a country and a state, we should be able to execute a project of this scale without it becoming so overbudget and delayed that we have to pivot to 19th-century tracks to keep the project alive.

2. The Double-Track Dilemma. According to FRA records, Union Pacific operates around 30 freight trains daily on this corridor, which is currently mostly single-track. Add to that the 15 daily Metrolink trains in each direction, and you have a capacity nightmare. You cannot "interline" high-speed rail with that level of freight and commuter traffic without a massive, expensive double-tracking project. If you don't double-track, the throughput simply isn't there to support a high-speed service.

3. The Prop 1A Legal Cliff. This is the big one. It will be politically exhausting in 2039 to ask for tens of billions of dollars for 28 miles of tunnels just to save 44 minutes. But here’s the kicker: it is legally required. Per the language of Prop 1A, the funding is tied to a system that achieves that 2-hour-and-40-minute (ish) travel time between LA and SF. Furthermore, the law requires a 5-minute headway capability. I repeat: a 5-minute headway is physically impossible on a shared, curvy, freight-heavy corridor like the Antelope Valley Line. If they don't build the tunnel, they are breaking the very law that gave them the money in the first place.

4. Where is the money coming from? It’s still unclear to me if the Authority can legally spend HSR-dedicated funds on Metrolink upgrades. When they funded the Caltrain electrification, it was justified because the final “blended” plan involves HSR trains running on those tracks forever. If the Antelope Valley Line is truly just an “interim” solution, can they legally justify the spend? Or is this “interim” plan actually a quiet admission that the tunnel might never happen?

   ---

I mean, yes?

At the end of the day, I absolutely believe that CAHSR must connect San Francisco to Los Angeles as quickly as humanly possible. So, yes, I support this decision, even if the details are sparse and the “unknowns” are incredibly loud.

I have always maintained that the Authority should have started with the hardest segments first: San Jose to Merced and Los Angeles to Palmdale. If you tackle the most difficult engineering feats at the beginning, you ensure the project’s survival. God forbid the project ever stops, you don’t want to be left with the “easy” Central Valley trackage while the massive mountain crossings remain unbuilt.

But we are where we are.

The Governor and the Statehouse need to stop treating this project like a “waterfall” construction schedule where funding only arrives in slow, sequential drips. We need the political will to provide more funding, faster. We should be boring those 28 miles of tunnels now, not waiting until the Initial Operating Segment opens in a decade.

I believe in the leadership at CAHSR, and with the current pressure coming from the federal level, the agency simply has to get this right. We need the interim solution to build the ridership, and we need the tunnel to fulfill the promise. And we need both of them, like yesterday.

Subscribe now

Share

Read more

I remember landing at that massive, modern airport (DTW) and realizing I’d need two buses and over 90 m…

Read more

The Northstar Commuter Rail Line, connecting Minneapolis to Big Lake, was a masterclass in the latter. As of last week, its life as a rail service has officially come to an end after the Metropolitan Council defu…

Read more

Watching a movie with me is nearly impossible, as my suspension of disbelief is nonexistent.

“Godzilla wouldn’t be able to walk on Park Avenue because the rail yard is directly below his feet, and the columns can’t hold his weight.”

“Do they have to do an environmental review to fix Chicago after the Transf…

Read more

That statement was as true as me telling you I have a wife waiting at home for me (I’m a gay man with a boyfriend, by the way).

Not only was Storrs, CT, h…

Read more

Read more

To accommodate this anticipated growth and the already heavy downtown traffic served by streetcars, the Providence City …

Read more

But Hartsfi…

Read more

That November, New York’s “forgotten borough” voted 2-to-1 to secede from the city. The reasons were plenty: frustration over underinvestment, poor city services, and the not-so-minor fact that Staten Island hosted the largest landfill in the world. The rebellion was stopped by Governor Cuomo and the state legislature, who feared that letting one borough go might unravel the city itself.

A few years later, after some political concessions, Mayor Rudy Giuliani ran for reelection on a promise that stuck: making the Staten Island Ferry free. For 30 years, the ferry between Staten Island and the southern tip of Manhattan has remained one of the best deals in American transit, completely free to ride.

History

The Staten Island Ferry began in 1817, when the Richmond Turnpike Company launched a steamboat service between Manhattan and Staten Island. Cornelius Vanderbilt, yes, that Vanderbilt, bought the company soon after, merging it with other ferry operations. By 1864, he sold it at a loss. The Baltimore & Ohio Railroad later purchased the service before the City of New York finally took control in 1905.

As with so many great public transit systems, the car changed everything. Ridership declined as people chose to drive instead of sail, and by 1967, the Staten Island Ferry stood alone as the city’s only remaining commuter ferry.

Today, it’s the busiest passenger ferry line in the world, a relic of the past that somehow became a backbone of modern New York.

Staten Island Railway Transit

After purchasing the Staten Island Ferry, Cornelius Vanderbilt financed the first rail line from Tottenville to Stapleton, responding to public pressure from island residents who wanted better transportation. The line made a modest profit but, like the ferry, eventually went into receivership. It was sold at foreclosure in 1872 to George Law, an entrepreneur, railroad investor, and one-time presidential candidate.

By the late 1870s, the railroad was barely operational. That’s when Erastus Wiman, a civic-minded Staten Island businessman, stepped in. On March 25, 1880, he organized the Staten Island Rapid Transit Company (SIR) and partnered with the Baltimore & Ohio Railroad to expand the network. Together, they built a new ferry terminal at what’s now St. George, on land donated by Law himself, who asked only that the terminal bear his name.

The North Shore Branch opened in 1886, and with Wiman’s advocacy, Congress funded the Arthur Kill Bridge, then the world’s longest drawbridge, connecting Staten Island to Cranford Junction via the Baltimore and New York Railway.

At the time, you could get from the Staten Island Ferry to Cranford (now on NJ Transit’s Raritan Valley Line) in just 17 minutes, a trip that takes 50 minutes by car or nearly two hours by transit today.

As freight traffic dwindled and ridership declined, the lines were gradually abandoned. In 1971, the Tottenville line was transferred to the MTA, where it continues to operate today as the Staten Island Railway (SIR) and is directly connected to the ferry.

(Note: I’ll cover both the costs of Staten Island secession and the island’s broader transit needs in future posts.)

Subscribe now

Operations

The Staten Island Ferry isn’t owned or operated by the MTA, it’s run by the New York City Department of Transportation (NYCDOT). The fleet consists of nine vessels, though only four operate at a time during peak hours. The service runs 24/7, departing every 30 minutes, with frequencies doubling to every 15 minutes during the morning and evening rush.

On an average day, the ferry moves about 46,000 passengers, more than any other ferry route in the world, and that’s with only one line. For comparison, most major commuter ferry systems operate between 4 and 35 routes to reach similar ridership totals.

(Note: pre-pandemic, the Staten Island Ferry Line itself would’ve been the second busiest commuter ferry system in the world).

The boats themselves are massive. Across four vessel classes, capacity ranges from 4,427 to 6,000 passengers per boat, making them the largest commuter ferries on the planet. That scale comes with a cost: it runs about $8,306.51 per revenue hour per vessel, or $16,613 per hour to operate the line under normal service, doubling to $33,226 per hour at peak.1

If every ferry were full, operating costs would amount to just $0.83 per passenger, compared to the average $2.30 per-trip subsidy across USA transit in 2022, an incredibly efficient way to move people across the harbor.

Subscribe now

Staten Island Ferry 2.0

Unlike most of my posts, which lean heavily on data, this idea is more observational. I’ve been thinking about where a Staten Island–style ferry system, high-capacity, frequent, and free, could actually work elsewhere.

Read more

You’re standing in one of the most beautiful train halls in the country, Chicago’s Union Station. You look up at the departure board, and it rivals a major European hub: six METRA lines, nineteen Amtrak routes, plus nine new regional lines and five “h…

Read more

Ouch. That reality stings, but it’s true. MARTA hasn’t added a single mile of new rail since 2000, the MTA has managed only a handful of new stations in decades, and Chicago’s Red Line extension has taken decades for just six miles of elevated trac…

Read more

It’s a nice airport with solid national connections, but it’s missing one big thing: ground transit. Unlike airports in New York, Providence, or Boston, Bradley is basically cut off. The only public option today is the 30X “Bradley …

Read more

Read more

Alongside the announcement came a project map, station list, and s…

Read more

Enter the next-gen Acela trains (

Read more

And without distraction, my mind …

Read more

Read more