The Line, Priced

Program status

Method

1. Benchmarks first. Research agents sweep the live record for what other projects actually did — cost, ridership, funding, wages, permitting durations — and each row lands with a source URL, a year-of-dollar, and a status (verified read from a primary; reported secondary; proxy a labelled stand-in).
2. Models second. Every line item, market row and funding source must name the benchmark keys it is scaled from. A number that cannot show its parent is deleted, not defended.
3. Adversarial review third. A red team is handed the model and the benchmark set and told to refute it. Findings are applied in full unless they fail arithmetic; each changed row carries a [REVISED] tag with the correction and the reason, and the originals survive only in version control.
4. Corrections travel. When a review moves a number, every page that cited it moves too — the 1h20 claim was corrected in the corridor paper, the modes table, the stopping figure and the movement corpus in the same pass.
5. Nothing is published while it is wrong in a known way. The public face of chiledo.com is a project notice until this program says otherwise.

Design basis

• Dedicated tracks end to end. Norfolk Southern runs 50–70 freight trains a day on the Chicago Line. Shared track is not a strategy; it is a terminal-approach tactic.
• 186 mph where the land allows it — the three rural segments, 129 miles — with 110–125 through suburban territory and conventional running into CUS. Above 125 mph, federal rule 49 CFR 213.347 permits no grade crossings at all: the fast segments are sealed corridors.
• The speed tier is proven forty miles away. Amtrak's Porter–Kalamazoo line has run 110 mph in revenue service since February 2012.
• Labor at Davis-Bacon package rates, statutorily required on FRA money (49 USC 22905/24312), from the actual 2026 wage determinations: $114/hr operating engineers in Cook County down to $44.62/hr laborers in St. Joseph County. The fast rural miles are cheap twice.
• Trip time honestly modeled: ~2h15 express for A (corrected twice: from 1h20, then from 2h10 once the 95 mph suburban averages were questioned); ~3h15 express / ~4h05 local for B-110, a diesel with stops averaging what Brightline actually averages. Today: 4h41.

The stack

Eight segments, priced by their governing constraint

Line items

Every row names its basis — the benchmark rows it is scaled from. Hover a basis tag for the benchmark's source.

What governs the schedule

1. No Corridor ID slot exists. Chicago–Toledo was not among the 69 corridors selected in December 2023; the nearest hooks miss one end each. The on-ramp is the next new-corridor application window. Steps: $500k scoping → Service Development Plan (10% match) → project pipeline (20% match).
2. STB authority is clean but eminent domain is not federal. An interstate line avoids Texas Central's jurisdictional decade, but condemnation power is state law — held separately in Illinois, Indiana, and Ohio by a railroad qualified to do business in each. Assemble the authority before the alignment is public, or become Texas Central: twelve years, a quarter of the parcels, zero track.
3. NEPA under the 2023 rules: two-year statutory EIS targets, 150-page limits, sponsor-prepared documents, FAST-41 dashboard status. Modeled 3 years tiered; 5+ if litigated. The Brightline West lesson: the fast path is an existing transportation corridor and inherited clearance — which is what the Dunes segment does, staying inside the South Shore footprint past the national park (the avoidance alternative is the alignment).
4. Seasonal survey windows bite: Indiana bat and massasauga surveys run on the species' calendar, not the project's. Missing a summer window costs a year, not a fee.

The choice set, after four reviews

The open question the reviews left — half-hourly-capable 186 mph infrastructure for an 8-round-trip service — answered by building the other railway and pricing it the same way. 125 mph is FRA Class 7: crossings are permitted with barrier systems, so no sealed corridor; diesel/hybrid sets, so no catenary; geometry that hugs the NS corridor; roughly twenty stations instead of eight. Same benchmarks, same 30% professional services, same 40% contingency.

1. B as first drawn was not a legal railway. 125 mph over 116 retained crossings: zero such crossings exist in US revenue service and no FRA-approved barrier product exists to buy. The honest B is B-110 — Class 6, quad gates and detection, 110 on the rural miles. Sealing the corridor for a true 125 costs +$5.4B and converges on A.
2. B gets fewer riders than A, not more. The local tier had assumed 26k boardings per small-town station — more than South Bend or Elkhart board today — for towns of 1,600 to 7,500. Actual Amtrak small-station median: ~9k. Corrected B central: 0.52M riders, $14.4M.
3. B's times were 30–50 minutes optimistic. A diesel with stops averages what Brightline averages (~80 mph). B-110 express ~3h15, local ~4h05 — the local saves ~35 minutes over tonight's overnight train. That is Chicago–St. Louis by construction.
4. B's saving is real but smaller. ~$14B of capital for a railway ~60 minutes slower with 40k fewer riders and ~$75M/yr more subsidy; ~$1.5B of the saving reappears as operating cost.
5. B's through-service case is MORE buildable than A's — diesel sets run onto the Michigan Line and NS to Cleveland today; electric sets cannot leave the wire. The spine premium is only 20 minutes either way.
6. The red team prefers neither. It prefers C: B-110 on A's tangents (the NS corridor is nearly tangent on the rural 129 miles, so B's geometry does not preclude 186 later), structures built with 25 kV clearance, 8–10 stations not 20 (keep Waterloo/Fort Wayne, drop the villages), ~$40B now and ~$8B later — a deferral, not a saving. And it names the option neither alternative priced: 0+, incremental capacity on the NS right-of-way itself, ~$8–9B, ~3h45, ~0.3M riders. B's local tier is 0+'s argument, not A's.

Who rides it

Three scenarios, each market row naming the actuals it is scaled from. The forecast-optimism literature is in the benchmark set — the spread is the forecast.

Who pays for it

The stack against the year-of-expenditure program, each source with its statute and its real-world precedent.

Open questions

1. Settled by four reviews: file B-110 ($40.6B, with Cs salvage) as the Service Development Plan build alternative; carry 0+ ($16.9B) as the honestly-priced baseline; A ($54.1B) is the network end-state. The open capital question is no longer which railway — it is the FD 36868 intervention and the funding stack.
2. Is Chicago–Toledo a project at all, or a segment? Every model improves sharply in the network case. The honest framing may be that this corridor is the middle third of Chicago–Detroit–Cleveland and should be studied as such.
3. Who pays the operating gap. ~$137M/yr central, forever. PRIIA 209 assigns it by train-mile; Indiana owes the majority. No precedent exists for Indiana paying anything like it.
4. The dedicated instrument. California's $1B/yr rides an existing carbon auction. The four-state equivalent — corridor road pricing, a compact sales-tax increment, something else — has no design, no sponsor and no vote.
5. Land before alignment. Every value-capture tool is useless to whoever shows up after the route is public. What entity assembles land along 234 miles of three states years ahead of a line? The 1827 grant answered this; the 2026 legal landscape does not.
6. The UP–NS merger. STB accepted the application May 28, 2026. The host railroad for every alternative in this file may be mid-acquisition — and merger conditions are historically where passenger concessions get extracted. The single largest new lever and the single largest new uncertainty; Options 0+ and C should be re-reviewed against whatever conditions emerge.

Sources and data

Studies the program draws on. CHECK: marks a source whose substance is known but which has not yet been read against the document and archived; nothing so marked may be quoted or page-cited on any page.

Data access behind the preview key: /api/estimate, /api/market, /api/register, /api/sources, /api/timeline, /api/route?from=&to=&year=. The storehouse shares its R2 archive with the sister projects and keeps its own D1 and vector index — one copy of each artifact in the world, three readings of what it means.

Change log