Show your work
Methodology & caveats
Opinionated calls, stated uncertainties. This is a casual blog analysis, not a census — but every number is sourced and reproducible.
The one-paragraph version
Each of the 48 contiguous states (plus DC) is assigned to the quadrant holding the majority of its landmass as the meme draws it — not real latitude, not population centroid. Then the handful of big metros that clearly sit across a line from the rest of their state are hand-corrected (Bay Area, Las Vegas, Northern Virginia, Houston). Population is the corrected headline metric; GDP and land area stay at whole-state resolution. All source numbers are one consistent vintage (CensusJuly 2025 estimates, 2025 nominal GDP).
Where the lines fall
The crosshair sits near the Kansas / Nebraska / Colorado / Missouri junction — roughly the geographic center of the lower 48, which is exactly where the state grid is densest. I did not solve for a precise latitude of each line: the meme's stroke isn't a parallel, and the metro corrections (Bay Area → A, NOVA → B) imply it isn't even constant latitude coast-to-coast. Instead, states the lines visibly bisect are theswing set, flagged and stress-tested.
Base state assignment
Alaska & Hawaii are excluded — the meme map shows only the contiguous 48. (For the record: AK ≈ 737k would extend corner A and holds Fort Greely's missile-defense interceptors; HI ≈ 1.43M anchors the Pacific Fleet. Neither is on the board.)
Swing states (bisected — flagged)
These sit on a line; each gets a base call, and the swing-flipscenario moves all of them to their alternate side at once (the worst case for the base call). Flagged swing states: CO, KS, MO, OK, WV. Even under a full flip, the population ranking B > D > C > A is unchanged — see the scenario chart on theeconomy page.
Metro corrections
A whole-state assignment misplaces big metros that sit across a line from their state's center of mass. Each metro's actual Census July-2025 population was fetched (never estimated) and moved. Tier 1 is the base case; Tier 2 is plausible-but-debatable and shown only in the sensitivity.
| Metro | People | Move | Tier |
|---|---|---|---|
| SF–Oakland–Fremont MSA (Bay Area ex-South Bay) | 4,630,041 | C → A | base |
| Las Vegas–Henderson MSA | 2,407,226 | A → C | base |
| Northern Virginia (Arlington/Fairfax/Loudoun/PWC) | 2,340,000 | D → B | base |
| Houston–Pasadena–The Woodlands MSA | 7,904,627 | C → D | base |
| Austin–Round Rock–San Marcos MSA | 2,620,945 | C → D | sensitivity |
| Dallas–Plano–Irving metro division | 5,641,795 | C → D | sensitivity |
- "Bay Area except South Bay" is the SF–Oakland–Fremont MSA, which already excludes San Jose (South Bay stays in C). It's afloor — adding the North Bay would push ~1M more from C to A.
- Northern Virginia = the four core NoVA counties (2.34M), a conservative named-jurisdiction floor; the full DC-metro Virginia footprint runs ~2.6–3.1M, which only widens B's lead.
- Texas is the swing engine. Houston is clearly east → D. Austin and the Dallas division sit right on the line; moving them shifts ~8.3M from C to D, so they're held as a Tier-2 sensitivity, not buried in the base case.
Deliberate simplifications
- Population is corrected; GDP & land are whole-state.Splitting metro GDP across lines would need metro-level GDP and a defensible allocation — out of scope. So GDP-per-capita and density are computed on the whole-state population basis, keeping numerator and denominator consistent.
- One vintage. Everything is Census-2025 / GDP-2025. Absolute numbers may lag a true 2026 count, but the shares — which the whole question turns on — are robust.
- Checksum. The four quadrants + DC sum to 339,614,767 (48 states + DC). The July-2025 estimate series runs slightly hot versus a ~335M gut figure; that's the vintage, not a double-count — each of the 49 units is summed exactly once.
Reproduce it all: compute.py +quadrant_states.csv ·quadrant_corrections.csv ·quadrant_summary.csv.