General Election Baseline

Austin Rogers (R) vs Amanda Marie Green (D) · FL-02 · 3 November 2026 loading…
Confidence tiers. Every headline figure on this site now carries one, because a certified count and a projection were rendering in identical cards. measured observed and reconciled against an independent source. derived arithmetic on measured values. modelled a projection with assumptions, carrying a range. estimated a weak or unvalidated basis, directional only. The persuadable pool is estimated: it applies a precinct's primary vote share to its registered Democrats, so it measures reachable VOLUME and is not a count of people. The November projection and its range live here.

Money

FEC candidate totals, cycle 2026. Every candidate below reports through the same date — a comparison across different coverage windows is not a comparison. Cash on hand is the number that matters the day after a primary; receipts count money already spent winning one.

Net of debt, the cash advantage inverts. Rogers holds $387,029 against $377,218 in debts, so his net is about $9,811. Green holds $28,038 with no debt. The 13.8x gross figure is real but it is not a spending advantage, and quoting it without the debt beside it is the kind of number that gets corrected in public. Use net internally.

Turnout model: the rate, checked against clean rolls

Vote-by-mail tracker (daily from the Division of Elections)

Media markets — November projection

Turnout history: primary vs general

Every baseline this project built was Republican primary turnout, which is the wrong electorate for November. The general history is now built from the turnout archive for 2018, 2020, 2022 and 2024, all-party, by county, with method mix. 2018 and 2022 are the comparable midterms; 2024 is reference only.

🔁 Primary vs General, side by side

Persuasion messaging

Segment-by-segment framings for soft Democrats and unaffiliated voters, built on North Florida's own economy rather than national talking points. Every framing is an untested hypothesis; the facts underneath are sourced and tiered.

🎙️ Soft D and NPA messaging

Where each nominee ran strong and weak

Share of their own party's primary vote in each county, against their own district-wide share. This is the share, not the vote count — Leon casts twelve times Liberty's ballots, so a raw count would just re-rank the counties by size.

The persuadable pool

No Party Affiliation voters plus Democrats who did not vote for Green. Green's weakness is measured in each precinct wherever the county publishes precinct results; where it does not, the county rate is used and the row says so. Ranked on absolute voters, because a percentage of a small precinct is not a canvassing plan.

Top 40 precincts by persuadable pool

NPA precinct targets

81,422 unaffiliated voters, scored on three things at once: how many there are, whether they actually turn out, and what terrain they sit on. Volume alone just ranks big precincts. Propensity is an index against each county's own average, measured from the 2022 general — the comparable midterm general — because a closed primary tells you nothing about voters who cannot vote in it.

Top 60 NPA precincts

Opposition research

Who Green is, what she has actually committed to, and the four attacks that would backfire. Internal only.

🔎 AMG research memo

The beaten Democrats: who is actually reachable

Three candidates took 66.5% of the Democratic primary between them. Their voters are not one group. A Democrat in Republican-leaning territory can hear a Republican message; a Democrat in a deep-blue campus precinct cannot, and the realistic outcome there is that they stay home in November rather than cross over. Raising the race's salience in that second set works against us, so the two are counted separately and never summed.

Zateslo — persuadable ground

R-leaning precincts only, ranked on NPA plus Zateslo-aligned Democrats. Polling place and address included so this can go straight to field.

A soft Democrat is not a Rogers voter. This ranks where persuadable volume sits, not where persuasion will work. ← back to the dashboard