Florida Vote Tracker 2026

Each county's projection starts from its past ballots: the Democratic plus Republican votes cast in the 2018 governor's, 2022 governor's and 2024 presidential races, each divided by 0.99 to account for votes cast for other candidates, since the history tables hold only the two major parties. The three totals are averaged with equal weight to give a historical expected ballot count for the county. Registration growth since 2018 is not reflected, because each past race enters as ballots cast and not as a share of that year's rolls, and the 2024 presidential year pulls the average up for a midterm. This year's pace is then applied: for every county with counted ballots, its counted votes divided by its historical estimate are compared with the same ratio across all reporting counties, the comparison is limited to between 0.7 and 1.3, and 15% of the gap moves that county up or down. All counties are then rescaled so the statewide total equals the plain three-election average, so the pace redistributes ballots between counties but never changes the statewide total. Each county is held to two limits: it cannot fall below the ballots already counted and it cannot exceed 95% of its registered voters. The statewide figure is the sum of the county projections divided by all registered voters, and each county's percentage is its projection divided by its own registered voters, both measured against a June 2026 snapshot of the voter rolls. The party split shown beside it estimates who is still to come from each county's unvoted registered voters by party, and it counts voters by registration, not by the candidate they vote for.

Each update (the 7 p.m. ET run) the model rebuilds the forecast in a fixed order. It starts with the live count by registered party, which sets each county's expected final turnout (its 2018, 2022 and 2024 average, nudged 15% by this year's early-vote pace) and the party mix of the voters still to come. The history side is each county's average two-party result in those three elections, weighted by projected turnout, then adjusted for how independents and each party's own voters are polling against their exit-poll baselines (about 49% Democratic among independents, 95% Democratic and 94% Republican loyalty). The polling side averages every poll: a poll's weight halves every 14 days, grows with sample size, and counts 60% if the pollster or sponsor is party-aligned and 35% if it is an internal poll released or paid for by a candidate or campaign, the same for Democratic and Republican polls, while nonpartisan pollsters count in full. The statewide target blends the two, with polls at about 84% of the weight today and rising toward 90% on Election Day, and a small campaign-money term (the same rule as the House, at most 1.5 points, from FEC filings for the Senate and press-reported state filings for Governor) is added to the fundamentals only, since polls already reflect spending, and one statewide shift then sets each county's result. For the House, each district starts from its 2024 presidential result on the new map, gets the same statewide swing (lopsided districts move less), and then a small campaign-money adjustment of at most 1.5 points from FEC filings (mostly through July 29), with money a candidate put in counted at half weight; only FL-08 and FL-09 have district polls, which carry a quarter of the weight. The weights are judgment calls that have not been tested against past elections, the forecast is a single number with no error range, and the exit-poll baselines come from the NBC (2018) and Edison/CNN (2022) exit polls and a 2024 table that does not name its pollster.