Election night — what the board should look like, hour by hour

Tue Aug 18 2026 · generated Aug 19 6:05 PM ET from the live model

This page is generated. Every number comes from site_context.json, the same file the dashboard and Scenario Lab read. Do not edit it by hand — the next pipeline run will overwrite you. It was static until Aug 16 and had drifted badly.

Read this before 7:00 PM. The first hour is going to look bad and it is supposed to.

Shares below are percent of BALLOTS CAST, not poll shares. A results board has no undecided column, so every reported share is larger than the poll number it came from. Until Aug 16 this page used poll scale and told you to expect Rogers ~23%. The correct expectation is 33.8%. Judging Tuesday against 23% would have read a ten-point shortfall as on track.

1. The district votes in two time zones

Zone	Closes	Counties	Projected ballots	Share
Eastern	7:00 PM ET	10 — Leon, Wakulla, Taylor, Gulf, Madison, Jefferson, Liberty, Gadsden, Franklin, Lafayette	40,015	48.2%
Central	8:00 PM ET	6 — Bay, Jackson, Walton, Washington, Holmes, Calhoun	42,999	51.8%

Bay is in the late zone. 21,776 projected ballots, 26% of the district, and none of it is on the board until 8:00 PM ET.

2. Projected turnout: 83,014 (99% of 2024)

	2026 projected	Banked so far
Mail	—	14,933
Early vote	—	30,306
Pre-Election-Day	—	45,239
Election Day	40,267	—
TOTAL	83,014	—

Band 72,498–90,964. Election Day carries 49% of the total, the most back-loaded shape in recent cycles.

3. The sequence

7:00 PM ET — Eastern counties close. Rogers's worst moment of the night.

Florida counties dump mail and early vote within minutes of closing, so the first big number is 10 counties' banked ballots at once.

~23,244 ballots on the board, 28% of the night
Leon alone is 52% of that dump
Rogers should show roughly 25.6%, running third
Expected order: Power > Murphy > Rogers > Gross

Leon alone is 44% of this. Rogers running third here is the FORECAST. Do not react to it.

He should be running THIRD. That is the expected result, not a bad one. Nobody should call anyone, concede anything, or brief a reporter off it.

8:00 PM ET — Central counties close. The correction.

~48,978 ballots on the board, 59% of the night
Rogers should show roughly 33.8%, running first
Expected order: Rogers > Power > Gross

Bay alone is 21,776 ballots, 26% of the night, and none of it exists on the board before 8 PM. Expect a six-point jump in one drop. This is the first readable moment.

9:00–10:30 PM ET — Election Day precincts, both zones

~74,712 ballots on the board, 90% of the night
Rogers should show roughly 33.8%, running first
Expected order: Rogers > Power

Election Day is more Rogers-favourable than the banked vote, so the drift from 8 PM should be slightly toward him.

Election Day is more Rogers-favourable than the banked vote, so the drift from 8:00 PM onward should be slightly toward him, not away.

4. Per-county reporting weight

County	Zone	Projected	Banked	Election Day	% of night
Bay	CT	21,776	12,791	8,985	26.2%
Leon	ET	18,784	10,649	8,135	22.6%
Jackson	CT	6,563	2,973	3,590	7.9%
Wakulla	ET	5,268	2,386	2,882	6.3%
Walton	CT	4,875	2,195	2,681	5.9%
Washington	CT	4,150	1,768	2,382	5.0%
Holmes	CT	3,259	1,323	1,935	3.9%
Taylor	ET	3,127	1,207	1,920	3.8%
Gulf	ET	3,093	1,679	1,413	3.7%
Calhoun	CT	2,376	1,171	1,206	2.9%
Jefferson	ET	2,370	965	1,404	2.9%
Madison	ET	2,294	1,199	1,096	2.8%
Gadsden	ET	1,848	905	943	2.2%
Franklin	ET	1,586	685	901	1.9%
Liberty	ET	1,548	815	733	1.9%
Lafayette	ET	97	36	61	0.1%

5. Tripwires — when to actually worry

Judge against the expectation, not against zero.

Checkpoint	Expected	Worry if
7:15 PM ET	Rogers ~25.6%, third	below 20.6%
8:15 PM ET	Rogers ~33.8%, first	below 29.4%
9:30 PM ET	Rogers ~33.8%, first	below 29.4%

The single most useful number all night: Bay's first report at 8:00 PM ET. Near 28% for Rogers and the night is over. At 22% or below the model is wrong and everything after is live.

6. The soft spots, stated plainly

No August crosstab was published. The Aug 10-11 poll gives a district-wide topline only, so every market split holds JULY's geography while scaling to the new district totals.

That matters most for Leon, which is 52% of the 7:00 PM dump. If the undercard's growth was district-wide rather than Tallahassee-concentrated, the 7:00 PM picture is wrong in Rogers's favour and he will look better than forecast.

The share scaling is uniform and the undecided pool is not. Panama City carries more undecided than Leon, so the 7:15 PM Eastern-only expectation is the softest of the three checkpoints. Treat the 8:15 PM number as the first one worth acting on.

Opponent percentages are not published on this page. They were built on a different basis than Rogers's share; scaling them to ballot units sent the 7:15 board past 100%. The running order is reliable, the opponents' exact percentages are not, so only the order is shown.

Generated 2026-08-19 18:05 ET · projection 83,014 · banked 45,239 · data through 2026-08-18.