Projected final count
RRepublicanDDemocratLighter bar = projected gain
If the rest of the count looks like...
The same remaining council votes, split using one batch at a time. Who wins the 4 seats, and who is first out.
Batch by batch
Each candidate's share of the votes in each batch. An even split across 11 candidates would be 9.1% each.
Mail vs in-person
Share of all HB council votes so far, by ballot type. In-person (vote center) ballots lean heavily Republican.
What's left to count (county)
From the Registrar's ballots-left page. Ballots handled at vote centers are likely to look more like the in-person column; that has not been confirmed yet.
Council votes per governor vote, by batch
Not every county ballot is from HB, and not every HB voter marks the council race. This ratio turns the county's ballots-left number into council votes left.
How the projection works
- Votes left. County ballots left x council votes per governor vote over the last batches. If the county hasn't posted its new ballots-left number yet, today's ballots are subtracted from the last posted number.
- Who gets them. The central projection splits votes left using the combined shares of the last batches that have full data for all 11 candidates.
- Chances. simulated finishes of the count. Each one varies the number of ballots left (-5% to +12%, since the county often finds more), the HB ratio, and the mix of each remaining daily batch, randomly blending the recent batches plus normal ballot-to-ballot noise. A candidate's chance is the share of simulations where they finish in the top 4.
- Limits. The model can only produce batches that look like a blend of recent ones. A batch that is much more in-person (Republican) or mail (Democratic) than anything seen lately would fall outside it.