Brookline Override 2026 โ Election Night Projection
As precincts report, a ridge regression fits 2026 results against five historical baselines: 2023 Q1 (Pierce), 2023 Q2A (Operating), 2023 Q2B (Override+Composting), 2018 Q1 (Operating Override), and 2018 Q2 (Debt Exclusion). 2018 data helps control for 2023 P6 anomalies (Pierce school in P6 inflated Q1 turnout there). Prior is weighted toward Pierce 2023 + 2018 Operating Override for a slightly pessimistic null hypothesis. Updates as precincts report.
๐ก Type 2026 Yes & No vote counts in the yellow cells as each precinct reports.
Model
Ridge regression initialized at prior โ waiting for precinct results.
Projected YES
โ
Projected NO
โ
Projected Margin
โ
No precincts reporting yet
How to use: as each precinct's results come in, type the 2026 Yes and No vote totals into that row. The "Reported" checkbox auto-checks once you enter both. Unreported precincts are projected from the 2023 baseline plus the average swing of reported precincts (weighted by current turnout). To override expected turnout for a precinct (e.g. you hear that turnout is way up), edit the "Expected Turnout" cell.