District is uncontested or non-competitive. Not tracked with itemized shift variables.
Each contested district starts from a baseline partisan lean (recent presidential results blended with down-ballot statewide races), then adds a statewide generic-ballot shift, an incumbency effect, a candidate-quality rating built from each candidate's past over- or underperformance, and a campaign-finance effect. The simulated elections share statewide errors, so districts rise and fall together.
The two independent challengers (HD-62 and HD-80) get a skewed candidate-quality term instead of a symmetric one: they're modeled as essentially never doing worse than a generic Democrat would, with a long tail of possible overperformance.
Spending in each district card is voter-contact spending from the pre-general campaign finance reports (Aug 5–Oct 22): the candidate's own campaign, plus outside money from party committees and independent groups that named the candidate. The finance effect gives incumbent spending less credit than challenger spending, since much of an incumbent's strength is already counted in the incumbency bonus and their candidate history.