An early step: the Florida 2022 gubernatorial forecast, before Pollstradamus.
Before Pollstradamus, there was a forecast for Florida’s 2022 governor’s race. That county-by-county map was an early experiment in turning electoral data into a prediction, and one step toward the idea of building this site.
Pollstradamus began in 2024 as a personal project at the intersection of data and electoral politics. It brings together forecasts for the presidency, Senate, and House, alongside writing about the campaigns and projections.
The aim is straightforward: Make a complicated election easier to follow. A forecast gives the available evidence a structure, with room for the uncertainty that comes with predicting how millions of people will vote.
Behind the forecast
Three sources. For the “magic.”
The forecast draws on a blend of polling, recent election results, and voter registration data.
01
Polling
State and national surveys offer a snapshot of voter preferences. They are the most frequently changing inputs to the forecast.
02
Election results
Recent results put new polling in context, grounding the forecast in how states and districts have actually voted.
03
Voter registration
Registration data adds context about the electorate. It changes more gradually than polling and does not, by itself, tell us how people will vote.
Which polls are included?
Polls must publish crosstabs—the detailed breakdowns of their responses—to be included. Surveys from partisan-aligned firms can qualify; internal campaign polls are excluded.
Looking back
The results are part of the story.
The 2024 report card looks back at the forecast’s performance, including where it fell short and what there is to learn.
