Accuracy
Is the model any good?
Comparisons against observed elections and published polls, including misses and contamination limits — not a promise of future accuracy. A simulation, not a poll.
Observed benchmark
A published measurement from a named source — Pew, AP VoteCast, or a certified election result.
Modeled estimate
What this engine produced for the same question, over the same period.
Miss
The point difference between the two. Signed in the record below, not just absolute.
Loading committed validation results…