2028 Electoral College Simulator
An interactive Monte Carlo simulator of the 2028 U.S. presidential election. Each update runs 10,000 simulations with correlated state errors and returns win probabilities, a distribution of electoral-vote outcomes, and the most likely tipping-point states.
The model
State errors are generated from a factor model with a default national correlation of ρ = 0.70:
error_i = ρ · national_shock · elasticity_i + √(1 − ρ²) · state_shock_i
predicted_margin_i = state_lean_i + national_margin · elasticity_i + state_override_i
State leans are computed from each state's average deviation from the national popular vote in 2016, 2020, and 2024. Elasticity measures how responsive a state is to national swings. Nebraska and Maine congressional districts are modeled separately.
Parameters
- baseCorrelation: 0.70
- stateErrorSD: 4.0 (low), 5.5 (medium), 8.0 (high)
- nationalErrorSD: 3.0 (low), 4.5 (medium), 6.5 (high)
Limitations
This is an educational tool, not a real forecast. It does not integrate polling data, candidate quality, or economic fundamentals, and state leans are static (no trend modeling).
Related projects
By G. Elliott Morris.