Gambit Forecaster is an election forecasting and research project developed and operated by Sam Massey, a Political Science student at Purdue University Northwest. This project is not affiliated with Purdue. The project comprises a suite of Monte Carlo forecasting systems that use structured empirical inputs and stochastic sampling to evaluate political outcomes. The election models rely on state-level vote share intervals, historical baselines, and defined uncertainty bounds to generate probabilistic distributions rather than single-point predictions. State ranges are centered on polling-based midpoints, and each model executes thousands of simulations to reflect realistic variation in voter behavior and the range of plausible political environments.
Simulations draw candidate vote shares within these intervals and apply layered adjustments representing national shocks, regional correlation, and state-specific variation. This hierarchical structure captures both local uncertainty and the correlated movement that defines modern elections, producing outcome distributions that reflect the full range of competitive scenarios.
The goal of Gambit Forecaster is to produce forecasts that are transparent, reproducible, and grounded in data, while providing a framework for evaluating uncertainty. The site also serves as an archive for applied modeling projects extending beyond elections, including systems that examine sports performance and market behavior. Each model is designed to show how changes in assumptions and inputs shift the distribution of outcomes.
The project will continue to expand as models are refined and new data becomes available. The site functions as both a public record of forecasting work and a workspace for improving the rigor, calibration, and clarity of probabilistic prediction.
