Case Study: Cornell University achieves refined probabilistic U.S. Senate midterm forecasts with Palisade's @RISK

Cornell Professor Takes Nate Silver Analysis of U.S. Senate Midterms a Step Further using Monte Carlo Simulation

Cornell University’s Johnson Graduate School of Management professor Lawrence W. Robinson set out to refine Nate Silver/FiveThirtyEight’s 2014 U.S. Senate forecast by quantifying the probability that Democrats would hold at least 50 seats. The challenge was to move beyond simple independent-race assumptions and model varying degrees of correlation across races, a task he addressed using Palisade’s @RISK Monte Carlo simulation software.

Using Palisade’s @RISK and FiveThirtyEight probabilities, Robinson ran thousands of correlated Monte Carlo scenarios to produce distributions and probability bounds: with 0% correlation the Democrats’ chance to retain control was 41%, with 100% correlation it was no more than 50%, and for correlation coefficients between 20%–85% the probability tightened to about 45% ± 0.19. Palisade’s tool made it practical to model intermediate correlations, quantify the impact, and demonstrate the method to students.

Cornell University

Lawrence W. Robinson

Professor of Operations Management