A0630
Title: Full house modeling: Rethinking fairness, extremes, and historical comparison in statistics
Authors: Daniel Eck - University of Illinois (United States) [presenting]
Abstract: Full house modeling (FHM) is a new statistical framework that addresses an age-old problem: how to fairly compare individuals across contexts that differ in size, structure, and opportunity. Originally developed to evaluate Major League Baseball players across eras, FHM has since evolved into a broader modeling paradigm with applications ranging from historical impact rankings to novel measures of extremity and fairness. A conceptual overview of full-house modeling is provided, with a particular emphasis on recent developments. It discusses how FHM enables context-adjusted rankings of historical figures and introduces paradox probabilities as a principled, distribution-free tool for quantifying the likelihood of mismatches between latent aptitude and observed achievement. A recursive extension of the model is also highlighted and, time permitting, share insights from an application to MLB batting averages before and after recent rule changes.
