Innovation Series: Advanced Science (ISSN 2938-9933, CNKI Indexed)

A Fairness–Excitation Tradeoff Study for Dancing with the Stars: Fan Vote Inference, Method Comparison, and a Hybrid Scoring System

Yanzhuo Wu, Lvheng Yang, Guangwu Ao*

School of Computer Science and Software Engineering, University of Science and Technology Liaoning, Anshan, 114051, China

Corresponding Author: Guangwu Ao

Abstract: Televised talent competitions often combine expert judging with audience voting, but the aggregation rule can create outcomes perceived as unfair when popularity overwhelms technical merit. Using 34 seasons of Dancing with the Stars (DWTS), we develop (i) a constrained inference model to estimate weekly fan votes consistent with observed eliminations, (ii) a season-wide comparison of the historical rank-based and percentage-based aggregation rules, and (iii) an optimized hybrid scoring system that explicitly balances fairness and audience excitement. The vote-inference model is posed as a quadratic program with elimination-consistency constraints and produces an overall elimination match rate of 86.3% across 337 elimination weeks. Uncertainty is quantified via feasibility-based 95% confidence intervals, which are narrower for high-visibility finalists and wider for low-profile contestants. Method comparison shows the rank-based rule amplifies the relative influence of fan votes and is associated with a higher controversy rate than the percentage-based rule. Finally, we propose a weighted hybrid score combining normalized judge rank with fan-vote share; an optimized weight (α=0.58 for technical merit) reduces controversy while retaining high vote dispersion as a proxy for viewer engagement. The framework is implementable with standard production data and provides a principled path to reduce recurring controversies without eroding audience participation.

Keywords: Dancing with the Stars; Fan vote inference; Rank aggregation; Convex optimization; Voting-system fairness; Audience excitement; Hybrid scoring

References