This study compares an inherently fuzzy decision framework-the Characteristic Objects Method (COMET)-and a traditional fuzzified methodology, Fuzzy TOPSIS, to evaluate how their underlying mathematical foundations influence gaming CPU rankings. Utilizing an empirical comparison, thirteen modern processors are evaluated across six technical and economic criteria: core count, thread count, thermal design power (TDP), cache, price, and average gaming frames per second (FPS). Architectural criteria are represented as crisp triangular fuzzy numbers , while FPS and price are modeled as fluctuating triangular fuzzy numbers . The Entropy Weighting Method (EWM) utilized for objective criterion weighting, which is integrated into Fuzzy TOPSIS, and a Simple Additive Weighting (SAW) benchmark solved via GAMS. Furthermore, a multi-scenario sensitivity analysis involving a (±10%) perturbation across price, TDP, and FPS was executed, alongside rank correlation measures (Spearman’s , Kendall’s τ, Goodman-Kruskal’s γ and WS coefficient). Results reveal a significant ranking divergence; Hierarchical COMET ranks the AMD Ryzen 7 9700X first, whereas Fuzzy TOPSIS prioritizes 3D V-Cache models such as the AMD Ryzen 7 5800X3D. Both models demonstrate robust stability under parameter perturbations, maintaining high rank correlation coefficients across all scenarios. Consequently, this paper demonstrates that final decision outcomes are more sensitive to the chosen methodological structure than to variations in input parameters, providing critical theoretical and practical insights for multi-constraint hardware selection problems.

Keywords

Fuzzy MCDM, Fuzzy TOPSIS, Hierarchical COMET, Entropy-weighting, Hardware Selection.