Is VARS more intuitive and efficient than Sobol’ indices?

Abstract The Variogram Analysis of Response Surfaces (VARS) has been proposed by Razavi and Gupta as a new comprehensive framework in sensitivity analysis. According to these authors, VARS provides more intuitive notion is much computationally efficient than Sobol’ indices. Here we review arguments critically compare the performance VARS-TO, for total-order index, against Jansen estimator. We argue that, unlike classic variance-based methods, lacks clear definition what an “important” factor is, show that alleged computational superiority does not withstand scrutiny. conclude while enriches spectrum existing methods analysis, especially diagnostic use mathematical models, it complements rather replaces estimators used