A paradoxical argument about domination (Q2297102)

The authors consider a well-studied problem: the Bayes estimator of the mean vector \(\Theta\) of a multivariate normal distribution under uncertain prior information when the covariance matrix \(\Sigma\) is unknown. The plug-in estimator is used in this study. The empirical Bayes estimator is compared with the best estimator described in [\textit{M. S. Srivastava} et al., J. Multivariate Anal. 96, No 1, 55--72 (2005; Zbl 1074.62039)].