Empirical Bayes estimation: When does $g$-modeling beat $f$-modeling in theory (and in practice)?

Re-assign community