Resampling methodologies have recently revealed to be very fruitful in the field of statistics of extremes. First, we mention the importance of the Generalized Jackknife to reduce bias. We next refer to the relevance of the Bootstrap in the estimation of a crucial tuning parameter in the area, related to the number k of upper-order statistics involved in the estimation of tails. Moreover, most of the estimators of parameters of rare and large events, among which we distinguish the extreme value index (EVI), are averages of statistics, based on k. Only for heavy tails, quite common in many areas of application, classes of reliable estimators of parameters of rare events based on adequate generalized means (GMs) are introduced and discussed. Together, these two resampling procedures and GMs enable the obtention of reliable semi-parametric estimates of any parameter of extreme events, like the primary parameter of extreme events, the EVI.