Abstract
We define the unit gamma-generalized family as a very competitive class to the well-known beta, Kumaraswamy, generalized odd log-logistic and odd Weibull generators, among others. We obtain some of its mathematical properties. Two simulation studies investigate the accuracy of the maximum likelihood estimators in members of the proposed family. We prove empirically the superiority of a sub-model of the new family over those of four generators by means of two real applications.