In some problems. however. B will be able to do just as well as A. In such a problem, the single tunction T = r(XXn) will in some sense summarize all the infomation contained in the random sample. and knowledge of the individual values of XXn will be irrelevant in the search for a good estimator of 0. A statistic T having this property is called a sufficient statistic. We shall now present the formal definition of a sufficient statistic