In the present work we study the stochastic dominance portfolio e ciency measures. The investor's risk attitude is given by the type of an utility function. If this information is unknown or a general investor is assumed, it is possible to use the stochastic dominance principle, in which the portfolio is only classi ed as e cient or ine cient. We build on the works of Post, Kuosmanen and Kopa, who formulated the criteria of portfolio e ciency for nonsatiate and risk averse investors. On the basis of these criteria, we de ne the second-order stochastic dominance (SSD) portfolio e ciency measures. We examine the properties of SSD ine ciency measures, which allow to compare SSD ine cient portfolios. We prove mutual relationships for the de ned SSD ine ciency measures. Eventually, we test the SSD e ciency of a US market portfolio on real-world US Stock Exchange data.