In an asset allocation problem, the investor seeks the combination of securities that best suit his needs in an uncertain environment. In order to determine the optimum allocation, the investor needs to model, estimate, and access and manage uncertainty. After a brief lull in literature on asset allocation, there have been considerable advances over the last fifty years. One of those responsible was Markowitz who presented the mean—variance approach (MV), as a modern and widely used tool on portfolio selection. However, MV assumes certain statistical performance of the stock markets which makes limitations for model application on small and emerging markets. Can MV or other modern assets allocation models assist investors on emerging markets in portfolio selection? How great errors investors are facing with in such cases? Would traditional 1/N allocation rule outperforms substantially the mean-variance portfolios?—questions are that this paper will try to answer. Research conducted 10 years ago, when standard MV approach was implemented on Montenegrin stock market and results obtained were compared with traditional 1/N approach, will be presented. The mean-variance optimal portfolios were constructed and overlooked measures of performance of this sophisticated strategy compared with the most naïve—1/N strategy. After 10 years since model was developed, author is in position to test the results in the real time. The main findings from the empirical datasets will be presented in this paper showing that emerging equity markets provide a challenge to existing models, however could be successfully applied using some extra features.