In this paper we examined adaptive markets hypothesis (AMH) using three factors we assumed that affect weak-form of market efficiency: observation period, time horizon represented by rolling window sizes and data aggregation level. We have analyzed market value weighted index MONEX20, which is proxy from Montenegro equity market, over 2004-2011 period. Rolling window analysis with fixed parameter in each window is employed to measure the persistence of deviations from a random walk hypothesis (RWH) over time. Actually, using rolling sample approach we checked whether short-range linear dependence is varying over time. This method was applied on the first order serial autocorrelation coefficients (AC1), as well as on runs test, since evidence on non-normality properties of MONEX20 suggests using non-parametric test. The evidence was found that all three factors impact degree of weak-form Montenegro equity market efficiency which has serious consequences on profit opportunities over time on this market.