The purpose of this paper is twofold: (a) to determine whether there is value premium in our sample of US stocks for the period May 1, 1969–April 30, 2011; and (b) to examine whether an additional screening to the first step of the value investing process can be employed to separate the outperforming value and growth stocks from the underperforming ones. In this paper, we document the following: We find a consistently strong and pervasive value premium over the sample period. We show that there are distinct differences between US exchanges which means that papers that aggregate all US exchanges under one umbrella may dilute findings and bias conclusions. The stocks of AMEX firms, high business risk firms and firms that report extraordinary items experience worse returns than the rest of the US stocks in our sample. We find that P/E based sortings produce better overall results than sortings based on P/B. We are able to construct a composite score indicator (SCORE), combining various fundamental and market metrics, which enable us not only to separate the winners from the losers among value and growth stocks, but also to predict future returns of value and growth stocks. SCORE portfolios give better results for sortings based on P/E and when we employed a cross-section–time series medians approach. Results remain robust for a time period out of sample, for negative P/E or P/B ratio firms and for the firms that were excluded from SCORE-based performance, namely, AMEX stocks, stocks with high business risk and firms that reported extraordinary items the year before. Finally, we provide evidence that the return of a portfolio strategy that buys (sells) stocks that rank low (high) in the composite score indicator has significant explanatory power in an asset pricing model framework and that such a strategy earns statistically significant positive returns.