We derive a stock ranking by applying a technical features based random forest model on an international dataset of liquid stocks. We show that portfolios based on an outperformance profitability ranking are more profitable than those constructed on the basis of predicted returns. When applying a decile split, equally (value) weighted long-short portfolios achieve a highly significant yearly six factor alpha of 23.49\% (17.51\%) and a Sharpe ratio of 2.37 (1.95). Unobserved risk factors identified via RP-PCA may not explain the outperformance. Moreover, we show that outperformance probabilities serve as a superior measure of future returns. Mean-variance portfolios of large stocks using our return measure are less volatile and more profitable than equally or value weighted portfolios. The results are not explainable by limits to arbitrage as they are robust to firm size, regional restrictions and non-crisis periods.