In theory technical analysis may work because of three reasons: 1) herding behavior, 2) prices being partially revealing and 3) sentiments. Our empirical cross-country evidence from 50 countries shows that technical analysis indeed works particularly well in countries that score highly on different proxies for herding and information uncertainty. As a crude indication our combined set of proxies explain around 65% of the variation in cross country profits from technical analysis strategies. Our result may explain why studies using data from different countries find mixed results on the profitability of technical analysis.