Picking which shares to buy and sell is not necessarily about being right all the time.

From an article:

This leads into the second finding: picking which shares to buy and sell is not necessarily about being right all the time. Even the best fund managers are frequently wrong, and Lagrange estimates his team are correct between 45 and 65 per cent of the time. The trick is to make more money when they are right than they lose when they are wrong. From Savage’s analysis, he says his team makes 1.1-1.2 times more when right than they lose when they are wrong – enough to make good money even if their odds of being right are no better than 50/50.

Trend following 101.