It's all a probability function. Nothing's going to be right 100% of the time. There are ways of increasing the probability (of having the correct directional bias), but in general, the higher the probability one manages to achieve with any given method, the fewer will be the trading opportunities presented by the method. Overall win-rate (which depends to a substantial extent on "getting the direction right") and overall profitability are two different parameters: it's easily possible to gain more, overall, from having slightly lower win-rates (due to slightly less accuracy of probable directional assessment), simply because of a greater frequency of trading opportunities. For this and other reasons, nobody should assume or imagine that "predicting overall direction correctly" is the holy grail of trading: trade-management is more important. Like so much else to do with statistics and probability, of course, to many people it's very counterintuitive: it would be easy to imagine that because "getting the overall direction right" is a prerequisite to making profit from any individual trade, that it must collectively be "the most important trading skill" ... but that's actually far from the truth. At its most simplistic level, and purely as a theoretical example, it's better to win an average of 10 pips an average of 6 times per day (or week, or month, or whatever) than it is to win an average of 20 pips an average of twice per day, etc.