You hear this predict/react distinction a lot. "I've given up trying to predict. It's enough to react successfully." And yet to react systematically we must begin with a set of predictions. What am I missing? answer1: You are missing the marketing factor, rhetorical argumentation. For 8 years I had been meeting clients discussing hedge funds. Clients have objections: But markets cannot be predicted, how can you make money in them? By trial and error you find that the contradictory statement: we do not predict markets, we react to them satisfies all investors who suspect it might be something deep (instead of the nonsense it is) and allows one to move over this objection and to go on to sell one's stuff. This contradiction makes sense to most people but then rational thought is not innate, it has to be taught, and it is not in most of the world. answer2: but still being wrong 25% of the time is a lot. It should not be considered trivial. Some questions need to be answered I think. Should we say that a random price series has a 75% chance of being declared random and a non random price series has a 25% chance of being random! Or should an experiment in head or tail be declared as non random because we observed sustainable trends in the series? 上面摘自wld的forum