I suppose the distinction is that you didn't necessarily see it coming but caught it when it was already in some stage of formation. But, as it relates to trading, you have no idea how well or how long that wave will endure. So you ride it as it is happening, rather than predicting how far it will go or whether it will be a smooth ride. Much like driving a car in unfamiliar territory, you react to the circumstances as you see them occurring. I don't normally associate driving with predicting; there may be a rare moment of "intuition" here and there based on years of experience, but mostly, you're reacting to your environment on the road.
You are correct, it's hard to gauge beforehand, counterproductive even, what to expect from markets. Trading is very much reacting to everchanging present conditions.
However, you still need to correctly
predict something in order to have the probabilities on your side (being on the right side of the market).
Any random reaction is simply not going to cut it!
Driving a car is a good analogy to this. While driving, you not only have to react to minute changes of direction, velocity, road position, viewpoint from the car, driving conditions, etc. You also need to correctly predict what's happening around you: other cars, their intentions, signals, people walking in the road, bicycles, jaywalkers, animals, clutter in the road, accidents and catastrophic events. For many of these, the driver need to correctly gauge and anticipate the decisionmaking and capabilities of others, not just slavishly following traffic rules, but correctly predict and adjust according to unwritten physical and social laws, in order to avoid adverse effects and consequences.
It is not merely reacting to changing conditions, because it may be undefined what the correct reaction is, according to traffic laws and laws of the land, but drivers are still expected to make the correct decisions and are made accountable for unnecessary accidents and damages. Humans are still working on computer systems that can do this in realtime, but it requires alot of raw sensors and TPU power. There's no definite solvable answer to the equation in sight, as there are no definite requirements defining the final end goal.