The physicist can drop a stone with a certain mass from a tower of a particular height and can predict that the next time she drops that rock from that tower, it will take the same amount of time to hit the ground.
Social science, which is most definitely science, must deal with the additional complexity of dealing with weird, quirky, fickle human beings. In a real sense, you and I are not the same people we were yesterday. Rocks and towers are much less interesting from that perspective.
Some physical or hard sciences, even without people, are lousy at predicting. Meteorology is a science, and there’s a reason for the old saw that the weather man is never right. What meteorology and social sciences have in common is they study complex dynamic systems where knowing the state of the entire system at any given time is difficult or even impossible and seemingly insignificant differences in precision can produce entirely different predictions.
Sometimes it is not the scale of complexity that makes a problem tough to tackle. I believe we agree that mathematics is a science, and I’ll make the modest claim that theoretical computer science is a branch of mathematics. We cannot predict in general whether an arbitrary Turing machine given a certain input will halt or run forever. Nevermind the simple question of whether a computer program will eventually terminate: Rice’s theorem tells us that in general we cannot predict whether an arbitrary program has any interesting property.
https://en.wikipedia.org/wiki/Rice%27s_theorem
Science is more than just high school physics problems.