The SAT isn't strong enough to predict anything. It can generally be answered by someone in their sophomore year at college or even their freshman, depending on what level of courses they are taking.

The problem finding a hard enough test with as little human intervention for assessments. Because human intervention brings with it subjectivity. This subjectivity was manageable when there weren't so many people applying for top schools (e.g. in the early 1900's). But right now its not.

SAT/ACT/GRE are no indicator of success. What this "study" is merely proving is that schools may have regressed in their rigor for grading hard courses.

Why does it matter if a college student, after three years of education, can do well on the entrance exam? Isn't that a given?