Correlation does not imply causation.
Practice may be very correlated with success at piano. But if talented people practice more (and untalented people, like me, quit, and semi-talented people become "committed amateurs"), then the correlation between success and practice is partly caused by the correlation between practice and talent.
Practice and talent may be less correlated in other fields. Chess may be more enjoyable even if you are only semi-talented. So the practice-success correlation may be lower, even if the causal effects of practice and talent are the same.
If you want to figure out the causal effects of practice and talent, you need a natural experiment that causes one to vary but not the other.
This is a very basic social science point, and yet it is amazing how much "research" gets published that blatantly ignores it. Enough to fill many many Malcolm Gladwell books.
(And as mentioned, the effects of practice and talent are likely complementary, adding further complexity that the simple correlative studies ignore).
