"Success" is a can of worms all its own, but lets assume for a moment some definition we can agree on, some people who have achieved something matching this definition, and a desire to match their achievement. If we find that a few people who "succeeded" in this way all did X* in their youth, what does that tell us?
It tells us that people who did X and then succeeded, did X. In other words, it tells us nothing at all. It's a tautology.
What if hundreds of thousands of other people also did X, and among them are thousands who wished for the same success, but failed? What if those thousands would have succeeded if only they hadn't done X? None of them succeeded, so we wouldn't hear about them.
If we can either survey the majority of all people who have had a specific kind of success, or we can take a truly random sample of them, and then find that having done X in their youth is something a disproportionate number of them have in commone, then it would be worth investigating these two hypotheses:
- Doing X gives you a significant advantages in achieving this sort of thing.
- The sorts of people who are likely to try this sort of thing, or likely to succeed at this sort of thing, also have a tendency to do X.
It's possible that both hypotheses are false, but we've at least established a reasonably likelihood that at least one of them is true.
But no random sample? No broad survey? Then the mere fact that someone successful did whatever it is they did, tells us nothing whatsoever about whether doing that thing helps you have that kind of success. It may even hurt your chances.
* Feel free to interpret X as literally X, if that amuses you.