If you can find someone descended from the same common ancestor and get them to take the test, it would be a strong indicator. Add the autosomal in and a paper trail. For near generations, autosomal probably works better. For further back, Y is more stable. I'm currently working to find a couple of direct male line relatives from earliest known ancestor in North America and some from a suspected father or sibling. This goes back 7 generations. Then I would know if they were relatives or not. This is a paper trail that, while not difficult, is time consuming and I get distracted.