I really enjoyed your argument. But why did you weaken it by reintroducing systematic cause:

>> if everyone in your family lives a really long time, there's a good chance you'll live a really long time too

This implies assumption of some sort of cause, which one would then expect to hold in other parts outside that cluster, too. Which I understood as: the island's longevity could be random -- but not really.