The only remaining condition is a valid understanding of "sufficiency" of evidence, which never seems to have a clear and convincing definition. Still, Popper and others have done well with falsifiability/testability. Those things certainly can be applied in the ship owner's situation; his theory that providence will protect a ship from being wrecked can be tested by examining the history of shipwrecks for instances of undeserving people dying in them. His theory that a boat's seaworthiness is constant and unrelated to wear, tear, and age can be tested through the same kind of investigation; search for ships that have sunk, after many successful voyages, due to poor maintenance.
In this analysis, "sufficiency" is defined through the application of a crucial and unbiased test, or through the examination of such evidence as already exists from earlier tests (whether those were practical tests or instructive examples from real situations).
The infinite regression spiral for this particular philosophical tangent begins right about when we start to try to define "crucial" and "unbiased." But the fact that a perfect definition eludes us is not an excuse for lazy thinking.
And Deadmeat, I think we're going to like you too.