I approve of your response, Douglas_Knight, but think that it is both incomplete and somewhat inaccurate.

The cost of underconfidence isn't necessarily or always an opportunity cost. It can be so, yes. But it can also be not so. You are making a subtle and mostly implicit claim of universality regarding an assertion that is not universally the case.

A strategy doesn't need to work in every possible contingency to be useful or valid.