Risk vs "statistical risk" is a distinction without meaning. "Statistical" is just a way of describing risk, not a different kind of risk.

I have wondered why people tend to erroneously believe that an observed outcome demonstrates that there was no risk, and something just now occurred to me. Triggered by that phrase "just statistical risk" it strikes me that people are thinking of risk as something inherent in a given situation. If everything follows physical laws (forget quantum mechanics) the outcome is always determined by the initial conditions. If nothing bad happens, then they reason that risk was zero. If something does happen, the risk was 100%. I think, when you get to the bottom of it, that's how people are seeing "risk" when they reason this way. Am I wrong?

That's not really what physical risk is though. It's not a physical property, not an inherent system state. Risk is a description which depends on what you don't know about the system. If you know every physical characteristic of a quarter, and of the surface it falls on, and the precise magnitude, position and direction of the force you use to flip it, a precise enough calculation would tell you whether it will land heads up or tails. There would be no risk in betting heads if you know it will land heads. But only because we don't know those things, we have a 50% risk of a bad outcome. We don't know, because the situation we describe is "flip a coin" and that includes ALL the variations of variables that we don't know.

The risk from a close pass is like that. Exactly like that. The risk of a "close pass" was no less just because the last one missed us. The risk of the next close pass is unaffected by the result of the last one. The "risk" that we're describing is the chance of *any* close pass causing an accident. Describing it more discretely does change the risk, but the outcome of one does not.