Even if you stretch Occam's Razor past it's limit and use it to establish preference between two theories that make different predictions but only about phenome

Even if you stretch Occam's Razor past it's limit and use it to establish preference between two theories that make different predictions but only about phenomenon that you personally can't practically check (i.e. a case such as this one), then you are still only establishing a preference for the "simpler" explanation and cannot actually rule out the other explanation without checking the difference in predictions (i.e. an investigation into the crash.)