That's certainly the common thought (that it would be impossible), but does it hold up in practice? Let's say a repeat of 9/11 happened, then judging by expect

That's certainly the common thought (that it would be impossible), but does it hold up in practice? Let's say a repeat of 9/11 happened, then judging by expected outcomes it would definitely be better to all rush the hijackers, even if half the plane died in subduing them it would be better than everyone dying (and more on the ground) if you let them do what they want. Today's example on the other hand, the best case scenario is what ended up happening (the hijacker/s don't have the goal of killing people, so there's a reasonable chance that everyone can survive), therefore is it worth taking the risk to stop them? In this situation, even one person dying while preventing them from taking control of the plane can actually be a worse outcome than leaving them to their negotiations. In terms of mathematical outcomes it's hard to say which course of action would be better: firstly because you couldn't give accurate probabilities, and secondly because the expected outcome might not match the best bet. For example, if preventing hijackers gives a 95% chance that at least one person dies, and leaving them to it gives a 1% chance that 50 people die, then played out over 1000s of times you're actually better not doing anything, but over a single time is it worth taking that risk? And then past that, even if these numbers were easy to analyse, how will people on board react when a.) emotions are running high and b.) rather than typing at my desk at work, I'm confronted with the situation that I'm the one risking my life to stop them. Even if I knew the odds meant they really needed to be stopped, I'm not sure I'd have the balls to run at men with weapons.