In actual reality you will never have any certainty about the outcomes of interactions like these. All you have control over is whether or not you hurt the person in front of you.
Because it reveals how you implicitly value human life and whether values can be traded or weighed against each other. There are moral theories that take various different positions on each permutation to the possible answers on these questions.
Thought experiments are about illuminating implicit values and thought processes.
I tend towards believing that people who bring up the "WHAT IF NOOK BOMB" type scenarios are really just looking for justification that their existing belief in torture is A-OK.
It's a fallacy because it's like a DOS attack on the argument - burning up most of the time on a absurdly small probability. Participants and viewers tricked by thinking "what's discussed most is what's most important".
[1] https://www.logicallyfallacious.com/tools/lp/Bo/LogicalFalla...
To see how debate can backfire, consider a currently-near-universal ethical proposition: it is unacceptable to torture innocent children. The contrived scenario for debate: there's a ticking nuclear time bomb hidden in an American city. The man who has the critical information can hold out against physical torture for the next few hours until it detonates. But he can be compelled to confess in time if you force him to witness the torture of his 5 year old niece. Imagine that the scenario has both options debated, repeatedly, in lecture halls, on op-ed pages. It's even depicted in fiction like 24 to persuade by storytelling rather than logic.
It doesn't matter that the world's actual torturers have never encountered a ticking nuclear time bomb, much less the improbable other circumstances that led to this dilemma. Just witnessing sober, respectable persons arguing the "in favor" position for innocent-child-torture -- no matter how dryly or laden with caveats -- is going to tend to weaken norms against innocent-child-torture. So will seeing the tactic succeed in fiction. (It doesn't really matter if excellent arguments against innocent-child-torture are also aired in the same venues. Practically everyone started out against it, even if their reasons were visceral rather than logical.) Some percentage of people, hopefully small, is going to come to "see the light" in favor of this contrived scenario's utilitarian calculus.
The really pernicious part: as with the vanilla ticking time bomb scenario, many of the persuaded will misapply the contrived scenario to justify real world torture. "The torturers know that an attack is imminent, and torturing this man will save a million lives" will get watered down by homeopathic magnitudes to "the torturers suspect that an attack is going to happen at some indefinite time, and torturing a child related to this man might save 20 lives if the torturers' guesses are correct." The effect on the body public of debating contrived scenarios that could justify atrocities is not to improve the public's reasoning or shore up their pre-existing intuitions. It's to erode their actual ethical judgment, to make them more open-minded regarding atrocities against children.
There is obvious value in these "contrived scenarios," as has been proven innumerable times in human history. You do realize that simplifying assumptions like "assume a spherical cow" have worked to produce an accurate model of the physical world, right?
Errr I think you're "proving too much" so to speak. Having someone in front of you who you know has details about a terrorist plot that could stop it, and you think finding those details is something you can't reason morally about? By your definition of certainty, you basically can't do anything with any degree of certainty unless its actions are immediate.
Call up an assassin and tell them to murder someone? Well who knows if she'll even go through with it, so that's morally fine.
(I'd also be grateful if you want to tell me what I'm getting wrong.)
And if you're in the same city as the bomb, whatever you can possibly do to them has a finite timespan and is therefore much more tolerable and resistable.
There's a shooter actively killing people.
Is it OK to shoot the shooter?
---
What's the benefit of doing physics problems "in a vacuum", etc? It lets you get to the essence of the question.
Similarly, a philosophical thought experiment lets you analyze your beliefs - rather than just a knee jerk ("No, I would never harm someone and anyone who harmed someone is not a morally decent human being"), you can come up with a more complete ethical framework for your decisions and judgments.
By identifying extremes, you can identify your priorities and understand the tradeoffs you will actually be faced with - what degree of harm are you willing to cause, and at what distance from "in front of me", to what benefit?
Now we're trying to establish the boundaries of that.
If your only means for investigating the nature of human relationships is via intellectual contrivances then you will always produce results which reaffirm your preexisting beliefs.
The first point to understand is that violence is a dicey business and never produces as tidy of outcomes as the intellectual justification for it expects. The second point to understand is that the pursuit of morality via intellectualization is counter-productive at best.
Do you have a better framework for pursuing morality? It seems like telling stories, asking questions, and thinking about them are the basis of all moral and ethical teaching, from religion to law school.