I fully agree with the advice to not allow yourself to be tied up by a stranger when alone. But the fact that the kid didn’t get murdered is evidence to the contrary. Elementary Bayesian reasoning.
If your initial estimate (the prior probability) of the chance of getting bad things done to you by a stranger is not excessively high to begin with, then “bad things not happening” is only weak evidence. However, if you are convinced that you have a 99% chance of being raped in that situation, a single case of “nothing happening” should make you lower your estimate significantly, in the absence of any other information.
A stranger with a straightjacket is a rather extreme case, but the same faulty reasoning can be applied whenever the evidence doesn’t fit your preconceived opinion.
Example: “If that particular member of minority group X didn’t cheat/rob/kill you, that’s all the more reason to be more careful around them next time”. Done. Prejudice reinforced.
And that’s why I am opposed to faulty reasoning even if I agree with the advice not to let yourself be tied up by a stranger. Faulty reasoning can and will be used against people if we allow it to thrive.
