My girlfriend has the ability to slit my throat at night without me having any way to resist, but I still think being around her is less risky than, say, walking across the freeway. A hopeless situation that is extremely unlikely to occur is less of a risk than a slightly-less-bad situation that is very likely.

In practice, it seems like the difficulty of getting eyeballs and converting them into sales is generally much more likely to kill your app than having Apple reject it. Optimizing for the latter case is not rational unless you have a strong reason to believe your app will be rejected.