The self-destruct is pointless, because you can't even verify that it was successful and you have no actual technology to protect (unlike say, US military drones which self-destruct to protect technological details of their construction - and the US would still prefer to commit a mission to bombing a crash site to be sure if they can).
Basically you answer your own problem: the most likely outcome of a self-destruct mechanism is that it goes off accidentally against your own forces, or gets exploited - while it would deliver no actual benefit to you.
Which is to say: when analyzing an adversaries likely actions you don't start by assuming they're stupid or irrational (which is different from whether or not their overall goals might be stupid or irrational).