This assumes we can come back from the failure without major consequences, and we'll still be in a position to leverage the learnings from that failure. I think the OP is pointing to one of those scenarios where that's less likely.