After this incident we can say, they clearly violated this rule. They cut corners. However, we have no clear view of all the incidents they prevented in the past because they violated that rule. We cannot see that there were twenty other times when violating this rule in exactly this way was exactly what kept things working.
This is the nature of hindsight bias. After the fact, it's completely obvious exactly which of the conflicting rules should have been followed. It is obvious because we now know the outcome.
If you're in the thick of an incident making high-pressure, and high-consequence decisions under duress there is no way to actually predict the outcomes that will become obvious only after-the-fact.
Taking a balanced approach is good universal advice for everything. But in this context, it misses the point of blameless retros.
Our cultural norm is to stop investigations once we find the culprit who cut-corners. The consequences of that systemic habit are that we fail to learn and understand all of the complexity that actually contributed to the failure. We cut short our own learning. What's more, we also effectively teach everyone in the system to conceal or withhold their full understanding in the next incident. So we learn only the most shallow details about what actually happened for this incident, and we suppress what we might learn in the next incident.
The core of the philosophy is assuming people come to work to do a good job, and when they don't do a good job, there are reasonable explanations that involve people making locally rational decisions. The goal is to make the system stable in a way that making locally rational decisions translates to globally good decisions.
has this ever, ever happened?
The problem is that some accidents do happen because people gamble and lose – or rather, because people cut corners is ways that probably won't matter but sometimes really, really do.
The best approach needs to balance both causes of error. I agree that most organizations tend to err too far in the "blame someone" direction – but that doesn't mean that it's impossible to err by going too far in the other direction.