Unfortunately, given I have only second-hand knowledge of these reports' existence, and given they are probably not in my native tongue, I'm not really in a position to validate what the situation was vis a vis the generators. However, generators positioned on the coast, behind a low seawall, in a country that is regularly inundated by tsunamis, appears to be an entirely predictable failure mode. Not quite to the degree of not bringing a parachute on a skydive, but in that direction.
It's like in New Orleans, where the emergency generators were located in the basements below sea level. There's nothing subtle about that. Sorry, but how daft can one get?
I disagree. Failure analysis does not start with "assume the seawall will never fail." It starts with "assume the seawall fails. What are the consequences?"
Given the location of those generators, it was obvious that a breach would take out CRITICAL generators, and it would have been cheap to remove that vulnerability.
Another cheap remedy would have moved the hydrogen vent pipe to vent outside, rather than INSIDE AN ENCLOSED SPACE FILLED WITH SPARKING ELECTRICAL EQUIPMENT.
The actual tradeoff would have been not just this one mitigation, but all the mitigations for all the equally probable risks. Including all the assessments to find out what those risks would have been.