You said what I meant, better than I did. I didn't mean they didn't guard against tsunamis at all, but rather that they only did what was economical, which is to ignore the ones you can't afford to protect against.
Every design engineer, including myself, who has had to design a defense against a threat of unknown size, has to decide, "Well, we can afford to defend against THIS much, but if nature throws more than that at us, we fold."
Economics has to win against over-design, or else the design never gets built. Unfortunately, in the case of 500-year events, over-design could still have been insufficient. This was one heck of a tsunami.
The design process here was a bit more horizontal than some other kinds might be. The primary question related to history of threat, the rest of them were along the lines of “okay, so if this fails, then what?”. Same kind of practicality limitations you mentioned. After four or five “then what’s” then, as you said, somebody has to conclude the probability is low enough or the design never gets built. “Then what” is potentially an infinite number otherwise.