That's not true in general. As the article notes, a "1000 year event" is an event which has a probability of occurring of .1% per year. Most disciplines of engineering prevent failures due to events with much lower probability than that.
The real issue, as you noted, is the difficulty and expense of building solutions to low probability events in civil engineering. It isn't hard to figure out how to prevent damage from a rarely occurring phenomenon, but it is hard to justify spending the money on the solution when everyone thinks it isn't likely to happen.
Most engineering solutions for flood control are based on a 200 year event. even those solutions would be insufficient for a 1000 year event