Some people greatly enjoy creating architectural works of art. Other want to give their dwarves best lives possible. Still others just enjoy designing things that work well. Some can have fun without threats.
Others, however, want to see threats and want to be tested. You can ask questions like "how much can I cut the "budget" on my secure entrance until it is no longer secure" or "how dangerous are my dwarves, comparatively?"
Or perhaps you like the to see the pretty teeth fly, enjoy blood-soaked ground aesthetically, think some flawed designs are still pretty (like a castle with water moat and raisable drawbridge).
I agree with the advice of deliberately doing things wrong on a learning principle, as long as one also does it right before or after. It's instructive how important something is, and doing something with intent along the lines "lets do this and see if/how many dead dwarves I get" is a lot different - I'd say freeing, even - feeling compared to "each and every dwarf is important, I must not let any of them die or I'll have failed as overseer".
With most dwarven civilizations and most problems, you can always throw more migrants at a problem.
(Still, that sewer is novel, given I've seen several succession fortresses abandon their upper levels because of half a dozen different sealable entrances, so twenty dozen unsealable entrances for 1vs1 sleeping dwarf vs titan/werebeast combat would certainly make one more keenly consider the rate of higher-quality threats.)
Anyway, some tags to consider: Flying, trapavoid, building destroyer, thief, merchant, visitor. A completely secure entrance should be able to not fall to any and all of the first 4 at the same time, while letting in the last two.
(Though trapavoid can almost be ignored unless you go to war with someone you used to trade with, I think - gremlins aren't much of a threat, for instance.)