What trade off would you choose between fixing a performance issue which is somewhat avoidable and fixing a game breaking crash bug? Because the latter gets priority and there’s never enough time to fix all of those before launch, let alone work your way up to closing out every last frame rate drop.
I think if you get 100k poly models in your (city builder) game in the first place (for any but the most amazing wonders) your process has failed spectacularly at some point.
