Your "1ST rule of Rocket Engineering" can also be stated: You always develop sub-optimal rockets.
Seems like a stupid rule to me.
If an engine goes out, or there is some other problem, you need extra fuel to accomplish the mission (increased gravity drag). So you have some extra fuel and extra delta v, and that's a good thing.
But if those events are rare -- and, eventually, they should be -- then you often have extra fuel. If you can use that fuel to return the craft intact to reuse and make more money, then I think that's a damn good idea. If you must burn the extra fuel, then you will lose the stage. It will cost the company more, but "less profit" is maybe an OK choice.
The goal is to optimize cost while maintaining very high reliability. For very high reliability, you need to understand worst case behavior. For optimizing cost, you need to make the common case cost efficient. Having extra delta v for anomalies and using that delta v to lower launch cost (via reuse) when no problems arise seems like smart engineering to me.