It seems like most of what you are saying is that you gain optimizations by, well... removing details from the game. I'd generally agree that running multiple sprite operations per frame is inefficient, but in this context it's also a creative solution to the problems coming from trying to make a 2.5D game in a 2D engine.
If I was doing this in Unity, I would just throw a pixel shader on a 3D scene and call it a day, but that seems to defeat the purpose of doing fun pixel artsy stuff by hand, which seems to be the dev's intention and driving motivation.
Most efficient ≠ artistic intention and engine limitations