It's not adding a layer of complexity, it's peeling off a layer of abstraction. With the "benefit" that it describes better what we observe and that it allows to successfully make new predictions.

What's your alternative to the scientific method exactly?