It makes sense that the first steps toward developing a sufficiently robust theory is making sure your model can recreate the various phenomena that you are int

It makes sense that the first steps toward developing a sufficiently robust theory is making sure your model can recreate the various phenomena that you are interested in modelling, no? The goal is to get to a theory that reflects the behavior of the universe, and figuring out how to emulate the various idiosyncracies inherent in how the universe evolves over time (finite speed of information propagation, etc.). Finding testable aspects of the model is only possible after actually applying the theory, seeing where it falls short, and identifying ways to validate ambiguous conclusions.