While I share a degree of your skepticism about models, i.e. GIGO is real, I think you are being far too dismissive. It has been demonstrated that models do have value and predictive power.

Watch a weather forecast lately? They are amazingly accurate compared to when I was a child. That's because the models got better and our data collection ability improved.

In "ultra complex" systems, where exact results depend on "chaos seeded" variables, models may never produce ""exact match" results with reality. Doesn't mean they are worthless, they give you the "shape" of things to come.