This is something I struggled with a lot. Internal me would scream, show me the numbers!!. But when I go into mathematical modelling, too many assumptions make

This is something I struggled with a lot. Internal me would scream, show me the numbers!!. But when I go into mathematical modelling, too many assumptions make the model very.. basic. (Like you say) However, if you aren't trying to control the system, but merely influence it - this helps figure out the leverage points, and what will happen with perturbations on said leverage points. As far as understanding a system goes, that is good enough to use it to your advantage. A formal discipline gets detached from reality (mathematical modelling). This I think rests somewhere in the middle. Or, atleast, that's what I've come to see.