A simulation takes three inputs: the institution as it actually operates, the regulatory environments it sits in, and a specific decision it is considering. It returns the consequences of that decision, the alternative configurations that would reach the same commercial objective, and an honest statement of where the answer depends on judgement rather than fact.
The unit of analysis is the decision, not the obligation. That is what separates simulation from research and from compliance tooling: the question is not 'what rules apply to us today' but 'what would change if we did this'.
For US financial institutions the practical value is comparison. Two designs that serve the same commercial goal can produce materially different positions, and the cheapest moment to discover that is before build.