We have recently posted how assumptions, left unquestioned can damage a project. It is similarly true for the product when we use models and simulation to generate our product.  In the course of building these models, we will know some things for certain.  Some attributes of the model we may think we know for certain but may in fact not be valid.

That is why we couple testing with our simulation and models. We will confirm those things we believe to be true and the many things we hope to be true, and probably uncover things we did not even consider to be true through testing.  Those attributes that are connected to the desired outcome, we must understand so we can accurately predict the outcome.  Below find a graphic that shows the process for building and understanding the model[1]:

Simulation Process

Assumptions are deadly when it comes to models.  If we believe our simulations and the many assumptions that went into creating are valid and just start building the product we have not learned much through simulation.  We build the models, run real world tests to ascertain the validity and veracity of the models, then update the models with what we have learned from the tests.  Eventually, the models and the testing results converge to an acceptable point where we can confidently generate the final iteration of the product.