Everything in words can be explained visually.
For simulations there seems to be two ends. A real simulation which models everything perfectly, or a fake simulation which is just a pretty moving picture for a theory. Both have their place. A real simulation can predict the future. A fake one can help teach theory. But I think there is something much more interesting in the middle.
A great scientific term is hypothesis generating research. Its science but in reverse. You look at uncontrolled data first then come up with questions about what you think is related. Later when you have a hypothesis you want to test, then you use normal science to control all the variables and run an experiment. The science part of the process gets all the credit as it is where we learn everything. But the first half is just as important. How do we come up with hypotheses?
Science is formal, academic, serious. Generating hypotheses is informal, quickly discredited, and playful. It is not serious so it get's overlooked. This is why we need simulations in the middle.
A completely fake simulation has no direct relation to the real world and we can't come up with hypotheses. A completely real simulation is indistinguishable from the real world and doesn't create hypotheses because it already accounts for all of them. What we need is half real, half fake. Real enough that it must exist in reality, and be bound by things like distance and time, but fake enough that we can make assumptions and see the changes in reality these lead to.