Do you feel this is similar to the idea that a problem can be approached either analytically or as a simulation?

Sometimes the math is too hard, so we create a world with some initial values and let things play.

The insights may then help us build the models or equations. But we couldn't get there just by thinking hard long enough.