I think there is a deeper idea implicit in Nick's post that is missing here. For something to be a good explanation, it doesn't just need to recreate the observed relationships, it needs to do so with a minimal amount of explanatory structure. For example, heliocentric models aren't better than geocentric ones because of some metaphysical preference towards the sun's reference frame, they are better because they explain the same observed phenomena with fewer free variables and computations.
The same is true for all kinds of simulations. If I designed a phys... (read more)
I think there is a deeper idea implicit in Nick's post that is missing here. For something to be a good explanation, it doesn't just need to recreate the observed relationships, it needs to do so with a minimal amount of explanatory structure. For example, heliocentric models aren't better than geocentric ones because of some metaphysical preference towards the sun's reference frame, they are better because they explain the same observed phenomena with fewer free variables and computations.
The same is true for all kinds of simulations. If I designed a phys... (read more)