Part of the scientific enterprise is to measure the material world and to explain its dynamics by means of models. However, not only is measurability of the world limited, analyzability of models is so, too. Most often, computer simulations offer a way out of this epistemic bottleneck. They instantiate the model and may help to analyze it. In relation to the material world a simulation may be regarded as a kind of a ‚Äúnon-material scale model‚Äù. Like any other scale model, it does not per se give any scientific explanation but is first in itself an object of scientific enquiry, a world. Since this world is numerical, it is a priori measurable. Its role in scientific explanation will be discussed.