Half of all floating-point values are between -1 and 1. Half of the remaining values have magnitude between 1 and 2. And so on... If we represent positions in k

Half of all floating-point values are between -1 and 1. Half of the remaining values have magnitude between 1 and 2. And so on... If we represent positions in kilometers, and we're talking about a solar system with radius on the order of a billion kilometers in diameter, and it becomes clear that we're wasting almost all of those floating point bits!