A lot of times you’ll have a problem with multiple unknowns and known correlations. In elliptical galaxies, for example, you know that:
- Size is correlated with velocity dispersion and brightness.
- But velocity dispersion is related to total mass
- And baryonic mass, which is bright, follows dark mass..
So you try to figure out how stellar mass is correlated with total mass (velocity dispersion) on average in different environments so you can feed into into your sub-grid model in a large-scale simulation. It gets messy, but by making sure all your known constraints match up, your answer can be maximally likely under the circumstances.
