Theoretical astrophysicsastronomers in physical terms. With this purpose, theoretical astrophysicists create and evaluate models to reproduce and predict the observations. In most cases, trying to figure out the implications of physical models is not easy and takes a lot of time and effort.
Theoretical astrophysicists use a wide variety of tools which include analytical models (for example, polytropes to approximate the behaviors of a star) and computational numerical simulations. Each has some advantages. Analytical models of a process are generally better for giving you insight into the heart of what is going on. Numerical models can reveal the existence of phenomena and effects that you would not otherwise see.
Theorists in astrophysics endeavor to create theoretical models and figure out the observational consequences of those models. This helps allows observers to look for data that can refute a model or help in choosing between several alternate or conflicting models.
Theorists also try to generate or modify models to take into account new data. In the case of an inconsistency, the general tendency is to try to make minimal modifications to the model to fit the data. In some cases, a large amount of inconsistent data over time may lead to total abandonment of a model.
Within the astronomical community, theorists are widely caricatured as being mechanically inept and unlucky for observational efforts. Having a theorist at an observatory is considered likely to jinx an observation run and cause machines to break inexplicably or to have the sky cloud over.
Some widely-accepted theories/models in astrophysics include the Big Bang, Cosmic inflation, dark matter, and fundamental theories of physics. An astrophysical theory which has some supporters but is widely seem to be at variance with observations is Plasma cosmology. An example of an astrophysical theory which not widely accepted but is considered viable enough to merit further work is Modified Newtonian Dynamics.