How can simulation are used as an observational method?
In principle a simulation would give the same result as an observation if you can input / account for all parameters that are involved. Keeping aside the practicality of accounting for a huge number of parameters, simulation allows us to account for the most important parameters (which we think are the most important) and after running it we can compare the result with real observation to see if those factors are the most important. If the observations and simulations agree we can run simulations by varying the parameters whichever way we like and it would amount to carrying out a real experiment with those parameters.
