words fail, simpulations, educated speculations, or whatever, the main thing is to preface these sorts of thing with some indication of there limitations and purely theoretical results, use them as such, and normalise the process of asking if the model has been verified with an observation