As a simple example, say I wish to calculate the product of a reaction between methyl iodide and triethylamine, then compare it with the product of benzyl iodide and triethylamine. My goal is to compare the products' relative stabilities (in a larger system this could be useful to compare reactivity options on a common substrate). I couldn't just calculate triethylmethylammonium's enthalpy and compare it with benzyltriethylammonium, right? I would expect the system with more electrons/bonds to always have a lower enthalpy value.
Alternatively, I could calculate all the reactants as well as products, including the iodide as a non-coordinating anion, then subtract the reactants from the products. This difference I could then use to compare the relative stability of the two products, am I correct in this? My worry with this approach is whether DFT can model the non-coordinating anion accurately, as well whether its result would mean anything without solvent effects. It feels like the rabbit hole can always go deeper, and I guess I am curious at what point I obtain qualitatively useful results.