Exactly. From a geocentric perspective, it's not wrong as long as you're explaining past events (which, as you noted, can be done by adding epicycles). The rea

Exactly. From a geocentric perspective, it's not wrong as long as you're explaining past events (which, as you noted, can be done by adding epicycles). The reason the model is not so useful (to paraphrase my college physics professor) is that you can't use it to predict things in advance with reasonable certainty (i.e. you don't know what epicycles you need until after you've observed it).