How the model reads him
Two ratings produce a baseline, then the matchup adjusts it: who lands more, who absorbs more, who controls where the fight happens. That's the whole shape of it for Stephen and for everyone else — no manual adjustments, no fighter the model likes.
What we'd flag
Our tape runs from 2015 across UFC and Contender Series bouts, so a long career elsewhere shows only the part we can measure. And variance in this sport is enormous — one punch settles a fight the better fighter was winning, which is why nothing here should be read off a single result. For Stephen specifically, Stephen has only 1 fight on our tape, so every rate here is built on very little.