Yankees and run differential are an interesting pairing. They had a short period of time where they really ran up their diff, and in that case you’re way more likely to underperform (because there isn’t much room to overperform), which makes sense because it also seems like an outlier. Then they’ve swooned since. It strikes me as a situation where their overall run diff might be overstating how good they are.
Has anyone ever compared the predictive power of raw run differential vs an adjusted run diff where you’re capped at, say, +8 per game?