What xG really measures (and what it doesn’t), the key misconception that trips up even seasoned pros, why its loudest critics are mostly wrong, and how to pull tons of xG data with Python.