Football seems like a particularly difficult sport to try the Moneyball approach with because it's largely a battle of team cohesion rather than individual skills, and even apparently novel and meaningful heuristics one might devise like "percentage of forward passes completed in opposition half over the course of a season" are likely to be heavily influenced by the tactical approach of the player's team. Even if your telemetry data is incredibly comprehensive, it's going to be difficult to teach a computer the difference between chances created because the timing of the pass was exquisite and chances succeeded because the covering defenders were out of position or made a glaring hash of the interception; not something you'll ever find coaches or commentators struggling to decide on.