Thoughts on data, football, and everything in between.
How do we deal with low-sample performances? How much should we expect a player's performance to change as he gets more minutes?