Football Analytics with Python & R

Preface

Life is about stories. For many, these stories revolve around sports, and football specifically. A lot of us remember playing games in the backyard with our family and friends, while some were lucky enough to play those games under the lights and in front of fans. We love our heroes and loathe our rivals. We tell these stories to help us understand the past and try to predict the future. Football analytics, at its core, allows us to use information to tell more accurate stories about the game we love. They give us ammunition for some of the game’s most vexing questions.

For example, do the most championships make the best quarterback? Or the best passing statistics? What do the “best passing statistics” even mean? Are there things that players do that transcend the math? If so, are they simply transcending our current understanding of the game mathematically, only to be shown as brilliant once we have new and better information? Who should your favorite team draft if it is trying to win this year? What about in three years from now? Who should your grandmother draft in fantasy football? Which team should you bet on during your layover in Las Vegas during a football Sunday?

Following the success of the Moneyball approach in baseball (captured by Michael Lewis in the 2003 book published by W. W. Norton & Company), people now increasingly use math and statistics to help them properly formulate and answer questions like these. While football, an interconnected game of 22 players ...