Rewriting the Playbook
with Data
Rewriting the playbook with data
PFF analyzes football in more detail than anyone else on the planet. And despite a team of over 500, there are limitations on what can be accomplished by just humans alone. By leveraging millions of data points from every game over the last 14 NFL seasons, and using ML models built on Amazon SageMaker on AWS, PFF is leading the way in how players are analyzed and evaluated.
It wasn’t always like this. PFF was founded in 2006 by Neil Hornsby, but his story starts in the early 1980s when Hornsby was inspired by a new TV show, broadcasting weekly NFL highlights in the UK.
Hornsby started reading everything he could on American football, and when he learned about a book by Paul Zimmerman, A Thinking Man’s Guide To Pro Football, he sent for it immediately. “That book sold me on football,” Hornsby said in a 2016 editorial he wrote for Sports Illustrated. “I’ve read it more times than I’ve read any other book in my life.” The book became a catalyst for Hornsby, cementing his love for the game and opening a universe of complexity and appreciation for all that the game could offer.
In 2003 Hornsby was inspired to follow Zimmerman’s footsteps in grading games and in 2006, took it a step further, while maintaining the spirit of Zimmerman’s high bar in accuracy and equity over conventional wisdom, by taking it online and crowdsourcing the analysis and grading of every player in every game, founding PFF in Luton, England.
By 2014 PFF had over 30 employees located across the globe and had managed to attract over 30% of the NFL teams as customers. This began to draw attention including that of a certain high-profile broadcaster, one Cris Collinsworth.