In Celebration of “The Hidden Game of Football”
When Kansas City and San Francisco take the field for Super Bowl LVIII, they will play in a competition that has been revolutionized by data analytics. With play sheets informed by advanced statistical analysis, today’s coaches pass more, kick less, and go for more two-point or fourth-down conversions than ever before.
In 1988, sportswriters Bob Carroll, Pete Palmer, and John Thorn proposed just this style of play in The Hidden Game of Football, but at the time baffled readers scoffed at such a heartless approach to the game. Football was the ultimate team sport and unlike baseball could not be reduced to pure probabilities. Nevertheless, the book developed a cult following among analysts who, inspired by its unorthodox methods, went on to develop the core metrics of football analytics used today: win probability, expected points, QBR, and more.
In the foreword to our new edition of The Hidden Game of Football, Football Outsider’s Aaron Schatz celebrates the significance of this book to the ongoing data analytics revolution.
It started for me in 1999, in a bookstore in New Jersey, which is strange because I don’t live in New Jersey. I was visiting a friend from college. I don’t remember why, but for some reason, we found ourselves in a local bookstore. Browsing through the sports section, I found a book called The Hidden Game of Football. I had no idea that book would change my life and eventually build my career. I grew up as a fan of baseball analytics. I read the Bill James Baseball Abstract every year, not to mention the book that preceded this one, The Hidden Game of Baseball. But by this time baseball wasn’t my favorite sport anymore— football was. Like a lot of fans in the Boston area, I had finally taken notice of the New England Patriots when Bill Parcells and Drew Bledsoe arrived in 1993 and made the team stand out in the Boston sports landscape for the first time. (Some people will tell you that Bostonians were fair-weather football fans until Bill Belichick and Tom Brady showed up, but it was Parcells and Bledsoe who really made the Patriots popular.)
However, I wasn’t looking for a book about football statistics, because I wouldn’t even have thought to look for one. All I knew were yards and touchdowns, the basic football statistics featured on national broadcasts. I didn’t know about The Hidden Game of Football, which was originally published in 1988 and then rereleased in 1998. I paged through it a bit, and it looked interesting enough. It’s not like football statistics were something that fans talked about much. But hey, I liked sabermetrics, I thought; I would probably like this.
I liked the book a lot. It changed how I watched the game of football and brought up all kinds of questions about how the NFL was covered in the early 2000s. In 2003, because I couldn’t get answers to those questions anywhere else, I launched a website called Football Outsiders. It was based primarily on the ideas I gleaned from reading The Hidden Game of Football. I still run that website twenty years later, and our main advanced metrics are still built around the ideas from Hidden Game.
The Hidden Game of Football was the first book to introduce the idea of situational football statistics to the public. The idea is so basic it seems self-evident. Of course, the definition of what makes for a successful play will change based on the down and distance—not to mention where the team is on the field. Coaches talked about football in terms of field position and “moving the sticks,” but football stats were just yardage totals and averages. There had to be a way to account for the fact that a 6-yard completion on third-and-5 was a better play than an 8-yard completion on third-and-10.
This book presented a number of ways to do that, methods that are still at the heart of football analytics today. Hidden Game introduced the idea of success rate, judging plays not just by yardage but also by down and distance. (Success rate compared to a league-average baseline eventually became Defense-adjusted Value Over Average [DVOA], the main metric at Football Outsiders.) Hidden Game featured the first public discussion of win probability analysis based on the current score, time remaining, and field position. And it was the first book to measure expected points added in the NFL, noting that each location on the field had a particular value depending on which team scored next. When Hidden Game showed that the defense, not the offense, was more likely to score next if the offense was backed up behind its own 25-yard line, it provided a mind-blowing new way to think about the game . . . but one that was also totally obvious once you read the explanation.
Critics have often written off statistical analysis in sports as just being about “fantasy numbers.” Reading Hidden Game, it was obvious that the authors’ interest in football went far beyond the yardage and touchdown totals of “skill players.” This book focused on what actually made teams win and lose ballgames. There was a chapter on offensive lines and an entire section on special teams. The authors also explored the history of football beyond the 1986 season covered in the book. They wrote about the development of the game and how football stats came to be so limited, they explained the history behind the quarterback passer rating, and they included an entire chapter on the Hall of Fame.
Going back over Hidden Game, it’s remarkable how many of the basic ideas of football analytics were already apparent to the authors. They suggested that teams go for it on fourth down instead of kicking field goals near the goal line and recommended adding sacks and sack yardage to passer rating so quarterback play would be measured more accurately. They also proposed moving the extra point back to make the game more interesting, which the NFL eventually did in 2015.
And years before I launched Football Outsiders with the same analysis, they showed why the cliché “when player X rushes for 100 yards, his team wins Y% of its games” is so wrong: teams that are losing late have to pass, while teams that are winning late run the ball to chew clock.
Not many people bought The Hidden Game of Football when it came out. It took a lot longer for advanced statistics to circulate in football than in baseball. The ideas really didn’t take hold at all until the internet. Football Outsiders started in 2003. Pro Football Focus began charting games in 2004. Brian Burke’s Advanced Football Analytics launched in 2007. And it’s only in the last few years that advanced analytics have really spread into front offices and onto NFL broadcasts.
But the conversation about football analytics has exploded in recent years. Now you can find all kinds of advanced analysis on Twitter thanks to nflfastR, a publicly available database of play-by-play that goes back to 1999. Want to know whether your team should have punted or gone for it on that fourth down? There’s a bot that updates numbers in real-time for every fourth down during every game on Sunday.
All thirty-two NFL teams now employ at least one analytics staff er, with some teams having up to eight or nine. In 2021, NFL coaches went for it on fourth downs nearly twice as often as historical averages based on situation. Over the last decade contracts have moved toward an analytical understanding of positional value, with running backs making less money and players involved in the passing game (quarterbacks, wide receivers, cornerbacks, and edge rushers) making more.
Meanwhile, the league does its own advanced stats with the Next Gen Stats project, which began tracking players using chips in their pads in 2016. This led to the Big Data Bowl competition, which introduced advanced metrics such as “rushing yardage over expected,” based on the location of all blockers and defenders at the time of the handoff. Rather than being hidden away on some rarely used website, these Next Gen Stats metrics are featured on broadcasts and in a big Amazon Web Services ad campaign. On some networks, you’ll even see win expectancy numbers on screen when there’s a big fourth-down decision to be made.
More than thirty years after it was first released, The Hidden Game of Football remains a solid introduction to the concepts of football analytics. The prose is readable and interesting, and the authors show a clear love for the game and its history. The math is laid out in a simple fashion and isn’t complicated. If more people had read the original Hidden Game, perhaps the football analytics revolution would have started much, much earlier. Instead, this book is a time capsule by three men who were thinking way ahead of everybody else about how to measure the quality of teams and players in the NFL— and how to use those improved measurements to win more games.
They say that not many people bought the first Velvet Underground record, but everyone who did started a band. Well, not many people bought The Hidden Game of Football, and none of those who did started a website since it was 1988 and they didn’t have the internet. But eventually the book found its way into my hands, and by that point it wasn’t difficult to start a website and gather a following writing about the NFL. Plenty of people who followed me got their start in football analytics because they read Football Outsiders. My eternal gratitude goes out to John Thorn, Pete Palmer, and Bob Carroll for being my Lou Reed and John Cale. You get to decide which one of you is Nico.
Aaron Schatz is editor-in-chief of Football Outsiders. He created many original statistical methods for NFL analysis such as DVOA and DYAR. The lead writer and editor on Football Outsiders Almanac, Aaron also writes for ESPN.com and ESPN+, and appears weekly during the NFL season on the Off The Charts podcast. He has written on sports, politics, and taxes for a number of publications including The New Republic, The New York Times, The Boston Globe, Slate, The American Prospect, and the Boston Phoenix.