This series has addressed the influence of analytics in today’s NFL. Now, I will specifically address the impact that analytics has had on the coaching aspect of today’s game.
As I mentioned, every NFL team uses analytics in some capacity. Some teams are using data in a prominent and more formal role more than others. However, extracting and using the data to assist in the decision-making process has been around since the inception of the game.
Relative to coaching, analytics have often been used to increase the overall impact and ultimately the efficiency of the job. Data has been used to assist coaches, not only with game-day situations but in the preparation as well.
Typically it has been the role of a lower-level assistant often called the “quality control coach” to extract the data in preparation for assembling the game plan for the week for the coaches. While “QC’s” are preparing the game plan for the week, analyzing the data allows them to look at things like an opponent’s tendencies on third downs by formation, an opponent’s favorite plays on third-and-short, their percentage of run/pass by formation and many other situations pertinent to situational football.
For game days, data is also used to assist the coaches with regards to an opponents tendencies. They show things like “Quarterback A” likes to throw short and to the left when throwing on first down from behind his own 50-yard line. But likes to throw it deep down the middle on third downs. Or when Derrick Henry aligns “home” and QB is under center, it’s typically up the middle. To illustrate this to coaches, I could use the data to pull up every defensive coverage against 11 personnel and even further support this notion with more data that solidified what we concluded. This ultimately gives the coaches what to expect on game day from a given formation.
Data also assist coaches on game day by alerting coaches when is a good time to call timeouts, when to accept or decline penalties or the probability of success when going for it on fourth down from a particular part of the field. So analytics has made the job of the QC more efficient by making it easier to locate specific situations while preparing the game plan for the week, without going through hundreds of old tapes to retrieve the same information.
When you see a coach on the sideline with a laminated call sheet, it not only has plays but certain scenarios and reminders pre-arranged by the director of analytics.
Every NFL team has someone responsible for the data. Some teams even go as far as to have their analytics director with a headset on game day with a direct line to the head coach. His job is to alert the coach during the game of all of these scenarios while they happen in real-time.
One of the most effective in game strategies created for coaches is the concept of “win probability.” The goal was to create a statistical model that would present the odds of a team's winning the game in every on-field situation. Win probability helps coaches on game day by suggesting what to do given a particular combination of circumstances.
For example, you’re on the goal line about to score. You are down by eight points with three minutes left in the game. What would we do to yield a winning outcome in the remaining three minutes? Should we score and go for two? Should we kick the field goal and onside kick or should we score, kick the extra point and trust our defense to give us the ball back?
There have been coaches who have been analytically driven in recent memory. Chip Kelly brought his innovative offense, sport science and mathematically based decision making to the NFL. He was an innovator with regards to correlating analytics with player tracking in the NFL. While he would have early success, he soon after found himself looking for another job.
One of the many complaints about Kelly was that he lacked emotional intelligence and the ability to build cohesion with people. This ultimately shows that you cannot simply rely on the data. The NFL is and has always been a people business. But if you want an example of someone who is a huge proponent of analytics look no further than Bill Belichick — arguably the greatest coach of this generation.
Whenever he is asked about the role that analytics has played in his decisions, he immediately gets defensive. While he doesn’t want to be known as a coach that’s analytically driven, I can assure you that all of his decisions with regards to who to draft, player tracking and game day evaluation relies heavily on data.
The guy behind the analytics is his childhood friend Ernie Adams, often described as extremely intelligent, extremely knowledgeable and very mysterious. His official title is the football research director but ask anyone associated with the Patriots what Adams’ role is and they will tell you, “I have no idea.” Adams and Belichick went to high school together at Phillips Academy.
Adams is a former NFL coach who reportedly made a lot of money on Wall Street as a municipal bond trader and later owned an investment firm. He did this until Belichick convinced him to come and work with him on the Cleveland Browns.
These days Adams can be seen watching practice while saying very little, charting trends and watching an enormous amount of film on the upcoming opponent. On game day, he has a headset directly linked to Belichick.
Kyle Van Noy described a moment he was on the sideline at a critical point in the game. On a critical down, the opponent ran the exact play Adams had said they run all week in practice. As the players came off the field, they looked at one another and said: “That’s the exact play Ernie said they’d run.” It is a copycat league, so if it’s good for Belichick, then it’s understandable that every team in the NFL has an analytics department in 2020.