Analytics and data have changed the way many people think about play calling in the NFL. While problems such as going for two point conversions or going for it on 4th down have made headlines, analytics could influence much more, such as pass and run rates, pre-snap motion, or even something as small as which gaps to attack when running. While a one size fits all approach to play calling will not lead to ideal results due to differing personnel, there are conclusions we can make from league wide data from the 2019 season. We are going to use EPA (expected points added) to measure the success of certain playcalls.
Simple Pass or Run Analysis
The analytics revolution in the NFL has gone hand in hand with the passing revolution in recent years. Team pass attempts, yards, and touchdowns have all gradually increased throughout the history of the NFL, and we are at a point where 600 pass attempts isn’t anything to squint at. It’s simple to see. While there are still arguments for running that we will get to, passing is simply more effective in most circumstances, even considering the higher chance of a sack or turnover (go Jameis!).
Aside from the small 4th down sample size, passing EPA is higher on every down, and it’s not particularly close. The league leaders in first half first down pass rate (to avoid situational reasons), was Kansas City at 66%, followed by the Falcons, Saints, and Giants at 55%. The league low was the Redskins at 37%. Not taking into account things like score, and quarter, there are few situations where rushing is actually more effective. The following graphs show EPA by yards to go in goal line and 3rd or 4th down situations.
We can see from both of these graphs that passing is more effective, and I believe most if not all teams should be rivaling the Chiefs in first down pass rate in neutral game scripts. There are, however, a lot of reasons to run for the sake of balance, and given the situation. It’s hard for even the most advanced analytics to determine how much running for the sake of balance and deception actually contributes to success. It’s just generally and rightfully assumed that some sort of balance is needed to prevent defenses from only focusing on the pass. It’s also assumed running helps set up play action passes, but studies such as this one from Football Outsiders show that it may be a myth. While a significant drop in running may show different results, the incremental differences in how often teams rush and how successful they are doing it doesn’t have a significant impact on that team’s play action success. I tend to believe deception and play-action success can be maintained with a 60+ percent first down pass rate. It’s also advantageous to run when ahead towards the end of the game, as teams often do. This is because it decreases the chance of a turnover and run plays take more time off the clock. However, this tends to make people think high rushing rates are associated with better teams, which is often true because those teams are generally leading. Seahawks Coach Pete Carroll misuses the concept as a reason not to throw (except for at the 2 yard line in the Super Bowl); however, for teams like the Ravens and 49ers, who trust their defense and can run effectively for multiple reasons, a run-oriented game plan can be successful; it just takes more working factors.
Running up the Middle
Play Calling isn’t as simple as passing or rushing because coaches also determine the major read for a pass or the gap to attack for a run. The following graph shows EPA for runs to the outside (end), more inside (tackle), and up the middle (guard).
Despite the clear boost in running outside, it was the least called of the three runs, with the highest percentage of runs behind the guard. The only team that had real success running up the middle was the Ravens, but that’s just what happens when you have a quarterback with more rushing yards than Saquon Barkley.
Passing Through the Middle
While where a pass is thrown is largely determined by who’s open, playcalls often play a role in influencing where the ball is thrown, especially with quarterbacks who don’t go through reads quickly (@Dak Prescott).
Passing left or right doesn't seem to make a difference, as expected, but passing through the middle tends to have a higher EPA. This could just be because there’s more space to create yards after the catch without being pinned to the sideline; however, it could still influence how teams scheme their first option.
Shotgun
Teams generally pass out of the shotgun and run from under center. That’s just been how football is played, but looking at the results, it makes little sense. Last season, there were approximately four times the passes from shotgun than under center, while twice as many rushes from under center. However, an analysis of success shows teams should be doing the reverse.
Teams are actually more successful passing from under center and rushing from the shotgun. This could be due to deception, as defenses are expecting a pass out of the shotgun, but even so, that deception could be an edge that teams exploit.