Predictably, the 2019 NFL regular season has produced quite a few surprises already, to say the least.  Few would have wagered that six weeks in, the vaunted LA Rams would flounder at 4-3, the 49ers would race away to a 6-0 start, or the Bills would own the second-best record in the AFC.  Some of the narratives surrounding this season pertain to teams’ abilities to either establish or stop the run — for example, the 49ers and Panthers have found success by battering and wearing down opposing defenses, while the Chiefs have stumbled due to an inability to stop the run and get offenses off the field.  Indeed, the New England Patriots won last year’s Super Bowl on the back of a physical run-first offense, with Tom Brady saving his heroics for clutch throws in fourth quarters and overtime.

However, most analytically-minded fans would immediately pooh-pooh this argument, aware that the pass is by far the more efficient offensive weapon. Plenty of teams have won Super Bowls without dominant or even average ground games, but the few champions without an explosive passing game or elite quarterback are so few and far between as to be revered for triumphing against the odds (think the 2000 or 2012 Ravens). Intrigued nevertheless, I decided to examine the numbers to determine whether the league was starting to reward run-first teams.

To cut a long story short — after running a series of regression analyses to compare the correlation between regular season wins and various basic and advanced stats during the 2018 regular season, it became clear that the analytics experts are still correct.  Three statistics in particular explained around half of the variation in regular season wins:

  1. The EPA of a team’s passing offense (53.9%)
  2. A team’s Offensive DVOA (51.0%) — The respective figures for the pass and run were about 43% and 33%
  3. The ANYA of a team’s starting quarterback (49.7%)

The equivalent advanced rushing statistics, on the other hand, had a coefficient of determination of around 33%, while the figure for defense was lower still. Thus, it follows that passing offenses are still the single most important factor in regular season success (simply think about how rarely the Patriots, Packers, or Steelers have missed the playoffs when healthy).

However, analyzing the EPAs of passing and rushing offenses and defenses during the playoffs reveals quite a different picture. Using game logs maintained by Pro Football Reference for each postseason from 2010 to 2018, one can begin to understand the “secret sauce” of the top teams in the NFL each year. I began by computing each playoff team’s postseason EPA and averaging the result to obtain a sense of how dominant pass and run offenses were that year. During the 2018 postseason, rushing offenses had an average EPA of 0.86, an improvement of over a full point compared to 2017 and good for second-best this decade.  By contrast, passing offenses nearly halved in productivity, from an extraordinary 9.64 expected points per game in 2017 to a mediocre 5.28 a year later.

Another potential method to measure the strengths of a pass or run first approach in a given year was to perform a similar analysis on the “net EPA” of each game. By subtracting the loser’s EPA stats from those of the winner, I was able to determine the distribution of these EPA figures between the teams that played in conference championship games and those that fell short. By this metric, the dichotomy is even more stark. In 2018, the winners of playoff games were only about 5 points better per game while passing the ball, good for dead last this decade. By contrast, they were more than 4 points per game better while rushing, the best figures this decade. They also won the time of possession battle by more than 10.5 minutes per game, nearly twice the margin of the next-closest year.

A similar analysis of Super Bowl winners mirrors this trend, albeit in a highly limited sample size per postseason. Despite the loss of MVP candidate Carson Wentz during the regular season, backup Nick Foles led the 2017 Eagles to the best average EPA rating of any Super Bowl winning passing offense this decade, with a below-average rushing attack and perhaps the worst defense of any of the championship winners. Curiously, Tom Brady’s 2017 Patriots mirrored this profile, with an elite passing offense hampered by an average run game and a poor defense (although it is worth noting that these elite offenses utterly brutalized each other’s defenses in the Super Bowl in combining for well over 1,000 total yards). Just a year later, the Patriots had improved markedly at both establishing the run and suffocating opposing offenses (barring the Chiefs’ 31-point second half in the AFC Championship), while the Rams were even more reliant on the run. Line graphs for both champions and runners-up are provided below.

This article does not claim that the passing game has somehow become irrelevant or less important than the running game — the 2015 Broncos and 2013 Seahawks, bolstered by all-time great defenses, have been the only champions this decade without a passing offense providing at least 11 expected points per game, compared to the 2018 Patriots’ rushing attack adding less than a field goal of value. However, it is worth pointing out that rushing the football is not quite as fruitless as it was previously. With a plethora of run-first teams like the 49ers, the Seahawks, the Bills, and the Ravens enjoying early success, the Chiefs may do well to take certain lessons from their heartbreaking AFC Championship Game defeat – especially with the Lions, Colts, and Texans all enjoying great success with adapted versions of Bill Belichick’s run-heavy, possession-oriented model. While each postseason covers a small sample of games, this analysis capably denotes the strengths and weaknesses of the top two teams (who each play at least three playoff games, nearly 20% of their regular season total).