December 18, 2011

Is Running The Football Even Worth The Effort?

     The average NFL pass gains 6.7 yards.  The average NFL run gains 4.3 yards.  So why do teams still run the ball over 40% of the time?

     The short answer:  Passes are very inconsistent.  If football was a game in which you had 10 plays to gain as many yards as possible, rushing plays wouldn't make any sense at all.  The 4-down structure (which in most cases is really a 3-down structure) makes passing very risky.  While it can get you quick chunks of yards, it's also likely to get you a quick 3-and-out.  The reason running plays are significant despite such a low average gain is that they usually yield some yards, which sets you up for the next down better than an incomplete pass.

     However, in accordance with my earlier rant on NFL coaches running outdated offenses, I wondered what would happen if a team decided to pass on every down.  To accomplish this, I set up an Excel spreadsheet that would simulate a drive using seven variables:

-Distance to the end zone
-Percentage of passing plays
-Completion percentage
-Yards per completion
-Interceptions per incomplete pass
-Sacks per pass play
-Yards per carry

     The first two are pretty much arbitrary.  Average values for the other five were fairly easy to find.  But there was something missing, a way to decide how far a specific pass completion or run would go.  For that I used three statistically normal games from a team that was near the middle of the league in passing and running averages.  (Thanks, Atlanta Falcons!)  I recorded the length of each run and each completion, and used that data to make a decent model of the distribution of those lengths.

     As a baseline, here's the results of 1280 drives for a league-average team with 57% passing plays, starting 70 yards from the end zone.  (I set it up to do 128 drives at a time.)

     (Field goal attempts are defined as drives which end within 33 yards of the end zone, which would be roughly a 50-yard field goal.  Points are calculated by giving 7 points for a touchdown and 2.5 for a field goal attempt, as 84% of field goals attempted in the NFL are made.)

     So the consistent nature of rushing plays is not influential enough to explain their widespread use.  What else could explain the lack of passing in the NFL?

     Another reason teams run the ball a lot is so their offense is "balanced".  That way, defenses don't know what to expect.  When they're able to predict whether the next play is a run or a pass, they're more likely to stop it.

     Let's assume that the league averages are only valid for a 50/50 offense, and changing to an all-passing offense results in, say, a 5% lower completion rate, one less yard per completion, and one more yard per run.  Changing to an all-rushing offense would have the opposite effect.  This is what the numbers would look like then:


     As you can see, too much passing would become harmful to the offense, and the optimal offense would consist of a little under 70% passes.  If each interception is said to cost the team one point (by making it easier for the other team to score), then the optimal percentage is closer to 65%.

     Now let's see what would happen for a team with a 5% better completion rate, 1 more yard per completion, and 1 less yard per run (something like this year's Giants):


     According to the data, this team should throw over 90% of the time!

     Now let's make the opposite adjustments to the average team (roughly the 2011 Vikings):


     This team should be running something close to a half-and-half offense.  What I find most interesting about these numbers is how close they are to one another.  With the average and pass specialist teams, run-heavy offenses fared much worse than normal ones.  The run specialists, though, can operate under any offense except the two extremes and be relatively successful.  This gives those teams an ability to change their offensive style depending on their opponent's defense, while the passing teams are less able to do so.

     So what have we learned from all of these numbers?  First of all, running plays are only useful because they make the pass more effective.  Secondly, NFL teams probably do not pass the ball as much as they should.  Thirdly, the best offenses are those that can pass the ball well, but the most versatile ones are those that run the ball much better than they pass it.  As coaches become more aware of the benefits of throwing the football often, the NFL is going to become even more high-scoring unless big changes are made.

No comments:

Post a Comment