Adjusted Stats 2009 Year in Review Part V

This is the fifth in a series of posts that will review the 2009 season (as well as 2008) from an adjusted stats perspective.

Tentative Publication Schedule

Part I: Which Stats Correlate Best to Winning? - 6/24/10
Part II: Drilling Down and Regression Models - 6/28/10
Part III: Passing Efficiency Formula - 6/30/10
Part IV: Conference Strengths and Pace - 7/2/10
Part V: Testing Conventional Wisdom - 7/16/10
Part VI: Team Matchups - 7/20/10
Part VII: Year-to-Year Changes - 7/22/10
Part VIII: Points per Yard Efficiencies - 7/26/10
Part IX: Data Dump - Team Rankings - 7/28/10
Part X: Data Dump - Team Reports - 7/30/10

Testing Conventional Wisdom

There are a few pieces of conventional wisdom regarding football success with which we’re all familiar. Depending on the conversation, you may hear that you have to be able to run the ball in order to win, that you have to be balanced in your play selection in order to win, or that teams that struggle to run the ball will have more trouble scoring in the red zone. Let’s take a look at these ideas in order.

The first has already been addressed to some extent in the previous posts. We’ve seen that total rushing per carry, while correlated positively to winning, is far less important than total passing per attempt. Immediately, then, we can conclude that being able to pass the ball well is more important than being able to run the ball well. The standard line has already been proven meaningless. Clearly being able to run the ball is more helpful than not. But if it’s less important than throwing it well, what’s the point of the conventional wisdom?

Running a regression analysis only further reinforces the point. Based on such an analysis, we can determine that in order to improve their expected winning percentage by the same amount that a 1 yard improvement in total passing per attempt would yield, a team would have to increase their total rushing per attempt by 2.2 yards if their passing remained constant. And while requiring 220% as much improvement seems daunting on its own, we also have to consider that the standard deviation in total passing per attempt is over 60% larger than it is in total rushing per carry. An improvement of one standard deviation in total passing per attempt is the equivalent of a 3.6-standard deviation improvement in total rushing per carry. These numbers were for 2009; the 2008 numbers show a similar but less extreme pattern. For 2008 a yard passing was worth 1.6 yards rushing while a standard deviation passing was worth 2.7 rushing. In either season it’s clear that being able to pass the ball was far more conducive to winning than being able to run it. And, perhaps more importantly, improving your running game would not increase your winning percentage as much as making the same relative improvement to your passing game. The regressions are given below (using adjusted stats):

For the second piece of conventional wisdom related to balanced play selection I assigned each team a balance score for their season totals. Teams that threw more often than they ran received negative scores while teams that ran more than they threw received positive scores with zero meaning perfect balance. Keep in mind that for correlation calculation purposes it is a team’s standard deviation ranking that matters. The balance score was equal to (Total Rushes – Total Passes)/(Total Rushes + Total Passes). Recall that using total rushes and passes merely moves sacks to pass totals so play selection is more accurately reflected. The correlation for this balance score to winning percentage for 2009 was 0.253, a positive yet possibly insignificant correlation. Furthermore, the positive correlation can be at least partially explained by the now familiar refrain of running when leading the game. 2008 had a 0.222 coefficient.

However, these numbers don’t reward balance; they actually show the correlation of running more than passing. We can reward true balance by instead dividing the lesser of the two playcall types by the larger. Perfect balance would therefore be equal to 1 while increasing imbalance will trend to 0. Doing this yields balance-to-winning coefficients of -0.086 for 2009 and -0.020 for 2008. Slightly negative correlations but certainly insignificant for each season tells us that balance as a playcalling goal is unnecessary at best and counterproductive at worst.

The final piece of conventional wisdom I wanted to test was the frequently repeated idea that you have to be able to run the ball to score in the red zone. As seen above, we would expect that being able to run the ball better would improve red zone performance. The real question, though, is whether running the ball is more important than throwing the ball in this situation as the conventional wisdom implies. Unfortunately data regarding touchdown conversions in the red zone was not available, so correlations to red zone scoring percentage were used instead. Below are the correlations of both running and passing efficiency with red zone scoring percentage:

Once again, consistent with basically every analysis we’ve run so far, it appears that passing proficiency is more important to success than running proficiency. Why then does the perception continue that lack of running ability is what fails teams? In my opinion it is simply another case of confirmation bias. When a team that either passes quite frequently or is not adept at running the ball fails to score in the red zone, the announcers and fans will simply assume that running plays or better running ability as a team would have solved the situation. But when a running team runs the ball and fails to score, those two groups focus on the reasons why the runs failed. A missed block, a fortunate defensive playcall, or any of several possible factors are to blame. When a pass attempt goes wrong then they should have run the ball. When a rush attempt goes wrong then they should have run the ball better.

Thoughts and comments are always welcome. Next up is a study on team matchups where I'll look at W/L records fro 2008 and 2009 based on which team has the rating edge in offense, defense, passing, rushing, and all combinations thereof.