THE CALL ON THE FIELD

In NFL games, penalties can change the course of a drive and ultimately the outcome of a game. We typically hear that the more disciplined teams are better bets because they have less penalties during the game. In this week’s Under the Hood analysis, we will use historical NFL data to determine if statistics can back the claim.

APPROACH & DATA

We are interested to find out if there is a correlation between a team’s performance and the number of penalties or penalty yards the team has. We have two datasets for this analysis. The first dataset is the penalties. The penalties dataset contains the type, yardage, and offending of every penalty since the 2000 season. The second dataset is our game dataset. The games dataset contains teams points-for and points-against along with the total penalty yards that the team committed.

    Play.Id            Team                            Description        Yards       
 Min.   :    27   OAK    : 2379   Offensive Holding          :11854   Min.   : 0.000  
 1st Qu.:172960   BAL    : 2179   False Start                :11161   1st Qu.: 5.000  
 Median :340605   PHI    : 2148   Defensive Offside          : 4617   Median : 5.000  
 Mean   :346016   STL    : 2138   Defensive Pass Interference: 3807   Mean   : 7.059  
 3rd Qu.:523748   DET    : 2099   Defensive Holding          : 3245   3rd Qu.:10.000  
 Max.   :697176   TEN    : 2098   Unnecessary Roughness      : 3018   Max.   :60.000  
                  (Other):49708   (Other)                    :25047                   
     Season          Week             Team        Points.For     Points.Agst     Penalty.Yds       Total.Pen     
 Min.   :2000   Min.   : 1.000   NE     : 285   Min.   : 0.00   Min.   : 0.00   Min.   :  0.00   Min.   : 1.000  
 1st Qu.:2004   1st Qu.: 5.000   IND    : 280   1st Qu.:14.00   1st Qu.:14.00   1st Qu.: 33.00   1st Qu.: 5.000  
 Median :2008   Median :10.000   BAL    : 278   Median :21.00   Median :21.00   Median : 50.00   Median : 7.000  
 Mean   :2008   Mean   : 9.525   GB     : 277   Mean   :21.74   Mean   :21.76   Mean   : 52.11   Mean   : 7.382  
 3rd Qu.:2012   3rd Qu.:14.000   PIT    : 277   3rd Qu.:28.00   3rd Qu.:28.00   3rd Qu.: 68.00   3rd Qu.: 9.000  
 Max.   :2015   Max.   :21.000   SEA    : 276   Max.   :62.00   Max.   :62.00   Max.   :188.00   Max.   :23.000  
                                 (Other):6827

PENALTIES PER GAMES

Before we jump into statistical calculations, let look at the distribution of a NFL team’s penalties per game.

First let’s look at histogram of penalty yardage. We can see that there is a bell-shaped curve for penalty yardage for a team. It is important to note that the distribution is slightly skewed with a mean of 52.114. The skewness coefficient is 0.719 and tells us that right tail of this distribution is longer than the left tail. This makes sense because it’s impossible to commit negative penalties.

If we look at just the number of penalties committed by a team in a game, we see a very similar distribution. The skewness coefficient for this distribution is 0.577, and like the yardage distribution, the distribution is skewed to the right.

CORRELATIONS

Let’s check the correlation between Penalties and Penalty Yards in a game. In the plot below, we can see that penalties and penalty yards are highly correlated. This makes sense, because the more penalties that a team commits, we expect a higher penalty yard total to be.

Let’s go one step further and see if there is a similar correlation between penalty yardage and a team’s performance. The plot below shows the correlations between a team’s score and the penalty yardage. Notice the contrast between the prior chart that shows a high correlation and this chart. In this chart, there isn’t any correlation between penalty yards and a team’s score. There is a very slight downward trend, however, this may be explained by the skewness of the distribution in the diagram.

Let’s look at all the variables in one chart to see if there are any correlations between penalties and scores. In the plot below, there isn’t a correlation between penalties and scores.

OFFENSIVE VS DEFENSIVE PENALTIES

Before we make a call on this, let’s compare the correlation between scores and the type of penalty (false starts, holding, etc). Maybe the big penalties have an impact on scores. First, we will look at common defensive penalties as they relate to Points Against for a team. The plot below shows that there isn’t any correlation between points scored and penalty yards.

It doesn’t look good for penalties as a predictor of scores. Before we make the call, let’s check the offensive penalties as they relate to points scores. Again, we see that there isn’t any significant correlation.

CONCLUSION

After further review, the call on the field is OVERTURNED. Penalties are not a statistical predictor of a team’s score. Simply using penalties to make your NFL bets would be a mistake. In future articles, we will dive deeper into penalties to find out if there other predictions that can be made with this statistic.