National Football League

I. INTRODUCTION

Most people in the United Sates watch and follow the National Football League (NFL). Of them a significant amount of these people that place wagers on games to predict who will win. The major gambling sites of Los Vegas and Atlantic City which accept these wager compared to there odds to who will win the games. As a fan of the NFL consider to place a wager on the game to gain a profit on the team that wins they need to consider there odds of winning. They should only consider placing wager on games they believe they know who the potential winners will be. These fans need to have a model to help them determine the winning team.

The NFL current has a sixteen game season. As the NFL season progress, team statistics are calculated based various events in the game. Some of the important statistics reported for each team are the total offensive yard, total defensive yards, total points for, total points against, turnover margin, offensive penalty yards, defensive penalty yards, and RTG-Team Quarterback passer rating. There are many different NFL analyses that use many different statistical data accumulated through out the season to predict the wining teams each week. NFL analyses discuss team statistics on the radio talk shows, television talk shows and internet web site.

In each NFL Game the offensive tries to move the ball down the field to cross the opponent's goal line. When the offensive crosses the goal line they score a touch down. After a touch down is score the offence has an option to kick an extra point or a two-point conversion. They may also score by kicking the football through the upright as a field goal. The most important position on the offence is the quarterback. The quarterback has the option to hand the ball to the running backs to run the ball or to throw the ball to a wide-received. The teams defensive try to stop the offensive from moving the ball down the field. As the game is played, the defensive may take the ball away from the offensive either from a fumble or interception. There are may rule that each player must follow to play their position. If they do not follow the rules then their team will be penalties. Each type of penalty has different number of yard the ball is moved for it.

Data collected for the 2008 NFL season by MSN and FOX sports provides the statistic information needed to evaluate each team. This team statistic information accumulated over the season will be used to make a model to predict the wining percentage.

II. METHODOLOGY

This paper looks at each NFL teams overall wining percent compared to the important team statistic to see what relationships exist, if any, and determine what team statistic have the greatest impact on the team winning percentage. The method used for this analysis was multiple regression. The general purpose of multiple regression is to learn more about the relationship between several independent, or predictor variables and a dependant, or criterion variables.

The importance of this study is to determine which team statistic is important in determine the teams winning percentage and then we can revise the model to best predict the winning percentage. The first model examined will use the eight team statistic previously mention.

Variables and Models Defined

Y X1 X2 X3 X4 X5 X6 X7 X8

PCT Total Off. Total Def. Points For Points Against Turn-over margin Yards Off. Pen. Yards Def. Pen. RTG-QB Passer Rating

With these variables been defined, models will need to be created to begin the study.

Multiple Regression Equation

Below will be the initial forms of models will be.

Y = B0 + B1X1 + B2X2 + B3X3 + B4X4 + B5X5 + B6X6 + B7X7 + B8X8

*Above are the initial forms the models. It is subject to change once results have been found and interpreted and finalized models will be created.

III. RESULTS

Initial Model

Regression result summary:

Y = NFL Team Winning Percentage

x1= Total Offensive Yards x5= Turnover margin

x2= Total Defensive Yards x6= Yard of Offensive penalties

x3= Points For x7 =Yard of Defensive penalties

x4= Points against x8 = RTG-QB Passer Rating

N = 32 (Sample size) F-ratio 22.8*

Adj. R2=.849

Ŷ=.32 +.002x1 +.000x2 -8.517x3 -.002x4 +.000x5 -.000x6 +.000x7 +.001x8

(3.5)* (0.2) (-0.3) (-4.0)* (0.0) (-0.9) (1.9) (0.3)

* Significant at 5% level

(T-ratios are presented in the brackets.)

*F = 22.8 > F Critical Value 2.37  Model is significant

It can be concluded that this model is significant with an F value of 22.8 and an Adjusted R Squared of .849. The model also proves to have some significant variables in it. With a T critical value of +/- 2.069, Total Offensive Yards and Points against all prove to be significant. This leaves Total Defensive Yards, Points For, Turnover margin, Yard of Offensive penalties, Yard of Defensive penalties, and RTG-QB Passer Rating insignificant in this model meaning they do not have a large enough impact in the equation.

Final Model

Regression result summary:

Y = NFL Team Winning Percentage

x1= Total Offensive Yards

x2= Points Against

x3= Yard of Defensive Penalties

N = 32 (Sample size) F-ratio 68.08*

Adj. R2=.866

Ŷ=0.3805 +0.0024x1 -0.0023x2 +0.0002x3