American Football Analysis
Student’s Name
Instructor
Institution
Date
American Football Analysis
American football is a unique sport in different considerations compared with various international team sports. Specifically, it is only in American football where team performance, not taking into account scores, is assessed by inches. There are several vital variables that impact a team’s performance which can be summed up as the ability to win games. These variables can be grouped into three major categories that include the general factors, the offensive factors, and the defensive factors. Under the general factors, some considerations may be towards the time of possession, the overall team effectiveness, the overtime first possession, among others. On the other hand, offensive factors may comprise yards per play, passing efficiency, first- and second-down conversions, third- and fourth-down conversions, penalties, among other factors. Lastly, the defensive play includes factors such as sacks, tackling efficiency, forced safety, red zone stops among others.
The present paper uses data from the NFL seasons 2017 to 2019 to examine the significance and the rating of four variables that include total points per game, Total 1st downs, Net Passing Yards, and Total offensive plays. A total of 25 teams is used in the present analysis. The teams are divided into two major blocks according to their ranks and an analysis of the four chosen variables is done to determine the difference between the two blocks. General descriptive statistics are carried out to first determine the mean and the standard deviation. Then t-tests and F-tests are obtained to help in comparing the means. A regression plot is then created to show the predicted rank. This study uses a statistical significance level of 0.05.
Results
| Variable/Statistic | Total Points Per Game | Total 1st Downs | Net Passing Yards | Total Offensive Plays |
| Mean (Block 1) | 44.43 | 357.25 | 4128.08 | 1044.83 |
| Mean (Block 2) | 39.2 | 316.4 | 3570.8 | 950.8 |
| t-test | 0.098 | 0.009 | 0.124 | 0.001 |
| Effect sizes | 0.726 | 1.210 | 0.678 | 1.551 |
The above table shows the mean of the two groups of data, the independent t-test p-values of the two blocks for each variable and the effect sizes for the T-tests. The mean for the upper group teams were 44.43 for the total points per game variable, 357.25 for the Total 1st Downs, 4128.08 for the Net Passing Yards, and 1044.83 for the Total Offensive Plays. In the lower ranked group, the mean were as follows: 39.2 for the Total Points Per Game, 316.4 for Total 1st Downs, 3570.8 for Net Passing Yards, and 950.8 for Total Offensive Plays. The t-test p-value for the means of the Total Points Per Game variable was 0.098, which is greater than the chosen 0.05 significance level, implying that the difference between the upper ranked group mean and the lower ranked group mean is not statistically significant.
The t-test for the Total 1st Downs means was 0.009, which is below the 0.05 significance level, meaning that the difference between the upper ranked group mean and the lower ranked group mean is statistically significant. The effect size is 1.210, which is large and thus the difference is not trivial. Third, the t-test p-value on Net Passing Yards is 0.124, which is greater than 0.05 thus the difference between the means is not statistically significant. Lastly, the t-test p-value on Total Offensive Plays is 0.001, which is below 0.05, implying that it is statistically significant. The effect size of 1.551 shows that there is a large effect.
The above plot shows a plot shows fitted trend lines to display the predictions of the rank. The y-axis shows the rank classification when comparing a team’s performance. From the above graph, there is a linear relationship between the actual rank and the predicted rank. The relationship is positive and the equation below shows the percentage variation.
The above equation is a derivation of a regression equation and particularly, shows the impact of the present rank of a team in predicting the future rank. The fitted trend lines show that within the next few prediction periods, a team’s performance is likely to remain the same.
Conclusion
A number of factors determine the performance of a team in American football. This study chooses four variables that can be used to determine how a team is ranked. These variables include Total Points per game, Total 1st downs, Net Passing Yards, and Total offensive plays. This paper finds that there is a significant difference between the upper ranked group mean and the lower ranked group mean of the Total 1st Downs variable and the Total Offensive Plays variable. The t-tests for both variables have large effect sizes implying that the difference between the means is not trivial. Further, this paper shows that there was a linear association between the actual and predicted rank.