Executive Summary

It is of little surprise that the New England Patriots are an outlier in overall win percentage from 2011-2020 – 74%! I became interested in this project when I learned more about the Bengals, who played in the Super Bowl this year. This past season, they had a win percentage of 59%, but in 2019-2020, their win percentage was 12.5%! After the poor season, the Bengals drafted Joe Burrow, widely regarded as one of the best quarterbacks in the league.

My project was to analyze the NFL draft, determining whether it affected team win percentage rankings. A Chi-squared analysis revealed that from the years 2011 to 2020, the teams were not equal in terms of win percentages. That likelihood is less than 0.01 percent! On average, among all 32 groups, there was no change in ordinal ranking for win percentages. However, there were significant changes in the bottom 16 teams.

Here is what I did. For every year except 2011, I identified the bottom 16 teams by win percentages. For each one, I calculated how much their win percentage rank changed from the previous year – this was an integer. I then averaged the rank changes for each season. Let us look at 2020 as an example. Out of the bottom 16 teams, the Washington Redskins improved the most from the previous season, moving from rank 30 to rank 17. On the opposite end of the spectrum, the 49ers and the Texans had the most substantial drop in win percentages. They moved from ranks two and eight to ranks 21 and 27. Overall, from 2012 to 2020, the bottom 16 teams improved their win percentages by an average of five rankings. As these teams receive the best players from the annual draft, it is unsurprising that they improve the most! In fact, on average, the bottom eight teams improve every year in terms of win percentage. (Note that the bottom eight teams can differ from year to year. Also, the average standard deviation of rank improvement is remarkably high at 9.50).

General managers are the people in charge of the draft picks for each team. I looked up the current GMs for each team and ranked how long they are currently tenured. There was no correlation with raw win percentage or ordinal rank of win percentage (R = 0.15, -0.20). The correlation makes sense; while head coaches are fired routinely after poor seasons, the general managers are less affected. Overall, I conclude that while the draft is effective at helping the worse teams, there is still luck involved. Sometimes top picks are flops, but players can also play better than anyone expected.

I spent about two to three hours each week on this project. I am satisfied with the work that I did. The bulk of the project was figuring out formulas in excel to sort and analyze the data. I had to refamiliarize myself with the IF and RANK operations and brush up on Chi-Squared. I also spent many hours double-checking my numbers, as most of them I entered by hand.

If I could have obtained the data more easily, I could have done more. While this was not the expectation, and I did not do this, I also would have entered the data faster if I had someone else read it to me while I typed it in. I was not familiar with R and did not believe I had sufficient time to learn practical applications for my project. An improved project might use more advanced software than Microsoft Excel. Finally, although I enjoy watching the super bowl, I am not what you would consider a fan of the activity. Thus, I do not know every minute detail about the draft processes or the hiring and firing of coaches and general managers. While I did an excellent job, the project might have been better with someone who knew more about the NFL.

All of the original data is from https://www.pro-football-reference.com/years/ . All other calculations (averages, standard deviations, rankings, etc.) were performed using Microsoft Excel 2016. If you want to contact me, I would love to chat! You can contact me at heisqu02@gettysburg.edu .