Written By: Carlo Duffy
When two NFL teams go at each other, home-field advantage can play a huge role in a game’s outcome. Green Bay’s Lambeau Field and Seattle’s CenturyLink Field (home of the “12th Man”) come to mind as places where their teams feed off the intimidating atmospheres for their opponents. But do teams really perform better when they are at home than away? To answer this question, I used Max Horowitz’s Detailed NFL Play-by-Play Data 2009-2016, which takes data from the regular season games from the 2009-2010 season to the 2016-2017 season, and I focused on each team’s points scored and points allowed for every game played during this time. Specifically, I looked at two sub-questions:
- Which teams significantly scored more points at home compared to away games?
- Which teams significantly allowed less points at home compared to away games?
To evaluate the first sub-question, I used the following simple linear regression model:
Points Scored in a Game-Avg Points Scored in a Given Year = a+b*(Home Indicator):
where Home Indicator is coded as 1 if the team is home and 0 if the team is away; a represents the predicted difference, on average, in points scored when the team is away; and b represents how much more or less this predicted points-scored difference is, on average, when a team is home rather than away. Based on these interpretations, I am most interested in the significance of the values of b. Equally as important, I subtracted the team’s Avg Points Scored in a Given Year from its Points Scored in a Game to account for season-to-season variation in team offensive performance. Roster moves, injuries, players’ improvements/declines, etc. can drastically alter a team’s offense from one year to the next.
After using the linear regression model, I compared the p-values of b for each of the 32 equations to a significance level of 0.05. Table 1 shows that, of the 32 teams, 12 were significant. Thus, these 12 teams’ offenses significantly performed better at home than away games.
Table 1: Teams with Significantly Better Offenses at Home
Likewise, to evaluate the second sub-question, I used the following SLR model:
Points Allowed in a Game-Avg Points Allowed in a Given Year = a+b*(Home Indicator):
where Home Indicator is coded the same, and a and b are similarly interpreted to the intercept and slope in the previous SLR, except the points-scored difference is replaced with the points-allowed difference. Again, I am most interested in the b estimates, and similar to the first SLR, I subtracted Avg Points Allowed in a Given Year from Points Allowed in a Game to account for season-to-season variation in team defensive performance.
Then, just as before, I compared the p-values of b for each of the 32 equations to 0.05. Table 2 shows that, of the 32 teams, 9 were significant. Thus, these 9 teams’ defenses significantly performed better at home than away games.
Table 2: Teams with Significantly Better Defenses at Home
Putting this all together, to illustrate which teams significantly have better offenses or defenses at home (or both), I plotted each team’s b-estimate for the points-scored model (scoreest) vs. each team’s b-estimate for the points-allowed model (defenseest). But if a team’s scoreest or defenseest value has a p-value greater than 0.05, I plotted that value as zero. This is all shown in Figure 1 (below).
Based on Figure 1, what can I conclude about NFL teams’ regular season performances from 2009 to 2016?
- The teams that have significantly better offenses and defenses at home are Green Bay, Seattle, Baltimore, and Buffalo.
- The teams that have only a significantly better offense at home are New Orleans, Detroit, Pittsburgh, Minnesota, Cincinnati, Dallas, St. Louis, and Miami.
- The teams that have only a significantly better defense at home are San Francisco, Indianapolis, New York (Jets), Carolina, and San Diego.
- The other 15 teams do not have significantly better offenses or defenses at home.