Summary of NFL Data

Team Summary

TEAM WINS TODIFF OYDS DYDS
Dallas Cowboys 12 14 407.0 351.0
Indianapolis Colts 9 14 347.1 343.2
Green Bay Packers 13 13 365.6 328.2
Arizona Cardinals 11 12 373.6 329.2
Minnesota Vikings 8 11 362.8 383.6
Tampa Bay Buccaneers 13 10 405.9 331.5
Buffalo Bills 11 8 381.9 272.8
New England Patriots 10 7 353.4 310.8
New Orleans Saints 9 7 304.5 318.2
Seattle Seahawks 7 5 323.9 379.1
Kansas City Chiefs 12 4 396.8 368.9
Houston Texans 4 3 278.1 384.4
Los Angeles Rams 12 2 372.1 344.9
Pittsburgh Steelers 9 2 315.4 361.1
Denver Broncos 7 1 330.5 326.1
Cincinnati Bengals 10 0 361.5 350.8
Miami Dolphins 9 0 307.0 337.5
Philadelphia Eagles 9 0 359.9 328.8
Los Angeles Chargers 9 -1 390.2 360.1
Cleveland Browns 8 -3 340.6 311.5
Tennessee Titans 12 -3 342.5 329.8
Detroit Lions 3 -4 322.6 379.8
San Francisco 49ers 10 -4 375.7 310.0
Washington Commanders 7 -5 323.6 359.3
Atlanta Falcons 7 -6 303.8 364.4
New York Giants 4 -8 287.3 354.8
Las Vegas Raiders 10 -9 363.8 337.2
Baltimore Ravens 8 -11 378.8 363.4
Carolina Panthers 5 -13 298.9 305.9
Chicago Bears 6 -13 307.4 316.7
New York Jets 4 -13 306.4 397.6
Jacksonville Jaguars 3 -20 305.4 353.1

Statistical Summary

STATISTIC
NFL_TEAM_TOT 32.0
NFL_WIN_AVG 8.5
NFL_TODIFF_AVG 0.0
NFL_OYDS_AVG 343.6
NFL_DYDS_AVG 343.6
  • Average wins of \(8.5\) ties to a 17 game season where 1 team’s win is another’s loss;
  • Average turnover differential of \(0\) ties to the idea that a +1 for one team is a -1 for another;
  • Average offensive and defensive yards equaling each other highlight a similar relationship as above - an offensive yard for one team is a defensive yard for another.

Correlation Matrix of 2021 NFL Data

  • WINS, TODIFF, OYDS, and DYDS show strong normality;
  • TODIFF, OYDS, and DYDS show statistically significant Pearson correlation with WINS response variable

Model 1 - Wins to Turnover Differential

\(WINS = \beta_0 + \beta_1 * TODIFF\)

  • \(\beta_1 = 0.21176\) suggests that +10 on a team’s turnover differential results in 2 extra wins
  • \(R^{2}_{ad} = 0.3968\) suggests that 40% of variability in number of wins is explained by a team’s turnover differential
## 
## Call:
## lm(formula = WINS ~ TODIFF, data = nfl_2021)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -5.1040 -1.3453  0.1371  1.5930  4.1665 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  8.46875    0.39910  21.219  < 2e-16 ***
## TODIFF       0.21176    0.04578   4.626 6.69e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 2.258 on 30 degrees of freedom
## Multiple R-squared:  0.4163, Adjusted R-squared:  0.3968 
## F-statistic:  21.4 on 1 and 30 DF,  p-value: 6.692e-05

Model 2 - Wins to Turnover Differential, Offensive & Defensive Yards

\(WINS = \beta_0 + \beta_1 * TODIFF + \beta_2 * OYDS + \beta_3 * DYDS\)

  • \(\beta_1 = 0.110826\) suggests that +10 on a team’s turnover differential results in 1 extra win;
  • \(\beta_2 = 0.045741\) suggests that +100 on a team’s offensive yards per game results in 4 to 5 extra wins;
  • \(\beta_3 = -0.025442\) suggests that +100 on a team’s defensive yards per game results in 2 to 3 extra losses;
  • \(R^{2}_{adj} = 0.7353\) suggests that 73.5% of variability in a team’s number of wins is explained by these three variables.
## 
## Call:
## lm(formula = WINS ~ TODIFF + OYDS + DYDS, data = nfl_2021)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -3.1444 -0.8479 -0.3528  0.9973  3.5624 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  1.494332   4.886279   0.306  0.76200    
## TODIFF       0.110826   0.034789   3.186  0.00353 ** 
## OYDS         0.045741   0.008639   5.295 1.24e-05 ***
## DYDS        -0.025442   0.009908  -2.568  0.01586 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 1.496 on 28 degrees of freedom
## Multiple R-squared:  0.761,  Adjusted R-squared:  0.7353 
## F-statistic: 29.71 on 3 and 28 DF,  p-value: 7.601e-09