Before the 2017 season, I wrote up an article about why we should’ve expected regression from the Cowboys. Though they were excellent in 2016, the Pythagorean Wins Expectations Theorem suggested that they outperformed their true team value. As I’ve discussed before, Pythagorean Wins is a pretty solid indicator of true team value, and a good predictor of regression to the mean.
“It is has been famously adjusted to estimate win percentages in the NFL, and has had quite a bit of accuracy doing so. Football Outsiders claims that between 1988 and 2004, 11 of the 16 Super Bowl winners also led the league in Pythagorean wins. They also claim that this equation is a valid predictor of improvement or regression during the following season. For example, when a team wins at least one full game more than the equation would project, they tend to regress the next year and vice-versa.”
Pythagorean Wins has been adjusted and crafted by many outlets, most recently by Football Outsiders. The current accepted equation is as follows:
((PF^2.37)/(PF^2.37+PA^2.37))*16
PF stands for “Points For” and PA stands for “Points Against.” 2.37 is the exponent which each value is raised to. This exponent has been derived by Houston Rockets General Manager Daryl Morey when he adapted the equation from the NBA to the NFL.
According to this equation, the 2016 Cowboys outperformed their true value by about 2 whole wins. Because of this, reasonable regression was expected. The 2016 Philadelphia Eagles under performed by 2 games according to their expected wins, and jolted ahead to a division title in 2017, correcting their previous performance.
Of course regressing all the way to 9 wins was surprising, but the results of the 2017 Cowboys might be cause for alarm. In 2017 the Cowboys scored 354 points and allowed 332. When you plug those numbers into the formula, you end up with 8.6 expected wins.
This means the Cowboys basically performed to their true team value in 2017 as they won just 0.4 more games than expected. According to Football Outsiders, when a team (45 teams included in this particular bundle) is within 0.0 – 0.5 of their expected wins they only change by an average of .27 wins the next season.
So, according to the numbers, the Cowboys are an 8-9 win team. While we know that injuries and suspensions forced some regression a year ago, these numbers do spell out how we all feel about the current roster. Hopefully the right changes can be made to outperform what Pythagorean Wins expects from the 2018 Cowboys.
