When I examined the Cowboys’ 2013 schedule, the first thing I did was check the first four and final four opponents. It’s vital for teams to start hot and gain confidence to propel them through the season and, as we’ve seen over the past half-decade, it’s obviously important to finish strong as well. When judging the strength of individual opponents, most analysts consider their record in the prior season as an indicator of their talent. In reality, there are stats much more effective than record at determining team strength. One of them is known as “Pythagorean Wins”—a simple theorem that determines team strength as a function of points for and points against. Pythagorean Wins provide an idea of how many games a team “should have” won based on how they scored and allowed points. So why is Pythagorean Wins superior to actual wins when analyzing the strength of the Cowboys’ opponents? Because it’s predictive. Teams who have more Pythagorean Wins than actual wins tend to perform better in the subsequent season, while those who outperform their “Pythagorean Expectation” usually regress. Actually, Pythagorean Expectation has over a 91 percent correlation with actual team wins, meaning it’s an extremely accurate way to predict future team strength. Let’s take a simple example: the most likely outcome for a team that scores 350 points and allows the same is obviously a .500 record, or 8-8. If that team actually finished 6-10, they likely got pretty unlucky, meaning they’ll probably improve in the next season solely because they’ll just be luckier. On the other hand, a team that scores the exact same number of points as their opponents but finishes 10-6 will be a strong candidate to regress in the next season. For the record, the Cowboys scored 376 points in 2012 but allowed 400, giving them a Pythagorean Expectation of 7.41 wins. Based on how they played, the most likely 2012 record for the Cowboys was 7-9, so they got lucky to finish 8-8.
Breaking Down the ‘Boys: A different way to look at the Cowboys’ 2013 schedule
Dallas Morning News | Apr 19