Analytics Stories

CHAPTER 26Will My Favorite Sports Team Be Great Next Year?

In 2015, the NFL Arizona Cardinals had a great season. They won 13 of 16 games and lost in the NFC championship game. Fans probably had high hopes for the 2016 season. Unfortunately for Cardinals fans, in 2016, the Cardinals fell back to earth and had a mediocre 7-8-1 season. This decline from excellence is not unusual. The eight teams that won at least 12 of 16 games in 2016 or 2017 averaged only 9.5 wins the next year. On the other hand, the six teams that won 4 or fewer games in 2016 or 2017 averaged winning 5.9 games the next season. The moral is that a good team is likely to be better than average next year but not as good as last year, and a bad team is likely to be worse than average next year but not as bad as they were last year. This is a notable example of regression to the mean.

Francis Galton and Regression to the Mean

Francis Galton (1822–1911) was a brilliant Englishman who made important contributions to many fields, including statistics, meteorology, psychology, and geography. Galton is credited with creating the concept of correlation and regression. Galton also coined the phrase “nature versus nurture.” Galton collected the following data for 898 children (see the file Galton.xlsx):

• Height of the mother and father
• Height of the child

Recall from Chapter 8, “Modeling Relationships Between Two Variables,” that the correlation between two data sets is always between –1 and +1. Galton observed that ...