In Exercise 9.26 on page 538 we consider simple linear models to predict winning percentages for NBA teams based on either their offensive ability (PtsFor = average points scored per game) or defensive ability (PtsAgainst = average points allowed per game). With multiple regression we can include both factors in the same model. Use the data in NBAStandings to fit a two-predictor model for WinPct based on PtsFor and PtsAgainst.

(a) Write down the fitted prediction equation.
(b) The Dallas Mavericks (2011 NBA Champion) had a 0.695 winning percentage over the season, scoring 100.2 points per game, while allowing 96.0 points against per game. Find the predicted winning percentage for the Mavericks using this model and compute the residual.
(c) Comment on the effectiveness of each predictor in this model.
(d) As a single predictor, PtsAgainst is more effective than PtsFor. Do we do much better by including both predictors? Choose some measure (such as s_є, SSE, or R²) to compare the simple linear model based on PtsAgainst to this two-predictor model.

Exercise 9.26 on page 538

A common (and hotly debated) saying among sports fans is ‘‘Defense wins championships.’’ Is offensive scoring ability or defensive stinginess a better indicator of a team’s success? To investigate this question we’ll use data from the 2010–11 National Basketball Association (NBA) regular season. The data stored in NBAStandings include each team’s record (wins, losses, and winning percentage) along with the average number of points the team scored per game (PtsFor) and average number of points scored against them (PtsAgainst).