In the years 1904, 1914, 1924, ... , 2014, the percentage of times the starting pitcher pitched a complete game were7: 87.6, 55.0, 48.7, 43.4, 45.2, 34.0, 24.5, 28.0, 15.0, 8.0, 3.1, 2.4.

a. The linear probability model has least squares fit Pˆ(Y = 1) = 0.6930 − 0.0662x, where x = number of decades since 1904. Interpret −0.0662.

b. Substituting x = 12 in the linear prediction equation, predict the proportion of complete games for 2024. The ML fit of the logistic regression model yields Pˆ(Y = 1) = 0.034 at x = 12. Which prediction is more plausible? Why?