Question

3. In a National Football League (NFL) regular season, each team plays 16 games. Let " wins" be the number of regular season wins by a team in a particular season (taking a tie as half a win.) Since on average teams win half their games, the distribution of team wins has mean 8. Assume the distribution of team wins stays about the same from year to year. There is a positive correlation between a team's wins one year and their wins 0.327.) Because of this, we c wins the previous year. anm an use regression to predict a team's win one year by using their a) Find the regression line to predict a team's wins one year from their wins the previous year. (Hint: You do not need to know the standard deviations, but if you cannot work out how to do the problem without standard deviations, make a reasonable guess.) (b) In 2013, Houston had 2 wins, while in 2014 they had 9 wins. In 2013, Dallas had 8 wins, while in 2014 they had 12 wins. Use regression to predict 2014 wins for Houston and Dallas based on their 2013 wins. Which team exceeded their prediction by a larger margin? (c) A cable sports analyst who does not know statistics suggests a different prediction systenm simply predict a team will win as many games one year as they did the previous year. Explain convincingly to the analyst why in the long run, this prediction system will not be as accurate as a regression line.

Answer 1

ctoges rnhl the ret n umber poa maskou ehoin un On yea 4 lot y team O model en det ribution on lam -Y 2-& f 2.

ev the enestion , the ovalytt auome tre pye dictan Cystem Rohich implio linear gulation chip betaon n y and y, hon, ch and theve no chonte fot arro ciatid an exptained-foith.co.彳nthe tre