"...Freakonomics meets ESPN."

—Alan Schwarz, author, The Numbers Game

Taking Measure of the Many Myths in Modern Sport
David Berri, Martin Schmidt, and Stacey Brook

 

 

 

Reviews | What's Inside | Where to Order | Stanford University Press

 

 

 

Home

Chapter Excerpts

Detailed Table of Contents

Preface Excerpt

Chapter 1 Excerpt

Chapter 2 Excerpt

Chapter 3 Excerpt

Chapter 4 Excerpt

Chapter 5 Excerpt

Chapter 6 Excerpt

Chapter 7 Excerpt

Chapter 8 Excerpt

Chapter 9 Excerpt

Chapter 10 Excerpt

 

 


 

 

 

 

 

Chapter Six: Shaq and Kobe

 

from The Accuracy of Our Methods, pp. 109-110

 

How accurate is this approach? We have already said that our model explains 95% of team wins. Unfortunately, this basic statement has little meaning for those not versed in regression analysis. To put the accuracy of our methods in perspective, we measured the wins produced by each player employed by the NBA for the 2003–04 season. We then summed wins produced by each player for each team. How does our summation of player wins produced compare to actual team wins?

 

The answer is reported in Table 6.8. The average difference between the summation of player wins produced and team totals is 1.67 wins. How does it do for the Lakers? The team won 56 games, yet our model predicts only 51.7 victories. The difference of 4.3 wins is actually one of the largest we observe. The team the Lakers played in the finals, the Pistons, had a predicted win total equal to 56.7. Remember our model is based on offensive and defensive efficiency. Given these results, perhaps people should not have been surprised when the Pistons defeated the Lakers in five games. Although playoffs are not a perfect test for determining the identity of the best team, our analysis of player production indicates that contrary to popular perception, the Lakers were not necessarily better than the Pistons in the regular season and should not have been heavy favorites prior to the 2004 NBA Finals.

 

Table 6.8

Evaluating the Accuracy of Wins Produced, 2003–04 Regular Season

Teams

Summation of

Player Wins Produced

Actual

Team Wins

Difference in

Absolute Terms

San Antonio

60.4

57

3.44

Detroit

56.7

54

2.75

Indiana

56.6

61

4.41

Minnesota

55.8

58

2.21

Sacramento

54.6

55

0.38

Dallas

53.1

52

1.15

LA Lakers

51.7

56

4.31

Memphis

47.8

50

2.23

New Jersey

47.5

47

0.52

Houston

45.9

45

0.90

Milwaukee

43.8

41

2.85

Denver

43.7

43

0.73

Miami

42.3

42

0.28

New Orleans

40.7

41

0.26

Seattle

39.3

37

2.31

Golden State

39.1

37

2.10

Portland

37.9

41

3.10

Utah

37.6

42

4.41

New York

37.1

39

1.89

Boston

36.8

36

0.82

Philadelphia

34.3

33

1.27

Cleveland

33.9

35

1.06

Toronto

33.1

33

0.09

Phoenix

30.6

29

1.64

Atlanta

28.5

28

0.55

LA Clippers

28.5

28

0.53

Washington

25.7

25

0.68

Chicago

23.7

23

0.70

Orlando

22.0

21

0.96

 

 

AVERAGE ERROR

1.67

 

 

Excerpts (c) 2006 by the Board of Trustees of the Leland Stanford Jr. University.  No further use, reproduction or distribution of this material is allowed without the written permission of the publisher.

 

Chapter Seven Excerpt