Tài liệu Practicing Sabermetrics pptx

Practicing Sabermetrics

ALSO BY

GABRIEL B. COSTA, MICHAEL R. HUBER,

AND JOHN T. SACCOMAN

Understanding Sabermetrics: An Introduction

to the Science of Baseball Statistics (McFarland, 2008)

Practicing

Sabermetrics

Putting the Science of

Baseball Statistics to Work

GABRIEL B. COSTA, MICHAEL R. HUBER,

AND JOHN T. SACCOMAN

McFarland & Company, Inc., Publishers

Jefferson, North Carolina, and London

LIBRARY OF CONGRESS CATALOGUING-IN-PUBLICATION DATA

Costa, Gabriel B.

Practicing sabermetrics : putting the science of baseball

statistics to work / Gabriel B. Costa, Michael R. Huber, and

John T. Saccoman.

p. cm.

Includes bibliographical references and index.

ISBN 978-0-7864-4177-8

softcover : 50# alkaline paper

1. Baseball—Statistical methods. 2. Baseball—

Mathematical models. I. Huber, Michael R., 1960–

II. Saccoman, John T., 1964– III. Title.

GV877.C68 2009

796.357—dc22 2009027463

British Library cataloguing data are available

©2009 Gabriel B. Costa, Michael R. Huber, and John T. Saccoman.

All rights reserved

No part of this book may be reproduced or transmitted in any form

or by any means, electronic or mechanical, including photocopying

or recording, or by any information storage and retrieval system,

without permission in writing from the publisher.

Cover image ©2009 Shutterstock

Manufactured in the United States of America

McFarland & Company, Inc., Publishers

Box 611, Je›erson, North Carolina 28640

www.mcfarlandpub.com

To Dr. Gerard Costa, my friend and my brother—GBC

To Terry, Nick, Kirstin, and Steffi:

let’s continue to live happily ever after—MRH

To JJS, MS, AJO, RMS, and, as always, MES—JTS

This page intentionally left blank

Table of Contents

Preface 1

1. What Is Sabermetrics and What Does It Do? . . . . . . . . . . . . . . . . . 5

2. Traditional Offensive Statistics: Hitting and Base-Stealing. . . . . . 11

3. Traditional Defensive Statistics: Pitching and Fielding. . . . . . . . . 20

4. Relativity and Normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

5. Park Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

6. Runs Created . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

7. Win Shares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

8. Linear Weights Batting Runs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

9. Linear Weights Pitching Runs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

10. Linear Weights Fielding and Base-Stealing Runs . . . . . . . . . . . . . 85

11. WHIP and Similar Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94

12. Weighted Pitcher’s Rating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

13. Base-Out Percentage and Total Average . . . . . . . . . . . . . . . . . . . 110

14. OPS, POP and the SLOB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

15. Total Power Quotient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

16. Isolated Power, Power Factor and Power Average . . . . . . . . . . . . 137

17. Power-Speed Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146

18. Range Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154

19. Hoban Effectiveness Quotient . . . . . . . . . . . . . . . . . . . . . . . . . . . 163

20. Equivalence Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172

21. Predicting with the Use of Regression. . . . . . . . . . . . . . . . . . . . . 179

vii

22. Higher Mathematics Used in Sabermetrics . . . . . . . . . . . . . . . . . 189

23. Potpourri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197

Appendix A: Abbreviations and Formulas 203

Appendix B: League Traditional Statistics and ABF Values 208

Appendix C: Technological Notes 214

Appendix D: Sabermetrics in the Classroom 223

Sources 225

About the Authors 227

Index 229

viii Table of Contents

Preface

Hello, my name is Gabe Costa. My co-authors, Michael R. Huber and

John T. Saccoman, and I are grateful that you are looking at this book. Prac￾ticing Sabermetrics is a follow-up to Understanding Sabermetrics, which was

published in early 2008 by McFarland.

Mike, John and I are professors of mathematics and life-long fans of the

national pastime. We have been fortunate to combine our interests in a sin￾gular way: by teaching courses on sabermetrics for over twenty years. The

term “sabermetrics” was coined by the noted baseball author and researcher

Bill James, who defined it as the search for objective knowledge about base￾ball (the “saber” part comes from the organization known as the Society for

American Baseball Research—SABR—which was established in 1971).

Our purpose for writing Practicing Sabermetrics is to give you an oppor￾tunity to familiarize yourself with the actual instruments or metrics used in

sabermetrics. Our goal has been to make our book as broad as possible and,

therefore, to reach as many people as possible. We assume the reader has

knowledge of the rules of baseball, is familiar with the fundamentals of alge￾bra and knows a tad about statistics. Chapter 22 is the only place where

advanced mathematics is introduced.

With very few exceptions, we have divided our chapters into three main

parts: an introduction of the specific concept or concepts; a number of care￾fully demonstrated problems involving the presented topics; a section where

you can actually practice sabermetrics, with the numerical answers provided.

By and large, the chapters are independent. That is, they can be read

out of order, so that the professor/teacher/student can “mix and match”—or

omit—topics as desired. We have also included a few chapters on advanced

sabermetrical themes and have added several pertinent appendices.

We trust our approach will be well received by the serious baseball fan

and by students taking courses on both the university and high school lev￾1

els. Sabermetrics has made serious inroads into academia during the past few

decades. The first course ever taught on sabermetrics was at Seton Hall Uni￾versity in 1988. Since then, the United States Military Academy, Bowling

Green University, and Quinnipiac University, among other institutions, have

offered related courses. We are also aware that the Massachusetts Institute of

Technology offers a program to middle school students dealing with the sci￾ence and mathematics of baseball. It is also our hope that parents and

guardians with children, who love baseball but dislike mathematics, would

see in our book a vehicle to encourage these children to learn mathematics.

Before I sign off and you hear from Mike and John, I would like to

acknowledge the following people to thank them for their support and assis￾tance with respect to this project: Colonel Michael Phillips and my colleagues,

the members of the Department of Mathematical Sciences at the United States

Military Academy at West Point; the Seton Hall University Priest Commu￾nity, ministered to by Monsignor James M. Cafone and the administrative

leaders of the same institution along with Dr. Joan Guetti and my colleagues

of the Department of Mathematics and Computer Science; baseball researcher

and historian Bill Jenkinson; Tony Morante of the New York Yankees; and

Linda Ruth Tosetti, the granddaughter of George Herman Ruth. Lastly, a note

of gratitude must be given to my archbishop, the Most Reverend John J.

Myers, J.C.D., D.D. In every sense, his blessing is most appreciated.

* * *

MICHAEL R. HUBER: One of our goals with this work has been to expand

the knowledge about the great former players of the national pastime. We

have tried to include many of the members of the Baseball Hall of Fame in

our examples and problems. The game has been a part of American culture

for over a century and a half, and many of the men who put their mark on

Major League Baseball did so long before we, the authors, were born. By

including mention of them, we hope to preserve their legacy. Many of the

measures we use were created to compare the best of the best. Those players

are enshrined in Cooperstown, and we felt it appropriate to create problems

broadcasting their success. We hope you enjoy the tidbits.

I must thank my co-authors Gabe and John, whose energy and passion

for both mathematics and studying baseball is contagious. They have indeed

made this a fun project for me. Extraordinary thanks go to Brandon Stern￾Charles and Joseph Dyer, two students of mine at Muhlenberg College. Bran￾don and Joe worked as summer research assistants, helping me collect data

and creating and verifying solutions to problems, mostly in the linear weights

chapters. They each hit a home run in their efforts.

2 Preface

I also want to thank my family for their support. My father, Erwin

Huber, taught me to appreciate the game of baseball when my brothers and

I were old enough to wear a glove or throw a ball. He taught me how to read

the box scores. He did what many fathers do—took us to practice, coached

our teams, helped the Little League organization as an umpire or by selling

booster tickets. I tried to pay him back by doing that for my children. Thanks,

Dad. I am grateful to my wife, Terry, and our children: Nick, Kirstin, and

Steffi. They never said no when asked to go to a baseball museum or attend

a game, whether driving a few hours to see an Army game or going to a minor

league or major league contest while on vacation, and they know that no one

leaves until the last out is made. Finally, I want to thank Father Gabriel Costa

again, for baptizing our granddaughter Riley and formally introducing her to

baseball with a New York Yankees bib after the ceremony. Grazie!

* * *

JOHN T. SACCOMAN: The baseball and mathematics have been lifelong

labors of love, and I am grateful for the opportunities that I have been given

to combine them. In particular, I am grateful to Seton Hall University, par￾ticularly the Department of Mathematics and Computer Science, for insti￾tuting and supporting the course in sabermetrics, and to my coauthors for

including me in their various sabermetrics endeavors, as a guest speaker, coau￾thor, panelist, and team teacher. In addition, I am grateful to my wife, Mary

Erin, for putting up with me through it all.

There is a wonderful tradition at the annual SABR meeting. A small

group will gather early in the morning and find a park in the city in which

to play catch. Playing catch is a pure expression of baseball companionship,

and one of life’s great pleasures for the baseball fan. I dedicate my efforts here

to the four people in my life with whom I most enjoyed playing catch: my

father, Dr. John J. Saccoman; my grandfather, Mario Saccoman; my cousin,

Anthony Ortega; and my son, Ryan Mario Saccoman.

Preface 3

This page intentionally left blank

CHAPTER 1

What Is Sabermetrics and

What Does It Do?

Introduction

Baseball is America’s game. In spite of the 1919 World Series scandal

involving the White Sox (known, thus, as the Black Sox) and the Reds, given

the many disgraceful decades when African Americans were barred from play￾ing in the Major Leagues, right up till the present controversies involving

steroids, the national pastime has endured. President Franklin D. Roosevelt

insisted that the game be played during World War II, so important was base￾ball to the American spirit.

Throughout the years we have all read novels about the game, watched

movies about the game and everyone still remember the words and music to

the song “Take Me Out to the Ball Game.” Baseball is ingrained in us.

For the past thirty years or so, however, the game of Baseball has been

looked at in new and different ways. When the yearly Bill James Baseball

Abstract (see our Sources at the back of the book) began to appear, people

started to look at the game differently. As was mentioned in the Preface, it

was James himself who coined the word “sabermetrics,” defining it as “the

search for objective knowledge about baseball.” By its very nature, this search

uses metrics; that is, instruments or tools to measure performances.

In their book The Hidden Game of Baseball, John Thorn and Pete Palmer

point out that while one may love baseball without numbers, the game itself

cannot be understood unless we bring numbers into the conversation. Over the

past three decades or so, a bevy of talented authors have published in this broad

area which we call sabermetrics. Writers such as Jim Albert, Jay Bennett, Bill

Jenkinson, Michael Lewis and G. Scott Thomas, to name but a few, have

looked at the national pastime in a myriad of ways. But they all use numbers.

5

In this book, you will be exposed to many measures. You will review

some of the traditional “old-school” statistics, such as Batting Average (BA)

and Earned Run Average (ERA), in addition to seeing newer metrics such as

Runs Created (RC), Linear Weights (LWTS) and the Power Speed Number

(PSN). By using these measures, it is hoped that a clearer picture emerges with

respect to whatever particular question is under discussion.

Let us consider an example. Suppose we want to compare pitchers from

two different eras, say the Washington Senators Hall of Famer Walter Johnson

and New York Yankees lefthander Whitey Ford. To do a “sabermetrical analy￾sis,” we would and could employ certain instruments. But exactly what meas￾ures should we use? Also, can we really compare players from different eras? What

about other considerations such as differences in the game due to changes in

the rules? How about other historical and contextual aspects, such as the fact

that Johnson never played a night game on the West Coast, nor did he ever

compete against African American players? Can these factors be “measured”?

We will return to these questions. Before we do, however, we must

emphasize the following point. We the fans must be made aware of the fact

that the degree of certainty in sabermetrics is not on the same order of as that

of pure mathematics. We do not prove theorems in sabermetrics. After all is said

and done, there is almost always a degree of subjectivity involving the inter￾pretation of our conclusions. Care must be exercised in our very choice of

measures, how they are applied and what one may derive from their use. In

a real sense, sabermetrics is as much of an art, as it is a science.

However, we can learn some things which were previously unclear or

unknown. We can gain some insights into questions like “Is a walk as good

as a hit?,” “Should we sacrifice and give up an out in order to get a runner to

second base?,” and “Was Hall of Famer Ty Cobb really a better all around

player than the icon we know as Babe Ruth?”

Let us now return to the Johnson vs. Ford question above. Can we pro￾ceed to make such a comparison and is there a formal process to follow in

answering such questions ... something like an algorithm? The answer is Yes!

The following list of ten suggestions may serve as a guide—it is not carved

in stone. You can modify this approach as you see fit.

Demonstrating Sabermetrics—10 Point Guide

1. Be careful to identify the question or questions under consideration.

What exactly is being asked or investigated? Can it be quantified or it is more

qualitative in nature? (See #8 below.)

6 Practicing Sabermetrics