## Curve Ball: Baseball, Statistics, and the Role of Chance in the GameJim Albert, Jay Bennett "... a smart and energetic collection of essays on baseball statistics. Curve Ball doesn't play misty-eyed homage to baseball's traditions and conventional wisdoms.... This is great stuff.... Curve Ball makes clear how pleasurable [stats] can be, and arguably how important, to view the great American game with real precision." -- The Wall Street Journal "Rating: 4.5 out of 5. Must own!" -- Baseballnotebook.com "In [Curve Ball] Albert & Bennett explain the game in ways the conventional press - even titans such as Bill James - cannot." -- Baseball America "[The book] illustrates how statistical reasoning can be useful in teasing out the role of chance from performance in baseball to better assess ability.... Curve Ball represents another advance in the genre of baseball and statistics books." -- Journal of the American Statistical Association There is a fascination among baseball fans and the media to collect data on every imaginable event during a baseball game and to use these data to try to understand characteristics of the game. But patterns in baseball data are difficult to detect due to the inherent chance variation that is present. This book addresses a number of questions that are of interest to many baseball fans - including how to rate players, predict the outcome of a game or the attainment of an attainment, make sense of situational data, and decide the most valuable players in the World Series. Curve Ball is directed to a general audience and does not assume that the reader has any prior background in probability or statistics, although knowledge of high school algebra will be helpful. |

### What people are saying - Write a review

#### LibraryThing Review

User Review - bluetyson - LibraryThingA look at baseball from a sports statistics and published mathematical analysis front. Interesting, but not as ground breaking as some of the amateur non university researchers came up with not too much later. A bit of an overview. Read full review

#### Curve ball: baseball, statistics, and the role of chance in the game

User Review - Not Available - Book VerdictBaseball is a fascinating game for the statistical analyst. On the surface it appears so simple and limited. But the more closely one studies the game, the more, it seems, there is to know. The ... Read full review

### Contents

Preface to the Paperback Edition | xiii |

Introduction | xv |

SIMPLE MODELS FROM TABLETOP BASEBALL GAMES | 1 |

Model Assumptions of AllStar Baseball | 8 |

Introducing the Pitcher | 9 |

The Independent Model | 15 |

The Interactive Model | 20 |

Which Model Is Best? | 24 |

THE CURVATURE OF BASEBALL | 209 |

The DLSI Simulation Model | 210 |

The Probability of Scoring Two Runs | 211 |

The Probability of Scoring No Runs | 213 |

A DLSI Example | 217 |

Lessons from the Simulation | 220 |

DLSI and Runs per Play | 226 |

Where Do We Stand? | 228 |

EXPLORING BASEBALL DATA | 27 |

A Batch of OnBase Percentages | 28 |

Simple Graphs | 29 |

Typical Values the Mean and the Median | 31 |

Measures of Spread Quartiles and the Standard Deviation | 32 |

Interesting Values | 34 |

A FiveNumber Summary | 35 |

OBPs of Offensive and Defensive Players | 37 |

Relationships Between Batting Measures | 38 |

Relating OBP and SLG | 39 |

What about Pitching Data? | 41 |

Strikeouts and Walks | 42 |

Looking at Strikeout Totals | 43 |

Defining a Strikeout Rate | 44 |

Comparing Strikeout Rates of Starters and Relievers | 47 |

Association Between Strikeouts and Walks? | 48 |

Exploring Walk Rates | 49 |

Comparing Walk Rates of Starters and Relievers | 50 |

INTRODUCING PROBABILITY | 51 |

Looking for Real Effects | 53 |

Predicting OBPs | 55 |

Probability Models | 57 |

Observed and True OBPs | 59 |

Learning about Batting Ability | 62 |

Estimating Batting Ability Using a Confidence Interval | 66 |

Comparing Hitters | 68 |

SITUATIONAL EFFECTS | 71 |

Surveying the Situation | 72 |

Looking for Real Effects | 74 |

Observed and True Batting Averages | 75 |

Batting Averages of the 1998 Regulars | 78 |

Two Models for Batting Averages | 79 |

Do All Players Have the Same Ability? | 80 |

A Model Using a Set of Random Spinners | 81 |

Situational Effects | 86 |

Turf vs Grass | 87 |

Scenario 1 No Situational Effect | 89 |

Scenario 2 Situational Bias | 90 |

Scenario 3 Situational Effect Depends on Ability | 91 |

Finding Good Models | 92 |

What Do Observed Situational Effects Look Like When There Is No Effect? | 93 |

The Last Five Years Data | 95 |

The No Effect Situations | 96 |

The Bias Situations | 98 |

The Ability Situations | 101 |

How Large Are the True Ability Effects? | 106 |

Game Situation Effects | 107 |

A Lot of Noise | 108 |

STREAKINESS OR THE HOT HAND | 111 |

Thinking about Streakiness | 112 |

Interpreting Baseball Data | 114 |

Moving Averages Looking at Short Intervals | 116 |

Runs of Good and Bad Games | 119 |

Numbers of Good and Poor Hitting Days | 120 |

Mr Consistent | 121 |

How Does Mr Consistent Perform During a Season? | 122 |

Mr Streaky | 126 |

How Does Mr Streaky Perform During a Season? | 129 |

Mr Consistent or Mr Streaky? | 131 |

Team Play | 134 |

A Consistent Team | 138 |

A Streaky Team | 140 |

Thinking about Streakiness Again | 143 |

MEASURING OFFENSIVE PERFORMANCE | 145 |

The Great Quest | 146 |

Runs Scored per Game | 148 |

Batting Average and Runs Scored per Game | 155 |

Slugging Percentage and OnBase Percentage | 158 |

Intuitive Techniques | 167 |

OnBase Plus Slugging OPS | 168 |

Batters Run Average BRA and Scoring Index DX | 170 |

Runs Created RC | 173 |

More Analytic Models | 175 |

AVERAGE RUNS PER PLAYS | 179 |

Least Squares Linear Regression LSLR | 180 |

Adding Caught Stealing to the LSLR Model | 186 |

Adding Sacrifice Flies to the LSLR Model | 189 |

The LndseyPalmer Models | 191 |

Palmer Enters the Picture | 202 |

Comparing the LSLR and LindseyPalmer Models | 204 |

Additive Models | 229 |

Product Models | 230 |

Player Evaluation in the Best Models | 232 |

Player Evaluation on an Average Team | 235 |

MAKING SENSE OF BASEBALL STRATEGY | 245 |

Lindseys Run Potential Talbe | 246 |

Old vs New Data | 247 |

A Second Important Table | 248 |

Stealing Second Base | 249 |

To Steal or Not to Steal | 251 |

A Different Criterion | 253 |

Stealing in Other Situations | 254 |

The Sacrifice Bunt | 255 |

Sacrifice Bunts in the 2001 World Series | 257 |

Should Curt Schilling Sacrifice? | 259 |

How About Craig Counsell? | 260 |

The Intentional Walk | 262 |

Compare the Costs | 263 |

MEASURING CLUTCH PLAY | 269 |

Clutch Hits | 271 |

Leading Off an Inning vs Not Leading Off | 275 |

Two Outs vs NoneOne Out | 277 |

Situation Evaluation of Run Production | 279 |

A New Criterion for Performance | 288 |

The Calculation of Win Probabilities | 295 |

Player Game Percentage PGP | 300 |

World Series Most Valuable Players | 307 |

The 2002 World Series | 311 |

Game 1 | 315 |

Game 2 | 317 |

Game 3 | 319 |

Game 4 | 320 |

Game 5 | 322 |

Game 6 | 323 |

Game 7 | 325 |

Looking to the Future | 326 |

PREDICTION | 327 |

Guessing | 328 |

Predicting 1999 Game Results | 329 |

How Good Were Our Predictions? | 331 |

Predicting the Number of McGwire and Sosa Home Runs | 333 |

A Simple Prediction Method | 334 |

A Spinner Model for HomeRun Hitting | 335 |

How Many AtBats? | 336 |

Binomial Probabilities | 337 |

What If We Dont Know Sosas True HomeRun Rate? | 338 |

Revising Our Beliefs about Sosas HomeRun Probability | 340 |

One Prediction | 341 |

Many Predictions | 344 |

Predicting Career Statistics | 347 |

Sosas HomeRun Probabilities | 348 |

How Long and How Many AtBats? | 349 |

Making the Predictions | 351 |

DID THE BEST TEAM WIN? | 353 |

The Big Question | 354 |

Describing a Teams Ability | 356 |

1871 to the Present | 357 |

Explanations for the Winning Percentages | 359 |

A Normal Curve Model | 361 |

Team Performance over Time Revisited | 363 |

A Mediocrity Model for Abilities | 365 |

A Normal Model for Abilities | 366 |

Weak Average and Strong Teams | 367 |

A Model for Playing a Season | 368 |

Simulating a Season | 369 |

Simulating an American League Season | 370 |

Simulating Many American League Seasons | 374 |

Performances and Abilities of Different Types of Teams | 376 |

Simulating an Entire Season | 380 |

Chance | 382 |

POSTGAME COMMENTS A BRIEF AFTERWORD | 385 |

TABLETOP BASEBALL GAMES | 389 |

AllStar Baseball | 390 |

StratOMatic Baseball | 392 |

Sports Illustrated Baseball | 394 |

397 | |

401 | |

405 | |

### Common terms and phrases

1993 World Series 2002 World Series ability Alomar's American League Anaheim Angels at-bats Barry Bonds baseball statistics batting averages boxplot calculation career Carter's chance Chapter column compute consistent hitters curve estimate evaluation fans Figure getting on base graph hitting data home runs home-run probability look LSLR Major League Mark McGwire McGwire National League number of games number of runs OBP values observed On-Base Percentage percent Phillies pitch plate appearances play players plot prediction probability of scoring RC/G reach base RMSE run potential run production run value runners in scoring Runs Created Runs per Game runs scored sacrifice flies scatterplot scoring position simulation single situational effects Slugging Percentage Sosa spinner stealing stemplot Strat-O-Matic strategy streaky hitter strikeout Table Team Runs team's Todd Zeile Toronto Troy Glaus true batting average weights win probabilities winning fractions World Series Zeile Zeile's