Book review: Scorecastingby Chris Jaffe
February 21, 2011
I haven't done a book review for THT in a while, so it's about time I get back into that swing of things. The book under the microscope this week is Scorecasting—full title: Scorecasting: The Hidden Influences Behind How Sports Are Played and Games Are Won
Put together by a pair of authors, University of Chicago professor Tobias J. Moskowitz and Sports Illustrated senior writer L. Jon Wertheim, its aim is (in the words of its own advertising copy) to be "equal parts Freakonomics, Moneyball, and Mythbusters."
The book itself is an enjoyable and intriguing series of essays taking on a series of fairly typical questions, trying to use statistics and psychology to come to some results that are often unexpected, but almost always well-reasoned.
Part of what makes this a nice, easy book to read also makes it a bit tricky to review. It's a series of essays—18 in all (not including the intro and epilogue)—covering 250 pages that you can pick up and put down at your leisure. You can read them in order if you want, or in any skitter-scatter sequence you prefer. That makes it nice to read, but reviewing is a bit tricky without giving a brief mini-Cliff Notes of all its chapters.
Let's start with the basics: it's a good book. Almost every essay within was a good one. There were one or two I didn't care for—a brief chapter comparing the Steelers and Pirates might be the low point—and a few more than just seemed "meh," but most are quite enjoyable and frequently thought-provoking.
The longer essays were generally better than the shorter ones. In part, that's because the authors had more time to really sink their teeth into what they're talking about. Then again, the longer ones are generally better because the authors just had more to say about a topic.
The book was at its best when it married its statistical investigations to psychological explanations. This comes into play in a few early chapters of the book focusing on loss aversion. In one chapter, the authors note that pitchers are more aggressive in 3-2 counts that began 0-2 than they are in ones that began 3-0. PITCHf/x data show they throw more fastballs if the count began 0-2.
Why? Loss aversion. It's a demonstrated psychological principle explaining that people dislike losing a dollar more than they like making it. A pitcher who was ahead 0-2 had that plate appearance lodged as an out in his mind, and was more upset about the possibility of giving up a walk than in a 3-2 count that began 3-0. In one case, he's losing what he thought he already had. In the other case, he's already playing with house money, so it's not terribly upsetting if he gives up ball four.
The concept of loss aversion comes up elsewhere in the book, explaining sports phenomenon as varied as why NFL coaches don't go for it very often on fourth down (despite statistical studies saying they should) to why Tiger Woods lost the 2009 PGA Champsionship to Y. E. Yang.
These sections are so illuminating not only because the authors explain WHAT happened but can take it a step further and explain WHY it did. Most sports statistical studies stop at the what (which is fine, but it's nice to read something going an extra leg).
Many chapters, however, just stick to the math to explain what the authors are talking about. For example, in the chapter on blocked shots in basketball, Moskowitz and Wertheim convincingly argue that Tim Duncan's 149 blocked shots in 2008 featured more value than Dwight Howard's 232 in that same season. They used play-by-play data to note various types of blocked shots and assigned values to those different plays. It's a nice piece, but it's pretty much just a numbers piece without much psychology in it.
The Scorecasters authors go through the play-by-play data and note that the value of a blocked shot on a jump shot is different from that of a layup, and both are different from goaltending. They go through the data to determine the value of each kind of blocked shot.
Sometimes, even the book's fairly generic chapters are better than I would've expected. For example, a chapter on NFL overtime has some fun suggestions on how to replace the coin toss. (I especially like the notion of turning into an auction using field position as currency—how far back are you willing to start to get the first shot at the ball?) Another chapter on the myth of streakiness livens up a fairly typical discussion on how people often misinterpret the role of luck by discussing an experiment in flipping a coin 200 times in a row.
There is one topic I really should cover: a pair of chapters focusing on home-field advantage in sports. First, these chapters are the heart of the book. Not only does the question get two chapters, but that duo weighs in at over 50 pages—over a fifth of the entire book. It's also quite possibly the most memorable and interesting part of the book. Last but not least, it's worth discussing this because these portions have already created some commentary here at THT, as Boss-man Studeman noted some concerns he had with it.
Let's start with that last point first. Boss-man Studes noted the section's main point—that home field advantage stems largely from officiating, as the on-court authorities in all sports appear influenced by the wisdom of the crowds that surround them—sounds uncannily familiar.
As well it should, because the exact same point is made in a book those of us around here are a little familiar with: The Hardball Times 2011 Baseball Annual. In it, John Walsh looked at the same PITCHf/x data the Scorecasters do and determined umps contribute 0.14 runs to home teams on average. Studes concludes:
I don't mean to complain, but it does bug me when book publishers and writers make big claims to sell books, but ignore the excellent research that has been performed by hard-core baseball analysts such as John, who has a PhD in Physics and has been a leading expert in baseball analysis for a while.
Studes has a point, but in this case it doesn't bother me too much. Looking at the specific issue of measuring the impact of umps upon home-field advantage, it strikes me as a ships-passing-in-the-night situation. Our book came out in late 2010. This book came out in early 2011. There's about a two-month window between them.
Some books, like the THT Annual, have a very small gap between being sent to the publisher and hitting the general public. Others, such as my own book Evaluating Baseball Managers, can take several months. Even if there was a window, it's pretty clear the Scorecasters already had completed their study before Walsh's came out.
More importantly, the section on home-field advantage is one of the parts where the authors go beyond discussing what happened to understand why it happened. Moskowitz and Wertheim explore the psychological term conformity influence, and go into the hows and whys of what that is, and how it applies across the board to officials in all sports, despite the fact that officials try to get their calls correct. The chapters on home-field advantage might be the best part of the book, given the attention the Scorecasters take in explaining the whats and whys of it.
That said, if I can excuse the Scorecasters for not mentioning Walsh's article in their book, Studeman's criticism does point to a trend in the overall book. The Scorecasters never really discuss previous research in these topics. That's fine when they present findings that are original, but that isn't always the case.
As I noted up top, many of the questions they ask are common ones and sometimes the answers are fairly common, too. Is defense really more important than offense in sports? Turns out they both matter. Is there an "I" in team? Turns out having superstars helps. Yeah, thanks.
I don't want to take this critique too far. The authors aren't plagiarizing anyone and are doing reasonable studies on their own to answer every question, but sometimes they're just reinventing the wheel.
And that leads to an interesting concern for me: are some of the chapters I found most enlightening the result of the Scorecasters finding out something original or because I didn't know of previous work on that subject? For example, there's a nice chapter showing that the NFL draft chart that nearly all teams use these days has some systematic flaws. Is that new or is it just new to me? I dunno.
Though I have that qualm, I really did enjoy this book. It's fun to read and has a lot of interesting studies that, as previously noted, often go beyond describing what happened and into understanding why it happened. I feel very comfortable in saying the large majority of THT readers would get a kick out of Scorecasting.
History instructor by day, statnerd by night, Chris Jaffe leads one of the most exciting double lives imaginable; with the exception of every other double life possible to imagine. Despite his lack of comic-book-hero-worthiness, Chris enjoys farting around with this stuff. His new book, Evaluating Baseball's Managers is available for order. Chris welcomes responses to his articles via e-mail. Oh, and now he's on twitter.