February 16, 2015

Book 12: The Numbers Game

The Numbers Game: Why Everything You Know About Soccer Is Wrong
Chris Anderson and David Sally

While I've found these aggressively contrarian statistical investigations into aspects of modern life to be somewhat hit and miss, I saw no obvious reason not to at least try The Numbers Game when I found it sitting on a shelf at the library. Unsurprisingly, perhaps, I found it to be a hit and miss volume itself, offering both the clever insights and egregious oversights (to say nothing of simplifications) I've come to expect in this type of book. Its basic framework is, as usual, fairly solid: Anderson and Sally are clearly keenly aware of what they are doing, leaving little room for doubt with their frequent references to the history of analytics-fueled Moneyball in various sports, but take less overt, annoying glee in their counter-intuitive findings than some of their more insufferable peers. Their book flows at a reasonable pace, and their findings likewise pop up in a sequence that not only makes sense but also allows authors and readers alike to build on insights implied by previous statistics. There is a narrative here, despite some of the more standalone aspects of this kind of work, and the authors' propensity to keep the big-picture story in mind during each of the book's many brief investigations ensures that it is accessible despite the math that drives it. Moreover, Anderson and Sally seem to strike a decent balance between getting lost in the statistics and interpreting them for readers (like myself) who haven't seen a regression analysis since high school.

It is, however, somewhat unclear whether the authors occasionally ignore their obvious knowledge of soccer in the pursuit of a particularly contentious claim. Despite their success in generating and maintaining the readers' interest, however, they are sometimes liable to fall into the familiar traps of corollary/causation confusion, and self-importance that plague their peers. Many of their calculations and resulting insights revolve around the typical binary win/loss system that soccer so uniquely avoids with the prevalence- and relative importance- of the draw. Though the authors do investigate some aspects of draws, they seem to ignore this third possible match outcome entirely at points where it appeared to me to be particularly relevant. Likewise, I was disappointed to find that the authors (and a former goalkeeper among them!) so casually dismissed the importance of high goal totals without even considering the implications of the goal differential as soccer's first-tier tiebreaker over competitions with multiple legs. Championships, relegations, and participation in continent-wide superleagues are routinely decided by those games where teams got or conceded what turned out to be a pivotal 4th or 5th goal, as irrelevant as it may have been (statistically or otherwise) within the context of the game at hand. That said, however, the authors' investigations of the relative importance and weight of goals were easily some of the most interesting, and relevant, parts of the book.

More pivotal, however, is the book's various treatments of what proves to be its central question (even if it is unacknowledged as such): is it better to play to win or to play not to lose? Anderson and Sally tackle this question from a variety of expected and unusual angles, eventually leaning toward a defensive-minded mindset (again with the goalkeeper's influence, no doubt) that they acknowledge is far from a hard-set rule. In doing so, they provide an interesting framework that spectators might consider when watching their favorite (or most hated) sides on and off the pitch, but do so without being excessively dogmatic or smug. Though the book is certainly susceptible to the usual pitfalls of its genre, the authors show a great deal of respect for their subject by weaving relevant historical facts into their narrative and providing examples from a variety of eras, countries, and leagues.

Some pitfalls, however, cannot be avoided, and one gets the sense that the authors do indulge in a fair bit of hand-waving throughout the book. As a rule, it lacks true transparency, and I'm not completely convinced that all of the more random-looking scatterplots do, in fact, indicate the trends the authors try to draw out of them. I left the book convinced that Anderson and Sally should have made more of an effort to explain the statistical methods behind their findings and conclusions, and that they possessed the skill to do so without bogging down the narrative or overwhelming their less scientifically inclined readership. The book may not be condescending, but inquisitive readers couldn't be faulted for wondering whether it might have been a bit more open to scrutiny from statisticians and laypersons alike. That said, however, the book is still a highly readable look into the potential implications of advanced analytics on a sport that may be even more hopelessly mired in tradition than the peers who have already embraced the revolution. The Numbers Game may be a product of its time and genre, but it will yield interesting insights to those readers who are willing to embrace analytics with a certain grain of salt.


Grade: B

No comments: