Scorecasting.

I apologize for the long delay between posts; we moved into our house last week and are finally settled, although far from unpacked.

I tweeted earlier today that I’ll be joining ESPN’s Baseball Today podcast as a co-host three days a week starting in mid-March. And, if you missed it, my preseason ranking of the top 50 prospects for this year’s Rule 4 draft went up last Thursday.

—

Tobias J. Moskowitz and L. Jon Wertheim’s Scorecasting: The Hidden Influences Behind How Sports Are Played and Games Are Won aims to be the Freakonomics of sports, a marketing angle made quite clear from the cover quote from Steven Levitt that calls Scorecasting the best book of its kind since Freakonomics, which is funny, since Levitt co-wrote that book. (And one wonders if the authors share an agent or an editor or something else.) My cynicism over the quotes aside, Scorecasting is a fun read, one that does a better job of challenging conventional wisdom than providing hard answers to hard questions, the sort of book that could make an old-school sports fan rethink some of his positions without requiring a background in behavioral economics. If you’re here, however, the odds are good that your mind is already open, in which case Scorecasting is more of an enjoyable lark but might leave you looking for more serious analysis than what the authors offer in a book aimed at the mainstream audience.

Wertheim and Moskowitz attack a number of questions over the course of the book, with the only unifying theme that these are questions that can be examined (if not actually answered) through some very rudimentary statistical analysis. For example, they examine the potential causes of home-field advantage, which is fairly persistent within sports but doesn’t seem to tie to attendance; whether icing the kicker is an effective strategy (I won’t reveal their answer, but have always found the practice unsportsmanlike); or whether momentum exists. The template for each essay – some just two or three pages, others thirty or forty – is standard: Explain the question and the conventional wisdom on the subject, discuss how they operationalized the variables, then present the results in text and graphical format, usually just showing some evidence telling us whether there’s a correlation between the independent and dependent variable. For example, in the momentum chapter (“The Myth of the Hot Hand”), they look at basketball, defining what a “hot” period of time constitutes (one, two, five, and ten-minute samples), then look at point differentials over the one, two, five, and ten minute periods immediately following a “hot” period. It’s not rigorous, but it will likely sway some of your opinions even if it doesn’t convince you.

The best essays in the book combine the Freakonomics-style analysis with interesting stories, like the chapter on the history of trades in the NFL draft (“Off the Chart”), which discusses the famous Mike McCoy chart on how to value draft picks in trade talks. The authors describe the chart’s genesis, early successes, propagation, and loss of usefulness once everyone had it, along with some potential explanations for the psychology behind incorrect valuations of draft picks. (Yet another reason why I’d like to see MLB allow teams to trade draft picks: It’s another way for smart front offices to create value.) Another essay (“Rounding First”) asks why we see more round numbers in seasonal statistics than you’d expect if the results were normally distributed, pointing to psychological and perhaps financial incentives that drive behavior in situations where the leverage (to the player, not the team) is increased.

Scorecasting is a text for the mass market, which means fewer numbers and more broad brush strokes in the book. I’m not the first to raise this objection, but the way the authors treat results that are merely indicative as if they’re conclusive is offputting if you realize what they’re doing and misleading if you don’t. For one thing, their analytical methods, while valid, are on the superficial side. For another, they often confuse correlation with causation, and even though I often agreed with their arguments on the causes of the effects they discovered, they meld those opinions with statements of statistical facts in a way that just isn’t warranted. It’s a marketing issue – the book wouldn’t sell if they just presented data paired with a lot of “draw your own conclusions” quotes – but it takes what could have been a serious work and makes it a popular one.

And some of their conclusions just aren’t supported by the analysis, at least when it comes to baseball. They offer throwaway comments on how a salary cap would increase parity in baseball without an ounce of evidence to justify the statements. They claim that PEDs improve baseball performance by showing that players who had been suspended for PED usage were more likely to be promoted to the next level, a lousy proxy for multiple reasons and one that makes their conclusion, “In addition to the science, the data support the claim that steroids work,” ignorant on both sides of its comma. I imagine that the authors glossed over similar controversies in other sports, enough that no matter your game of choice you’ll find something in the book to annoy you.

You should read Scorecasting, though, in spite of its shortcomings. Moneyball was equally flawed, perhaps more so, and yet it launched a quiet revolution not just within the industry but within the fan base, an inflection point that I believe saw a major increase in the number of students of the game who began pursuing and publishing their own analyses, with some even finding themselves entering the industry as a result. I could see Scorecasting as a similar spur to innovation in the analysis of sports, and in the way sports are covered. One thing that Scorecasting does confront, without ever explicitly saying so, is ignorance. If you say “X causes Y,” others will look for a way to verify it, so don’t make the statement without trying to verify it yourself.