Wednesday, September 7, 2011

a book I read (5)

Moneyball, by Michael Lewis

So I’ll admit, I was less ‘excited’ to read this book and more like ‘at least I can say I’ve read it when I get into a discussion about sports and analytics and the person asks me if I’ve read Moneyball’. But I actually really enjoyed it! So yay for that.

I think the strength of the book is that Lewis manages to frame the story very effectively — Billy Beane as the one guy who actually learned something from the pratfalls of conventional wisdom, and his team is the scrappy, can-do bunch of cast-offs nobody else wanted who go on to achieve far above expectations that we all want to feel like we can relate to.

The problem I see with the book is indeed embodied in the standard objection to it (“the players and prospects it touts mostly suck!”), but isn’t quite the same thing. Essentially, the book collapses two distinct breakthroughs in the study of baseball into one thing, the success of the Oakland As, when they are best understood separately.

First is Billy Beane’s enduring contribution in promoting sabermetrics — the most significant (canonical?) example being the understanding that OBP and slugging are better measures of offensive output than average. By more accurately assessing the qualities that are associated with winning baseball games, the As had an advantage, and were able to spend less money per win — at least until other teams started to catch on. Since the book came out, indeed, other teams have embraced a variety of new and in some cases proprietary statistics to try to suss out undervalued attributes, and the As are no longer perennial overachievers. Essentially, it was a one-off — you benefit from discovering the Secrets of Winning until everyone else finds out, too.

The other management technique the book deals with is how to wheel and deal, essentially. For example, the As trade their closer (a position overvalued because it has its own statistic, saves) for prospects, acquire a non-closer relief pitcher (who is cheap relative to his skill level because he has no statistic to accumulate), make him the new closer, and sell high a year or two down the line. Or they target players who are stuck in AAA because their major-league team has a star player ‘blocking’ their advancement. Or they play a lot of young players while they’re on (inexpensive) rookie contracts and then let them get overpaid elsewhere later in their careers, while the As receive compensatory draft picks. These can be employed for as long as other teams are buying what they’re selling — and given the closer-related trade rumors this summer and Tampa Bay’s NINE compensatory draft picks last draft, they have proven sustainable.


So the As returning to mediocrity isn’t necessarily an indictment of either of these techniques. What it may be an indictment of is part of the scrappy, can-do-ness that I haven’t talked about — the team’s willingness to overlook the qualities that make someone look like a “typical” baseball player or not. The book discusses these draft selections — a catcher who has great plate discipline but is overweight, a right-handed pitcher who is too short for the majors — as examples of Billy Beane Versus Conventional Wisdom, but they also serve as a great hook, being the most relateable of the professional athletes for their supposed ‘defects’. And don’t get me wrong, I got caught up in it, too. Particularly with Beane’s own career held up as the other side of the coin, Lewis builds what sounds like a convincing case.

Unfortunately, this is just bad analysis. The essential story is this: the As front office regressed data from a sample of baseball player-looking baseball players, and found out that OBP was crucial to winning. Then they found a guy who doesn’t look like a baseball player, but hits with a high OBP. They conclude that that guy will be crucial to winning, even though “looks like a baseball player” seems to have endogeneity problems within their model. In pitching terms, they noticed that tall guys with control do better than tall guys without it, so they found a not-tall guy with control and expected the same results.

That’s why we ended up with 5000-word blog posts assessing the careers of a bunch of career minor-league players standing in as reviews of moneyball the philosophy. Meanwhile, the Nationals were supposedly offered an above-average center fielder (underpaid through 2014 on an extension of his rookie contract) for a pitcher who’s on pace to throw 70 innings this year (but accumulates lots of saves)… and don’t accept!


  1. spongedinosaur posted this