Tim Marchman on Moneyball and the 2012 A's

By Jason Wojciechowski on October 8, 2012 at 11:00 PM

To get my mind off of the impending doom that is the playoffs: Tim Marchman had a piece up at Slate yesterday that takes another look at Moneyball and how the 2012 A's compare to the team depicted in that book. I tweeted about it and immediately got a lot of push-back both from smart people who I respect greatly and from people who I don't know, but who seem quite intelligent in their tweets to me. This tells me that a deeper dive than "this is interesting" is necessary.

First, let me lay out what I think Marchman is saying in bullet point form. This is the place to start with your reading of the piece — that is, do you disagree with my understanding of his argument? Fortunately, the piece is straightforwardly and clearly structured, so I can essentially give a bullet for each paragraph or two seriatim and, I think, capture the meaning.

  1. Many people have compared this A's team to the 2002 one depicted in Moneyball.

  2. The Moneyball A's employed novel theories to build a great collection of talent, saw it not fully come to fruition, and saw the theories be stolen.

  3. The actual A's of 2002 were just as much driven by Mark Mulder and Tim Hudson as they were by the Moneyball players.

  4. Michael Lewis's core thesis was merely the tautology that success comes from acquiring things that are worth more than they cost.

  5. The A's were not unlike the Twins of the same era: minor leaguers who developed well, supplemental stars, and veteran role players.

  6. Beane moved away from the tautology and "embrac[ed] indeterminacy" by acquiring volume and waiting for some of that volume to work out.

  7. Examples of this in 2012: Jonny Gomes, Chris Carter, Sean Doolittle, Brandon Moss. Even Yoenis Cespedes is a roll of the dice, not necessarily a shrewdly chosen undervalued asset.

  8. The core idea of Moneyball is that the world makes sense. The core idea of the 2012 A's is that sometimes things come together.

The reason why I think this is interesting is because the typical stories around Moneyball amount to saying "It was about undervalued assets, not on-base percentage" and then pointing out how everyone else caught up to the game and thus the market tightened up and made it that much harder for Beane to get what he needed to compete, as his budget hadn't changed in the meantime. Marchman's piece looks at Moneyball from a different perspective, though hardly a fully revisionist one, and notes that the essence of Lewis's story is that things are in the general manager's control and thus that via application of smart business methods, the unfair game can be won.

Which, stealing liberally from my Twitter mentions, does not mean the piece is perfect.

First, there may be some conflation of the 2002 team and Moneyball. Note how paragraphs 2-4 above are about Moneyball, then paragraph 5 describes how the team in real life was different from how Michael Lewis's narrative portrayed it, but then in paragraph 6, Marchman claims that Beane actually had been behaving as Lewis stated, and then moved away from it.

The difficulty with the actual 2002 team and the Moneyball version of that team shading into each other in Marchman's piece is that it raises questions about whether Beane really was behaving differently before, say, 2007. As Ken Arneson pointed out on Twitter, Beane acquired Scott Hatteberg, Carlos Pena, and Jeremy Giambi to play first base — that looks an awful lot like embracing indeterminacy long before the shift supposedly occurred. Other supposedly shrewd acquisitions can be viewed the same way: Chad Bradford, for instance, was undervalued, but he was also easily dumped if he didn't work out, no different from the Sean Doolittle conversion project.

If the piece ignored how Beane actually built his teams and made no claim about a shift, the argument would be cleaner: the thesis then is simply "the 2012 team is not built on the principles espoused in the book."

But then comes in another Ken Arneson tweet: this is an apples to oranges comparison. I think Ken might be right about that, and we might even be able to take it one step further: it's an apples to fictional oranges comparison. What I think underlies this point is the basic question: so what? What value is there in comparing a real-life team built based on some set of rationales and principles to the narrative depiction of a team based on some other set of rationales and principles that is, charitably, not fully descriptive of what was actually going on?

I don't think this argument completely destroys Marchman's piece, though. Or at least not the idealized version of the piece that I may have constructed. Pretending, again, the "cleaner" argument I described above, I think there's some value in pointing out that successful low-budget baseball teams (or at least the A's) are not built by a roomful of geniuses outsmarting the world. They're built by grabbing as many good players as they can manage and finding upside where they can and hoping the team comes together enough to make a run at the postseason.

Now, to make that argument really work, you'd like to see it fleshed out. Most ideally, you'd like a full account of how the 2002 team was not actually built the way Lewis describes, but rather through the serendipity of high draft picks who stayed healthy and developed and scrap-heap pickups coming together at the same time to form a winning team.

Even then, though, you might be left with two contrasting narratives and no way to choose between them. Are Sean Doolittle and Chad Bradford the results of enough dice rolls that eventually had to come up sixes, or are they the result of shrewd evaluation that other teams missed? Chris Carter, Brandon Moss, Kila Ka'aihue, Daric Barton vs. Jeremy Giambi, Carlos Pena, Scott Hatteberg — which one is killer insight and which one is throwing pasta at the wall?

(I should note that these are obviously stylized notions of how the front office is going about its job. Embracing indeterminacy does not mean acquiring any old random players and crossing your fingers — it means accepting that outcomes aren't fully predictable, no matter how good your process is. Conversely, Lewis's narrative does not, I don't think, actually posit that baseball is a solvable game in the sense that wins can be guaranteed for as long as a team is more intelligent than other teams.)

One significant distinction between the early aughts A's and the current ones is that the team now is in no small part the residue of a bunch of trades of stars for prospects (and the stars were the results of such trades themselves). It's not clear, though, that the distinction actually contains any difference — the Moneyball A's were built from high draft picks because the team in earlier years didn't have enough good players to win any games, much less be traded for good younger players.

In the end, I think my Twitter compatriots may be right: Marchman's piece doesn't fully hang together. That said, I don't think it's worthless — it is, as I mentioned, in (on?) Slate, a publication that is not part of the sabermetric mainstream. A very large number of casual baseball fans surely read it. Certainly more than read FanGraphs. To the extent that they completely buy into Marchman's argument, maybe we have quibbles and disagreements with them. However, it strikes me that there is a very important, very base-level takeaway for said casual fans, which takeaway we (if you're reading this particular blog, you are very likely "we") should probably applaud Marchman for putting in their heads: that large, enormous, huge amounts of baseball outcomes, at all levels from the pitcher working his craft up to the GM working his, are not determined but are instead more or less unpredictable.