The economics of Moneyball

At, I have a new piece up with Kevin Grier, excerpt:

Here’s something funny about the Moneyball strategy: It is bringing us a world where payroll matters more and more. Spotting undervalued players boosts their salaries and makes money more important for the general manager; little did Billy Beane know that in the long run he would be strengthening the hand of the large home-market teams, such as the Yankees. From 1986 to 1993, payroll explained 2.2 percent of the variation in team winning percentage, and that meant spending more money yielded little return in terms of quality on the field. In the 2004 to 2006 seasons, after the Moneyball revolution was under way, payroll explained 27.1 percent of the variation in team winning percentage, which means a stronger reason to spend more.

One of our sources you will find here, a very good paper by Jahn Hakes and Skip Sauer.


All this rests on the assumption that Moneyball ever actually happened.

Well it did, but the thing everyone forgets is Oakland had been way overperforming payroll even before Moneyball, their scouts, represented as triglodytes in the movie, were some of the best in baseball.

I really can't stand how Sorkin's idiotic tropes so consistently get confused with reality.


Thanks! An iphone on the road is a terrible way to write.

Err, shouldn't this be viewed as a piece of evidence in favor of Moneyball happening?

If Moneyball did happen, this is exactly what you'd predict-- discovered and widely reported arbitrage goes away.

It doesn't "rest on the assumption" it argues in favor of it.

The As didn't win from '99-'03 because of the "Moneyball" players. They won't because they had a core of about a dozen young players, both hitters and pitchers. These guys accounted from about half their runs and two thirds of their wins.
As for the sudden increase in value of on-base percentage, remember that Moneyball was a management and baseball, like another business, is run by the Pointy Haired Boss from Dilbert.

Who's talking about the sudden increase in value of on-base percentage, Ted? The evidence is the massive increase in the effect of payroll on wins.

Of course the price of on-base percentage going up wouldn't tell you whether or not on-base percentage was actually useful. But the fact that spending became more correlated with winning certainly implies that management starting getting more for its money somehow.

Yes, but there's many reasons for the change and I really question how much you can attribute to one overrated management book and one overrated general manager. For example, the divisions became much more balanced following re-alignment after the strike. All the biggest spenders were put in two divisions.

I really question how much you can attribute to one overrated management book and one overrated general manager

It's clear that you and I are using "Moneyball" to refer to different things. I do blame the movie and the incorrect popularization of what went on. You're arguing against the popularized conception of it, which is not what Tyler and Angus wrote about.

For example, the divisions became much more balanced following re-alignment after the strike. All the biggest spenders were put in two divisions.

But wouldn't this cause the opposite affect? If all the biggest spenders are in the same division, one should expect the effect of payroll on winning percentage to decrease, not increase. The schedules are unbalanced, you play extra games against teams in your own division.

You're committing a tremendous error if you think that Moneyball was just about the high OBP low strikeout bad fielding players initially identified as undervalued at one point.

Admittedly, it's the same error that many other people, including those who made the film, do.

Saying that "Moneyball happened" doesn't mean "on-base percentage was (and always is) undervalued." It's saying that "management got a lot better at measuring talent instead of randomly signing players because they 'looked like ballplayers' or looking for skills that didn't correlate with success."

The move from "payroll explained 2.2 percent of the variation in team winning percentage," to "payroll explained 27.1 percent of the variation" is strong evidence that management in all of baseball got much better at evaluating the type of talent that led to wins, and exactly how much it contributed.

Beane's specific theories could even have been total luck, but the overall advancement in sabermetrics and baseball management's use of it is real.

You have the pattern recognition of the scouts on the one hand and the cold rational methodology of Beane on the other hand. There is probably an optimum that uses both, but that lacks drama and this way someone will get to write a book about the scouts taking it back from the soulless Beane counters.

Why would Bean care about that, his goal was to win this season, and maybe build something for the next couple of seasons. Considering every year every good player he had was bought off him, I doubt he had any reason to think about the three-four year time horizon.

Agreed. Oh and less we forget about the pitching staff of Oakland at the time. Also include the fact that Miguel Tejada was present.

If I follow the argument correctly, in the 1980s teams tried to spend money to hire the best players, but they had little idea who the "best" players were, so much of that extra money was wasted.

In the 00s, general offices had a better grasp of statistics, so when they pay a high price to get a good player, they are more likely to actually get a good player.

I think Beane himself has argued this, and said he has to look harder now for market failures.

How much of front office inefficiency in the 1980s is explained by the Yankees alone? Yankees spending is a big part of the market and was extreme erratic at best in the 1980s. Though they had another championship draught in the 00s, the ridiculously paid teams often had the best regular season records in those years and were much better than the ridiculously paid 80s teams.

First of all, Beane DID know that his edge was temporary, and he downplayed his organization's success in public forums to try to keep the secret from getting out. Of course, ultimately it did and now Oakland stinks again. Second of all, 1986 to 1993 seems like an unfair control group for this study because it includes several years of the collusion period, when the owners conspired to shut down free agency. So that would tend to reduce variation in payrolls and fundamentally alter the relationship between payroll and performance.

This is true. Michael Lewis spent exactly one day with the As. (Read it again, you'll notice.) I suspect Beane sniffed out the mission and had him banned, but its clear that Lewis did not have access to Beane or DePodesta beyond that one day in the draft room. Beane did not want this book about him and his strategy.

I'm assuming DePodesta was not pleased either since he wasn't in the movie.

"Beane did not want this book about him and his strategy"

Then he was fixated on the wrong percentages.

"Michael Lewis spent exactly one day with the As. (Read it again, you’ll notice.) I suspect Beane sniffed out the mission and had him banned, but its clear that Lewis did not have access to Beane or DePodesta beyond that one day in the draft room."

Are you sure about that? This article suggests quite the opposite.

Lewis has often claimed that he spent a large number of days with the As and with Beane. He lives about 15 min from the A's stadium and is obviously a sports fan. Need a source for the claim that he spent only one day.

Need a source hes a sports fan.

I personally had the chance to ask Beane why he let Lewis stick around to write the book. Beane said that he initially thought that Lewis was writing a short article for the NY Times Magazine and it would have limited impact. After a couple of months it occurred to him that Lewis had visited far more than he needed to for an article so they asked him and learned that Lewis was writing a book. He joked that by then it was too late to stop him and everyone had looked forward to seeing Tabitha (Lewis' wife) on her visits.

The cost comparisons of closers are flawed to the point that that portion of the article should be HEAVILY revised.
It ignores the fact that the guys who are making very little money are making that amount (in many cases) due to the way that MLB free agency is structured - players are cost-controlled early in their careers, and then become eligible over time for arbitration and free agency.
What you're comparing when you compare costs (in many cases) is the effectiveness of a team's scouting and player development and luck in having drafted well, not the way in which different teams value players.

I really hope when you say "2004 to 2006", you're looking at total salary across those 3 seasons, and comparing it with total W/L record. I hope you're not looking at each team-season from those three years individually.

You seem to misunderstand how salaries of players who are not eligible for arbitration works. Neftali Feliz and Jason Motte wouldn't be paid the amounts they got on an open market.

Thats not exactly a fair response. The Rangers and cardinals certainly could have paid for an experienced veteran free agent reliever to be their closer -- instead they just used young, cheap fireballers to fill the role who, a decade or two ago, might not have been placed in that position because they were thought to lack the gut to handle the 9th inning.

Feliz did lack the guts. That's part of why the Rangers lost.

I get what you're saying, but this is a bad example. They didn't need to pay for a good free agent reliever because they had good in-house relievers who they would have had to pay a lot of money to acquire on the market. It was not choosing cheap/bad players because they recognized the limited impact the players in those positions had. In fact, if they had made that realization Feliz should have been a starter already.

Agree with Cliff and Dan. Example: The Red Sox, whom the authors say are sticking to the closers-are-valuable strategy because they can afford it, just let their pricey closer walk and replaced him with a pre-free agency substitute in Daniel Bard. The closer market is a bit of a crapshoot nowadays--plenty of teams (not the Phillies!) have realized that you don't need to pay a closer $50 million to get three outs with the bases empty in the 9th inning up three runs. But contending teams still need a good pitcher there, and not every one has a cheap, good, young fireballer hanging around.

I would probably argue that middle-relievers are one of the big market inefficiencies these days--that is, the smart teams avoid overpaying for the services of a "proven" middle reliever, given the evidence that past performance of middle relievers is only very weakly predictive of future performance.

Isn't the simpler explanation that where once, small-salary teams could compete without spending by exploiting an under-recognized talent (e.g., OBP), now that everyone's caught wind, that particular talent is no longer under-recognized and cheap?

This is a good story.

The Beane-as-lucky-stockpicker analogy seems to me odd. It's not as if all GMs throughout time have picked random stats as the next big thing and Beane got lucky in the one he hit on. And I also think it's odd to say that "little did Billy Beane know that in the long run he would be strengthening the hand of the large home-market teams." I don't think Beane ever had any delusions about being the only GM who would ever adopt sabermetrics, and thus wouldn't he have to understand that he is strengthening the hand of the Yankees by exposing to them one of their inefficiences?

As someone who read Bill James back in the 1980's I have a difficult time believeing the moneyball myth.

The information was widely available before and after.

Furthermore, there was a pretty good correlation between player reputation and value created per Bill James back in the 1980s. There were a few noted consistent exceptions like walks undervalued and stolen bases over valued, but this did not affect all that many players significantly.

And I'd also add that the hypothesis could be tested far more directly seeing if individual player salaries have become more correlated to various sabermetrics like Runs Created.

Gene Mauch famously put Brian Downing in lead-off, and Earl Weaver's book on strategy covers a lot of modern "innovations", as does Weaver's opinion of Terry Crowley.[Definitely NSFW]. I suppose the exception prove the rule.

So picking up 100 dollar bills laying on the ground makes the market more efficient!? WHO KNEW!?

Dollars per save has always been a lousy way to measure a closer's value, since even before the sabermetric era produced much better metrics. If the $500K/year closer on your team blows three-quarters of his saves, he is not as cost-effective as the $2M/year guy on your division rival that doesn't blow any saves.

In any case, Feliz and Motte make close to the league minimum because they are young players, subject to the many arcane rules of MLB that place them firmly under the control of their respective teams. Their salaries do not reflect the value of their contribution to their teams (except maybe to establish a lower bound), and they won't until they become eligible for free agency or at least arbitration.

The A's realized this, and it's one of the central points of Moneyball - that draft picks are more valuable than most teams thought, not because the players were especially great, but because draftees' salaries were kept artificially cheap by the CBA.

Gains from innovation! Er, in hitting and catching a ball.

Bill James, Billy Beane, and Michael Lewis all covered up the most significant innovation of recent drugs: performance-enhancing drugs. Talking about the importance of on-base average plus slugging percentage was the perfect cover for playing juicers, like the Giambi Brothers for Billy Beane in Oakland and Ramirez and Ortiz under Bill James in Boston.

A great point.

They've been around since the 1940s.

They didn't have the training regiments and the medical support that, say, Barry Bonds received. So if they used steroids, it didn't have the impact that was seen with someone like Ken Caminiti.

Damn revisionists always mucking up our elegant economics.

I don't buy the claim that Moneyball has anything to do with payroll becoming more important. Revenue inequality increased because of cable tv revenue in the mid90s, long before moneyball. The Yankees and Braves were the first and were dominant. Of course payroll differences matter more when the magnitude of the inequality is greater.

Comments for this post are closed