January 31, 2012
Managing Expectations: Baseball's Next Big Inefficiency
“The management and analysis of data, whether it be scouting reports, statistics, medical information or video, is a critical component of our operation. We look forward to developing a customized program that utilizes the most advanced and efficient technology available in the marketplace today to facilitate quicker, easier and more accurate access to all the sources of information we use to make baseball decisions.”—Cubs President of Baseball Operations Theo Epstein, January 2012
“[Statistical analysis] helps but doesn’t tell the whole story of the game. There is a lot of gut feeling you got to make. If you have a stat and see a flashing number and you see that this guy is doing very good against this other guy, you can use that in a game during a key situation. Yes. But we cannot just depend on stats alone. You got to depend on many other things… I don’t like to become a fantasy manager. The goal for a good manager is to have players who are able to manage themselves on the field.”—Unsuccessful Cubs managerial candidate Sandy Alomar Jr., November 2011
“I do my due diligence and video work and prepare as much as anybody. As far as the stats, those are what they are, and we can use them to our advantage. It's a big part of the game now. It's helping us win a lot of ballgames, the stats and the matchups. That's just part of the game now, and you use what you can."—Successful Cubs managerial candidate Dale Sveum, November 2011
Last month, I tried to size up the state of the game’s front offices in an article called “Keeping Up with the Friedmans.” My point, to the extent that I had one, was that hiring a smart, saber-savvy GM with an interdisciplinary background is no longer innovative. Treating team-building like a science instead of an art is now the norm. Anteing up for an executive capable of putting a progressive process in place was once a suitable subject for a bestselling book, but by the time Moneyball hit theaters, Oakland’s story had spawned so many sequels that the A’s edge had evaporated. In 2012, a baseball team being smart isn’t the exception. It’s the rule.
It’s easy to see this process in action. A couple weeks after my article appeared—not that I’m taking any credit—new Astros GM Jeff Luhnow hired the Cardinals’ Sig Mejdal as director of decision sciences.* A couple weeks after that, Houston made a scientific decision to hire Mike Fast, who joined former public PITCHf/x analysts like Joe P. Sheehan, Dan Fox, and Josh Kalk in transitioning to a top-secret team job. The Cubs, having already imported much of Boston’s former braintrust, partnered with Bloomberg Sports to build “a state-of-the-art player evaluation system.”** Just like that, two teams that had been criticized for being slow to adapt suddenly fell into lockstep with the rest of the league’s forward thinkers, leaving precious few backward thinkers behind.
*Until recently, the Astros were a team that mostly tried to decide things like, “Which Phillies player and/or prospect should we try to trade for?” and “How much money can we offer Brandon Lyon without making him suspicious?” Now, not only do they have decision-making down to a science, but they’re probably asking better questions, too.
**Disclaimer: I also work for Bloomberg Sports. I don’t know if I had to mention that, but it seemed like something a professional journalist would say, and technically I’m supposed to be one.
We’ve established that teams are getting smarter. It’s becoming increasingly difficult for them to discover something their competitors don't already know. And as I wrote in December, “the less variation there is among GMs, the stronger the correlation between spending and winning will become.” But maybe what teams know isn’t all that matters. Maybe what they do with what they know now matters more. Before we give up on the underdogs and leave baseball to the big-market teams, it’s worth pausing to consider whether the biggest inefficiency out there is right in front of our faces.
“I think one of the pieces of our strategy going forward is to make decisions based on the best and most timely pieces of information that you can have at your fingertips at the time you need to make the decision.”—Astros GM Jeff Luhnow, January 2012
As long as there’s been baseball, there’s been baseball’s tendency toward specialization.
It used to be that baseball was played by amateurs. That’s not to say the players weren’t skilled, but they didn’t devote all their time to playing and practicing. They spent their days doing whatever unsavory and unhygienic things 19th-century people did to make money, taking breaks to play baseball. Then people started paying to watch them play. Once someone realized there might be some money in being good at baseball, the first openly all-professional team came along and beat the amateurs 89 consecutive times. Now playing baseball is an occupation in its own right, and when those of us who pursued other occupations want to play, we're forced to seek out other people who aren’t very good at it, either.
For a while, some players pitched and also played other positions. Some of them were among the best at both. But hitters got better at hitting, and pitchers got better at pitching, and before long—the occasional Brooks Kieschnick aside—it became almost impossible to split your time between pitching and hitting and have a career. Later, the American League let roughly half the pitchers stop pretending they could still compete at the plate. Hitters continued to improve, and teams reacted—maybe overreacted—by adding many more pitchers, each of whom pitched less. Monte Ward pitched 76 percent of the innings for the 1879 Providence Grays, who used two pitchers. Justin Verlander was the hardest-working pitcher in baseball last season and was on the mound only 17 percent of the time for the Tigers, who used 26.
Specialization isn’t unique to baseball. You’re reading these words on an advanced piece of technology that’s probably a total mystery to you. Admit it—you have no idea how your computer works. Don’t feel bad. I wrote the words you’re reading, and I don’t know either. Not only could you not assemble your computer from scratch, but you probably couldn’t put it together if you had all the parts. You just press the power button, and it turns on. If one day you press the power button and it doesn’t turn on, you’ll call someone else who knows how to fix it. And that’s okay. If you’d taken the time to learn how to build a computer, you wouldn’t know everything else you know. Whatever it is you do know, the people who can fix your computer probably don’t. (Except for your passwords and the websites you’ve been browsing. Those, they probably know.)
Now that I’ve mentioned it, you might feel miffed that you’re so dependent on something you don’t understand. But damage to your ego aside, the system works. If we all get good at something, we can collectively be better at more things. And whether we all understand how we do it or not, we can build amazingly complex machines capable of handling incredibly demanding tasks that we couldn’t perform without them. Then we can use all that processing power to read articles about baseball on pixels instead of on paper.
What does not knowing how computers work have to do with baseball’s biggest inefficiency? It’s less noticeable, but coaching staffs have undergone the same transformation that players and society as a whole have. In baseball’s early days, there was only one coach: the manager. As managing became more demanding, skippers began to delegate, led by John McGraw. That meant more coaches. At first, all-purpose coaches appeared, and then they too became more specialized. Pitching coaches. Bullpen coaches. Base coaches. Hitting coaches. Eventually, coaching staffs were capped at six (not including the manager), but teams are starting to find even that allowance restrictive. St. Louis has two hitting coaches, only one of whom wears a uniform. The Cardinals won’t be the only ones this season, since the Padres think that idea sounds swell.
*Edit* Another sign of coaching specialization I meant to mention: we haven't seen a player-manager in the majors for almost 30 years.
But managers aren’t slacking off with all that extra help around. With millions of dollars riding on the outcome of every game, it’s even more important that they keep their players motivated and in line, a task that probably hasn’t been made easier by sky-high salaries, a powerful Players Association, and the plethora of languages, nationalities, and ethnicities that make every major-league clubhouse a miniature Babel. On top of that, they have to spend hours each day providing injury updates, explaining decisions, and answering questions asked by mouth-breathers like me. They have to break bad news to struggling players. They have to report to their superiors. And on top of all that—which amounts to more than a full-time job—they’re supposed to manage the games.
For the most part, they’re very good at managing those games, though you might not know it to watch one on TV with a typical fan. The typical fan has been known to claim that anyone (him, for instance) could do better in the dugout. The typical fan is wrong about that. But someone probably could.
In a chapter in our upcoming Extra Innings book called, “How Can We Measure the Impact of Managers?”, Steven Goldman looks into how we can—well, you can probably guess. What he found is that the worst managers may have cost their teams several wins in a single season through poor tactical decisions—namely, sac attempts, intentional walks, and low-percentage stealing. Managerial seasons that bad don’t happen very often—skippers who cost their teams several wins were the Brandon Woods of bad managing. But there were some decisions that Steven’s analysis didn’t consider: starting the best possible players, removing starting pitchers at the optimal times, replacing them with the best possible relievers, pinch-hitting, hitting and running, ordering lineups, and the like.
It’s hard to say how many wins a typical manager costs his teams through decisions we know to be suboptimal. But we do know a lot about how tactical choices affect a team’s chances of victory, and teams know more than you or I (unless you work for a team, in which case congratulations on knowing things the rest of us don’t and please stop hiring our authors). And what we know suggests that managers are either costing their teams wins or failing to secure them, depending on your point of view.
I don’t entirely trust myself to reach that kind of conclusion. Most of my clubhouse visits have been brief. I’m a second-hand sabermetrician. I sometimes mentally mispronounce “misled” as “my-sild” thanks to a childhood reading session gone wrong. In short, I might not be right. So to sanity-test the theory that teams are hamstringing themselves by giving old-school managers relatively free rein when it comes to in-game moves, I did what any internet writer with delusions of grandeur and a brand-new BBWAA card on the way would do: I polled several front office sources.* Here’s what I asked them:
*And by “several,” I mean “seven.” Why seven? Because that’s roughly when I ran out of front office sources. Kevin Goldstein talks to seven front office sources before he gets out of bed in the morning. Let no one say I’m well-connected.
The responses ranged from “1-2” to “5-10.” The average was approximately three wins. That seems reasonable, if not conservative, given how many crucial managerial decisions come in high-leverage spots. It’s generally accepted that a manager can’t make a bad team good or a good one bad, and that’s probably true. But three wins is an awful lot. Three wins cost a fortune on the free agent market—much more than any manager might hope to make. And adding these three wins wouldn’t require any elusive discovery—finally figuring out defense, say, or learning how to keep pitchers healthy. These three wins are achievable with what we already know.
Let me make a couple things clear. I don’t think managers are dumb, and I don’t think anyone with a win expectancy table could do their jobs better. But I do think teams might be well-served by altering what “manager” means, which they’ve done plenty of times in the past. Thirty years ago, it made sense for the manager to pull all the strings. Front offices weren’t any better-equipped to make tactical decisions than the manager was. Things are different now. Managing people is just as important, and managing the media is more important than ever. But more of our hard-won knowledge about what makes teams win has found its way to the front office than the manager’s office, which suggests that more of the decisions should, too.
If anything, I'm arguing for the same beer and tacos served in the front office to be displayed in the dugout. Those three wins aren’t worth tuning out the intangibles. I’m not suggesting that teams simply tell the smartest stat guy on hand to take over, or that they drape a jersey over a laptop, insert some tobacco into its ExpressCard slot, and let it simulate every move before making one. (Laptops aren’t allowed in the dugout, though that’s probably the least of the problems with that idea.) But there are a few less drastic things they can do instead.
Most feasibly, teams can target a Joe Maddon type, a managerial candidate who can connect in the clubhouse but also has the open-mindedness to practice what the front office preaches. There may not be many Maddons out there, but there must be more on the way. Remember those Cubs quotes at the top of the page? There’s a reason why only one of those guys got the job, and it likely has a lot to do with a willingness to make data-driven decisions.
A team could also reinvent the bench coach role. Instead of letting the manager bring on his own best drinking buddy to sit by his side, a team could make the bench coach its conduit for real-time tactical input. Maybe that six-coach limit could be expanded to include a coach devoted to tactics alone. More ambitiously, a manager could be brought on with the understanding that making moves that go against the sabermetric book isn’t part of his purview. If that arrangement were to become known, of course, it could damage the manager’s credibility in the clubhouse, jeopardizing everything else he’s paid to do. So there’s that to consider. But it might also prevent (or delay) a firing in the future, when a progressive front office comes to realize that it can't live with a manager who doesn't view the game the same way.
I probably don’t need to tell you that there are a number of other reasons why most of these things haven’t happened, and why they probably won’t happen anytime soon. For one thing, there’s the fact that they haven’t been done before. A managerial candidate accustomed to making all the moves himself wouldn’t take kindly to having that responsibility stripped from him, which might force an early adopter of the new managerial order to bypass the most attractive options on the market. (Remember how Hollywood Art Howe reacted to Hollywood Billy Beane’s instructions to start Scott Hatteberg?) But being a general manager used to be a one-man job, too—now, even the smartest executive has to rely on a sizeable group of assistants who oversee certain aspects of his club’s operation, freeing the GM to make more media appearances and tend to other recently-added aspects of the job. If a GM can surrender some control to run a team more efficiently, a manager can make the same sacrifice.
There’s also the fact that managerial hirings aren’t always the GM’s call. Owner interference won’t go away as long as rich guys with big egos and business backgrounds can buy baseball teams—in other words, forever, or at least as long as there are baseball teams to buy. It would take a combination of a particularly pliable owner and an especially autonomous GM to push for a fundamental restructuring of the balance of power between coaching staff and front office. Of course, if a team was crazy enough to try the “College of Coaches,” anything is possible. I don't know from football, but my understanding is that most head coaches leave the tactical heavy lifting to their offensive and defensive coordinators, while they limit themselves to giving pep talks, looking angry on the sidelines, and taking occasional Gatorade baths. There might be something to that.
Just as it once might have seemed unthinkable that a manager would outsource most of his hands-on coaching to a group of subordinates, it seems far-fetched now that one might not make all of his team’s tactical moves. But in a business as competitive as baseball, potential advantages don’t stay unexploited for long. Teams are spending more and more money to acquire both data and the teams of executives qualified to distill it into actionable information. Front-office types become as impatient with misguided managerial moves as anyone watching at home—if not more so, since their own jobs, reputations, and investments of time are at stake. Managers will always be important, since their presence in the clubhouse means they know things a statistical model might not. But look at that Luhnow quote again. If teams are aiming to make “decisions based on the best and most timely pieces of information that you can have at your fingertips,” how much longer will they entrust that information to someone liable to go with his gut?