BP Idol
I sent in for it, but since mine didn’t make the final cut and apparently isn’t going to be invited to even be one of those one-shot “interesting, but not quite good enough” articles, I figure, hey, easy blog post! Anyway, it’s all about managing and the future of statistics in evaluating it. Enjoy!
Managing: The Great Subjective Oasis
“Okay, lets walk this next batter, set up the double play.”
“Oh Hank, I can strike out this lapdancer in my sleep.”
“No, it’s not good strategy! Walk this gal and pitch to the one with the ridiculous implants up next!”– King Of The Hill, “Take Me Out Of The Ballgame”
Sabermetrics has penetrated almost every front office to this point, whether it is used extensively like in Boston or given lip-service as it is in Kansas City. The majority of the analytical war is over, and natural evolution will follow in front offices across America as Willy Taveras and his ilk are slowly marginalized. The internet has provided better access to educational material for fans, and they’re in the weaning process to slowly get off the influence of old school baseball clichés.
However, there remains one area of organized baseball that still seems to be a safe haven from statistical analysis: managing. Whether it’s a baseball site, football site, basketball site, team-specific or general, one of the staples of the fan experience is that we all get together and talk about how stupid our manager is. You can get it from a Nationals blog or a Red Sox blog, you can read about it in a Pirates game in mid-April or the World Series.
Despite the fact that we all love to critique it, very little statistical research has actually been employed to determine the skills of a manager. Baseball Prospectus 2008 sums it up bleakly with “There are many aspects of a manager’s job that defy enumeration, and as a result, few bother to try. For the most part, only two statistics are ever brought to bear in evaluating manager’s performance: wins and losses.” Indeed, the fancy new redesign at baseball-reference has brought us so many batting and pitching statistics that our heads spin, but the managing statistics remain pretty simple: wins, losses, finish.
In 2006, The Book was published by Tom Tango, Mitchell Lichtman, and Andrew Dolphin. A revolutionary tome, it set out to detail in great analysis what optimal strategies would be if playing by the theoretical “book”, which unsurprisingly differed greatly from what the actual conventional wisdom of the time has gone with. While trying to do some research to soothe my head about Jerry Manuel’s early season proclivity to bring in five relievers a game, I reread parts of The Book and it made me wonder how hard it would be to concoct even a rudimentary rating system to look at.
The amount of subjectivity one would have to remove to get real results out of the project is staggering. I would think most fans expect the end rankings to purport that most every manager in baseball was doing his team a disservice, making it a situation where the “least bad” was really the best. There would also be several suggestions in The Book that simply would be unfeasible to lay solely at the hands of the manager. The “three starters, two swingmen” philosophy is something that would probably require more of an organization-wide commitment than simply a managerial decision, for example. Free agent starters would probably be unwilling to sign if they wound up having their innings cut.
But is it possible?
So What Could Be Used To Rank Managers?
Perhaps the simplest matter is their lineup. Hitters are the more predictable creature, and given the amount of research that The Book provides on the subject, it would seem pretty easy to judge a manager on his lineup choices. Kudos can be given for batting the best hitters where they should hit, or for running a successful platoon. However, you wind up with questions such as “what about players with no major league track record?” or “what happens with a player who has been in a three-month slump that may wind up being more?” that make even this task seem a little harder than you’d normally think.
Starting pitching gets a little more complicated. Is the manager skipping his fifth starter as often as possible? If he’s not, does he start after June when he’s got a better handle on how his team is playing? The larger problem is that every pitcher has different stuff and tendencies. Does this guy have the stamina to even get me to 90 pitches? Or is his stuff so unaffected that he could go 130? If this guy lets runners on, is his stretch pitching so bad that I need to go get him in any situation? The lineup you could judge on a team level fairly quickly, but for the rotation you need to look at each pitcher as an individual.
With substitutions, you run into an even greater array of subjective problems. Perhaps the manager hasn’t used the best pinch-hitter on his bench? Or perhaps the manager used the third-best, but justified it because his hitter hits fastballs and the pitcher throws mainly fastballs? Is he making his relief corps make short appearances when they could handle long ones? Is he properly leveraging situational pitchers, or is he leveraging them to the point where it costs his team because they can’t throw to enough batters? Is he consciously using his best reliever in the highest leverage situations?
Strategically, there are even more decisions. You can’t just have a blanket no-bunt rule, but the idea of a manager bunting too much or not enough is also, to a point, team dependent. Is he bunting in good situations for his team? Does he sometimes let his average or good-hitting pitchers swing away in bunt situations? Is he shifting against everyone he should be shifting against? Does he abuse the intentional walk or pitchout?
As you can see, such a project would be a challenge. Not only would it encapsulate every little piece of objective data that the statistical community has been building for years, but it would also have cutoff points and limitations that would limit its effectiveness as an analytical tool unless we had the same knowledge each manager has.
The Future
This wall shall also crumble one day. While all we really have these days to hold onto besides The Book are vague hintings of Billy Beane running the show from his office or Manny Acta reading Prospectus, managing will probably be the next forefront of sabermetric focus. I would bet quite a bit of money that we’ll see some sort of basic managerial analytical tools pop up sometime in the next five years.
But will it reach the owners? This is a much touchier subject than player acquisition. With that, you could point to a much greater cost savings and there was a clear divide between the business side of baseball and the actual baseball lifers. This doesn’t save the same amount of money, guarantee any improvement to the club, and is quite a bit riskier in terms of human interaction. Bring in a statistical guy who can’t run a fielding drill or demand the respect of the players, and the clubhouse could divide. Bring in someone as a surrogate tactical manager and you risk alienating the actual manager, perhaps at least to the point where he ignores the tactical suggestions.
Pardon the tacky fantasy/politics association, but this situation calls for a uniter. Perhaps there is a youth or player among us who can read the stats and also carry himself as a baseball lifer. That will be the next barrier. It may take some time, as it usually does in life, but logic and education find a way to beat tradition eventually. When that day comes, we’ll probably all still think we know better. But at least we’d have reason to pause and think about it.
