Tuesday, September 24, 2013

Mattingly, And Why He Might Matter As Manager

Up until midway through June, many Dodgers fans were ready to pull the ripcord on Don Mattingly. So now that he's guided (or at least presided) over a historic run and the Dodgers' first playoff berth since 2009, it's a good time to consider how much of an impact he might have had on that turnaround.

ESPN the Magazine's Peter Keating did just that in the September 16, 2013 issue, in his column "Do Managers Really Matter?" (link insider only). And Mattingly was actually front and center as the example:

UNTIL JUNE 22 of this year, the nation's sports writers tripped all over themselves to excuse Don Mattingly's managerial record -- a mediocre 198-197 to that date. "Mattingly has proven to be a victim of circumstance" was one typical take. Then LA ripped off 42 wins in 50 games, and the same pundits started talking up Mattingly for NL Manager of the Year. This kind of thing drives me bonkers. If Mattingly wasn't to blame when the Dodgers failed to meet expectations, why does he deserve so much credit for vaulting them into first place? It is recency bias and a man crush, not sound statistical judgment.

But the Mattingly bandwagon does make you wonder: When it comes to evaluating managers, what is sound statistical judgment? There's no sabermetric berbenchmark like Wins Above Replacement for managers; it's surprisingly hard to track their effect on players or teams. That leaves many statheads hostile to the idea that skippers have any significant impact at all. (Just think of Moneyball's portrayal of Billy Beane the genius vs. Art Howe the spluttering doofus.) As Sabermetric Research blogger Phil Birnbaum has put it, "There can't be a whole lot of manager influence in temporarily increasing a player's talent." [...]

But maybe [a lack of statistical correlation in overperformance, year-over-year, is] the point. Perhaps a manager's true value is not found amid strategic dexterity but in a hitter belting 30 home runs, for whatever reason, when we expect him to hit 15. When Davey Johnson gives regular playing time to teenagers like Dwight Gooden or Bryce Harper, we should credit him for the production of his young players. When veterans rejuvenate their careers playing for Tony La Russa, we should view their stat lines as incorporating his impact. "I think statheads have to be a little careful when they say managers don't matter," Nate Silver told me last November. "What if you can get guys to play toward the higher end of their performance curve?"

This is the idea that Hardball Times columnist Chris Jaffe examined in his excellent book, Evaluating Baseball's Managers. Jaffe studied a huge database that Birnbaum developed, which stretches back to the 1890s and projects how well players should have performed in each season. (Essentially, Prince Fielder's estimates for 2009 are a weighted average of his stats for 2007, '08, '10 and '11.) He figured out how much each team in every season exceeded its aggregate expectations, or fell short of them, then credited the difference to its manager. And he found that over time, managers' totals didn't regress to zero; they grew, and the longer a skipper managed, the more his hitters and pitchers tended to beat projections, indicating skill was involved. Jaffe's book came out in 2010 and ranked Joe McCarthy as the greatest manager of all time and La Russa as the best since World War II.

So whether or not Mattingly wins NL Manager of the Year (my vote would be for Pittsburgh's Clint Hurdle), his value doesn't come from his curfew rules or batting orders. It comes from keeping Hanley Ramirez happy and playing Yasiel Puig -- generating loads of unexpected W's.

It's funny to look at this article now, a couple weeks after it was published, now that the Dodgers have slowed a bit and there appear to have been some managerial gaffes in the last ten games (Juan Uribe bunting in the ninth with none out and two on, for example?). Is Mattingly really getting the most out of his players, like the stories of Tommy Lasorda's gift? I suppose we'll find out soon enough, if this is truly Donnie's gift as well.

0 comments: