Printer-friendly

Contact Author

August 27, 2012

Pebble Hunting

The Most Fleeting Way to Win

by Sam Miller

Pedro Strop was a utility infielder who spent three years in short-season ball and never slugged as high as .350 before the Rockies decided, in 2006, to try his live arm on the mound. In his first try as a pitcher, he struck out 22, and walked just two, in 13 innings. But the Rockies released him two years later, and the Rangers picked him up, waited him out, and eventually promoted him to the big leagues. He had an ERA over 7, with 22 walks in 27 big-league innings, when they traded him to Baltimore as the non-famous half of an August waiver-period trade for Mike Gonzalez.

That was a year ago this week, and Strop has been far more significant in this year’s pennant race than Gonzalez ever was with Texas. He has the second-best ERA in the American League (minimum 30 innings), and among relievers, his win probability added is sixth, ahead of all but four closers. He has the fifth-best groundball rate in the majors, a fastball that averages 97 mph, and no platoon split worth worrying about. The Orioles entered play Sunday tied for a Wild Card spot, though their playoff odds were just 13.4 percent. It’s probably no exaggeration to say that, without Strop's work this year, those odds would be close to zero.

You could say the same about the Orioles closer Jim Johnson, and about the side-arming set-up man Darren O’Day, who rank first and 10th in WPA. You could absolutely say that about the Orioles' bullpen, which has the league’s third-best ERA, and which has been the most obvious reason that the Orioles have gone 23-6 in one-run games, which has been the obvious reason they have managed to get this far despite a similar run differential to the Royals'. When a team's record strays from its run differential, the simplistic assumption is that it's about the bullpen, or luck. But from a team-building perspective, that might be redundant. A good bullpen is ephemeral, and the correlation from one year to the next is basically insignificant.

That’s not hyperbole. The correlation between AL teams’ bullpen ERAs in 2011 and 2012 is an insignificant .06. The correlation between teams’ bullpen ERAs in 2010 and 2011 was actually negative, slightly. That oversells the point a bit: the correlation varies greatly from year to year. But over the past 10 years, the average year-to-year correlation is a very weak .18. (Teams’ starting-pitching ERAs, by contrast, are much steadier. The correlation between 2011 and 2012 is .74, and the average correlation over the past decade is .58.) Considering that I’m using a stat that isn’t even park-adjusted—that is, simply playing in the same ballpark should provide some consistency from year to year—this should be discouraging news for Orioles fans hoping to relive the O’Day-to-Strop-to-Johnson magic next year.

The AL's top three bullpens last year—Yankees, Angels, Mariners—rank seventh, 12th, and fifth in ERA this year. The top three bullpens this year—Rays, A’s, Royals—ranked sixth, seventh, and eighth in the AL last year. The Orioles were 13th last year.

Using the Angels as an example, we can see how bullpen success one year can be so fleeting. The Angels' bullpen in 2011 was a testament to Mike Scioscia’s ability to find roles in which marginal arms could succeed. That’s what I would focus on if I were face to face with Mike Scioscia, and I wanted him to like me. But the other way of saying it is that the Angels' bullpen last year comprised:

• A groundball specialist who far, far, far outperformed his peripherals thanks to a .204 BABIP with runners on base;

• The rest of this article is restricted to Baseball Prospectus Subscribers. Not a subscriber? Click here for more information on Baseball Prospectus subscriptions or use the buttons to the right to subscribe and get access to the best baseball content on the web.

  Cancel anytime. That's a 33% savings over the monthly price! That's a 33% savings over the monthly price! Already a subscriber? Click here and use the blue login bar to log in.