The Giants just signed Jeremy Affeldt to a two-year, $8 M contract. Affeldt is a B-level free agent, so the Giants won’t surrender a draft pick. More on this later, but first — Ha! Cruel fates. Just before the news broke I finished writing the post below, which I’ll let you read for a few chortles and galumphs.
*********
Because this Hot Stove season will almost certainly find a shiny new free-agent relief pitcher wrapped up under Lou Seal's Christmas tree, I need to submit to the inevitable. It's going to happen, kids, so let's get with the program and make a best guess which of the relievers makes the most sense.
The names floated so far are Bobby Howry, Juan Cruz, and Joe Beimel, with a teaspoon of Jeremy Affeldt added for flavor. But I’ve read Affeldt wants a chance to start, so I’ll ignore him for now. (Forget those guys, I say: Sign this enticing prospect.)
Plenty of statistical options exist to evaluate relievers. As perhaps you’ve heard, Joe Beimel has only allowed one home run in his last 116 innings. That’s a strong mark in his favor.
But other stats don’t cut the mustard at all: ERA, saves and holds are just plain silly, and if you don't know why, here's an extreme example: Joe the Bullpenner enters with his team up by three, two outs, and the bases loaded. He gives up a bases-clearing double, but the third runner is thrown out at the plate by a hair. Pretend this happens a lot. He gets the save (or hold), his ERA is 0.00, and he sucks.
I’m going to use Baseball Prospectus's WXRL, a fancy stat that measures not only a reliever's effectiveness but factors in the pressure (or leverage) of the situation. It’s expressed as an ERA-like number, except the higher the better. Most of the top relievers by this measure are closers: Brad Lidge, Mariano Rivera, Frankie Rodriguez, Joakim Soria, Joe Nathan, Bobby Jenks and Brad Ziegler (who took over for Huston Street in Oakland) are all top-ten, ranging from a 7.59 (Lidge) to 4.39 (Ziegler). Brian Wilson just missed, ranking 13th with 3.53. More points of comparison: Sergio Romo was next in the Giant bullpen with 1.85, and Tyler Walker was third with 1.06.
Let’s see where this year’s free agents rank.
Juan Cruz was Arizona’s fourth-best reliever in 2008 with a measly 1.12. He fared worse in 2007, even worse in 2006. He’s got a great arm and strikes out a lot of batters, but will he suddenly thrive in a key setup role? Paying him $12 M over three years (call it the Marte precedent) to find out seems a foolish gamble.
Bobby Howry had negative WXRL in 2008, which means he subtracted wins from his team even though he racked up the second-most appearances in the Cubbie pen. In 2007 he fared a lot better, with a 3.13 WXRL. In 2006 he led the team with 2.17. Just for fun, let’s go back two more years to his Cleveland days. In 2005 he had 3.11, in 2004, which seemed injury-shortened, he barely cleared zero. Conclusion? If this year was an aberration, Howry seems reliable when healthy. He turned 35 in August, by the way. Doesn’t sound like Marte money to me.
Joe Beimel racked up the most appearances out of the Dodger pen this year but only scored a 1.32 WXRL. The previous two years he scored 2.20 and 2.44.
I realize that WXRL depends on how a pitcher is used, and how a pitcher is used depends on his manager. It’s possible that Juan Cruz is an 8th-inning revelation just waiting for someone to trust him. But it’s more likely he’s never gotten lots of highly-leveraged innings because his previous performance hasn’t warranted that trust.
Giving him the chance is all well and good. The problem is the money. You don’t pay $12 M over three years to conduct that kind of socio-economic experiment. But as I noted in the top of this post, Brian Sabean seems absolutely determined to give an undistinguished relief pitcher $10 M and change, so if it has to be one of these three, how about Howry for two years, $8 M, plus an option?
