Thursday, March 6, 2008
Taking the bullpen by the horns
Posted by Ted Cox on 03.06.08 at 12:05 PM
All right, having looked at the Cubs through the statistical prism of The Bill James Gold Mine 2008, let's see what it says about the White Sox. First, the bullpen was awful last season. No surprise there. Excluding the closers -- and Bobby Jenks was exceptional -- Sox relievers posted a 5.98 earned-run average. The only stunning fact is that there was a worse team in the American League: the Tampa Bay Devil Rays, with a 6.33 ERA. Just how badly that horrendous bullpen hurt the Sox is shown in another nifty stat: the Sox were ahead more often than not through each of the first four innings last year. (Specifically, through four innings, they were ahead 71 times, behind 64 times, and tied 27 times.) That's an amazing figure for a team that wound up 72-90. Of course, from the fifth inning on, when the middle relief typically entered the game, things got progressively worse. Here's hoping Scott Linebrink and Octavio Dotel are cures for what ailed the bullpen.
Yet the collapse of the Sox wasn't just the bullpen's fault. James points out that every Sox batter with more than 150 at-bats compiled an OPS (on-base plus slugging percentage, a key run-scoring indicator) below his career average. (The only exception was Rob Mackowiak, and when you figure in what he did after being dealt to the San Diego Padres, he too came in below his career average.) That's a team-wide batting slump. And with the Sox' reliance on the home run -- and lack of team speed -- they had a miserable 13-44 record in games they were held without a homer, a .228 winning percentage.
Unfortunately, general manager Kenny Williams didn't address this problem by bringing in a run-producing leadoff man -- unless you count Orlando Cabrera, who may not even hit leadoff. The good news there is that the Sox are due for a team-wide return to the norm in hitting. Still, the law of averages is not something a general manager should generally hang his hopes on.
like
dislike
