Weighing runs prevented vs. runs created

Mac also loves Castillo's defense. He says this:

"Our major concern is not what either one [Hill or Castillo] is going to do offensively," McClendon said. "Are you going to catch the ball? Because this is not a perfect world. If it's one thing I know, if we're going to be competitive, we can't be giving teams 28, 29, 30 outs a game. It's not going to work."

We explored this subject a little bit in a previous post. I think Mac might be overestimating the effect a defensive upgrade at second base might have on run prevention. 2 or 3 outs per game? Adding 3 outs per game is the equivalent of giving the team another inning. In the national league, teams score about 750 runs per year batting in roughly 1450 innings (which is a hair under nine innings per game; extra inning games more or less make up for the eight innings of hitting that home teams do in a victory). Adding an inning a game by giving away three outs with bad defense would increase runs allowed by 11% ((1450+162)/1450) or 83 runs for an average NL offense. That seems like a world of hurt to blame on one player's defense. Even if poor defense at second costs a team one out per game, that's a free one-third of an inning which should hurt runs allowed by what, 1/3 of 11 or about 4%, or 28 runs total. Even that sounds too high by double. The folks at Baseball Prospectus study this as much as anyone. Even Aramis Ramirez's butchery at third base only cost the team 12 runs, they estimate. Unless Castillo's health is much more promising, and unless his hitting is pretty close to Hill's, his defense alone is not enough to keep him over Hill. Let's split the difference in Mac's argument - say the defense costs 2 outs per game or 56 runs over the course of a season - and put it another way. The difference between a hitter who generates 110 runs and a hitter who generates 166 runs in the same number of plate appearances - say 600 for the sake of an example - is roughly the difference between someone producing .237 / .299 / .351 and someone producing .289 / .340 / .465. Mac is basically arguing that a 650 OPS middle infielder can help a team to score more runs than they allow just as much as a 805 OPS middle infielder with bad defense.

Hey, when I put that way, I think maybe Mac's right. I could see that being true. [Update: The sober second thought says no way, not unless that 805 OPS MIF fields the way I would.] Maybe Aramis Ramirez's play at third cost the team more than 12 runs after all. For a 750 run, 1450 inning offense, saying Ramirez only added 12 runs to the other side's score is like saying he gave away 23 free innings with his 23 errors and general inability to snag just out-of-reach grounders or turn double plays.

OK, I'm confused. I don't think we'll understand exactly how many runs a slick fielder prevents over a bumbler until we can express those runs in a series of equivalent statements that all sound persuasive.

In other news, speaking of runs prevented, Bob Dvorchak explains the wide range of starter-bullpen arrangements the Pirates have been discussing. Turns out they have discussed putting Vogelsong in the pen, a move I suspected about a week ago. Obviously that's not the best solution right now. For the record, I think using Oliver Perez as a fifth starter and swing man is an excellent decision since it will limit his innings, keep him under the wing of the big-league coaches, and not tempt him to cheat on his mechanics as I'm sure he can do when he faces AAA hitting.

[Ed. note: This post was the second half of the previous post. The first digression nows looks substantial enought to justify re-publishing with a new title.]