If I wanted to, I could write this post real quick. "Regression to the mean." It works, right? 2009 saw David Aardsma post the best strikeout rate and the best home run rate of his career. On that basis alone, I would be justified in saying that Aardsma's due to get worse, and it's not a lot of fun when your closer gets worse. When a player's season stands out from his track record, the probability is that he will inch back in the direction of his track record the next year.
And that would get the job done. Regression to the mean is the reason why David Aardsma makes me nervous. But it's hard to attach emotions to a statistical phenomenon. It's much easier to attach emotions to something like this:
Let's ignore Aardsma's strikeouts and walks for the moment. Let's ignore everything about David Aardsma but the fly balls. David Aardsma allows a lot of fly balls. Of the 178 balls that hitters put in play against Aardsma in 2009, 96 were fly balls, and 37 were line drives. Only a quarter of those balls in play stayed on the ground, which means a lot of balls were getting a lot of air. 2009 David Aardsma was one of the most extreme air pitchers in baseball.
Fly balls can be good. Infield fly balls are automatic outs. Safeco's a big park, and the Mariners have an awesome outfield defense. Fly ball pitchers stand to benefit the most from the Seattle environment. But fly balls also go for home runs, and they go for home runs quite often.
A very small amount of David Aardsma's fly balls went for home runs in 2009. Only four of them, to be precise. Four home runs on 96 fly balls and 133 balls in the air. When we look at a rate like that, we should almost immediately think "regression", and that's correct, but for the sake of being thorough, I wanted to bring something else to your attention.
In 2009, David Aardsma allowed 17 balls in play to at least the warning track. 13 of them stayed in the yard, and four of them left the park. The 17 is an approximate figure based on MLB's Gameday hit locations, but it should be pretty accurate.
Now, on its own, we don't have any idea what this means. We don't know if 4/17 is good, or bad, or normal, or what. But let's look at the rest of the 2009 Mariners pitching staff, using the same hit location tool. For the sake of simplicity, I'm only going to look at how the guys did at home. It takes too much time to look at how they did in every single ballpark, and I don't think limiting them to home appearances should have any major effect on the results.
In 2009, the rest of the Mariner pitching staff allowed 136 balls in play to at least the warning track at home, and 77 of them left the park.
Aardsma: 24%
Rest of staff: 57%
This is not a perfect measure, by any means. The ball-in-play tool shows where balls were fielded, not where they were hit. I had to eyeball the dot locations. The warning track is an arbitrary landmark. And so on, and so forth. You shouldn't go quoting that 24% or 57% to your friends. These are just approximations I made when calculating a statistic that doesn't exist.
However, I think this shows something. We all remember Aardsma allowing a bunch of fly balls to the track last year, and intuitively, we see those as luck, or unsustainable. If you have a fly ball pitcher, and he doesn't allow a lot of home runs, you might be able to explain it away by saying that he induces weak fly ball contact. Maybe he gets a lot of infield flies and shallow outfield pops. But here we're talking about balls to the track. Balls to the track require solid contact, and when you're talking about balls hit that close to the wall, I have to think that, at that point, it's essentially a coin flip whether the ball goes over or not.
While I haven't run any numbers or performed an exhaustive study, it doesn't make any sense to me that a pitcher could possess the ability to allow a bunch of fly balls to the track, and no farther. Felix: 8/12. Bedard: 4/5. RRS: 6/10. Washburn was 5/19 in Seattle, but then he went to Detroit and went 11/13. Balls hit that well tend to clear the fence, and though this is completely new to me and I'm willing to entertain the possibility that Aardsma can just induce slightly weaker contact than others, I find that to be incredibly unlikely. I find it more likely that Aardsma possesses no such special skill, and that as an extreme fly ball pitcher, more of the balls that didn't clear the fence in 2009 are going to end up clearing the fence in 2010.
It's not a guarantee. This is just an idea of mine. And even if Aardsma does give up more home runs, he should still be a good reliever. He changed his pitching style last year, throwing a lot of high fastballs to righties and outside fastballs to lefties, and he generated a ton of strikeouts while somewhat limiting his walks. That's good. But the problem is that Aardsma's our closer, and as closer, his failures become more visible. What if his home run rate doubles next year? Every single one of those additional four homers could mean a blown save, or a loss. Aardsma saved 38/42 games last year, and people liked him. It's a thin line between a good closer and a shaky one, though, and when you combine walks and fly balls in the ninth, you're playing with fire.
As closer, David Aardsma's got a dangerous set of skills. He made it work in 2009, but I think it's fair to call into question the sustainability of his success. While Aardsma looks set to be a fine reliever again in 2010, his skillset and the spotlight on the closer role means we may never be too far from seeing Mark Lowe or Brandon League get promoted.