Analyzing relief pitchers is a vexing business in large part because of their small sample sizes. A limited amount of measurements -be it coin flips, dice rolls or batters faced- makes it more difficult to nail down the actual probability of outcomes. It is risky to draw any firm conclusions from the following numbers, but I find the possible trends interesting. Take a look at Aardsma's profile from when he joined Seattle in 2009 through the 2010 All Star Break and then afterwards:
2009 - '10 Break: 26% K, 10% BB, 27% GB, 85% fastballs, 11.1% SwStr
Post 2010 Break: 26% K, 15% BB, 42% GB, 67% fastballs, 11.3% SwStr
There are three major differences here. The ground balls are way up (small sample alert, just 48 balls in play), the walks increase by half and the fastball percentage plummeted.
We have mentioned the shift in fastball philosophy before -see here and here- but the vast jump in ground balls is new to me. While that's encouraging, there is also a troubling rise in walks. How much of a role the additional off speed pitches played in that spike is hard to tell precisely. While Aardsma is more capable of throwing fastballs in the zone, the percentage of pitches that were called a ball actually dropped in the second half of 2010. There is also a good chance that sample sizes are exacerbating the issue since the changes in how often Aardsma threw outside the strike zone or for a ball do not seem to merit such a dramatic increase in walks.
Blanket statements such as "walks bad" and "ground balls good" are fine for a sound bite, but fail to give us a complete picture of the trade off. How do the two above pitching lines compare? How much of a higher ground ball rate does it take to offset a higher walk rate?
Using xFIP, the answer is mixed depending on whether you use ground ball rate or fly ball rate to estimate the expected number of home runs allowed. Looking at both gives us a range of guesses, which is better. By ground ball rate, Aardsma's xFIP is 0.23 runs higher post All Star Break and by fly ball rate, it is 0.01 runs lower. I would lean more toward using fly ball rate anyways from a philosophical standpoint, but it looks like no, there are just too many additional walks. However, it is critically worth noting that dropping the walk rate down to 14% makes it an even split so we're dealing with a close call.
Pitchers are constantly changing and evolving so with these small samples and an entire off season and surgery in between, I don't want to draw any conclusions about Aardsma's effectiveness going forward in 2011.
What I can say is that is that Aardsma has always been one skill away from being a truly excellent reliever. He could not quite get the strikeouts high, the walks low and the fly balls under control. This late season change gives me hope that he might be on his way to getting there. There are a couple ways, such as improved control from more reps or even simple regression, that he could lower his walk rate without sacrificing the gains in the other two rates. Aardsma might be on the verge of a step forward as a pitcher. He's had two out of three key skills for a while now, bringing in all three could transform him from above average to legitimately great.