Strategy room: Stats say Romero to keep rolling, Zito to falter in 2010
Last week, we took an in-depth look at the metric of batting average on balls in play, or BABIP. This week, we will apply some of those principles and findings to analyze some starting pitchers off to unexpected starts, both good and bad. The concept of expected BABIP will be introduced and compared to the hurler's actual BABIP, in an attempt to gauge how lucky or unlucky the starter has been.
But first, here is the explanation of expected BABIP. As was discussed last week, each type of batted ball has a probability of becoming a hit. With some routine algebra, it is possible to use these probabilities along with the pitcher's actual hit distribution, their homers per fly ball and their strikeout rate to come up with what their BABIP should be. The calculation can be modified to account for outs derived from double plays, caught stealing, pickoffs, sacrifices, etc., basically outs recorded without a corresponding at bat. However, there is no consideration of the defense backing the pitcher. That said, this process is not intended to be presented at a SABR convention, but simply to aid in evaluating some pitching performances.
If a pitcher's expected BABIP exceeds their actual BABIP, an argument can be made they have been lucky, and vice versa. The key, however, is all this does is help put in perspective what has already occurred. The inputs of hit distribution and HR/FB will undoubtedly change. How you feel those may change is a better indicator of future performance.
• The first pitcher we will examine is
Thus far, he is sporting a disappointing ERA of 6.02 with a WHIP of 1.489. Fans of some advanced sabermetrics will suggest his xFIP of 3.61 and tERA of 5.63 suggest Harang has been unlucky. These metrics focus on the skill elements of pitcher's performance, normalizing some of the luck elements to compute what should be an expected ERA. Since these both result in a number less than Harang's actual ERA, the difference is chalked up to some misfortune. Using the above described expected BABIP formula, Harang's BABIP should be about .333, but presently it sits 25 points higher at .358. What this suggests is at least part of the reason Harang's expected ERA is lower than his real one is that based on his present distribution of batted balls, he has allowed more hits that probability dictates.
While this is promising for Harang's owners, even if his BABIP corrects to .333, he is still going to struggle, as that is a whopping 30 points higher than league average. The reason his expected BABIP is so high is he is carrying a very high line drive percentage of 23.3. Unfortunately for Harang's owners, this is not too far off from his career mark, meaning he usually sports a BABIP higher than league average. In fact, his career mark is .317, over 17 points higher than the .300 league average discussed last week. Further complicating matters is Harang usually allows more fly balls than grounders but in the small sample of 2010, this is reversed. These additional grounders are further raising his BABIP. But since Harang is allowing 18.9 percent of his fly balls to leave the yard, well above the 11 percent league average, these additional grounders have helped limit the number of home runs.
As is often the case in small samples, Harang's numbers are all over the place. And while our new expected BABIP suggests Harang has been unfortunate, there are enough caution signs on to refrain from recommending him as a buy low candidate even with the excellent strikeout and walk peripherals. For example, it is a bit disturbing that he historically carries a high line drive percent. This alone is enough to temper expectations. If a different pitcher were carrying such a high rate, the contention would be that it will regress closer to the league average of about 18 percent, but not with Harang.
On the flip side, however, is the hope he continues inducing more grounders than normal while his HR/FB regresses toward his usual 12 percent or so. If this happens, the fact his fan rate is high and walk rate is low will result in a big positive ERA correction. But that is a lot of what-ifs to count on.
To sum up Harang's expectations for the rest of the 2010 campaign, he is not a guy to target as a buy-low candidate, despite his stellar strikeout and walk rates as there is too much else going on to be able to expect considerable better results. However, there is enough there to warrant stashing him away if you can acquire him cheaply, most notable a higher ground ball rate which should result in fewer home runs.
OK, now it is time to let you in on a little secret. It does not take a PhD in common sense to realize Harang is a risky proposition, despite excellent strikeout and walk rates. While the theory behind expected BABIP might be elegant, it is merely a piece of the puzzle, and a rather small one at that. Pardon the elaborate set-up, but the intended lesson here is that no matter how complicated or telling, no single metric can be used by its lonesome. It is best to use a bunch of different metrics in conjunction with each other to best paint the entire picture. With that in mind, let's take a brief look at some other early season pitching performances.
• One of the American League's surprises has been Toronto's
Some may consider Romero to be a sell high candidate. If you have such an owner in your league, get in on that action. Let us assume some regression and see how that will affect Romero's numbers using the expected BABIP tool. How about we drop his K/9 to 8.0 and normalize his HR/FB to 11 percent. We will leave his BB/9 and hit distribution as is, which is reasonable based on his history. His expected performance is still quite impressive. His BABIP should approach .310 and his HR/9 should rise to .8, but in conjunction with a still excellent K-rate, the increase in ERA should only be to the mid 3s.
• One of the American League's biggest disappointments is Boston's
• The final subject of today's analysis will be the rejuvenated