Happy New Year! Stat Head is back after a week off to talk about some more not-so-obvious stats that can help you improve your player evaluations and help you win a fantasy baseball championship. Today we’re going to talk about line drive rates for pitchers.

When we talk about batted ball types, we are referring to ground balls, outfield fly balls, infield fly balls (aka pop-ups) and line drives. You’ll occasionally hear about fliners (the type of balls that are somewhere between and fly ball and a line drive), but these are the primary four. Of these four, line drives fall for hits at the highest rate. Here is a table breaking down the hit percentages on these batted ball types over the past four years.


And here is a table breaking down the BABIP of each batted ball type for the past four years.


As is plain to see, if all a batter is trying to do is get a hit, a line drive is what he should be hoping for. And since pitchers don’t like to give up hits, it should also be pretty obvious that they want to allow as few line drives as possible.

The problem is, pitchers don’t actually have much control over their line drive rates. As with most of the other stats we looked at thus far, pitchers who deviate too far from the league average tend to regress going forward. Here are the league average rates for the past four years.

These are calculated by dividing line drives by the total number of line drives, fly balls, and ground balls. This formula – which will be used throughout this article when we refer to line drive rate or line drive percentage – excludes events like bunts that don’t say much about a pitcher’s skill level.


Let’s use regressions analysis to see just how much control pitchers have over their line drive rates. In this analysis, we’ll examine a pitcher’s line drive rate from consecutive years. If there is a strong relationship between the two years, we can assume that line drive prevention (or lack thereof) is a repeatable skill. If not, we can see just how much external noise influences a pitcher’s line drive rate. We’ll use data from 2004-2007 and include all pitchers with at least 75 innings pitched in both Year 1 and Year 2.

Correlation Coefficient: 0.15
R2: 0.02
Adjusted R2: 0.02
P-value: 0.003
Level of Significance: 1%

What this shows us is that there is a weak relationship between a pitcher’s line drive rate from one year to the next and that only 2% of its movement can be predicted using his corresponding statistic from the previous year. It is still significant, though, showing that pitchers can control it to a degree.

Because it isn’t easily predicted, though, we can say that much of a pitcher’s line drive rate is composed of “luck” or, as we called it two weeks ago, unexplained variation. Because of this, it is best to assume that a pitcher should put up a line drive rate near league average, and that those who put up rates at the extremes are getting either lucky or unlucky.

I should note that the above analysis is a bit misleading. While the majority of pitchers can’t control their line drive rates, there are a couple of subsets who have just a bit more control over it than the rest. Pitchers with extreme ground ball or fly ball rates are able to prevent line drives a little better than pitchers with non-extreme GB/FB tendencies. Below are the charts illustrating this:


You can see that as a pitcher’s ground ball rate or fly ball rate rises to an extreme level (which is something pitchers do have a lot of control over), his line drive rate subsequently falls. Not dramatically, but significantly. This shows that we need to evaluate these types of pitchers on a slightly different plane, though the general concept still applies.

This means that if a pitcher with a 55% ground ball rate posts a 12% line drive rate, we should still expect regression. Even though pitchers with good ground ball rates can prevent line drives to an extent – as is the case with normal pitchers – if the line drive rate is too far from the appropriate baseline, there will be regression.

Let’s check out a few prime candidates for regression in 2008.


As is always the case, make sure you check out a pitcher’s complete skill set and his other ‘luck indicators’ before making a decision. None of the stats we’ve looked at thus far should ever be used alone. Isolated, they have a limited amount of predictive power, but when we look at the overall picture that these indicators paint we can get a good feel for the overall direction a player’s surface stats will take.