More Graph Fun: Anatomy of a ground ball

Ever wondered what a pitcher can do to induce a ground ball? You may already know much of what is to come, but let's investigate how ground ball percentage varies depending on pitch location, movement, velocity and the handedness of both the batter and the pitcher.

First, let's start by looking at the strikezone. You will see a graph of pitch location, with the heat level indicating the percentage of ground balls hit per all pitches in play for a given area. An excellent GB% is about 50% and higher, indicated by yellowish green to red. A poor GB% is about 40% and lower, indicated by teal to dark blue.

Click on the images for a larger view. I think you can take a lot out of these graphs by looking at the general idea -- the thumbnails should be sufficient as long as you remember which graph is which. I have included platoon splits for completeness, and for the sake of consistency will have the pictures ordered from left to right the same in each topic:

To no one's surprise, pitches lower in the strikezone to just below the strikezone tend to be hit for the most ground balls. Balls toward the upper middle of the strikezone, and especially for RHB up and inside, there is a very low probability that the pitch will result in a ground ball. Low and away, just outside the strikezone, appears to be consistently the best place to attempt to locate a pitch in order to induce a ground ball.

Now, let's look at pitch movement. I have broken down pitches into handedness splits again, labeled the same left to right as before. I have also broken down by pitch speed. I attempted to use MLB's default pitch types, but there were too many errors for it to be taken seriously (it labels 91 mph cutters as curveballs for David Robertson, for example).

The first four graphs on this row are all pitches thrown between 90-100 mph. We're looking at fastballs, sinkers, very very fast change-ups and very fast sliders. I have indicated on each graph the general area for where we would classify each pitch type. The color scale has also been altered slightly, anything greenish to red is considered to be good for getting grounders and anything a little darker than teal is considered bad for getting grounders.

What exactly do we see? For all at-bats, a pitch with less rise and more sink will generally have a greater chance of ending up as a ground ball. In same-handed situations (LHP vs. LHB or RHP vs. RHB), a pitch with a lot of arm-side run will have a good chance of becoming a grounder. Once a pitch reaches a certain threshold, say about 10 inches of break toward the batter, it will have a 60% of resulting in a ground ball. This makes intuitive sense since sinkers and two-seam fastballs are most effective in same-handed at-bats.

Now let's look at all pitches from 80-90 mph -- these now include slow fastballs/sinkers (i.e. Jamie Moyer), average speed changeups, most sliders, and curveballs. There is a real diversity of pitch types here.

Just like before, it becomes obvious that the more sink and the less rise on a pitch, the more likely it will be hit on the ground. We still see the strong arm-side run splits for same-handed at-bats that we did before. It is interesting to see that sliders, thrown from 80-90 mph (as most are), tend not to induce ground balls whereas sinkers/change-ups will. At at this speed range, a pitch must have either some sink or some run for it to induce a ground ball.

Here are pitches from 70-80 mph. We really only have Tim Wakefield's fastball, change-ups, a few sliders and many curveballs. Also, there might be some surprise data points for RHP due to Wakefield's knuckleball.

We find less of a split between batter type -- here it is mostly about the handedness of the pitcher. Southpaws with slow dropping curveballs or slower change-ups with good sink will get the job done in this range. Righties are much the same. In the RHP vs. RHB graph, the spot of red and yellow may be due to right-handed specialists who typically submarine and sink the ball (think Chad Bradford) or perhaps the pesky knuckler. Sink plus run is a good thing. We get the same sort of patch for lefties albeit slightly less well-defined.

Last, but not least, here we have pitches ranging from 60-70 mph. All that is left are really slow change-ups, curveballs and sliders.

At this speed, it becomes very difficult to induce a ground ball. These pitches really float to the batter, so perhaps they hang so much that they just become popups or fly balls. Only curveballs to righties from left-handed pitchers seem to have consistent success when the ball is pitched this slowly.

In summary...

None of this analysis is particularly surprising nor groundbreaking. It does, however, validate much conventional baseball wisdom in the process. Pitches with more sink and more arm-side run in same-handed at-bats have always been considered to be the way to induce a grounder. An increased velocity appears to slightly induce more grounders (the green on the graphs turns teal as you go down the page). This should be cautioned though by stipulating that balls thrown harder tend to naturally gain rise at the same time -- so it's not just a matter of throwing harder. It's doing that while maintaining good ground ball-friendly movement that may increase success.

Further, pitches located down in the zone are more likely to be hit closer to the ground. Low and away is truly a sweet-spot for maximizing GB%.

The only thing that I did not expect to see was that grounders are hit less often on the inside part of the plate in opposite-hand ABs compared with same-handed ABs. Perhaps this goes back to the effectiveness of something like two-seem fastballs, which are deadly on the inside part of the plate, but only in those same-handed at-bats.