It’s the bottom of the eighth inning. Men are on first and third base, there’s one out and your team is down by one run. The opposing team has one of the best ground-ball pitchers on the hill, and the infield is playing back and is looking for a double play. All you need is a fly ball to tie the game and significantly swing your chances of winning.
So who do you want at the plate?
It’s likely that the opposing manager will either bring in a ground-ball specialist or just tell the pitcher to stay away from pitches that could be hit in the air to the outfield. Knowing who you’d want to hit requires an understanding of what pitches are the most likely to induce a ground ball — and what hitters manage to hit fly balls against those pitches most often.
Josh Weinstock examined the characteristics of ground-ball pitches in 2011. Generally speaking, ground balls are heavily dependent on their vertical location and vertical break. In terms of vertical location, ground-ball rates increase the lower the pitch is thrown. Pitches lower than two feet above the plate induce ground balls at a rate more than 50%. Josh also found that the more sink a pitch has, the more likely it induces a ground ball. But there was a whole bunch of overlap between sink and vertical location.
I determined what pitches to categorize as ground-ball-likely by comparing the fly-ball to ground-ball (FB/GB) ratios of pitches that where lower than two feet and had vertical break (pfx_z in the PITCHf/x database) between 0 and 2.5 (since Josh showed ground-ball rates cresting around 60% in that range) to those pitches that did not (2008 to 2012). Similarly, I compared pitches that were lower than two feet from the plate and those that were higher, without regard for sink.
|pz<2 AND pfx_z 0-2.5||20,880||61%||22%||18%||17%||4%||0.30||0.37||0.28|
|Non-pz<2 AND pfx_z 0-2.5||635,513||44%||46%||29%||19%||8%||0.66||0.84||0.56|
The chart above includes traditional ground-ball, fly-ball and line-drive percentages, but it also includes pop up percent (PU%). Additionally, I wanted to isolate “true” fly-ball rate from pop ups, so FB(noPU)% is simply the number of non-pop up fly balls, divided by all balls in play. Finally, I calculated the ratio of true fly balls to ground balls (FB(noPU)/GB) as well as the ratio of true fly balls to ground balls, plus pop ups (FB/(GB+PU)). The logic of FB/(GB+PU) is that I wanted to control for hitters who didn’t just hit fly balls but who also hit a large number of pop ups (since pop ups are the worst kind of ball a batter can hit in play).
As Josh’s work suggested, you don’t pick up much by including sink. Fly-ball percentages decrease by only two percentage points (20%, versus 18%). Additionally, the sample sizes are much larger without the sink (~150,000, versus ~21,000), so I’ll just stick with pitches lower than two feet.
Now, in these situations, opposing managers are likely to instruct pitchers to keep the ball down to try and induce a ground ball — or they can simply bring in a ground-ball specialist. In either case, a hitter is likely to see a higher percentage of ground-ball-likely pitches in these situations. Because of that, I calculated the batted-ball distributions for all hitters since 2008 for pitches less than two feet off the ground.
Here are the top-25 hitters, since 2008, at hitting fly balls on ground-ball pitches. This sorted by FB/(GB+PU), relative to league average:
Can’t beat Bengie Molina or Rod Barajas when it comes to increasing your odds of a fly ball. Molina and Barajas have been 131% and 109% better, respectively, than league average when it comes to the ratio of fly balls to ground balls or pop ups on ground-ball pitches. Hitters such as Alfonso Soriano, Justin Morneau, Colby Rasmus and Carlos Quentin have also been quite good at turning ground-ball pitches into non-pop-up fly balls since 2008.
Looking over the list, we do see a fair number of fly-ball-hitting power hitters. Does that mean that the ability to hit ground-ball pitches in the air is simply a function of being a fly-ball hitter?
Well, sort of.
I correlated a hitter’s FB% and FB(noPU)% for ground-ball pitches with their overall percentages and found correlations of .62 and .61, respectively. Those are decent correlations — and I mean that almost 40% of the variation can be explained simply by a hitters overall FB%. But there is still a large percentage left unexplained. It may be that what makes up the gap — as our own Bradley Woodrum has hypothesized — is a hitter’s swing-plane.
So there you have it. The (or, at least, an) answer to the question: Who do you want up with a chance to drive in the tying or go-ahead run from third with a fly-ball when there’s a risk of a double play and the opposing manager has his best ground-ball pitcher on the mound.
Bill leads Predictive Modeling and Data Science consulting at Gallup. In his free time, he writes for The Hardball Times, speaks about baseball research and analytics, has consulted for a Major League Baseball team, and has appeared on MLB Network's Clubhouse Confidential as well as several MLB-produced documentaries. He is also the creator of the baseballr package for the R programming language. Along with Jeff Zimmerman, he won the 2013 SABR Analytics Research Award for Contemporary Analysis. Follow him on Twitter @BillPetti.