Measuring predictability in hitters

Honing the unpredictable can be a great trait in baseball. For a pitcher, becoming a master of sequencing can make up for shortcomings in velocity. For a baserunner, toying with the battery can increase the probability of a successful stolen base. For a hitter, distributing the baseball all over the field and avoiding certain batted ball trends can make it harder for a team to align a defense. Focusing on that, unpredictability leads to a greater chance of contact turning into a hit. This makes sense because if a hitter avoids trends, this will lower the likelihood of a shift, or at least an effective one for that matter.

When a shift has been put into place, the league-wide wOBA sits at .290 on balls in play, compared to the expected wOBA of .334. When a shift is not being utilized, the league hits for a .294 wOBA, compared to an expected wOBA of .325. The differential in actual and expected wOBA increases in the alignment of a shift. Hitters who are prone to the shift will tend to get more unlucky.

This can be backed up by comparing the differential in wOBA - xwOBA (on balls in play) with percentage of pitches that a hitter sees a shift in. There shows to be a decent correlation among qualified hitters this season (r = .25). The ten hitters with the highest rate of shifts underperformed their xwOBA by -.080 points. The ten hitters with the lowest rate of shifts underperformed their xwOBA by an average of -.042 points. The ten hitters at the top are almost twice as unlucky as the ten hitters at the bottom.

Baseball Savant

As I mentioned at the top, unpredictability is the clear goal to fighting this off. In the main spectrum of public data, we don’t have a way to measure a hitter’s unpredictability. Luckily, we have tracked data that we can compound together to fix this.

Looking at Baseball Info Solutions batted ball data (dating back to 2002) available on FanGraphs, you see three main sections. The type of batted ball (line drive, ground ball, fly ball), the field the batted ball was hit to (pulled, center, opposite), and the quality of the batted ball (soft, medium, hard).

To measure level of predictability, the section we’ll be looking at is the field each batted ball is hit to. But first, if you want an idea of what the ideal unpredictable hitter would look like, it’d be that he simply distributes the ball evenly to all three fields (33.3 percent pull, 33.3 percent center, 33.3 percent opposite). You can see where making a shift on this would be hard. The highest level of unpredictability a hitter could have would be hitting every batted ball to the same field. In this hypothetical scenario, the hitter would struggle to find a hit with the hypothetical seven fielders standing the direction his batted ball is heading.

To find the level of predictability in a hitter, we simply find the absolute value from 33.3 percent of each of the three distribution stats for the hitter. Adding all three of those absolute values up would get us a spray score (for lack of a better term). The lower the spray score is, the less predictable a hitter is.

To build up an idea, the lowest spray score in recorded history among qualified hitters is Derek Jeter, who pulled 32.8 percent of his batted balls, hit 34.1 percent to center, and knocked 33.1 percent to the opposite field. Without coincidence, Jeter ran a .342 BABIP from 2002 on. The highest spray score in that time belongs to Marcus Thames. His BABIP going back to 2002 was .275.

With a larger sample size of 874 hitters, there proves to be a good correlation between BABIP and spray score (r = .452).

Here are the 10 lowest spray scores, with the mentioned Jeter topping the list.

And then the 10 lowest...

With teams further advancement in shifting, being an unpredictable hitter has mattered more than it ever has. Of course, if the shift ever ends up being banned, none of this will ever matter again. But for now, hitters like Whit Merrifield, Elvis Andrus, and Nick Markakis will continue to find offensive value by possessing the trait of unpredictability at the plate.

Patrick Brennan loves to research pitchers and minor leaguers with data. You can find additional work of his at Royals Review and Royals Farm Report. You can also find him on Twitter @paintingcorner.