Could A High BABIP Be A Sign of Good Pitching?

Back when Defense Independent Pitching stats were originally developed, BABIP was treated almost like a trash bin “other” category for plate appearances that didn’t end with one of the so called “Three True Outcomes.” Strike outs, Walks, and Home runs were observed to have far greater predictive power than other stats with regards to runs scored, and pitcher performance could be accurately estimated using only these three metrics, regardless of other factors. That’s the theory anyway, and it has evolved somewhat to account for new research that has popped up over the years. However, with the Statcast data now publicly available, we now have an unprecedented granular view of batted ball data, and many people, including myself, have developed various methods to apply exit velocity and launch angles to predict both offensive and pitching performance.

I’ve previously written about xOBA, xBABIP, scFIP, VH and PH, each of which aim to estimate end of season results given the (publicly available) Statcast data on each given batted ball. Note, this is all done on a batted ball by batted ball basis, then summed up at the end for each pitcher, team, or what have you. I’m not using average launch angles or average exit velocities to calculate these things.  In calculating and applying these stats I’ve noticed that while xOBA has pretty decent year to year predictive value (2 = .22) and excellent predictive value within a given year (r = .78), xBABIP does not.  Nor does standard BABIP, even though xBABIP is pretty good at predicting BABIP itself.  The two stats, xBABIP and xOBA, are calculated using nearly identical methods, adding just one single step at the end of xOBA, weighting the batted balls by the linear weights you can find here.  How could this be?  

The following is a group of pitchers who have a FIP less than or equal to 3.50, BABIP higher than .300, and a minimum of 80 IP in 2016. This filter gives us a group of pitchers who are having good to great seasons while maintaining an above average BABIP.

Looking through the list I see four names that stick out to me immediately: Noah Syndergaard, Jose Fernandez, Jacob deGrom, and Gerrit Cole. Each is a high quality ace pitcher. Especially Syndergaard and Fernandez, who are having excellent seasons. However, each of these guys is sporting a pretty high BABIP. Especially Fernandez. Look at that thing! .342 BABIP is pretty crazy for an ace pitcher, right? Something has to be wrong. You may have pulled up his player page thinking I’ve made an error or a typo. Nope! His BABIP is actually .342. And three of the guys on this list are Mets, a team with a notoriously weak defense, surely that must have something to do with it, right?

Here’s an image I created for my original post about xBABIP a few months ago. You’ll see areas with high BABIP are denoted by the red side of the spectrum, while areas with low BABIP are denoted by blue side.  I apologize to those of you with vision issues who may have trouble discerning the colors, I’ll explain it shortly.  

I like to think of this chart as having three tiers:

Tier 1a: The outfield areas down the lines. Defenders rarely if ever play the lines in the outfield, so any line drive or shallow fly ball that enters that area will likely land as a hit. High fly balls may or may not be caught, depending on the situation, giving those areas of the field about .700 BABIP, give or take.

Tier 1b: The area up the middle beyond the infield, but too shallow for an outfielder to catch. This area is about a .600 BABIP.

Tier 2a: The outfield areas just over the second basemen and short stop’s heads, but too shallow for the right and left fielders to catch. This area tends to have about a .400-.500 BABIP.

Tier 2b: The deep outfield gaps, the areas too far into the outfield to be caught by an average fielder. These tend to have more of a .300-.400 BABIP.

Tier 3a: The big blue regions in the outfield where the left, center, and right fielders play. These areas tend to have BABIPs around .100.

Tier 3b: The areas around the infielders, like with the outfielders these tend to have BABIPs around .100.

Looking at the chart, you can see the shadows of each fielder. The shallow part of the infield is full of weird artifacts which are caused by the way fielding locations are measured. Unfortunately I don’t have access to the more accurate data regarding the exact landing locations for balls, and I am forced to use the Gameday fielding locations instead. But, hey, lets not make excuses and roll with what we have.

One thing to keep in mind with this chart, though, the bluest, lowest BABIP areas represent average batted ball velocity. If you sit back and think about it a little, it should be obvious, but it may not leap out of the chart when you first see it, so I thought I’d bring it up. Fielders play where the ball is most likely to land. And, of course, that landing spot is decided in part by average exit velocity. Average launch angle, too. You can also factor in stuff like whether a batter is more likely to push or pull the ball, on and on, but averaged out over every batter in every stadium, the way it has been done for this chart here, it represents the average exit velocities and average launch angles.

So, what happens when you decrease the average exit velocity but hold everything else constant? Well, you would expect the balls to have a shorter flight time, and as a result they will not travel quite as far. You would expect balls that would have landed in Tier 3a to instead land in Tier 2a or Tier 1b, effectively adding 300-600 points to the BABIP for those particular batted balls. You’d also expect balls from Tier 2b to fall into tier 3a, dropping the BABIP of those balls by about 250 points, give or take. You’d see a few from Tier 1b to fall into 3b, too, which would reduce their BABIP by a few hundred points as well.

All in all, subtracting exit velocity, while keeping everything else constant could, in many cases, actually serve to increase the BABIP for a pitcher, by pulling balls closer to the infield, away from the fielders.

Sure, you could counter act this by placing your defenders a few steps in towards the infield when these pitchers are on the mound, but we’ve noticed over the years that outfielders tend to gravitate more towards comfort in the outfield as opposed to what, perhaps, might be a more ideal defensive alignment. It also tends to look pretty bad when you have an ‘easy catch’ sail over your head into the gap for a double or even a triple because you were playing shallow. Even if playing shallow registered a few extra easy outs during the course of that same game.

Not all pitchers or batters will reduce exit velocity in a vacuum, of course. If you change the launch angle as well you could start changing the landing locations in pretty significant ways, depending on whether you’re hitting the ball more vertically, more towards left field, or what have you. For example, if you hit the ball higher into the air you give fielders more time to catch it, so fly ball pitchers with weak contact tend to be difficult to hit for high BABIP, while those who give up an equal number of fly balls hit on a more shallow trajectory could see an increase in BABIP.