Drew Pomeranz and Beating BABIP

Drew Pomeranz is in the midst of a breakout season. He’s already surpassed his season high for innings and his ERA is a very low 2.47, while his FIP is a low — if not quite as low — 3.15. Those very good numbers netted the San Diego Padres a very good pitching prospect recently in the form of Anderson Espinoza.

Much of Pomeranz’s newfound success has been attributed to the addition of a cutter to his repertoire, which Jeff Sullivan detailed just before the trade last week. One notes, however, that the success is aided by a .240 BABIP and 80.8% left-on-base rate. Even if those numbers aren’t sustainable, the 3.15 FIP indicates Pomeranz’s success is real. But there’s reason to believe that Pomeranz isn’t as susceptible to regression as the average pitcher. Or there’s reason, at least, to believe that the Red Sox believe he isn’t.

Speaking with WEEI’s John Tomase, former major-league pitcher and current Red Sox assistant pitching coach Brian Bannister has indicated that Pomeranz’s cutter makes it more likely that he’ll sustain some of his batted-ball suppression in Boston.

From Tomase’s piece:

[Bannister] explained that like knuckleballers, whose BABIP numbers tend to skew low, pitchers who feature cutters tend to outperform league average on balls in play. He knows this because he did it over his first two years in the big leagues, posting BABIPs between .239 and .249.

“I was an example of it,” Bannister said. “[Cutters] generate a different batted-ball profile. There’s just different weak contact in there. Some guys it’s popups. Sometimes you get gyro-spin and it’s almost like a knuckleball. I mean, knuckleballers beat BABIP. It’s not always a given that a full regression is going to occur. When I look at a guy, if there’s a cutter involved or a knuckleball involved, you just can’t say for sure. I know a lot of people look at those two numbers — left on base percentage and the BABIP — and say, ‘Oh, he’s going to get worse in the second half.’ It’s not always a given.”

While we know pitchers tend to gravitate towards league average when it comes to BABIP, some pitchers are better than others at limiting hits on balls in play. Pop ups, like Bannister mentioned, can be a good way to induce easy outs. Fly balls and ground balls have different expected batting averages. Given a large enough sample size, we might be able to deduce which pitchers have these type of skills. With a smaller sample, perhaps looking at pitch types would help us determine which pitchers are likely to produce low BABIPs and thus more likely to outperfrom their fielding-independent numbers.

Regardless of whether Pomeranz’s cutter is a BABIP-beater, it’s a good pitch for the lefty. He generates whiffs on the pitch 16% of the time per Brooks Baseball. One quarter of the at bats ending on a cutter against Pomeranz have been strikeouts. The pitch likely helps his fastball and curve by providing an intermediary option of which hitters must remain aware. As far as the pop ups go, seven of the 35 balls in play against Pomeranz’s cutter this season have ended as pop ups. The pitch works for Pomeranz, but does it suppress BABIP?

Looking into 2016’s Statcast data at Baseball Savant, we can seek a preliminary answer. The simple chart below shows BABIP, batting average, and ISO on the cutter compared to the overall numbers this season:

2016 BABIP on Cutters
BABIP BA ISO
Overall .303 .256 .162
Cutter .296 .254 .144

In a sample of more than 5,000 cutters this season, there appears to be a seven-point difference between cutter BABIP and league BABIP. Batting average is going to include strikeouts and home runs, but the numbers are closer there while the ISO is much lower on the cutter. While a seven-point difference might not seem like a lot, it does lend some credence to the cutter’s possibility of reducing BABIP.

However, when we include other pitch types, we see we must go further to get better information. The chart below shows other pitches with more than 5,000 at-bats this season and their respective averages.

BABIP by Pitch in 2016
BABIP BA ISO
4-seam .314 .275 .189
2-seam .314 .291 .160
Overall .303 .256 .162
Cutter .296 .254 .144
Curve .289 .208 .135
Slider .287 .214 .141
Change .282 .242 .161

So while the cutter has a lower BABIP than average, so do the curve, slider, and change. While the cutter does seem to suppress BABIP a bit overall on its own, there doesn’t seem to be anything inherently special about the pitch at this point. That said, returning to Bannister’s point, he never suggests that the pitch itself is responsible for suppressing BABIP. After all, Pomeranz has thrown the pitch only around 12% of the time this season and around 20% of the time since he used it heavily. That isn’t going to be enough to have a dramatic effect on BABIP. Bannister’s argument seems to be that pitcher’s featuring the cutter tend to produce lower BABIPs.

To test Bannister’s theory, first I looked at qualified pitchers from last season, and I separated them into two groups. First, I took the qualified starting pitchers who threw the cutter at least 10% of the time, and separated them from the rest of the pitchers. Here is how the two groups performed last season.

Cutter Users v Non-Cutter Users in 2015
K/9 BB/9 HR/9 BABIP LOB% GB% HR/FB ERA FIP
Cutter Group (19) 7.5 2.5 1.00 .282 74.2% 43.7% 10.8% 3.61 3.85
Non-Cutter Group (58) 8.0 2.6 0.93 .296 74.0% 46.8% 11.0% 3.67 3.67

Look at that. The group featuring the cutters had a much lower BABIP. The LOB% was the same, but the lower BABIP was enough to help them beat their FIP by 0.24 while the other group posted an average BABIP and had an ERA matching the FIP. Now let’s try 2014:

Cutter Users v Non-Cutter Users in 2014
K/9 BB/9 HR/9 BABIP LOB% GB% HR/FB ERA FIP
Cutter Group (24) 7.3 2.3 0.87 .295 73.5% 44.7% 9.6% 3.58 3.65
Non-Cutter Group (63) 7.5 2.6 0.85 .289 74.1% 45.2% 9.4% 3.51 3.66

What was true for 2015 wasn’t necessarily true for 2014: the cutter group here actually allowed a higher BABIP. Nothing conclusive here in either direction.

Let’s try to use a larger sample. For the next set, I looked at the five-year period between 2011 and 2015 and looked at qualified starters who threw the pitch more than 10% of the time. I looked at 50 players who were throwing the cutter and 129 players who were not, and weighted the results by innings, as some players had as few as 300 innings while some had a bit more than 1,000 innings pitched. This is how the groups played out:

Cutter Users v Non-Cutter Users: 2011-2015
K/9 BB/9 HR/9 BABIP LOB% GB% HR/FB ERA FIP
Cutter Group 7.3 2.5 0.97 0.293 72.8% 45.0% 10.6% 3.81 3.81
Non-Cutter Group 7.4 2.8 0.96 0.293 73.1% 45.1% 10.5% 3.85 3.85

It is possible a deeper dive or looking at more specific characteristics of players might yield different results, but based on the information above it doesn’t appear that pitchers emphasizing the cutter have any special ability to suppress BABIP compared to the rest of their pitching brethren. It is possible that Pomeranz has a skill set that can reduce BABIP, but it would appear too early to draw that conclusion based on the above information.





Craig Edwards can be found on twitter @craigjedwards.

10 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Chicago Mark
7 years ago

Good stuff Craig. Could we also expect some BABIP regression due to park, league and division changes? Not to mention any possibility of fatigue.