Generating Weak Contact: Bringing It All Together

by
September 23, 2021

After more than two days of pouring over pitcher batted ball data, I better write something before I am fired. In the end, I found nothing groundbreaking. Popups and groundballs good… everything else (i.e. line drives and flyballs) is bad. The change I’d recommend going forward is to move to a more granular grading of batted to bins, like barrels, based on outcomes.

I started the analysis with five core ideas on how pitchers can generate weak contact.

In the end, I reverified that groundballs and popups are desirable outcomes. Additionally, I found that generate chase contact leads to weak contact… that could be a groundball or popup. Finally, the horizontal movement lead to some weak contact but most of its results were probably accounted for with the popups and groundballs.

From all the work, I’ve found it’s nearly impossible to constantly generate desirable batted results. With popups, the conditions that create them also generate home runs (i.e. high HR/BIP values). With downward hit groundballs, some will be hit with a little zip and go for hits (i.e. higher BABIP). Here is a comparison of our groundballs and popup (FB%*IFFB%) to the two new metrics (min 20 IP).

While GB% is a decent proxy for the new value (r-squared of .74), the popup value is not even close. If there is one change I’d like it’s the high flyball grouping becoming a standard available value.

Using simple linear regression, I compared the four values (High%, Low%, O-Contact Rate, Horizontal movement) to the pitcher batted ball metrics (vsISO, BABIP, HR/BIP, HR/FB%, ERA-xFIP, and added ERA-Steamer ERA).

Here are the r-squared values with all for factors for pitchers and those with 200 and 500 balls in play.

R-squares for Pitcher Batted Ball Factors
Batted Ball Metric All BIP >= 200 BIP >= 500 BIP
BABIP 0.10 0.11 0.15
HR/FB 0.01 0.01 0.03
HR/BIP 0.14 0.25 0.28
ERA-xFIP 0.05 0.07 0.13
ISO 0.12 0.19 0.20
ERA-sERA 0.08 0.09 0.08

And here is the correlation with the non-significant factors removed.

R-squares for Remaining Pitcher Batted Ball Factors
Batted Ball Metric All BIP >= 200 BIP >= 500 BIP Factors Kept
BABIP 0.10 0.11 0.15 Low, High
HR/FB 0.00 0.00 0.02 Low
HR/BIP 0.14 0.25 0.27 Low, High
ERA-FIP 0.05 0.07 0.13 Low, High, O-Contact
ISO 0.10 0.18 0.19 Low
ERA-sERA 0.08 0.09 0.08 Low, High, O-Contact

First, I acknowledge that horizontal pitch variance is a small factor, its results are seen in the groundball and flyball numbers. As for the out-of-zone swing and contact rate, it’s a smaller factor (and could lead to popups and groundballs) and I’m having problems easily incorporating it into a usable variable. I don’t have an answer yet but maybe someday one will materialize. In the meantime, know that if a pitcher gets hitters to chase out of the zone for contact, the pitcher may have better than expected batted ball metrics.

While some of the factors are significant, the tug in different directions with the “popup” group leading to more HR/BIP while the “groundball” grouping leading to fewer. The back and forth pull is a major factor in getting to the bottom of any good answer. At times, it may seem that getting to an answer shouldn’t matter, but being on either side is better than the middle where home runs and line drives live.

For one final comparison, I just added the two new batted ball rates together, grouped them into 5% point intervals, and here are the median results.

Batted Ball Factors for Combined Weak Contact
Interval BABIP HR/FB HR/BIP ISO ERA-Projected ERA
>= 50% .292 12.6% 4.4% .145 -0.29
45% to 50% .291 13.9% 4.9% .150 0.06
40% to 45% .295 12.9% 4.8% .152 0.00
< 40% .296 12.9% 6.6% .150 0.08

The one key is that while there may be traits that limit hard contact, there is so much noise (e.g. home parks and defense) that the differences are barely noticeable.

finally, for reference, here are the 2021 leaders and laggards

2021 Leaders & Laggards in Generating Weak Contact
Name IP Low% High% Combo BABIP HR/FB HR/BIP vsISO ERA WHIP
Framber Valdez 123 52.9% 6.7% 59.6% .277 19.6% 2.9% .105 3.07 1.27
Kyle Gibson 171 40.6% 15.3% 56.0% .279 10.6% 3.0% .121 3.51 1.23
Corbin Burnes 158 32.6% 23.3% 55.9% .308 5.5% 1.7% .069 2.34 0.94
Logan Webb 132 41.4% 13.4% 54.8% .303 13.8% 2.6% .098 2.79 1.10
Vladimir Gutierrez 111 22.6% 31.8% 54.4% .274 14.6% 5.7% .209 4.53 1.37
Sandy Alcantara 194 32.8% 21.5% 54.3% .265 12.8% 3.5% .133 3.05 1.06
Pablo Lopez 101 30.7% 23.5% 54.2% .293 12.3% 3.8% .144 3.03 1.09
Adrian Houser 131 37.1% 17.1% 54.1% .261 14.5% 3.1% .114 3.43 1.29
Nick Pivetta 144 21.4% 32.4% 53.8% .290 13.8% 5.9% .187 4.63 1.34
Walker Buehler 195 24.2% 29.5% 53.6% .247 10.4% 3.6% .132 2.58 0.98
Trevor Bauer 107 21.3% 32.3% 53.5% .220 16.0% 7.5% .192 2.59 1.00
Charlie Morton 170 30.6% 22.8% 53.5% .267 13.1% 3.7% .114 3.49 1.06
Lance McCullers Jr. 150 34.3% 19.1% 53.4% .284 11.5% 3.1% .120 3.11 1.25
Clayton Kershaw 115 32.7% 20.6% 53.3% .281 13.5% 4.4% .142 3.27 0.97
Gerrit Cole 169 24.0% 29.3% 53.2% .303 12.7% 5.2% .137 3.03 1.03
Josh Fleming 100 39.0% 14.1% 53.1% .292 13.1% 3.3% .144 5.01 1.35
Anthony DeSclafani 158 27.2% 25.8% 53.0% .262 11.7% 4.2% .133 3.23 1.09
Luis Garcia 150 20.1% 32.9% 52.9% .283 10.6% 4.3% .148 3.23 1.15
Jordan Montgomery 149 28.1% 24.8% 52.9% .301 10.7% 3.8% .120 3.55 1.24
Adbert Alzolay 120 26.6% 26.3% 52.9% .270 22.7% 7.4% .213 4.80 1.18
Median 30.7% 23.4% 53.6% .280 12.9% 3.7% .133 3.23 1.12
Martin Perez 110 28.1% 20.1% 48.1% .338 16.4% 5.1% .185 4.83 1.52
Julio Urias 174 19.5% 28.5% 48.0% .277 10.0% 4.0% .135 3.10 1.03
Paolo Espino 102 20.8% 27.1% 48.0% .282 12.4% 5.2% .191 3.94 1.20
Lucas Giolito 167 19.0% 28.9% 47.9% .273 15.2% 6.2% .185 3.70 1.12
Freddy Peralta 139 17.6% 30.2% 47.8% .228 9.2% 4.3% .130 2.65 0.96
Steven Matz 140 25.9% 21.9% 47.8% .313 12.1% 4.0% .142 3.84 1.32
Alec Mills 110 27.9% 19.8% 47.7% .325 14.0% 3.9% .156 4.83 1.42
Michael Wacha 114 20.7% 26.9% 47.6% .329 18.7% 6.6% .209 5.49 1.39
Jorge Lopez 121 30.4% 17.0% 47.3% .340 20.0% 5.6% .192 6.07 1.63
J.A. Happ 142 17.4% 29.8% 47.1% .310 13.7% 6.2% .230 6.02 1.48
Adam Wainwright 196 28.1% 18.2% 46.3% .256 10.7% 3.2% .127 2.89 1.03
Cole Irvin 169 17.6% 28.3% 45.9% .301 9.4% 3.5% .141 3.99 1.31
Rich Hill 148 19.1% 26.6% 45.7% .276 11.6% 4.8% .172 3.87 1.19
Chris Bassitt 151 21.4% 24.3% 45.6% .273 9.9% 3.6% .132 3.22 1.05
Eduardo Rodriguez 146 20.4% 25.1% 45.5% .359 13.2% 4.6% .171 4.97 1.39
Chris Paddack 108 20.8% 24.4% 45.2% .311 12.9% 4.5% .175 5.07 1.26
Ryan Yarbrough 144 18.0% 27.2% 45.2% .298 13.3% 5.1% .194 5.30 1.27
Matt Harvey 127 21.7% 23.4% 45.1% .331 12.3% 4.3% .183 6.27 1.54
Antonio Senzatela 150 26.3% 18.7% 45.0% .316 9.3% 2.4% .135 4.14 1.29
Joe Ross 108 21.7% 23.0% 44.6% .277 15.5% 5.5% .164 4.17 1.22
Median 20.8% 24.7% 46.7% .306 12.7% 4.5% .172 4.15 1.28

While some names seem out of place in the “good” (e.g. Pivetta and Gutierrez) and “bad” (e.g. Wainwright and Bassitt), the top group is outperforming the bottom group in the three most important variables, BABIP, HR/BIP, and vsISO.

As I stated early on in the article, I’m still working through the information. The next step is to see if the weak contact is not being fully incorporated into projections. While I rarely push for new stats (I think many should disappear), will try to see if the Dark Overlord will make the Weak%, or at least a new popup group, available here on the FanGraphs.





 

Jeff, one of the authors of the fantasy baseball guide,The Process, writes for RotoGraphs, The Hardball Times, Rotowire, Baseball America, and BaseballHQ. He has been nominated for two SABR Analytics Research Award for Contemporary Analysis and won it in 2013 in tandem with Bill Petti. He has won four FSWA Awards including on for his Mining the News series. He's won Tout Wars three times, LABR twice, and got his first NFBC Main Event win in 2021. Follow him on Twitter @jeffwzimmerman.

Comments are closed.