Which Statcast Measures Correlate Best with Pitcher HR/FB and BABIP?

Note: As was the case in a previous analysis of Statcast measures and their correlation with power metrics for hitters, I owe a debt of gratitude to Alex Chamberlain. He did a lot of heavy lifting for this column, running the correlations and discussing interpretations with me.

It won’t be the first or last time, but I did a silly thing on Twitter. In announcing a pick for the Pitcher List Experts Mock, I decided to tout the player I chose by citing one of his achievements, as captured by a Statcast metric.

(Justin, by the way, made his pick very promptly.)

Though I had yet to analyze how well Statcast measures correlate with pitcher outcomes, I made the assumption that Trevor Williams’ ability to limit average exit velocity on flyballs and line drives in 2017 and 2018 was a meaningful harbinger of success to come in 2019. My thinking was that his extreme success in squelching exit velocity was not only linked to his low HR/FB (8.0 percent, both at home and on the road) and BABIP (.261) from this season, but to his ability to put up similar numbers next season.

That wasn’t entirely right. The Pearson’s r correlation between average exit velocity on flies and liners and BABIP is just .14, which is not statistically significant (p < .05). As the table below shows, I would have been best off looking at Williams’ average launch angle and at the percentage of batted balls hit off him that had an exit velocity of at least 95 mph. As it turns out, he was just outside the top third in average launch angle and among the 10 percent of pitchers (min. 400 BBE) that had the lowest percentage of 95-plus mph batted balls in 2018. There were good reasons for Williams to be better than average at preventing hits on balls in play after all.

Correlations Between Statcast Metrics and BABIP for Single Seasons, 2015-2018
Measure BABIP
95 MPH% 0.27
GB EV 0.21
Avg FB Dist 0.20
Overall Avg EV 0.20
FB/LD EV 0.14
Avg HR Dist 0.10
Max EV 0.07
Max Dist 0.04
Brls/PA % -0.07
Brls/BBE % -0.12
Overall Avg Dist -0.22
LA (°) -0.32
Minimum 400 BBE, n = 567, except for Avg FB Dist (min 80 flyballs, n = 309). Correlations that are statistically significant (p < .05) are highlighted in yellow.

Better yet, Williams’ ultra-low average exit velocity on flies and liners did bode well for a low HR/FB. That, along with average flyball distance and barrel rate (per batted ball and per plate appearance), registered a Pearson’s r coeffcient greater than 0.40, easily clearing the bar for statistical significance.

Correlations Between Statcast Metrics and HR/FB for Single Seasons, 2015-2018
Measure HR/FB
Avg FB Dist 0.55
Brls/BBE % 0.42
Brls/PA % 0.42
FB/LD EV 0.41
95 MPH% 0.37
Max Dist 0.27
Overall Avg EV 0.23
Max EV 0.12
Avg HR Dist 0.10
GB EV 0.07
Overall Avg Dist -0.09
LA (°) -0.23
Minimum 400 BBE, n = 567, except for Avg FB Dist (min 80 flyballs, n = 309). Correlations that are statistically significant are highlighted in yellow.

None of these findings say anything about whether these Statcast measures are predictive of future ability to limit home runs on flyballs or base hits on balls in play. After running year-to-year correlations going back to the first year of Statcast data in 2015, it is clear that launch angle is the most reliable of these indicators from one season to the next. Among the Statcast measures, it has the strongest correlation with BABIP (and as one would expect, it’s a negative relationship), so it is our best predictor of a pitcher’s future BABIP. Current-year BABIP is a poor predictor of next year’s BABIP, so we should rely on launch angle to decide whether we should expect a pitcher to be an over- or under-performer on balls in play for the coming season.

Year-to-Year Correlations For Selected Pitcher Stats, 2015-2018
YEAR TO YEAR
LA (°) 0.80
95 MPH% 0.36
FB/LD EV 0.32
Brls/PA % 0.31
Brls/BBE % 0.29
GB EV 0.26
Avg FB Dist 0.24
BABIP 0.15
HR/FB 0.06
SOURCE: FanGraphs and Statcast.
Correlations that are statistically significant are highlighted in yellow.

HR/FB has an even lower year-to-year correlation than BABIP does. Fortunately, the Statcast measures that are strong single-season correlates of HR/FB — most notably, average flyball distance, barrel rates and average exit velocity on flies and liners — have reasonably strong year-to-year correlations. While average flyball distance is the strongest of these correlates for HR/FB in a single season, it does have the weakest year-to-year correlation of any of the Statcast measures included in the table. As noted above for the Trevor Williams example, we can confidently rely on exit velocity on flies and liners to predict future HR/FB ratios for pitchers.

It should be noted that if we have multiple years of favorable average flyball distance data, we can be fairly certain that the pitcher in question has had the skill of limiting home runs on flyballs over the longer haul. Of all of the Statcast measures included here, average flyball distance has the strongest correlation with HR/FB over the last four seasons. For the same time period, average exit velocity on ground balls was the strongest correlate for BABIP — even stronger than launch angle and percentage of batted balls with an exit velocity of at least 95 mph.

Correlations Between Statcast Metrics and BABIP and HR/FB, Cumulatively for 2015-2018
Measure BABIP HR/FB
Avg FB Dist 0.27 0.50
FB/LD EV 0.20 0.42
95 MPH% 0.43 0.39
Overall Avg EV 0.33 0.32
GB EV 0.46 0.29
Brls/PA % -0.06 0.24
Max EV 0.22 0.21
Brls/BBE % -0.12 0.19
Max Dist 0.07 0.04
Avg HR Dist 0.29 -0.07
Overall Avg Dist -0.30 -0.28
LA (°) -0.39 -0.38
Minimum 1500 BBE, n = 71. Correlations that are statistically significant are highlighted in yellow.

With this knowledge in hand, the 2018 Statcast leaderboards offer some insight as to which pitchers are likely to help or hurt themselves as a result of extreme BABIP rates and HR/FB ratios in 2019. Marco Estrada has long been the poster child for high launch angle and low BABIPs, and he was the launch angle leader this season (min. 400 BBE). Right behind him was Matthew Boyd, whose .258 BABIP and 1.16 WHIP may not regress much. Clayton Richard and Dallas Keuchel had the lowest launch angles, yet they escaped with BABIPs of .289 and .300, respectively. Both could record higher rates in 2019.

The Mets’ Big Three of Noah Syndergaard, Zack Wheeler and Jacob deGrom recorded the lowest average exit velocities on flies and liners, so all three could be primed to maintain the sub-9.0 percent HR/FB ratios they posted in 2018. Right behind them were Ryan Yarbrough, Miles Mikolas and Kyle Freeland, all of whom complied lower-than-average HR/FB ratios that could be sustainable. The three highest average exit velocities on flies and liners belonged to Richard, Matt Harvey and Luis Castillo. That is particularly distressing news for Castillo, who ranked only behind Jon Gray for the highest HR/FB among qualifed starting pitchers.

Those who are skeptical of Freeland have another Statcast trend to consider. Aside from Kenta Maeda, no one has allowed a lower average exit velocity on ground balls going back to 2015 than the Rockies’ lefty has (min. 500 ground balls). Maybe his .287 career BABIP at Coors Field is not a fluke. Just behind him is Chase Anderson, who has a .274 BABIP over the last four years. At the other end of the leaderboard is James Paxton, who has allowed an average exit velocity of 87.6 mph on grounders. It may not be a coincidence that he owns a .311 BABIP since 2015.

Over the last four seasons, no one has limited average flyball distance more than Eduardo Rodriguez has (min. 250 flyballs). He looks like a solid option when pitching at Fenway Park, where he has compiled a 9.2 percent HR/FB and 3.89 xFIP over the last two seasons. His dampening of flyball distances has not helped him much on the road, with frequent visits to Yankee Stadium, Oriole Park at Camden Yards and Rogers Center, as he has a 13.3 percent HR/FB and 4.27 xFIP in away games going back to 2017. Masahiro Tanaka trails only Chad Bettis for the honor of the highest average flyball distance, so owners should not look for him to improve on this season’s 17.7 percent HR/FB or 3.75 ERA.

We hoped you liked reading Which Statcast Measures Correlate Best with Pitcher HR/FB and BABIP? by Al Melchior!

Please support FanGraphs by becoming a member. We publish thousands of articles a year, host multiple podcasts, and have an ever growing database of baseball stats.

FanGraphs does not have a paywall. With your membership, we can continue to offer the content you've come to rely on and add to our unique baseball coverage.

Support FanGraphs




Al Melchior has been writing about Fantasy baseball and sim games since 2000, and his work has appeared at CBSSports.com, BaseballHQ, Ron Shandler's Baseball Forecaster and FanRagSports. He has also participated in Tout Wars' mixed auction league since 2013. You can follow Al on Twitter @almelchiorbb and find more of his work at almelchior.com.

newest oldest most voted
Ryan DC
Member
Member
Ryan DC

Great stuff. I hope we get a check-in on these groups of pitchers during the season.