Which Statcast Measures Correlate Best with Power Metrics?

Note: Many thanks to Alex Chamberlain, who provided the correlations cited in this column, as well as insights regarding some of the relationships.

As I have searched for ways to spot undervalued power sources over the last few seasons, I have relied heavily on several Statcast metrics that are available on Baseball Savant. I have leaned especially hard on average flyball distance. While the leaderboard typically includes several players who are proven power sources, it has also featured some players who appear to be undervalued. For example, Scott Schebler, Kendrys Morales, Trey Mancini, Tim Beckham and Mitch Moreland all finished in the top 20 percent in average flyball distance in 2017 (min. 50 flyballs), and that gave me a little extra confidence to give them a try in 2018. For stretches, Morales and Moreland paid some dividends, but it was far from a foolproof method.

Yet it’s a metric that reflected well on Max Muncy, Brandon Nimmo and Johan Camargo when they were still widely available, as we headed into the middle portion of the 2018 season. Though adding any one of these hitters based on a small sample of flyballs was risky, the gamble paid off, particularly with Muncy, who finished in the top 12 in Roto value (per ESPN) at three different positions.

It turns out that average flyball distance correlated strongly with HR/FB, ISO and hard contact rate in 2018. As the table below shows, the only Statcast metric that had a stronger correlation with HR/FB was barrel rate, both as a proportion of batted ball events and plate appearances. The Pearson’s r coefficient for all three metrics with HR/FB was 0.80 or higher. The only other metric that is similarly associated with HR/FB is average exit velocity on flyballs and line drives, which had a Pearson’s r of 0.79,

Correlations Between Statcast Metrics and HR/FB, ISO and Hard Contact Rate, 2018
HR/FB ISO Hard%
Brls/BBE % 0.85 0.84 0.75
Brls/PA % 0.84 0.85 0.75
Avg FB Dist 0.80 0.72 0.72
EV FB/LD 0.79 0.72 0.75
Dist Max 0.68 0.63 0.52
EV Avg (Overall) 0.67 0.64 0.73
EV Max (Overall) 0.61 0.53 0.49
Overall Avg Dist 0.44 0.68 0.59
Avg HR Dist 0.43 0.38 0.34
95 MPH+ 0.41 0.49 0.47
EV GB 0.34 0.31 0.43
SOURCE: Baseball Savant
Includes cases with a minimum of 150 batted ball events. Average flyball distance is for hitters with at least 50 flyballs.

Barrels per batted ball event and barrels per plate appearance were also the most strongly correlated of these metrics with ISO. Average flyball distance and average exit velocity on flyballs and line drives are also strongly correlated with ISO, though the gap between them and the barrel rate measures was wider than they were for HR/FB. There is greater parity among all four measures in terms of their correlation with hard contact rate, with each having a coefficient in the 0.71-to-0.76 range. Not surprisingly, average exit velocity for all hit balls is also one of the strongest correlates of hard contact rate. Since hard contact is hard contact, regardless of launch angle, there is no need to parse out ground balls when looking at the relationship between exit velocity and hard contact rate.

It may seem obvious that there would have been a strong positive relationship between exit velocity on flyballs and line drives and measures of power production (i.e., HR/FB, ISO). Yet there is reason to question that relationship, as there is evidence to suggest the ball may have been de-juiced in 2018 (or at the very least, that it took a higher frequency of hard contact to produce a given level of power production). Even with the possibility that batters were hitting a de-juiced ball this past season, differences in exit velocity on airborne balls (and by extension, differences in barrel rate) made a substantial difference in how much extra-base power was generated.

Not only are barrel rates and exit velocity on flyballs and line drives strongly correlated with HR/FB, ISO and hard contact rate, but they appear to be reliable indicators of what is to come in the following season. Exit velocity on flyballs and line drives had the strongest year-to-year correlation between 2015 and 2018 with a coefficient of 0.82, and barrels per batted ball event were slightly behind at 0.80. Barrels per plate appearance also correlate strongly from year to year, though not quite as much, clocking in with a coefficient of 0.76.

Year-to-Year Correlations, 2015-2018
Measure Year-to-Year Correlation
EV FB/LD 0.82
Brls/BBE % 0.80
EV Avg (Overall) 0.78
EV Max (Overall) 0.77
Brls/PA % 0.76
Overall Avg Dist 0.76
Avg FB Dist 0.65
95 MPH+ 0.63
EV GB 0.61
Dist Max 0.56
Avg HR Dist 0.17
SOURCE: Baseball Savant
Includes cases with a minimum of 150 batted ball events. Average flyball distance is for hitters with at least 50 flyballs.

The year-to-year correlation coefficient for average flyball distance from 2015 to 2018 was 0.65, which is still fairly strong. Yet aside from average home run distance, the year-to-year correlation for average flyball distance was not especially stronger than it was for any of the Statcast measures in the table. While average flyball distance is not a bad tool for finding cheap power in the coming season, barrels per batted ball event and exit velocity on flyballs and line drives are even better.

A quick scan of the barrel/BBE rankings from 2018 reveals a number of hitters who could have some power upside in 2019. Jake Cave, Chad Pinder and Ian Happ stand out among this group, as all ranked among top 10 percent of hitters with at least 150 batted balls. All finished with an ISO below .210, though each was impeded by a strikeout rate that was well above the major league average. None of the three is assured of regular playing time, but if they can become fixtures in their team’s lineups and cut back on strikeouts, they could be inexpensive power sources.

The EV FB/LD leaderboard has a lot of overlap with the barrel/BBE leaderboard, but Jorge Soler (.202 ISO) and Aaron Altherr (.152 ISO) are somewhat surprising entries among the top 25. Altherr may need a change of scenery to find playing time, but he may not need a bandbox like Citizens Bank Park in order to resuscitate his power numbers. Soler will need to stay healthy, but his power potential makes his name one to add to your list of late-round fliers.

We hoped you liked reading Which Statcast Measures Correlate Best with Power Metrics? 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
NickGerli
Member
Member
NickGerli

Good stuff. Paying attention to the Brls/BBE stat early in the season can lead to you to some diamonds in the rough (Muncy, e.g.) before other people catch on.

A big Pinder fan as well. I wonder if Oakland gives him full playing time at second with Lowrie likely moving on.

RonnieDobbs
Member
RonnieDobbs

It can lead you a lot of directions. I recall a time when Taylor Motter was all the rage of these digital pages. Would a handful of lucky swings not lead to that as well?