Remeasuring the Hit Tool: A Look at Estimated Bat Speed, Collision Efficiency, and Smash Factor

“The hit tool sits atop the great pyramid of tools, trumping its own off-spring–power–as well as the three remaining tools in a position prospect’s physical cache: speed, glove, and arm. The hit tool is the simple measure of how often a ball is properly squared up, driven with authority, and deposited into the field of play.”

Jason Parks, Extra Innings: More Baseball Between the Numbers from the Team at Baseball Prospectus, 2012

In the world of baseball scouting, a hit tool grade for a prospect reigns supreme. A relatively small difference in the grade could be the difference between a player with world-class power hitting 40 home runs at the major league level or having trouble surpassing High-A. In the public realm, whether it be prospect analysis, fantasy baseball, writers, or just general fans of the game, the term is integrated in public player evaluation.

I could be wrong and maybe this is just my evaluation, but it does feel like the term gets oversimplified by getting directly tied to how many a times a player swings-and-misses or strikes out. It’s easy to assume that a player that strikes out a fourth of the time they come up to bat has a below-average hit tool, I’ve probably been guilty of this same exact thing in the past. But it’s not always the accurate assumption.

But as shown in the quote above, hit tool is more than just a hitter’s swings-and-misses and strikeout totals. The analysis for hit tool doesn’t stop when a player *does* make contact. Things like a hitter’s barrel control need to be considered more in the public realm.

Why barrel control doesn’t get as much attention as contact-skills is very understandable. Analyzing it isn’t as simple as putting total strikeouts over total plate appearances. In the Statcast-era, fans and public analysts have been overwhelmed with all the new measurements produced for player evaluation. But for hitter evaluation, there’s unfortunately not many data points past exit velocity and launch angle that are publicly available.

Bat speed is a very crucial measurement for a hitter’s development. While Trackman and Statcast do not track it, other technological resources such as BlastMotion do. And thanks to the past research of many smart people, we can reverse engineer a hitter’s bat speed using publicly available Statcast data to generate a hitter’s estimated bat speed, with a useful level of confidence that it is measuring what it is intended to measure.

Back in 2003, Dr. Alan Nathan wrote a research paper that characterized the performance of baseball bats (there’s lots of advanced physics terms in this, but I highly recommend the read). In that paper, he gave out his formula for calculating a hitter’s exit velocity.

EV = eAvball + (1 + eA)vbat

This is where eA equals the collision efficiency of the bat, vball equals the velocity of the incoming pitch, and vbat equals the speed of the swing.

David Marshall over at the community page of FanGraphs did a good job of breaking down the process of how he reverse engineered bat speed, citing that because collision efficiency and and exit velocity have a linear relationship, and that because we can assume that the average of the upper-echelon and lower-echelon of a hitter’s exit velocities can respectively match up with a 0.21 and -0.10 collision efficiency, a linear regression formula can be ran on a player’s average exit velocity to calculate their average collision efficiency. 

I calibrated eA for each hitter with at least 100 BBE in a season by estimating that the average of the top 15 BBE by exit velocity corresponds to eA =0.21 and the average of the bottom 15 BBE by exit velocity corresponds to eA = -0.1 for each player. Since eA and EV are related linearly, we can compute eA from EV for each player. Finally, I will assume that every player uses a standard 34 in., 32 oz. bat. Since Nathan’s study used a 34 in., 31 oz. bat, I subtracted 0.42 MPH from the estimated swing speeds, because every extra ounce reduces that bat speed by about 0.42 MPH

To perform this analysis myself, I decided to calculate the average 8th percentile exit velocity for each hitter in a season since 2015, along with their average 92nd percentile exit velocity. With the 92nd percentile average exit velocity, a fair enough assumption can be made that the hitter is squaring the ball up perfectly (collision efficiency = 0.21), while with an 8th percentile average exit velocity, the hitter is missing the ball with the barrel of the bat significantly (collision efficiency = -0.1).

As for the results, here were the top 20 hitters since 2015 in average 92nd percentile exit velocity (minimum 50 batted balls total). The total amount of qualified hitter seasons equals 2,662.

Player NameSeasonBatted Balls (n=)92nd Percentile Exit Velocity
Giancarlo Stanton202052116.0
Giancarlo Stanton2016270115.5
Giancarlo Stanton2015181115.1
Giancarlo Stanton2018390115.0
Aaron Judge2017338114.6
Giancarlo Stanton2017437113.9
Franchy Cordero201880113.6
Aaron Judge2018247113.3
Nelson Cruz2017425111.9
Nelson Cruz2018377111.9
Aaron Judge2019227111.8
Gary Sanchez2019262111.7
Joey Gallo2017252111.7
Miguel Sano202099111.6
Daniel Palka2018250111.5
Joey Gallo2018275111.5
Mark Trumbo2016428111.5
Joey Gallo2019125111.4
Jorge Alfaro2018203111.3
Chris Iannetta201991111.2

Next, the bottom 20 in average 92nd percentile exit velocity.

Player NameSeasonBatted Balls (n=)92nd Percentile Exit Velocity
Dee Strange-Gordon20206193.4
Ender Inciarte20209394.4
Jake Elmore20165594.7
Billy Hamilton201921294.9
Luis Sardinas20157195.0
Billy Hamilton201835095.2
Ronald Torreyes20188295.4
Pete Kozma20156695.7
Dee Strange-Gordon201752496.0
Jake Elmore201511396.1
Billy Hamilton201530096.1
Daniel Castro20159196.2
Billy Hamilton201743496.2
Tony Kemp20208196.3
Ben Revere201726896.4
Jesmuel Valentin20185696.4
Billy Hamilton201629996.5
Chris Stewart201710396.6
Darwin Barney201726996.6
Billy Burns201626396.6

Doing the same for average 8th percentile exit velocity, here’s the top 20.

Player NameSeasonBatted Balls (n=)8th Percentile Exit Velocity
Corey Seager20188580.3
Giancarlo Stanton201518179.6
Kyle Schwarber201515278.6
Justin Ruggiano20157878.3
Mike Brosseau20205578.2
Greg Bird201510578.2
Matt Olson201839777.9
Mitch Haniger20167677.6
Nelson Cruz201642477.4
Joey Gallo201827577.4
Joey Wendle20167477.1
Miguel Sano201517676.9
Mike Trout202014476.9
Miguel Cabrera201810876.8
Alex Avila20169176.8
Tommy La Stella20157676.7
Tyler Saladino20188076.7
David Ortiz201544676.6
Teoscar Hernandez20166676.5
Jose Bautista201544776.5

The bottom 20 in average 8th percentile exit velocity.

Player NameSeasonBatted Balls (n=)8th Percentile Exit Velocity
Roman Quinn20206834.7
Cedric Mullins202010235.2
Garrett Hampson202010740.9
Adalberto Mondesi202014841.7
Danny Jansen20209148.0
Giancarlo Stanton20205250.2
Alex Gordon202012651.3
Ryan Rua20177753.9
Hanser Alberto202018954.0
Charlie Tilson20188554.3
Nicky Delmonico201710754.8
Andrew Romine20187555.0
Ichiro Suzuki201716055.1
Trent Grisham202015155.2
Nick Franklin20178355.8
John Hicks201712256.4
Roberto Perez201714556.7
Lane Adams20177157.0
Edward Olivares20207257.2
Dee Strange-Gordon20206157.2

Now with the 8th and 92nd percentile exit velocity data, it’s time to run separate linear regressions for each player to estimate their average collision efficiency. Let’s use 2020 Fernando Tatis Jr. as an example. The x-values for his regression would be his 8th and 92nd percentile exit velocities, so 75.9 and 110.1. Those are matched up against the y-values for collision efficiency, -0.1 and 0.21. With a slope (0.009) and intercept (-0.788) generated for him, we can plug his average exit velocity (95.9) into a linear regression formula and have a fitted value of 0.082, his estimated average collision efficiency.

Player NameSeasonBatted Balls (n=)Average Collision Efficiency
Danny Jansen2020910.109
Luis Torrens2020540.106
Josh Rojas2019760.105
Martin Prado20171190.100
Kelby Tomlinson20171430.097
Elvis Andrus2020880.096
Asdrubal Cabrera20173950.096
Willie Calhoun2020840.096
Nick Ahmed20171280.095
Zack Granite2017850.095
Joey Bart2020620.095
Ty Kelly2017670.095
Ke’Bryan Hayes2020640.095
Freddie Freeman20201760.095
DJ LeMahieu20201760.095
Steve Clevenger2016510.094
Dixon Machado2015530.094
Nomar Mazara2020920.094
J.P. Crawford2018770.094
Jose Abreu20201860.094

Now with an estimated average collision efficiency and publicly available pitch velocity and exit velocity data, we can rework Dr. Nathan’s exit velocity formula to create an estimated bat speed formula.

Estimated Bat Speed = ((Average Exit Velocity – ((Average Pitch Velocity) * Average Collision Efficiency)) / (1 + Average Collision Efficiency)) – 0.42

Finally, here are the results. The top 20 estimated bat speeds since 2015.

Player NameSeasonBatted Balls (n=)Average Bat Speed
Giancarlo Stanton201518188.4
Corey Seager20188586.0
Joey Gallo201827585.0
Giancarlo Stanton201627084.6
Steven Moya20165784.5
Nelson Cruz201642484.4
Kyle Schwarber201515284.4
Aaron Judge201733883.9
Franchy Cordero20188083.9
Matt Olson201839783.8
Pedro Alvarez201623683.6
Mitch Haniger20167683.1
Aaron Judge201922783.1
Jose Bautista201544783.0
Nelson Cruz201930782.8
Jorge Soler201615782.8
Miguel Sano201517682.8
Matt Chapman201837882.7
Jung Ho Kang201911482.7
Nelson Cruz201837782.7

With estimated collision efficiency and bat speed values, there are now measurements to help assess a player’s ability to control the barrel of their bat and drive the ball, two key components of the hit tool. Just needing contact frequency skills, what if they all could be combined to create a more complete quantitative measurement of the hit tool?

That’s where Smash Factor comes into play.

Smash Factor = 1 + (Exit Velocity – Bat Speed) / (Pitch Speed + Bat Speed)

A few days ago, Noah Thurm, Dan Aucoin, and Max Dutto of Driveline released a highly-informative piece that explained the metric.

Smash Factor measures the collision efficiency of the bat and ball at contact, in essence telling us how much of a swing’s bat speed was converted into exit velocity. In simpler terms, balls that are “squared up” with minimal deflection or glancing at contact will have the highest collision efficiencies, and therefore the highest Smash Factors. 

They also detailed how the metric can be a more complete measurement of a hitter’s hit tool once foul balls and whiffs are considered.

Considering Smash Factor just on batted balls, however, only covers the quality part of our contact skill evaluation. By assigning whiffs and fouls a Smash Factor of 0, taking a player’s average describes both how often and how well they make contact. This is where the value added by Smash Factor is clearest—K% focuses on at-bat level performance, BABIP describes batted-ball luck, and Z- and O-Contact rates only apply to proportions of all pitches seen. Smash Factor is usable on every pitch a batter swings at, making it faster to reliability and a more robust single measure of hitter skill.

Finally, with all the data and estimations in place, Smash Factor (with the fouls and whiffs considered) can be calculated, giving what seems to be the most complete statistical measurement of a hitter’s hit tool. Since 2015, here are the top 20 Smash Factors in a season, minimum 200 batted balls. As you can see, Ben Revere, Andrelton Simmons, and David Fletcher reign king in this metric.

Player NameSeasonBatted Balls (n=)Overall Smash Factor
Ben Revere20163000.642
Ben Revere20155120.605
Andrelton Simmons20184900.605
Ben Revere20172680.600
Andrelton Simmons20154590.589
Mookie Betts20175520.588
Andrelton Simmons20164010.585
David Fletcher20195000.584
David Fletcher20182500.578
Ben Zobrist20154280.575
Mookie Betts20165850.566
Martin Prado20165270.563
Johnny Giavotella20162990.561
Eric Sogard20152860.559
Denard Span20152150.559
Michael Brantley20154610.557
Andrelton Simmons20193290.557
J.J. Hardy20163370.554
Martin Prado20154350.550
Tommy La Stella20192560.548

Investigating the year-over-year correlation for overall smash factor shows the metric is a consistent skill measurement for hitters, as expected. The r^2 for hitters with back-to-back seasons with at least 200 batted balls equals 0.7.

With a new and improved measurement for hit tool, we can stack the new metric up against more simple and traditional stats tied to hit tool, such as strikeout rate. Looking specifically at 2019 hitters with at least 250 plate appearances (n = 208), there seems to be a very strong relationship between strikeout rate and Smash Factor (r^2 = 0.8).

But while there’s a solid relationship between the two, some hitters on the extreme find themselves as outliers, as Smash Factor would either lower or heighten their perceived hit tool (K%+). Some examples being…

  • Jeff McNeil: 139 K%+, 98 Smash Factor+
  • Luis Arraez: 165 K%+, 131 Smash Factor+
  • Daniel Vogelbach: 83 K%+, 103 Smash Factor+
  • Javier Baez: 73 K%+, 96 Smash Factor+
  • Adalberto Mondesi: 70 K%+, 87 Smash Factor+

The hit tool is the most important skill a hitter can have in this game. Yet, varying definitions of the term and lack of publicly available data have made it something not easy to analyze. But thanks to the great research of others, estimated bat speed, collision efficiency, and Smash Factor make it easier.


Related Work

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create your website with WordPress.com
Get started
%d bloggers like this: