Submitted by Cassie Reilly-Boccia1*, Travis Ficklin2*, Robin Lund3*
1* Director of Research and Development at Athletes Warehouse in Pleasantville, NY
2* Assistant Professor of Movement and Exercise Science at the University of Northern Iowa
3* Associate Professor of Movement and Exercise Science at the University of Northern Iowa
Cassie Reilly-Boccia is a former member of the National Champion University of Alabama softball team and is the Director of Research and Development at Athletes Warehouse in Pleasantville, NY. Travis Ficklin is an Assistant Professor of Movement and Exercise Science at the University of Northern Iowa. Robin Lund is an Associate Professor of Movement and Exercise Science at the University of Northern Iowa.
The purpose of this study was to describe basic kinematic variables of the swing and the relationships that exist between these variables in Division I female softball players. These variables included bat velocity (BV), bat quickness (BQ), and bat acceleration (BA). Video data were collected for all swings during a 15-game softball tournament in which six NCAA Division I teams played. High-speed video cameras recording at 300 Hz were located along the first and third base lines recording every pitch. Data from 1,099 swings were analyzed for bat velocity (BV), bat quickness (BQ), and bat acceleration (BA). BQ and BV were calculated by video analysis and digitization. All swings were rank ordered by BA and assessed for relationships among BV, BQ, and BA. Descriptive statistics (mean ± SD) were calculated for all swing kinematic variables. Pearson product moment correlations were used to examine relationships among the swing kinematic variables. Alpha was set at (p<0.05) for all tests. Mean BV for all swings was 28.77 ± 4.94 m/s, mean BQ for all swings was 0.208 ± 0.042 s, and mean BA for all swings was 144.39 ± 38.44 m/s2. When observing correlations of all swings, BV and BQ unexpectedly had an inverse relationship. When grouping swings into homogenous strata based on BA, BQ, and BV proved to have a significant positive correlation.
Key words: softball, bat velocity, bat quickness, bat acceleration
Historically, one of the most difficult actions to complete in sports is to successfully hit a baseball or softball. Athletes and coaches from both sports have searched for countless ways in order to gain supreme technical proficiency in this task. Bat quickness (BQ, s) and bat velocity (BV, m/s) are two aspects of a swing that can be evaluated in order to define a successful hitter.
Bat quickness is defined by the time it takes to complete the swing (9). To calculate BQ, the number of frames between swing onset and bat contact with the oncoming ball were counted and multiplied by 1/300s, which is the time of one frame. Bat velocity is measured in m/s and is measured at the instant the bat makes contact with the ball. Theoretically, if a hitter can complete the swing in a shorter amount of time (decreased BQ), this will allow the hitter more time to evaluate the incoming pitch thus make a better decision about whether or not to commit to the incoming pitch (3). An increase in BV will potentially lead to an increase in ball velocity off the bat, thus increase the chance of the batted ball being recorded as a hit (5). Previously, a study found that a higher BV would require more time thus a higher value for BQ. Conversely, a quicker swing that was completed in less time (lower BQ), would have a slower BV (9).
An effective swing is dependent on executing the phases of the swing via a kinetic chain occurring in a sequential order. Proper technique theoretically leads to a hitter optimizing BV and BQ. A novel way of evaluating the relationship between BQ and BV is to consider bat acceleration (BA). Bat acceleration is an average acceleration of the bat head calculated by dividing the BV in m/s of any swing at contact by the swing BQ time in seconds. Greater BA values come from some combination of maximizing BV and minimizing BQ.
Previously, of the swing research that exists, most studies have been conducted with male athletes in baseball, with very few done with female athletes in softball. Specifically, Stellar and his colleagues (1993) (9) evaluated BQ and BV in Major League Baseball (MLB) professional male athletes. Despite the obviously physiological differences that exist between males and females, the constraints of baseball and softball should demand similar swing kinematics from hitters in both sports. Despite the pitching velocities being significantly different between the two sports, due to the pitching distance in softball being shorter than that of baseball, the reaction times for hitters are also similar. However, no such study evaluating BQ, BV (measured from the end of the bat head), or BA of in game hitting performance has been completed in softball. Therefore, the purpose of this study is to describe the swing kinematics (BQ, BV, and BA) and the relationships that exist between them in Division I female softball players.
Subjects for this study were members of six NCAA Division I softball teams, ranging in Ratings Percentage Index (RPI) at the time of data collection from 1 to 217. The teams in this study were ranked at the following RPI: 1, 59, 71, 73, 161, and 217. The participants in this study were between the ages of 18 and 23 years old. All data collection procedures were approved by the university Institutional Review Board.
Video data were collected for all pitches during a 15-game softball tournament in which the six NCAA Division I teams played. For every pitch of the tournament, video cameras (JVC GC-PX1, Tokyo, JP) shooting at 300 Hz were used to capture any swing made. Video was converted to the AVI format for use in Maxtraq (Innovision Systems, Inc., Columbiaville, MI) software for digitizing. Digitized data were exported in comma separated value files and analyzed using custom Matlab software (Mathworks, Natick, Massachusetts).
Cameras were positioned on the first-base and third-base sides of the field with their optical axes aligned with the front edge of home plate. The first-base-side camera was used to capture any swings made by right-handed hitters, while the third-base-side camera was used to capture any swings made by left-handed hitters. Video was recorded for every pitch thrown using the camera pertaining to each hitter. This resulted in over 5000 video recordings. For analysis, only videos made of full swings were used. Left-handed slap attempts, balls or strikes taken (not swung at) by the batter and all pitches that hit the batter were excluded. Additionally, if contact was made in front of the batter box or behind the back corner of the plate, the swing was excluded from analysis due to the out-of-plane motions of the bat head, and therefore unreliable positional data, that resulted. Based on these criteria, 1,099 swings made by 80 separate players were analyzed.
These video clips were trimmed and converted to the AVI format, then transferred to the Maxtraq program for digitizing. Digitizing and calculations were executed using a method that has been shown to yield valid velocity data for swings recorded from the camera orientation used (4). eference points were digitized at the front inside corner of both batters boxes and the back corner of home plate. To mark the onset of the swing, the head of the bat was digitized in the frame at which the swing started. The onset of the swing was judged to happen in the frame where the bat head began initial trajectory toward contact with the ball. Then, the bat head was digitized at contact or, in the case of a miss, in the frame at which it had attained the same horizontal position as the ball. The bat head was also digitized four frames prior to the frame of contact. Data were exported in a comma separated value file containing the coordinates of the reference points, the location of the bat head at contact, the location of the bat head five frames prior to contact, and the location of the bat head at the onset of downswing.
All calculations were carried out using custom software written in Matlab. To make position and velocity calculations, a reference distance was calculated. The centroid of the batters box corners was calculated, and based upon the known distance from the back corner of the plate to this centroid (1.4351m), a scale factor for each trial was calculated for conversion from pixels to meters. To calculate BV, the horizontal displacement of the bat head between contact and five frames prior to contact was converted to meters and divided by 1/75s, which is the time spanned by four frames. This time span was chosen in order to get a valid velocity at contact while minimizing the percent error of digitizing that could occur if only single frame spans were to be used (4). To calculate BQ the number of frames between swing onset and contact were counted and multiplied by 1/300s, which is the time of one frame. BA was calculated by dividing BV by BQ to yield acceleration in m/s/s.
Pearson product moment correlations were used to determine the strength of the relationships in question. Alpha was set at p<0.05 for all tests. SPSS 21 (International Business Machines Corp, USA) was used for all analyses.
Descriptive statistics of swing kinematic variables for all swings can be found in Table 1. The mean BV for all swings was 28.77 m/s (± 4.94 m/s). The mean BQ for all swings was 0.21 s (± 0.042 s). The mean BA for all swings was 144.30 m/s2 (± 38.44 m/s2). The descriptive statistics of the swing kinematic variables collapsed for all players can be found in Table 2. The mean BV for all players was 29.12 m/s (± 2.19 m/s). The mean BQ for all players was 0.20 s (± 0.03 s). The mean BA for all players was 148.54 m/s2 (± 22.37 m/s2). As expected, BA had significant relationships with both BV (0.65) and BQ (-0.85) since BA was calculated from these variables.
Correlations between all swing kinematic variables for the collapsed averages calculated for each player can be found in Table 3. No significant relationship between BV and BQ existed. The correlation between BV and BQ for collapsed swings was
Due to the findings of the current study, it is evident that the basic swing kinematic variables of a Division I female softball player and the relationships that exist between these variables have considerable applications to the sport. The variables described for Division I softball players can now be compared to those of previous studies completed with MLB hitters. These findings can continue to improve the way athletes in both sports are coached and developed.
The results of the current study indicated that Division I softball players had similar swing kinematic variables to MLB professional hitters. Despite the physiological differences that exist, there can be several reasons to explain why there are similarities. The characteristics of the bat differ greatly in both sports; the average weight of a fast pitch softball bat is much less compared to that of an MLB bat. The range for a collegiate softball bat weight and length is 23-28 ounces and 32-34 inches. A profession baseball bat weight and length is 32-34 ounces and 31-34 inches (6). 6However, the length and weight of the bat alone do not make the bat more or less difficult to swing.
The moment of inertia of a bat is dependent on the mass of the bat and how that mass is distributed with respect to the pivot point (8). The further the mass of the bat is distributed from the pivot point – approximately at the mid-grip of the bat as the bat is held in the hands – the larger the moment-of-inertia will be. The larger the moment-of-inertia, the more difficult it is to swing the bat. The way in which the weight is distributed for a softball bat compared to a baseball bat leads to a softball bat feeling easier to swing compared to the bats used by MLB (8). Additionally, a study found a direct relationship between the speed of the ‘sweet-spot’ (the point of the bat that is most advantageous to make contact with the ball) and the speed of the bat center of mass. This would conclude that if a hitter had the sufficient strength and swing technique to handle a greater moment of inertia, this could result in a larger bat velocity via improved bat acceleration for the hitter (2). Because of these differences in length, weight, and moment-of-inertia, the similar swing parameter values found in this study can be explained despite a lack of female collegiate player strength compared to that of the major league counterpart.
Total reaction time given to hitters (the time between the release of the ball by the pitcher to the instant at which contact needs to be made) is comparable between elite-level baseball and softball hitters. For example, a 95-mph fastball pitched from a 60 ft 6 in distance in baseball has comparable travel time to that of a 66 mph softball pitched from 43 ft in softball making the reaction time for hitters in each sport to be approximately 0.40 s (1). It is therefore reasonable that the BQ values for the subjects in the present study would need to be similar to those of elite baseball hitters in previous studies (1,8). Because of the reduced mass and moment of inertia of the softball bats, it is reasonable to witness similar BV values generated during the BQ times of swing. Indeed, many coaches do not distinguish between a baseball or softball swing. Instead, one swing is taught that is viable for both sports (7). There is nothing that limits the female player from having identical swing mechanics to those of a male hitter. The biological strength differences are then minimized by the differences in bat properties and the rules of each game.
As noted previously, the study conducted by Stellar, House, DeRenne, and Blitzblau (1993) (9) reported significant relationships between BV and BQ, namely that the two variables are directly related. That is, as BQ increases, there is more time for the hitter to generate a higher BV. As BQ decreases, the hitter is left with a smaller amount of time, and thus generates a smaller BV. In the current study, this relationship did not materialize (Figure 1). Instead, a significant inverse correlation was observed between BV and BQ. In other words, the greater BV was associated with shorter BQ, meaning the best velocities were being generated in the least time.
This unexpected result was most likely due to the heterogeneous nature of the population of swings that were examined. As mentioned, the range in RPI for this convenience sample was 1-217, meaning the data included kinematic swing variables from some of the best, but also from some of the worst teams in Division I softball. This variety in talent may explain why the best velocities in BV were associated with the smallest swing times in BQ (likely better players on better teams) and vice versa (worse players on worse teams).
To explore this hypothesis, a secondary analysis of the data was performed. In an attempt to group like swings with one another, all swings were sorted by BA and placed into ten strata of similar size based on percentiles. Bat quickness and BV correlations for each strata can be found in Table 5. As hypothesized, the relationship between BV and BQ became positive within a stratus of similar swings. However, the strength of these relationships within some of the strata was unexpected. BQ and BV correlations produced coefficients that ranged from 0.379 – 0.975. However, these correlations were found to be weakest in the top percentile group and lowest percentile group. The strength of the relationship between BV and BQ is clear in the middle eight strata. This strength of relationship begins to deteriorate in the extreme strata enough so that these swings disrupted the overall expected model for BV and BQ. These relationships can also be observed in Figure 2.
Figure 3 displays data of BQ vs. BV with the top and bottom 10th percentiles removed from the data set. By doing so, a positive significant correlation of 0.136 is observed. Figure 4 displays data of BQ vs. BV with the top and bottom 20th percentiles removed from the data set. By doing so, an even stronger positive significant correlation of 0.428 is observed. This shows that when homogenous swings are grouped with one another, the expected direct relationship between BV and BQ materializes. This also suggests that within this current sample, the best swings in terms of BV were also executed in the shortest times. These were swings probably executed by stronger, higher level players, and future research should examine the strength training implications for these kinematic variables in executing a swing.
In conclusion, the swing kinematic variables described in this study clearly show that similarities exist between Division I softball hitters and MLB baseball hitters. The likeness in the swing parameters between both males and females despite obvious differences in relative strength can be explained by the bat properties and the constraints of each game.
Interesting relationships materialized when evaluating correlations within the swing kinematic variables. Unexpectedly, an overall negative relationship between BQ and BV became evident when analyzing all swings. However, when these swings were percentile ranked and each swing grouped with other like swings, there was a clear positive correlation between BV and BQ as expected. It appears the heterogeneous nature of the convenient sample contributed to the overall negative correlation between these swing kinematic variables. Future studies should take a look at the predictive power of these swing kinematic variables on in game statistical performance variables.
APPLICATIONS IN SPORT
In coaching baseball and softball athletes, coaches should resist the urge to evaluate bat velocity alone. Being careful that a hitter does not use too much swing time in maximizing velocity will aid the hitter in pitch selection and becoming a more consistent hitter. It is evident that the combination of both BV and BQ plays a significant role in a swing, and future research should be conducted to verify this.
|Table 1. Descriptive Statistics for All Swings|
|Bat velocity (m/s)
Bat quickness (s)
Bat acceleration (m/s2)
|Table 2. Descriptive Statistics for All Players|
|Bat velocity (m/s)
Bat quickness (s)
Bat acceleration (m/s2)
|Table 4. Correlations of Swing Kinematics for Collapsed Swings|
|Table 5. BQ/BV Correlated by Percentile Strata|
Figure 1. BQ vs. BV means for every swing executed during the tournament.
Figure 2. BV and BQ Relationships Stratified by Deciles
Figure 3. BV vs. BQ Relationships for Swings in 10th-90th Percentile
Figure 4. BV vs. BQ Relationships for Swings in 20th-80th Percentile
- Brenkus, J. (Producer). (2007). Sports Science: Episode 7 (Motion Picture). United States: Base Productions.
- Bahill, A.T. (2004). The ideal moment of inertia for a baseball or softball bat. IEEE Transactions On Systems, Man, and Cybernetics – Part A: Systems and Humans. University of Arizona, Tuscon AZ.
- DeRenne, C. (1993) High-tech hitting. Laguna Hills, CA: West Pub Co.
- Ficklin T.K., and Lund R. (2014). Validation of a two-dimensional video method for measuring in-game softball bat velocity. World Congress on Biomechanics. (Boston, MA).
- Fortenbaugh, D.M. (2011). The biomechanics of the baseball swing (Doctoral dissertation). University of Miami Library, Miami, FL.
- Noble, L., and Eck, X. (1985). Empirical determination of the axis of percussion of softball and baseball bats. Biomechanics Champaign, IL. pp. 516-520.
- RightView Pro. (2009). Baseball and softball training with Mike Candrea [Video file]. Retrieved from http://www.rightviewpro.com/about-us/founders- staff/candrea.
- Smith, L., Broker, J., & Nathan, A. (2003). A study of softball player swing speed. Sports Dynamics Discovery and Application. RMIT University, Melbourne, Australia, pp. 12-17.
- Stellar, T., House, T., DeRenne, C., & Blitzblau, A. (1993). The absolutes of hitting: Dynamic balance, kinetic links, axis of rotation, bat lag. (Motion Picture) United States: BioKinetics Inc.