Seeing or Doing?

Pitch Recognition of Batters versus Pitchers: A Preliminary Report

Yin-Hua Chen, Pei-Hong Lee, Yu-Wen Lu and Nai-Shing Yen

Research Center for Mind, Brain and Learning, National Chengchi University

No. 64, Sec. 2, Zhi-Nan Rd., Wen-Shan District, Taipei 11605, Taiwan

Keywords: Baseball, Pitch Recognition, Action Anticipation, Pitcher, Batter.

Abstract: In this study we tackled the question: between the experience of seeing or doing the movement, which one

is more important in understanding the observed movement? We thus asked batters and pitchers, in high and

intermediate skill levels, to identify the type of pitch that was edited in difference lengths. In general, we

found that advanced players showed significant higher accuracy and lower uncertain rate than the

intermediate players, particularly in viewing short pitch sequences. These results reflected the requirement

of fast sports such as baseball, in which players have to make a correct decision quickly rather than staying

uncertain. Moreover, advanced batters showed the tendency of being more accurate than advanced pitchers,

though the difference did not reach statistical significance possibly due to small sample size. In consistency

with the previous studies, all players showed higher accuracy in identifying the strike pitches when they

could see longer sequence of the pitch motion and the baseball trajectory (Paull & Glencross, 1997). In sum,

our results supported the notion that when understanding an observed movement, the perceptuo-motor

experience reacting to it is more important than the actual motor experience of the observed movement.

1 INTRODUCTION

In anticipating the action of the opponent in sports, it

has been shown that the experience plays an

important role (review see Williams, Davids, &

Williams, 1999). That is, due to the accumulated

experience, a skilled athlete knows where to view in

the opponent and then makes the best use of the

information extracted to act or react to the opponent

(e.g., Farrow & Abernethy, 2003; Aglioti et al.,

2008). For example, advanced baseball batters pay

close attention to the pitcher’s motion particularly in

the pitcher’s shoulder, elbow and wrist, and then

switch the focus to the ball trajectory with fewer

fixations than the intermediate batters for making the

batting decision (e.g., Hubbard & Seng, 1954; Shank

& Haywood, 1987; Takeuchi & Inomata, 2009).

In a recent study, it was investigated the ability

to predict the fate of actual or fake soccer penalty

kicks between goalkeepers, kickers and novices

(Tomeo et al., 2012). Goalkeepers showed higher

accuracy for fake actions as compared to kickers and

novices. Kickers were even more confused by the

fake actions than goalkeepers and novices. The

authors concluded that goalkeepers could

outperform kickers and novices due to their visual

rather than motor expertise. However, we thought

that goalkeepers should be considered as “visuo-

motor” experts since they are trained to “perceive

and react” to the penalty kicks. Kickers, instead,

don’t have to intercept the penalty kick even though

they are capable of doing a fool action.

In fact, baseball and football can be considered

very special sports because the players have two

very distinctive roles. Take baseball player for

example: the pitcher is responsible for throwing the

pitch and the batter has to bat and run. The two roles

of players have developed very specialized

perceptual and motor expertise depending on the

task required in the match. Thus, we would like to

investigate whether the pitcher (who possesses the

expertise of performing the pitch motion) or the

batter (who possesses the perceptuo-motor expertise

of intercepting the pitch) could better recognize

whether the pitch is a strike or a ball. We thus asked

elite pitchers and batters to identify whether a pitch

is a strike or a ball and compared their performance

with intermediate players. The pitch sequence was

edited in different lengths to see whether different

amount of the information of the baseball trajectory

could differently influence the pitch identification

Chen Y., Lee P., Lu Y. and Yen N..

Seeing or Doing? - Pitch Recognition of Batters versus Pitchers: A Preliminary Report.

DOI: 10.5220/0005144700170024

In Proceedings of the 2nd International Congress on Sports Sciences Research and Technology Support (icSPORTS-2014), pages 17-24

ISBN: 978-989-758-057-4

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

for pitchers and batters with high and low expertise

levels.

We expected that advanced players, both

pitchers and batters, would be more accurate than

intermediate players due to their much more

experience in pitch identification in general.

Furthermore, batters would be more accurate than

pitchers, particularly in advanced group, because

they were the so-called perceptuo-motor experts in

this task in the match.

2 METHODS

2.1 Participants

We recruited 9 high-level pitchers (hereafter HP;

mean age=21; training years=12.52; hours per

week=16.19

) and 18 high-level batters (hereafter

HB; mean age=20; training years=10.22; hours per

week=23.94) from highly ranked Taiwanese

university baseball team in this study. Most of them

had the experience of participating in international

competition. Moreover, a group of intermediate

pitchers (n=7; hereafter IP; mean age=23; training

years=5.43; hours per week=4.14) and intermediate

batters (n=12; hereafter IB; mean age=24; training

years=5.79; hours per week=6.63) were recruited as

control groups. All participants were right-handed

males, and with the height about 180 cm to have

similar strike zone. This study was approved by the

Research ethics committee of National Taiwan

University and was in accordance with the

Declaration of Helsinki; participants gave written

informed consent.

2.2 Stimuli

The stimulus sequences were colour video clips

(wmv format) of baseball pitches of 2 skilled

pitchers. The 2 skilled pitchers were asked to throw

four-seam fastballs to the strike zone of a 180-cm

right-handed batter from the pitcher’s mound toward

the catcher, given a draw situation of full count (2

strikes & 3 balls), 2 out, and full base at the last

inning. The video sequences were taken from the

right-handed batter’s perspective using video camera

(SONY HDR-XR150; 30 frames/s; setting see

Figure 1). 9 strikes and 9 balls thrown by each

pitcher were recorded, making a total of 36 (2

pitchers x 2 types of pitch x 9 throws) different

throws. Whether the pitch was a strike or a ball was

judged by a skilled catcher on site. The criterion of

recruiting the 2 skilled pitchers and catcher was the

same as the criterion of recruiting the advanced

skill-level batters. The average speed of the throws

was controlled at around 115 km/hr by a speed gun.

We then edited each video in 12 different lengths,

which include the windup preparation phase and the

pitching phase till the moment of the baseball

released from the pitcher, or 33, 67, 100, 133, 167,

200, 233, 267, 300, 333, and 367 ms after the

baseball released from the pitcher, respectively.

2.3 Task

The task is twofold. Right after viewing the pitch,

participant had to decide whether he would swing

the bat or not (to bat, not to bat, or uncertain) by

pressing the response key 1, 2, or 3 with index,

middle or ring finger. Immediately after this batting

decision, he had to recognize the pitch type (strike,

ball, or uncertain) again by pressing one response

key with its corresponding finger. The response key

(1, 2, or 3) assigned to the answer of batting decision

(to bat, not to bat, or uncertain) were

counterbalanced between participants. The response

key of the answer of batting decision (to bat, not to

bat, or uncertain) was always combined with the

response key of answer of pitch recognition (strike,

ball, or uncertain) following the nature of batting a

strike and not batting a ball. All of the responses had

to be made in 2.5 s, or the trial would be skipped.

We reminded participants to respond as quickly as

possible, but we emphasised accuracy over speed.

Figure 1: The display of experimental apparatuses: the

blue filled box indicates the position of video camera.

icSPORTS2014-InternationalCongressonSportSciencesResearchandTechnologySupport

2.4 Procedure

Before testing, we demonstrated the video sequences

of the 9 strike and 9 ball pitches of each pitcher to

the participant. The video sequences were longer

than the testing stimuli because they were terminated

at the moment of 200 ms before the baseball was

caught by the catcher. This procedure was applied to

let the participants familiar with the strike zone

judged by the catcher and to let participants adapted

to the scene filmed by the video camera. We then

explained the task to the participant and the

participant could practice at least 10 trials to make

sure that the task is fully understood.

In each trial, the participant was presented with a

fixation cross displayed on a white background and

located in the centre of the screen (1024x768, 60Hz)

for 1 s. Next, the video clip of the pitch was played.

When the video clip terminated, participant had to

decide whether to bat or not and to recognize the

pitch type. The inter-stimuli interval (ISI) was 1 s

(See Figure 2). There were 432 (2 pitchers x 2 types

x 9 pitches x 12 video lengths) trials, randomly

divided into 8 runs, to be completed. Between each

runs, participant could have a short break of 3-5

minutes. The entire experiment took approximately

1.5 hr. The experimental protocol was written using

Eprime 2.0. The response and response time of

participants were registered for data analysis.

2.5 Data Analysis

We calculated the correct, incorrect, and uncertain

response in percentage of each participant in each

experimental condition. The data was then entered

into 3 separate repeated-measures mixed-model 3-

way (4 groups x 2 types of pitch x 12 lengths of

video) ANOVAs for the correct, incorrect and

uncertain response, respectively. For all ANOVAs,

group was the between-subject factor, and type of

pitch and length of video were within- subject

factors. The threshold for significance was set at p

< .05. A Bonferroni adjustment was used for

multiple comparisons. SPSS 20.0 was used for

statistical analysis.

3 RESULTS

In Figure 3, we demonstrated the percentage of

“correct”, “incorrect” and “uncertain” response of

pitch identification of four groups for strikes and for

balls, in different lengths of video sequence of the

pitch. The statistics were reported in Table 1.

3.1 Correct Response

The ANOVA detected a significant main effect of

group, with advanced players showing higher

accuracy than intermediate players (mean value of

60%, 62%, 55%, and 55% for HP, HB, IP, and IB,

respectively). Post-hoc analyses indicated that HB

showed significantly higher accuracy than IB. There

was also a significant main effect of pitch type, with

higher accuracy in strikes than in balls. Moreover,

we found a significant pitch type-by-group

interaction, with HB being more accurate than IB

particularly for strikes. And, all groups except for IB

showed higher accuracy in strikes than in balls (see

Figure 4 top panel). The main effect of length of

video was also significant, showing that all players

were more accurate when they could see longer

pitch sequence. Video length-by-group interaction

was also significant, for that HB showed

significantly higher accuracy than intermediate

players (both HP and HB) especially for short videos

(length 1~4; see Figure 4 middle panel). The video

length-by-pitch type interaction was also significant,

with higher accuracy in strikes than in balls

particularly for long videos (length4, and 6~12). The

3-way interaction was not significant.

Figure 2: The procedure of a trial.

SeeingorDoing?-PitchRecognitionofBattersversusPitchers:APreliminaryReport

Figure 3: The percentage of “correct”, “incorrect” and “uncertain” response of pitch identification for strikes and for balls

of four groups (HP: high-level pitcher, HB: high-level batter, IP: intermediate pitcher, IB: intermediate batter) after viewing

different lengths of video sequence of the pitch. The 12 different video lengths showed the windup preparation phase and

the pitching phase until the moment of the baseball released from the pitcher, or 33, 67, 100, 133, 167, 200, 233, 267, 300,

333, and 367 ms after the baseball released from the pitcher, respectively.

icSPORTS2014-InternationalCongressonSportSciencesResearchandTechnologySupport

Table1: Statistic s of all main effects and interaction effects for correct rate, incorrect rate, and uncertain rate.

Index Effect F value p value Pairwise comparisons

Correct rate

(%)

Group F

(3,42)

= 3.67 p < .05 HB > IB

Pitch type F

(1, 42)

= 38.55 p < .001 Strike > ball

Pitch type-by-group

interaction

(3,42)

= 3.29 p < .05

HB > IB for strikes;

pitch type effect for all groups

expect for IB

Video length

(11, 462)

136.61

p < .001 Long > short

Video length-by-group

interaction

(33,462)

= 4.58 p < .001

H > I, in length1&2;

HB > I, in length3 & 4

Video length-by-pitch

type interaction

(11, 462)

= 18.54 p < .001

Strike > ball, in long videos

(length 4, and 6~12)

Incorrect rate

(%)

Group F

(

)

= 3.60 p < .05 HP > IP

Pitch type F

(

)

= 42.35 p < .001 Ball > strike

Pitch type-by-group

interaction

(3, 42)

= 3.00 p < .05 HB > IB, for balls

Video length F

(

473

)

= 5.89 p < .001

Video length-by-group

interaction

(33, 462)

= 2.77 p < .001

H > I, in length1;

HP > I, in length3;

HP > IP, in length5

Video length-by-pitch

type interaction

(11, 462)

= 16.26 p < .001

Ball > strike, in long videos

(length 4~12)

Uncertain rate

(%)

Group F

(

)

= 5.90 p < .005 I > H

Video length F

(

473

)

= 89.01 p < .001 Short > long

Video length-by-group

interaction

(33, 462)

= 5.85 p < .001

I > H, in length1&2;

I > HB, in length3

Video length-by-pitch

type interaction

(11, 462)

= 2.65 p < .005

Strike > ball, in length3;

ball > strike, in length11&12

H = high-level pitcher and batter; I = intermediate pitcher and batter; HP = high-level pitcher; HB = high-level batter; IP =

intermediate pitcher; IB = intermediate batter.

3.2 Incorrect Response

The ANOVA detected a significant main effect of

type of pitch, in which balls were identified with

more mistakes than strikes. The main effect of group

was also significant, with HP showing higher

inaccuracy than IP (mean value of 36%, 32%, 28%,

and 29% for HP, HB, IP, and IB, respectively). A

significant pitch type-by-group interaction was

detected. Post-hoc analyses indicated that HB

showed higher inaccuracy than IB particularly for

balls. And, all groups except for IB had higher

inaccuracy for balls than for strikes (see Figure 5 top

panel). The main effect of length of video was also

significant. Moreover, video length-by-group

interaction was also significant, with advanced

players selectively showing higher rate than the

intermediate players in length 1, 3, and 5 (see Figure

5 middle panel). The interaction between the main

effect of pitch type and video length was also

significant. This interaction was due that the

inaccuracy for the strikes decreased as the video

became longer, while the tendency was opposite for

the balls (see Figure 5 bottom panel). The 3-way

interaction was not significant.

3.3 Uncertain Response

We found a significant main effect of group was also

significant, with advanced players, both pitchers and

batters, showing lower uncertain rate than the

intermediate players (mean value of 5%, 7%, 17%,

and 16% for HP, HB, IP, and IB, respectively). The

main effect of length of video was also found

significant, with higher rate for the short videos. The

interaction between the main effect of type of pitch

and the length of video was also found significant.

This interaction was due to a higher rate for strikes

than balls only in length 3 (see Figure 6 bottom

panel). Video length-by-group interaction was also

SeeingorDoing?-PitchRecognitionofBattersversusPitchers:APreliminaryReport

significant, for advanced players, no matter pitchers

or batters, showed lower rate than intermediate

players especially for short videos (length 1~3; see

Figure 6 bottom panel). The other two-way or three-

way interactions were not found significant, and

their p values were far from significant level.

Figure 4: Four groups’ average correct response for two

types of pitch (top panel) and for viewing different lengths

of pitch sequence (middle panel); all players average

correct response for strikes and for balls in different video

lengths (the bottom panel). *p < .05. Error bars indicate

standard errors.

Figure 5: Four groups’ average incorrect response for two

types of pitch (top panel) and for viewing different lengths

of pitch sequence (middle panel); and all players average

incorrect response for strikes and for balls in different

video lengths (the bottom panel). *p < .05. Error bars

indicate standard errors.

icSPORTS2014-InternationalCongressonSportSciencesResearchandTechnologySupport

Figure 6: Four groups’ average uncertain response after

viewing different lengths of pitch sequence (top panel);

and all players average uncertain response for strikes and

for balls in different video lengths (the bottom panel). *p

< .05. Error bars indicate standard errors.

4 DISCUSSION

In this study we tackled the question: in terms of

understanding an observed movement, is the

perceptuo-motor experience reacting to it more

important than the actual motor experience of the

observed movement? Baseball pitchers and players

could be considered as the best candidates to study

this topic since pitchers are trained to throw the pitch,

and batters are trained to intercept it. When they are

required to identify the type of the pitch, who can

show the higher accuracy? We recruited pitchers and

batters, with advanced and intermediate skill levels,

to study this topic.

We found that advanced batters showed the

highest accuracy among the four groups, particularly

when only very short pitching sequence was

presented to them for pitch identification. They were

significantly more accurate than the intermediate

players (both pitchers and batters). Advanced

pitchers were slightly less accurate than the

advanced batters (even though without reaching

statistical significance), and intermediate players

were the worst. Between intermediate pitchers and

batters, there was no significant difference. This

result was consistent with previous finding that

video-motor experts (goalkeepers) were more

accurate than the motor experts (kickers) (Tomeo et

al., 2012) and high level players were more accurate

and faster in their response than intermediate players

(Williams et al., 1999). However, the lack of

significant difference between the 18 advanced

batters and the 9 advanced pitchers could be due to

the unbalanced sample size. We expect to find a

statistical difference when we will recruit more

advanced pitchers in the future. As the indifference

between intermediate pitchers and intermediate

batters, we thought it is reasonable because they

have not developed such great difference since their

training experience was not so different as compared

to advanced players.

In addition, in consistency with the previous

studies, we found that our players generally showed

higher accuracy in identifying the strike pitches

when they could see longer sequence of the pitch

motion and the baseball’s trajectory (Paull &

Glencross, 1997). However, no such difference was

found for ball pitches. For explaining this result, we

would like to note that the final position where the

balls passed were not always very far from the

striking zone (in the perspective of the catcher). For

several balls, the difference to the striking zone was

only the size of half or one baseball. In this case,

these balls were easily to be identified as the strikes,

especially because we set experiment situation as in

a full count (2 strikes & 3 balls), 2 out, and full base

at the last inning.

Furthermore, intermediate players showed the

higher uncertain rate than the advanced players

particularly for short videos. This finding was of

importance because it revealed the requirement of

fast sports such as baseball, in which players have to

make a quick decision rather than staying uncertain.

Last, we found that advanced pitchers showed the

higher inaccuracy rate than the intermediate players.

However, we would not take it as they were more

erroneous than the intermediate players. Instead, we

interpreted it as the result of identifying a pitch as a

strike or a ball, rather than giving an uncertain

response.

SeeingorDoing?-PitchRecognitionofBattersversusPitchers:APreliminaryReport

In sum, at this moment our results suggested that

advanced players were generally more accurate than

intermediate players. Moreover, there seemed to the

difference between pitchers and batters, particularly

in high levels. Whether the difference is significant

or not has to be confirmed by recruiting more

pitchers and comparing their performance with

batters. This difference could be important because

it supports the notion that athletes (batters) can

better perform the task that they train better than the

athletes (pitchers) that acquire the actual motor

experience in doing the movement (Newell, 1986).

A further analysis of the players’ response time will

be done to better understand their ability in pitch

identification.

ACKNOWLEDGEMENTS

We would like to thank coach Shih-Kuei Huang

from Chinese Culture University, coach Wen-Nan

Liao from University of Taipei, coach Jung-Tang

Kung from National Taiwan Sports University, and

coach Po-Hsiu Lin from National Taiwan Normal

University for their help in recruiting baseball

players and their useful insights in experiment setup

and discussion.

REFERENCES

Aglioti, S. M., Cesari, P., Romani, M., & Urgesi, C., 2008.

Action anticipation and motor resonance in elite

basketball players. Nature Neuroscience, 11, pp.

1109–1116.

Newell, K. M., 1986. Constraints on the development of

coordination. In Wade M, Whiting HTA (editors):

Motor Development in Children: Aspects of

Coordination and Control (pp.341-360). Dordrecht,

Germany: Martinus Nijhoff.

Farrow, D. & Abernethy, B., 2003. Implicit perceptual

learning and the significance of chance comparisons:

A response to Jackson. Journal of Sports Sciences, 21,

pp. 511-513.

Farrow, D., & Abernethy, B., 2003. Do expertise and the

degree of perception-action coupling affect natural

anticipatory performance? Perception, 32, pp. 1127-

1139.

Paull, G., & Glencross, D., 1997. Expert perception and

decision-making in baseball. International Journal of

Sport Psychology, 28, pp. 35-56.

Takeuchi, T., & Inomata, K., 2009. Visual search

strategies and decision making in baseball batting.

Perceptual and Motor Skills, 108 (3), pp. 971–980.

Tomeo, E., Cesari, P., Aglioti S.M., & Urgesi, C., 2012.

Fooling the Kickers but not the Goalkeepers:

Behavioral and Neurophysiological Correlates of Fake

Action Detection in Soccer. Cerebral Cortex Advance

Access .

Williams, A. M., Davids, K., & Williams, J. G., 1999.

Visual perception and action in sport. London: E. & F.

N. Spon.

icSPORTS2014-InternationalCongressonSportSciencesResearchandTechnologySupport