Video Analysis Application to Assess the Reaction Time in an
ATP Tennis Tournament
Lucio Caprioli
1a
, Francesca Campoli
1b
, Saeid Edriss
1c
, Elvira Padua
2d
, Emilio Panichi
1e
,
Cristian Romagnoli
2f
, Giuseppe Annino
3g
and Vincenzo Bonaiuto
1h
1
Sports Engineering Lab., Dept. Industrial Eng. Univ. Rome Tor Vergata, Rome, Italy
2
Dept. of Human Science & Prom. of Quality of Life, San Raffaele Rome Univ., Rome, Italy
3
Dept. of Medicine Systems, Univ. Rome Tor Vergata, Rome, Italy
Keywords: Video Analysis, Tennis, Reaction Time, Visual Anticipation, ATP Tennis Tournament.
Abstract: 2D Video analysis is often used in tennis to analyze the players' technique or issues related to game tactics.
This paper applies video analysis to assess the reaction time in tennis matches. Fifteen subjects were examined
(26.20 ± 4.75 years old, weight 79.13 ± 5.67 kg, height 184.40 ± 5.30 cm, BMI 23.26 ± 1.19), all with an ATP
ranking between the #130 position and the #1066 position updated on the day of the sampling. The average
RT was 0.248 ± 0.07 s. The longer reaction times were recorded at the first stroke after the serve, while the
shorter were in defensive situations when the opponent was attacking or playing a volley, and the examined
player often anticipate by starting the mouvement even before the opponent's stroke. The reaction times of
high-level tennis players were found to be very short, often less than 120ms in defensive actions. These results
prompt us to consider the importance of kinetic perceptual skills such as reaction speed and anticipation in
tennis training.
1 INTRODUCTION
In tennis, many physical and mental (Casale, 2003;
Castellani Alberto et al., 1996; Cei Alberto, 2015;)
abilities are involved in the performance of high-level
athletes, who are called upon to solve complex motor
problems in a short-time through sprints and
explosive actions often performed in precarious
balance (Issurin, 2010; Matveev, 2001). Indeed, it
appears that, by the tennis players' performance
model, various coordinative and perceptive kinetic
abilities, such as reaction, anticipation, and
transformation, play a fundamental role (Fox et al.,
1993; Schönborn, 1999). Scanning signals in
advance, such as understanding the spot on the court
where the opposing tennis player is about to address
a
https://orcid.org/0009-0005-4049-5225
b
https://orcid.org/0009-0004-1342-5881
c
https://orcid.org/0009-0000-0224-8294
d
https://orcid.org/0000-0001-5227-2567
e
https://orcid.org/0009-0003-6591-1147
f
https://orcid.org/0000-0003-0904-634X
g
https://orcid.org/0000-0001-8578-6046
h
https://orcid.org/0000-0002-2328-4793
the ball, allow one to be in the right place and at the
right time to mechanically implement and promote
attack strategies (Singer & Negri, 1984). Reaction
ability is the coordinative ability to react quickly (as
quickly as possible) and correctly to given stimuli.
The reaction time (RT) is the latency period between
the occurrence of a stimulus and its response action
(Koch et al., 2018; Schmidt & Lee, 2019; Sternberg,
1969), and is given by five components:
a. production of a stimulus in the sensory receptor;
b. transmission of the stimulus to the Central
Nervous System;
c. processing (evaluation of the stimulus, choice of
response, and formation of the effector signal);
d. sending effector signal to the muscle;
e. muscle response (Janssen, 2015; Zelaznik, 1996).
Caprioli, L., Campoli, F., Edriss, S., Padua, E., Panichi, E., Romagnoli, C., Annino, G. and Bonaiuto, V.
Video Analysis Application to Assess the Reaction Time in an ATP Tennis Tournament.
DOI: 10.5220/0012184900003587
In Proceedings of the 11th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2023), pages 151-157
ISBN: 978-989-758-673-6; ISSN: 2184-3201
Copyright © 2023 by SCITEPRESS Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)
151
Simple and choice RT can be achieved in sports. In
the first case, the reaction is predefined, and a precise
and well-known motor plan corresponds to the
stimulus. Simple RT is often automated, and thus any
processing time is absent (it involves eliciting simple
conditioned reflexes to decrease the period between
stimulus and response). During choice RT, it is
necessary to perform a counter action appropriate to
the situation, which is unexpected, that is being
determined. The choice RT is always longer than the
simple RT because there is a cognitive-rational
processing phase between the perception of the
stimulus and the execution of the motor action. In
particular, it depends on the number of variables, that
is, the number of possible stimulus-response
alternatives. HICK's law describes the relationship
between the logarithm of the number of stimulus-
response options and RT; it shows that as the number
of stimulus-response pairs increases, RT increases
proportionally (Janssen, 2015; Proctor & Schneider,
2018). In reading the INPUT, being able to choose,
among many, the most helpful information is a more
relevant cognitive factor, which can allow for early
reading of what is about to happen. That is why in the
competition, the perception can occur before and
during the stimulus. Motor anticipation is the ability
to intuit a movement from the form of the action that
precedes it (Meinel et al., 1984). There are two forms
of motor anticipation, and the first is based on the
experience of prior game situations, such as an
opponent's tendency to serve wide from the left can
prompt the responding player to anticipate the move,
or visual anticipation, which is based on the visual
reading of the opponent's movement. Visual
anticipation is the ability to make accurate predictions
from partial or incomplete visual information
(Montagne et al., 2008). Examples of such
information might be the direction of a player's gaze
as they are about to shoot a penalty, the tennis player's
throwing of the ball on serve, the position of the feet,
or the movement of the racket and trunk before
impact (Shimizu et al., 2019). In other disciplines,
such as basketball, it has been found that athletes
more successfully predict the direction of free throws
to the basket with greater anticipation and accuracy
than other individuals (like coaches or sports
journalists and novices) (Aglioti et al., 2008). Visual
anticipation is fundamental in the motor response
process of the tennis player and is proportional to the
athlete's level of experience. In recent years, some
research has been conducted to understand what
factors most influence these perceptual processes in
sports. In studies performed with spatial occlusion of
the opponent's body parts (legs, trunk, arms, and
racket, etc.) worse accuracy and slower response
times were found in videos with occlusion of the ball
and trunk (Costa et al., 2023). Anticipation can also
be observed in an isolated act, such as correctly
predicting the speed and placement of a thrown object
(e.g., a ball) that allows the athlete to be in the right
place to repel or intercept it (Meinel et al., 1984;
Shimizu et al., 2019). During training, reaction and
anticipation skills can be trained with the help of new
technological aids (Senatore & Buzzelli, 2022).
Several research studies on reaction speed have been
conducted in the last two decades, and various off-
field measurement tools exist. However, there is not
as extensive a description of systems for measuring
reaction speed during competition, at least in tennis.
Video-based methods for testing RT have been used
successfully in some sports, such as karate, in
previous publications (Mudric et al., 2015).
2 METHODS
This paper uses video analysis to measure and
evaluate RT in tennis matches made by fifteen male
professional players during the ATP Challenger
"Castel del Monte" tournament in November 2022
(ATP Tour, 2022), played on indoor hard court
surface, with Artengo TB930 balls. The sample of
players had an age of 26.20 ± 4.77 years, weight of
79.13 ± 5.67 kg, height of 184.40 ± 5.30 cm, and body
mass index of 23.26 ± 1.19. All subjects had an ATP
ranking between the #130 and #1066 positions
updated on the day the measurements were done. The
analysis was conducted using video analysis software
(BIOMOVIE ERGO, 2023) on video acquired at 240
fps by a WOLFANG Action Camera that was placed
behind the central court. For the measurement of RT,
the time between the impact of the opposing player
and the first movement of the examined player was
measured (the first movement coincides with the
rotation of the shoulder line in the direction of the
displacement, of the foot descending from the split-
step or with a counter-movement of the contralateral
leg). The impact of the opponent (the starting point of
the ball) was taken as the zero point in the timeline
because it is the stimulus to which the player reacts.
The player's first movement is mainly evidenced by
the rotation of the inside foot in the direction of the
displacement on landing or in the moments
immediately following the split step (Fish, 1983).
icSPORTS 2023 - 11th International Conference on Sport Sciences Research and Technology Support
152
Figure 1: First movement lateral shift to the right.
Figure 2: First movement lateral shift.
This shift can occur in the direction of the
displacement or in the opposite direction to unbalance
the weight of the body in that direction (counter-
movement) with either the inside or outside foot
(Figure 1, Figure 2) (Vuong et al., 2022). Four main
types of step patterns are encountered: the Jab Step
(Figure1, Figure3d), Pivot Step (Figure2), Gravity
Step, and Counter Step (Figure3f). The Jab Step tends
to be the most common and most efficient in small
movements, and the Counter Step, on the other hand,
has been found by recent studies to be the most
effective in jerking to retrieve difficult balls (Vuong
et al., 2022).
Video Analysis Application to Assess the Reaction Time in an ATP Tennis Tournament
153
(
a
)
(
b
)
(c) (d)
(
e
)
(
f
)
Figure 3: (a), (c), and (e) flight phase during the split step of three different players; (b), (d), and (f) first movement.
(
a
)
(
b
)
(
c
)
Figure 4: First displacement detected by racket movement in response shot: a) split step, b) landing, c) first movement.
In other cases, the first movement is evidenced by the
rotation of the shoulders (UNIT-TURN) (Fish, 1983;
Groppel et al., 1986) or the opening of the racket
(Figure 4). In this case, the first movement (UNIT
TURN) is calculated when the racket begins to move
in the direction of displacement, i.e., to the right for
forehand and to the left for backhand in right-handers,
and vice versa for left-handers. The timer function of
the video analysis software was used to measure
reaction times. Twenty game situations were
analyzed (all by the same examiner), mixed between
rallies, serve response, first shot after serve,
icSPORTS 2023 - 11th International Conference on Sport Sciences Research and Technology Support
154
attack/defense, and volley play for each player. Single
factor ANOVA was used to compare the means
between the differtent game situations. Pearson's
correlation coefficient was used to compare the
correlation between reaction time and player ranking,
and the t-test was used to examine the statistical
significance of the difference between two groups of
players distinguished by ATP ranking.
3 RESULTS
Since the opponent's impact gives time zero in the
timeline, negative values were found when the player
anticipates the movement (Table 1). The average RT
was 0.248 ± 0.07 s, correlating perceptibly with
playing level. Specifically, the correlation index was
found to be a trivial 0.15. The ten tennis players with
ATP rankings between #130 and #400 had an average
RT of 0.246 ± 0.07 s, and the subjects with rankings
between #400 and #1066 had an average RT of 0.250
± 0.07 s (Figure 5).
Figure 5: Average RT in players with ATP rankings
between #130 to #400, and subjects with rankings between
#400 to #1066).
The longest RTs were recorded at the first stroke after
the serve, averaging 0.280 ± 0.05 s (0.278 ± 0.05 s for
players with better ranking and 0.282 ± 0.05 s for the
second group). The shortest in defensive situations
when the opponent was attacking or playing a volley
and the player examined anticipate by starting the
move even before the opponent's stroke at times. In
this case, they averagely reacted in 0.069 ± 0.18 s
(Figure 6).
Table 1: Measurement of RTs in the fifteen players examined. Text color refers to: black = baseline rally, red = return, green
= 1st shot after serve, light blue = opponent volley, purple = volley.
Pla
y
er ATP Challen
g
er Castel del Monte 2022 Measurements of reaction time
# Wei
g
ht Hei
g
ht A
e
R
ankin
g
ATP
B
est Ran
k
#1 #2 #3 #4 #5 #6 #7 #8
1 78 185 26 130 Sin
g
les 105 0.245 0.212 -0.250 -0.108 0.383 0.308 0.320 0.237
2 82 185 18 251 Sin
g
les 246 0.270 0.241 0.245 0.325 0.258 0.300 0.220 0.245
3 84 191 25 277 Sin
g
les 272 0.220 0.212 0.241 0.354 0.200 0.154 0.312 0.266
4 86 191 28 286 Sin
g
les 211 0.245 0.170 0.279 0.204 0.233 0.254 0.187 0.225
5 82 185 28 293 Doubles 271 0.204 0.270 0.204 0.275 0.220 0.279 0.250 0.313
6 81 185 29 295 Sin
g
les 262 0.262 0.270 0.295 0.250 0.241 0.200 0.175 0.270
7 73 188 25 323 Sin
g
les 220 0.200 0.254 0.337 0.366 0.220 0.258 0.208 0.366
8 78 185 20 334 Sin
g
les 261 0.325 0.337 0.375 0.175 0.362 0.345 0.233 0.325
9 87 185 35 423 Sin
g
les 49 0.270 0.229 0.275 0.266 0.362 0.212 0.270 0.254
10 77 188 26 442 Sin
g
les 259 0.295 0.316 0.254 0.258 0.200 0.262 0.245 0.320
11 80 188 23 452 Sin
g
les 313 0.304 0.245 0.287 0.258 0.308 0.250 0.287 0.350
12 86 185 36 479 Sin
g
les 33 0.245 0.129 0.237 0.250 0.270 0.266 0.179 0.300
13 70 175 24 575 Sin
g
les 557 0.229 0.250 0.166 0.325 0.175 0.220 0.195 0.212
14 70 175 26 602 Sin
g
les 536 0.245 0.225 0.220 0.212 0.216 0.229 0.295 0.204
15 73 175 24 1066 Sin
g
les 1027 0.262 0.275 0.229 0.208 0.260 0.237 0.245 0.200
Measurements of reaction time
#9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 Average Median Stand.
Dev.
0.250 0.333 0.266 0.287 0.195 0.262 0.258 0.183 0.262 0.270 0.208 0.287 0.220 0.260 0.15
0.241 0.283 0.258 0.208 0.333 0.258 0.270 0.341 0.125 0.250 0.245 0.241 0.258 0.254 0.05
0.166 0.183 0.104 0.233 0.200 0.225 0.270 0.208 0.137 0.354 0.187 0.204 0.222 0.210 0.07
0.229 0.254 0.225 0.258 0.233 0.250 0.245 0.191 0.212 0.204 0.237 0.287 0.231 0.233 0.03
0.220 0.300 0.191 0.325 0.183 0.241 0.258 0.233 0.329 0.108 0.295 0.337 0.252 0.254 0.06
0.237 0.200 0.220 0.283 0.270 0.195 0.233 0.241 0.183 0.237 0.245 0.300 0.240 0.241 0.04
0.250 0.370 0.216 0.266 0.241 0.237 0.412 0.187 0.250 0.270 0.283 0.266 0.273 0.256 0.06
0.229 0.345 0.254 0.275 0.250 0.216 0.308 0.262 0.150 0.250 0.158 0.258 0.272 0.260 0.07
0.245 0.266 0.200 0.241 0.333 0.358 0.237 0.304 0.295 0.212 0.366 0.283 0.274 0.268 0.05
0.279 0.291 0.279 0.325 0.283 0.250 0.295 0.225 0.250 0.220 0.325 0.170 0.267 0.271 0.04
0.270 0.258 0.158 -0.187 0.250 0.258 0.287 0.270 0.245 0.283 0.250 0.258 0.244 0.258 0.11
0.270 0.212 0.166 0.258 0.283 0.291 0.300 0.225 0.250 0.354 0.291 0.216 0.250 0.254 0.05
0.262 0.258 0.250 0.233 0.258 0.241 0.262 0.220 0.237 0.208 0.245 0.250 0.235 0.239 0.03
0.220 -0.060 0.000 0.233 0.300 0.191 0.320 0.187 0.254 0.254 0.329 0.325 0.220 0.227 0.10
0.208 0.245 0.275 0.233 0.254 0.254 0.245 0.291 0.370 0.241 0.291 0.362 0.259 0.250 0.04
Average:
_
0.248 0.250 0.07
Video Analysis Application to Assess the Reaction Time in an ATP Tennis Tournament
155
Figure 6: General average RT, in first-shot situations after
serve, defense, and return.
Although the best RTs occurred in defense situations,
the players’ move did not always occur in the correct
direction, because the anticipation has a relative
margin of error. Due to any fast wrist movement until
the impact, even a few milliseconds before,
everything can change. In the serve response phase,
RT was close to the average value of 0.245 ± 0.03 s.
The average RT in response in tennis players with
ATP rankings between #130 and #400 was 0.242 ±
0.03 s, while in the second group with rankings
between #400 and #1066 it was 0.248 ± 0.03 s
(Figure7).
Figure 7: Average RT of various game situations in players
with ATP rankings between #130 to #400 and in subjects
with rankings between #400 to #1066.
The largest percentage difference between the two
groups was found in the serve response situation. i.e.,
2.48%. In the average, a more negligible difference in
the order of 1.63% is shown (Figure 8). Although
some differences were found in the two ranking
groups, none of them was statistically significant.
While the differences reaction times found in various
game situations were found to be statistically
significant at the ANOVA p<0.001.
Figure 8: Percentage difference in various game situations
between players with ATP rankings between #130 to #400
and between #400 to #1066.
4 CONCLUSIONS
In this paper, 2D video analysis was successfully
applied to assess RTs in tennis matches. RTs of high-
level tennis players have been shown to be very short,
sometimes less than 120 ms, especially in defensive
actions, when the player often starts before the
opponent's attacking shot. This result prompts us to
consider the importance of perceptual kinetic skills,
such as reaction speed and anticipation, in tennis
training. No relevant correlations were found between
reaction time and ranking. It would be interesting to
enlarge the sample and involve elite top10 players in
future works. Further development of this study may
focus on techniques for training perceptual skills in
tennis players with measurement of the possible
improvement of RTs in matches, investigating also
the effectiveness of different types of first movement.
Another possibility of new work advancements could
involve 3D video capture using binocular cameras
(Zanela et al., 2022) and analysis techniques based on
artificial intelligence (Vincenzo Bonaiuto et al.,
2023).
ACKNOWLEDGEMENTS
The authors are grateful to the organization of the
ATP Challenger "Castel del Monte" in particular in
the person of Vincenzo Ormas, and the ATP referees
Carmelo Di Dio and Riccardo De Biase.
REFERENCES
Aglioti, S. M., Cesari, P., Romani, M., & Urgesi, C. (2008).
Action anticipation and motor resonance in elite
icSPORTS 2023 - 11th International Conference on Sport Sciences Research and Technology Support
156
basketball players. Nature Neuroscience, 11(9), 1109–
1116.
ATP Tour. (2022). ATP Challenger “Castel del Monte”
tournament in November 2022.
https://www.atptour.com/en/tournaments/andria/6911/
overview/
BIOMOVIE ERGO. (2023). [Computer software].
http://www.infolabmedia.eu/
Casale, L. (2003). PHYSICAL TRAINING FOR TENNIS
PLAYERS. Professional Tennis Registry.
Castellani Alberto, D’Aprile Angelo, & Tamorri Stefano.
(1996). Tennis Training. Società Stampa Sportiva.
Cei, A. (1998). Psicologia dello sport. Il mulino.
Cei Alberto. (2015). Allenarsi per vincere. Calzetti
Mariucci.
Costa, S., Berchicci, M., Bianco, V., Croce, P., Di Russo,
F., Quinzi, F., Bertollo, M., & Zappasodi, F. (2023).
Brain dynamics of visual anticipation during spatial
occlusion tasks in expert tennis players. Psychology of
Sport and Exercise, 65, 102335.
Cox, R. H. (1994). Sport psychology: Concepts and
applications. Madison. WI: Brown and Benchmark.
Fish, D. (1983). Footwork. Journal of Physical Education,
Recreation & Dance, 54(5), 27–30.
Fox, E. L., Bowers, R. W., Foss, M. L., & others. (1993).
The physiological basis for exercise and sport. Brown
& Benchmark.
Groppel, J. L., Conroy, B., & Hubb, E. (1986). SPORTS
PERFORMANCE SERIES: The mechanics of the
tennis forehand drive: Suggestions for training the
tennis player. Strength & Conditioning Journal, 8(5),
5–11.
Issurin, V. B. (2010). New horizons for the methodology
and physiology of training periodization. Sports
Medicine, 40, 189–206.
Janssen, S. (2015). The determinants of reaction times:
Influence of stimulus intensity [Master’s Thesis].
University of Waterloo.
Koch, I., Poljac, E., Müller, H., & Kiesel, A. (2018).
Cognitive structure, flexibility, and plasticity in human
multitasking—An integrative review of dual-task and
task-switching research. Psychological Bulletin,
144(6), 557.
Matveev, L. P. (2001). Teoría general del entrenamiento
deportivo. Editorial Paidotribo.
Meinel, K., Gulinelli, M., & Schnabel, G. (1984). Teoria
del movimento: Abbozzo di una teoria della motricità
sportiva sotto l’aspetto pedagogico. Società stampa
sportiva.
Montagne, G., Bastin, J., & Jacobs, D. (2008). What is
visual anticipation and how much does it rely on the
dorsal stream? International Journal of Sport
Psychology, 39(2), 149–156.
Mudric, M., Cuk, I., Nedeljkovic, A., Jovanovic, S., &
Jaric, S. (2015). Evaluation of Video-based method for
the measurement of reaction time in specific sport
situation. International Journal of Performance
Analysis in Sport, 15(3), 1077–1089.
Proctor, R. W., & Schneider, D. W. (2018). Hick’s law for
choice reaction time: A review. Quarterly Journal of
Experimental Psychology, 71(6), 1281–1299.
Schmidt, R., & Lee, T. (2019). Motor learning and
performance 6th edition with web study guide-loose-
leaf edition: From principles to application. Human
Kinetics Publishers.
Schönborn, R. (1999). Advanced techniques for
competitive tennis. (No Title).
Senatore, F., & Buzzelli, S. (2022). Development of
reaction times in young tennis players using the
SensoBuzz application. ITF Coaching & Sport Science
Review, 30(88), 34–38.
Shimizu, T., Hachiuma, R., Saito, H., Yoshikawa, T., &
Lee, C. (2019). Prediction of future shot direction using
pose and position of tennis player. Proceedings
Proceedings of the 2nd International Workshop on
Multimedia Content Analysis in Sports, 59–66.
Singer, R. N., & Negri, P. (1984). L’apprendimento delle
capacita’motorie. Societa’stampa sportiva.
Sternberg, S. (1969). The discovery of processing stages:
Extensions of Donders’ method. Acta Psychologica, 30,
276–315.
Vincenzo Bonaiuto, Giuseppe Annino, Francesca Campoli,
Lucio Caprioli, Saeid Edriss, Elvira Padua, Emilio
Panichi, Cristian Romagnoli, Naomi Romagnoli, &
Andrea Zanela. (2023). Assessing Sports Performances
Using an Artificial Intelligence-Driven System.
Vuong, J.-L., Edel, A., Voß, P., & Ferrauti, A. (2022).
Effectiveness and Kinematic Analysis of Initial Step
Patterns for Multidirectional Acceleration in Team and
Racquet Sports. Journal of Human Kinetics, 85(1), 13–
22.
Zanela, A., Schirinzi, T., Mercuri, N. B., Stefani, A.,
Romagnoli, C., Annino, G., Bonaiuto, V., & Cerroni, R.
(2022). Using a Video Device and a Deep Learning-
Based Pose Estimator to Assess Gait Impairment in
Neurodegenerative Related Disorders: A Pilot Study.
Applied Sciences, 12(9), 4642.
Zelaznik, H. N. (1996). Advances in motor learning and
control
. Human Kinetics.
Video Analysis Application to Assess the Reaction Time in an ATP Tennis Tournament
157