Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
Procedia Engineering 112 ( 2015 ) 540 – 545
1877-7058 © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University
doi: 10.1016/j.proeng.2015.07.240
ScienceDirect
Available online at www.sciencedirect.com
7th Asia-Pacific Congress on Sports Technology, APCST 2015
Movement variability of professional pool billiards players on
selected tasks
Philipp Kornfeinda*, Arnold Bacaa, Thomas Boindla, Andreas Kettlgrubera and Gerald
Gollnhuberb
aFaculty of Sport Science, University of Vienna, 1150 Vienna, Austria
bAustrian Pool-Billard Association, Carinthia, Austria
Abstract
Parameter values characterizing the motion of the cue during impact in pool billiards have been determined for selected shots. 20
elite players performed 18 predefined tasks comprising follow, draw, stop shots and breaks. 3D-kinematics were obtained using a
motion analysis system comprising 8 cameras operating at 250 Hz and a high speed camera capturing with 5000 Hz.
Longitudinal accelerations of the cue stick were recorded with 5 kHz using an accelerometer, mounted on the butt cap of the cue.
Coefficients of variation for the parameter values obtained range from 3.9 % (height of impact point of maximum follow shot) to
58.3 % (elevation angle of 10-ball break). The average cue stick motion is basically non-accelerated (-0.060 +/- 0.508 ms-2) at
ball impact for all of the tasks except for the breaks (3.918 +/- 0.164 ms-2). Despite the high number of DOF in the input
configuration the tested pool billiard players achieved very similar outcomes.
© 2015 The Authors. Published by Elsevier Ltd.
Peer-review under responsibility of the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University.
Keywords: Biomechanics; motion analysis; normative profile
1. Introduction
The physics behind the game of billiards is well understood [1]. Biomechanical aspects have, however, rarely
been considered. In pool billiards, where the focus of this particular study has been set, the object of the game is to
* Corresponding author.
E-mail address: philipp.kornfeind@univie.ac.at
© 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University
541
Philipp Kornfeind et al. / Procedia Engineering 112 ( 2015 ) 540 – 545
strike the white cue ball into the colored balls and then subsequently pocket them. Successful pool billiard players
are required to have efficient fine motor skills coupled with excellent repeatability in order to handle the different
challenges on the table [2]. When performing identically tasks, international top players show recognizable inter-
individual differences in their playing technique. These variations are notably related to the range of motion and the
sequence of segment movements, which lead to the question about the existence of a common playing technique.
Hence, the knowledge of subject independent movement parameters determining a successful shot technique is of
interest in order to derive general principles for coaches and youth players. It may be hypothesized that individual
movement strategies are necessary to compensate differences in anthropometrics and physiological properties.
Moreover, it may be assumed that due to the physical principles governing the impact between cue stick and ball,
top elite players show a similar resulting motion of the cue stick at impact. To the best knowledge of the authors,
kinematic data describing this motion have not been published so far. The objective of the study described within
this paper, which has been performed as part of a larger research project, is to examine this expectation.
2. Methods
All measurements were done during an international tournament in St. Johann (Austria Open, Salzburg,
Austria).A group of 20 top elite players (height: 180.2 ± 6.7 cm, mass: 80.3 ± 13.5 kg), all of them ranked under the
top 80 of the Euro-tour of the European Pocket Billiard Federation, 5 of them former world champions, performed
18 predefined tasks (A-R) comprising follow (A-F), draw (G-L), stop shots (M-N) and breaks (O-R). A follow shot
is (normally) done by hitting the cue ball above center. The cue ball stops momentarily then follows the object ball's
direction upon contact with the object ball. In a draw shot the cue ball stops momentarily then draws away from the
object ball towards the player upon contact with the object ball. A stop shot occurs when the cue ball comes to a
direct stop after making square contact with the object ball. The break shot is the first shot, which is used to separate
the object balls which have been racked together.
Each predefined position of the balls has been previously marked on the table’s cloth to guarantee identical
conditions for all participants (see Figures 1-6).
Fig. 1. Short follow shot. A: Follow 2 diamonds. B: Follow 6 Fig. 2. Long follow shot. D: Follow 2 diamonds (player is allowed
diamonds (including one rail). to touch the rail. E: Follow 4 diamonds (including one rail).
Fig. 3. Short draw shot. G: Draw 2 diamonds. H: Draw 4 Fig. 4. Long draw shot. J: Draw 1 diamond. K: Draw 3-4
diamonds (player is allowed to touch the rail). diamonds.
542 Philipp Kornfeind et al. / Procedia Engineering 112 ( 2015 ) 540 – 545
Fig. 5. Stop shot. M: Stop over 2 diamonds. N: Stop over 4 Fig. 6. 10-ball break. O: From the middle (breaker). P: From the
diamonds. Tolerance is a ball movement of 0.5 cm. left wing (breaker). Q: From the left wing (house cue). R: From
the middle (house cue). Maximum power with a straight hit.
All players had to succeed in one given task within up to three attempts and were allowed to use their own sports
equipment (shooting cue, break cue). 3D-kinematics were captured at 250 Hz using a motion analysis system
comprising 8 cameras (Vicon, Oxford). Longitudinal accelerations of the cue stick were recorded with 5 kHz using
an accelerometer, mounted on the butt cap of the cue. In addition, video imaging of the cue tip and the cue ball were
done at 5000 Hz with a high speed camera (IDT, Tallahassee) positioned perpendicular to the shot direction.
The following parameters were analyzed:
1. vimpact…Velocity of the cue at ball impact
2. φimpact…Elevation angle of the cue at ball impact
3. himpact…Height of impact point
4. timpact…Contact time between cue and ball during impact
5. aimpa…Acceleration of the cue at ball impact
vimp is defined as the 3D-velocity of a marker point attached near the top of the cue, himp is the height of the first
contact point between cue and ball, aimp as the acceleration into the longitudinal direction of the cue at impact. vimpact,
φimpact and himpact were determined using the motion analysis system, timpact from the high speed video recordings and
aimpact from the accelerometer data. It was assumed that all parameters follow a normal distribution. In order to find
out, if there is evidence against this assumption Shapiro-Wilk tests were performed (significance level: 0.05). Mean
values, standard deviations and coefficients of variation were calculated.
IBM SPSS (ver. 21) was used to test for normal distribution.
3. Results
Mean values and standard deviations (SD) for all parameters and shot types are depicted in Figures 7 to 11.
Coefficients of variation (CV) are given for the ratio-scaled variables (all but aimpact) and are written into the
respective bars. The asterisk (*) indicates a significant difference to normal distribution.
543
Philipp Kornfeind et al. / Procedia Engineering 112 ( 2015 ) 540 – 545
Fig. 7. Cue velocity at impact.
Fig. 8. Cue elevation angle at impact.
Fig. 9. Height of impact point. The horizontal line indicates the vertical central line of the cue ball.
544 Philipp Kornfeind et al. / Procedia Engineering 112 ( 2015 ) 540 – 545
Fig. 10. Contact time between cue and ball.
Fig. 11. Cue acceleration at impact.
It should be noted that the accuracy of the contact time is limited by the temporal resolution of the recording high
speed camera (5000 Hz).
CV for vimpact were lowest (8 % on average) whilst largest CV values were observed for φimpact (33 % on average).
For himpact CV was 9 % on average. The lowest timpact values were observed for the breaks. This well corresponds to
the highest vimpact values. The mean CV for timpact for all types of shots was 21 %. The average cue stick motion was
basically non-accelerated (-0.060 +/- 0.508 ms-2) at ball impact for all of the tasks except for the breaks (3.918 +/-
0.164 ms-2).
4. Discussion
Because of the high skill level of the participants, the collected data may serve as a reference for coaches and
athletes as well as for future studies. Main differences between the parameter values characterizing the shot types
can well be inferred from Figures 7 to 11. Short follow shots, for example, show larger values for himpact than long
follow shots, short draw shots lower values for vimpact than long draw shots and draw shots lower values for himpact
545
Philipp Kornfeind et al. / Procedia Engineering 112 ( 2015 ) 540 – 545
than follow shots. These differences follow physical principles of the tasks to be performed and will not be
discussed in more detail, here.
The values of vimpact for maximum follow and draw shots were almost identical (~4.5 m/s). Also himpact showed
quite similar values for all types of draw shots (~17.4 mm). The occurrence of these values may be caused by
physical boundary conditions during the interaction between cue stick and ball (e.g. static friction). Higher cue stick
velocities or lower vertical impact positions on the cue ball may result in a bad shot and are probably limiting factors
in pool billiards.
The distribution of timpact significantly differs from a normal distribution in almost all cases. This may partly be
reasoned by the insufficient timely resolution of the high-speed system used for recording the motions. Respective
results have therefore to be interpreted with care. Different properties of the material used for the cues’ tip might
also have influenced these results.
Significant deviations to the normal distribution of aimpact were mainly caused by one outlier. Whereas players
basically hit the ball with a non-accelerated cue stick for all the tasks but the breaks, one player showed a different
behavior in accelerating the cue before impact. aimpact was negative in most cases revealing a deceleration before
impact. Nevertheless, a non-accelerated motion of the cue stick during impact is clearly prevalent. Comparable
observations have been made for hitting devices in other sports, e.g. in tennis [3].
Concerning inter-individual variability of the players, vimpact and himpact show average CVs below 10 %. Although
CV for φimpact is comparatively high, the mean SD of this parameter for all types of shots is 1.2 deg. The distributions
of himpact and φimpact do not significantly differ from normal distributions for all types of shots, those of vimpact only in
cases, where SD is very small (see Figure 7). It may therefore be concluded that top elite players do not basically
differ in their cue stick motion at impact. The mean values of vimpact, himpact and φimpact might well characterize the
playing behavior of top elite players.
5. Conclusion
Despite the high number of DOF in the input configuration the tested pool billiard players achieved very similar
outcomes even though individual players selected a different strategy for some tasks.
References
[1] J. Jankunas, R.N. Zare, Why some pool shots are more difficult than others, Resonance 19(2) (2014) 116-122.
[2] P. Kornfeind, A. Baca, D. Kuzdas, G. Gollnhuber, M. Neumann, Investigation of subject independent movement parameters in professional
pool billiard, in: A. De Haan, C.J. De Ruiter, E. Tsolakidis (Eds.), Book of Abstracts of the 19th Annual Congress of the European College
of Sport Science, 2014, p. 81.
[3] H. Hatze, Forces and duration of impact, and grip tightness during the tennis stroke, Medicine and Science in Sports and Exercise, 8 (1976),
88-95.