Using 3D Motion Capture to Analyse Ice Hockey Shooting Technique

on Ice

Mikael Swarén

1,2,3

, Quirin Söhnlein

, Thomas Stöggl

and Glenn Björklund

5,6

Royal Institute of Technology, Department of Mechanics, Stockholm, Sweden

Swedish Olympic Academy, Stockholm, Sweden

Dalarna University, Department of Sport, Fitness and Medicine, Falun, Sweden

University of Salzburg, Department of Sport and Exercise Science, Salzburg, Austria

Swedish Sports Confederation, Stockholm, Sweden

Mid Sweden University,

Department of Health Sciences, Östersund, Sweden

Keywords: Slap Shot, One Timer, Kinematics.

Abstract: This study investigates the feasibility to use a passive marker motion capture system on ice to collect 3D

kinematics of slap shots and one timers. Kinematic data were collected within a volume of 40x15x2 m by 20

motion capture cameras at 300 Hz, a resolution of 12 megapixels and a mean residual for all cameras of

3.4±2.5 mm, at a distance of 11.6 m. Puck velocity, blade velocity, ice contact time and distance to the puck

were analysed for ten consecutive shots for each technique, for two professional ice hockey players. The total

mean puck velocity was 38.0 ± 2.7 m/s vs. 36.4 ± 1.0 m/s. (p=0.053), for one timers and slap shots

respectively. One player had higher puck velocity with one timers compared to slap shots 40.5 ± 1.0 m/s vs.

36.9 ± 1.0 m/s (p=0.001). Puck contact time was longer for slap shots than for one timers, 0.020 ± 0.002 s vs.

0.015 ± 0.002 s, (p<0.001). The motion capture system allowed continuous kinematic analyses of the puck

and blade velocities, ice contact times and detailed stance information. The results demonstrate the

possibilities to use motion capture systems to collect and analyse shooting kinematics on ice, in detail.

1 INTRODUCTION

Ice hockey is a physical demanding sport with high

intensity and tempo that obliges the players to have

excellent physical condition and high levels of

precision and technical skills. The shooting technique

is an essential keystone in ice hockey and the two

most commonly used shooting techniques in ice

hockey are the slap and the wrist shots. The slap shot

generates the highest puck velocities (33 - 36 m/s),

whereas wrist shots generate lower puck velocities

(20 - 28 m/s) but generally with higher accuracy

compared to slap shots (Michaud-Paquette, Magee,

Pearsall, & Turcotte, 2011; Villasenor, Turcotte, &

Pearsall, 2006; Worobets, Fairbairn, & Stefanyshyn,

2006).

The slap shot is generally divided into six

different phases: i) the backswing, ii) downswing, iii)

preloading, iv) loading, v) release and vi) the follow

through (B. Kays & Smith, 2014; Pearsall,

Montgomery, Rothsching, & Turcotte, 1999). During

preloading, the stick contacts the ice several inches

behind the puck to initiate the bending of the stick.

The puck is then impacted during the loading phase,

which increases the bending of the stick. The built-up

strain energy in the stick is then transferred to the

puck during the release phase.

A wrist shot uses less swing compared to a slap

shot and the player starts with the puck near the heel

of the blade. The puck is propelled forward as the

player does a forward sweeping motion while

creating shaft bending by applying a downward

pressure. The puck then rolls of the blade as the shaft

recoils (B. Kays & Smith, 2014; Pearsall et al., 1999).

Several studies have used high speed video or

motion capture to investigate these two different

shooting techniques as well as how the stick stiffness

affects the puck velocity (Frayne, Dean, & Jenkyn,

2015; Goktepe, Ozfidan, Karabork, & Korkusuz,

2010; Hannon, Michaud-Paquette, Pearsall, &

Turcotte, 2011; Michaud-Paquette et al., 2011;

Pearsall et al., 1999; Villasenor et al., 2006).

However, to the authors knowledge there are no

studies investigating a third shooting technique, the

“one timer”, which is similar to a slap shot but with

the difference that the player meets the puck with an

immediate slap shot, without touching or trying to

204

Swarén, M., Söhnlein, Q., Stöggl, T. and Björklund, G.

Using 3D Motion Capture to Analyse Ice Hockey Shooting Technique on Ice.

DOI: 10.5220/0008351602040208

In Proceedings of the 7th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2019), pages 204-208

ISBN: 978-989-758-383-4

control the puck first. The one timer is a shooting

technique that often initiates a rapid change of

direction of the opposing defenders and is frequently

used in power plays where at least one opposing

player is serving a penalty.

Previous ice hockey studies on ice, have mostly

used high speed cameras to capture kinematic data

while studies using 3D motion capture systems have

been executed off-ice by simulating real ice condition

by using e.g. synthetic ice (Frayne et al., 2015;

Goktepe et al., 2010; B. T. Kays & Smith, 2017;

Michaud-Paquette et al., 2011; Worobets et al.,

2006).

Stidwall et al. (2010) investigated skating kinetics

and kinematics on ice and on synthetic ice and found

that synthetic ice permits comparable mechanics for

on-ice forward skating. However, it is fair to assume

that the large difference in friction coefficients

between ice and synthetic ice (µ = 0.003 to 0.007 vs.

µ = 0.27) can affect the mechanics in technical

exercise such as shooting and agility drills. It can also

be hypothesised that shooting kinematics, stance and

distance to the puck at impact differ between slap

shots and one timers as one timers depends greatly on

timing and how the player positioning himself in

relation to the incoming puck, to achieve a good

shooting technique. Hence, there is need for a 3D

kinematic system that can be used on ice to acquire

realistic, detailed and accurate ice hockey shooting

analyses.

Upjohn et al. (2008) stated that passive marker

systems for motion capture have limited applications

for collecting data in a field setting due to the large

capturing volume needed for ice hockey and

problems with controlling ambient lighting.

However, new and better motion capture systems

have developed during the last years which allow data

capturing in poor lighting conditions, e.g. outside in

sunlight and/or in vast capturing volumes with large

quantities of unfavourable reflections. Shell et al.

(2017) and Renaud et al. (2017) used modern motion

capture systems to analyse ice hockey skate starts on

ice. Still, the use of a marker-based motion capture

system to collect 3D kinematics of different ice

hockey shooting techniques on ice has yet to be

tested.

The main purpose of the current study was to

investigate the feasibility to use a passive marker

motion capture system to collect 3D kinematics of

professional ice hockey players on ice while

performing shooting exercises. A secondary aim was

to investigate possible differences between the

conventional slap shot and the one timer shot with

regard to puck velocity, blade velocity and distance to

the puck at puck contact.

2 METHOD

Two professional ice hockey players were recruited

(player A and B) from a team in the Deutsche

Eishockey Liga (DEL) in Germany. Both players

were right hand shooters and used their own skates,

gloves, helmets and sticks. Soft reflective markers

were attached to the skates at the front tip, the heel

and at the position of the lateral malleolus. Six

markers were placed on each stick where four

markers were placed on the shaft, 0.30, 0.57, 1.1 and

1.5 m below the top of the shaft and two markers were

place on the heel and the tip at the backside of the

blade. For velocity measurements, markers were

placed on every puck. Each marker had a diameter of

15 mm and was attached by double-sided tape and

additionally fixed with tape around the base to avoid

movement of the markers.

The players performed a 15 min individual warm

up, including the various shooting techniques.

Following the warm up, to familiarise themselves

with the test setup, both players completed 20 shots.

During the data collection, each player performed 10

slap shots and 10 one timers, in opposite order. With

the one timers, the puck was hit directly from a pass,

coming from a 45 degrees angle to the right. All 10

shots in each series were executed in one sequence

without any rest between each shoot. For the

slapshots, all 10 pucks where lined up next to each

other and with the one timers, the next puck was

passed directly after the previous shot. The players

had a five minutes rest between the two different

shooting techniques.

Data were collected within a volume of 40 x 15 x

2 m by 20 Qualisys Uqus 7+ cameras (Qualisys AB,

Gothenburg, Sweden) at 300 Hz and with a resolution

of 12 mega pixels. Eight cameras were positioned on

high tripods, placed on isolated plates on the ice to

prevent the tripods from melting down into the ice.

The remaining 12 cameras were placed on top of the

safety glass surrounding the rink (Figure 1).

Protective padding was positioned behind the goal

to collect the pucks and to protect the boards and glass

from being hit by a puck and hence cause camera

movements. The measurement volume was calibrated

by using a hand-held T-wand, consisting of two

reflective markers at each end, with a known distance

between them. The orientation of the coordinate

system was performed by placing an L-frame at the

Using 3D Motion Capture to Analyse Ice Hockey Shooting Technique on Ice

205

Figure 1 A: Overview of half of the capture volume where the cameras are highlighted with circles. Photo courtesy of

Reimund Trost, Qualisys AB. B: 3D view of the capture volume where the player’s stick, markers on the skates and pucks

are visible.

decided origin. The mean residual for all cameras was

3.4 mm, s = 2.5 mm, at a distance of 11.6 m from each

camera.

2.1 Data Analysis

Stick blade velocity was determined by the resulting

speed of the marker, positioned at the tip of the blade.

The velocity data for all shots in each session were

synchronised concerning the time of puck contact.

The puck contact time was determined from the 3D

kinematic data.

Shooting stance was calculated as the distance, d,

between the left and right lateral malleolus markers,

in the xy-plane (the transverse plane) by equation 1.

 







 











 







(1)

where x

, y

, x

and y

are the xy-coordinates for the

left and right ankle markers respectively.

The distance from the player to the puck was

defined as the shortest distance between the puck

marker and the line between the left and right ankle

markers in the xy-plane. The line was constructed by

using the basic slope-intercept equation for a straight

line through two points, Eq. 2 and Eq. 3.

  



   





(2)

where m is calculated by:

 





 







 



(3)

The general form of a linear equation is:

      

(4)

and the distance D from a point P

) to this line

can be calculated by Eq. 5.

 







 



 









 



(5)

The data analyses from the two players and for

each condition were averaged separately for each

player. All data were checked for normality,

calculated with conventional procedures and

presented as means () and standard deviations (±

SD). The coefficient of variance was calculated as:

  





(6)

and presented in percent.

Significant effects between shooting techniques

were investigated with a two-sided paired Student t-

test and a unpaired two-sided Student t-test between

players, choosing an alpha level of 0.05 as criterion

for significance. Hedge’s g effect size (ES) was

further calculated and ranked as low (0.2), medium

(0.5) and high (0.8+) (Thomas, Salazar, & Landers,

1991) to determine the meaningfulness of the

differences between one timers and slap shots

between and within the same player.

3 RESULTS

Mean maximal puck velocity for slap shots versus

one timers for player A and B, respectively, were 36.0

± 0.9 m/s vs. 35.4 ± 1.0 m/s (p = 0.4, ES = 0.54) and

36.9 ± 1.0 m/s vs. 40.5 ± 1.0 m/s (p = 0.001, ES =

3.2). The total mean puck velocity showed a trend

towards lower velocity for slap shots compared to one

timers 36.4 ± 1.0 m/s vs. 38.0 ± 2.7 m/s, (p = 0.053,

ES = 0.79). Both players had a lower coefficient of

variation when shooting slap shots compared to one

timers 9% and 7% vs.14% and 34%.

icSPORTS 2019 - 7th International Conference on Sport Sciences Research and Technology Support

206

The blade velocities for slap shots and one timers

can be seen in Figure 2 A-B and the puck contact time

for slap shots was significantly longer compared to

one timers, 0.020 ± 0.002 s vs. 0.015 ± 0.002 s, (p <

0.001, ES = 2.2).

The perpendicular mean distance to the puck at

impact for slap shots versus one timers for player A

and B respectively, was 0.95 ± 0.05 m vs. 0.86 ± 0.08

m (p = 0.07, ES = 1.2) and 0.97 ± 0.09 m vs. 0.95 ±

0.07 m (p = 0.4, ES = 0.2).

Figure 2: Areas of variance around the mean (±SD)

illustrating instantaneous blade velocity for both players for

slap shots (A) and for one timers (B). The position for when

the blade contacts the ice is represented by the dashed line

a, and for the puck contact position by the dashed line b.

4 DISCUSSION

In contrast to previous studies, 3D kinematics were

successfully acquired from ice hockey shooting

exercises on ice by using passive markers and motion

capture cameras. The motion capture system allows a

higher sampling rate (>240 Hz) compared to

previously used measurement systems on ice which

have used different setups of video cameras with a

sampling rates of 60 – 200 Hz (Goktepe et al., 2010;

Lafontaine, 2007).

The results for the slap shot puck velocities and

contact times are in line with the results by previous

studies (B. T. Kays & Smith, 2017; Lomond,

Turcotte, & Pearsall, 2007; Villasenor et al., 2006;

Worobets et al., 2006), even if lower puck velocities

also have been reported by Wu et al. (2003) and Woo

et al. (2004). The characteristics of the blade velocity

are shown in Figure 1 A-B and are similar compared

to previous studies (B. T. Kays & Smith, 2017;

Villasenor et al., 2006).

It is interesting that the trend towards higher puck

velocity for one timers is achieved with shorter puck

contact time, compared to slap shots. This is most

likely accomplished by a higher pre-loading of the

stick that increases the stored potential energy in the

shaft and affects the puck velocity. However, future

studies are necessary to understand the mechanisms

which allow similar and even higher puck velocities

but with shorter puck contact times.

Neither blade velocity nor stance showed any

significant differences between the two shooting

techniques. However, it is interesting to note that both

players had a higher repeatability, indicated by a

lower CV (7% and 9% vs. 34% and 14%), for slap

shots compared to one timers, which Figure 1 A-B

also demonstrates. Even though no significant

difference, it is worth noticing that player B’s higher

puck velocity for one timers, is achieved without any

difference in stance, whereas player A had very

similar puck velocities but with a shorter stance with

one timers. Future studies should investigate if a

maintained stance is a performance indicator to

increase one timers shooting performance.

Only two players and two different shooting

techniques were analysed, making it difficult to

perform any reliable statistical analysis. Still, small

differences between the different shooting techniques

and players can be detected and analysed by using a

motion capture system, such as the one used in the

present study.

Even though all markers were placed with great

care, several came loose during the trials, most likely

due to high accelerations when the stick impacted the

ice and the puck. However, if using stronger adhesive

tape and additional reflective markers, the motion

capture setup in the present paper could be used to

analyse stick bending and the recoil effect when

shooting on ice. Stick bending and the recoil effect

were investigated by Villaseñor et al. (2006) but they

only used one high-speed video camera,

perpendicular to the sagittal plane, and synthetic ice.

Compared to using video cameras, the methodology

in the present paper appears to allow detailed 3D

kinematic analyses on ice, which entails the correct

friction coefficient between the blade and the ice.

Previous studies have only used synthetic ice to

simulate on-ice condition when analysing stick and

puck kinematics during slap, which plausible affects

Using 3D Motion Capture to Analyse Ice Hockey Shooting Technique on Ice

207

the results (Frayne et al., 2015; B. T. Kays & Smith,

2017; Lomond et al., 2007; Villasenor et al., 2006).

The ability to measure shooting kinematics

accurately on ice is vital in order to analyse ice

hockey biomechanics and to identify key

performance indicators for different shooting

techniques. The presented method is shown to

perform well for analysing ice hockey kinematics and

shooting performance on ice. However, the capturing

volume requires a large number or cameras and the

long time for setting up cameras should be taken in

consideration for future studies. Still, 3D motion

capture data enable numerous possibilities to

investigate ice hockey shooting kinematics on ice.

5 CONCLUSIONS

The use of a 3D motion capture system, as the one

used in the present study has the potential to enable

new possibilities to accurately analyse ice-hockey

shooting kinematics and shooting performance on ice.

REFERENCES

Frayne, R. J., Dean, R. B., & Jenkyn, T. R. (2015).

Improving ice hockey slap shot analysis using three-

dimensional optical motion capture: A pilot study

determining the effects of a novel grip tape on slap shot

performance. Proceedings of the Institution of

Mechanical Engineers, Part P: Journal of Sports

Engineering and Technology, 229(2), 136-144.

doi:10.1177/1754337114562096

Goktepe, A., Ozfidan, I., Karabork, H., & Korkusuz, F.

(2010). Elbow but not knee joint kinematics can be

assessed using photogrammetric methods during a non-

stationary slap shot in ice hockey. Scientific Research

and Essays, 5(3), 339-342.

Hannon, A., Michaud-Paquette, Y., Pearsall, D. J., &

Turcotte, R. A. (2011). Dynamic strain profile of the ice

hockey stick: comparisons of player calibre and stick

shaft stiffness. Sports Engineering, 14(2-4), 57-65.

doi:10.1007/s12283-011-0072-5

Kays, B., & Smith, L. (2014). Field Measurements of Ice

Hockey Stick Performance and Player Motion.

Procedia Engineering, 72, 563-568.

Kays, B. T., & Smith, L. V. (2017). Effect of ice hockey

stick stiffness on performance. Sports Engineering,

20(4), 245-254. doi:10.1007/s12283-017-0232-3

Lafontaine, D. (2007). Three-dimensional kinematics of the

knee and ankle joints for three consecutive push-offs

during ice hockey skating starts. Sports Biomech, 6(3),

391-406. doi:10.1080/14763140701491427

Lomond, K. V., Turcotte, R. A., & Pearsall, D. J. (2007).

Three-dimensional analysis of blade contact in an ice

hockey slap shot, in relation to player skill. Sports

Engineering, 10(2), 87-100. doi:10.1007/bf02844206

Michaud-Paquette, Y., Magee, P., Pearsall, D., & Turcotte,

R. (2011). Whole-body predictors of wrist shot

accuracy in ice hockey: a kinematic analysis. Sports

Biomech, 10(1), 12-21. doi:10.1080/14763141.2011.

557085

Pearsall, Montgomery, Rothsching, & Turcotte. (1999).

The influence of stick stiffness on the performance of

ice hockey slap shots. Sports Engineering, 2(1), 3-11.

doi:10.1046/j.1460-2687.1999.00018.x

Renaud, P. J., Robbins, S. M. K., Dixon, P. C., Shell, J. R.,

Turcotte, R. A., & Pearsall, D. J. (2017). Ice hockey

skate starts: a comparison of high and low calibre

skaters. Sports Engineering, 20(4), 255-266.

doi:10.1007/s12283-017-0227-0

Shell, J. R., Robbins, S. M. K., Dixon, P. C., Renaud, P. J.,

Turcotte, R. A., Wu, T., & Pearsall, D. J. (2017).

Skating start propulsion: three-dimensional kinematic

analysis of elite male and female ice hockey players.

Sports Biomech, 16(3), 313-324. doi:10.1080/14763

141.2017.1306095

Stidwill, T. J., Pearsall, D., & Turcotte, R. (2010).

Comparison of skating kinetics and kinematics on ice

and on a synthetic surface. Sports Biomech, 9(1), 57-64.

doi:10.1080/14763141003690237

Thomas, J. R., Salazar, W., & Landers, D. M. (1991). What

is missing in p less than .05? Effect size. Res Q Exerc

Sport, 62(3), 344-348. doi:10.1080/02701367.1991.

10608733

Upjohn, T., Turcotte, R., Pearsall, D. J., & Loh, J. (2008).

Three-dimensional kinematics of the lower limbs

during forward ice hockey skating. Sports Biomech,

7(2), 206-221. doi:10.1080/14763140701841621

Villasenor, A., Turcotte, R. A., & Pearsall, D. J. (2006).

Recoil effect of the ice hockey stick during a slap shot.

J Appl Biomech, 22(3), 202-211.

Woo, T., Loh, J., Turcotte, R., & Pearsall, D. (2004). The

Ice Hockey Slap Shot, Elite versus Recreational. Paper

presented at the 22nd International Symposium of

Biomechanics in Sports, Ottawa, Canada.

Worobets, J. T., Fairbairn, J. C., & Stefanyshyn, D. J.

(2006). The influence of shaft stiffness on potential

energy and puck speed during wrist and slap shots in ice

hockey. Sports Engineering, 9(4), 191-200. doi:10.

1007/BF02866057

Wu, T. C., Pearsall, D., Hodges, A., Turcotte, R., Lefebvre,

R., Montgomery, D., & Bateni, H. (2003). The

performance of the ice hockey slap and wrist shots: the

effects of stick construction and player skill. Sports

Engineering, 6(1), 31-39. doi:10.1007/BF02844158

icSPORTS 2019 - 7th International Conference on Sport Sciences Research and Technology Support

208