Content uploaded by Martin Valcke
Author content
All content in this area was uploaded by Martin Valcke on Jan 02, 2016
Content may be subject to copyright.
420
Learning opportunities in 3 on 3 versus 5 on 5 basketball
game play: An application of nonlinear pedagogy
ISABEL B. TALLIR*, RENAAT PHILIPPAERTS, MARTIN VALCKE**,
ELIANE MUSCH*, MATTHIEU LENOIR*
Ghent University, Belgium
(*)Department of Movement and Sports Sciences,
(**)Department of Educational Studies
This study investigates the differential learning opportunities in 5 on 5 versus
3 on 3 basketball game play. Video-analysis of the game performance of thirty bas-
ketball players (10-11 years) resulted in significantly higher scores on all game per-
formance components (GPC’s: cognitive decision making component (DM), motor
skill execution efficiency (MSEfficiency) and motor skill execution efficacy (MSEf-
ficacy) component), indicating more learning opportunities during 3 on 3 game
play. The actual game performance level, showed only significantly higher scores for
the percentage of positive decisions for cutting actions in the 5 on 5 condition.
Future research is needed to indicate to what extent learning results are easier or
faster attained when using small sided games, based on the nonlinear pedagogy
framework, and second which is the optimal game play situation to assess game per-
formance, and this for players of a different game performance level or for different
stages.
KEY WORDS: Game Performance, Motor skill acquisition, Small-sided games,
Task constraints.
Introduction
The constraints-led dynamical system approach, building further on the
information-movement coupling principle, is promoted as a framework for
understanding the process of motor skill acquisition in sport and exercise
(Davids, Button, & Bennett., 2008; Araújo, Davids, Bennett, Button, &
Chapman, 2004). This constraints-led approach views influential factors
within the learning environment as constraints that guide the acquisition of
movement coordination and control (Newell, 1986, 1996). Modelled as
Correspondence to: Isabel B. Tallir, Department of Movement and Sports Sciences, Ghent
University, Watersportlaan 2, B-9000 Gent, Belgium. E-mail: isabel.tallir@ugent.be
Int. J. Sport Psychol., 2012; 43: 420-437
doi: 10.7352/IJSP 2012.43.420
dynamical systems, team sports display characteristics of complexity due to
the potential for interactions that emerge between performers over time. The
decisions and the actions of a single player become dependent on what
neighboring players (either teammates or opponents) are doing and on the
immediate events prior to a sub-phase emerging. This contextual depen-
dence in behaviour signifies that player interactions are not deterministic
(entirely predictable), nor completely random (entirely variable). Conse-
quently, invasion game play can be characterized as a nonlinear dynamical
system (Chow et al., 2006) since it is composed of many interacting parts
(e.g., players, ball, referees, court dimensions) (e.g., Gréhaigne, Bouthier, &
David, 1997; McGarry et al., 2002).
Based on the influences of motor learning frameworks on motor skill
acquisition and decision-making nonlinear pedagogy was presented by
Chow and colleagues (2006) as a methodology for games teaching, capturing
how phenomena such as movement variability, self-organization, emergent
decision making, and symmetry-breaking occur as a consequence of interac-
tions between agent-agent and agent-environment constraints. Nonlinear
pedagogy highlights the interactive role that key constraints (i.e., performer,
task and environmental) play in learning contexts to shape emergent move-
ment behaviors which arise during practice (Chow et al., 2006, 2007; Davids,
et al., 2008). Since sports are dynamic non-linear systems and as such, sport
skill learning should also be non- linear (Chow et al., 2007). This can be
achieved by pedagogical manipulation of three elements; game constraints
(such as changing game rules), performer constraints (such as how the per-
former is permitted to move), and environmental constraints (such as chang-
ing spaces or equipment). Chow et al. (2009) noted that one of the strengths
of tactical instructional approaches, such as Teaching Games for Under-
standing (TGfU) (Griffin & Butler, 2005), is that it enables learners to prac-
tice in a managed environment with all key information sources present, so
that perceptual and action processes in learners can become tightly coupled
during practice. The lack of variability in the traditional “closed” training
drills reduces the opportunity for players to learn how to adapt movement
solutions to changing environmental demands. While closed drills may pro-
vide a simplified environment that allows players to execute skills with
increased precision and reduced error, modified games provide players with
an opportunity to better calibrate the execution of the skill with relevant and
reliable perceptual variables, such as the locations of defenders relative to
teammates (Passos, Araújo, Davids, & Shuttleworth, 2008). Therefore,
unlike the traditional approaches, tactical instructional approaches advocate
a student-centered emphasis for learning tactics and skills in modified games
421
(e.g., Griffin, Butler, Lombardo & Nastasi, 2003). This is an perfect illustra-
tion of the motor learning principle of information-movement coupling
which proposes that the motor learning of the skills needed in games should
involve the process of task simplification, rather than traditional methods of
part-task decomposition (Davids et al., 2008). Task simplification refers to a
process whereby scaled-down versions of tasks are created in practice and
performed by learners to simplify the process of information pick-up and
coupling to movement patterns. In these scaled-down modified game con-
texts, important information-movement links are maintained in practice and
are not disrupted in practice task design. This implies that the information
available to be actively explored by players during practice must represent
the same task and environmental constraints that exist during performance.
Otherwise, the information-movement couplings that emerge during prac-
tice will be attuned to perceptual variables different from those available
during performance or in other words decision-making practices should be
based on performer-environment interactions rather than a traditional per-
former- or task-centered approach (Chow et al., 2009).
A logical consequence of the evolution to more student-centered
instructional approaches, such as for example Teaching Games for Under-
standing (TGfU) and application of motor learning principles for games
learning was a comparable shift in research focus. Whereas the initial studies
were process/product studies, the focus of the more recent research ques-
tions is more about the knowledge construction of the individual learners in
relation to their learning environment (Richard & Wallian, 2005).
The purpose of the present study was to investigate if players experience
more and/or different learning opportunities while playing 3 on 3 half-court
versus 5 on 5 full court basketball game play or in other words if the manip-
ulation of the task constraints (here number of players) results in a differen-
tial number and type of learning opportunities.
Methods
PARTICIPANTS
Four junior competition basketball teams or 42 players of 11-12 years old participated in
this study. Game play of 30 players (23 boys, 7 girls) (mean age = 11.08 ± 0.55 years) was
analysed. Each team was divided in two subgroups (team A and team B) to play against each
other during the 5 on 5 (and 3 on 3) games (see table 1). These subgroups remained the same
for the two assessment moments. To assure an equal distribution of play level in both sub-
groups the coach of each team subjectively ranked the players from 10 to 5, where 10 indi-
422
423
cated the player with the highest play level and 5 was used to indicated the player with the low-
est play level (see table 1: (player)level). During the 5 on 5 and the 3 on 3 game play the players
always marked an opponent of an comparable play level of the other subgroup of their own
team. Only one-on-one defence was permitted, between the pairs of attackers and defenders
as established previously by the coaches. This last rule, one-on-one defence, was important in
that the matching of opposing players enabled differences in technical ability to be more eas-
ily controlled. Of each team only the players with highest rankings (numbers 1 to 30) were
included in data-analysis (see table I)
PROCEDURE
Data collection. The study consisted of two assessment moments orga-
nized on two different testing days during which the players played respec-
tively 5 on 5 and 3 on 3 basketball game play sessions of 5 minutes. Both
assessment moments were organized for each team on a day players had no
basketball training or were involved in a competition match. At the begin-
ning of each testing day, players started with a warming-up of 10 minutes.
Players participated in four 5 on 5 game play session of 5 minutes, with 5
minutes of rest between the game play sessions. This procedure was repli-
TABLE I
Team Compositions Of The Game Play Sessions During The 3 On 3 And The 5 On 5
Assessment Moments
Assessment moment 1 Assessment moment 2
(5 on 5) (3 on 3)
Team Session Playerslevel Playerslevel Playerslevel Playerslevel
team A team B team A team B
11(1)10-(2)8-(3)7-(4)6-(31)5(5)9-(6)8-(7)8-(8)6-(32)51-2-3 5-6-7
2 (1)10-(2)8-(3)7-(4)6-(31)5(5)9-(6)8-(7)8-(8)6-(32)51-31-4 5-32-8
3 (1)10-(2)8-(3)7-(4)6-(31)5(5)9-(6)8-(7)8-(8)6-(32)52-3-4 6-7-8
4 (1)10-(2)8-(3)7-(4)6-(31)5(5)9-(6)8-(7)8-(8)6-(32)51-2-31 5-6-32
21(9)10-(10)8-(11)6-(12)8-(33)7(13)9-(14)7-(15)6-(16)8-(35)79-10-11 13-14-15
2 (9)10-(10)8-(11)6-(12)8-(34)7(13)9-(14)7-(15)6-(16)8-(35)734-12-33 13-16-35
3 (9)10-(10)8-(11)6-(12)8-(33)7(13)9-(14)7-(15)6-(16)8-(35)710-11-12 14-15-16
4 (9)10-(10)8-(11)6-(12)8-(34)7(13)9-(14)7-(15)6-(16)8-(35)79-10-33 13-14-35
31(17)10-(18)9-(19)8-(36)9-(37)6(20)10-(21)8-(22)7-(38)8-(39)517-18-40 20-21-38
2 (17)10-(18)9-(19)8-(36)9-(37)6(20)10-(21)8-(22)7-(38)8(40)417-19-37 20-22-39
3 (17)10-(18)9-(19)8-(36)9-(37)6(20)10-(21)8-(22)7-( 38)8-(39)518-19-40 21-22-38
4 (17)10-(18)9-(19)8-(36)9-(37)6(20)10-(21)8-(22)7-(38)8-(40)417-18-37 20-22-39
41(23)8-(24)10-(25)8-(26)9-(41)6(27)9-(28)10-(29)7-(30)9-(42)523-24-25 27-28-29
2 (23)8-(24)10-(25)8-(26)9-(41)6(27)9-(28)10-(29)7-(30)9-(42)523-26-41 27-30-42
3 (23)8-(24)10-(25)8-(26)9-(41)6(27)9-(28)10-(29)7-(30)9-(42)524-25-41 28-29-42
4 (23)8-(24)10-(25)8-(26)9-(41)6(27)9-(28)10-(29)7-(30)9-(42)523-24-26 27-28-30
Note 1. Of each team only the players with highest ranking (numbers 1 to 30) were included in data-analysis.
Note 2. Italic numbers were not included in the game performance analysis.
cated for the 3 on 3 game play sessions which occurred one week later. Here
the players played two or three game play sessions of 5 minutes. Players were
told to play 3 on 3 (or 5 on 5) whereby they had to cover the same player (of
an equal game performance level) during each 5 minutes game session (see
table 1). All game sessions were videotaped to allow post hoc analysis of
game performance. To put the entire playing area on the screen 4 digital
video cameras (Sony Handycam DCR-HC20E PAL) were used. The 4 cam-
eras were positioned at 3 meter height at both sides of the center line and
were pointed in the direction of the basketball goal.
The dimensions of the 5 on 5 basketball court was 14m x 26m, whereas
for the 3 on 3 games the court dimensions were 14m x 13m. This implies a
relative playing area per individual player of 36.4 m2(= 364m2 /10) in the 5
on 5 game play sessions versus 30.3 m2(= 182 m2/ 6) in the 3 on 3 game play
sessions. In both game play situations two basketball goals were used.
Players’ heart rates were recorded every 5 seconds during all the game
play sessions by means of a heart rate monitor (Polar type RS400). This
allowed direct measurement of a physiological response to both game play
situations. Heart rates monitored during the rest period were not analysed.
Coding instrument. The coding instrument of Tallir et al. (2007) was
used to asses every observable offensive action, both on- and off-the-ball. For
every offensive action three game performance components (GPC’s) were
assessed, namely a decision-making component (DM), a motor skill execu-
tion efficiency component (MSEfficiency) and a motor skill execution effi-
cacy component (MSEfficacy). Consequently, each action resulted in a posi-
tive or a negative score for each of the three game performance components.
The DM component was coded positive/negative when a player took a cor-
rect/incorrect decision in a particular game situation. For the MSEfficiency
component the different aspects of the executed skill were coded positive /
negative when the skill was executed technically correct/incorrect, respec-
tively. The MSEfficacy component was coded positive/negative if the action
had a successful / unsuccessful outcome. The entire coding instrument
included 95 categories and is presented in appendix 1. Content validity of
this instrument was assured in Tallir et al. (2007). The test-retest reliability
coefficients of the instrument in this former study were above 0.95 and the
Cronbach’s inter-observer reliability coefficients were higher than 0.73 for
the three components.
An advantage of this coding instrument (see figure) compared to the
Game Performance Assessment Instrument (GPAI; Oslin, Mitchell, & Grif-
fin, 1998) is that it made it possible to determine, not only the total number
424
425
of learning opportunities, but also the proportion, of both the positive and
the negative scores, of the four game performance categories (in case: con-
trol, scoring and creating scoring opportunities (CSO: dribble, pass, cut-
ting), setting up an attack) because this coding instrument identified every
off- and on the ball-action (Tallir et al. 2007) (see appendix 1).
Data-analysis. The software “Catmovie” was used to analyse the offen-
sive (on and off-the-ball) game performance. Catmovie (http://www-cam-
pus.uni-r.de/edu1/catmovie/) is a software program that allowed coding all
the categories of the GPC (DM, MSEfficiency and MSEfficacy) for every
game play action. This software program made it possible to assess all the
necessary components and their categories on screen while watching a con-
tinuously repeated interval of 5 seconds of the game session before moving
on to the next interval. Data-analysis was done in this manner for one player
at a time during the entire game play session of 5 minutes. Since the number
of 3 on 3 game play sessions (2 or 3 sessions for every player) was not identi-
cal to the number of 5 on 5 game play sessions (4 sessions for every player)
the average numbers of the three GPC (DM, MSEfficiency and MSEfficacy)
per game play session were calculated. Further, as a reflection of game per-
formance, the average proportion of positive scores for the three GPC and
the four game performance categories per game play session will be reported
in percentages to indicate the distribution of positive and negative scores.
Only the positive percentages are reported since the negative percentages are
the remaining percentages to 100%.
Statistical analysis. Data were analysed using the Statistical Package for
the Social Sciences, version 19.0 (IBM Corporation, New York, USA). Level
of significance for all statistical analyses was set at a .05 alpha level and the
size of effect was provided by partial Eta squared (ηp2). Post-hoc LSD test
were used to further analyse the main effects.
A repeated measures ANOVA, with one independent (condition: 3-3 vs.
5-5) was used to compare, firstly the differences in the average numbers per
session of the three GPC (DM, MSEfficiency, MSEfficacy), secondly the dif-
ferences in the average positive proportional values of the GPC (DM, MSEf-
ficiency, MSEfficacy) per session.
In a next step the three GPC were analyses in more detail. For the three
GPC (DM, MSEfficiency, MSEfficacy) a RMANOVA with repeated mea-
sures of the average numbers per session of the game performance categories
(in case: control, scoring and CSO (dribbling, passing, cutting actions), set-
ting up an attack) was executed, followed by a RMANOVA with repeated
426
measures of the positive proportional values of the game performance cate-
gories (in case: control, scoring and CSO (dribbling, passing, cutting
actions), setting up an attack).
To compare the physiological response of the players to both game play
situations the heart rate results during the 3 on 3 and 5 on 5 game play ses-
sions a RMANOVA was used to compare the average heart rates of the play-
ers in both game play situations.
Results
For the three GPC a main effect condition was found F(3,27)=78,69, p
<.001, ηp2= .90. Significantly higher averages numbers were found during
the 3 on 3 condition compared to the scores during the 5 on 5 condition (see
table II). Effect sizes of the average numbers of the three GPC’s were.86, .89
and .88 for DM, MSEfficiency and MSEfficay respectively.
Comparison of the average positive proportional values of the average
numbers of the three GPC indicated that for every GPC the positive pro-
portional values were not significantly different in both game conditions
F(3,27)= .74, p = ns, ηp2= .08 (see table III).
The former results were analyzed more detailed at the level of the four
game performance categories (in case: control, scoring and CSO, setting up
an attack). The average numbers per session of the decision making compo-
TABLE II
Average Numbers Per Session Of The Three Game Performance Components (GPC’s: DM, MSEfficiency,
MSEfficacy) During 5 On 5 And 3 On 3 Game Play Together With The P-Values And Effect Sizes (
Η
p2)
5on5 3on3 Fp
ηp2
DM 29.88 (9.49) 50.35 (8.18) 175.22 .001 .858
MSEfficiency 92.23 (40.87) 164.87 (38.81) 239.82 .001 .892
MSEfficacy 30.20 (11.33) 52.09 (10.27) 203.53 .001 .875
Note. Standard deviation in brackets.
TABLE III
Average Positive Proportional Values (In Percentages) Of The Three Game Performance Components
(GPC’s: DM, MSEfficiency, MSEfficacy) During 5 On 5 And 3 On 3 Game Play
5on5 3on3
DM 80.15 (10.45) 79.78 (12.76)
MSEfficiency 85.83 (5.83) 83.97 (12.71)
MSEfficacy 78.91 (7.03) 76.75 (13.01)
Note. Standard deviation in brackets.
427
nents of the game performance categories were significantly different F(6,24)=
60,06, p <.001, ηp2= .94 between both game conditions. For all game perfor-
mance categories, except for setting up an attack significantly higher scores
were found in the 3 on 3 condition compared to the scores during the 5 on 5
condition (see table IV). For the game performance category setting up an
attack scores were significantly higher in the 5 on 5 condition compared to
the 3 on 3 condition for the DM and the MSEfficacy component. For the
MSEfficiency higher scores were found for setting up an attack in the 5 on 5
condition. Effect sizes for the game performance categories varied between
.36 and .86.
TABLE IV
Average Numbers Per Session Of The Three Game Performance Components (GPC’s: DM, Msefficiency,
Msefficacy) Of The Game Performance Categories (In Case: Control, Scoring And Creating Scoring
Opportunities (CSO: Dribbling, Passing, Cutting Actions), Setting Up An Attack) During 5 On 5 And 3
On 3 Game Play Together With The P-Values And Effect Sizes (
Η
p2)
5on5 3on3 Fp
ηp2
Decision-making component
Control [cat 09-10]5.43 (2.58) 9.85 (2.73) 142.15 .001 .83
Scoring [cat 11-12]1.65 (1.07) 4.30 (1.74) 107.54 .001 .79
Creating scoring Dribbling [cat 13-16]4.18 (2.32) 7.19 (2.34) 88.09 .001 .75
opportunities (CSO) Passing [cat 14-17]5.23 (2.66) 8.93 (2.55) 66.11 .001 .70
Cutting [cat 15-18]2.75 (1.54) 4.21 (1.43) 31.11 .001 .52
Setting up an attack [cat 19-20-21-22-23-24]5.29 (1.52) 3.58 (.96) 24.69 .001 .46
Motor skill efficiency
Control [cat 25-26-27-28-29-30-31]4.71 (2.22) 7.07 (1.87) 58.78 .001 .67
Scoring [cat 35-36-37-38-39-40-41-42-43-44-45-46]1.66 (1.07) 4.31 (1.77) 104.39 .001 .78
Creating scoring Dribbling 3.21 (1.89) 5.55 (2.07) 78.08 .001 .73
opportunities (CSO) [cat 47-48-49-50-51-52-
53-54-55-56]
Passing 3.92 (2.10) 6.34 (1.81) 49.45 .001 .63
[cat 57-58-59-60-61-62]
Catching 5.49 (2.44) 9.92 (2.51) 162.73 .001 .85
[cat 63-64-65-66-67-68]
Cutting 2.47 (1.51) 3.89 (1.57) 25.49 .001 .47
[cat 69-70-71-72-73-74]
Setting up an attack [cat 75-76-77-78]3.02 (.68) 4.34 (1.38) 23.02 .001 .44
Motor skill efficacy
Control [cat 79-80]5.24 (2.58) 9.69 (2.69) 174.03 .001 .86
Scoring [cat 81-82]1.68 (1.07) 4.24 (1.73) 90.14 .001 .76
Creating scoring Dribbling [cat 83-84]3.76 (2.20) 6.52 (2.54) 78.15 .001 .73
opportunities (CSO) Passing [cat 91-92-93]9.88 (4.52) 17.336 (4.38) 125.82 .001 .81
Cutting .95 (.63) 1.45 (.60) 16.24 .001 .36
[cat 85-86-87-88-89-90]
Setting up an attack [cat 94-95]8.72 (2.75) 6.03 (1.38) 23.63 .001 .45
Note. Standard deviation in ( ) brackets. Detailed description of the game performance categories can be
found in the appendix and is noted in []brackets.
428
In a next step the focus is on the positive proportional values of the game
performance categories. Comparison of the average positive proportional
values of the decision making components of the game performance cate-
gories (in case: control, scoring CSO (dribbling, passing, cutting actions),
setting up an attack) resulted in a main effect of condition F(6,24)= 3.34, p
<.05, ηp2= .46 (see table V). Significantly higher averages positive propor-
tional values of the decision making component during the 3 on 3 condition
were only found for the game performance category CSO cutting actions
compared to the 5 on 5 condition F(1,29)= 12,59, p <.001, ηp2= .30 (see table
V). Identical analyses were executed on the average positive proportional
values of the motor skill efficiency and efficacy components of the game per-
formance categories. Results for the motor skill efficiency component F(6,24)=
.36, p = ns, ηp2= .08 and for the motor skill efficacy component F(7,23)= 2.17,
p = ns., ηp2= .40 showed no significant differences between both conditions
when comparing the average positive proportional values of the both com-
ponents of the game performance categories (see table 5).
TABLE V
Average Positive Proportional Values Of The Three Game Performance Components (GPC’s: DM,
Msefficiency, Msefficacy) Of The Game Performance Categories (In Case: Control, Scoring And Creating
Scoring Opportunities (CSO: Dribbling, Passing, Cutting Actions), Setting Up An Attack) Together
With The P-Values And Effect Sizes (
Η
p2)
5on5 3on3 Fp
ηp2
Decision-making component
Control 87.37(13.61) 85.88(14.68) .33 ns .011
Scoring 76.79(29.93) 86.66(15.15) 2.33 ns .074
Creating scoring Dribbling 73.62(25.63) 77.84(19.02) .96 ns .032
opportunities (CSO) Passing 85.70(10.60) 85.32(11.00) .03 ns .001
Cutting 57.01(27.08) 43.08(21.94) 12.59 .001 .303
Setting up an attack 81.78(14.10) 82.84(13.69) 0.14 .ns .005
Motor skill efficiency
Control 92.85 (6.22) 90.09 (8.63) 3.96 ns .120
Scoring 88.19 (2.75) 95.79 (3.79) 2.85 ns .090
Creating scoring Dribbling 91.74 (17.70) 90.37 (17.59) 0.09 ns .003
opportunities (CSO) Passing 86.92 (7.89) 85.82 (10.58) .27 ns .009
Catching 90.26 (8.68) 91.10 (7.20) .45 ns .015
Cutting 41.15 (25.08) 32.59 (21.44) 6.69 ns .188
Setting up an attack 69.26 (19.18) 72.24 (18.61) .92 ns .031
Motor skill efficacy
Control 99.44 (2.12) 98.86 (3.31) 1.36 ns .045
Scoring 48.51 (27.07) 45.95 (21.28) .14 ns .005
Creating scoring Dribbling 90.09 (20.12) 92.46(18.24) .21 ns .007
opportunities (CSO) Passing 91.57 (5.21) 92.66 (4.43) .72 ns .024
Cutting 29.34 (27.04) 30.31 (24.01) .06 ns .002
Setting up an attack 72.30 (19.81) 73.39 (18.79) .12 ns .004
Note. Standard deviation in brackets.
Comparison of the average heart rates in both game play conditions
F(1,28)= 7,0, p <.05, ηp2= .20 resulted in significantly higher heart rates in the
5 on 5 condition (M = 185.07 ± 10,76) compared to the 3 on 3 condition (M
= 180.80 ± 12,10).
Discussion
A crucial prerequisite for (motor) learning to occur is exercising, and more
specifically the frequent repetition of a skill (Newell, 1996). For games teach-
ing, this means it is crucial to select appropriate learning activities (Silverman,
2003) that reflect players’ developmental readiness and that allow certain
aspects of the game to come into play more often so that players get more in-
game repetitions on key tactics and motor skills (Metzler, 2000). So, in games
teaching the learning activities should contain constraints that are representa-
tive of those that players will face during game play (Renshaw et al., 2010) for
an integrated development of both components of game performance. In the
absence of a theoretical framework a similar statement was already made by
Turner and Martinek (1992) two decades earlier, namely that through game
play more adaptable schemata for motor skills may be developed.
The general purpose of this study was therefore to compare the amount
of learning opportunities in the 3 on 3 modified game compared to the 5 on
5 full basketball game. For the assessment of the number of learning activi-
ties and the related game performance the game performance coding instru-
ment of Tallir et al. (2007) was used. This instrument distinguishes three
GPC’s (the decision-making component, the motor skill execution efficiency
component and the motor execution efficacy component) of every observed
offensive action, on and off-the-ball and was consequently preferable to the
Game Performance Assessment Instrument (GPAI) (Oslin et al., 1998)
where all “decisions made” are assessed as one game component, which
makes it impossible to identify in more detail whether a decision is related to
dribbling, passing or shooting.
The significantly higher scores on the three GPC in the 3 on 3 game play
condition indicated that in this game play situation players experienced more
learning opportunities compared to the 5 on 5 full game play condition. This
confirms Metzlers’ (2000) statement that small-sided or modified games (in
this study the 3 on 3 game play), contain more in-game repetitions of key tac-
tics and motor skills. Capel (2000) noted that small-sided games provide
more learning opportunities compared to the full game. When the focus is on
the analysis of the different game performance categories (control, score,
429
CSO and setting up an attack) the larger amount of learning opportunities in
the 3 on 3 game play situation is only found for the game performance cate-
gories control, score, CSO. This may originate from the smaller amount of
players which indirectly offered the players more space and time to make
decisions and to execute these decisions. For the game performance category
setting up an attack the numbers in the DM component and the MSEfficacy
component were higher in the 5 on 5 game play situation, while the MSEffi-
cacy component showed higher scores in the 3 on 3 game play situation. In
the 5 on 5 condition players may experience the need for setting up more
than in the 3 on 3 game play situation because of the crowed situation with
ten players on the playing field. The actions related to setting up an attack are
however not executed correctly, whereas they show nevertheless have a suc-
cessful outcome. This finding shows similarities to the variability principle of
the constraints-led approach (Davids et al., 2008). However, it should be
acknowledged that the assessment of the MSEfficiency component was not
detailed enough in this study to effectively draw this conclusion.
If the increase in learning opportunities found in this study results in an
improvement of game performance remains an unanswered question. Com-
parison of the game performance of the players in both game play situations
showed that the positive proportional values of the three GPC’s are not sig-
nificantly different in both game play situations, except for the cutting
actions. In the 5 on 5 game play the proportion of positive cutting actions in
the decision-making component was significantly higher as compared to the
3 on 3 game play. This may originate from the fact that during 5 on 5 game
play players execute more actions off the ball since they have less opportuni-
ties to execute on the ball actions because of the larger number of players.
The results of the present study underline the importance of future
research to investigate to what extent learning results are easier or faster
attained when applying the non-linear pedagogy principles (Davids et al.,
2008). It is necessary to evaluate the impact of the manipulation of a task
constraint, e. g. the rules of the game, the equipment used, size of the playing
area, and number of players involved during a learning experiment on games
teaching. This should add empirical support to the current lack of strong evi-
dence (Strean and Bengoechea, 2002) in favour of the more student-centered
instructional approaches such as Teaching Games for Understanding
(TGfU). It was stated a decade ago by Rink (2001) that selection of the learn-
ing task may be one of the most important decisions made by physical edu-
cation teachers and this still holds true as illustrated by Chow et al. (2009)
who mentioned that it is a challenge for future research to extend under-
standing of nonlinear pedagogy principles in games teaching research.
430
Another unanswered question in this domain is the search for the opti-
mal game play situation to assess game performance or what game situation
with its specific constraint reflects the developmental status of the players
whose game performance is to be assessed. This nonlinear pedagogy research
topic may even be extrapolated to investigation of the statement made by Sil-
verman (2003) that players will learn more if they get developmental appro-
priate practice during the game play learning activities holds.
Overall, the findings of this study showed that 10-11 year old players
experienced more learning opportunities during 3 on 3 game play as com-
pared to 5 on 5 game play. Since creating learning situations with a lot of
learning opportunities is not an isolated objective, in physical education
lessons as well as in competitive training sessions, the physical load was mea-
sured in both game play situations. Results showed that players had signifi-
cantly lower average heart rates while playing 3 on 3 game play as compared
to the 5 on 5 full game. Heart rates between 170 and 190 BPM are an indi-
cator of high physical exertion, whereas heart rates above 190 BPM occur
during maximal physical exertions. Since the heart rate results of the players
in this study are both covered by the zone of high physical exertion the dif-
ference found between both game play conditions should be nuanced.
Apparently, the smaller dimensions of the basketball court in the 3 on 3
small-sided games, resulting in the lack of long runs during counterattacks,
did not result in clear differential physical requirements compared to the full
game. The conclusion with regard to the physical activity and fitness aims in
physical education lessons as well as in basketball training is that full games
as well as small-sided games can be used to obtain these goals, but that small-
sided games (in case 3 on 3) have the additional advantage of more potential
for learning (and improving) decision-making and motor skill execution.
A limitation of the present study is the fact that the dynamic environ-
mental interactions are not taken into account (Chow et al 2007) in the analy-
sis of game performance. In a recent study of Vilar, Araujo, Davids and But-
ton (2012) the ecological dynamics framework is proposed as a framework to
substantiate insights in successful and unsuccessful performance in game
play. The assessment of the MSEfficacy component of game performance in
this study is far more rudimentary. In fact, it is limited to the assessment of
the observed outcomes of the decisions as they are executed by players dur-
ing game play (Turner & Martinek, 1999).
A final suggestion for future research is related to the fact that repetition,
an important component of exercising, is related to the number of learning
opportunities and is, indeed, a very important prerequisite for learning to
occur. However, one should avoid that players are guided continuously dur-
431
432
ing game play, culminating in the development of teacher or coach depen-
dent performers (Turner & Martinek, 1992). Therefore it is important that
learning activities, such as small-sided games, provide players with a maxi-
mum of in-game repetitions of motor skills without losing the focus on prac-
tising decision-making skills and thus enticing players to continuously reflect
on their game performance. Therefore the instructions and feedback the
learners receive also need to be carefully chosen. According to Metzler
(2000) players’ motivation and thus their learning results depend on the pres-
ence of challenging learning activities. An interesting aspect for future
research may be to investigate the motivational aspects of these small-sided
games compared to the full game.
433
Date Condition
Participants’ number
Game performance Decision-making
component
Motor skill execution
component
Effectiveness
component
Control Pivoting in the direction
of the basket. (cat09)
Not pivoting or not piv-
oting in the direction of
the basket. (cat10)
Holding the ball with two
hands. (cat25)
Knees, hip and elbow
bent. (cat26)
Feet parallel and aimed
at the basket. (cat27)
Pivoting without travel-
ling foul. (cat28)
Pivoting according to the
position of the defender.
(cat29)
Holding the ball with one
hand or holding the ball
with two hands but close
to the floor. (cat30)
Knees, hip and elbow not
bent. (cat31)
Feet not parallel and not
aimed at the basket.
(cat32)
Pivoting with travelling
foul. (cat33)
Pivoting without taken
the position of the
defender into account.
(cat34)
Player stays in possession
of the ball. (cat79)
Player looses the ball.
(cat80)
Scoring Standing close to the bas-
ket and trying to score
when there is no
defender nearby. (cat11)
Standing close to the bas-
ket and not trying to
score when there is no
defender nearby.
Standing far away from
the basket and trying to
score while there was free
space to dribble closer to
the basket.
Standing under the bas-
ket and trying to score.
Trying to score while a
team-mate was in a
favourable position.
Trying to score while
there is close defence.
(cat12)
Feet parallel and aimed
at the basket. (cat35)
Holding the ball in
shooting pocket. (cat36)
Overhand shooting.
(cat37)
Bow in trajectory of the
ball. (cat38)
Clear flexion-extension
movement. (cat39)
Ball hits square on the
basket or the ring. (cat40)
Feet not parallel and not
aimed at the basket.
(cat41)
No shooting pocket.
(cat42)
Not shooting overhand.
(cat43)
No bow trajectory of the
ball. (cat44)
Clear flexion-extension
movement is missing.
(cat45)
Ball misses square on the
basket or the ring. (cat46)
Ball ends in the basket.
(cat81)
Ball misses the basket.
(cat82)
APPENDIX
The game performance coding instrument (Tallir et al. 2007)
(Continued)
434
Date Condition
Participants’ number
Game performance Decision-making
component
Motor skill execution
component
Effectiveness
component
Creating
Shooting
Opportunities
Dribbling
Passing
Catching
Dribbling to take the
free space to the basket.
Dribbling to create
space. (cat13)
Dribbling on the spot.
Not dribbling while
there was free space to
the basket.
Dribbling while a team-
mate stands free in a
favourable scoring posi-
tion. (cat16)
Pass to a team-mate
who stands free and/or
in a more favourable
position. (cat14)
Pass while there was
free space to dribble to
the basket.
Pass to a team-mate
who does not stand
free. Pass to a team-
mate while there was a
scoring opportunity.
(cat17)
No travelling foul at the
start of the dribble.
(cat47)
No travelling foul dur-
ing the dribble. (cat48)
No travelling foul at the
end of the dribble.
(cat49)
Dribble with view on
the game. (cat50)
Travelling foul at the
start of the dribble.
(cat51)
Travelling foul during
the dribble. (cat52)
Travelling foul at the
end of the dribble.
(cat53)
Dribble with the back
aimed at the game.
(cat54)
Not dribbling while
moving with the ball.
(cat55)
Useless dribble (e.g.
when catching a ball).
(cat56)
In the cutting direction.
(cat57)
Not too high, not to far.
(cat58)
Passing with two hands
(chest or bounce pass).
(cat59)
Not in the cutting
direction. (cat60)
Too high, to far, not far
enough. (cat61)
Not passing with two
hands (no chest or
bounce pass). (cat62)
In the running direc-
tion. (cat63)
Not too high, not too
far. (cat64)
With two hands. (cat65)
Not in the running
direction. (cat66)
Too high, too far.(cat67)
Not with two hands.
(cat68)
Dribble ends in a scor-
ing opportunity. Players
stays in possession of
the ball. (cat83)
Player looses the ball.
(cat84)
Ends in ball possession.
(cat91)
Ball possession (after
control with dribble).
(cat92)
Ball is lost. (cat93)
Ends in ball possession.
(cat91)
Ball possession (after
control with dribble).
(cat92)
Ball is lost. (cat93)
(Continued) APPENDIX
(Continued)
435
Date Condition
Participants’ number
Game performance Decision-making
component
Motor skill execution
component
Effectiveness
component
Creating
Shooting
Opportunities
Cutting Cutting to the basket
after giving a pass.
Not cutting to the bas-
ket while the player
with the ball undertakes
an action to the basket.
(cat15)
Cutting to the basket
while the player with
the ball undertakes an
action to the basket.
Not cutting after giving
a pass.
Running behind the
player with the ball.
Cutting and returning
immediately. (cat18)
Cutting immediately
after giving a pass.
Not cutting when there
is an action to the bas-
ket. (cat69)
Asking the ball while
cutting. (cat70)
Eye contact while cut-
ting. (cat71)
Not cutting immedi-
ately. Cutting while
there is an action to the
basket. (cat72)
Not asking the ball
while cutting. (cat73)
No eye contact while
cutting. (cat74)
Leads to a good passing
opportunity. (cat85)
Leads to ball posses-
sion. (cat86)
Leads to a scoring
opportunity. (cat87)
Does not lead to a good
passing opportunity.
(cat88)
Does not lead to ball
possession. (cat89)
Does not lead to a scor-
ing opportunity. (cat90)
Setting up an attack Moving to lose the
defence.
Player is free in the
around the spots.
(cat19)
Not moving when he
can receive the ball.
(cat20)
Player can receive the
ball left and right from
the player with the ball.
(cat21)
Remain standing with
defence in the passing
lane. (cat22)
Free but too far away
from the player with the
ball. (cat23)
Free, but not remain
standing.
Two players on one side
of the ball. (cat24)
Change of speed and
direction. (cat75)
Free, remain standing.
(cat76)
No change of speed and
direction.
Remain standing in a
useless position. (cat77)
Free, but not remain
standing. (cat78)
Player can receive the
ball. (cat94)
Player cannot receive
the ball. (cat95)
(Continued) APPENDIX
REFERENCES
Araújo, D., Davids, K., Bennett, S, Button, C, & Chapman, G. (2004). Emergence of sport
skills under constraint. In A. M. Williams, & N.J. Hodges (Eds.), Skill acquisition in
sport: Research, theory and practice (pp. 409-433). London: Routlegde, Taylor & Francis.
Capel, S. (2000) ‘Approaches to teaching games’ in S. Capel and S. Piotrowski (eds). Issues in
physical education (pp. 81-98) London and New York: Routledge Falmer, Taylor & Fran-
cis Group.
Chow, J. Y., Davids, K., Button, C., Shuttleworth, R., Renshaw, I., & Araújo, D. (2006). Non-
linear pedagogy: A constraints-led framework to understand emergence of game play
and skills. Nonlinear Dynamics, Psychology and Life Sciences, 10(1), 74-104.
Chow, J., Davids, K., Button, C., Shuttleworth, R., Renshaw, I., & Araújo, D. (2007). The role of
nonlinear pedagogy in physical education. Review of Educational Research, 77(3), 251-278.
Chow, J.Y., Davids, K., Button, C., Renshaw, I., Shuttleworth, R., Uehara, L. A. (2009). Non-
linear pedagogy: implications for teaching games for understanding (TGfU). In T. F.
Hopper, J. Butler, & B. Storey (Eds.), TGfU: Simply good pedagogy: understanding a com-
plex challenge. Ottawa: Physical Health Education Association (Canada).
Davids, K., Button, C. & Bennet, S. (2008). Dynamics of skill acquisition: a constraints-led
approach. Champaign, IL: Human Kinetics.
Gréhaigne, J. F., Bouthier, D., & David, B. (1997). Dynamic-system analysis of opponent rela-
tionships in collective actions in soccer. Journal of Sport Sciences, 15(2), 137-149.
Griffin, L. L., Butler, J., Lombardo, B., & Nastasi, R. (2003). An introduction to teaching games
for understanding. In J. Butler, L. Griffin, B. Lombardo & R. Nastasi (Eds.), Teaching
games for understanding in physical education and sport. VA: NASPE Publications.
Griffin, L. L., & Butler, J.(2005). Teaching Games for Understanding – Theory, research and
practice. Champaign, IL: Human Kinetics.
McGarry, T., Anderson, D. Wallace, S., Hughes, M. & Franks, I. (2002). Sport competition as
a dynamical self-organizing system. Journal of Sport Sciences, 20, 771-881.
Metzler, M. W. (2000). Instructional models for physical education. Boston: Allyn and Bacon.
Mitchell, S.A., Oslin, J.L., & Griffin, L.L. (1995). The effects of two instructional approaches
on game performance. Pedagogy in practice. Teaching and coaching in physical education
and sport, 1, 36-48.
Newell, K.M. (1986). Constraints on the development of coordination. In M.G. Wade, &
H.T.A. Whiting (Eds.), Motor development in children. Aspects of coordination and con-
trol (pp. 341-360). Dordrecht, Netherlands: Martinus Nijhoff.
Newell, K.M. (1996). Change in movement and skill: Learning, retention and transfer. In M.L.
Latash, & M.T. Turvey (Eds.), Dexterity and its development (pp. 393-430). Mahwah, NJ:
Erlbaum.
Oslin, J. L., Mitchell, S. A., & Griffin, L. L. (1998). The game performance assessment instru-
ment (GPAI): Development and preliminary validation. Journal of teaching in physical
education, 17, 231-243.
Passos, P., Araújo, D., Davids, K., & Shuttleworth, R. (2008). Manipulating constraints to
train decision making in rugby union. International Journal of Sport Science & Coaching,
3(1), 125-140.
Renshaw, I., Chow, J. W., Davids, K. & Hammond, J. (2010). A constraints-led perspective to
understanding skill acquisition and game play: a basis for integration of motor learning
theory and physical education praxis? Physical Education and Sport Pedagogy, 15(2), 117-
137.
436
437
Richard, J., & Wallian, N. (2005). Emphasizing student engagement in the construction of
game performance. In L. L. Griffin, & J. Butler (Eds.), Teaching Games for Understand-
ing – Theory, research and practice. Champaign, IL: Human Kinetics.
Rink, J. (2001). Investigating the assumptions of pedagogy. Journal of Teaching in Physical
Education, 20, 112-128.
Silverman, S. (2003). The pedagogy of motor skill learning: Teachers and students. In A. Laker
(Ed.), The future of physical education (pp. 102-120). London: Routledge, Taylor & Fran-
cis Group.
Strean, W. B., & García Bengoechea, E. (2003). Beyond technical vs. tactical: Extending the
games teaching debate. In J. Butler, L. Griffin, B. Lombardo, & R. Nastasi (Eds.), Teach-
ing games for understanding in physical education and sport: An international perspec-
tive..VA: NASPE publications.
Tallir, I. B., Lenoir, M., Valcke, M., Musch, E. (2007). Do alternative instructional approaches
result in different game performance learning outcomes ? Authentic assessment in
varying game conditions. International Journal of Sport Psychology, 38, 263-282.
Thomas, K. T., French, K. E., & Humphries, C. A. (1986). Knowledge development and sport
skill performance: Directions for motor behaviour research. Journal of Sport Psychology,
8, 259-272.
Turner, A.P., & Martinek, T.J. (1992). A comparative analysis of two models for teaching
games (technique approach and game-centered (tactical focus) approach). International
Journal of Physical Education, 29(4), 15-31.
Turner, A.P., & Martinek, T.J. (1999). An investigation into teaching games for understanding:
Effects on skill, knowledge and game play. Research Quarterly for Exercise and Sport,
70(3), 286-296.
Vilar, L., Araujo, D., Davids, K.,& Button, C. (2012). The role of ecological dynamics in ana-
lyzing performance in team sports. Sports Medicine, 42(1), 1-10.
Manuscript submitted May 2012. Accepted for publication September 2012.
