In many invasion sports such as Australian football (AF), agility is important to either evade a defender or put pressure on a player in possession of the ball. Agility has typically been defined as the ability to change body direction and position rapidly (2). The Planned Australian Football League (AFL) agility test is one example of a test used for talent identification in Australian football and involves 5 planned changes of direction around poles. However, more recently, it has been recognized that agility performance is also influenced by perceptual and decision-making factors (3,14,18). There is mounting evidence that the ability to quickly and accurately react to an opponent's movements is associated with high-level performance. For example, a number of studies have compared tests involving “reactive agility” with tests that require preplanned changes of direction with no perceptual and decision-making components. This research generally shows that the reactive agility tests (RAT) are better able to discriminate higher-level performers from lower-level athletes in netball (4), rugby league (5,13), and Australian football (15).
Expert rugby union players are less susceptible to deceptive actions from an opponent changing direction than inexperienced players are (8). Perceptual skill is also important in soccer, with higher-level players producing faster and more accurate responses to an opponent dribbling a ball than players with less expertise (11,16,17). The source of this advantage is the ability of the higher-skilled players to use advanced kinematic cues emanating from the biomechanics of their opponent's action (16). Semiprofessional soccer players exhibited superior perceptual and decision-making abilities when presented in static images or on film displaying soccer-specific tasks than lower level players exhibited (6). In contrast, there were no substantial differences between the player groups in non–soccer-specific perceptual tasks such as reaction time. These findings suggest that the cognitive skill of experts is not generic in nature but requires the ability to locate and use highly sport-specific visual information.
One example of an RAT involves a live tester producing a side step as a stimulus for the athlete to produce a change of direction response (5,15). Although this field test possessed adequate reliability, the variability associated with the live tester movements can influence the overall agility score (20). An alternative approach to overcome this limitation is the use of life-size video clips of an attacker projected on a large screen to provide a standardized stimulus for the athlete to respond to (4,13). This approach to testing reactive agility has not yet been evaluated with Australian football players. Therefore, the main purpose of this study was to assess the concurrent criterion validity and test–retest reliability of a new video-based RAT designed specifically for Australian football. A secondary aim was to determine if an RAT containing a non–sport-specific stimulus and the Planned AFL agility test could discriminate between elite and secondary school junior Australian football players. We hypothesized that the new RAT involving a sport-specific stimulus would be better able to discriminate between higher and lower level players (and therefore possess concurrent criterion validity) than tests that use either a generic stimulus or do not use any stimulus to change direction. The results of this study will therefore inform practitioners about the importance of including a stimulus in agility tasks and the nature of that stimulus in designing better agility training activities and testing protocols.
Experimental Approach to the Problem
Elite and subelite junior Australian football players were compared on 3 agility tests to determine what characteristics of the tests were associated with the level of competition. One agility test involved planned changes of direction, whereas the other 2 involved reacting either to a generic stimulus (arrow) or a sport-specific stimulus to change direction. By comparing a higher and lower performance group of players, an insight into the association between the 3 types of agility tests and sports performance reveals concurrent criterion validity. This approach to analyzing the utility of agility tasks has been used extensively in previous research (4,5,13,15,17).
A subgroup of athletes performed the new reactive tests on 2 occasions to establish the test and retest reliability. All tests were conducted while the athletes came together for a training and educational camp. This scenario meant that the players were involved in various nonfatiguing activities in small groups as directed by the team coaches and resulted in somewhat smaller and uneven sized groups in some tests. The study was designed to provide insights into the utility of the new RAT and the characteristics of agility needed for Australian football performance.
The participants were 50 male junior Australian football players (age 16.5 ± 0.8 years, height 189.6 ± 6.1 cm, and body mass 79.3 ± 6.8 kg). This group comprised elite athletes (n = 35) from a National talent squad and lower performance group (n = 15) from a private secondary school. The higher performance group players were considered the best players nationally and were provided with a structured strength and conditioning program to supplement normal football training. The lower performance school players represented their school in a local competition, and all subjects were free of injury and were just commencing a preparation phase at the time of testing. The study was approved by the Australian Institute of Sport (AIS) and University's Ethics Committees.
All the tests were conducted in the morning between 9:30 and 11:30 am. The athletes were encouraged to arrive at all tests in a hydrated state and had refrained from exercise for 18 hours. All athletes consumed a self-selected breakfast 2–3 hours before the agility testing sessions. Alcohol consumption was not allowed at any time before or during the testing period.
The RAT required the participants to respond to the movements of a player shown on a video clip displayed in life size on a 3- × 3-m rear projection screen. The footage presented an elite junior footballer running while carrying a ball and performing a side step to change direction to either the left or the right. There were 3 views of this simulated attacker; front-on, an oblique view at about 45° from the left or right of the screen, and a rear view. Previous RATs performed with a live tester were restricted to a front-on view without a ball, so the new sport-specific test was designed to include a ball and multiple views as would be observed in a game.
The participants were instructed to move forward and, while watching the player projected on the screen, run to the left or right as fast as possible in response to his movements, as if briefly chasing him. The initial forward movement triggered the playing of the video clip, and the participant typically moved forward approximately 4 m and then changed direction by ∼45° followed by a 4-m sprint to complete the test (Figure 1). The total time of the movement was recorded by an electronic time gate system (Swift Performance, Wacol, Australia) with a resolution of 0.01 seconds. The time recorded began the instant the video clip was triggered and finished the instant the player ran through the left or right timing gate. After a standardized warm-up, each participant performed 4 practice trials followed by 6 test trials consisting of 2 of each of the 3 views and equal number of trials to the left and the right. The warm-up lasted approximately 10 minutes and consisted of run-throughs over 15 m in a straight line and runs with 4 changes of direction in a zig-zag pattern and gradually increasing in intensity from 50 to 100% of maximum effort.
The other version of this RAT used the same procedure except that a left or right arrow on the screen was provided as a stimulus instead of a video clip of a player. Six trials were again performed with the left and right directions presented randomly. This test was designed to identify if an RAT containing a directional non–sport-specific stimulus (arrow) could discriminate the 2 playing groups.
The sequence of the clips was presented in the same order for each participant but each athlete was tested on their own and was not permitted to discuss the details of the test with the other players until all the tests were completed. The video clips were controlled by custom-made “React” software (AIS, Canberra, Australia). The mean of the 6 trials was subsequently used as the score for each individual. The third test was the Planned AFL agility test, involving 5 changes of direction described in detail elsewhere (19). Briefly, this test required the athletes to weave around 5 poles that were approximately 1 m in height and was timed with the same electronic timing system as used for the reactive tests. The faster of 2 trials was recorded as the criterion time. Complete recoveries between trials of approximately 30 seconds were provided for all tests to minimize the effects of fatigue.
The data were analyzed for descriptive statistics and the 2 playing groups were compared on each test with a 1-way analysis of variance, with the significance level set at 0.05. The magnitude of standardized differences between groups was described using the thresholds of Hopkins (7). A Pearson correlation was calculated to describe the relationship between the new RAT and the RAT (arrows). The test and retest reliability was estimated by calculating the typical error (TE), the coefficient of variation (with 90% confidence limits), and the intraclass correlation (ICC).
The descriptive statistics, p values, and effect sizes for each group are shown in Table 1. The elite group was significantly better in the new RAT than the school footballers were. In contrast, there was no significant difference between the groups in the RAT (arrows), whereas the school group was moderately better in the Planned AFL agility test. The correlation between the new RATand the RAT (arrows) was r = 0.504 (p = 0.079).
The reliability statistics are shown in Table 2. The TEs for the new RAT and the RAT (arrows) were 0.07 seconds (0.05–0.11 seconds) and 0.09 seconds (0.06–0.15 seconds), respectively.
The major finding in this study was that better players were clearly superior in the new RAT, which supports the concurrent construct validity of this test. It appears that the characteristics of the test such as the movement pattern and perceptual and decision-making skills are highly relevant to the playing ability of junior AF players. This finding is consistent with that of previous research using an RAT with a live tester, which was also related to expertise in AF players (15). The critical finding of this study was that the elite group did not perform any better on the RAT (arrows) despite this test involving the same movement pattern. It appears that the superiority of the elite group of players resides in their ability to perceive and use the information displayed in the video footage of the simulated attacker changing direction. It is likely that the higher National level players were able to use advanced kinematic cues to anticipate the direction of the simulated attacker and respond and move more quickly. This skill could not be employed when reacting to the arrows in the RAT (arrows) test. Anticipatory skill has also been shown to be related to playing ability in soccer (17), rugby league (13), and netball (4).
Our findings are consistent with those of research on soccer showing that better players are faster and more accurate at responding to sport-specific dribbling and passing images but not better than players with less expertise in reacting to a light stimulus (6). Oliver and Meyers (9) reported a very high correlation (r = 0.93, p < 0.001) between a planned agility test and a test involving a similar movement pattern requiring a response to a light stimulus. This observation indicates that the inclusion of a light stimulus does not significantly alter the nature of planned change of direction task. In this study, the correlation between the new RAT and the RAT (arrows) was r = 0.504, representing only 25% common variance. We interpret this to mean that the task of reacting to a player moving is very different to the task of responding to the arrow stimulus. Good agility performance appears to require the perception and processing of sport-specific cues rather than reaction to a nonspecific generic stimulus.
The exact cues that are used by elite performers in this task are not yet known but would be of interest so that players could potentially be coached to “read” their opponents' movements more effectively. Previous research has tracked eye movements to provide insights into the visual search strategies of soccer goalkeepers when observing penalty kicks (12). However, the cues used in other sports and the fact that they are restricted to a front-on view have limited application to the complex agility scenarios used in this study. More naturalistic work using representative task designs are required to better understand the link between perception and resultant anticipatory action in team sports.
The finding that the National talent squad performed moderately worse than the school group in the Planned AFL agility test was unexpected (Table 1). Previous assessment of elite junior players in an Australian Institute of Sport AFL Academy program over a 6-year period (n = 219) yielded a mean time for this test of 8.77 ± 0.43 seconds (unpublished observations, AIS). The mean time in this study (9.08 seconds) was considerably slower and may in part be explained by the fact that the National talent squad comprised “development” players who were of a slightly lower standard than the Australian Institute of Sport AFL Academy players previously tested. Another possibility is that players from the National talent squad were slightly fatigued before our testing because they were in a training camp. Although this could partly explain their lower performance in the Planned AFL agility test, it would also be expected to impair their performance on the other tests. Despite the possible presence of fatigue, the National talent squad was far superior to the school group in the new RAT.
The fact that the elite group was not better than the school group in the planned AFL agility test may also be explained by the characteristics of this test. First, it involves changes of direction of up to 180°, that is, the player runs up around a pole and back to the original direction. Time–motion analysis of AFL games (1) has shown that typical changes of direction are <90°. Another feature of this test is that it involves preplanned movements whereby the athlete has to locate the position of a pole and adjust his steps to move around the obstacle without touching it. The resulting movements can appear quite unnatural and are different from the agility demands of the game. Previous research using the planned AFL agility test has shown that this test is only slightly related to playing performance in junior footballers (10,19). However, a study of fitness characteristics of AFL Draft Combine players (who are 1–2 years older than the players in this study) (10) reported a mean time of 8.57 seconds for the planned AFL agility test, a performance that is superior to the players in this study. Therefore, more work is needed to refine planned and RATs to consistently differentiate players of different levels of expertise and seniority.
Although the results of this study clearly point to the advantages of the new RAT, the use of this test must take into consideration its reliability and practicality. The reliability of this test was not very good, as indicated by the ICC of 0.33 (Table 2). The TE of 0.07 seconds suggests that the test can be used confidently to detect moderate to large changes in reactive agility, but refinements are needed to identify small differences. One potential source of variability was a degree of uncertainty of the players about how to approach the initial forward movement. Some athletes seemed to move quickly, whereas others hesitated while waiting for the change of direction stimulus. Although these variable approaches were observed, the use of an optimum movement strategy can be considered as a relevant aspect of agility skill that should be included in a test. Another factor is the degree of familiarization of the players to the new tests. A suggested refinement to the RAT is to include a series of familiarization trials for players. More work is required to refine both reactive and planned agility tests to improve their sensitivity in detecting small differences and changes in agility performance of football and other team sport players.
Elite junior Australian football players were distinguished from their secondary school counterparts by the new RAT that required a fast change of direction in response to the movements of a simulated attacker. Given that there was no substantial difference between playing groups on a reactive test involving a generic stimulus to change direction (arrows), it is concluded that agility performance is influenced by the ability to perceive and use sport-specific kinematic cues from an opponent. This conclusion was reinforced by the finding that the AFL agility test involving preplanned movements with no reactive component was not associated with players' expertise. Training and testing reactive agility should contain sport-specific rather than generic stimuli. An example of a training activity would be small-sided games that require multiple changes of speed and direction in response to the movements of other players, in both attacking and defensive roles. In contrast, agility drills involving nonspecific stimuli such as flashing arrows, lights, or verbal stimuli such as a coach calling out a direction are not specific to sports and are therefore not recommended.
There was no external financial support provided for this research.
1. Dawson, B, Hopkinson, R, Appleby, B, Stewert, G, and Roberts, C. Player movement patterns and game activities in the Australian Football League. J Sci Med Sport
7: 278–291, 2004.
2. Draper, JA and Lancaster, MG. The 505 test: A test for agility in the horizontal plane. Austr J Sci Med Sport
March: 15–18, 1985.
3. Ellis, L, Gastin, P, Lawrence, S, Savage, B, Sheales, A, Stapff, A, Tumilty, D, Quinn, A, Woolford, S, and W. Young. Protocols for the physiological assessment of team sport players. In: Physiological Tests for Elite Athletes
. Gore, CJ, ed. Champaign, IL: Human Kinetics, 2000. pp. 128–144.
4. Farrow, D, Young, W, and Bruce, L. The development of a test of reactive agility
for netball: A new methodology. J Sci Med Sport
8: 52–60, 2005.
5. Gabbett, TJ, Kelly, JN, and Sheppard, JM. Speed, change of direction
speed, and reactive agility
of rugby league players. J Strength Cond Res
22: 174–181, 2008.
6. Helsen, WF and Starkes, JL. A multidimensional approach to skilled perception
and performance in sport. Appl Cogn Psychol
13: 1–27, 1999.
8. Jackson, RC, Warren, S, and Abernethy, B. Anticipation skill and susceptibility to deceptive movement. Acta Psychol
123: 355–371, 2006.
9. Oliver, JL and Meyers, RW. Reliability and generality of measures of acceleration, planned agility, and reactive agility
. Int J Sports Physiol Perform
4: 345–354, 2009.
10. Pyne, DB, Gardner, AS, Sheehan, K, and Hopkins, WG. Fitness testing and career progression in AFL football. J Sci Med Sport
8: 321–332, 2005.
11. Reilly, T, Williams, AM, Nevill, A, and Franks, A. A multidisciplinary approach to talent identification in soccer. J Sport Sci
18: 695–702, 2000.
12. Savelsbergh, GJP, Williams, AM, Van Der Kamp, J, and Ward, P. Visual search, anticipation, and expertise in soccer goalkeepers. J Sport Sci
20: 279–287, 2002.
13. Serpell, BG, Ford, M, and Young, WB. The development of a new test of agility for rugby league. J Strength Cond Res
24: 3270–3277, 2010.
14. Sheppard, JM and Young, WB. Agility literature review: Classifications, training, and testing. J Sport Sci
24: 919–932, 2006.
15. Sheppard, JM, Young, WB, Doyle, TLA, Sheppard, TA, and Newton, RU. An evaluation of a new test of reactive agility
and its relationship to sprint speed and change of direction
speed. J Sci Med Sport
9: 342–349, 2006.
16. Ward, P and Williams, MA. Perceptual and cognitive skill development in soccer: The multidimensional nature of expert performance. J Sport Exerc Psychol
25: 93–111, 2003.
17. Williams, AM and Davids, K. Visual search strategy, selective attention, and expertise in soccer. Res Q Exerc Sport
69: 111–128, 1998.
18. Young, WB, James, R, and Montgomery, I. Is muscle power related to running speed with changes of direction? J Sports Med Phys Fitness
42: 282–288, 2002.
19. Young, WB and Pryor, L. Relationship between pre-season anthropometric and fitness measures and indicators of playing performance in elite junior Australian Rules football. J Sci Med Sport
10: 110–118, 2007.
20. Young, WB and Willey, B. Analysis of a reactive agility
field test. J Sci Med Sport
13: 376–378, 2010.