Agility literature review: Classifications, training and testing
J. M. SHEPPARD
1
& W. B. YOUNG
2
1
Australian Institute of Sport, Belconnen, ACT and
2
School of Human Movement and Sport Sciences, University of Ballarat,
Ballarat, VIC, Australia
(Accepted 4 November 2005)
Abstract
At present, no agreement on a precise definition of agility within the sports science community exists. The term is applied to
a broad range of sport contexts, but with such great inconsistency, it further complicates our understanding of what trainable
components may enhance agility. A new definition of agility is proposed: ‘‘a rapid whole body movement with change of
velocity or direction in response to a stimulus’’. Agility has relationships with trainable physical qualities such as strength,
power and technique, as well as cognitive components such as visual scanning techniques, visual scanning speed and
anticipation. Agility testing is generally confined to tests of physical components such as change of direction speed, or
cognitive components such as anticipation and pattern recognition. New tests of agility that combine physical and cognitive
measures are encouraged.
Keywords: Direction change, speed, anticipation, acceleration
Introduction
Sprint training is an integral part of the overall training
for track and field athletes, as well as field and court
sports. Most sprint training focuses on drills and
conditioning to develop acceleration and top speed
straight sprinting (Blazevich, 1997a, 1997b; Delecluse,
1997; Donati, 1996; Francis, 1997; Knicker, 1997;
Luchtenbern, 1990; Sheppard, 2003, 2004; Young,
1995; Youn g, Benton, Duthie, & Pryor, 2001a). Past
research and reports have been published regarding
sprinting phases, including the acceleration, maximal
speed and speed endurance phases (Burggemann
& Glad, 1990; Enoka, 2002; Kyrolainen, Komi, &
Belli, 1999; Mann, 1981).
The current paradigm of speed development
is undergoing change in the sport science commu-
nity, wherein a greater emphasis is being placed
not just on acceleration, top speed and speed
endurance training, but also on change of direc-
tion speed drills (Fulton, 1992; Gambetta, 1996;
Moreno, 1995; Sayers, 2000; Twist & Benicky,
1996). This represents an emphasis on the specificity
of training with specific movement patterns, as
straight sprint training appears to have little or no
influence on the improvement of sprinting that
involves changes of direction (Young, McDowell,
& Scarlett, 2001b). Additional support for this is
evidenced by a weak relationship between straight
sprint performance and change of direction speed
performance (Baker, 1999a; Buttifant, Graham, &
Cross, 1999; Clark, Martin, Lee, Fornasiero, &
Quinn, 1998; Tsitskarsis, Theoharopoulus, & Garefis,
2003; Young, Hawken, & McDonald, 1996).
Many field and court sports involve some straight
sprinting, but more often repeated short sprinting
with changes of direction. The ability to sprint
repeatedly and change direction while sprinting is a
determinant of sport performance in field and court
sports, as evidenced by time and motion analysis,
validation of testing batteries for elite and non-elite
performers, and coaching analysis for sports such as
rugby (Docherty, Wenger, & Neary, 1988; Meir,
Newton, Curtis, Fardell, & Butler, 2001), field
hockey (Keogh, Weber, & Dalton, 2003) and soccer
(Reilly, Williams, Nevill, & Franks, 2000). Con-
sidering that field and court sports generally include
these changes of direction in response to a stimulus
(e.g. another player’s movement, movement of play
or the ball), it would seem important to provide
testing and training that mimics this demand to
increase specificity.
This article is a review of the literature that is
relevant to defining agility, its relationships with
other trainable qualities, and the testing of agility.
We propose a new definition of agility that recognizes
both the cognitive and physical components involved
in agility for sport.
Defining agility
At present, there is no consensus among the sports
science comm unity for a clear definition of agility.
Agility has classically been defined as simply the
ability to change direction rapidly (Bloomfield,
Ackland, & Elliot, 1994; Clarke, 1959; Mathews,
1973), but also the ability to change direction rapidly
and accurately (Barrow & McGee, 1971; Johnson &
Nelson, 1969). In more recent publications, some
authors have defined agility to include whole-body
change of direction as well as rapid movement and
direction change of limbs (Baechle, 1994; Draper &
Lancaster, 1985).
Even more confusing has been the introduction of
the term ‘‘quickness’’ (Baker, 1999a; Moreno,
1995), which is seemingly used interc hangeably for
both agility and change of direction speed. Quickness
has been identified as ‘‘a multi-planar or multi-
directional skill that combines acceleration, explo-
siveness, and reactiveness’’ (Moreno, 1995). This
definition suggests that quickness consists of cogni-
tive and physical reactive abilities and explosive
acceleration. If this is an identifiable physical quality,
then one might infer that quickness is a component
of agility, as the proposed definition (Moreno, 1995)
for quickness does not include deceleration or
changing direction. However, the available literature
includes skills and tests that involve changing
direction and deems these to be quickness drills
and tests (Baker, 1999a; Moreno, 1995).
Currently, the term quickness is used a great deal
in North American sports settings, and has been the
topic of several presentations and workshops mar-
keted towards athletes and coaches. The term
quickness is also used extensively on the world-
wide-web in reference to training methods for field-
sport athletes. Although the exact definition of
quickness is unclear, its use wi ll be avoided in the
current article, as it is seemingly vague.
In addition, the term ‘‘cutting’’ has been used with
reference to a directional change during a sprint
movement (Bernier, 2003; Besier, Lloyd, Ackland, &
Cochrane, 2001a; Besier, Lloyd, Cochrane, &
Ackland, 2001b; Colby et al., 2000; McClay et al.,
1994). Unlike the term quickness, cutting seemingly
refers only to the specific portion of a directional
change where the athlete’s foot contacts the ground
to initiate the change of direction.
The difficulty in finding an accepted definition of
agility could be the result of the multiple factors,
from various disciplines within sports science, which
influence agility performance. A biomech anist might
view agility in terms of the mechanical changes
involved in altering body position. A motor learning
scientist in sports psychology might view agility in
terms of the information processing involved in
visual scanning, decision making and reaction to a
stimulus to change directions, as well as the process
involved in learning and retaining the appropriate
motor skill. Strength and conditioning coaches might
define agility in terms of the physical qualities
involved in changing direction. The differences seen
in definitions of agility could simply be due to the
perspective of various authors, and th eir individual
expertise and background. A comprehensive defini-
tion of agility would recognize the physical demands
(strength and conditioning), cognitive processes
(motor learning) and technical skills (biomechanics)
involved in agility performance.
In 1976, Chelladurai proposed a thorough defini-
tion of agility, noting that although there was
agreement on the importance of agility in many
sports, there were many varied definitions of agility.
Furthermore, Chelladurai noted that none of these
definitions included appropriate recognition of the
perceptual and decision-making components that are
involved in many sports. The author outlined a
classification of agility so that tasks were deemed to
be simple, temporal (no spatial uncertainty, but
temporal uncertainty), spatial (no temporal uncer-
tainty, but spatial uncertainty), or universal
(temporal and spatial uncertainty) (Table I).
Defining various forms of agility performance,
such as simple, temporal, spatial and universal, is
unique in the literature. In particular, movements
like the sprint start in athletics, which are considered
agility tasks (Chelladurai, 1976), could be described
as involving reaction time and velocity, as reaction
time is defined as the minimum time from the
presentation of a stimulus to the onset of a response,
with velocity being defined as the rate of change in
position with respect to time (Enoka, 2002). How-
ever, in the context of Chelladurai’s (1976) complete
group of definitions, simple, temporal, spatial and
universal agility provide a unique framework for the
understanding of the demands of many sports. When
viewed from their simplest to most complex, tasks
can be classified into one of the four categories
outlined by Chelladurai (1976). This framework
could be useful for coaches and sport scientists to
classify sporting skills, thereby allowing an improved
understanding of the sub-component s involved.
Most research on agility testing has applied the
term ‘‘agility’’ to describe any dynamic sporting
action that involves a change in body position
(Draper & Lancaster, 1985; Fulton, 1992; Hastad
& Lacy, 1994). The application of the term agility
varies, but has included lunges (Cronin, McNair, &
Marshall, 2003), a 3-yard run forward and back from
a stationary start (Hoyle & Holt, 1983), climbing
over and under a track and field hurdle (Alricsson,
Harns-Ringdahl, & Werner, 2001), sprinting for-
ward, stopping and returning from a 1808 turn
(Draper & Lancaster, 1985), simple hopping move-
ments (Booher, Hench, Worrell, & Stikeleather,
1993), but most commonly sprinting with directional
changes (Fulton, 1992; Gabbett, 2002; Gambetta,
1996; Meir et al., 2001; Reilly et al., 2000; Rigg &
Reilly, 1987; Twist & Benicky, 1996). According to
Chelladurai (1976), all of these movements could be
classified as simple agility only, in that there is no
temporal or spatial uncertainty involved.
Recently, Young, James and Montgomery (2002)
outlined a comprehensive definition of agility as it
related to running sports such as football codes. The
researchers addressed the multi-faceted influences
involved in agility performance. In particular, the
authors outlined that there are two main components
of agility – change of direction speed and perceptual
and decision-making factors. Within these two main
components, sub-components exist, as outlined in
Figure 1.
Table I. Classifications of agility (modified from Chelladurai, 1976).
Agility classification Definition Example of sporting skill
Simple No spatial or temporal uncertainty Gymnast’s floor routine: pre planned activity,
initiated when the athlete desires, with movements
that the athlete has pre planned. Stimulus is the
athlete’s own movement and the physical domain in
which they are executing the skill
Temporal Temporal uncertainty, but movement is
pre planned (spatial confidence)
Athletics sprint start: pre planned activity, initiated in
response to a stimulus (starter’s pistol) wherein there
is no certainty as to exactly when the pistol will fire
Spatial Spatial uncertainty, but timing of movement
is pre planned (temporal confidence)
Volleyball or racquet sport service receive: the umpire
determines a narrow window of time wherein the
server must serve the ball to the opponent. However,
there is no certainty on the part of the receiver as to
where the service will be directed
Universal Spatial and temporal uncertainty Ice hockey or football: during offensive and defensive
plays, the athletes cannot anticipate with certainty
when or where opposition players will move to
Figure 1. Universal agility components (modified from Young et al., 2002).
In addition to their classification of agility for
running sports, Young et al. (2002) included the
term ‘‘change of direction speed’’ not only as a
component of agility, but also to describe movement
wherein no reaction to a stimulus is required. In
other words, some conditioning exercises could be
classified as change of direction speed exercises
(sprints with changes of direction), while others
could be classified as agility (sprints with directional
changes in response to a stimulus).
Based on a review of literature that attempts to
classify agility, it is obvious that several inconsisten-
cies exist. There is clearly a trend among coaches and
sport scientists to apply agility in a liberal manner,
seemingly wherever a task involves dynamic move-
ment requiring athleticism. This practice retards our
understanding of the unique nature of skills that are
applied in various settings. For example, if we are to
accept that a pre-planned task such as an obstacle
course is an agility task (Pandorf et al., 2003), yet we
also accept that a reactive evasion drill for team sport
athletes requires agility, how do we measure agility?
And, within each circumstance, what factors are
involved in agility performance? Sports scientists and
coaches will be unable specifically to target the
agility-related needs of various athletes if these sub-
classifications of agility are not understood.
A simpler definition of agility could be established
by using an exclusion criterion, rather than an
inclusion criterion such as that proposed by
Chelladurai (1976) (Table I). While Chelladurai
defined four levels within the context of agility, a
more straightforward definition could assist coaches
and sport scientists in communication, research,
training and testing of agility.
If we accept that agility involves a response to a
stimulus, then it is appropriate to eliminate the use of
the classifications within agility that Chelladurai
(1976) proposed. Tasks could be identified by
describing the skill itself, using a biomechanical or
physiological perspective. For example, a shot-put
motion does not involve a response to any stimulus,
although it has been described as an agility task
(Chelladurai, 1976). This ability to pre-plan the skill,
although requiring decision making, does not involve
a response to a stimulus. From a cognitive perspec-
tive, this type of activity is referred to as a closed skill
(Cox, 2002; Murray, 1996). Instead of classifying
this as an agility task, we could describe the muscular
and biomechanical motions involved. This will
eliminate the grouping of such varying tasks into
the classification of agility.
Another example often described as an agility
running drill involves athletes running patterns
around stationary objects (Gambetta, 1996). There
is no reaction to a stimulus in these drills, and
therefore training and testing in this manner will
simply develop or evaluate cha nge of direction speed,
rather than agility performance.
Ultimately, the cognitive components involved in
tasks that have traditionally been described as agility
(e.g. athletics sprint start, shot put, zig-zag runs)
differ greatly from tasks that cont ain significant
uncertainty of time or space (e.g. react ing to a spike
in volleyball, evading an opponent in football). To a
great extent, the execution of many skills that have
traditionally been deemed as agility have an auto-
matic response, and therefore little or no uncertainty
(Murray, 1996). From a cognitive perspective, these
are closed skills and uncertainty is limited.
Open skills require athletes to respond to sensory
stimuli around them, and the response is not an
automated or rehearsed response (Cox, 2002). To
provide further clarification, the example of a sprint
start is useful. When a sprinte r is set in the blocks, he
or she will initiate movement in response to the audio
stimulus of the starter’s pistol. However, the
response is one that can be rehearsed and therefore
pre-planned. Although this skill has been referred to
as an agility task (Chelladurai, 1976), it is not an
open skill, and therefore is not an agility task.
By adopting an exclusive definition, the clarity and
specificity of wording within the sporting community
will increase. In other words, tasks are either
accepted as agility tasks or they are not. We propose
a new definition of agility for sport as fol lows: ‘‘a
rapid whole-body movement with change of velocity
or direction in response to a stimulus’’. This
definition respects the cognitive components of
visual scanning and decision making that contribute
to agility performance in sport (Abernethy, Wood, &
Parks, 1999; Chelladurai, 1976; Young et al., 2002),
as well as the physical performances involved in
acceleration, deceleration and changes of direction in
evading an opponent, sprints with changes of
direction to contact a ball or player, or initiation of
whole-body movement in response to a stimulus. To
be considered an agility task, the movement will not
only involve change in speed or direction, but must
also be an open skill, wherein a reaction to a stimulus
is involved and the movement is not specifically
rehearsed.
This definition is therefore not dependent on
directional change, as with previous definitions
(Baechle, 1994; Chelladurai, 1976; Draper &
Lancaster, 1985; Johnson & Nelson, 1969; Semenick,
1990). For example, agility could describe a soccer
player who rapidly accelerates or decelerates in a
straight line to evade an opponent, as this action is
not pre-planned, would be in response to the move-
ments of the opposing player (stimuli) and is an
open skill.
Physical relationships with agility
With a clear understanding of what an agility task
is, we can proceed to examine the physical factors
that can potentially be traine d to improve agility per-
formance. Almost all existing literature that has
attempted to describe relationships with some
measure of agility or training to improve agility has
used a timed task involving one or more changes of
direction, also known as change of direction speed.
Relationship between straight sprinting speed
and change of direction speed
Anecdotally, it would appear that many strength and
conditioning coaches believe that there is indeed a
strong relationship between straight sprinting speed
and change of direction speed, as some articles and
many training sessions tend to address both qualities
simultaneously. However, research evidence to sup-
port this view could not be found.
For example, in comparing the relationship be-
tween performance of the Illinois agility test and a
20-m sprint, Draper and Lancaster (1985) reported a
statistically significant low to moderate correlation
(r ¼ 0.472). The Illinois agility test is a timed task
involving some straight sprinting and multiple
direction changes around obstacles.
Young et al. (1996) also investigated the relation-
ships between speed and change of direction speed
among Australian Rules football players. In this
study, the researchers compared straight sprinting,
sprinting while bouncing a football, sprinting with
three planned directional changes at 908 angles,
sprinting with three planned direc tional changes at
908 angles while bouncing a football, and sprinting
with three directional changes at 1208 angles. The
results supported the researchers’ hypothesis in that
the correlations between sprint and agility tests were
all very low, indicating that sprinting, sprinting while
bouncing a ball and sprinting while changing
direction were distinct and specific qualities.
Using similar planned change of direction tests,
Baker (1999a) examined the performance differences
of elite and developmental rugby league players.
The results of the study indicated that the two
groups were similar in their straight running speed,
but that the elite players performed better in tests
that involved change of direction. The results of
Baker’s (1999a) study support those of Young et al.
(1996), in that sprinting and agility are separate
physical qualities. Additionally, similarly poor corre-
lations (r ¼ 0.33) were reported with soccer players
(Buttifant, Graham, & Cross, 1995) when comparing
change of direction speed (CODS) test performance
and straight sprint test performance. The participants
were tested on a 20-m straight sprint as well as a
generic CODS test, involving four directional
changes, of approximately 20 m.
Based on the similar results presented by Baker
(1999a), Buttifant et al. (1999), Draper and Lancaster
(1985) and Young et al. (1996), straight sprint testing
appears not to be related strongly to sprinting with
changes of direction testing. Furthermore, and
perhaps most importantly, straight sprint training
does not improve performance in sprints with
changes of direction (Young et al., 2001b). In this
rare training study comparing sprint training with
CODS performance (Young et al., 2001b), no
significant improvements were reported in CODS
performance after a chronic period of sprint training.
If sprinting and sprinting with directional changes
were strongly related, and if speed had a causal
relationship with change of direction speed, the sprint
training intervention used by Young et al. (2001b)
would not only have improved straight speed (as was
reported) but also change of direction speed. Gen-
erally, the more changes of direction, the less the
transfer from the straight sprint training to change of
direction speed. Also, CODS training had limited
transfer to straight speed, providing clear evidence for
the specificity of speed and CODS training (Young
et al., 2001b). It might be hypothesized that straight
sprint training would contribute even less to perfor-
mance in an agility test that requires decision making.
Another consideration that is releva nt to field and
court sports involving complex skills (running with a
ball, dribbling, etc.) is that sprinting while perform-
ing a skill further increases the complexity of the task.
This increase in complexity affects an athlete’s
performance, as evidenced by weak relationships
between straight sprinting ability and the ability to
perform complex tasks such as dribbling a basketball
(Tsitskarsis et al., 2003) or bouncing an Australian
Rules Football (Young et al., 1996). Based on this
consideration, tests and training that address skill
demands could increase validity.
Leg strength qualities and change of direction speed
Many strength and conditioning coaches bel ieve that
strength and power measures and sprinting perfor-
mance are strongly linked (Blazevich, 1997a, 1997b;
Johnson, 1996; Luchtenbern, 1990; Sheppard, 2003,
2004), as correlations in the literature generally
suggest moderate to strong relationships (Baker,
1999b; Young, McLean, & Ardagna, 1995; Young
et al., 1996). However, as noted previously, straight-
sprinting speed and speed while changing direction
appear to be distinct physical qualities (Buttifant
et al., 1999; Draper & Lancaster, 1985; Young et al.,
1996, 2001b). Therefore, one cannot infer that the
