Ramana Devi A/P Ananthan

M20152001096

Article Review - In search of invariant athletic behaviour in sport: An example

from championship squash match-play

Author (APA Style)

McGarry, T. & Franks, I. M. (1996). In search of invariant athletic behaviour in sport:

An example from championship squash matchplay. Journal of Sports
Sciences, 14(5), 445-456.

Introduction

Research of invariant athletic behaviour in sports was carried out by McGarry and
Franks (1996) to identify important invariant playing of athletic behaviours in sports.
Identification of invariant responds of athletes can help in improving future athletes
performance in sports. Now lets have a look on the review about this research in
detail.

Purpose of the study

This research was carried out to identify important invariant playing of athletic
behaviours in sports. This kind of analysis of competitive sport performance from
priori observation is frequently used to prepare an athlete for upcoming contests.
This practice assumes that the observed athletic behaviours are invariant, in that the
data from earlier competitions could be applied to the next.

Instrument

Various data collecting instrument was used in this experiment such as transcribing
and observation. For making an accurate observation researchers used checklist so
that they can get enough data to proceed the research. Other than that, video
recording tools and also computer software was used to identify invariant athletic
behaviour in sports. In addition, a weighted kappa coefficient (K"W) was used as the
dependent psychometric measure in both studies.
 Ramana Devi A/P Ananthan
M20152001096

Research Participants / Subjects

This research involves eight international male squash players (RM, JK, AK, CD,
RN, RT, TN, DH). Athletic behaviours of these eight players was analysed from
semi-finals and final of the 1987 World Mens Open Championship and the quarter-
finals, semi-finals and final of the 1988 Canadian Mens Open Championship.

Data Collection

The data were recorded from a set of 15 shot types, 16 position cells and three rally
outcomes such as winner, error and let according to the following rally structure:
Player - Position - Serve - Position - Shot - Position - Shot - Outcome. The server of
the rally was identified, as where his initial court location from where the ball was
played and the corresponding stroke. The position where the opponent's shot
response occurred, as well as his corresponding shot, was then recorded and this
process repeated until a rally outcome was assigned to its associated shot. All rallies
from a match were recorded in this manner and analysed after establishing that the
data collection procedures were both reliable and valid.

Other than that, a weighted kappa coefficient (K"W) was used as the
dependent psychometric measure in both studies and set at a criterion level K"w >
0.80, which indicates very high agreement for behavioural observation studies. Both
studies met this criterion in each of their measures of position, shot and outcome.

Data Processing

The 15 shot types were collapsed. Straight and cross-court conditions were
collapsed for each shot type and volley-boasts were collapsed with short volleys.
This led to seven distinct shots such as serve, drive, boast, drop, lob, volley(short),
volley(long) which in conjunction with court location or that shot's own antecedent,
formed the preceding shot conditions. These distinct shots, except for the serve, also
formed the shot responses, which were then further categorized according to their
response function to the preceding shot.
 Ramana Devi A/P Ananthan
M20152001096

Results and discussion

According to the researchers, different analyses of championship squash matches

from 1987 and 1988 in search of invariance and thus general signature of athletic
behaviour. The results show that once the previous condition to an athletic response
become more detailed, by either accounting for where on court the previous shot
was played to, or by pairing the previous shot with its own antecedent, the
corresponding response becomes more predictable (P > 0.25). This is the case for a
squash champion when the forehand and backhand characteristics of his shot
response are accounted.

In summary, these results support the interpretation that a further detail of the
preceding shot condition, be it on court location or the antecedent shot, elicits an
increased invariant response from each player, although the degree of response to
each preceding condition differs according to the particular player. This leads to the
interesting possibility that invariant playing signatures are produced by different
players in response to different preceding conditions. These results demonstrate the
importance of qualifying the analytic appropriately.

Analysis used

The 1987 and 1988 match data were grouped by player and by year to provide some
protection against sampling error. This research was also involving two way x2
analysis for inference of stability of shot response to the preceding condition. The
two independent variables were shot response, year and the dependent variable was
the frequency of shot response. Two-way x2 analyses were then conducted for each
preceding condition as determined by the particular analysis and the available data.

The x2 analysis tests for evidence of differences between data sets, in this
case shot response profiles. If, as in this study, a failure to detect any evidence of a
difference is to be interpreted as evidence of no difference then it is necessary to
inflate P from traditional values. A critical P-value of 0.25 was selected in keeping
with that of earlier studies (McGarry and Franks, 1994). A value P> 0.25 was
therefore interpreted as an invariant response and P < 0.25 interpreted as a variant
response. The same data were analysed in different ways and various 'playing
profiles' were established for each player. The various analyses differed in their
 Ramana Devi A/P Ananthan
M20152001096

preceding conditions by some qualification of the previous shot, either through

tagging the ball's subsequent court position or through prefacing the previous shot
with its own antecedent. A player's 1987 and 1988 playing profiles were then
analysed for consistency of shot response to the particular preceding condition using
two-way x2 analysis. A playing profile comprises a two-dimensional frequency matrix.

Research Findings

The finding of this research shows that the champion is more consistent if his
forehand and backhand shot responses to a preceding shot are analysed separately
shows that the level of analytic detail is an important consideration if the derived
information from sport analysis is to prepare an athlete reliably for future sport
competition. Last but not least, we consider sport competition from an ecological
perspective with regard to the natural organization and stability of non-linear systems
in an attempt to explain the empirical data.

Conclusions

The application of objective data to inform the coaching process may therefore, in
itself, be of little value. It is immediately evident, for example, that winners and errors
discriminate successful performance in squash and tennis and lead to sport
accomplishment. Our task is to find the critical features that reliably produce these
behaviours. Thus the main priority of sport analysis is to identify the variant and
invariant features of sport performance if athletic behaviour in future sport
competition is to be explained and optimized.