Essentially, performance analysis (PA) is about creating a valid and reliable record of performance employing systematic observations that can be analysed with a view to facilitating change. In the absence of such an approach, coaches are liable to form biased opinions of their athletes’ or players’ performances, which may lead to incorrect substitution decisions or training prescriptions.
The process of PA relies on two distinct sports science disciplines:
- Notational/match analysis, which uses means to record aspects of team performance;
- Biomechanics, which revolves around the sporting impact of body movements.
These two areas of interest practise similar methods of data collection and rely profoundly on IT for data analysis. The main similarity however, involves the use of measured observation during or after an event to quantify performance in a reliable and valid method.
Hand notation systems are generally very accurate however require considerable learning time, particularly for complex systems. Furthermore, the data must then be translated into a meaningful form for the coach or athlete. It may take up to forty hours to process the data from one squash match.
Computerised notation however can tackle these problems effectively. Data processing may be accelerated and accessed immediately. Coaches can easily understand the presentation of data in a graphical format. The increasing sophistication, and reducing cost, of video systems has greatly enhanced post-event feedback, from playback with subjective analysis by a coach to detailed objective analysis by means of notation systems (Brown and Hughes, 1995).
Within any given sporting event, strength, stamina and motivation play a crucial role as improved training regimes bring athletes closer to the limits of what is physically possible. The difference between winning and losing becomes increasingly hinged on technique, for example the block start of a 100m sprint. Computers simply serve to optimise these techniques to the advantage of the athlete.
PA is a relatively new concept having only established itself in the 1990’s. Facilitated by advances in information technology and digital photography resources, PA is now acknowledged to enhance performance at all levels of sporting participation. Coaches are now considering the demand for PA as a separate branch of sports science given that observation and analysis clearly constitute an integral part of the training process.
Information technology has touched nearly every aspect of our lives and has become increasingly involved in sport. Computers are now an essential part of designing sports equipment from training shoes to golf clubs. But now even the human touch of the sports coach is being challenged by technology.
Coaching athletics and gymnastics have always been one of the most demanding coaching jobs and as the focus shifts to the finest details of technique, the power of the computer can prove the deciding factor between a gold or silver medal.
The movement of the human body underpins much in sport science. Until recently, an athlete’s movement could only be scrutinised by the eagle eyes of the coach. The arrival of the video camera however changed this. Slow motion playback and provided a new tool for analysing the smallest movements that influence performance. Over the last five years, improvements in the speed and graphical capabilities of computers have taken this work a stage further. A computer can now analyse a video film of an athlete at work and for the first time in history, coaches are able to construct detailed three dimensional simulations of exactly how the human body moves.
An athlete’s body is modelled as a linked system of segments and in any set activity e.g. jumping the hurdle or the pole vault, their actual bodily movements can be compared with the optimum movements need to perfect the jump or hurdle. Their training movements can then be analysed repeatedly until the athlete executes the technique exactly.
Computer models of more complex gymnastic movements, like a twisting somersault, can also be used to help develop exercises that will promote the co-ordination and movement skills necessary for an athlete to perform the movement. These exercises can be perfected before trying the final movement, making it much easier to master and dramatically reducing the risk of injury.
The simple model (fig.1.) of the coaching process can be extended (fig.2.) to include the recent developments in computer and video technology.
Fig.2. A more Complex Scheme of the Coaching Process
The concept is used to identify and measure a range of ‘performance indicators’ that have a real bearing on the outcome of any given sport, thus providing a better understanding of how success can be achieved.
These include:
- tactical indicators (patterns of play)
- technical indicators (technique/performance)
- physiological indicators (intensity profiles)
- psychological indicators (arousal, motivation)
Prior to deciding which performance indicators require attention in analysing the performance of athletic performance, consultation of a technical expert in the particular sport may be deemed necessary to identify indicators that are known to contribute to successful performance. These chosen indicators will define the design of the PA system itself.
Initially, the most important step is to create a logical structure to the game itself. This involves defining a range of possible actions within the game and linking these actions with possible outcomes, thus describing the sequential path the performance can take. The identification of these sequential paths permits comparisons that may lead to emerging patterns. For example, the lob shot in a tennis match may prove much more successful than the slam.
Performance analysts have tended to focus on tactical and technical indicators and, in doing so, have contributed to the understanding of the physiological, psychological and tactical demands of many sports (). For example, in golf, performance may be assessed by the number of greens hit in regulation.
Indicators provide simple information that can be used to describe and define a particular performance. The presentation of data however must be used in context, since in isolation, can give a distorted impression of performance. For example, striker A and striker B have both scored four goals in four games. It would be easy to assume both are performing well. However, if striker A has sixteen shots on goal to striker B’s eight shots, the formers success ratio is 4:1 compared with a much more impressive ratio of 2:1 for the latter.
Performance comparisons of individuals and teams are straightforward and accurate if performance indicators are expressed in terms of ratios. These may take the form of possession to turnovers, winners to errors, and passes taken to passes completed.
An individual’s performance profile may also be rendered distorted if the correct comparisons are not made. Many profiles differ according to the opposition. For example, presenting an individual soccer midfielder’s performance may be misleading without a comparison with the opposing midfielder’s data.
Performance data for an individual can be presented in three ways to evaluate success (3):
- In relation to the opponent’s data. This allows for a direct comparison with the opposition, but could be misleading if the players are not of a similar standard;
- In relation to players of the same standard. This allows for comparison between equals, which is useful providing data of this nature is available or can be compiled;
- In relation to their own profiles of previous performances. Over the course of a number of competitive matches, a normative profile of a player or team can be created for comparative purposes. A player can then be assessed against his own normative profile to assess the relative merits of his latest performance.
There are two main methods of coding the observations made within a sporting situation: ‘live coding’ and ‘post play coding’. The former requires a high degree of competency in coding a sporting situation, with video footage fed directly into a laptop and coded via the keyboard as events unfold within the training session or game. This method is particularly useful where information is required on the spot, for example to correct unique body movements thus preventing injury. Post play coding is similar to live coding, in the information being presented on the laptop; only information may be slowed down or reviewed repeatedly to ensure a high degree of accuracy in observation.
In reality, the entire process of PA can be very successful and provide invaluable information the coach. To confirm this, I conducted my own mini research into analysing the performance of a football striker. My objective as discussed with the team coach was to improve feedback to the player regarding their individual performance.
The first stage of the analysis focused on the role and function of the centre forward to obtain logical understanding of the striker’s involvement within the team’s tactics.
Using video-recording equipment, I filmed the player on a ‘player cam’ basis for two matches. The study conducted remained covert to ensure player movements were not due to the awareness of analysis, known as the Hawthorne effect. Overt study would enable the player to manipulate results rendering them invalid.
An emerging tactic that occurred frequently throughout the two games involved the striker gaining possession of the ball, laying the ball off to the midfielder’s who would transfer play to the wings resulting in a cross or through ball for strikers to achieve a shot on goal. It became obvious that this form of play was heavily dependent on the striker’s play, and essentially formed the key to attacking strategies. This seemed an obvious focal point for analysis (fig.3.).
-
Possession → won/lost? → How won/lost? → Position on pitch?
-
Response → pass/lose possession/foul play? → How? → Position on pitch
-
Shot on goal → header/foot? → On target/off target → Successful/unsuccessful?
Fig.4. Movement Sequences that required analysis
The pre analysis consultation involved identifying a performance profile for the player. This was constructed on the path (fig. 3.) and the movement sequences (fig. 4.) which would expose the strengths and weaknesses of the individual player.
I found myself in the fortunate position of having access to computerised notational match analysis software package Nordulus Observer Pro. Coding the matches manually post competition, and the results obtained from the analysis were relayed to the coach. This data was then communicated personally to the player along with recommendations for improvement.
The performance profile, built from the analysis emphasised the striker’s response upon gaining possession of the ball. Possession was gained 26 times during the game (Table 1.)
Table 1. Player’s Varying Response
The analysis also identified the number of header’s won/lost, shots on and off target and the number of times possession was won and lost.
- 8 headers won, 3 in midfield, 3 in attack and 2 in defence;
- 7 headers lost, 4 in midfield and 3 in attack;
- 6 attempts on target, 2 headers and 4 with the foot, 2 successful strikes;
- 3 attempts off target, from a header and 2 strikes;
- Possession won 7 times; 5 through closing down, 2 winning tackles;
- Possession lost 2 times in midfield through being closed down.
The performance profiles identified the personal strengths and weaknesses of the individual players, providing a technical focus for future training sessions. On presenting this information to the coach, it was suggested that training sessions should be skill specific, in aiming for enhanced performance.
- Work on the players’ ability to maintain possession of the ball
- Improve the link up play with midfield players to help decrease the number of possessions lost and maintain fluency within the attack;
- Work on converting goal-scoring opportunities into goals, particularly headers.
- Work to the strikers’ strengths of making successful use of possession when the ball is played into their feet.
The points were incorporated into the training sessions, and I hope to analyse one final game in the near future to establish the value of the process. Other research in this area however has demonstrated significant improvement by the player’s and substantial progress towards individual goals (Bishop, 2003).
The coach agreed the project had been a success and had highlighted weaknesses they hadn’t previously been aware of. The player responded well to the feedback and targets and the process gave rise to a second project involving the central midfielders. The results of the analysis can be retained for future reference and even form a basis for future research undertaken in this particular area.
To conclude, as new world records are measured in hundredths of a second and fractions of a centimetre, the use of computers in perfecting performance will only increase. Countries and individual teams which can put large amounts of computing power at the disposal of their athletes will have a distinct advantage over those who can’t. And just as Formula One motor sport as drawn the technological line the preserves some real competition, other sporting bodies may well consider the need to bring in similar legislation in the future.
Journal of Sports Science (1991), 9 (3) pp285-297
Journal of Sport Behaviour (1986), 9, pp34-45
Journal of Sports Science (2002) 20, pp739-754
References
-
Coakley, J.J. (2001) Sport in Society: Issues and Controversies. 7th Edition.
-
Magill R.A (1998) Motor Learning: Concepts & Applications. 5th Edition. McGraw Hill.
-
Performance Coaching: The Individualisation of training programmes (2004) http://www.coachesinfo.com/article/8/ [accessed 5th February, 2004].
-
Schmidt A, Wrisberg A (2000) Motor Learning and Performance. 2nd Edition.
-
Training Principles (2003) http://www.brianmac.demon.co.uk/trnprin.htm [accessed 4th February, 2004].
-
Wesson K, Wiggins N, Thompson G, Hartigan S (2000) Sport and PE. 2nd Edition.