Secondary Logo

Journal Logo

Original Research

An Evaluation of the Physiological Demands of Elite Rugby Union Using Global Positioning System Tracking Software

Cunniffe, Brian1,2; Proctor, Wayne2; Baker, Julien S1,3; Davies, Bruce1

Author Information
Journal of Strength and Conditioning Research: July 2009 - Volume 23 - Issue 4 - p 1195-1203
doi: 10.1519/JSC.0b013e3181a3928b
  • Free

Abstract

Introduction

Rugby union is an intermittent high-intensity sport, in which activities that call for maximal strength and power are interspersed with periods of lower intensity aerobic activity and rest (26). Until recently, it has not been possible to collect objective data on player work rates in situ other than via use of heart rate (HR) monitoring. Traditionally, the majority of studies have investigated game demands in rugby union through time-motion analysis systems incorporating the use of game video recordings (24,9,10,14,30). Problems using video recordings may occur however as a result of errors involved in categorization of locomotor activity. This is important because rugby union is a dynamic intermittent sport with many gait changes during game phases. Furthermore, notational analysis systems are largely dependent on trained users, and considerable subjectivity may exist when interpreting data. Therefore, accurate performance assessment may be technically difficult given the complex interactions of players and the varied nature of game play.

With the development of Global Positioning System (GPS) technology for use in sport, investigators can now evaluate training loads and activity profiles of players on the field. This is achieved via use of portable tracking devices, which permit quantitative measurement of activity profiles through traditional GPS triangulation methods and accessory accelerometer software. Positional data are normally achieved by comparing the signal travel time of radio frequency signals sent from the orbiting satellite and the GPS receiver worn by the player/athlete. The distance to the satellite is then calculated by multiplying the signal travel time with the speed of light. By calculating the distance to at least 4 satellites, the exact position can be trigonometrically determined (19). Changes in speed (velocity) are usually determined via the Doppler shift method, that is, measurements of the changes in signal frequency due to movement of the receiver. With this, an opportunity exists to gain valuable data on game demands in team sports like rugby union through objective distance and speed calculations not previously available using the same recording system. Through investigation of game demands, training methods can be targeted to mimic the positional requirements and physiological loads, thus optimizing player conditioning to the energy demands of the sport.

“On the field” GPS technology has been used previously in Australian Football League to profile positional demands and enhance knowledge on injury potential. To the author's knowledge, these data have not been published within scientific literature. Limited data using this type of technology exist in rugby union despite its initial use within domestic southern hemisphere rugby. Furthermore, evaluation of game demands at the elite level is currently difficult as such communication devices are not allowed during competitive league and cup games.

With this in mind, the aim of this pilot study was to gather information on rugby union forward and back play at the elite level and demonstrate the potential use of GPS technology in the assessment of the games' physiological demands. It should be noted that this study focuses on a small subject number, the goals being to provide some insight into the contemporary demands of rugby union with an additional focus on future research. An improved understanding of the game and training loads may help facilitate best practice advice for player management issues and appropriate training periodization.

Accuracy of Global Positioning System Technology

Limited data exist on the accuracy and reliability of current GPS software estimation of true distance and velocity. In a study using a generic GPS logger, it has been stated that errors in data logging may increase during activity over circular paths most likely because of underestimations in speed (35). A recent study comparing current GPS software with timing gates in soccer demonstrated that both methods produced comparable speed and distance data in a linear running protocol (28). In the same study, it was concluded that GPS data recording at 1 Hz seemed appropriate for calculating distance at lower velocities but that greater error in estimation may occur at higher velocities. The accuracy of this system in calculation of distances traversed has been previously shown to be within 4.8 (15) and <1% (22) in estimation of true distance measured using a trundle wheel. In the latter study, GPS technology was shown to be accurate in the assessment of speed (within 0.01 m·s−1 of true value). Limited data exist on accuracy of GPS in calculation of distance and speed in field activities requiring repeated changes in running intensity.

Methods

Experimental Approach to the Problem

To examine game play at the elite level, data in this study were taken from players during an out-of-season competitive 80-minute game. This involved 2 teams normally participating in the Celtic (Magners) league and Guinness Premiership, respectively. Both leagues represent the highest standard of club play among the Celtic nations (Ireland, Scotland, and Wales) and England.

Subjects

Data were obtained from 3 home team players (age, 25.0 ± 3.6 years; weight, 104.6 ± 10.4 kg; height, 193.3 ± 9.7 cm; o2max, 53.3 ± 2.1 ml·kg−1·min−1; mean ± SD). These included 1 back (out-half) and 2 forwards (back row and lock). However, as one of the players (lock forward) only participated in a quarter of the game, discussion will concentrate on the 2 players with full data sets. Forward and back players were selected in an attempt to investigate game play differences for different playing positions. Before participation in the study, players provided informed consent and were made aware of their ability to withdraw from testing at any time. Ethical procedures for the study were obtained from the University of Glamorgan Ethics Committee. All players were fully habituated and familiarized with the data collection systems. This was done on several occasions during training sessions before the actual game itself.

Procedures

Players were asked to wear an individual GPS unit (mass: 80 g; dimensions: 91 × 45 × 21 mm) encased within a protective harness between the player's shoulder blades in the upper thoracic-spine region (estimated). Players also wore a HR transmitter belt (Polar Electro, Kempele, Finland) to incorporate HR data. This was recorded synchronously (1-second intervals) with the GPS tracking device (SPI Elite; GPSports Systems, Canberra, Australian Capital Territory, Australia). Devices were switched on 5 minutes before the start of the game and turned off immediately after the game had ended. Data stored included time, velocity (calculated via Doppler shift), distance, position, direction, HR, and number and intensity of player impacts as measured in “g” force. GPS data were recorded at 1 Hz and accelerometry (tri-axis) data at 100 Hz, respectively. After collection, data were downloaded to a personal computer where further analysis was carried out via use of the system software provided by the manufacturer (Team AMS; GPSports, V1.2).

Measures

Heart Rate and Locomotor Activity

Recorded game HRs were categorized into 6 HR zones based on each player's known maximum HR (HRmax) monitored using an incremental treadmill running test in the laboratory. HR zones were as follows: (a) 0 to 60% HRmax, (b) 60 to 70% HRmax, (c) 70 to 80% HRmax, (d) 80 to 90% HRmax, (e) 90 to 95% HRmax, and (f) 95 to 100% HRmax. Total values for HR exertion were achieved using a similar weighting system carried out by Edwards (16).

Frequency and duration of locomotor efforts were evaluated from the time spent in 6 player speed zones. Allocations of speed zones were those thought typical of varying locomotor categories during intermittent team sport. These were as follows: (a) standing and walking (0-6 km·h−1), (b) jogging (6-12 km·h−1), (c) cruising (12-14 km·h−1), (d) striding (14-18 km·h−1), (e) high-intensity running (18-20 km·h−1), and (f) sprinting (>20 km·h−1). The above categories were later divided into 2 further locomotor categories to provide a crude estimate of player work to rest ratios: (a) low-intensity activity (0-8 km·h−1) and (b) moderate- and high-intensity activity (>8 km·h−1). This categorization was based on data obtained from a similar previous study in Australian rules football using GPS software (34).

Body Load and Game Impacts

Player impact data (intensity, number, and distribution) were gathered from accelerometer data provided in “g” force. Intensity of impacts was graded according to the following scaling system provided by system manufacturers: 5-6G: light impact, hard acceleration/deceleration/change of direction; 6-6.5g: light to moderate impact (player collision, contact with the ground); 6.5-7g: moderate to heavy impact (tackle); 7-8g: heavy impact (tackle); 8-10g: very heavy impact (scrum engagement, tackle); and 10+g: severe impact/tackle/collision.

Computation of player body load during exercise also involved use of the above acceleration zone forces. Body load was calculated automatically using the system software provided by the manufacturers.

Estimation of Energy Expenditure

Information regarding estimation of energy expenditure (EE) was obtained from continuous measurement of HR during the game. In turn, corresponding values for o2 were estimated from player's individual HR-o2 relationships (2), which were obtained during a standard incremental running protocol (o2max test) collected on the players in our laboratory. Rates of energy expenditure were subsequently calculated using methods similar to those previously shown in soccer (3,17) and rugby league (8). Non-playing periods (halftime, warm-up, and cooldown) were omitted from all analyzed GPS and energy expenditure data.

Statistical Analyses

Given the nature of the present investigation and small subject number, data presented below are of a descriptive nature only. Where appropriate, the duration for each activity is presented as the mean and SD.

Results

Heart Rate

Players recorded mean and peak HRs of 172 and 200 b·min−1, respectively, during the game (Figure 1). Both players reached preestablished maximum heart rate (HRmax) during the game. The back spent more time at 80 to 90% HRmax (42%) than the forward (27.7%), whereas the forward spent more time at 90 to 95% HRmax (15.4%) than the back (4.7%) (Table 1). Mean HR was higher in the first half than the second half (173 vs. 169 b·min−1; Table 2).

T1-21
Table 1:
Percent (%) time spent in each HR zone.*†
T2-21
Table 2:
Game data per position and half.*†
F1-21
Figure 1:
Graph showing HR data at 1-second intervals per half for both players. Values are expressed as percentage (%) of maximum heart rate (HRmax). HR = heart rate.

Motion Analysis

During the game time of 83 minutes, 72% was spent standing and walking, 18.6% jogging, 3.3% cruising, 3.8% striding, 1% high-intensity running, and 1.2% sprinting (mean data for both players). These values represent a work to rest ratio of 1:5.7. Players covered on average 6,953 m during the game (Table 2). Of this distance, 37% (∼2,800 m) was spent standing and walking, 27% (∼1,900 m) jogging, 10% (∼700 m) cruising, 14% (∼990 m) striding, 5% (∼320 m) high-intensity running, and 6% (∼420 m) sprinting (Table 3). The majority of moderate to intense accelerations occurred over running intervals of 4 to 6 seconds with little difference between player positions.

T3-21
Table 3:
Distance (meters) covered by each player in designated speed zones.*

Acceleration data (Table 4) refer to the number of times the players changed velocity in defined categories over 1-second time intervals. Changes in velocity over 1.5, 2.0, 2.5, and 2.75 m·s−1 correspond to changes in running speed of 5.4, 7.2, 9, and 10 km·h−1 in 1-second intervals, respectively.

T4-21
Table 4:
Game acceleration and speed data per position and half.*

During the game, the players encompassed 742 changes in tempo, occurring approximately every 3 to 4 seconds. The back entered the high-speed zone (>20 km·h−1) on a greater number of occasions (34 vs. 19) than the forward (Table 5). In turn, the forward entered the lower speed zone (6-12 km·h−1) on a greater number of occasions than the back (315 vs. 229) but spent less time standing and walking than the back (66.5 vs. 77.8%).

T5-21
Table 5:
Number of entries/surges by players into each speed zone.*

Players reached maximum speeds of 28.7 km·h−1 (back) and 26.3 km·h−1 (forward), respectively. Peak speeds for both players occurred during the second half (Table 4). Both players' work to rest ratios (average, 1:5.7) were also lower during the second half (Table 2), indicating less recovery time (i.e., time spent below 8 km·h−1) between play periods. Average player running speed over the game was 4.2 km·h−1, values greater during the second half for both players (Table 4). Within-half comparisons revealed that values for maximum speed, average speed, total distance covered, and peak HR were higher at the start of the each playing (20 minutes) quarter (Q) such that values for Q1 > Q2 and Q3 > Q4 (both players). Cross-quarter comparisons revealed that values for the above variables were highest during the third quarter of the game, that is, first 20 minutes after halftime.

Body Load and Game Impacts

Both players received a large number of impacts during the game with positional differences observed between the number of impacts received by the back and forward (798 vs. 1,274). Grouping of game impacts within the latter 3 categories (heavy + very heavy + severe) revealed that the forward was involved in 60% more high-level impacts than the back (Table 6). Furthermore, 66% of the high-level impacts received by the forward occurred during the second half. This resulted in greater overall body load and body load per minute for the forward player (Table 6).

T6-21
Table 6:
Game impact and body load data per position and half.*†

Discussion

To the author's knowledge, this is the first study to evaluate player demands during a competitive game of rugby union using objective “on the field” software. During the game, players covered an average distance of 6,953 m (83.7 m·min−1) (Table 2). These values are less than distances reported in professional soccer players (118 ± 7.5 m·min−1) using similar GPS technology (1) but greater than previous estimations of running distance in rugby union (10,30). Similar to the study by Roberts et al., the present data suggest that backs travel greater total distances during a game than their forward counterparts (7.6% further in the current study). Interestingly, both players recorded greater running distances in the second half of the game (6.7% back, 10% forward), indicating that deterioration in running ability or, perhaps, depletion of energy reserves was not an issue in this player group. During game activity, players performed 87 moderate-intensity runs (>14 km·h−1) (18) over an average distance of 19.7 ± 14.6 m. Along with running a greater total distance, the back entered the high-speed zone (>20 km·h−1) on a greater number of occasions (34 vs. 19; Table 5) than the forward. The back also covered a greater total distance sprinting (>20 km·h−1, 524 vs. 313 m) when compared with the forward. Not surprisingly, values for total sprinting distance observed in the current study (elite senior players) are substantially greater than those reported previously in elite U19 rugby for backs (253 ± 45 m) and forwards (94 ± 27 m) (10). Overall, the data would suggest that backs participate in a greater amount of higher intensity locomotor work when compared with forwards, although the forward was found to cover a greater average distance per sprint burst activity (15.3 m back, 17.3 m forward) in the present investigation.

Several studies have demonstrated that estimated total work performed (quantified by HR and movement patterns) is lower for backs than forwards (24,10,11). This is thought to occur despite the fact that forwards spend more time in the lower speed zones. Data in the present study revealed that the forward entered the lower speed zone (6-12 km·h−1) on a greater number of occasions than the back (315 vs. 229; Table 5) but spent less time standing and walking than the back (66.5 vs. 77.8% of total time; Table 7). It should however be noted that the percentage of time or effort exerted by the forwards in static activity and tackling was not measured in the current study and such activity would be anticipated to significantly contribute to game workload in this player group (30). Overall, the above findings demonstrate that the back participated in more anaerobic high-intensity activity interspersed with longer recovery periods in the lowest speed zones, whereas the forward spent more time in the moderate speed zones as recovery time between high-intensity activities. This may have implications for positional specific training requirements.

T7-21
Table 7:
Percent (%) time spent by players in each speed zone.*

The typical sprint distances of 15 to 20 m in the current study and number of intense accelerations (Table 4) imply that the ability to accelerate quickly is highly important in professional rugby union. Of interest, the majority of intense accelerations did not occur from standing starts (0 km·h−1), implying that quick changes in player running gait are of essence in game performance. During the game, approximately 10% of game time was spent performing intense locomotor activity. This corroborates previous studies using less objective methods of analysis (9,10). Because the current study did not include time spent performing intense game-specific efforts and utility movements, it is likely that this figure is greater in modern day rugby union. Nevertheless, the longest continuous time recorded above speeds of 20 km·h−1 was just 7 seconds (46.6 m), with the majority of high-intensity work periods below 6 to 7 seconds in duration. These data support previous findings where typical sprint distances of just 10 to 20 m have been shown (10,11). Such findings point to the contribution of the anaerobic energy system, in particular the phosphocreatine system during high-intensity activity of which is interspersed with long periods of lower intensity activity, primarily aerobic in nature. Interestingly, there was a general trend for both players' speed and distance to decrease after intense 5-minute play periods. It is not known what effect, if any, interval or aerobic type training may have on repeatability of intense efforts during rugby play as measured through GPS technology. This is important because improved aerobic fitness has been implicated in sprint and recovery and fatigue resistance (6,25,32,33). Future investigation using a larger data set on typical work rates in elite rugby union may be of interest. With this, it may be possible to determine minimal requirements in aerobic fitness and role of o2 kinetics in recovery from short-term anaerobic efforts.

Previous research has shown that despite a disparity in distances (assessed through 10-minute intervals) covered between game halves, a rugby player's ability to perform high-intensity activity with increasing game duration is not limited (30). In the present study, total running distance, peak running speed, number of peak accelerations, average speed, and peak HR were all greater during the second 40-minute play period. Cross-quarter (∼20 minutes) comparisons revealed the third playing quarter (after halftime) to be most intense. This was reflected by lower work to rest ratios and high-intensity (>18 km·h−1) running distance per minute. Interestingly, these results compare favorably with previous work, which has shown that most injuries occur in the third quarter of the game (5). These data suggest that player fatigue was not a factor between halves. Indeed, of the high-level impacts undertaken by the forward, 66% of them occurred during the second half of play. Furthermore, players reached maximum speeds of 28.7 km·h−1 (back) and 26.3 km·h−1 (forward), both of which occurred during the second half. Expressing these values as a percentage of individual peak running speed (measured using GPS 2 weeks before the study) revealed that both players were capable of reaching 92% of their peak running speed. This shows that players do reach maximum levels of locomotor activity during rugby play despite previous game activity.

In this study, the average work to rest ratio during the game was 1:5.7, indicating that for every 1 minute of running, there was almost 6 minutes of lower intensity activity. Figures drop for both players during the second half, indicating that play periods were more frequent with less recovery (Table 2). Although work to rest ratios provide important information on demands of the sport, in the case of rugby union, player work to rest ratios calculated from player locomotor activity may underestimate actual work time. Considerable time spent in specific game-related phases such as pushing/pulling in rucks/mauls/scrums may register as low-intensity activity using current GPS technology despite intense static player efforts. Although the above ratio provides information on the intermittent nature of elite rugby union, it may not provide a true reflection of player work rates, in particular for forwards. Combining objective GPS data with qualitative analysis of time spent in non-running exertion and utility movements may help in establishing more defined work to rest ratios and setting of fitness goals.

Although the game in the present investigation was outside normal competition, it was played between 2 of the topsides normally participating in the Celtic League and Guinness Premiership. The game occurred at the end of preseason training before commencement of the regular club season. Both teams contained a large number of first-team regulars, were evenly matched, and the game served as an important element in seasonal preparation. Therefore, data, although limited by subject number, do provide some insight into game demands at the top level in European rugby. Our results suggest that players exercise at ∼80 to 85% o2max (Table 8) during the course of the game. This is similar to those reported in rugby league (80% o2max) (8) and higher than values observed in Gaelic footballers (∼72% o2max) (29) during competitive matches. It is possible that elevations in HR may have overpredicted aerobic demand because it has been suggested that changes in HR may not accurately reflect changes in energy cost occurring over short-term high-intensity activities (20). Factors other than oxygen uptake such as environmental temperature, emotions, continuity of exercise, and perhaps more importantly muscle actions and body position can influence HR response to exercise. With particular reference to rugby, players are required to exert forces dynamically and statically during various game activities. These activities often demand both upper- and lower-body musculature, for example, during scrummaging, rucking, and mauling. In such cases, elevations in HR may not accurately predict actual oxygen uptake. Indeed, it has been shown that when muscles act statically in straining-type exercise, HRs are consistently higher compared with dynamic leg exercise only at a particular oxygen uptake (21,27). This should be taken into account when prescribing training drills based on HR. Nevertheless, measurement of game HR does provide a useful index of overall physiological strain. Mean game HR was 172 b·min−1 (∼88% HRmax; Figure 1), higher than values of 166 ± 10 b·min−1 recorded within semiprofessional rugby league (8).

T8-21
Table 8:
Estimated values for energy expenditure, oxygen consumption, and percentage of maximal oxygen consumption (Table 1o2max) between position and half.*

Extrapolation of laboratory-based HR-o2 relationships in the EE during intermittent activity has been previously shown to reflect metabolic expenditure during soccer activity (17). Furthermore, the HR-o2 regression has been shown to be a good predictor of aggregate responses to irregular exercise including vigorous anaerobic activity (7). Using this method, data above (Table 8) show that estimated values for EE were 6.9 and 8.2 MJ for the back and forward, respectively. These values correspond to ∼13 metabolic equivalents and are similar to those reported in semiprofessional rugby league (7.9 MJ) but ∼25% greater than those reported in professional soccer players (31,3). It is perhaps not surprising that EE is so high in professional rugby union given the nature and intensity of the game, involvement of total body musculature, and, most importantly, player size. This is significant because energy cost of locomotion increases directly with increasing body mass (23). In this study, players weighed 92 kg (back) and 107 kg (forward), respectively. Nevertheless, potential errors in estimation of EE using this method may occur and have been the subject of recent attention (13). This “averaging out” approach may be criticized on the basis that the regression line is based on steady state responses, conditions not found in intermittent sport (12). Consequently, the current EE data should be regarded as a crude estimation only and does not take into account resting energy expenditure. Nonetheless, the data suggest that replenishment of energy after a game is of great importance. This has obvious implications for maintenance of muscle mass during a season and for replenishment of energy reserves between games. Previous research has shown that there is a 2-day delay in muscle glycogen replenishment after a game of soccer despite administration of a high carbohydrate diet (4). Therefore, repeated exposure to heavy exercise should be monitored closely so as to avoid any adverse effects on player well-being.

Practical Applications

The data presented in this case study are of a descriptive nature only, are limited by subject number, and do not reflect variations in game activity/player demands, which may occur within and between participation levels. Further data on players from different playing levels, positions, and teams will help in defining physiological demands and evolutionary trends. Nevertheless, the current report does provide insight into the intense intermittent nature of elite rugby union. These findings seem to confirm that the contemporary rugby union player runs longer and harder than previously thought, data of which have not been described previously using this technology. Such data may have important applications in terms of training replication of game demands, conditioning of player groups, and in the evaluation of overall game stress.

Use of GPS-accelerometry technology offers a valuable insight into physiological demands during match play, not previously available through HR-based collection methods and video analysis. Further work using this technology, in particular detailed analysis of accelerometer and player impact data, may help fitness experts in evaluating player work rates outside that of traditional locomotor activity. Combination of GPS software with game recordings may provide more insight into categorization of forces/accelerations received/exerted during the many contact elements within the game. Appropriate classification of these contact loads may help in devising individual recovery programs specific to the player in question.

Acknowledgments

The author wishes to thank the players and staff of Llanelli Scarlets RFC, Wales, for participation and facilitation of this study. No grant aid was received in conjunction with this work, and no conflicts of interest are declared.

References

1. Álvarez, JCB and Castagna, C. Activity patterns in professional futsal players using global position tracking system. J Sports Sci Med 10: 208, 2007.
2. Astrand, PO and Rodahl, K. Textbook of Work Physiology. New York, NY: McGraw-Hill, 1977.
3. Bangsbo, J. Energy demands in competitive soccer. J Sports Sci 12: S5-S12, 1994.
4. Bangsbo, J, Mohr, M, and Krustrup, P. Physical and metabolic demands of training and match-play in the elite football player. J Sports Sci 24: 665-674, 2006.
5. Bathgate, A, Best, JP, Craig, G, Jamieson, M, and Wiley, JP. A prospective study of injuries to elite Australian rugby union players. Br J Sports Med 36: 265-269, 2002.
6. Bogdanis, GC, Nevill, ME, Boobis, LH, and Lakomy, HK. Contribution of phosphocreatine and aerobic metabolism to energy supply during repeated sprint exercise. J Appl Physiol 80: 876-884, 1996.
7. Bot, SDM and Hollander, AP. The relationship between heart rate and oxygen uptake during non-steady state exercise. Ergonomics 43: 1578-1592, 2000.
8. Coutts, A, Reaburn, P, and Abt, G. Heart rate, blood lactate concentration and estimated energy expenditure in a semi-professional rugby league team during a match: A case study. J Sports Sci 21: 97-103, 2003.
9. Deutsch, MU, Kearney, GA, and Rehrer, NJ. Time-motion analysis of professional rugby union players during match-play. J Sports Sci 25: 461-472, 2007.
10. Deutsch, MU, Maw, GJ, Jenkins, D, and Reaburn, P. Heart rate, blood lactate and kinematic data of elite colts (under-19) rugby union players during competition. J Sports Sci 16: 561-570, 1998.
11. Docherty, D, Wenger, HA, and Neary, P. Time motion analysis related to the physiological demands of rugby. J Hum Mov Stud 14: 269-277, 1998.
12. Drust, B, Atkinson, G, and Reilly, T. Future perspectives in the evaluation of the physiological demands of soccer. Sports Med 37: 783-805, 2007.
13. Dugas, LR, van der Merwe, L, Odendaal, H, Noakes, TD, and Lambert, EV. A novel energy expenditure prediction equation for intermittent physical activity. Med Sci Sports Exerc 37: 2154-2161, 2005.
14. Eaton, C and George, K. Position specific rehabilitation for rugby union players. Part I: Empirical movement analysis data. Phys Ther Sport 7: 22-29, 2006.
15. Edgecomb, SJ and Norton, KI. Comparison of global positioning and computer-based tracking systems for measuring player movement distance during Australian football. J Sci Med Sport 9: 25-32, 2006.
16. Edwards, S. High performance training and racing. In: The Heart Rate Monitor Book. Edwards, S, ed. Sacramento, CA: Feet Fleet Press, 1993. pp. 113-123.
17. Esposito, F, Impellizzeri, FM, Margonato, V, Vanni, R, Pizzini, G, and Veicsteinas, A. Validity of heart rate as an indicator of aerobic demand during soccer activities in amateur soccer players. Eur J Appl Physiol 93: 167-172, 2004.
18. Hartwig, TB, Naughton, G, and Searl, J. Defining the volume and intensity of sport participation in adolescent rugby union players. Int J Sports Physiol Perform 3: 94-106, 2008.
19. Larsson, P. Global positioning system and sport-specific testing. Sports Med 33: 1093-1101, 2003.
20. Little, T and Williams, AG. Measures of exercise intensity during soccer training drills with professional soccer players. J Strength Cond Res 21: 367-371, 2007.
21. Maas, S, Kok, ML, Westra, HG, and Kemper, HC. The validity of the use of heart rate in estimating oxygen consumption in static and in combined static/dynamic exercise. Ergonomics 32: 141-148, 1989.
22. Macleod, H and Sunderland, C. Reliability and validity of a global positioning system for measuring player movement patterns during field hockey. Med Sci Sports Exerc 39: 209-210, 2007.
23. McArdle, WD, Katch, FI, and Katch, VL. Exercise Physiology: Energy, Nutrition, and Human Performance (6th ed.). Baltimore: MD, Lippincott Williams and Wilkins, 2006.
24. McLean, DA. Analysis of the physical demands of international rugby union. J Sports Sci 10: 285-296, 1992.
25. McMahon, S and Wenger, HA. The relationship between aerobic fitness and both power output and subsequent recovery during maximal intermittent exercise. J Sci Med Sport 1: 219-227, 1998.
26. Nicholas, CW. Anthropometric and physiological characteristics of rugby union football players. Sports Med 23: 375-396, 1997.
27. Patterson, R and Pearson, J. Work-rest periods: Their effects on normal physiologic response to isometric and dynamic work. Arch Phys Med Rehabil 66: 349-352, 1985.
28. Portas, M, Rush, C, Barnes, C, and Batterham, A. Method comparison of linear distance and velocity measurements with global positioning satellite (GPS) and the timing gate techniques. J Sports Sci Med S10: 7-8, 2007.
29. Reilly, T and Doran, D. Science and Gaelic football: A review. J Sports Sci 19: 181-193, 2001.
30. Roberts, SP, Trewartha, G, Higgitt, RJ, El-Abd, J, and Stokes, KA. The physical demands of elite English rugby union. J Sports Sci 26: 825-833, 2008.
31. Shephard, RJ. The energy needs of a soccer player. Clin J Sport Med 2: 62-70, 1992.
32. Tomlin, DL and Wenger, HA. The relationship between aerobic fitness and recovery from high intensity intermittent exercise. Sports Med 31: 1-11, 2001.
33. Tomlin, DL and Wenger, HA. The relationships between aerobic fitness, power maintenance and oxygen consumption during intense intermittent exercise. J Sci Med Sport 5: 194-203, 2002.
34. Wisbey, B and Montgomery, P. Quantifying AFL Player Game Demands Using GPS Tracking. FitSense Australia, 2005. Available at: http://www.fitsense.com.au/downloads/GPS%20Research%20Report%20-%202005.pdf. Accessed April 20, 2008.
35. Witte, TH and Wilson, AM. Accuracy of non-differential GPS for the determination of speed over ground. J Biomech 37: 1891-1898, 2004.
Keywords:

work-rates; game analysis; GPS technology; team sport

© 2009 National Strength and Conditioning Association