Article Review #1

O

T

R a

b

c

a

A R R A A

K G T T M

1

i r t f n m H a s

v t C

c

1 h

Journal of Science and Medicine in Sport 17 (2014) 562–566

Contents lists available at ScienceDirect

Journal of Science and Medicine in Sport

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j s a m s

riginal research

he acceleration dependent validity and reliability of 10 Hz GPS

ichard Akenhead a,b,∗, Duncan French a, Kevin G. Thompson c, Philip R. Hayes a

School of Life Sciences, Northumbria University, Newcastle upon Tyne, United Kingdom Newcastle United Football Club, Newcastle upon Tyne, United Kingdom Faculty of Health, University of Canberra, ACT, Australia

r t i c l e i n f o

rticle history: eceived 7 April 2013 eceived in revised form 2 August 2013 ccepted 23 August 2013 vailable online 29 August 2013

eywords: lobal positioning ime motion eam sports onitoring

a b s t r a c t

Objectives: To examine the validity and inter-unit reliability of 10 Hz GPS for measuring instantaneous velocity during maximal accelerations. Design: Experimental. Methods: Two 10 Hz GPS devices secured to a sliding platform mounted on a custom built monorail were towed whilst sprinting maximally over 10 m. Displacement of GPS devices was measured using a laser sampling at 2000 Hz, from which velocity and mean acceleration were derived. Velocity data was pooled into acceleration thresholds according to mean acceleration. Agreement between laser and GPS measures of instantaneous velocity within each acceleration threshold was examined using least squares linear regression and Bland–Altman limits of agreement (LOA). Inter-unit reliability was expressed as typical error (TE) and a Pearson correlation coefficient. Results: Mean bias ± 95% LOA during accelerations of 0–0.99 m s−2 was 0.12 ± 0.27 m s−1, decreasing to −0.40 ± 0.67 m s−1 during accelerations >4 m s−2. Standard error of the estimate ±95% CI (SEE) increased

−1 −2 −1

from 0.12 ± 0.02 m s during accelerations of 0–0.99 m s to 0.32 ± 0.06 m s during accelerations >4 m s−2. TE increased from 0.05 ± 0.01 to 0.12 ± 0.01 m s−1 during accelerations of 0–0.99 m s−2 and >4 m s−2 respectively. Conclusion: The validity and reliability of 10 Hz GPS for the measurement of instantaneous velocity has been shown to be inversely related to acceleration. Those using 10 Hz GPS should be aware that during accelerations of over 4 m s−2, accuracy is compromised.

© 2013 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.

. Introduction

The utilisation of global positioning system (GPS) technology s now widespread within team sports, allowing practitioners and esearchers to measure and monitor the movement patterns of raining and competition.1–3 Accelerations and decelerations are requent within team sports,3–5 and can elicit a high metabolic and euromuscular demand,6,7 making the accuracy of their measure- ent of practical importance to practitioners and researchers.3,5

owever it is unclear whether commercially available 10 Hz GPS re valid and reliable for measuring the spectrum of accelerations een within maximal sprinting.1,8

To date fourteen studies examining the validity or reliability of

arious versions of commercially available GPS for measuring dis- ance and speed have been published (for reviews see Aughey1 and ummins et al.9). Common findings of validation studies include

∗ Corresponding author at: School of Life Sciences, Northumbria University, New- astle upon Tyne NE1 8ST, United Kingdom.

E-mail address: [email protected] (R. Akenhead).

440-2440/$ – see front matter © 2013 Sports Medicine Australia. Published by Elsevier L ttp://dx.doi.org/10.1016/j.jsams.2013.08.005

measurement accuracy being related to the speed and distance of the measurement, with accuracy compromised when measuring short distances and/or high speeds. To date only the work of Varley et al.,8 has examined the validity of GPS during maximal acceler- ation, finding the accuracy of instantaneous velocity decreases as acceleration increases. However, due to differences in GPS man- ufacturer, latitude, the criterion task, utilised methodology, GPS sampling rate, and statistical analysis, a global statement on the validity of GPS for measuring team sports movements is not possi- ble.

Some researchers have utilised precision instruments such as VICON tracking systems,10 Theodolite11 and radar guns8,12 as criterion measures. However we believe that the methodology employed in previous validation studies may limit the extent to which findings can be accepted. For example, all previous stud- ies have had human participants travel at various speeds over a pre-determined distance whilst wearing one or more GPS devices.

This approach may be problematic as movement patterns such as medio-lateral displacement during gait13,14 and leaning during changes of direction15 create discrepancies between the trajectory of the GPS receiver (worn between the scapulae of the participant)

td. All rights reserved.

and M

a i t

t e t b s r a G f t b

i 2

c s h

2

a ( A p p A a

a & o T s p

s fi p c l

F s

R. Akenhead et al. / Journal of Science

nd the intended trajectory.10 Whilst the ecological validity is mportant, the criterion validity must also be understood in order o appreciate the contextual efficacy of GPS technology.

A further consideration is the method used to demarcate the ask within the GPS manufacturers’ software.1 The most commonly mployed method is to utilise timing gates to measure the dura- ion of the trial,16–18 which is then applied to the GPS velocity trace eginning when reported velocity rises above zero. However, mea- urement artefacts caused by leading/trailing limbs can affect the eliability of the time measurement. We believe a more accurate pproach is to utilise the tri-axial accelerometer present in many PS devices to demarcate the start and end of the trial. We have

ound (unpublished data) that visual inspection of the anteropos- erior force curve provides an objective and reliable marker of the eginning and end of the trial (Supplementary Figure 1)

Supplementary material related to this article can be found, n the online version, at http://dx.doi.org/10.1016/j.jsams. 013.08.005.

The aim of this study was to examine the acceleration dependent riterion validity and inter-unit reliability of 10 Hz GPS for mea- uring instantaneous velocity during maximal acceleration using a ighly standardised protocol to eliminate extraneous movement.

. Methods

A custom aluminium monorail system was built to provide linear track with a 10 m working length. Two GPS receivers MinimaxX S4, firmware 6.75, Catapult Innovations, Melbourne, ustralia) spaced 30 cm apart were securely mounted in an upright osition to a platform with the antenna free from obstruction. The latform was able to move freely along the rail on bearings (Fig. 1). ll testing took place on an outdoor natural grass football pitch in n open space free from tall buildings.

GPS receiver velocity was recorded using a laser which sampled t 2000 Hz (optoNCDT ILR 1191, Micro-Epsilon Messtechnik GmbH

Co, Germany). The laser provided displacement to a resolution f 0.001 m with manufacturer reported typical error of ±0.002 m. he laser was positioned 5 m behind the start of the track with the ighting beam focussed on the centre of a 0.1 m × 0.2 m reflective anel secured to the platform.

Maximum displacement was limited to precisely 10 m using teel brackets at either end of the monorail. The distal bracket was

tted with steel compression springs to dampen the impact of the latform. The distance of 10 m from the starting position to first ontact with the springs was verified first by tape measure and ater by laser measurement.

ig. 1. Photograph of the 12 m monorail (A) with sliding platform (B) which was towed ecured to a bracket which was bolted to the platform.

edicine in Sport 17 (2014) 562–566 563

To produce ecologically valid acceleration profiles, displace- ment of the GPS receivers was achieved via towing the sliding platform during maximal 10 m sprints. One male professional foot- ball (soccer) player (21 y, 85 kg, 1.79 m) volunteered to take part and provided written informed consent. Experimental procedures received approval from the Institutional Ethics Committee. The participant completed 15 trials of 10 m maximal sprints from a sta- tionary 2-point athletic stance, each separated by a recovery period of 1 min, towing the platform via a lightweight 10 m non-stretch cord combined with a 2 m stretchable cord fastened around the waist with a customised harness. The participant wore studded football boots and was instructed to sprint maximally for 10 m, and to come to a complete stop within a marked 2 m deceleration zone. Although the compliance of the stretchable cord was minimal, it was included to prevent discomfort resulting from recoil caused by the platform striking the distal compression springs.

Data from each trial was identified in the respective manufac- turer supplied software and exported to Microsoft Excel. Laser data consisted of time and displacement measured at 2000 Hz, which was then re-sampled via rolling average and exported at 100 Hz using the manufacturer’s software. For laser data, the beginning of the trial was identified when platform displacement increased above 5.00 m (laser was positioned precisely 5 m behind the plat- form). From this point, data was then re-sampled to 10 Hz by selecting every 10th data point to provide data that could be com- pared directly with 10 Hz GPS data. The end of the trial was taken as the moment displacement reached 15.00 m.

GPS data consisting of time, “raw” and “smooth” velocity was exported at 10 Hz. Raw velocity (GPSRAW) refers to the velocity cal- culated only from time and GPS positional data. Smooth velocity (GPSSMOOTH) refers to the velocity calculated from time, GPS posi- tional data, plus a unique manufacturer specific algorithm designed to improve accuracy (algorithm unknown). Within the GPS soft- ware (Sprint 5.0, Catapult Innovations, Melbourne, Australia) the onset of a trial was identified as the instant anteroposterior force measured by the integrated 100 Hz tri-axial accelerometer increased above zero (as shown by the “forward” trace within the software). The end of the trial was taken as the instant anteroposte- rior force was seen to sharply increase as the result of impact with the distal compression springs (Supplementary Figure 1)

10 Hz velocity data from GPS and laser were synchro- nised for each trial. Additionally, synchronised velocity data

from both GPS devices was pooled according to the calculated mean acceleration taken from criterion 10 Hz data. For exam- ple, all data points that were collected when mean acceleration was between 0 and 0.99 m s−2 were grouped for analysis of

by the athlete during the 10 m sprints. Two GPS devices spaced 30 cm apart were

564 R. Akenhead et al. / Journal of Science and Medicine in Sport 17 (2014) 562–566

Table 1 Validity of 10 Hz GPS receivers for measuring instantaneous velocity during a 10 m maximal sprint.

Acceleration (m s−2 ) Intercept ± 95% CI Slope ± 95% CI SEE ± 95% CI (m s−1 ) Bias ± 95% LOA Samples (n)

GPSSMOOTH GPSRAW GPSSMOOTH GPSRAW GPSSMOOTH GPSRAW GPSSMOOTH GPSRAW

0–1 0.11 ± 0.09 0.14 ± 0.11 0.99 ± 0.02 1.00 ± 0.02 0.12 ± 0.02 0.19 ± 0.02 0.12 ± 0.27 0.13 ± 0.37 185 1–2 0.01 ± 0.09 0.22 ± 0.9 0.98 ± 0.02 1.01 ± 0.02 0.16 ± 0.02 0.17 ± 0.02 0.08 ± 0.31 0.20 ± 0.32 141 2–3 −0.10 ± 0.10 −0.03 ± 0.17 0.97 ± 0.03 0.93 ± 0.04 0.18 ± 0.03 0.30 ± 0.04 0.02 ± 0.35 0.32 ± 0.63 69 3–4 −0.77 ± 0.10 −0.20 ± 0.19 0.81 ± 0.04 0.88 ± 0.05 0.19 ± 0.02 0.29 ± 0.04 −0.10 ± 0.28 0.22 ± 0.70 32 >4 −0.64 ± 0.28 −0.59 ± 0.23 0.86 ± 0.16 0.70 ± 0.08 0.32 ± 0.06 0.36 ± 0.06 −0.40 ± 0.62 0.28 ± 0.94 30 All data 0.27 ± 0.05 −0.09 ± 0.07 0.93 ± 0.01 0.97 ± 0.02 0.19 ± 0.01 0.29 ± 0.01 0.03 ± 0.43 0.20 ± 0.57 457

D the lea a .

a z o 2 D t o b d i

a s m i l p e m m c E

3

6 M i M o r

9 0 r s

T I

D o

ata presented are the intercept, slope and standard error of the estimate (SEE) of nd criterion data. Mean bias and 95% limits of agreement (LOA) are also presented

cceleration-dependent validity. For the current study, acceleration ones were operationally defined, and the following thresh- lds were examined: 0.0–0.99 m s−2 (0–1), 1.0–1.99 m s−2 (1–2), .0–2.99 m s−2 (2–3), 3.0–3.99 m s−2 (3–4), and 4.0–7.0 m s−2 (>4). ata from 4.0–7.0 m s−2 were combined within one group due to

he low number of data points present for accelerations in excess f 4 m s−2. Thresholds were chosen based upon previous research y Osgnach et al.,7 and unpublished data collected from our group emonstrating the frequency and magnitude of accelerations dur-

ng football training and match play. Least squares linear regression was conducted to establish the

greement between criterion and GPS measures. Intercept and lope of the regression line ±95% confidence intervals were deter- ined. Standard error of the estimate (SEE) ±95% confidence

ntervals and bias ±95% limits of agreement (LOA) were also calcu- ated to provide measures of agreement between the criterion and ractical measures. Inter-unit reliability was expressed as typical rror (TE) and as a coefficient of variation (CV). Pearson product oment correlation coefficients were also calculated to provide a easure of inter-unit agreement. All statistical procedures were

onducted in PASW Statistics v18.0 for Windows and Microsoft xcel.19,20

. Results

Mean ± SD laser derived peak velocity achieved was .6 ± 0.2 m s−1 with a mean trial duration of 2.19 ± 0.03 s. ean ± SD acceleration at t = 0.2 s was 5.93 ± 0.1 m s−2. Test-

ng was conducted under a clear blue sky with no cloud cover. ean ± SD number of connected satellites and horizontal dilution

f position (HDOP) for both GPS devices was 13 ± 1 and 0.9 ± 0.1 espectively.

Mean ± 95% CI distance reported by the GPS devices was

.99 ± 0.08 m. SEE ± 95% CI for distance measurement was .0 ± 0.0 m. Mean bias ± 95% LOA was −0.01 ± 0.43 m. Inter-unit eliability as expressed by TE ± 95% CI was 0.03 ± 0.03 m. The Pear- on correlation coefficient was 0.98 ± 0.02.

able 2 nter-unit reliability of 10 Hz GPS receivers for measuring instantaneous velocity during a

Acceleration (m s−2 ) TE (m s−1 ) CV (%)

GPSSMOOTH GPSRAW GPSSMOOTH

0–1 0.05 ± 0.01 0.11 ± 0.02 0.7 ± 0.1 1–2 0.06 ± 0.01 0.18 ± 0.04 1.1 ± 0.1 2–3 0.09 ± 0.01 0.16 ± 0.03 2.2 ± 0.2 3–4 0.10 ± 0.01 0.44 ± 0.15 3.9 ± 0.4 >4 0.12 ± 0.01 0.43 ± 0.15 9.1 ± 1.0 All Data 0.08 ± 0.00 0.21 ± 0.04 3.1 ± 0.2 ata presented are comparisons of inter-unit reliability from the two GPS devices used in f variation (CV) and a Pearson product moment correlation statistic.

st squares linear regression calculated for the agreement between 10 Hz GPS data

An acceleration dependent shift in measurement bias was observed within the GPSSMOOTH measurements. Velocity tended to be overestimated during low accelerations (0–1 and 1–2), and underestimated during greater accelerations as shown by the mean bias, regression slope and y-intercept (Table 1 and Supplementary Figure 2). The regression slope ± 95% CI of GPSSMOOTH included the value of 1.0 for all acceleration thresholds except 3–4 and >4 m s−2.

Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsams. 2013.08.005.

Validity was greatest during accelerations of 0–1 m s−2 and decreased in an acceleration-dependent manner (Table 1). Over- all, GPSSMOOTH demonstrated increased accuracy compared to GPSRAW. Mean SEE ± 95% CI was greater in GPSRAW compared to GPSSMOOTH (0.29 ± 0.01 m s−1 vs. 0.19 ± 0.01 m s−1) as was bias ± LOA (0.20 ± 0.57 m s−1 vs. 0.03 ± 0.43 m s−1)

There was a shift in the intercept of the regression line from positive to negative suggesting that the small systematic over- estimation of instantaneous velocity during lower accelerations transitions to an under-estimation during higher acceleration for both GPSSMOOTH and GPSRAW (Table 1). There was a similar shift in measurement bias from positive to negative as calculated by the Bland–Altman analysis (Supplementary Figure 3). However this trend was evident only in GPSSMOOTH, with GPSRAW bias remaining positive for all acceleration categories. 95% limits of agreement increased in a broadly acceleration dependent pattern for both GPS measures.

Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jsams. 2013.08.005.

Absolute and relative inter-unit reliability was found to decrease with increasing acceleration for both GPSSMOOTH and GPSRAW. Typ- ical error was lower in GPSSMOOTH compared to GPSRAW in all

acceleration bands (Table 2). The Pearson product-moment cor- relation coefficient exhibited a trend to decrease as acceleration increased for GPSSMOOTH. An acceleration dependent reduction in the correlation coefficient was not evident for GPSRAW (Table 2).

cceleration.

Pearson correlation Samples (n)

GPSRAW GPSSMOOTH GPSRAW

1.8 ± 0.3 0.96 ± 0.01 0.79 ± 0.1 60 3.5 ± 0.7 0.95 ± 0.01 0.76 ± 0.15 56 3.7 ± 0.7 0.91 ± 0.02 0.87 ± 0.09 52

31.2 ± 12 0.91 ± 0.02 0.70 ± 0.50 33 47.4 ± 8. 0.87 ± 0.04 0.75 ± 0.20 32 15.6 ± 2.5 0.99 ± 0.00 0.98 ± 0.00 233

the current study. Reliability is expressed as absolute typical error (TE), a coefficient

and M

4

c G fi G a s

i l o t G s i e a

d a s o G d 2

a g G t i r h fi i

a b p 6 s b

f i a t

E

w i a i r E w m p t i a b p

R. Akenhead et al. / Journal of Science

. Discussion

This study is the first to examine the acceleration-dependent riterion validity and inter-unit reliability of Catapult S4 10 Hz PS receivers for measuring instantaneous velocity. The primary nding was that validity and inter-unit reliability of the 10 Hz PS devices tested appears acceleration-dependent, with greater cceleration reducing the validity and reliability of velocity mea- urement.

The difference in SEE ± 95% CI of measurements taken dur- ng accelerations ranging from 0 to 1 m s−2 and 3 to 4 m s−2 was ow (0.12 ± 0.02 m s−1 vs. 0.19 ± 0.04 m s−1). However, the non- verlapping confidence intervals suggest that the differences are rue (Table 1). Similarly, differences in SEE between GPSSMOOTH and PSRAW also demonstrate true differences. SEE for GPSSMOOTH mea- urement in the >4 m s−2 category was 0.32 ± 0.06 m s−1, a marked ncrease from the broadly similar SEE values of the preceding accel- ration categories which strongly suggests compromised validity bove >4 m s−2.

The slope of the regression line ± 95% CI for GPSSMOOTH did not iffer from the line of identity for categories up to and including ccelerations of 3 m s−2 (Table 1). For accelerations over 3 m s−2 the lope ± 95% CI was less than 1.0, indicating a true underestimation f instantaneous velocity during higher accelerations. The slope of PSRAW regressions demonstrated a similar trend, however the true eviation from the line of identify occurred for accelerations over

m s−2. Inspection of the calculated inter-unit reliability expressed as

coefficient of variation (TE as a percentage of the mean) sug- ests that GPSSMOOTH is more reliable than GPSRAW (Table 2). PSSMOOTH exhibits sufficient reliability (CV < 5%) for accelera-

ions up to and including 4 m s−2. For GPSRAW the 5% threshold s breached for accelerations >3 m s−2. As with validity, inter-unit eliability appears acceleration dependent and is compromised at igher accelerations. Practitioners should be aware of the current ndings, and in line with previous suggestions avoid using devices

nterchangeably.10,18

From examination of laser-derived acceleration, mean ± SD cceleration at t = 0.2 s was 5.93 ± 0.1 m s−2 which is similar ut understandably lower than reported mean ± SD values for eak acceleration of medium level male sprinters at t = 0.2 s of .42 ± 0.61 m s−2.6 The similarity of these values provides strong upport for the ecological validity of the protocol used for assessing rief maximal sprints.

Given the calculated validity of the GPS receivers examined or measuring 10 Hz instantaneous velocity, it is possible to make nferences regarding the efficacy of these devices for measuring cceleration. Error in mean acceleration (Errora) measurement by he devices used in the current study can be defined by Eq. (1).

rrora = (ErrorV1 − ErrorV0) × 1/�t (1)

here ErrorV0 and ErrorV1 is the error for measuring initial and final nstantaneous velocity respectively, and �t = the time between v0 nd v1 (dwell time). However, as demonstrated by Eq. (1), error n the measurement of instantaneous velocity does not necessarily esult in an error in reported mean acceleration. For example, if both rrorV0 and ErrorV1 are equal, the calculated mean acceleration ill be correct. In practical terms, mean acceleration measure- ent accuracy increases as �t increases, but sensitivity to brief

eaks in acceleration will be compromised. To minimise error, the ime period over which the acceleration is calculated (�t) can be

ncreased. Due to the apparent reduction in validity during higher ccelerations (>4 m s−2) reported in the current study, it may not e advisable to use the studied 10 Hz GPS units to measure acute eak accelerations of such magnitude.

edicine in Sport 17 (2014) 562–566 565

Direct comparison between the findings of the current study and previous studies is difficult due to differences in protocol and the categorisation of data into acceleration thresholds. In general agreement with the current study, Varley et al.,8 found validity of the same 10 Hz devices to be greater during lower changes in veloc- ity, although data was not categorised in acceleration thresholds but rather by starting velocity of the acceleration effort. Addition- ally, in the study of Varley et al.,21, acceleration did not commence from a stationary start but from a “rolling” start of at least 1 m s−1. Thus, peak acceleration, although not reported, was likely lower than in the current study. The current study builds on the study of Varley et al.,21 to provide a better understanding of the efficacy of 10 Hz GPS for measuring instantaneous velocity and therefore acceleration.

The experimental protocol utilised in the current study allows for a more confident interpretation of the findings through the com- bination of an appropriate criterion measure, precise methods of demarcation for synchronisation of data points, and the elimination of measurement error due to the standardisation of GPS receiver trajectory. The current study employed linear 10 m maximal accel- erations from a stationary start as they represent sport-specific situations in which the greatest accelerations occur. Validity and reliability during decelerations were not examined as it was not possible with the current protocol. However, as the variable exam- ined was instantaneous velocity, it may be reasonable to assume that the current findings also extend to decelerations. With the continual development and refinement of GPS technology, future validation studies will be warranted. Researchers should seek to standardise experimental protocols where possible to allow for greater confidence in the interpretation of results.

5. Conclusion

The validity and inter-unit reliability of 10 Hz GPS for the mea- surement of instantaneous velocity has been shown to be inversely related to acceleration. Those using the Catapult S4 10 Hz model of GPS should be aware that during accelerations of over 4 m s−2, accuracy is compromised. In the examined GPS unit, GPSSMOOTH offers enhanced validity and inter-unit reliability in comparison to GPSRAW. Researchers using GPS should select and justify both the acceleration thresholds and the time period over which mean acceleration is calculated.

Practical implications

• Practitioners should be aware that the GPS technology used in the current study may be unsuitable for the measurement of instan- taneous velocity during high magnitude (>4 m s−2) accelerations.

• Practitioners and researchers investigating team sports should be aware of the impact of dwell time duration (�t) on the accuracy and sensitivity of measuring acceleration derived from instanta- neous velocity.

• To eliminate inter-unit variation, GPS devices should not be used interchangeably.

Conflicts of interest

None of the authors have any conflicts of interest to declare.

Disclosure statement of funding

No external funding was received for this project.

5 and M

A

b

a t C U e

s (

R

1

1

1

1

1

1

1

1

1

1

66 R. Akenhead et al. / Journal of Science

cknowledgements

Funding for this study was provided by Newcastle United Foot- all Club and the School of Life Sciences, Northumbria University.

We would like to thank Dr. Glyn Howatson for his continued ssistance and his role in securing funding for equipment. Special hanks are also due to Dr. Roger Penlington and the department of omputing, Engineering and Information Sciences at Northumbria niversity for their help with the design and construction of the quipment used.

The laser instrument used as the criterion measure in the current tudy was supplied and operated free of charge by Micro Epsilon Micro Epsilon, UK).

eferences

1. Aughey RJ. Applications of GPS technologies to field sports. Int J Sports Physiol Perform 2011; 6(3):295–310.

2. Brewer C, Dawson B, Heasman J et al. Movement pattern comparisons in elite (AFL) and sub-elite (WAFL) Australian football games using GPS. J Sci Med Sport 2010:618–623.

3. Young WB, Hepner J, Robbins DW. Movement demands in Australian rules football as indicators of muscle damage. J Strength Cond Res 2012; 26(2): 492–496.

4. Varley MC, Aughey RJ. Acceleration profiles in elite Australian soccer. Int J Sports Med 2013; 34(1):34–39.

5. Akenhead R, French D, Hayes PR et al. Diminutions of acceleration and decel- eration output during professional football match play. J Sci Med Sport 2013 [in press].

6. di Prampero PE, Fusi S, Sepulcri L et al. Sprint running: a new energetic approach. J Exp Biol 2005; 208(Pt 14):2809–2816.

7. Osgnach C, Poser S, Bernardini R et al. Energy cost and metabolic power in elite soccer: a new match analysis approach. Med Sci Sports Exerc 2010; 42(1):170–178.

2

2

edicine in Sport 17 (2014) 562–566

8. Varley CM, Fairweather IH, Aughey RJ. Validity and reliability of GPS for mea- suring instantaneous velocity during acceleration, decceleration, and constant motion. J Sports Sci 2011; 30(2):121–127.

9. Cummins C, Orr R, O’Connor H et al. Global positioning systems (GPS) and microtechnology sensors in team sports: a systematic review. Sports Med 2013:1–18.

0. Duffield R, Reid M, Baker J et al. Accuracy and reliability of GPS devices for mea- surement of movement patterns in confined spaces for court-based sports. J Sci Med Sport 2009; 13(5):523–525.

1. Gray AJ, Jenkins D, Andrews MH et al. Validity and reliability of GPS for measuring distance travelled in field-based team sports. J Sports Sci 2010; 28(12):1319–1325.

2. Johnston RJ, Watsford ML, Pine MJ et al. The validity and reliability of 5-Hz global positioning system units to measure team sport movement demands. J Strength Cond Res 2012; 26(3):758–765.

3. Schache AG, Blanch P, Rath D et al. Three-dimensional angular kinematics of the lumbar spine and pelvis during running. Human Mov Sci 2002; 21(2):273–293.

4. Orendurff MS, Segal AD, Klute GK et al. The effect of walking speed on center of mass displacement. J Rehabil Res Dev 2004; 41(6):829–834.

5. MacLeod H, Morris J, Nevill A et al. The validity of a non-differential global posi- tioning system for assessing player movement patterns in field hockey. J Sports Sci 2009; 27(2):121–128.

6. Coutts AJ, Duffield R. Validity and reliability of GPS devices for measuring move- ment demands of team sports. J Sci Med Sport 2010; 13(1):133–135.

7. Jennings D, Cormack S, Coutts AJ et al. The validity and reliability of GPS units for measuring distance in team sport specific running patterns. Int J Sports Physiol Perform 2010; 5(3):328–341.

8. Petersen C, Pyne D, Portus M et al. Validity and reliability of GPS units to monitor cricket-specific movement patterns. Int J Sports Physiol Perform 2009; 4(3):381–393.

9. Hopkins WG. Analysis of validity by linear regression (Excel spreadsheet), A new view of statistics. sportsci.org: Internet Society for Sport Science, 2000, cited available from: http://www.sportsci.org/resource/stats/xvalid.xls [accessed 18.12.12].

0. Hopkins WG. Measures of reliability in sports medicine and science. Sports Med 2000; 30(1):1–15.

1. Varley MC, Fairweather IH, Aughey RJ. Validity and reliability of GPS for mea- suring instantaneous velocity during acceleration, decceleration, and constant motion. J Sports Sci 2011; 30(2):121–127.