Tracking Systems in Team Sports

Tracking Systems in Team Sports: Back to Basics

Lorena Torres-Ronda 1 ,Jo Clubb 2 3 , Emma Beanland 4
1
Spanish Basketball Federation, Spain
2
Global Performance Insights, UK; University of Technology Sydney, Australia
3
University of Technology Sydney, Australia
4
Sports Performance, Buffalo Bills, Buffalo, U.S.

Technology | Monitoring systems | External load | Data | Sport science

Headline GPS device constantly receives the exact time, based on the
he desire to deepen the understanding of sports perfor- atomic clock, via radio signals transmitted by these satellites
T mance has led to the proliferation of technology. Within
this, tracking technology enables the objective capture of work
at the speed of light (6). The difference in time between the
satellite and GPS receiver is used to calculate the signal travel
completed by athletes independent of their internal charac- time and therefore, distance from the satellite to each receiver
teristics, which is labelled external load (1). Such data en- (6). A minimum of four satellites is required (nine is preferred)
ables practical applications that include quantifying physical to calculate the exact position via trigonometry, expressed as
demands, objectifying training prescription, monitoring indi- latitude and longitude (7).
vidual training load, and assisting with rehabilitation. Once the position is known, the displacement over a given
epoch can be used to calculate instantaneous speed (velocity
of a movement) (5). Velocity is calculated via the Doppler-
Aim shift method, which refers to a change in frequency of the
Several types of tracking systems have emerged since the satellite signals caused by the movement of the receiver (8).
1970s, while include notational analysis (2), multi-camera Doppler-shift is frequently preferred to positional differentia-
systems (3), and microtechnology devices (4). Professionals tion (the derivative of distance caused by the change in device
tasked with managing these systems have a responsibility to location) owing to greater accuracy and lower error (9). Us-
understand data collection, processing, and analysis as it per- ing the Doppler-shift method, distance and acceleration can
tains to athlete preparation within their specific setting. be derived from velocity (10).
In order to select and utilise impactful technologies, under- There are three key factors that influence GPS quality: i)
standing the methodologies, requirements, and differences be- the line of sight from GPS receiver to satellite, ii) the signal
tween systems should be thorough. Knowledge of hardware strength and iii) the horizontal dilution of precision (HDoP)
(e.g., satellites/anchors, receivers/devices, network require- (11). The line of sight is affected by structural interferences
ments) and software (e.g., system operation, firmware version, (e.g., stadium obstructions) as GPS signal will not travel
metric definition) will better prepare users to understand the through metal or concrete (7). The signal strength is de-
subsequent advantages and disadvantages of the data. During termined by both the number of satellites connected to the
the technology selection process, practitioners should consider receiver and their orientation within the atmosphere, with a
the context, regulations, and culture of their specific sport signal greater than 40 decibels ideal (12). Manufacturers have
and organisation. Thereafter, practitioners must be knowl- attempted to improve data quality through the integration
edgeable in data collection, processing, analysis, visualisation, of both the United States GPS and the Russian GLONASS
and communication, in order to transform data into meaning- (GLObal NAvigation Satellite System), which together dou-
ful and actionable information for athlete management. bles the number of available satellites. Finally, the HDoP
In this paper, we provide an overview of the emergence and refers to the geometrical arrangement of satellites, whereby
development of different tracking systems. Considerations for one satellite is positioned directly overhead and the others are
a systematic approach to data processing and collection are evenly spaced, with less than 1.0 representing ideal positioning
also presented. (11).
GPS validity and reliability can be compromised by move-
ments that require shorter durations, higher velocities and/or
Types of Tracking Systems more complexity (5). Such movements may be the most crit-
ical to a sporting context, as well as being the most physi-
In the following section, the methodologies of the most com-
cally demanding. Therefore, practitioners are presented with
monly utilised player tracking technologies within team sport
a paradox in that the most vital movements may be the least
settings are discussed. By means of comparison, Table 1 pro-
accurately quantified. In an attempt to address this, hardware
vides an overview of these characteristics. Here, we define
improvements, including higher sampling frequencies (13), up-
tracking systems as those that quantify external load within
dated GPS chipset and antenna providers, and more power-
the training environment. Therefore, monitoring systems that
ful microprocessors have enhanced the quality of GPS data
quantify internal load (e.g., heart rate), track segmental move-
(7). Additionally, manufacturers frequently release enhanced
ments (e.g., force plates or motion capture) or capture training
firmware that control hardware operation, but practitioners
load response (e.g., subjective wellness) are not included.
should be aware such changes may affect the data output (10).

Global Positioning Systems

Global Positioning Systems (GPS) provide measures of dis-
placement, velocity, and acceleration by using a satellite net-
work, which was initially used by the US Department of De-
fense, and have since been validated for use in sport (5). A

sportperfsci.com 1 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

Optical Tracking Systems the need of manual operators (16). By employing a camera
Optical tracking (OT) uses multiple cameras placed around system, OT does not require the athletes to wear any technol-
the playing environment to collect computer video analysis ogy, which is especially pertinent in sports that prevent the
that determines the 2-dimensional coordinates (2D; x, y) and use of these in competition.
trajectories of all moving objects (3). The camera setup, in- Despite the appeal of a non-invasive solution, OT presents
cluding position, orientation, zoom, and field of vision, are some limitations (Table 1) (17). Given the system perma-
fixed upon installation to ensure an all-encompassing vision nence, teams may be limited to competition and/or home
of the playing area, with at least two cameras covering ev- venue data. Users may not receive tracking data until 36 hours
ery area (14). Sampling frequencies vary between systems after competition, due to post-event processing, although some
with between 10 and 25 Hz reported (3, 15). Early OT was systems now offer real-time analysis (17). Despite extensive
semi-automated as it required operators to code activities (3). research on early OT accuracy (3, 14), more recent technology
However, more recent solutions utilise machine learning and has been introduced into sporting competition despite a lack
computer vision techniques to calculate location data without of peer-reviewed precision research (16).

Table 1. Comparison of Characteristics Across Different Tracking Systems.

GPS OT LPS IMU

The travel time of Anchors placed

Camera system Multiple microsensors
radio signals around the
around the playing detect movement on
METHODOLOGY transmitted by the environment
environment tracks the body upon which
satellites to GPS triangulates radio
coordinates they are placed
tags/devices. signals with tags

Displacement, Reception time of

Accelerometer =
distance via known Trajectories of moving signal used to
acceleration
time between Satellite objects from which determine location
Gyroscope = angular
MEASUREMENT and GPS tag. Velocity distance, speed, and from which distance,
velocity
via Doppler-shift. acceleration are speed, and
Magnetometer =
Acceleration derived derived acceleration are
orientation
from velocity derived

Raw accelerometry
SAMPLING 1-15 Hz (Greater than
10-25 Hz data at 500-1000 Hz,
FREQUENCIES 10 Hz recommended)
reported as 100 Hz

DIMENSIONS 2D (x,y) 2D (x,y) 2D (x,y) 3D (x,y,z)* 3D Triaxial (x,y,z)

PLAYING OBJECT
In progress Yes Yes #N/A
TRACKING

SETTING Outdoor Indoor, Outdoor Indoor, Outdoor Indoor, Outdoor

PORTABILITY Yes Lacking Lacking Yes

High costs,
Accuracy in demanding
Susceptibility to Alone will not provide
movements that infrastructure
potential electronic detail of global
KEY require shorter requirements, lack of
interference and/or movement tasks
LIMITATIONS durations, higher portability, post-event
signal instability, lack regarding overall
velocities and/or more processing, lack of
of portability distance/speed.
complexity peer-reviewed validity
and reliability studies

Radio Frequency Identification

Local Positioning Systems
Radio frequency identification (RFID) technology has been
Local positioning systems (LPS) provide location information
commercially available since the 1970s however, it was not
on an object or a person within a space covered by a local area
until the early 1990s that it entered sport, initially for timing
network (15, 18). In sport settings, LPS triangulate signals
mass running events. RFID refers to a wireless system com-
between anchors placed in an elevated position and electronic
posed of antennae (or anchor nodes) that emit radio waves,
tags worn by the athlete. The system provides information
and tags that return signals to communicate their identity
about athlete position at a given time by their 2D or, in some
cases, 3-dimensional (3D) coordinates.

sportperfsci.com 2 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

(17). The reception time between the antennae and tag is segment to which the sensor is placed, expressed as G-force
synchronised and used to determine location, with the data (where 1 G = 9.81 m.s−2 ) (23).
immediately processed by a central computer (17) . Types In order to provide timely data processing, accelerometers
of RFID systems differ by radio frequency signal; low- (30-500 usually report sampling at 100 Hz (24), however often the raw
KHz), high- (3-30 MHz), ultra-high-, and microwave frequency data sampling is much higher (e.g., 500 Hz – 1000 Hz). The
(including ultra-wideband; see below). signal obtained from the sensor is usually filtered to reduce
The advantages of RFID include commensurate perfor- noise. Physical activity can be obtained by classifying the vol-
mance across indoor and outdoor venues (19) and advances ume and intensity of the signal in a specific time interval (25).
in accuracy and processing speed over other tracking devices In addition, manufacturer-specific, vector-magnitude metrics
(20). However, as with other LPS, RFID is most reliable when derived from the triaxial accelerometer, such as PlayerLoadTM
the system is permanent therefore, it lacks portability and re- or BodyLoadTM (26), are often used to represent external load
quires installation that can be costly (21). When a compatible (23).
portable system is available, it can involve a time-consuming The combination of data from accelerometers, gyroscopes,
setup and requires calibration for 3D data. Additionally, if and magnetometers, coupled with increased sampling rates
a playing implement is used and not endowed with a tag, it and machine learning, is increasing the accuracy of fine move-
is not possible to provide tactical information related to the ment detection (27). As such, the ability to capture locomotor
object. activities such as sprinting and changes of direction (28), as
well as sport-specific movements, including rugby collisions
(29), ice hockey strides (30) and specific skill-based activities
Ultra-wideband Local Positioning System (31), have been enhanced. The detection of such demands,
Ultra-wideband (UWB) is a type of RFID that transmits short along with the ability to derive these measures while both in-
pulses via radio signals, specifically, using a bandwidth greater doors and outdoors, highlight a number of advantages of IMU
than 500 MHz or greater than 25% of the centre carrier fre- over other technologies.
quency (21). It provides greater precision than other RFID
systems, owing to better ability of the signal to pass through
obstacles and greater resistance to interference (21). UWB Data Processing
also provides lower power consumption, higher data rate trans- Often not enough attention is placed on the data processing
mission and may cost less than alternatives (22). techniques that calculate the myriad of metrics available via
tracking technology. Filtering, threshold selection, and effort
detection rules are important components for defining high-
Inertial Measurement Units speed and high-intensity efforts, which are of great interest to
Inertial measurement units (IMU) typically consist of an ac- practitioners given their association with neuromuscular de-
celerometer, a 3D gyroscope (to determine angular veloc- mands, performance outcomes and fatigue (32, 33). However,
ity) and/or a magnetometer (to measure the orientation with these definitions are often not included within the literature
the Earth). Most athlete IMU contain a triaxial piezoelec- and there exists a lack of uniformity across and within sports.
tric linear accelerometer that measures 3D (anterior-posterior, Becoming familiar with these settings will enable practitioners
medial-lateral, and vertical) acceleration of the body or body to select settings relevant to the context of their environment.

Fig. 1. Select high-speed running thresholds (A) from a group of studies with similar subjects (sport, sex) and (B) a
visual comparison of distance within corresponding bands applied to a single hypothetical tracking file.
Ab = absolute thresholds; Rel = relative thresholds; LS = low speed band (light grey); HS = high-speed (dark grey);
VHS = very high-speed (black)

sportperfsci.com 3 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

Filtering qualitative descriptors to be assigned, typically based on a

Raw, unfiltered data is subject to a degree of noise that adds form of intensity (e.g., very low, low, moderate, high, and very
error to a signal. Therefore, to accurately analyse velocity high) (39). However, there is a lack of consensus in defining
or acceleration data, filtering is required to ensure the signal these thresholds and consequently large discrepancies exist,
resulting from human motion is detected (34). A variety of even within the same sport (37) (Figure 1).
filters can be applied to velocity data however, manufacturers Traditionally, thresholds have been determined arbitrar-
may not disclose information regarding the technique used. ily by using absolute measures however, growing interest ex-
Acceleration derived from velocity is also smoothed at a time ists in individualised thresholds (40). An overestimation of
interval set within the software (e.g., 0.2-0.5 s), whereby the high-speed distance has been demonstrated with an absolute
wider the time interval, the greater the smoothing of the data. threshold compared to individualised thresholds, determined
Most manufacturers recommend frequent updates to apply by a treadmill test to exhaustion (41). However, Lovell and
bug fixes, feature enhancements and filtering improvements Scott (2018) highlighted the additional challenge practition-
and yet, slight adjustments to these settings can have a no- ers face in obtaining an appropriate and up-to-date maximum
table effect on data output (10, 35). While filtering is often value for each individual to apply to relative thresholds (42).
discussed in relation to GPS, it is important to note that OT While manufacturers may now enable both absolute and rel-
and LPS may also use filtering techniques that merit consid- ative threshold settings, it is important practitioners assess the
eration. demands of their sport and population when selecting appro-
priate bands. Manufacturer defaults may not always be appli-
cable for practitioners within their sporting context. While,
Thresholds and Bands equal bandwidth thresholds, such as 5km/h bands, have been
Velocity and acceleration efforts, particularly those at greater advised (37), new data mining techniques provide a potential
magnitudes, are of interest to practitioners given the increased avenue for defining more appropriate band settings, although
importance on match outcomes and energetic demands of such meaningful differences in outputs between different statistical
movements respectively (36, 37). Threshold selection refers to approaches remain (7, 42).
a series of lower and upper limits that define a band, which
are commonly set across metrics (38). These bands allow for

Fig. 2. Acceptance criteria demonstrated for velocity effort detection where, 1) minimum effort duration, or ‘dwell
time’, is set at 0.5 seconds, and 2) time out window is set to greater than 1 second. Note, these time windows are for
illustration only and may differ accord to manufacturer. Green shows the acceptance criteria is met; Red shows that it
is not. HS = high-speed; VHS = very high-speed

velocity or acceleration may oscillate around a threshold. It is

Effort Detection imperative practitioners seek out these settings to understand
While distance-aggregates within bands require only a lower how their system is defining efforts, especially when integrat-
and upper threshold limit, effort detection requires multiple ing data from different systems.
rules. These may include determining time periods for the
minimum effort duration (MED) and a time out window be-
tween efforts (Figure 2). The MED, or ‘dwell time’, determines Data Collection and the Need for Systematic Process
the minimum duration by which the athletes must exceed a Having “unpacked the black box” (7) to understand data pro-
threshold to count as an effort (10). A MED of 0.5 s has been cessing techniques employed, the onus is now on the practi-
suggested to detect high-speed running efforts (37). Histor- tioner to establish a consistent approach to their workflow.
ically, detection has required a one or two band drop below Their primary responsibility is to ensure their process enables
the current threshold post-MED to register an effort. A time accurate data collection, especially if decisions around player
out window can specify the minimum time required to elapse management are to be based upon such information. Further-
before an effort is counted as a separate effort to the previous more, promoting clean and organised data will enable more
one. These rules are an attempt to ensure potential erroneous time to be spent on analysis, interpretation and communica-
spikes are not incorrectly captured as efforts and to avoid mul- tion. A summary of these considerations is shown in Table
tiple effort detection for what is in reality a single effort, as 2.

sportperfsci.com 4 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

Pre-session Considerations where possible (11, 35). Practitioners should build time into
Despite that manufacturers often deliver systems in an imme- their workflow prior to the session to allow them to validate
diately operational state, it is imperative to review the set- tag switch on, athlete mapping, and anchor/satellite visibility.
tings before use. It is recommended to review features such The specific location of the tag on the athlete can vary de-
as the synchronisation time (especially when receiving new pending on the manufacturer and the algorithms used for cal-
tags), system modes (e.g., indoor/outdoor), and the data pro- culating metrics. For garment setup, it is worth noting that
cessing settings discussed in section 3. Practitioners may wish while custom-built pockets in jerseys have become common-
to document their preferred settings and data collection steps, place, they may increase incidental unit movement, as demon-
in order to ensure a replicable workflow, especially if other strated by increases in accelerometer loads across standardised
colleagues may also be tasked with running the system. running, agility, and tackling drills, compared to data collec-
Considerations relating to tag setup in preparation for track- tion via manufacturer-provided vests (43). This provides an
ing a session are provided in Table 2. Since questionable example of an accuracy-practicality trade-off, whereby prac-
between-device reliability has previously been demonstrated, titioners should give pause to the most practical solution for
it is recommended to assign the same tag to the same athlete their setting without too much loss in precision.

Table 2. Checklist for Promoting Clean Tracking Data.

Session Time Point Considerations for Practitioners

Have I reviewed the manufacturer instructions?

System Setup Have I reviewed, where available, the filtering and smoothing utilised
by this manufacturer?

Have I determined and set suitable thresholds/bands and effort

detection settings for my population?

Have I tested the system operation and data collection process prior
to use with athletes?

Where necessary, have I synchronised the time on all tags?

Is every tag charged, mapped and labelled for the assigned athlete?
Where possible, have I used the same tag for the same athlete?
Pre-session
Is every tag set to turn on at a sufficient time before the session?
Where automated turn-on not possible, have I planned when and how
I will switch the tags on?

Is the garment suitable sized for the assigned athlete and holding the
tag in the correct position and/or orientation?

Have I verified tag switch on prior to or at the start of the session?

Within-session
Have I created the drills with accurate start and stop times?

Have I managed each athlete correctly (e.g., benching, rehab, extra

workload, etc.)?

Have I correctly named and categorised the session and drills

according to a systematic typology? Can I use a pre-populated list in
the software or on the computer?

Have I checked for errors and removed bad data?

Post-session
Have I logged data precision information (e.g., signal strength, HDoP)
and software and firmware versions?

Have I successfully integrated the data with reporting and/or data

management systems?

sportperfsci.com 5 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

In-session Considerations to enable uniformity and speed-up naming. For the afore-
While tracking a session, there are often software features mentioned applications to be accurate, the drill timings and
available that can promote clean data, minimise administra- participants need to be precisely aligned.
tive time after the session, and facilitate insightful and efficient Perhaps the greatest challenge facing the practitioner during
data analysis. Firstly, there is an opportunity to label the ses- a session within the team setting is tracking each individual.
sions with attributes that assist in retrospective analysis by Athletes may need removing from a drill or entire session if
describing the session (see Table 3). they cease to participate, to ensure data is not skewed by a
A drill typology opens potential avenues for analysis, in- missing participant. Alternatively, athletes regularly carry out
cluding workload comparison (44), assessing readiness by nor- extra drills before and/or after the main team session, which
malising drill outputs by player (45), and monitoring athlete should be captured for longitudinal load monitoring. Such
response using standardised drills (46). Table 4 outlines po- work should be labelled appropriately to allow for removal
tential categories and associated examples from a range of from analysis when only structured, coach-led training is of
sports that may be considered. Maintaining consistency in interest. This also applies to athletes undergoing rehabilita-
naming structure, including the case used (e.g., upper, lower, tion, who should have their sessions captured and named in
sentence), is advised. Where possible, it is recommended to a consistent manner that enables inclusion or exclusion from
set up a pre-populated list of drill categories and/or names, analysis as required.

Table 3. Examples of Session Level Attributes.

Category Examples

Game/Match, Training/Practice, Rehabilitation,

Session Type
Fitness Test

Category D+1/D-2, G-3, Strength, Day Before Game

Pitch 1, Field 2, Main Court, Rink 2, Indoor,

Location
Stadium

Surface Grass, Synthetic, Sand, Court

Equipment Boots/Cleats, Trainers/Sneakers, Pads, Helmets

D = Day; G = Game

Table 4. Examples of Drill Typology

Category Examples

Team, Individual, Technical, Tactical, Physical,

Type
Rehabilitation

Warm-up, Possession, Shooting, 3rd Down,

Drill/Period
Scrimmage, Defensive

10v10+GKs, 7v7, 9 on 7, 1 on 1s, 5v5+5+GK,

Numbers
Individual

80x50yd, 30x20m, Half Court, Full Court,

Space
Penalty Area

Descriptors Coach encouragement, Walk-through, Two-touch

GK= Goalkeepers

sportperfsci.com 6 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

Nov;48(11):2549-75.
Post-session Considerations
Where the workflow detailed above is attentively followed, the 5. Aughey RJ. Applications of GPS technologies to field
post-session data management processes should be stream- sports. Int J Sports Physiol Perform. 2011 Sep;6(3):295-310.
lined. Nevertheless, practitioners should investigate the qual-
ity of the data post-session for errors and spikes. Where 6. Larsson P. Global positioning system and sport-specific
the system allows, descriptive information relating to the testing. Sports Med. 2003;33(15):1093-101.
data quality (e.g., GPS signal strength and HDoP) should be
logged. 7. Malone JJ, Lovell R, Varley MC, Coutts AJ. Unpacking
Tracking data may be integrated with an athlete data the Black Box: Applications and Considerations for Using
management system to allow for analysis and communica- GPS Devices in Sport. Int J Sports Physiol Perform. 2017
tion alongside other sources of information. The practitioner Apr;12(Suppl 2):S218-S26.
should consider how missing data (e.g., for a specific drill due
to an error with the system or a completely missing session 8. Schutz Y, Chambaz A. Could a satellite-based navigation
due to a device not working/being worn) will be handled, system (GPS) be used to assess the physical activity of indi-
particularly for the purposes of longitudinal load monitoring. viduals on earth? Eur J Clin Nutr. 1997 May;51(5):338-9.
Interested readers are directed to a review by Broman and
Woo (47) in which data organisation techniques (e.g., variable 9. Townshend AD, Worringham CJ, Stewart IB. Assessment
names, categories, and units, managing date formats, struc- of speed and position during human locomotion using nondif-
turing tidy data, and using a data dictionary) are discussed. ferential GPS. Med Sci Sports Exerc. 2008 Jan;40(1):124-32.

10. Varley MC, Jaspers A, Helsen WF, Malone JJ. Method-

Practical Applications ological Considerations When Quantifying High-Intensity Ef-
• While practitioners may be tempted to dive straight into forts in Team Sport Using Global Positioning System Tech-
tracking data analysis, this article demonstrates the im- nology. Int J Sports Physiol Perform. 2017 Jan 04:1-25.
portance of understanding the characteristics of different
systems and settings. 11. Jennings D, Cormack SJ, Coutts AJ, Aughey RJ. GPS
analysis of an international field hockey tournament. Int J
• Each technology differs in how it tracks athletes and subse- Sports Physiol Perform. 2012 Sep;7(3):224-31.
quently, each offers varying advantages and disadvantages
for tracking external load. In particular, how a technology 12. Witte TH, Wilson AM. Accuracy of non-differential GPS
and/or manufacturer processes the data should be consid- for the determination of speed over ground. J Biomech. 2004
ered within the sporting context. Dec;37(12):1891-8.

• The filtering processes, velocity/acceleration thresholds for 13. Scott MT, Scott TJ, Kelly VG. The Validity and Relia-
band selection, and effort detection rules should be ex- bility of Global Positioning Systems in Team Sport: A Brief
plored, given that the manufacturer-specific defaults may Review. J Strength Cond Res. 2016 May;30(5):1470-90.
not be best suited for the specific sport, team, or athlete(s)
in question. 14. Di Salvo V, Collins A, McNeill B, Cardinale M. Valida-
tion of Prozone ®: A new video-based performance analysis
• The potential applications for tracking data will only be en- system. Int J Perf Anal Spor. 2006 2006/06/01;6(1):108-19.
abled if a systematic process to data collection is employed.
This process can include utilising a pre-session checklist, 15. Linke D, Link D, Lames M. Validation of electronic per-
along with a sport-specific typology for session and drill formance and tracking systems EPTS under field conditions.
classification, to ensure consistent and accurate collection. PloS one. 2018;13(7):e0199519.

16. McLean BD, Strack D, Russell J, Coutts AJ. Quantify-

Twitter: Lorena Torres-Ronda (@lorenatorres07), Jo Clubb ing Physical Demands in the National Basketball Association-
(@JoClubbSportSci), Emma Beanland (@EmmaBeanland1) Challenges Around Developing Best-Practice Models for Ath-
lete Care and Performance. Int J Sports Physiol Perform.
2019 Apr 1;14(4):414-20.
References
1. Halson SL. Monitoring training load to understand fatigue 17. Carling C, Bloomfield J, Nelsen L, Reilly T. The role
in athletes. Sports Med. 2014 Nov;44 Suppl 2:S139-47. of motion analysis in elite soccer: contemporary performance
measurement techniques and work rate data. Sports Med.
2. Reilly T, Thomas V. A motion analysis of work-rate in 2008;38(10):839-62.
different positional roles in professional football match-play. J
Hum Mov Stud. 1976;2:87–97. 18. Woellik H, Mueller A, Herriger J. Permanent RFID Tim-
ing System in a Track and Field Athletic Stadium for Training
3. Castellano J, Alvarez-Pastor D, Bradley PS. Evaluation and Analysing Purposes. Procedia Engineering. 2014;72:202-
of research using computerised tracking systems (Amisco and 7.
Prozone) to analyse physical performance in elite soccer: a
systematic review. Sports Med. 2014 May;44(5):701-12. 19. Sathyan T, Shuttleworth R, Hedley M, Davids K. Validity
and reliability of a radio positioning system for tracking ath-
4. Whitehead S, Till K, Weaving D, Jones B. The Use of letes in indoor and outdoor team sports. Behav Res Methods.
Microtechnology to Quantify the Peak Match Demands of the 2012 Dec;44(4):1108-14.
Football Codes: A Systematic Review. Sports Med. 2018

sportperfsci.com 7 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

20. Carling C, Williams AM, Reilly T. Handbook of Soccer sess peak accelerations during walking, jogging and running.
Match Analysis: A Systematic Approach to Improving Per- Eur J Sport Sci. 2015;15(5):382-90.
formance. J. Sports Sci. Med. 2006;5(1):171.
35. Buchheit M, Allen A, Poon TK, Modonutti M, Greg-
21. Angrisani L, Arpaia P, Gatti D. Analysis of localiza- son W, Di Salvo V. Integrating different tracking systems in
tion technologies for indoor environment. IEEE International football: multiple camera semi-automatic system, local posi-
Workshop on Measurement and Networking (M&N. 2017:1-5. tion measurement and GPS technologies. J Sports Sci. 2014
Dec;32(20):1844-57.
22. Alarifi A, Al-Salman A, Alsaleh M, Alnafessah A, Al-
Hadhrami S, Al-Ammar MA, et al. Ultra Wideband In- 36. Varley MC, Gabbett T, Aughey RJ. Activity profiles
door Positioning Technologies: Analysis and Recent Advances. of professional soccer, rugby league and Australian football
Sensors (Basel, Switzerland). 2016 May 16;16(5). match play. J Sports Sci. 2014 Dec;32(20):1858-66.

23. Boyd LJ, Ball K, Aughey RJ. The reliability of MinimaxX 37. Sweeting AJ, Cormack SJ, Morgan S, Aughey RJ. When
accelerometers for measuring physical activity in Australian Is a Sprint a Sprint? A Review of the Analysis of Team-
football. Int J Sports Physiol Perform. 2011;6(3):311-21. Sport Athlete Activity Profile. Front Physiol. 2017 2017-
June-20;8(432).
24. Barrett S, Midgley A, Lovell R. PlayerLoad: reliabil-
ity, convergent validity, and influence of unit position dur- 38. Dwyer DB, Gabbett TJ. Global positioning system
ing treadmill running. Int J Sports Physiol Perform. 2014 data analysis: velocity ranges and a new definition of sprint-
Nov;9(6):945-52. ing for field sport athletes. J Strength Cond Res. 2012
Mar;26(3):818-24.
25. Migueles JH, Cadenas-Sanchez C, Ekelund U, Delisle
Nyström C, Mora-Gonzalez J, Löf M, et al. Accelerometer 39. Coutts AJ, Duffield R. Validity and reliability of GPS
Data Collection and Processing Criteria to Assess Physical Ac- devices for measuring movement demands of team sports. J
tivity and Other Outcomes: A Systematic Review and Prac- Sci Med Sport. 2010 Jan;13(1):133-5.
tical Considerations. Sports Med. 2017 Sep;47(9):1821-45.
40. Reardon C, Tobin DP, Delahunt E. Application of In-
26. Gómez-Carmona CD, Pino-Ortega J, Sánchez-Ureña B, dividualized Speed Thresholds to Interpret Position Specific
Ibáñez SJ, Rojas-Valverde D. Accelerometry-Based External Running Demands in Elite Professional Rugby Union: A GPS
Load Indicators in Sport: Too Many Options, Same Practical Study. PLoS One. 2015;10(7):e0133410.
Outcome? Int J Environ Res Public Health. 2019;16(24):5101.
41. Lovell R, Abt G. Individualization of time-motion anal-
27. Gurchiek R, Rupasinghe Arachchige Don H, Watagoda ysis: a case-cohort example. Int J Sports Physiol Perform.
L, McGinnis R, Werkhoven H, Needle A, et al. Sprint Assess- 2013 Jul;8(4):456-8.
ment Using Machine Learning and a Wearable Accelerometer.
J Appl Biomech. 2019 01/24;35:1-16. 42. Lovell R, Scott D, Park L. Soccer velocity thresholds: do
we really know what’s best? Science and Medicine in Football.
28. Meylan C, Trewin J, McKean K. Quantifying Explosive 2019 01/02;3:85-6.
Actions in International Women’s Soccer. . Int J Sports Phy
Perf. 2017:310-5. 43. McLean BD, Cummins C, Conlan G, Duthie G, Coutts
AJ. The Fit Matters: Influence of Accelerometer Fitting and
29. Hulin BT, Gabbett TJ, Johnston RD, Jenkins DG. Wear- Training Drill Demands on Load Measures in Rugby League
able microtechnology can accurately identify collision events Players. Int J Sports Physiol Perform. 2018 Sep 1;13(8):1083-
during professional rugby league match-play. J Sci Med Sport. 9.
2017 Jul;20(7):638-42.
44. Schelling X, Torres L. Accelerometer Load Profiles for
30. Douglas A, Johnston K, Baker J, Rotondi MA, Jam- Basketball-Specific Drills in Elite Players. J. Sports Sci. Med.
nik VK, Macpherson AK. On-Ice Measures of External Load 2016 Dec;15(4):585-91.
in Relation to Match Outcome in Elite Female Ice Hockey.
Sports (Basel). 2019 Jul 16;7(7). 45. Lacome M, Simpson B, Buchheit M. Monitoring training
status with player-tracking technology. Still on the road to
31. Chambers R, Gabbett TJ, Cole MH, Beard A. The Use Rome. Part 2. Aspetar J. 2018 06/21;7:64-6.
of Wearable Microsensors to Quantify Sport-Specific Move-
ments. Sports Med. 2015 Apr 3;45(7):1065-81. 46. Lacome M, Simpson BM, Cholley Y, Lambert P, Buch-
heit M. Small-Sided Games in Elite Soccer: Does One Size Fit
32. Bishop D, Girard O, Mendez-Villanueva A. Repeated- All? Int J Sports Physiol Perform. 2018 May 1;13(5):568-76.
sprint ability - part II: recommendations for training. Sports
Med. 2011 Sep 1;41(9):741-56. 47. Broman K, Woo K. Data organization in spreadsheets.
Am Stat. 2018;72.1:2-10.
33. Girard O, Mendez-Villanueva A, Bishop D. Repeated-
sprint ability - part I: factors contributing to fatigue. Sports
Med. 2011 Aug 1;41(8):673-94.
Copyright: The article published on Science Performance
34. Wundersitz DW, Gastin PB, Richter C, Robertson SJ, and Science Reports are distributed under the terms of
the Creative Commons Attribution 4.0 International License
Netto KJ. Validity of a trunk-mounted accelerometer to as- (http://creativecommons.org/licenses/by/4.0/), which permits un-

sportperfsci.com 8 SPSR - 2022 | Mar | 159 | v1

Tracking Systems in Team Sports

restricted use, distribution and reproduction in any medium, pro- cation waiver (http://creativecommons.org/publicdomain/zero/1.0/)
vided you give appropriate credit to the original author(s) and the applies to the data made available in this article, unless otherwise
source, provide a link to the Creative Commons license, and indicate stated.
if changes were made. The Creative Commons Public Domain Dedi-

sportperfsci.com 9 SPSR - 2022 | Mar | 159 | v1