Available via license: CC BY 4.0
Content may be subject to copyright.
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 1
GOLD SCORE ATHLETICS: TALENT DETECTION MODEL
FOR TRACK AND FIELD
GOLD SCORE ATHLETICS: MODELO DE DETECÇÃO DE TALENTOS PARA O ATLETISMO
GOLD SCORE ATHLETICS: MODELO DE DETECCIÓN DE TALENTOS PARA EL ATLETISMO
Guilherme Eugênio van Keulen1,2
(Physical Education Professional)
Francisco Zacaron Werneck3
(Physical Education Professional)
Emerson Filipino Coelho3
(Physical Education Professional)
Caio Márcio Aguiar3
(Physical Education Professional)
Luciano Miranda2, 4
(Physical Education Professional)
Jorge Roberto Perrout de Lima2
(Physical Education Professional)
1. Institute of Social Sciences,
Education and Zootechnics (ICSEZ)
of the Universidade Federal do
Amazonas (UFAM), Parintins,
AM, Brazil.
2. Graduate program from
the Physical Education and
Sports College (FAEFID) of the
Universidade Federal de Juiz de
Fora (UFJF), Juiz de Fora, MG, Brazil.
3. Universidade Federal de Ouro
Preto (UFOP), School of Physical
Education (EEF), Laboratory of
Exercise and Sports Studies and
Research (LABESPEE), Ouro Preto,
MG, Brazil.
4. Military School of Juiz de Fora
(CMJF), Juiz de Fora, MG, Brazil.
Correspondence:
Jorge Roberto Perrout de Lima
Graduate program from
the Physical Education and
Sports School (FAEFID) at the
Universidade Federal de Juiz de
Fora (UFJF)
n/n, Rua José Lourenço Kelmer,
Campus Universitário - São Pedro,
Juiz de Fora, MG, Brazil. 36036-900.
jorge.perrout@gmail.com
ABSTRACT
Introduction: Talent detection is a dynamic and multifactorial process that must start at school. Objective:
Create a mathematical model for evaluating the sporting potential of schoolchildren for athletics in speed,
throwing, and endurance events and to test its psychometric properties. Methods: 2871 schoolchildren of
both sexes, from 11 to 17 years old, from a military school participated. Between 2015 and 2019, students
were submitted to a multidimensional battery of tests containing anthropometric, physical-motor, psycholo-
gical, socio-environmental, maturational, and performance indicators. In addition, ten teachers evaluated the
students regarding the intangibles aspects of their sporting potential and the expectation of future success
during this period. Adopting analytical and heuristic procedures, the Gold Score Athletics was created – linear,
hybrid (tests + coaches´ eye), and weighted index, according to each indicator’s importance, depending on
the event type. Results: In the model validation sample (n = 1384), 13.9%, 16.6%, and 11.7% of boys and 10.9%,
10.1%, and 9.1% of girls were classified as high potential (Gold Score ≥ 60) for speed, throwing and endurance
events, respectively. Internal consistency (r = 0.76 to 0.82) and diagnostic stability were high (r = 0.72 to 0.81).
The Gold Score Athletics for sprinters, throwers, and long-distance runners, both for boys and girls, was higher
in students selected for a national competition when compared to those not selected (p < 0.001; d: 0.95 a 1.44)
– construct validity – and higher in medalists in an athletics competition, held two years after diagnosis, when
compared to non-medalists (p < 0.05; d: 0.62 a 1.87) – predictive validity. Conclusion: The Gold Score Athletics
is a valid and reliable scientific model for evaluating the sport’s potential of schoolchildren, being useful in the
talents detection for Athletics. Level of Evidence II; Diagnostic study.
Keywords: Track and Field; Physical Fitness; Statistics.
RESUMO
Introdução: A detecção de talentos é um processo dinâmico e multifatorial que deve começar pela escola. Ob-
jetivo: Criar um modelo matemático de avaliação do potencial esportivo de escolares para as provas de velocidade,
lançamentos e resistência no atletismo, e testar as suas propriedades psicométricas. Métodos: Participaram 2871
escolares de ambos os sexos de 11 a 17 anos de um colégio militar. Os alunos foram submetidos a uma bateria de
testes multidimensionais, contendo indicadores antropométricos, físico-motores, psicológicos, socioambientais,
maturacionais e de desempenho. 10 professores avaliaram os alunos quanto aos aspectos intangíveis do potencial
esportivo e a expectativa de sucesso futuro. Adotando procedimentos analíticos e heurísticos, criou-se o Gold Score
Athletics – índice linear, híbrido (testes + olho do treinador) e ponderado, de acordo com a importância de cada
indicador em função do tipo de prova. Resultados: Na amostra de validação do modelo (n = 1384), 13,9%, 16,6% e
11,7% dos meninos e 10,9%, 10,1% e 9,1% das meninas foram classificados como elevado potencial (Gold Score ≥
60) para provas de velocidade, lançamentos e resistência respectivamente. A consistência interna (r = 0,76 a 0,82)
e estabilidade do diagnóstico foram elevadas (r = 0,72 a 0,81). O Gold Score Athletics para velocistas, lançadores e
corredores de longa distância, para ambos os sexos, foi maior nos estudantes selecionados para uma competição
nacional quando comparados aos não selecionados (p < 0,001; d: 0,95 a 1,44) – validade de construto – e maior nos
medalhistas em uma competição de Atletismo, realizada dois anos após o diagnóstico, quando comparados aos não
medalhistas (p < 0,05; d: 0,62 a 1,87) – validade preditiva. Conclusão: O Gold Score Athletics é um modelo científico
válido e fidedigno de avaliação do potencial esportivo de escolares, sendo útil na detecção de talentos para o Atletismo.
Nível de Evidência II; Estudo diagnóstico.
Descritores: Atletismo; Aptidão Física; Estatística.
RESUMEN
Introducción: La detección de talentos es un proceso dinámico y multifactorial que debe iniciarse en la escuela.
Objetivo: Crear un modelo matemático para evaluar el potencial deportivo de escolares para pruebas de velocidad,
lanzamiento y resistencia en atletismo, y probar sus propiedades psicométricas. Métodos: Participaron 2871 escolares
de ambos sexos de 11 a 17 años de una escuela militar. Los estudiantes fueron sometidos a una batería de pruebas
multidimensionales, que contenían indicadores antropométricos, físico-motores, psicológicos, socioambientales,
Associate Editor responsible for the review process: André Pedrinelli
Original article
Artigo originAl
Artículo originAl
TRAINING
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 2
madurativos y de desempeño. 10 docentes evaluaron a los alumnos sobre los aspectos intangibles del potencial
deportivo y la expectativa de éxito futuro. Adoptando procedimientos analíticos y heurísticos, se creó el Gold Score
Athletics, índice lineal, híbrido (pruebas + mirada del profesor) y ponderado, según la importancia de cada indicador
según el tipo de prueba. Resultados: En la muestra de validación del modelo (n = 1384), el 13,9%, 16,6% y 11,7% de los
niños y el 10,9%, 10,1% y 9,1% de las niñas fueron clasificados como de alto potencial (Gold Score ≥ 60) en velocidad,
lanzamiento y eventos de resistencia. La consistencia interna (r = 0,76 a 0,82) y la estabilidad diagnóstica fueron
altas (r = 0,72 a 0,81). El Gold Score Athletics para velocistas, lanzadores y corredores de fondo, para ambos sexos, fue
mayor en los estudiantes seleccionados para una competición nacional en comparación con los no seleccionados (p
< 0.001; d: 0,95 a 1,44) – validez del constructo – y mayor en medallistas en una competición de atletismo, realizada
dos años después del diagnóstico, en comparación con los no medallistas (p < 0,05; d: 0,62 a 1,87) – validez predic-
tiva. Conclusión: El Gold Score Athletics es un modelo científico válido y fiable para evaluar el potencial deportivo de
los escolares, siendo útil en la detección de talentos para el Atletismo. Nivel de Evidencia II; Estudio diagnóstico.
Descriptores: Atletismo; Aptitud Física; Estadística.
Article received on 03/12/2022 accepted on 08/22/2022
DOI: http://dx.doi.org/10.1590/1517-8692202430012022_0147i
INTRODUCTION
The search for sports talent is part of the daily life of coaches, managers,
sports clubs, and Sports Science.1-4 The identification of sports talent is a
step in the long-term training process that aims to detect young athletes
with high potential to become elite athletes.5,6 When this process is car-
ried out at school with young people not yet involved in systematic and
specialized sports practice, it is called talent detection. Several countries
have systematic models for identifying and developing talented young
athletes.7 In athletics, this topic has been widely investigated by research-
ers,
8-10
specially to develop models and tools that enable greater accuracy
and success in identifying talent.11-15 At school, however, no models are
found to estimate the potential of children for athletics.
Identifying and developing sporting talent is one of the pillars of
international sporting success.
16,17
The existing consensus is that this
process should begin at school.18-20 The school should promote sports
practice both from the perspective of a physically active lifestyle and
social inclusion and for talent development. Besides offering all students
supervised, diversified, and quality sports practice, the Physical Education
teacher must provide the appropriate development conditions for those
with high sporting potential.18,21,22 In this sense, the school must evalu-
ate the sporting potential of students as a first step in discovering new
talents, using a multidimensional, longitudinal, and inclusive approach.23,24
Every child and adolescent have a sporting potential that must be
evaluated to be adequately developed. This sporting potential results
from a dynamic interaction of multiple indicators related to the individual,
the task, and the environment, such as anthropometric, physical-motor,
psychological characteristics, skills, family support, quantity, and quality
of training, which change over time and determine long-term sporting
performance.25-28 The scientific method has contributed to the under-
standing of the intervening factors that lead the young talented athlete
to become an elite athlete throughout the training process.27
The scientific talent identification models aim to identify young people
with high sporting potential, guide them to the sports that best suit their
profile, select the most talented, and predict future success.
29
Talent is
known to be identifiable, and that future performance can be predicted.
However, it is a difficult and often inaccurate process,30 since the trajec-
tory of athlete development to high performance is often not linear.
31
However, advances in data science have allowed the development of
expert systems for detecting sports talent, combining objective data from
athletes obtained through tests of tests and performance in competition,
the subjective assessment made by coaches, and statistical modeling.
32,24
Sport Interactive in the UK,
34
Sport Talent in Croatia,
32
and the Flemish
Sports Compass in Belgium33 are examples of talent identification models
used in schools. In Brazil, the Z-Celafiscs Strategy35 and, later, PROESP -
Projeto Esporte Brasil
36
are precursor scientific models for talent detection.
Aiming to improve the previous models, researchers at the Universidade
Federal de Ouro Preto developed the Golden Athletes® Project to validate
an intelligent system for multidimensional and longitudinal assessment
of the sporting potential of children and adolescents.24,37
The Projeto Atletas de Ouro® began at the Military School of Juiz de
Fora (CMJF) in 2015, intending to identify students with high sports skills,
map their strengths and weaknesses, guide them to the modalities more
appropriate to their profile, assisting teachers in the process of long-
term sports training. Using a battery of general and multidimensional
tests in schoolchildren aged 11 to 17, a model for assessing sporting
potential was created, including biological maturation and subjective
assessment by teachers. Since then, specific talent detection models
have been developed for soccer, basketball, swimming, and trampoline
gymnastics.24 Continuing the development of the tool, further studies
are needed to model other sports, including athletics.
Track and field athletics comprises running, jumping, throwing, com-
bined events, field races, mountain races, and athletic walking.
38
Each type of
event requires a specific profile for high performance, which implies young
athletes’ guidance, selection, and development processes.
39,40
Sprinters
and jumpers, for example, have apparently developed muscle mass and
high power of lower limbs; throwers are taller and have high body mass
associated with upper limb strength, while fundists have though aerobic
capacity, movement efficiency, and a low percentage of body fat.41,1 The
morphological and physical capacity differences between the types
of tests make athletics a sport of high analytical complexity.
42,43
In this
sense, systematic methods for evaluating multiple indicators of sporting
potential, linked to the long-term training process, can help the young
person choose the sport most appropriate for his profile.33
Athletes’ success in such distinct events is determined by a diverse set
of morphological and motor characteristics such as height, limb length,
strength, aerobic capacity, power, and speed, articulated with technical
aspects specific to each event.
44
Young athletes who possess an optimal
combination of the performance indicators of the sport and who respond
favorably to training and competition have a greater chance of future suc-
cess.3,33 For this, it is necessary to continuously monitor biological maturation,
physical growth, and physiological and motor adaptations in response to
the training provided.
14
Moreover, coaches’ knowledge adds value to talent
identification models, especially in defining which indicators should be
evaluated and the relative importance of each one in talent development.10,40
Scientific evidence has contributed to a better understanding of the
process of identifying and developing talent in track and field, based on
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 3
studies of athletes from different tests,47,46 comparison of athletes from
different competitive levels,
47,48
performance prediction,
40,14,15
proposals
for identifying and selecting young athletes,
8,13,9,6,49
talent development
models11,
24
and longitudinal studies of athletes’ career development.
50,12
However, there is a need to develop new tools for talent detection
in athletics, especially in the Brazilian school setting, considering the
interaction of multiple indicators of sporting potential.44,51
Given the above, the objective of this study was to create a math-
ematical model for assessing the sporting potential of schoolchildren
in athletics, specifically for the speed, throwing, and endurance tests,
and to test its psychometric properties. The hypothesis is that the model
will be valid and reliable for estimating the athletic potential of school-
children for athletics.
MATERIALS AND METHODS
This study is part of the longitudinal research called “Projeto Atletas de
Ouro®: Avaliação Multidimensional e Longitudinal do Potencial Esportivo de
Jovens Atletas” (Multidimensional and Longitudinal Assessment of Sporting
Potential in Young Athletes), approved by the Research Ethics Committee
of the Universidade Federal de Ouro Preto (CAAE: 32959814.4.1001.5150).
In this study stage, the mathematical model for assessing the sporting
potential of young athletes in track and field was developed.
Research Model
A total of 2871 CMJF schoolchildren participated in the study, 1312
female and 1559 male, aged 11 to 17 years, assessed from 2015 to 2019.
The sample for the construction of the mathematical model was com-
posed of 1487 records of tests performed in female (n = 659) and male
(n = 828) CMJF students, aged between 11 and 17 years, evaluated in
the period from August 2015 to June 2017. One thousand one hundred
eighteen records were of students who participated only in physical edu-
cation classes, 113 were athletics practitioners, and 256 practiced other
sports. In the teachers’ evaluation of the sporting potential, 33.2% were
classified as high potential, and athletics was suggested as the sport with
the highest probability of future success for 16.2% of the students. In
addition, 37.9% of the students had competitive experience, 20.6% had
already won a medal at least at the municipal level, and 9.7% had par-
ticipated in the Friendship Games - a national school competition. The
mathematical model validation sample consisted of 1384 test records
(653 girls and 731 boys) conducted from March 2018 to March 2019. A
total of 10 teacher-coaches from the Physical Education Section of CMJF
(mean age 41.0±8.0 years and mean time of experience 12.5±9.8 years)
specialized in different sports (soccer, volleyball, orienteering, swimming,
fencing, military triathlon, basketball, volleyball, handball, and track and
field) and with academic backgrounds - undergraduate (n=2), specialist
(n=3), and master’s (n=5) - participated, most of them former athletes.
The inclusion criteria for participation in the study were: age range of 11
to 17, being enrolled and regularly attending classes at CMJF, and being
present on the day of data collection. The students who did not hand in the
TCLE signed by their guardian or who refused to participate were excluded,
as well as those who presented any physical or clinical condition that inter-
fered with the performance of the tests. The consent of the legal guardians
and the agreement students obtained before participation in the study.
Instruments and Procedures
Between 2015 and 2019, the schoolchildren performed an annual
battery of multidimensional tests to measure anthropometric, physical-mo-
tor, psychological, environmental, and maturational indicators related to
sporting potential. In addition, the PE teachers evaluated their students on
the intangible aspects of sporting potential and the expectation of future
success (Chart 1). The indicators of sporting potential evaluated, as well as
the procedures for the tests, and measurements are described in Miranda
et al. (2019),52 Ribeiro Júnior et al. (2019)53 and Werneck et al. (2020).24
The battery of tests was applied during the students’ physical education
class time, lasting approximately 90 minutes, on three different days. The
data were collected from Monday to Friday between 09:00 and 12:30.
The evaluation was done by properly trained professionals, with fixed
evaluators selected for each test. On the first day, a lecture was held in
the CMJF auditorium, where the testing protocol and the collection of
socio-demographic information and sporting experience of the students
were explained, under the supervision of the Physical Education teachers.
On the second day, anthropometric measurements were collected and
physical-motor tests were performed, in the form of a circuit, in the gym.
On the third day, a 20-meter back-and-forth race test was performed to
evaluate cardiorespiratory endurance. Then, the collected data were sto-
red in an electronic spreadsheet, using Excel® software, version Windows
10.0. Finally, procedures were performed for organizing, validating, and
debugging the data, creating new variables, standardizing, and creating
graphs to present the individual student results.
Chart 1. Factors and indicators of sporting potential evaluated by the battery of tests
of the Golden Athletes Project® in schoolchildren.
Factors Indicators Unit/Classification
Test Battery
Anthropometric
Body Mass kg
Height cm
Expected adult height (EAP) cm
Sitting Height cm
Length Lower Limbs cm
Spread cm
Body Fat %
Physical-Motor
Handgrip strength test kgf
Medicine ball throwing test (2kg) m
Countermovement vertical jump cm
10m and 20m speed race s
Sitting and Reaching Flexibility Test cm
20m back and forth race m / VO2max
Psychological
SOQ - Motivational Orientation
(competitive, winner, determined) pts
ACSI-28 - Coping Skills
(coping with adversity, performance
under pressure, goals/mental
readiness, concentration, worry-free,
confidence/motivation, coachability)
pts
Perceived Athletic Competence pts
Environmental
Sports experience (train, practice
time, sports preference)
yes / no
years and months
/ type of sport
Competitive level local/regional, state,
national, international
Victory in competition
Socioeconomic level (ABEP) pts / A, B, C, D, E
Level of physical activity pts
Family Participation pts
Parents’ sports practice yes / no
Athlete in the family yes / no
Maturational
Percentage EAP achieved %
APS Z-score achieved delayed, normomature,
advanced
Maturity offset years
PVC Age years
Subjective Teacher Evaluation
Teacher’s
Perspective
Sporting Potential Likert Scale from 1 to 5
Intangible Aspects pts
A description of the test procedures and measurements can be found in Miranda et al. (2019), Ribeiro Júnior et
al. (2019) and Werneck, Coelho, and Ferreira (2020).
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 4
Modeling the sporting potential of schoolchildren
Statistical modeling aims to model phenomena that have uncer-
tainties and extract knowledge for decision-making.54 It represents the
operational way scientific models for identifying sports talent quantify
the potential of young people. The assumption is that it is possible to
identify sports talents and predict future performance with some pro-
bability of success, aiming to help physical education teachers in the
decision-making process related to developing their student’s potential.
A phenomenon that cannot be observed directly is evaluated using
indicators. In practice, we estimate the sporting potential by diagnosing
personal and environmental characteristics inherited, acquired, measured,
and observed. Once analyzed, we can estimate the student’s potential for
a certain sport. The modeling of the sporting potential of schoolchildren,
therefore, comprises a procedure that aims to obtain a valid and reliable
estimate of the sporting potential of students for a particular sport from
the analytical and heuristic processing of multiple indicators of sporting
talent and is represented by a mathematical equation.24
Building the model
In this study, the construction of the operational model for assessing
the athletic potential for athletics was divided into steps, according to
the method described by Werneck et al. (2020)24 and Werneck, Coelho,
and Miranda.
51
In the first step, a univariate descriptive analysis was per-
formed to calculate and measure the data’s central tendency, dispersion,
position, and distribution. Next, quantitative variables were normalized,
according to the procedures adopted by the Z-Celafiscs Strategy.35 The
Z score of tests in which performance was against time was reversed
(10-meter speed run, for example), so that r values always represented
higher performance. To calculate the Z score, the mean and standard
deviation were used as reference values, by age group and sex, accord-
ing to the study by Miranda et al. (2019).
52
Based on the standardized
normal distribution, the Z score for each indicator was converted to
the corresponding percentile value so that the indicator score ranged
from 0 to 100%. Students with exceptional performances above the
90th percentile (P90) in the indicators relevant to performance in speed,
throwing and endurance modalities had a bonus score in their final score.
In the second step, a top-down performance prediction approa-
ch55 was adopted to investigate differences and similarities among the
schoolchildren, relationships among the variables, and possible perfor-
mance and potential sporting determinants. For this purpose, bivariate
statistical analyses were performed (t-tests and Chi-square test) to find
differences in the various indicators between students practicing athletics
vs. non-practicing, students with high potential according to teachers’
subjective perception vs. low potential, students with potential for future
performance in athletics vs. another modality, and students selected
for the Friendship Games vs. not selected, as well as medal winners vs.
non-medal winners. In this exploratory analysis, the statistical signifi-
cance and the size of the effect (practical relevance) of the differences
found for each indicator analyzed was observed, as one of the criteria
for choosing the variables to be included in the mathematical model.
In the third stage, we adopted a bottom-up approach to acquiring
knowledge of the factors necessary for developing young elite athle-
tes.55 To this end, we used two sources of information: the knowledge
of coaches(experts)10 and the available scientific literature. Based on
the athletics performance model and previous studies, a questionnaire
was administered to 10 Brazilian track and field coaches to investigate
the degree of importance attributed to different factors and indicators
that determine performance in different track and field events.10 Also,
retrospective longitudinal studies with elite athletes and Olympians
were reviewed to obtain evidence and search for patterns regarding the
characteristics that explain athletic success in track and field.
In the fourth stage, based on the analytical and heuristic procedures
used in the previous stages, we operationalized the estimation of sporting
potential through an index called Gold Score Athletics. The Gold Score Athle-
tics is a standardized index ranging from 0 to 100%, obtained by a linear
equation, composed of 6 factors with 24 indicators, presented in Chart 2.
The relative importance of factors and indicators of sporting potential was
defined from exploratory data analysis, literature review, and expert knowled-
ge, varying between the speed, throwing, and endurance events. Therefore,
Gold Score Athletics is a hybrid multidimensional and multidisciplinary model
that combines observed test performance and developmental potential
assessed by coaches, generating a quantitative estimate of schoolchildren’s
athletic potential for athletics. The criteria for determining sporting talent
was Gold Score >80% in the mathematical model.
Calculation Gold Score Athletics
For each factor, the percentile values of each indicator are added toge-
ther and multiplied by their respective weights. Then, the result is divided
by the sum of the weights of the indicators (
∑βFi
). Then the results obtained
in each factor are multiplied by their respective weights, and divided by
the sum of the factor weights (
∑FαF
). The student with scores above the 90th
percentile (P90) in the indicators relevant to performance in the question
test (speed, throwing or endurance) and the variables sport preference,
sport indicated by the teacher, and somatic maturation enter the equation
as an adjustment factor. Equation 1 defines Gold Score Athletics.
Equation 1
Where, GS is an individual’s Gold Score Athletics. F are the factors
of the sport potential,IFi is the i-th indicator evaluated by the battery of
tests corresponding to the factor F, βFi is the weight of the i-th indicator
of the factor F, and αF is the weight of the factor. βi ∈ {1,2,3...10} and αF ∈
{1,2,3,4}. FA is the adjustment factor.
The classification of the schoolchildren in Gold Score Athletics was
defined using the following criteria: <40% Developing Sport Potential;
40-59% Average Sport Potential; 60-80% High Sport Potential; >80%
Sport Potential of Excellence.
Model Validation
For construct validity, the outcome was considered whether the
student was selected for the Friendship Games from 2018 to 2019. For
predictive validity, it was considered as an outcome of whether the
student was a medalist or not in the athletics competition held in the
Internal Games of the Military School of Juiz de Fora in 2021.
Chart 2. Factors and indicators used in the mathematical model for calculating Gold
Score Athletics.
Factors Indicators
F1-
Anthropometric
Predicted Adult Height, Body Mass
and Size.
F2- Physical-Motor Medicine ball Throw, Countermovement jump,
Speed 20m, Endurance, Motor Talent.
F3-Psychological Perceived Competence, Competitive, Winning,
Determined, Confidence/Motivation, Trainability, and Coping.
F4-Environmental Participation in Training, Competitive Level, Family Support
F5-Intangible Intangible Aspects, Sports Potential.
F6-Performance Victory in Competition
Setting Sports Preference. Sport indicated by the
Teacher/Trainer, Maturational Stage.
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 5
Statistical Analysis
The data were described using mean ± standard deviation (quantita-
tive variables) and percentages (qualitative variables). Cronbach’s Alpha
correlation coefficient measured the model internal consistency. The
intraclass correlation coefficient (ICC) was used to analyze the stability
of the diagnosis with a 12-month interval between the first and second
evaluations. Construct Student’s t-test assessed validity and criterion
validity. Finally, the effect size was calculated by Cohen’s d. All analyses
were done in IBM SPSS software version 24.0 (IBM Corp., Armonk, NY).
A value of p≤0.05 was adopted for statistical significance.
RESULTS
Table 1 presents the general characteristics of the students who prac-
tice athletics and the Gold Score Athletics comparison to non-athletes.
It was possible to observe that in the analysis of sprinters, 46.9% of the
male students were identified as developing potential, 39.3% as medium
potential, 11.4% as having high potential, and 2.5% as having potential for
excellence. In females, 56.4% were identified as developing potential, 32.8%
as medium potential, 9.8% as having high potential, and 1.1% as having
excellent potential. For the thrower category, the percentage values found
for the male schoolchildren were: 46.4% identified as developing potential,
37.1% as medium potential, 14.0% as high potential, and 2.6% as excellent
potential. For the schoolgirls, they were: 47.8% were identified as developing
potential, 42.1% as medium potential, 9.2% as high potential, and 0.9% as
excellent potential. In the analysis regarding the sporting potential of male
fundists, 52.4% were classified as developing potential, 35.8% as medium
potential, 8.6% as high potential, and 3.1% as the potential for excellence.
In the analysis of the sporting potential of female fundists, the values found
were: 60.2% as developing potential, 30.8% as medium potential, 7.7% as high
potential, and 1.4% as excellent potential. The data are presented in Figure 1.
Statistically significant differences were observed in all comparisons,
both for boys [sprinters vs. throwers (p= 0.004); sprinters vs. (p= <0,001);
Table 1. Comparison of multidimensional indicators of sporting potential and Gold Score Athletics of 13 to 17-year-old school athletes and non-athletes.
Indicators
Male Female
Athletes (n = 103) Schoolchildren
(n = 410) p-value d Athletes (n = 72) Schoolchildren
(n = 439) p-value d
Chronological age (years) 15.6±1.3 15.0±1.3 <0.001 0.46 15.0±1.3 15.0±1.3 0.92 0.01
Birth Quartile (1ºQ) 22.3% 22.7% 0.42 0.07 19.4% 28.0% 0.16 0.10
Anthropometric
Body Mass (kg) 60.0±9.1 61.5±13.5 0.19 0.11 52.5±9.2 54.7±10.9 0.11 0.20
Height (cm) 170.5±7.0 169.2±7.9 0.13 0.16 159.6±5.9 160.1±6.0 0.45 0.08
Width (cm) 174.0±8.2 172.7±9.1 0.19 0.14 161.4±6.4 161.9±7.4 0.54 0.06
Fat Percentage (%) 14.1±6.8 17.2±7.1 <0.001 0.43 21.6±4.4 24.0±5.7 0.001 0.42
Physical-Motor
Flexibility (cm) 25.2±8.1 22.8±9.1 0.02 0.26 32.2±8.4 29.1±8.1 0.005 0.38
Hand Grip (kgf ) 36.6±8.5 33.4±8.9 0.001 0.36 25.5±5.3 24.9 ±5.7 0.46 0.10
Medicine Ball Throw (m) 5.2±0.9 4.8±0.9 <0.001 0.44 3.6±0.48 3.4±0.48 0.01 0.41
Vertical jump (cm) 33.8±7.4 28.8±6.6 <0.001 0.75 24.4±4.2 21.0±4.2 <0.001 0.80
Speed 20 m (s) 3.28±0.27 3.49±0.28 <0.001 0.75 3.71±0.26 3.98±0.31 <0.001 0.87
20m back and forth race (m) 1470.5±410.1 1047.7±330.6 <0.001 1.27 887.4±250.0 645.0±223.3 <0.001 1.08
VO2max (ml/kg/min) 50.4±5.3 45.4±4.8 <0.001 1.04 43.0±4.1 39.6±4.1 <0.001 0.82
Maturational
Predicted adult height (cm) 176.7±6.7 178.3±6.7 0.04 0.23 162.9±5.1 163.3±5.5 0.60 0.07
PAH (%) 96.5±3.5 94.9±4.1 0.001 0.39 97.9±1.9 98.0±1.9 0.76 0.05
Z-score 0.60±0.77 0.59±0.74 0.85 0.01 -0.72±1.08 -0.68±1.17 0.75 0.03
Status Maturacional
Delayed 0.0% 1.9%
0.37
0.06 45.7% 38.8%
0.55 0.05Normomature 72.7% 72.9% 47.1% 53.4%
Advanced 27.3% 25.2% 7.2% 7.8%
MO (years) 1.42±1.1 1.0±1.2 0.001 0.35 2.1±0.9 2.2±0.9 0.74 0.11
APHV (years) 14.1±0.7 14.0±0.6 0.005 0.16 12.9±0.6 12.8±0.6 0.51 0.16
Psychosocial
Coping Skills 12.4±2.8 10.4±3.0 <0.001 0.66 11.6±3.1 9.5±2.7 <0.001 0.77
Perceived Competence 7.2±1.6 6.2±1.7 <0.001 0.59 6.9±1.5 5.9±1.6 <0.001 0.21
Winner 2.6±0.8 2.5±0.8 0.003 0.12 3.5±0.7 3.4±0.8 0.37 0.12
Determined 4.3±0.7 4.0±0.9 <0.001 0.33 4.3±0.5 3.8±0.8 0.002 0.62
Competitive 4.1±0.8 3.5±0.8 <0.001 0.75 3.9±0.7 3.3±0.8 <0.001 0.75
Family Support 26.2±7.3 24.6±8.2 0.01 0.19 26.5±8.0 22.0±8.4 <0.001 0.53
Coach Evaluation
High Sports Potential (%) 20.1% 79.9% <0.001 0.32 14.1% 85.9% <0.001 0.25
Intangibles 31.0±10.0 25.6±8.3 <0.001 0.65 27.3±8.2 22.3±7.9 <0.001 0.63
Gold Score Athletics
Sprinters (%) 55.8±17.7 37.5±13.3 <0.001 1.37 59.5±16.6 37.7±13.5 <0.001 1.61
Throwers (%) 52.7±15.2 38.4±13.3 <0.001 1.07 54.4±13.8 38.4±11.6 <0.001 1.37
Fundists(%) 55.6±17.2 35.8±11.8 <0.001 1.67 57.7±16.2 36.8±12.5 <0.001 1.67
(%PAH: Attained percentage of predicted adult height; MO: Maturity Offset; APHV: Age at peak height velocity.
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 6
sprinters vs. fundists (p= <0,001)] and for girls [sprinters vs. fundists
(p= <0,001); sprinters vs. throwers (p= <0,001); fundists vs. (Figura 2)
The internal consistency of the sport potential factors ranged from
0.76 (female fundists) to 0.82 (male fundists and throwers) and the sta-
bility of the sport potential factors ranged from 0.72 (Female throwers)
to 0.81 (throwers), as presented in Table 2.
The construct validity and criterion validity were satisfactory, and it
was observed that the selected students ( Table 3) and the medal-winning
students had higher Gold Score Athletics. The difference between non
-
-selected and non-medal-winning students was statistically significant
(Table 4). From a practical point of view, the observed differences were
of moderate to high magnitude.
DISCUSSION
In the present study, a linear, multidimensional, hybrid, computerized
mathematical model was developed that provides a valid and reliable
estimate of the sporting potential of school athletes in track and field,
called the Gold Score Athletics. In addition, the model presented satis-
factory psychometric properties, regarding internal consistency and
diagnostic stability after 12 months, validly discriminating schoolchildren
of different competitive levels and the highest level of victory in the
competition. Gold Score Athletics is a technological innovation that
combines a battery of tests, biological maturation, the coach’s eye, and
statistical modeling, forming an expert system to support teachers and
coaches in their decision-making the process of identification, selection,
and development of sports talent in school, with strong social impact
and relevance to physical education and school sports.
The development of Gold Score Athletics corroborates previous
studies that have also developed school-age talent identification models,
for example Sport Interactive in the UK,2 Sport Talent in Croatia32 and the
Flemish Sports Compass in Belgium.33
To predict future success in athletics, Henriksen, Stambulova,
and Roessler (2010)
40
developed a performance prediction mo-
del based on an ecological and holistic approach that encourages
practitioners to broaden their focus beyond themselves to help
them in a successful transition to the professional level. In Poland,
Maszczyk, Zając, and Ryguła (2011)
15
used neural models to establish
a predictive relationship for throwing sports outcomes, and found
a good relationship between the Perceptron network model and
outcome prediction. Another study that proposed predicting future
success was conducted by Liu and Schutz (1998).14 These authors
proposed identifying the best mathematical model and data set to
predict future athletic performance. But to date, it appears that Gold
Score Athletics is the first mathematical model that uses a multidi-
mensional test battery, an assessment of biological maturation, an
analysis of environmental and psychological factors, and a subjective
assessment by teachers.
To be more accurate in predicting talent, the performance and current
condition of the schoolchild must be evaluated through physical-motor
tests. Still, one must also consider what is expected concerning his or her
development and prospects. The preliminary modeling performed by
the Gold Athletes Project®, it was shown that the proposed test battery
measures current performance. At the same time, the coaches’ opinion
Table 2. Internal consistency and stability after 12 months of the sport potential
factors and the Gold Score Athletics in school athletics athletes.
Gold Score
Athletics
Internal Consistency Stability
Mean ± SD Alpha Baseline After 12
months CCI (IC95%)
Male
Sprinters 42.8±15.6 0.81 41.4±14.5 44.1±16.4 0.78 (0.71-0.83)
Throwers 44.0±15.8 0.82 42.8±15.2 45.2±16.2 0.81 (0.75-0.85)
Fundists 41.6±15.6 0.82 41.0±15.0 42.2±16.1 0.78 (0.72-0.83)
Female
Sprinters 40.0±14.7 0.76 40.3±14.1 39.7±15.3 0.77 (0.70-0.83)
Throwers 41.7±13.6 0.76 41.5±12.9 41.9±14.3 0.72 (0.63-0.78)
Fundists 38.8±14.6 0.77 39.7±14.2 38.1±15.1 0.75 (0.67-0.81)
Figure 1. Histogram of the Gold Score Athletics of male (n = 731) and female (n = 653) schoolchildren for the speed, throwing, and long-distance running events.
Figure 2. Gold Score Athletics bar-and-error graph of male (n = 731) and female
(n = 653) schoolchildren for the speed, throwing, and cross-country running events.
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 7
estimates the development potential, so they should be analyzed toge-
ther for better understand athletes’ sporting potential.
37
Thus, Gold Score
Athletics, the mathematical model proposed in this study, confirms the
hypothesis that sporting talent is identifiable and measurable. It can be
estimated through a linear equation and the main factors and indicators
of sporting potential.
As athletics is composed of a great variety of disciplines, consequently,
it demands different athletic profiles from its participants. Considering
this characteristic, the Gold Score Athletics presents itself as a valid
model to discriminate the schoolchildren regarding their potential for
each group of tests, since different scores were identified according to
the group of tests analyzed.
The construct validity of the Gold Score Athletics was evidenced when
it was observed that the scores developed to compare schoolchildren
selected for the games were higher than the scores of schoolchildren
who had not been selected. It was also observed that the Gold Score
Athletics showed predictive validity since the scores of medal-winning
schoolchildren were higher than those of non-medal-winning school-
children, demonstrating that the predictive data initially collected by
the model for athletic talent was confirmed through the positive per-
formance results (being a medal winner). These results confirm that the
model, for being multidimensional, presents greater effectiveness than
an individual analysis of teachers and coaches, or even of models that
consider, for example, only the physical-motor aspects, corroborating the
statements of the study by Baker, Cobley, Schorer, and Wattie (2017).29
The results also corroborate several studies in athletics that found
statistically significant differences between athletes in different events45,46
and competitive levels
47,48
and that have investigated performance
predictor variables.40,14,15
CONCLUSION
The Gold Score Athletics presents itself as a dynamic tool, with a holis-
tic characteristic, applied to the process of identification and development
of sports talent in school, since it qualitatively and quantitatively analyzes
a large number of multidimensional characteristics associated with the
sports potential of students practicing or not athletics, becoming an
essential tool for talent detection and for predicting future performance.
Gold Score Athletics presents itself as an important tool, as it is applica-
ble in schools and institutions that select, identify, promote, and develop
school athletes. With the use of the computerized system, based on the
Gold Score Athletics, it is possible to recognize those students who have
the greatest potential for excellence in athletics, to increase investments
in financial and human resources in the training process of these students,
to optimize training to improve the potential and minimize weaknesses,
supporting possible decisions of teachers and coaches in the inclusion
or exclusion during the development process of the sport.
However, it is necessary to apply systematic evaluations, avoiding
hasty judgments based only on cross-cutting diagnoses, guaranteeing
development opportunities to all students. The model allows highlighting
the best-performing schoolchildren at the time of evaluation, but also to
conduct the results so that the schoolchildren are offered the best conditions
to develop at the limit of their potential, considering that they may present a
superior performance in the future, also considering the maturation process.34
As limitations, it should be noted that the normative values refer to the
sample itself, making it possible to generalize the results only to the com-
petitive level in which the study participants are inserted. Nevertheless, the
difficulty of detecting sports talent is inherent to the theme. Furthermore,
new studies may contribute to a better understanding of the phenome-
non, allowing constant adaptation of the model proposed in this study.
We conclude that the Gold Score Athletics is a valid and reliable
model for estimating and assessing the sporting potential of school
athletes. And, given the evidence of validity and stability of the diagnosis
performed, it can be said that the modeling of the sporting potential
proposed in this study has shown promise as an instrument to systematize
the identification of sporting talent for athletics in school.
ACKNOWLEDGMENTS
We appreciate the Universidade Federal de Ouro Preto and the Military
School of Juiz de Fora for the Research Cooperation Agreement and the
Graduate Program in Physical Education at the Federal Universidade de
Juiz de Fora for the financial support for the publication of the article.
All authors declare no potential conflict of interest related to this article
Table 3. Comparison of Gold Score Athletics for the speed, throwing, and distance
running events in schoolchildren selected and not selected for the Friendship Games
- a national level school competition.
Factors / Models Group p-value D
Selected Not Selected
Male n = 91 n = 640
Sprinters 55.2±14.7 41.0±14.9 0.000* 0.95
Throwers 56.8±14.8 42.2±15.1 0.000* 0.96
Fundists 54.5±15.1 39.8±14.8 0.000* 0.99
Female n = 77 n = 576
Sprinters 55.5±14.1 37.9±13.5 0.000* 1.30
Throwers 57.1±12.8 39.6±12.3 0.000* 1.42
Fundists 55.2±16.8 36.6±12.9 0.000* 1.44
(*statistically significant difference, p<0.05; d: effect size)
Table 4. Comparison of the Gold Score Athletics for the speed, throwing, and distance
running events in medal-winning and non-medal-winning schoolchildren in a school
athletics competition 2 to 3 years after diagnosing athletic potential.
Factors / Models Competition Performance p-value d
Medalists Non-Medalists
Male n = 24 n = 201
Sprinters 54.3±22.2 44.8±15.1 0.006* 0.62
Throwers 56.7±20.5 45.9±15.5 0.002* 0.69
Fundists 54.8±24.1 43.0±14.5 0.001* 0.81
Female n = 20 n = 160
Sprinters 63.5±13.1 40.4±14.0 0.000* 1.65
Throwers 55.0±11.9 43.8±12.5 0.000* 0.89
Fundists 64.0±13.7 41.1±12.2 0.000* 1.87
(*statistically significant difference, p<0.05; d: effect size)
AUTHORS’ CONTRIBUTIONS: Each author contributed individually and signicantly to the development of the manuscript. GEVK: substantially contributed to the writing of the manuscript, the design of
the paper, the critical review of the intellectual content, and the interpretation of the results; FZW: participated in the approval of the nal version, substantially contributed to the design of the paper, critical
review of the intellectual content, and performed the statistical analysis of the data and interpretation of the results. OBE: contributed substantially to the conception of the work and critical review of the
intellectual content. Caio Márcio de Aguiar: contributed to the data collection and critical review of the intellectual content. LM: contributed to data collection and critical review of intellectual content. JRPL:
participated in the nal approval of the version and contributed substantially to the design of the work, critical review of the intellectual content, and interpretation of the results.
REFERENCES
1.
Brown J. Sport talent: how to identify and develop outstanding athletes. Champaing: Human Knetics; 2001.
2. Abbott A, Button C, Pepping GJ, Collins D. Unnatural selection: talent identification and development
in sport. Nonlinear Dynamics Psychol Life Sci. 2005;9(1):61-88.
3.
Issurin VB. Evidence-Based Prerequisites and Precursors of Athletic Talent: A Review. Sports Med.
2017;47(10):1993-2010.
4.
Johnston K, Baker J. Waste reduction strategies: factors affecting talent wastage and the efficacy of
Rev Bras Med Esporte – 2024; Vol. 30 – e2022_0147
of 8
Page 8
talent selection. Front Psychol. 2019;10.
5.
Pankhurst A, Collins D. Talent identification and development: The need for coherence between research,
system, and process.Quest. 2013;65(1):83-97.
6. Schiffer J. Children and youths in athletics. New Stud Athl. 2008;23(3):7-18.
7.
Platonov V. Theoretical and methodological background for sports selection and orientation in modern
elite sports. Science in Olympic Sport. 2018;3:24-51.
8. Babić V, Viskić-Stalec N. A talent for sprinting--how can it be discovered and developed. Coll Antropol.
2002;26(Suppl):205-19.
9. Mirkolaee EG, Razavi SMH, Amirnejad S. A mini-review of track and field’s talent-identification models
in Iran and some designated countries. Ann Appl Sport Sci. 2013;1(3):17-28.
10.
Aguiar CM, Coelho EF, de Paula HE, Ferreira RM, de Lima JRP, Werneck FZ. Determinantes do desempenho
no atletismo: uma perspectiva dos treinadores. Conexões. 2022;20:e022004.
11.
Ben-zaken S, Meckel Y, Lidor R, Nemet D, Eliakim A. Genetic profiles and prediction of the success of
young athletes’ transition from middle- to long-distance runs: an exploratory study. Pediatr Exerc Sci.
2013;25(3):435-47.
12.
Grossocordón JG. The Royal Spanish Athletics Federation program of talent identification, development
and assistance for young athletes. New Stud Athl. 2003;18(3):35-45.
13.
Ijzerman J, Damen T, Koens G, Collée T. Improving talent identification and development in young
distance runners. New Stud Athl. 2008;23(3):35-48.
14.
Liu Y, Schutz R. Prediction models for track and field performances. Meas Phys Educ Exerc Sci.
1988;2(4):205-23.
15.
Maszczyk A, Zając A, Ryguła I. A neural network model approach to athlete selection. Sports Eng.
2011;13(2):83-93.
16.
De Bosscher V, De Knop van Bottenburg M, Shibli S, Bingham J. Explaining international sporting
success: an international comparison of elite sports system and policies in six countries. Sport Manag
Rev. 2009;12(3):113-36.
17. Vaeyens R, Lenoir M, Williams AM, Philippaerts RM. Talent identification and development programmes
in sport: current models and future directions. Sports Med. 2008;38(9):703-14.
18.
Bailey R, Morley D. Towards a model of talent development in physical education. Sport Educ Soc.
2006;11(3):211-30.
19.
Fisher R. Education and talent identification and development in sport. In: Fisher R, Bailey R, editors.
Talent identification and development – The search for sporting excellence. In Perspectives – The
multidisciplinary series of physical education and sport science. Berlim: ICSSPE; 2008. p. 224.
20.
Prieto-Ayuso AJC, González-Villora S, Fernanández-Rio J. Are physical Education lessons suitable for sport
talent identification? A systematic review of the literature. Int J Environ Res Public Health. 2020;17(6):2-12.
21.
Gaya ACA, Torres L, Cardoso V. Detection of sporting talents in school physical education: from aversion
to fascination. In: Soares Y, editors. Sports training - multifactorial aspects of performance. Rio de Janeiro:
Medbook; 2014. p. 263-74.
22.
Platvoet S, Elferink-Gemser MT, Baker J, Visscher C. Physical education teachers’ perceptions of sport
potential: Development of the scale for identification of sport potential (SISP). Annals of Research in
Physical Activity & Sport. 2015;6:63-79.
23.
Cobley S, Till K, O’Hara J, Chapman C. Variable and changing trajectories in youth athlete develop-
ment: further verification in advocating a long-term inclusive tracking approach.J Strength Cond Res.
2014;28(7):1959-70.
24.
Werneck FZ, Coelho EF, Ribas PR, Miranda L, Ferreira RM, Ribeiro Júnior DB, et al. Sistema Atletas de
Ouro - modelagem do potencial esportivo. In: Werneck FZ, Coelho EF, Ferreira RM, editors. Manual do
jovem atleta: da escola ao alto rendimento. Curitiba: CVR; 2020. p. 231-55.
25.
Campbell E, Irving R, Poudevigne M, Dilworth L, McFarlane S, Ismail O, et al. Contextual factors and sporting
success: The relationship between birth date and place of early development on the progression of
Jamaican track and field athletes from junior to senior level. PLoS One. 2019;14(12):e0227144.
26.
Phillips E, Davids K, Renshaw I, Portus M. Expert performance in sport and the dynamics of talent
development. Sports Med. 2010;40(4):271-83.
27.
Rees T, Hardy L, Güllich A, Abernethy B, Côté J, Woodman T, et al. The Great British Medalists Project:
a review of current knowledge on the development of the world’s best sporting talent. Sports Med.
2016;46(8):1041-58.
28. Thomas CE, Abbott G, Gastin PB, Main LC. Construct validity and reliability of the Talent Development
Environment Questionnaire in Caribbean youth track and field athletes. PLoS One. 2020;15(1):e0227815.
29. Baker J, Cobley S, Schorer J, Wattie N. Routledge Handbook of Talent Identification and D evelopment
in Sport. Milton Park: Taylor & Francis; 2017.
30. Till K, Baker J. Challenges and [Possible] Solutions to Optimizing Talent Identification and Development
in Sport. Front Psychol. 2020;11.
31. Gulbin J, Weissensteiner J, Oldenziel K, Gagné F. Patterns of performance development in elite athletes.
2013;13(6):605-14.
32. Papić V, Rogulj N, Pleština V. Identification of sport talents using a web-oriented expert system with a
fuzzy module. Expert Syst Appl. 2009;36(5):8830-38.
33. Pion J. The Flemish Sports Compass: from sports orientation to elite performance prediction. Ghent,
Belgium. Thesis – Ghent University. Faculty of Medicine and Health Sciences; 2015.
34. Abbott A, Collins D. A theoretical and empirical analysis of a ‘state of the art’ talent identification model.
High Abil Stud. 2002;13(2):157-78.
35.
Matsudo VKR. Prediction of future athletic excellence. In: Bar-or O, editor. The Child and adolescent
athlete. Oxford: Blackwell Science; 1996. p. 92-109.
36.
Gaya ACA. Talento esportivo: estudo de indicadores somatomotores na seleção para o desporto de
excelência. Revista Perfil. 2002;6(6):86-96.
37.
Werneck FZ, Coelho EF. Pré-Startup Atletas de Ouro: Um Novo Modelo de Ecossistema para Identificação
de Talentos. In: Miragaya A, et al, editors. Tecnologia, inovações e startups no esporte – agenda olímpica
2020 na Prática. Rio de Janeiro: Ciência Moderna; 2020.
38.
Confederação Brasileira de Atletismo [http://www.cbta.org.br]. O atletismo: origens. Transinf. [Acesso
em 2021 dez 18]. Disponível em: http://www.cbta.org.br/atletismo/origens.asp.
39.
Aoki K, Kohmura Y, Sakuma K, Koshikawa K, Naito H. Relationships between field tests of power and
athletic performance in track and field athletes specializing in power events. Int J Sports Sci Coach.
2015;10(1):133-44.
40. Henriksen K, Stambulova N, Roessler KK. Successful talent development in track and field: considering
the role of environment. Scand J Med Sci Sports. 2010;20(2):122-32.
41.
Abraham G. Analysis of anthropometry, body composition and performance variables of young athletes
in southern region. Indian J Sci Technol. 2010;13(12):1210-3.
42.
Stoffels F, Kober RS, Pupo JD, Rocha Junior I, Mota CB. Analysis of kinematic variables of running in young
sprinters. Revista Portuguesa de Ciências do Desporto. 2007;7(1):59-67.
43. Thomas CE, Chambers TP, Main LC, Gastin PB. Motives for dropout among former junior elite Caribbean
track and field athletes: a qualitative investigation. Front Sports Act Living. 2021;3:1-13.
44.
Freitas JV, Werneck FZ, Souza RS, Castro PHC, Figueiredo AJ, Lima JRP. Maturation, morphological,
motor and technical characteristics of under 16 female track and field athletes. Rev Bras Cineantropom
Desempenho Hum. 2020;22:1-12.
45.
Boccia G, Moisè P, Franceschi A, Trova F, Panero D, La Torre A, et al. Career Performance trajectories in track
and field jumping events from youth to senior success: the importance of learning and development.
PLoS One. 2017;12(1):e0170744.
46.
Sedeaus A, Marc A, Marck A, Dor F, Schipman J, Dorsey M. BMI, a Performance Parameter for Speed
Improvement. PLoS One. 2014;9(2):1-7.
47.
Boccia G, Cardinale M, Brustio PR. World-Class Sprinters’ Careers: Early Success Does Not Guarantee
Success at Adult Age. Int J Sports Physiol Perform. 2020;16(3):367-74.
48.
Kearney PE, Hayes PR. Excelling at youth level in competitive track and field athletics is not a prerequisite
for later success. J Sports Sci. 2018;36(21):2502-9.
49. Tønnessen E, Svendsen IS, Olsen IC, Guttormsen A, Haugen T. Performance development in adolescent
track and field athletes according to age, sex and sport discipline. PloS One. 2015;10(6):e0129014.
50.
Gibbons T, McConnnell A, Foster T, Riewald ST, Peterson K. Reflections on success: U.S. Olympians
Describe the Success Factors and Obstacles that Most Influenced their Olympic Development. 2003.
51.
Werneck FZ, Coelho EF, Miranda L. PROJETO ATLETAS DE OURO®: uma inovação na detecção de talentos
esportivos. E-legis. 2022;15(n.º esp.):89-104.
52.
Miranda L, Werneck FZ, Coelho EF, Ferreira RM, Novaes JS, Figueiredo AJB, et al. Motor talent and
biological maturation in schoolchildren from a military college. Rev Bras Med Esporte. 2019;25(5):372-8.
53. Ribeiro Junior DB, Vianna, JM, Lauria AA, Coelho EF, Werneck FZ. Sports potential modeling of young
basketball players: a preliminary analysis. Rev Bras Cineantropom Desempenho Hum. 2019;21:1-12.
54. Breiman L. Statistical modeling: the two cultures. Stat Sci. 2001;16(3):199-215.
55. Régnier G, Salmela J, Russel S. Talent detection and development in sport. In: Singer RN, Murphey M,
Tennant LK, editors. Handbook of Research on Sport Psychology. Canada: MacMillan; 1993. p. 290-313.
