Journal of Strength and Conditioning Research

Reliability and concurrent validity of seven commercially available devices for the
assessment of movement velocity at different intensities during the bench press
--Manuscript Draft--

Manuscript Number: JSCR-08-11315R2

Full Title: Reliability and concurrent validity of seven commercially available devices for the
assessment of movement velocity at different intensities during the bench press

Short Title: Feasibility of seven devices to measure velocity

Article Type: Original Research

Keywords: linear position transducer; linear velocity transducer; smartphone application, inertial
measurement units; Velocity-based training; testing

Corresponding Author: Amador García Ramos, PhD

Faculty of Sport Sciences, University of Granada.
SPAIN

Corresponding Author Secondary

Information:

Corresponding Author's Institution: Faculty of Sport Sciences, University of Granada.

Corresponding Author's Secondary

Institution:

First Author: Alejandro Pérez-Castilla

First Author Secondary Information:

Order of Authors: Alejandro Pérez-Castilla

Antonio Piepoli

Gabriel Delgado-García

Gabriel Garrido-Blanca

Amador García Ramos, PhD

Order of Authors Secondary Information:

Manuscript Region of Origin: SPAIN

Abstract: This aim of this study was to compare the reliability and validity of seven commercially
available devices to measure movement velocity during the bench press exercise.
Fourteen men completed two testing sessions. The bench press one-repetition
maximum (1RM) was determined in the first session. The second testing session
consisted of performing three repetitions against five loads (45-55-65-75-85% of 1RM).
The mean velocity was simultaneously measured using an optical motion sensing
system (Trio-OptiTrack™; “gold-standard”) and seven commercially available devices:
1 linear velocity transducer (T-Force™), 2 linear position transducers (Chronojump™
and Speed4Lift™), 1 camera-based optoelectronic system (Velowin™), 1 smartphone
application (PowerLift™), and 2 inertial measurement units (PUSH™ band and
Beast™ sensor). The devices were ranked from the most to the least reliable as
follows: (I) Speed4Lift™ (coefficient of variation [CV] = 2.61%), (II) Velowin™ (CV =
3.99%), PowerLift™ (3.97%), Trio-OptiTrack™ (CV = 4.04%), T-Force™ (CV = 4.35%),
Chronojump™ (CV = 4.53%), (III) PUSH™ band (CV = 9.34%), and (IV) Beast™
sensor (CV = 35.0%). A practically perfect association between the Trio-OptiTrack™
system and the different devices was observed (Pearson’s product-moment correlation
coefficient (r) range = 0.947-0.995; P < 0.001) with the only exception of the Beast
sensor (r = 0.765; P < 0.001). These results suggest that linear velocity/position
transducers, camera-based optoelectronic systems and the smartphone application
could be used to obtain accurate velocity measurements for restricted linear
movements, while the inertial measurement units used in this study were less reliable
and valid.

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation
Title Page (Showing Author Information)

Reliability and concurrent validity of seven commercially available devices for the

assessment of movement velocity at different intensities during the bench press

Running Title: Feasibility of seven devices to measure velocity

Authors: Alejandro Pérez-Castilla,1 Antonio Piepoli,2 Gabriel Delgado-García,1 Gabriel

Garrido-Blanca,2 and Amador García-Ramos1,3

Institutional Affiliations:
1
Department of Physical Education and Sport, Faculty of Sport Sciences, University of

Granada, Granada, Spain.

2
Department of Health Sciences, Faculty of Health Sciences, University of Jaen, Jaen, Spain.
3
Department of Sports Sciences and Physical Conditioning, Faculty of Education, CIEDE,

Catholic University of the Most Holy Concepción, Concepción, Chile.

Corresponding author:

Amador García-Ramos. Department of Physical Education and Sport, Faculty of Sport

Sciences, University of Granada, Spain. Department of Sports Sciences and Physical

Conditioning, Faculty of Education, CIEDE, Catholic University of Most Holy Concepción,

Chile. Tel: (+34) 677815348, Fax: (+34) 958244369, E-mail: amagr@ugr.es

Acknowledgments:

The authors thank all the participants who selflessly participated in the study. This study was

supported by the Spanish Ministry of Education, Culture and Sport under a pre-doctoral grant
(FPU15/03649) awarded to APC and by the University of Granada under a postdoctoral grant

(perfeccionamiento de doctores) awarded to AGR.

Manuscript ( NO AUTHOR INFORMATION - Manuscript Text
Pages, including References and Figure Legends)

Feasibility of seven devices to measure velocity

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27 Reliability and concurrent validity of seven commercially available devices for the
28
29 assessment of movement velocity at different intensities during the bench press
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
1
65
 Feasibility of seven devices to measure velocity

ABSTRACT
1
2 This aim of this study was to compare the reliability and validity of seven commercially
3
4
5 available devices to measure movement velocity during the bench press exercise. Fourteen
6
7 men completed two testing sessions. The bench press one-repetition maximum (1RM) was
8
9
10 determined in the first session. The second testing session consisted of performing three
11
12 repetitions against five loads (45-55-65-75-85% of 1RM). The mean velocity was
13
14
15 simultaneously measured using an optical motion sensing system (Trio-OptiTrack™; “gold-
16
17 standard”) and seven commercially available devices: 1 linear velocity transducer (T-
18
19
Force™), 2 linear position transducers (Chronojump™ and Speed4Lift™), 1 camera-based
20
21
22 optoelectronic system (Velowin™), 1 smartphone application (PowerLift™), and 2 inertial
23
24 measurement units (PUSH™ band and Beast™ sensor). The devices were ranked from the
25
26
27 most to the least reliable as follows: (I) Speed4Lift™ (coefficient of variation [CV] =
28
29 2.61%), (II) Velowin™ (CV = 3.99%), PowerLift™ (3.97%), Trio-OptiTrack™ (CV =
30
31
32 4.04%), T-Force™ (CV = 4.35%), Chronojump™ (CV = 4.53%), (III) PUSH™ band (CV =
33
34 9.34%), and (IV) Beast™ sensor (CV = 35.0%). A practically perfect association between the
35
36
37
Trio-OptiTrack™ system and the different devices was observed (Pearson’s product-moment
38
39 correlation coefficient (r) range = 0.947-0.995; P < 0.001) with the only exception of the
40
41 Beast sensor (r = 0.765; P < 0.001). These results suggest that linear velocity/position
42
43
44 transducers, camera-based optoelectronic systems and the smartphone application could be
45
46 used to obtain accurate velocity measurements for restricted linear movements, while the
47
48
49 inertial measurement units used in this study were less reliable and valid.
50
51
52
53
54 Keywords: linear position transducer; linear velocity transducer; smartphone application,
55
56 inertial measurement units; velocity-based training; testing.
57
58
59
60
61
62
63
64
2
65
 Feasibility of seven devices to measure velocity

INTRODUCTION
1
2 Velocity-based resistance training has gained in popularity over recent years due to the
3
4
5 proliferation of different commercially available devices (e.g., linear position transducers,
6
7 inertial measurement units [IMUs], smartphone applications, etc.) that are supposed to
8
9
10 accurately measure movement velocity (3,5). It has been proposed that the monitoring of
11
12 barbell velocity could be an appropriate alternative to prescribe the training load as compared
13
14
15 to the traditional approach that requires the determination of the one-repetition maximum
16
17 (1RM) (16,32). The use of movement velocity to prescribe the training load is justified by the
18
19
strong and linear relationship that has been reported for multiple exercises between
20
21
22 movement velocity and the %1RM (12,26,31). In this regard, instead of determining the 1RM
23
24 through a single maximal lift or by a set of repetitions to failure, the load can be prescribed to
25
26
27 match the desired velocity (16,34). Despite the encouraging applications of velocity-based
28
29 resistance training (20), little research is available comparing the reliability and validity of
30
31
32 different commercially-available devices used in training and research to monitor movement
33
34 velocity.
35
36
37
38
39 From a scientific standpoint, the three-dimensional (3D) motion capture has been
40
41 recognized as the “gold-standard” instrument to measure movement velocity (22,36).
42
43
44 However, since this technology is not practical or affordable for strength and conditioning
45
46 professionals, others devices are typically used in practice when implementing the velocity-
47
48
49 based resistance training approach. The linear position transducer has been the most used
50
51 device in scientific research (2,5,9,13). The linear position transducer consists of an
52
53
54 isoinertial dynamometer with a cable that is typically attached to the barbell and it derives
55
56 velocity from the recorded displacement-time data using the inverse dynamic approach (17).
57
58
More recently, a linear velocity transducer named “T-force™” (T-Force™ system, Ergotech,
59
60
61
62
63
64
3
65
 Feasibility of seven devices to measure velocity

Murcia, Spain) has been made commercially available, which directly provides velocity
1
2 measurements through the recording of electrical signals that are proportional to the cable’s
3
4
5 extension velocity (33). It is reasonable to speculate that the linear velocity transducer could
6
7 be more precise than linear position transducers since it is known that the successive
8
9
10 manipulation of raw data increases measurement errors (25,30). However, it remains
11
12 unexplored whether the reliability of velocity outputs significantly differs between linear
13
14
15 position and linear velocity transducers, as well as their concurrent validity with respect to
16
17 the “gold standard” 3D motion capture.
18
19
20
21
22 It should be acknowledged that linear position/velocity transducers are not always
23
24 practical or affordable. The need to attach the cable to the barbell restricts exercise selection
25
26
27 to the ones predominantly performed in a vertical direction (5). Another common drawback
28
29 of linear position/velocity transducers is their high price ( 2,000 US dollars), which may
30
31
32 limit their use to laboratory-based or professional sport settings (6,23,24). However, it should
33
34 be noted that a new linear position transducer named “Speed4Lift™” (Speed4Lift™, Madrid,
35
36
37 Spain) have appeared on the market with a considerably lower price (340 US dollars),
38
39 although there are no available data regarding its reliability and validity. As an alternative to
40
41
42
linear position/velocity transducers, wearable technologies are increasingly gaining
43
44 popularity in the field of strength training and conditioning (3,5,11,28).
45
46
47
48
49 One of the wearable devices that have recently appeared on the market is named
50
51 “Velowin™” (Velowin™, DeporTeC, Murcia, Spain). Velowin™ is a camera-based
52
53
54 optoelectronic system designed to measure movement velocity by the tracking of an infrared
55
56 reflective marker placed in the barbell. A high reliability and concurrent validity of the
57
58
59 Velowin™ to measure movement velocity has been reported during the free-weight back
60
61
62
63
64
4
65
 Feasibility of seven devices to measure velocity

squat exercise (11,21). The main advantage of Velowin™ as compared to linear

1
2 position/velocity transducers is that it does not require to be attached to the barbell through a
3
4
5 cable and, therefore, this would eliminate the risk of cable rupture (21). However, the
6
7 Velowin™ cost ( 625 US dollars) and limited portability (e.g., a PC software is needed)
8
9
10 could limit its use for many strength and conditioning professionals. Many practitioners can
11
12 only afford more practical devices such as smartphone applications or IMUs (2,28).
13
14
15
16
17 A smartphone application named “PowerLift™” has been designed to monitor
18
19
20 movement velocity by the manual inspection of a slow motion video recording by the
21
22 smartphone high-speed camera (4). The high reliability and validity of PowerLift™ to
23
24
25
monitor mean velocity has been confirmed in exercises such as the bench press, full-squat
26
27 and hip-thrust (3,4). However, the main limitation of PowerLift™ is that it does not provide
28
29 real-time velocity feedback since coaches are required to indicate manually the start and end
30
31
32 of the concentric phase. The PUSH™ band (PUSH™ band, PUSH Inc., Toronto, Canada)
33
34 and Beast™ sensor (Beast™ sensor, Beast Technologies Srl., Brescia, Italy), which are
35
36
37 composed by the combination of 3-axis accelerometers and 3-axis gyroscopes, are two of the
38
39 IMUs most commonly used in research and practice (3,5). An advantage of IMUs is that they
40
41
42
are able to account for the anteroposterior displacement that is frequent during free-weight
43
44 exercises (28), while linear position/velocity transducer cannot distinguish the direction of
45
46 the cable displacement. However, due to the discrepancies found between the studies that
47
48
49 have evaluated the validity of the PUSH™ band (3,5,36), as well as due to the scarce number
50
51 of studies that have examined the reliability and validity of the Beast™ sensor (3), more
52
53
54 research is needed to explore the feasibility of both IMUs devices.
55
56
57
58
59
60
61
62
63
64
5
65
 Feasibility of seven devices to measure velocity

To address the existing gaps in the literature, the present study was designed to
1
2 provide a comprehensive analysis of different devices (i.e., linear velocity transducer, linear
3
4
5 position transducers, camera-based optoelectronic system, smartphone application and IMUs)
6
7 that are being used in practice for the measurement of movement velocity during resistance
8
9
10 training. Specifically, the objective of this study was to compare the reliability and validity of
11
12 of seven commercially available devices to measure movement velocity during the bench
13
14
15 press exercise. We hypothesized that the devices would be ranked from the most to the least
16
17 reliable and valid as follows: (I) linear velocity transducer; (II) linear positions transducers;
18
19
(III) camera-based optoelectronic device, (IV) smartphone application and (V) IMUs. The
20
21
22 results of this study should provide practical information for strength and conditioning
23
24 coaches regarding the reliability and concurrent validity of different devices that can be used
25
26
27 in practice for the assessment of movement velocity.
28
29
30
31
32 METHODS
33
34 Experimental approach to the problem
35
36
37
This study was designed to explore the reliability and concurrent validity of seven
38
39 commercially available devices for the measurement of movement velocity. Subjects
40
41 completed two testing sessions separated by 48-72 hours. The 1RM in the bench press
42
43
44 exercise was determined in the first testing session. The second testing session consisted of
45
46 performing three repetitions against five different loads (45%, 55%, 65%, 75%, and 85% of
47
48
49 1RM). The mean velocity of the barbell was measured using an optical motion sensing
50
51 system (V120:Trio, OptiTrack™, NaturalPoint Inc., USA) that was considered the gold-
52
53
54 standard in the present study (27,38). In addition, the mean velocity was also measured by
55
56 seven commercially available devices: 1 linear velocity transducer (T-Force™ system,
57
58
Ergotech, Murcia, Spain), 2 linear position transducers (Chronojump™ Boscosystem,
59
60
61
62
63
64
6
65
 Feasibility of seven devices to measure velocity

Barcelona, Spain; and Speed4Lift™, Madrid, Spain), 1 camera-based optoelectronic system

1
2 (Velowin™, DeporTeC, Murcia, Spain), 1 smartphone application (PowerLift™) and 2 IMUs
3
4
5 (PUSH™ band, PUSH Inc., Toronto, Canada; and Beast™ sensor, Beast Technologies Srl.,
6
7 Brescia, Italy). The two repetitions with higher mean velocity recorded by the Trio-
8
9
10 OptiTrack™ at each load were used for calculating intra-session reliability (3,21), while only
11
12 the repetition with the highest mean velocity recorded at each load by the Trio-OptiTrack™
13
14
15 was used for validity analyses.
16
17
18
19
Subjects
20
21
22 Fourteen physically active men (age: 22.9 ± 1.6 years; body height: 1.76 ± 0.06 m; body
23
24 mass: 76.9 ± 7.8 kg; bench press 1RM: 86.1 ± 11.9 kg) volunteered to participate in this
25
26
27 study. Subjects were recruited from a fitness centre and all of them were familiarized with the
28
29 bench press exercise prior to the beginning of the study. None of them suffered from physical
30
31
32 limitations, health problems or musculoskeletal injuries that could compromise tested
33
34 performance. Subjects were instructed to avoid any strenuous exercise two days before each
35
36
37
testing session. They were informed of the study procedures and signed a written informed
38
39 consent form prior to initiating the study. The study protocol adhered to the tenets of the
40
41 Declaration of Helsinki and was approved by the Institutional Review Board.
42
43
44
45
46 Testing procedures
47
48
49 The first testing session was used for anthropometric measures as well as to determine the
50
51 1RM during the concentric-only bench press exercise following an incremental loading test
52
53
54 (10). The standardized warm-up consisted of jogging, self-selected dynamic stretching and
55
56 joint mobilization exercises, and one set of five repetitions performed against external load of
57
58
17 kg (mass of the unloaded Smith machine barbell) during the bench press exercise.
59
60
61
62
63
64
7
65
 Feasibility of seven devices to measure velocity

Thereafter, the external load was incremented from 10 to 1 kg until the 1RM load was
1
2 reached. The average number of loads tested was 8.9 ± 1.3. The inter-set rest was set to 4 min
3
4
5 and 1-2 repetitions were performed with each load.
6
7
8
9
10 The second testing session began with the same warm-up described for session 1.
11
12 Afterwards, subjects performed the concentric-only bench press exercise against five relative
13
14
15 loads (45%, 55%, 65%, 75%, and 85% of 1RM) that were implemented in an incremental
16
17 order. Lower loads (i.e., < 45% of 1RM) were not tested because they are generally not used
18
19
in training with the bench press exercise, while higher loads (i.e., > 85% of 1RM) were
20
21
22 excluded to avoid high fatigue that could compromise reliability analyses. Three repetitions
23
24 were executed with each load. Inter-repetition rest was set to 15 s and inter-set rest was fixed
25
26
27 to 4 min. The barbell was held by the safety stops of the Smith machine during the recovery
28
29 periods. Subjects were encouraged to lift the barbell at the maximum possible velocity.
30
31
32
33
34 The standard five-point body contact position technique (head, upper back, and
35
36
37
buttocks firmly on the bench with both feet flat on the floor) was followed in the two testing
38
39 sessions. Subjects self-selected the grip width, which was measured and kept constant on
40
41 every lift. The barbell was held by the safety stops of the Smith machine 1-2 cm above
42
43
44 subjects' chest at the level of the sternum. From that position, subjects were instructed to
45
46 perform a concentric-only movement as fast as possible until their elbows reached full
47
48
49 extension. Two spotters were responsible for lowering the barbell after each repetition.
50
51
52
53
54 Measurement equipment and data acquisition
55
56
57
58
59
60
61
62
63
64
8
65
 Feasibility of seven devices to measure velocity

The measurement devices were not synchronized and they separately collected the mean
1
2 velocity of the same repetitions (Figure 1). The specific characteristics of each device are
3
4
5 provided below:
6
7
8
9
10 **Figure 1 near here**
11
12
13
14
15 -Trio-OptiTrack™ (V120:Trio, OptiTrack™, NaturalPoint Inc., USA): an optical
16
17 motion sensing system which included three infrared and precalibrated cameras fixed on a
18
19
rectangular frame that give 3D position data of a reflective marker at a sampling rate of 120
20
21
22 Hz. Raw data of marker position in space were acquired using the software Motive v.1.5.0
23
24 (OptiTrack, NaturalPoint Inc., USA) and then analysed in Microsoft Excel (Microsoft,
25
26
27 Seattle, WA, USA). Instantaneous velocity was calculated by the differentiation of the
28
29 displacement data with respect to time. The reflective marker was placed on the left side of
30
31
32 the barbell and the Trio-OptiTrack™ was positioned at a distance of 2.5 m from the marker.
33
34 -T-Force™ (T-Force™ system, Ergotech, Murcia, Spain): an isoinertial dynamometer
35
36
37
which consist of a cable-extension linear velocity transducer interfaced with a personal
38
39 computer by means of a 14-bit resolution analog-to-digital data acquisition board.
40
41 Instantaneous velocity was automatically calculated at a sampling rate of 1,000 Hz by the
42
43
44 custom software v.2.28. The cable was vertically attached to the right side of the barbell
45
46 using a Velcro strap.
47
48
49 -Chronojump™ (Chronojump™ Boscosystem, Barcelona, Spain): an isoinertial
50
51 dynamometer which consist of a cable-extension linear position transducer attached to the
52
53
54 barbell interfaced with a personal computer at a sampling rate of 1,000 Hz. Raw data were
55
56 exported from custom software v.1.6.2 and then analysed in Microsoft Excel (Microsoft,
57
58
Seattle, WA, USA). Instantaneous velocity was calculated by the differentiation of the
59
60
61
62
63
64
9
65
 Feasibility of seven devices to measure velocity

displacement data with respect to time. The cable was vertically attached to the right side of
1
2 the barbell using a Velcro strap.
3
4
5 -Speed4Lift™ (Speed4Lift™, Madrid, Spain): an isoinertial dynamometer which
6
7 consist of a cable-extension linear position transducer attached to the barbell. Data were
8
9
10 directly recorded by the differentiation of the displacement data with respect to time at a
11
12 sampling rate of 1,000 Hz via wi-FI connection with an Android smartphone using
13
14
15 Speed4Lift application v.4.1. The cable was vertically attached to the left side of the barbell
16
17 using a Velcro strap.
18
19
-Velowin™ (Velowin™, DeporTeC, Murcia, Spain): an optoelectronic system which
20
21
22 included an infrared camera interfaced with a personal computer that measured displacement
23
24 of a reflector fixed to the barbell. Data were directly recorded from the custom software
25
26
27 v.1.6.314 by the differentiation of the displacement data with respect to time at a sampling
28
29 rate of 500 Hz. The Velowin was placed at a distance of 1.7 m from the infrared reflector and
30
31
32 it was calibrated according to the manufacturer’s instructions.
33
34 -PowerLift™: a smartphone v.6.0.1 application which involves the frame-by-frame
35
36
37
manual inspection of a slow motion video recording by the smartphone high-speed camera at
38
39 a frequency of 240 frames per second and a quality of 720 pixels. The mean velocity was
40
41 computed as the individual range of motion (i.e., vertical displacement of the barbell from the
42
43
44 initial [≈ 1 cm above the subject's chest] to the final [elbows at full extension] position)
45
46 divided by the lifting time (i.e., time between two frames selected by the user). The
47
48
49 smartphone (iPhone, Apple Inc., California, USA) was held by a researcher in portrait
50
51 position and recorded each lift from the front of the subject at approximately 1.5 m.
52
53
54 -PUSH™ band (PUSH™ band, PUSH Inc., Toronto, Canada): a wearable wireless
55
56 IMUs which consist of a 3-axis accelerometer and a gyroscope that provided 6 degrees of
57
58
freedom in its coordinate system (2). Data were directly recorded by the integration of the
59
60
61
62
63
64
10
65
 Feasibility of seven devices to measure velocity

acceleration data with respect to time at a sampling rate of 200 Hz via Bluetooth 4.0 LE
1
2 connection with a smartphone (iPhone, Apple Inc., California, USA) using PUSH™
3
4
5 application v.1.1.26. The PUSH™ band was worn on the subject’s dominant forearm
6
7 immediately inferior to the elbow crease with the main button located proximally (2,5).
8
9
10 -Beast™ sensor (Beast™ sensor, Beast Technologies Srl., Brescia, Italy): a wearable
11
12 wireless IMUs which included a 3-axis accelerometer, gyroscope and magnetometer. Data
13
14
15 were directly recorded by the integration of the vertical acceleration with respect to time at a
16
17 sampling rate of 50 Hz via Bluetooth 4.0 LE connection with a smartphone (iPhone, Apple
18
19
Inc., California, USA) using Beast™ application v.2.3.7. The Beast™ sensor was placed on
20
21
22 the barbell using a built-in magnet (3).
23
24
25
26
27 Statistical analyses
28
29 Descriptive data are presented as means and standard deviations, while the coefficient of
30
31
32 variation (CV) of the five loads is presented through their median value. Reliability was
33
34 assessed for each individual load by the CV and the intraclass correlation coefficient (ICC
35
36
37
model 3,1). Acceptable reliability was determined as a CV < 10% and a ICC > 0.70 (7). The
38
39 median CV value of the five loads was calculated to compare the reliability between the eight
40
41 devices examined in the present study. To interpret the magnitude of differences observed
42
43
44 between two CVs, a criterion for the smallest important ratio was established as higher than
45
46 1.15 (37). Bland-Altman plots were constructed to explore the concurrent validity of the
47
48
49 seven commercially available devices respect to the Trio-OptiTrack™ system. Since we
50
51 observed proportional bias in six out of seven comparisons (r2 > 0.1) (1), the data were log-
52
53
54 transformed before calculating the Pearson’s product-moment correlation coefficients (r)
55
56 (18). The criteria to interpret the strength of the r coefficients was as follows: trivial (< 0.1),
57
58
small (0.1-0.3), moderate (0.3-0.5), high (0.5-0.7), very high (0.7-0.9), or practically perfect
59
60
61
62
63
64
11
65
 Feasibility of seven devices to measure velocity

(> 0.9) (18). Statistical significance was accepted at P < 0.05 level and confidence limits
1
2 were set at 95%. All reliability assessments were performed by means of a custom
3
4
5 spreadsheet (19), while other statistical analyses were performed using the software package
6
7 SPSS (IBM SPSS version 22.0, Chicago, IL, USA).
8
9
10
11
12 RESULTS
13
14
15 Mean velocity values only reached an acceptable reliability at all loads for the Trio-
16
17 OptiTrack™ system, the linear velocity/position transducers and the smartphone application
18
19
(CV < 6.24% and ICC > 0.73) (Table 1). Based on the comparison of the median CVs of the
20
21
22 5 loads, the devices were ranked from the most to the least reliable as follows: (I)
23
24 Speed4Lift™ (CV = 2.61%), (II) Velowin™ (CV = 3.99%), PowerLift™ (3.97%), Trio-
25
26
27 OptiTrack™ (CV = 4.04%), T-Force™ (CV = 4.35%), Chronojump™ (CV = 4.53%), (III)
28
29 PUSH™ band (CV = 9.34%), and (IV) Beast™ sensor (CV = 35.0%).
30
31
32
33
34 ****Table 1 near here****
35
36
37
38
39 Bland-Altman plots revealed low systematic bias and random errors (≤ 0.05 m·s-1) for
40
41 the T-Force™, Chronojump™, Speed4Lift™, Velowin™, and PowerLift™ as compared to
42
43
44 the Trio-OptiTrack™ system (Figure 2). Both IMUs showed larger random errors (PUSH™
45
46 band = 0.06 ms-1 and Beast™ sensor = 0.21 ms-1). Heteroscedasticity of the errors was
47
48
49 observed for all devices with the only exception of the Speed4Lift™ (r2 = 0.007). A
50
51 practically perfect association between the Trio-OptiTrack™ system and the different devices
52
53
54 was observed (r range = 0.947-0.995; P < 0.001) with the only exception of the Beast™
55
56 sensor (r = 0.765; P < 0.001) (Figure 3).
57
58
59
60
61
62
63
64
12
65
 Feasibility of seven devices to measure velocity

****Figure 2 near here****

1
2 ****Figure 3 near here****
3
4
5
6
7 DISCUSSION
8
9
10 This study compared the reliability and concurrent validity of seven commercially available
11
12 devices for the measurement of movement velocity during the bench press exercise
13
14
15 performed in Smith machine. The different devices were ranked from the most to the least
16
17 reliable as follows: (I) Speed4Lift™, (II) Velowin™, PowerLift™, T-Force™,
18
19
Chronojump™, (III) PUSH™ band, and (IV) Beast™ sensor. The concurrent validity of the
20
21
22 T-Force™, Chronojump™, Speed4Lift™, Velowin™ and PowerLift™ with respect to the
23
24 Trio-OptiTrack™ system was practically perfect. The two IMUs, especially the Beast™
25
26
27 sensor, showed the lowest concurrent validity (i.e., lower r coefficients and larger random
28
29 errors) as compared to the Trio-OptiTrack™ system. The Speed4Lift™ was the only device
30
31
32 that did not report heteroscedasticity of the errors. The results of the present study speak in
33
34 favour of the Speed4Lift™ as the most reliable and valid device for the measurement of
35
36
37
movement velocity during the bench press exercise performed in a Smith machine.
38
39 Collectively, the results of this study suggest that all devices, with the exception of the IMUs
40
41 (especially the Beast™ sensor), could be used to obtain accurate velocity measurements for
42
43
44 restricted linear movements. However, strength and conditioning professionals should be
45
46 aware of the presence of heteroscedasticity between these devices, which could limit the
47
48
49 interchangeable use of different devices.
50
51
52
53
54 Linear position/velocity transducers have been routinely used for training and testing
55
56 purposes (5,15,29). These devices have been considered the “gold-standard” to measure
57
58
barbell velocity in many studies (2,3,5). Specifically, the linear velocity transducer used in
59
60
61
62
63
64
13
65
 Feasibility of seven devices to measure velocity

the present study (T-Force™) has been frequently used to validate other devices that were
1
2 designed to measure movement velocity (2,13,14). The preferential use of linear velocity
3
4
5 transducers as compared to linear position transducers could be justified because the direct
6
7 measurement of movement velocity is expected to provide more accurate velocity outputs
8
9
10 (30). However, there is scarce information regarding the comparison of the reliability of
11
12 velocity outputs between linear velocity and linear position transducers. Contrary to our
13
14
15 hypothesis, the linear velocity transducer (i.e., T-Force™) was not more reliable than the two
16
17 linear position transducers used in the present study (i.e., Chronojump™ and Speed4Lift™).
18
19
In addition, to the best our knowledge, this is the first scientific article that has provided data
20
21
22 regarding the validity of the Chronojump™ and Speed4Lift™. It should be noted that the
23
24 Speed4Lift™ was the most reliable device and the only device that did not report
25
26
27 heteroscedasticity of errors with respect to the Trio-OptiTrack ™ system. These results
28
29 together with its lower price ( 340 US dollars) and excellent portability (the software is
30
31
32 installed in a smartphone app that is wireless connected with the hardware) place the
33
34 Speed4Lift as an accurate, cost-effective and practical device for the measurement of
35
36
37 movement velocity. It is worth noting that the heteroscedasticity of errors observed for the
38
39 other devices compromises their interchangeability. The T-force™ and PowerLift™ showed
40
41
42
progressively larger values than the Trio-OptiTrack™ with increasing velocities, while
43
44 opposite results were observed for the Chronojump™, Velowin™, PUSH™ band, and Beast
45
46 Sensor™.
47
48
49
50
51 The velowin™ has been recently developed as a more affordable and practical device
52
53
54 to measure movement velocity during resistance training exercises. However, to date, only
55
56 two studies have examined its reliability and concurrent validity to measure movement
57
58
59 velocity. García-Ramos et al. (11) found a comparable reliability (CV = 4.29-4.60%) and
60
61
62
63
64
14
65
 Feasibility of seven devices to measure velocity

high validity (r = 0.97-0.98) between the T-Force™ and Velowin™ systems during the free-
1
2 weigh back squat exercise. Laza-Cagigas el al. (21) also observed a high validity (concordant
3
4
5 correlation coefficient = 0.96) and reliability (CV = 7.3% and ICC = 0.97) of the Velowin™
6
7 system to measure mean velocity during the free-weigh back squat exercise with respect to a
8
9
10 3D system. In line with these findings, our results showed a high and comparable reliability
11
12 and validity of the Velowin™ system as compared to the data provided by linear
13
14
15 position/velocity transducers during the bench press exercise performed in a Smith machine.
16
17 More recently, an affordable ( 11 US dollars) smartphone application named PowerLift™
18
19
20 has been made commercially available to measure movement velocity. The results of this
21
22 study corroborate the findings of Balsalobre-Fernández et al. who showed a comparable
23
24
25
reliability (ICC = 0.93-0.99) and high validity (r = 0.94-0.98) of PowerLift™ with respect to
26
27 the data provided by a linear position transducer (SmartCoach Europe, Stockholm, Sweden)
28
29 during free-weight exercises (3,4). Although it should be acknowledged that it is possible to
30
31
32 obtain accurate measurements of mean velocity with PowerLift™, the necessity of
33
34 individually select the start and end points of each repetition may be unpractical when many
35
36
37 athletes need to be assessed. In addition, the no provision of real-time velocity feedback
38
39 should also be considered as a limitation of compared to the other devices analysed in the
40
41
42
present study (39).
43
44
45
46 Wearable technology such as the PUSH™ band and Beast™ sensor are being
47
48
49 increasingly used in the strength and conditioning field. These two IMUs have been recently
50
51 validated to measure movement velocity during a variety of resistance training exercises
52
53
54 (2,3,5,36). In line with our results, the PUSH™ band provided highly valid measurements of
55
56 mean velocity when the data of several loads were combined for the analysis during the
57
58
59 bench press performed in a Smith machine as well as during the free-weight shoulder press
60
61
62
63
64
15
65
 Feasibility of seven devices to measure velocity

and biceps arm curl exercises (r > 0.86) (2,36). However, the validity of the PUSH band
1
2 seems to be compromised when the data of individual loads are analysed separately (specially
3
4
5 for loads  80% of 1RM) (5). The PUSH™ band device should be therefore considered with
6
7 some caution given the controversial results and the lower reliability reported in the present
8
9
10 study. Regarding the Beast™ sensor, to the best of our knowledge, only one study has
11
12 investigated the reliability and validity of this device with respect to the data provided by a
13
14
15 linear position transducer (SmartCoach Europe, Stockholm, Sweden), reporting both a very
16
17 high reliability (ICC > 0.95) and validity (r > 0.98) during the bench press, full-squat and
18
19
20 hip-thrust exercises (3). On the contrary, our results suggest that the Beast™ sensor is the
21
22 least reliable and valid device among all the commercially available devices analysed in the
23
24
25
present study. It is plausible that the lower sampling frequencies of the IMUs or the need to
26
27 integrate acceleration-time data to obtain velocity values could have caused their lower
28
29 reliability (5,28). Therefore, more evidence about the feasibility of wearable technology in
30
31
32 the field of velocity-based resistance training is needed.
33
34
35
36
37 Several limitations and directions for future research should be considered. First, it
38
39 should be noted that the mean velocity is not the only velocity variable used in practice. The
40
41
42
mean propulsive velocity (i.e., average velocity from the start of the concentric phase until
43
44 the acceleration of the bar is lower than gravity) and maximum velocity (i.e., maximum
45
46 instantaneous velocity value reached during the concentric phase) have been commonly
47
48
49 recommended for training and testing (8,35). Therefore, future studies should consider
50
51 expanding the analysis of the reliability and validity of different devices to these variables. It
52
53
54 should be noted that the mean velocity was the only variable analysed in the present study
55
56 since it was the only common variable for the eight devices analysed. Second, the PUSH™
57
58
59 band was placed on the subject’s forearm, while the other devices were attached to the
60
61
62
63
64
16
65
 Feasibility of seven devices to measure velocity

barbell. Therefore, while the measurement of the PUSH™ could be affected by

1
2 anteroposterior movements, the displacement of the other devices was restricted to the
3
4
5 vertical direction. Another factor that could have promoted the lower reliability and validity
6
7 of the IMUs is their lower sampling frequency. In this regards, it is plausible that the
8
9
10 reliability and validity of the PUSH™ could be improved with the current 2.0 version which
11
12 has a higher sampling frequency (1,000 Hz) and it can be directly attached to the barbell.
13
14
15 Future studies should examine whether our findings could be applicable to free-weight
16
17 exercises in which the displacement of the barbell is not restricted to the vertical direction.
18
19
Finally, since one of the main applications of the use of velocity during resistance training is
20
21
22 the prediction of the 1RM (12,26,31), future studies should examine the precision of different
23
24 commercially available devices for predicting the 1RM during basic resistance training
25
26
27 exercises.
28
29
30
31
32 PRACTICAL APPLICATIONS
33
34 The devices were ranked from the most to the least reliable as follows: (I) Speed4Lift™, (II)
35
36
37
Velowin™, PowerLift™, T-Force™, Chronojump™, (III) PUSH™ band, and (IV) Beast™
38
39 sensor. All devices presented a high concurrent validity with respect to the Trio-OptiTrack™
40
41 system with the only exception of the Beast™ sensor, while the Speed4Lift™ was the only
42
43
44 device that did not report heteroscedasticity of the errors. Taken together, these results
45
46 suggest that the Speed4Lift™ is the most appropriate device for the measurement of
47
48
49 movement velocity during the bench press exercise performed in a Smith machine. Note that
50
51 linear velocity transducers, linear position transducers, camera-based optoelectronic systems
52
53
54 and smartphone application could all be used to obtain accurate measurements of mean
55
56 velocity during the bench press exercise performed in a Smith machine, although the
57
58
presence of heteroscedasticity of errors should be in mind. The two IMUs present a lower
59
60
61
62
63
64
17
65
 Feasibility of seven devices to measure velocity

reliability and validity and, consequently, more caution should be taken when using PUSH™
1
2 band and Beast™ sensor devices for implementing velocity-based resistance training
3
4
5 programs.
6
7
8
9
10 REFERENCES
11
12 1. Atkinson, G and Nevill, AM. Statistical methods for assessing measurement error
13
14
15 (reliability) in variables relevant to sports medicine. Sport Med 26: 217–238, 1998.
16
17 2. Balsalobre-Fernández, C, Kuzdub, M, Poveda-Ortiz, P, and Campo-Vecino, JD. Validity
18
19
and reliability of the push wearable device to measure movement velocity during the
20
21
22 back squat exercise. J Strength Cond Res 30: 1968–1974, 2016.
23
24 3. Balsalobre-Fernandez, C, Marchante, D, Baz-Valle, E, Alonso-Molero, I, Jimenez, SL, and
25
26
27 Munoz-Lopez, M. Analysis of wearable and smartphone-based technologies for the
28
29 measurement of barbell velocity in different resistance training exercises. Front
30
31
32 Physiol 8: 649, 2017.
33
34 4. Balsalobre-Fernández, C, Marchante, D, Muñoz-López, M, and Jiménez, SL. Validity and
35
36
37
reliability of a novel iPhone app for the measurement of barbell velocity and 1RM on
38
39 the bench-press exercise. J Sports Sci 36: 64–70, 2018.
40
41 5. Banyard, HG, Nosaka, K, Sato, K, and Haff, GG. Validity of various methods for
42
43
44 determining velocity, force and power in the back squat. Int J Sports Physiol Perform
45
46 12: 1170–1176, 2017.
47
48
49 6. Conceicao, F, Fernandes, J, Lewis, M, Gonzalez-Badillo, JJ, and Jimenez-Reyes, P.
50
51 Movement velocity as a measure of exercise intensity in three lower limb exercises. J
52
53
54 Sports Sci 34: 1099–1106, 2016.
55
56 7. Cormack, SJ, Newton, RU, McGuigan, MR, and Doyle, TLA. Reliability of measures
57
58
obtained during single and repeated countermovement jumps. Int J Sports Physiol
59
60
61
62
63
64
18
65
 Feasibility of seven devices to measure velocity

Perform 3: 131–144, 2008.

1
2 8. García-Ramos, A, Haff, G, Padial, P, and Feriche, B. Reliability of power and velocity
3
4
5 variables collected during the traditional and ballistic bench press exercise. Sport
6
7 Biomech 17: 117–130, 2018.
8
9
10 9. García-Ramos, A, Jaric, S, Pérez-Castilla, A, Padial, P, and Feriche, B. Reliability and
11
12 magnitude of mechanical variables assessed from unconstrained and constrained
13
14
15 loaded countermovement jumps. Sport Biomech 16: 414–526, 2017.
16
17 10. García-Ramos, A, Padial, P, Haff, GG, Argüelles-Cienfuegos, J, García-Ramos, M,
18
19
Conde-Pipó, J, and Ferriche, B. Effect of different interrepetition rest periods on
20
21
22 barbell velocity loss during the ballistic bench press exercise. J Strength Cond Res 29:
23
24 2388–2396, 2015.
25
26
27 11. García-Ramos, A, Pérez-Castilla, A, and Martín, F. Reliability and concurrent validity of
28
29 the Velowin optoelectronic system to measure movement velocity during the free-
30
31
32 weight back squat. Int J Sports Sci Coach 13: 737–742, 2018.
33
34 12. García-Ramos, A, Pestaña-Melero, F, Pérez-Castilla, A, Rojas, F, and Haff, G.
35
36
37
Differences in the load-velocity profile between 4 bench press variants. Int J Sport
38
39 Physiol Perform 13: 326–331, 2018.
40
41 13. Garnacho-Castano, MV, Lopez-Lastra, S, and Mate-Munoz, JL. Reliability and validity
42
43
44 assessment of a linear position transducer. J Sports Sci Med 14: 128–136, 2015.
45
46 14. Gomez-Piriz, PT, Sanchez, ET, Manrique, DC, and Gonzalez, EP. Reliability and
47
48
49 comparability of the accelerometer and the linear position measuring device in
50
51 resistance training. J Strength Cond Res 27: 1664–1670, 2013.
52
53
54 15. González-Badillo, JJ, Rodríguez-Rosell, D, Sánchez-Medina, L, Gorostiaga, EM, and
55
56 Pareja-Blanco, F. Maximal intended velocity training induces greater gains in bench
57
58
press performance than deliberately slower half-velocity training. Eur J Sport Sci 14:
59
60
61
62
63
64
19
65
 Feasibility of seven devices to measure velocity

772–781, 2014.
1
2 16. González-Badillo, JJ and Sánchez-Medina, L. Movement velocity as a measure of
3
4
5 loading intensity in resistance training. Int J Sports Med 31: 347–352, 2010.
6
7 17. Harris, NK, Cronin, J, Taylor, KL, Boris, J, and Sheppard, J. Understanding Position
8
9
10 Transducer Technology for Strength and Conditioning Practitioners. Strength Cond J
11
12 32: 66–79, 2010.
13
14
15 18. Hopkins, WG, Marshall, SW, Batterham, AM, and Hanin, J. Progressive statistics for
16
17 studies in sports medicine and exercise science. Med Sci Sports Exerc 41: 3–13, 2009.
18
19
19. Hopkins, W. Calculations for reliability (Excel spreedsheet). A new view of statistics,
20
21
22 2000. Available from: http://www.sportsci.org/resource/stats/relycalc.html#excel.
23
24 21. Bryan, JM, Patrick A., I, Stephen P., S, Mann, J, Ivey, P, and Sayers, S. Velocity-Based
25
26
27 Training in Football. Strength Cond J 37: 52–57, 2015.
28
29 21. Laza-Cagigas, R, Goss-Sampson, M, Larumbe-Zabala, E, Termkolli, L, and Naclerio, F.
30
31
32 Validity and reliability of a novel optoelectronic device to measure movement
33
34 velocity, force and power during the back squat exercise. J Sports Sci, 2018. Epub
35
36
37
ahead of print.
38
39 22. Lorenzetti, S, Lamparter, T, and Luthy, F. Validity and reliability of simple measurement
40
41 device to assess the velocity of the barbell during squats. BMC Res Notes 10: 707,
42
43
44 2017.
45
46 23. Loturco, I, Kobal, R, Moraes, JE, Kitamura, K, Cal Abad, CC, Pereira, LA, and
47
48
49 Nakamura, FY. Predicting the maximum dynamic strength in bench press: the high
50
51 precision of the bar velocity approach. J Strength Cond Res 31: 1127–1131, 2017.
52
53
54 24. Loturco, I, Pereira, LA, Winckler, C, Santos, WL, Kobal, R, and McGuigan, M. Load-
55
56 velocity relationship in national paralympic powerlifters: a case study. Int J Sports
57
58
Physiol Perform, 2018. Epub ahead of print.
59
60
61
62
63
64
20
65
 Feasibility of seven devices to measure velocity

25. McMaster, DT, Gill, N, Cronin, J, and McGuigan, M. A brief review of strength and
1
2 ballistic assessment methodologies in sport. Sports Med 44: 603–623, 2014.
3
4
5 26. Munoz-Lopez, M, Marchante, D, Cano-Ruiz, MA, Chicharro, JL, and Balsalobre-
6
7 Fernandez, C. Load-, force-, and power-velocity relationships in the prone pull-up
8
9
10 exercise. Int J Sports Physiol Perform 12: 1249–1255, 2017.
11
12 27. Nagymáté, G and Kiss, RM. Application of OptiTrack motion capture systems in human
13
14
15 movement analysis. A systematic literature review. Recent Innov Mechatronics 5,
16
17 2018.
18
19
28. Orange, ST, Metcalfe, JW, Liefeith, A, Marshall, P, Madden, LA, Fewster, CR, and
20
21
22 Vince, RV. Validity and reliability of a wearable inertial sensor to measure velocity
23
24 and power in the back squat and bench press. J Strength Cond Res, 2018. Epub ahead
25
26
27 of print.
28
29 29. Pareja-Blanco, F, Rodríguez-Rosell, D, Sánchez-Medina, L, Sanchis-Moysi, J, Dorado,
30
31
32 C, Mora-Custodio, R, et al. Effects of velocity loss during resistance training on
33
34 athletic performance, strength gains and muscle adaptations. Scand J Med Sci Sports
35
36
37
27: 724–735, 2017.
38
39 30. Pérez-Castilla, A, Feriche, B, Jaric, S, Padial, P, and García-Ramos, A. Validity of a
40
41 linear velocity transducer for testing maximum vertical jumps. J Appl Biomech 33:
42
43
44 388–392, 2017.
45
46 31. Pérez-Castilla, A, García-Ramos, A, Padial, P, Morales-Artacho, A, and Feriche, B.
47
48
49 Load-velocity relationship in variations of the half-squat exercise: Influence of
50
51 execution technique. J Strength Cond Res, 2017. Apub ahead of print.
52
53
54 32. Sakamoto, A, Sinclair, PJ, and Naito, H. Strategies for maximizing power and strength
55
56 gains in isoinertial resistance training: Implications for competitive athletes. J Sports
57
58
Med Phys Fitness 5: 153–166, 2016.
59
60
61
62
63
64
21
65
 Feasibility of seven devices to measure velocity

33. Sánchez-Medina, L and González-Badillo, JJ. Velocity loss as an indicator of

1
2 neuromuscular fatigue during resistance training. Med Sci Sports Exerc 43: 1725–
3
4
5 1734, 2011.
6
7 34. Sánchez-Medina, L, González-Badillo, JJ, Pérez, CE, and Pallarés, JG. Velocity- and
8
9
10 power-load relationships of the bench pull vs bench press exercises. Int J Sports Med
11
12 35: 209–216, 2014.
13
14
15 35. Sanchez-Medina, L, Perez, CE, and Gonzalez-Badillo, JJ. Importance of the propulsive
16
17 phase in strength assessment. Int J Sports Med 31: 123–129, 2010.
18
19
36. Sato, K, Beckham, GK, Carroll, K, Bazyler, C, Sha, Z, and Haff, GG. Validity of wireless
20
21
22 device measuring velocity of resistance exercises. J Trainology 4: 15–18, 2015.
23
24 37. Stewart, AM and Hopkins, WG. Consistency of swimming performance within and
25
26
27 between competitions. Med Sci Sports Exerc 32: 997–1001, 2000.
28
29 38. Vasconcelos, F, Peebles, D, Ourselin, S, and Stoyanov, D. Spatial calibration of a 2D/3D
30
31
32 ultrasound using a tracked needle. Int J Comput Assist Radiol Surg 11: 1091–1099,
33
34 2016.
35
36
37
39. Weakley, JJ, Wilson, KM, Till, K, Read, DB, Darrall-Jones, J, Roe, G, Phibbs, PJ, and
38
39 Jones B. Visual feedback attenuates mean concentric barbell velocity loss, and
40
41 improves motivation, competitiveness, and perceived workload in male adolescent
42
43
44 athletes. J Strength Cond Res, 2017. Epub ahead of print.
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
22
65
 Feasibility of seven devices to measure velocity

FIGURE LEGEND
1
2 Figure 1. Distribution of the measurement devices during the testing protocol. (1) Trio-
3
4
5 OptiTrack™, (2) T-Force™, (3) Chronojump™, (4) Speed4Lift™, (5) Velowin™, (6)
6
7 PowerLift™, (7) PUSH™ band, and (8) Beast™ sensor.
8
9
10
11
12 Figure 2. Bland-Altman plots for the measurement of mean velocity between the Trio-
13
14
15 OptiTrack™ system and the seven commercially available devices: T-Force™ (upper-left
16
17 panel), Chronojump™ (upper-right panel), Speed4Lift™ (upper middle-left panel),
18
19
Velowin™ (upper middle-right panel), PowerLift™ (lower middle-right panel), PUSH™
20
21
22 band (lower middle-left panel), and Beast™ sensor (lower panel). Each plot depicts the
23
24 averaged difference and 95% limits of agreement (dashed lines), along with the regression
25
26
27 line (solid line).
28
29
30
31
32 Figure 3. Relationship of mean velocity values between the Trio-OptiTrack™ system and the
33
34 seven commercially available devices: T-Force™ (upper-left panel), Chronojump™ (upper-
35
36
37
right panel), Speed4Lift™ (upper middle-left panel), Velowin™ (upper middle-right panel),
38
39 PowerLift™ (lower middle-right panel), PUSH™ band (lower middle-left panel), and
40
41 Beast™ sensor (lower panel). The Pearson’s correlation coefficient (r) was calculated using
42
43
44 the log-transformation since the assumption of homoscedasticity was violated.
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
23
65
Table

Table 1. Reliability of mean velocity values obtained from the Trio-OptiTrack™ method and

seven commercially available devices at different loads during the bench press exercise.

Load Mean velocity CV ICC

Device
(%1RM) (ms-1) (95% CI) (95% CI)
45 0.84 ± 0.05 3.47 (2.52, 5.59) 0.73 (0.35, 0.90)
55 0.71 ± 0.05 2.22 (1.57, 3.76) 0.93 (0.79, 0.98)
Trio-OptiTrack™ 65 0.58 ± 0.05 4.04 (2.93, 6.50) 0.84 (0.57, 0.95)
75 0.46 ± 0.05 4.15 (2.98, 6.85) 0.83 (0.54, 0.95)
85 0.35 ± 0.05 4.64 (3.37, 7.48) 0.88 (0.67, 0.96)
45 0.83 ± 0.06 2.48 (1.78, 4.09) 0.90 (0.70, 0.97)
55 0.70 ± 0.05 1.82 (1.32, 2.93) 0.95 (0.84, 0.98)
T-Force™ 65 0.59 ± 0.05 4.35 (3.15, 7.01) 0.78 (0.45, 0.93)
75 0.49 ± 0.05 4.78 (3.43, 7.89) 0.77 (0.40, 0.92)
85 0.37 ± 0.05 4.90 (3.55, 7.90) 0.87 (0.64, 0.96)
45 0.90 ± 0.05 2.31 (1.67, 3.72) 0.87 (0.64, 0.96)
55 0.76 ± 0.05 2.09 (1.51, 3.36) 0.90 (0.71, 0.97)
Chronojump™ 65 0.60 ± 0.07 6.24 (4.47, 10.3) 0.72 (0.31, 0.90)
75 0.47 ± 0.06 4.53 (3.25, 7.48) 0.85 (0.58, 0.95)
85 0.34 ± 0.05 5.65 (4.05, 9.32) 0.86 (0.60, 0.95)
45 0.88 ± 0.06 2.61 (1.80, 4.77) 0.87 (0.55, 0.96)
55 0.75 ± 0.04 2.39 (1.73, 3.85) 0.84 (0.57, 0.94)
Speed4Lift™ 65 0.63 ±0.05 2.42 (1.69, 4.25) 0.93 (0.78, 0.98)
75 0.51 ± 0.05 3.92 (2.81, 6.47) 0.81 (0.49, 0.94)
85 0.38 ± 0.05 3.41 (2.38, 5.98) 0.94 (0.78, 0.98)
45 0.79 ± 0.04 2.85 (2.02, 4.83) 0.84 (0.55, 0.95)
55 0.70 ± 0.06 3.97 (2.81. 6.74) 0.85 (0.57, 0.96)
PowerLift™ 65 0.58 ± 0.05 4.91 (3.48, 8.33) 0.74 (0.32, 0.92)
75 0.51 ± 0.05 3.69 (2.58, 6.48) 0.87 (0.58, 0.96)
85 0.40 ± 0.04 4.97 (3.47, 8.71) 0.85 (0.54, 0.96)
45 0.91 ± 0.06 2.89 (2.09, 4.65) 0.83 (0.56, 0.94)
55 0.77 ± 0.05 3.27 (2.35, 5.40) 0.79 (0.45, 0.93)
Velowin™
65 0.64 ± 0.06 3.99 (2.86, 6.59) 0.83 (0.53, 0.94)
75 0.51 ± 0.05 6.01 (4.36, 9.69) 0.68 (0.26, 0.89)
85 0.38 ± 0.06 7.64 (5.54, 12.3) 0.69 (0.27, 0.89)
45 0.79 ± 0.07 5.02 (3.56, 8.52) 0.69 (0.22, 0.89)
55 0.63 ± 0.07 7.84 (5.39, 14.3) 0.46 (-0.27, 0.81)
PUSH Band™ 65 0.46 ± 0.08 9.34 (6.77, 15.0) 0.78 (0.45, 0.92)
75 0.31 ± 0.06 14.6 (10.0, 26.6) 0.50 (-0.21, 0.82)
85 0.24 ± 0.06 19.1 (13.7, 31.5) 0.47 (-0.09, 0.80)
45 0.82 ± 0.32 33.4 (23.5, 58.9) 0.29 (-0.42, 0.72)
55 0.70 ± 0.27 24.2 (15.6, 53.2) 0.64 (-0.28, 0.89)
Beast sensor™ 65 0.51 ± 0.21 35.0 (22.6, 77.1) 0.30 (-0.91, 0.77)
75 0.34 ± 0.16 40.2 (28.1, 70.5) 0.31 (-0.40, 0.73)
85 0.23 ± 0.15 54.9 (38.9, 93.2) 0.27 (-0.38, 0.70)

CV, coefficient of variation; ICC, intraclass correlation coefficient; 95% CI, 95% confidence

interval.
Figure 1_R1
(5)
(7)

(1)

(8)

(2)

(4) (6) (3)

Figure 2_R1

Differences (Trio-OptiTrack™ – Chronojump™)

Differences (Trio-OptiTrack™ – T-force™)

0.15 0.15
Systematic bias ± random error: -0.01 ± 0.03 ms-1 Systematic bias ± random error: -0.03 ± 0.03 ms-1
r² = 0.266 r² = 0.508
0.10 0.10

0.05 0.05

0.00 0.00

-0.05 -0.05

-0.10 -0.10

-0.15 -0.15
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
Mean Trio-OptiTrack™ and T-force™ Mean Trio-OptiTrack™ and Chronojump™
Differences (Trio-OptiTrack™ – Speed4Lift™)

Differences (Trio-OptiTrack™ – Velowin™)

0.15 0.15
Systematic bias ± random error: -0.04 ± 0.02 ms-1 Systematic bias ± random error: -0.05 ± 0.03 ms-1
r² = 0.007 r² = 0.247
0.10 0.10

0.05 0.05

0.00 0.00

-0.05 -0.05

-0.10 -0.10

-0.15 -0.15
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
Mean Trio-OptiTrack™ and Speed4Lift™ Mean Trio-OptiTrack™ and Velowin™
Differences (Trio-OptiTrack™ – PUSH band™)
Differences (Trio-OptiTrack™ – PowerLift™)

0.15 0.30
Systematic bias ± random error: -0.01 ± 0.05 ms-1
r² = 0.437
0.10 0.20

0.05 0.10

0.00 0.00

-0.05 -0.10

-0.10 -0.20 Systematic bias ± random error: 0.10 ± 0.06 ms-1

r² = 0.232

-0.15 -0.30
0.20 0.40 0.60 0.80 1.00 0.15 0.35 0.55 0.75 0.95
Mean Trio-OptiTrack™ and PowerLift™ Mean Trio-OptiTrack™ and PUSH™ band
Differences (Trio-OptiTrack™ – Beast sensor™)

0.75

0.50

0.25

0.00

-0.25

-0.50
Systematic bias ± random error: 0.05 ± 0.21 ms-1
r² = 0.426
-0.75
0.05 0.35 0.65 0.95 1.25
Mean Trio-OptiTrack™ and Beast™ sensor
Figure 3_R1

1.00 1.00
y = 1.069x - 0.051 y = 0.881x + 0.046
r = 0.995; P < 0.001 r = 0.991; P < 0.001
Trio-OptiTrack™ (ms-1)

Trio-OptiTrack™ (ms-1)
0.80 0.80

0.60 0.60

0.40 0.40

0.20 0.20
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
T-force™ (ms-1) Chronojump™ (ms-1)
1.00 1.00

y = 0.985x - 0.030 y = 0.912x + 0.001

r = 0.994; P < 0.001 Trio-OptiTrack™ (ms-1)
Trio-OptiTrack™ (ms-1)

0.80 0.80 r = 0.993; P < 0.001

0.60 0.60

0.40 0.40

0.20 0.20
0.20 0.40 0.60 0.80 1.00 0.20 0.40 0.60 0.80 1.00
Speed4Lift™ (ms-1) Velowin™ (ms-1)
1.00 1.00
y = 1.205x – 0.130 y = 0.822x + 0.187
r = 0.970; P < 0.001 r = 0.947; P < 0.001
Trio-OptiTrack™ (ms-1)

Trio-OptiTrack™ (ms-1)

0.80 0.80

0.60 0.60

0.40 0.40

0.20 0.20
0.20 0.40 0.60 0.80 1.00 0.00 0.20 0.40 0.60 0.80 1.00
PowerLift™ (ms-1) PUSH™ band (ms-1)
1.00

0.80
Trio-OptiTrack™ (ms-1)

0.60

0.40
y = 0.456x + 0.341
r = 0.765; P < 0.001
0.20

0.00
0.00 0.30 0.60 0.90 1.20 1.50
Beast™ sensor (ms-1)