Round 1

Reviewer 1 Report

This is an interesting manuscript, with practical applications, which is perfectly suitable for the Journal. The aim of the study was to compare five commonly used on-field systems for measuring (or calculating) running speed during the acceleration phase of sprinting. Subsequently, the horizontal force-velocity relationships were calculated and compared.

Concurrent validity results overall showed a relative systematic error of 0.86 to 2.28% for maximum and theoretically maximal velocity variables and 4.78 to 12.9% for early acceleration variables. The inter-trial reliability showed low coefficients of variation (< 5.74%), which were similar between all systems. It was concluded that all systems tested can be considered relevant to measure maximal velocity and to calculate force-velocity mechanical outputs.

Comments to be addressed:

Lines 109-110: Please confirm the sampling frequency of the Stalker Radar (46.875 Hz). Please explain in the manuscript how you treated this data (e.g. smoothing etc)

Lines 133-135: please include the value of the delay

Lines 135-137: Please specify what exactly the video analysis included

Fig. 1: In the linear encoder data, the brief acceleration and deceleration phases of each step are evident, while this was not the case in the laser velocity data. Please explain this discrepancy with a couple of sentences in the Discussion. Could it be due to movement of the string connected with the encoder? Please specify where exactly it was attached on the runner's body.

Fig. 1: The left part of Fig. 1 is too small. Please rescale it.

Please briefly explain in the Discussion how the positioning quality, the average horizontal dilution of precision and the number of satellites may influence the results obtained by the use of GPS to measure speed and acceleration (and therefore the F-V profile). This is important for the practitioners, as they need to understand the magnitude of possible errors.

Fig. 2: change "limits of agreements" to "limits of agreement"

Please discuss the possible effects of the 2.28+/-1.49% difference between timing gates and laser. What effects can this have on the assessment of sprinting. Similarly, please discuss the larger differences found for Tau and Fo between systems (5.5 to 12.0%). The explanation in lines 247-252 about the fitting method reduce the reliability of this approach. Please clarify.

Please clarify the sentence in lines 290-292 of the Conclusion. How exactly should practitioners "interpret their own data obtained...in light of these results"? Please be specific and elaborate on that using 2-3 sentences.

 

 

Author Response

Paris, 25/10/2022

 

Dear Associate Editor,

 

We thank the referee for their interest and constructive remarks regarding our manuscript “Sprint acceleration mechanical outputs derived from position- or velocity-time data: a multi-system comparison study” (sensors-1966520).

We have revised the manuscript according to his recommendations and your editorial suggestion that are in bold. Amended sentences are in italic with the additional wordings in red.

 

On behalf of all authors,

 

Jean Slawinski

 

 

Reviewer N°1

We thank you for reviewing our manuscript and for providing comments and suggestions that have helped us to improve the manuscript. We have considered your remarks and made amendments when necessary in the revised manuscript. We appreciate your further perusal of the revised manuscript.

We have provided our responses to your comments that are in bold. Amended sentences are in italic with the additional wordings in red.

Comments to the Author

This is an interesting manuscript, with practical applications, which is perfectly suitable for the Journal. The aim of the study was to compare five commonly used on-field systems for measuring (or calculating) running speed during the acceleration phase of sprinting. Subsequently, the horizontal force-velocity relationships were calculated and compared.

Concurrent validity results overall showed a relative systematic error of 0.86 to 2.28% for maximum and theoretically maximal velocity variables and 4.78 to 12.9% for early acceleration variables. The inter-trial reliability showed low coefficients of variation (< 5.74%), which were similar between all systems. It was concluded that all systems tested can be considered relevant to measure maximal velocity and to calculate force-velocity mechanical outputs.

We are grateful for your valuable comments. We’ve made corrections to improve manuscript’s readability.

 

Lines 109-110: Please confirm the sampling frequency of the Stalker Radar (46.875 Hz). Please explain in the manuscript how you treated this data (e.g. smoothing etc)

Yes, we confirm the sampling frequency of the Stalker Radar at 46.875 Hz.

Data were analysed as follows:

Lines 111-113: Raw data outliers were deleted. Then, the cleaned data were fitted using the exponential model proposed and validated by Samozino and colleagues [5] in order to compute sprint mechanical outputs.

 

Lines 133-135: please include the value of the delay

Thank you for this relevant comment. The value of the delay has been added as follow.

Lines 133-139: Due to this location, the time delay between the first propulsive action of the participant (i.e. determined when the thumb of the forward hand took off the ground) and the crossing of the timing gates was determined frame by frame by visual inspection from the video recordings using an iPhone 8 (iOS 13.7, Apple Inc., USA). Video were recorded at 240 fps from a left lateral view located at 5 m of the start line. The time delay was computed on 36 sprint trials and its value was 0.25 ± 0.06 s.

Lines 135-137: Please specify what exactly the video analysis included

Ok. Video analysis was used in order to compute the aforementioned delay.

 

Fig. 1: In the linear encoder data, the brief acceleration and deceleration phases of each step are evident, while this was not the case in the laser velocity data. Please explain this discrepancy with a couple of sentences in the Discussion.

Could it be due to movement of the string connected with the encoder?

Please specify where exactly it was attached on the runner's body.

Thanks for this comment. You are right, the linear encoder data was the raw data that explain the velocity variation of the runner while the laser velocity data considered was the smoothed data. However, it is difficult to conclude that the observed velocity oscillation in the linear encoder are completely due to movement of the string connected with the encoder and attached to a waist belt.

The following sentence have been modified in the methods part :

Lines 91-93: However, this extra 1-kg resistance did not affect the inter-system comparisons for running velocity and derived data, since all systems measured the same running movement with the runner connected to the encoder via a cable attached to a waist belt.

 

Fig. 1: The left part of Fig. 1 is too small. Please rescale it.

Ok, done.

 

Please briefly explain in the Discussion how the positioning quality, the average horizontal dilution of precision and the number of satellites may influence the results obtained by the use of GPS to measure speed and acceleration (and therefore the F-V profile). This is important for the practitioners, as they need to understand the magnitude of possible errors.

Thanks for this very interesting and important comment. We have added the following sentences to the discussion:

Lines 289-295: The number of satellites and the average horizontal dilution of precision are key indicators to ensure that the data collected can be analysed and interpreted with confidence. Indeed, poor quality data (e.g. less than 6 satellites connected to the devices and/or horizontal dilution of precision greater than 1) [23] could give erroneous values of speed and acceleration measures that can influence the interpretation of the sprint mechanical variables computed. Therefore, practitioners should systematically take into account these indicators to analyse the data with confidence.

 

Fig. 2: change "limits of agreements" to "limits of agreement"

Ok, done.

Please discuss the possible effects of the 2.28+/-1.49% difference between timing gates and laser.

It’s difficult to compare the tools betwwen them but the intra-comparison still valable.

What effects can this have on the assessment of sprinting. Similarly, please discuss the larger differences found for Tau and Fo between systems (5.5 to 12.0%).

Thanks for this relevant comment.

As the rate of velocity increase is higher in the early acceleration phase, this will exacerbate the difference found between the systems, compared to the lower rate of change in velocity by the end of the acceleration. These results confirms the interest of intra-system comparisons.

This explanation has been included in the discussion part as follows.

Lines 237-241: Since the rate of velocity increase is higher in the early acceleration phase, this likely exacerbates the difference found between systems, compared to the lower rate of change in velocity by the end of the acceleration. Contrary to maximum velocity variables, these results confirm the reliability of intra-system comparisons and suggest that comparisons of athlete’s values obtained from different systems should take these differences into account.

 

The explanation in lines 247-252 about the fitting method reduce the reliability of this approach. Please clarify.

Due to the exponentional rate of change of velocity is higher at the beginning of the sprint, this latter has a greater influence on the parameters computed by the model. This influence is even higher as the the sample rate of the system is low.

For more clarity, the following sentence has been added:

Lines 252-255: In line with the interpretation of good agreement for maximal velocity values, the lower agreement for acceleration variables could also be due to the very high rate of change in velocity during early acceleration. This influence is even higher when the sample rate of the system is lower.

 

Please clarify the sentence in lines 290-292 of the Conclusion. How exactly should practitioners "interpret their own data obtained...in light of these results"? Please be specific and elaborate on that using 2-3 sentences.

Ok, thanks for your comment. Please find below the following sentences that we propose:

Lines 303-308: Only intra-systems comparisons will allow the most accurate interpretation for inter-athletes or intra-athlete changes. Alternatively, inter-systems comparisons (for example between different research studies) must be interpreted cautiously due to the percent differences observed here and the different data sampling and processes. Some of our recent unpublished works show that higher speed thresholds to determine the sprint start (e.g. 1 m.s^-1) lead to higher inter-systems reliability.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

It would have been interesting to include MySprint App in the analysis

Author Response

Paris, 25/10/2022

 

Dear Associate Editor,

 

We thank the referee for their interest and constructive remarks regarding our manuscript “Sprint acceleration mechanical outputs derived from position- or velocity-time data: a multi-system comparison study” (sensors-1966520).

We have revised the manuscript according to his recommendations and your editorial suggestion that are in bold. Amended sentences are in italic with the additional wordings in red.

 

On behalf of all authors,

 

Jean Slawinski

 

Reviewer: 2

We thank you for reviewing our manuscript and for providing comments and suggestions that have helped us to improve the manuscript. We have considered your remarks and made amendments when necessary in the revised manuscript. We appreciate your further perusal of the revised manuscript.

We have provided our responses to your comments that are in bold. Amended sentences are in italic with the additional wordings in red.

 

Comments to the Author

It would have been interesting to include MySprint App in the analysis

Yes, application has been validated against cells and radar by Romero-Franco et al.

A lot of systems have been compared in the present study (linear encoder, laser, radar, GPS and timing gates) and adding this one was considered but not feasible. Moreover, as my sprint is based on the split times like the timing gates, we think that it was enough to have the timing gates .

Author Response File: Author Response.pdf