Ingo Sandau, Helmi Chaabene, Urs Granacher

• This study examined the concurrent validity of an inverse dynamic (force computed from barbell acceleration [reference method]) and a work-energy (force computed from work at the barbell [alternative method]) approach to measure the mean vertical barbell force during the snatch using kinematic data from video analysis. For this purpose, the acceleration phase of the snatch was analyzed in thirty male medal winners of the 2018 weightlifting World Championships (age: 25.2±3.1 years; body mass: 88.9±28.6 kg). Vertical barbell kinematics were measured using a custom-made 2D real-time video analysis software. Agreement between the two computational approaches was assessed using Bland-Altman analysis, Deming regression, and Pearson product-moment correlation. Further, principal component analysis in conjunction with multiple linear regression was used to assess whether individual differences related to the two approaches are due to the waveforms of the acceleration time-series data. Results indicated no mean difference (p > 0.05; d = −0.04)This study examined the concurrent validity of an inverse dynamic (force computed from barbell acceleration [reference method]) and a work-energy (force computed from work at the barbell [alternative method]) approach to measure the mean vertical barbell force during the snatch using kinematic data from video analysis. For this purpose, the acceleration phase of the snatch was analyzed in thirty male medal winners of the 2018 weightlifting World Championships (age: 25.2±3.1 years; body mass: 88.9±28.6 kg). Vertical barbell kinematics were measured using a custom-made 2D real-time video analysis software. Agreement between the two computational approaches was assessed using Bland-Altman analysis, Deming regression, and Pearson product-moment correlation. Further, principal component analysis in conjunction with multiple linear regression was used to assess whether individual differences related to the two approaches are due to the waveforms of the acceleration time-series data. Results indicated no mean difference (p > 0.05; d = −0.04) and an extremely large correlation (r = 0.99) between the two approaches. Despite the high agreement, the total error of individual differences was 8.2% (163.0 N). The individual differences can be explained by a multiple linear regression model (R2adj = 0.86) on principal component scores from the principal component analysis of vertical barbell acceleration time-series waveforms. Findings from this study indicate that the individual errors of force measures can be associated with the inverse dynamic approach. This approach uses vertical barbell acceleration data from video analysis that is prone to error. Therefore, it is recommended to use the work-energy approach to compute mean vertical barbell force as this approach did not rely on vertical barbell acceleration.…