Training with Weightlifting Derivatives, Making Diamonds, and Fooling Yourself
📝 Weekly paper summary
Training with Weightlifting Derivatives: The Effects of Force and Velocity Overload Stimuli (Suchomel et al., 2020)
Timothy J Suchomel
Category
Randomized Trial
Context
Researchers and practitioners often use Olympic weightlifting movements, such as the clean and jerk or snatch, to promote speed-related training adaptations. Traditionally, these lifts include a "catch" phase, whereby after the initial propulsion of the barbell from the ground (the first and second pull), the trainee must then get back under the bar to complete the lift.
However, researchers and practitioners have questioned whether they should include the "catch" phase for non-weightlifting athletes, given the technical proficiency required to complete the whole lift. Instead, some have argued that weightlifting derivatives that do not require the catch phase (e.g., a clean pull) can further promote strength- and speed-related training adaptations. Researchers and practitioners propose that the improved training adaptations result from two main factors. First, trainees would not have to "cut their pull short" (i.e., change how they perform the first and second pull and reduce how much they propel the bar off the floor). Second, practitioners would have more flexibility to modulate load- and velocity-overloads using pulling variations. With pulling variations, athletes can use loads above their 1RM weightlifting movements or achieve faster movement velocities than the traditional Olympic lift.
Currently, there is limited longitudinal evidence comparing the training adaptations between programs that use whole Olympic weightlifting movements versus their pulling derivatives. Specifically, it's unclear whether any potential improvements in training adaptations when using weightlifting derivative is from using the derivative itself or due to what the derivative affords (i.e., the potential to use a more comprehensive array of training loads and velocities with pulls versus whole lifts). Therefore, the purpose of this study was to compare the training adaptations between three training programs: 1) uses traditional Olympic weightlifting movements (CATCH), 2) uses weightlifting pulls with loads matched to the traditional Olympic weightlifting movement program (PULL), and 3) uses weightlifting pulls with a broader array of loads and velocities than groups 1 and 2 (Overload group; OL).
Correctness
The study recruited male collegiate athletes ranging from 18 to 28 years old, a representative sample for most practitioners using these lifts. However, it would have been an additional strength if the researchers had included female athletes.
The researchers implemented the training program for ten weeks, which is sufficient to elicit minor to moderate adaptations in these athletes.
However, there are two main statistical issues to consider when interpreting these findings:
- The authors expressed their data as a percent change. Representing improvements as a within-person change (e.g., X% improvement from baseline) is problematic since regression to the mean, floor/ceiling effects, and measurement error can skew the data. More relevant to this analysis expressing the data this way likely inflated the authors' specified \(\alpha=0.05\) (false positive rate) to around \(0.75\). For more information, I'd strongly suggest reading Bland and Altman's 2011 paper on the topic.
- There seem to be inconsistencies between the p-values and the effect sizes reported in the study. For example, the authors noted that "10-m improvements were greater for the OL group compared with the CATCH group (p= 0.026, g=1.32), but not the PULL group (p=0.121, g=1.35)". However, in a study with a fixed sample size, the distribution of the p-value should be monotonically related to the effect size (i.e., larger effect sizes = smaller p-values). Since the sample size was fixed and equal between groups in this study, it's unclear why the larger effect size also yielded a larger p-value. There could be an explanation for this, but it's not mentioned within the paper and would require a deeper dive into the authors' data. Regardless, I would have preferred the authors tested for interaction effects in a linear model.
Contributions
- Given the statistical challenges outlined above regarding the p-values and effect sizes, I think it's best to report the general trends seen in the "raw" data provided by the authors. Mainly, the average improvements in CATCH and PULL were similar, but OL's improvements were generally greater. Integrating these findings with some previous work suggesting there is no difference in adaptations between CATCH and PULL, it seems the benefits of adopting pulling variations are not due to the variation itself, but what it can afford during the training program (i.e., a wider array of loads and movement velocities).
- Although, it would be great to have updated analyses with more rigorous statistics to verify this assertion.
🧠 Fun fact of the week
Most people know that we can "make" diamonds by subjecting coal to intense heat and pressure. However, did you know you could make diamonds out of peanut butter? Scientists Dan Frost and colleagues were interested in studying how crystalline structures are formed in the earth's lower mantle. To test one of their research questions, they conducted an experiment to replicate an effect that would pull CO2 from the atmosphere into the deeper layers of the earth, which leaves behind carbon that can form diamonds. To do this, the researchers subjected a carbon-rich material, peanut butter, to these high pressures and created diamonds. I think this is worthy of a Bill Nye gif:
🎙 Podcast recommendation
This podcast episode is for those interested in biomechanics and markerless motion capture systems!
🗣 Quote of the week
"The first principle is that you must not fool yourself and you are the easiest person to fool."
- Richard Feynman