Stop Terminating Sets With Velocity Loss in Free-Weight Exercises
Only have a second? Check out the takeaway below. Have 5 minutes? Check out the rest of the newsletter.
TRAINING TAKEAWAY: While potentially useful as a descriptive training variable, intraset velocity loss lacks acceptable session to session agreement in the free-weight back squat, raising doubt about its practical utility as a set termination method.
Background
Terminating sets after a desired amount of intraset velocity loss has gained considerable attention in recent research. In a positive light, velocity loss offers an objective method to autoregulate proximity to failure and relative volume load (reps x sets x % of 1RM) by altering the number of repetitions performed within a given set relative to the maximum number of repetitions possible. Indeed, multiple studies have compared groups terminating sets at different velocity loss thresholds and provide evidence that velocity loss seems to modify training outcomes (1,2,3). Due to its objectivity, one could make the argument that velocity loss rectifies some of the limitations of subjective autoregulation methods like RIR-based RPE.
Despite the theoretical advantages of utilizing velocity loss to terminate sets, an important assumption to acknowledge is the stability of a given threshold to result in a similar percentage of possible repetitions performed. Studies carried out in exercises with minimal degrees of freedom (e.g., Smith-machine squat and Smith-machine bench press) have shown acceptable session to session reliability across multiple loads (1,2); however, these exercises are less commonly used by strength athletes alongside velocity based training. Thankfully, a recent study by Jukic and colleagues comprehensively investigates this concept in the free-weight back squat.
Study Overview
Fifty one resistance trained participants were recruited to participate in the current study with a back squat 1RM relative to body mass of 1.79 ± 0.35 and 1.25 ± 0.30 in males and females, respectively. Each participant completed 4 different experimental sessions:
Familiarization of study procedures
1RM test
Repetitions to failure
Repetitions to failure (repeat of visit 3)
The methods of visits 1 and 2 are self-explanatory, whereas visits 3 and 4 require more detail. Specifically, participants completed 3 sets to failure with 90%, 80%, and 70% of 1RM in a non-randomized order (i.e., descending load) with maximal concentric intended velocity. Throughout each of these sets, the velocity of every single repetition was collected. Following at least 72 hours of rest, an identical testing session was performed to evaluate how the relationship of repetitions performed and velocity loss changes session to session.
Results
The authors performed an extremely comprehensive analysis with many useful comparisons; however, I’m going to focus on findings I believe are the most practical. Specifically, the authors assessed the session to session agreement of the percentage of possible repetitions that were performed upon reaching a given velocity loss threshold using a Bland Altman analysis. This analysis estimates the overall mean difference between the observed values and constructs limits of agreement that are expected to contain the difference between measurements for 95% of pairs of future measurements. For example, this analysis estimates the difference in the percentage of possible repetitions performed at 80% of 1RM and a 30% velocity loss threshold between visit 3 and visit 4.
The authors also performed an equivalence test in which the measurements would be considered statistically equivalent if the confidence intervals of the limits of agreement (i.e., upper and lower dots and intervals at each load) lie entirely within the equivalence bounds. The equivalence bounds were set at a ±10% difference in the percentage of possible repetitions performed at a given velocity loss threshold.
As you can see in Figure 1, the confidence intervals of the limits of agreement fall considerably outside of the equivalence bounds in each of the loads investigated. In fact, the confidence intervals of the limits of agreement lie completely outside of the equivalence bounds of ±10%.
Ultimately, this suggests that a given velocity loss threshold did not result in practically acceptable agreement of the percentage of possible repetitions performed across training sessions.
Practical Applications
This means that terminating sets at the same velocity loss threshold (e.g., 30% velocity loss) with the same load (e.g., 80% of 1RM) can result in large fluctuations in the proximity to failure and relative volume load performed. This variability makes it difficult to integrate velocity loss thresholds into program design as the recovery cost for sets at a given velocity loss would likely fluctuate to a similar degree.
Moreover, velocity loss relationships are exercise, load, and set specific, making their construction cumbersome and potentially cost-prohibitive. Ultimately, velocity loss thresholds do seem to influence acute and chronic responses to training, but may be best used descriptively (and on the program level) rather than prescriptively (and on a set-by-set level). Instead, lifters should utilize other set termination methods (subjective RIR-based RPE, percentage based training, and absolute velocity thresholds) and can optionally monitor intraset velocity loss as a proxy of intraset fatigue.