

Cluster sets may be a valuable programming strategy in flywheel training, particularly when the goal is sustaining velocity and power output across sets. Emerging research, alongside Exerfly App data, suggests that short intra-set rest periods can help maintain more consistent concentric performance, which directly influences eccentric resistance. For coaches aiming to maximize output while managing fatigue, cluster configurations offer a practical and data-supported option within flywheel training.
Flywheel training is an effective and efficient method for increasing strength, power, and measures of athletic performance across different populations (3). Much like traditional resistance training, flywheel training is usually divided into sets of pre-defined numbers of repetitions with several minutes of rest in-between. But there is growing interest in cluster sets, which break up each set into clusters of reps with short “intra-set” rest periods in-between.
As an example, if a traditional flywheel training set involves 8 consecutive reps, a cluster set variation may break the 8 reps into 4-rep clusters with 30 seconds of rest between. While either set structure can be beneficial, there’s evidence that cluster sets have appealing features such as the ability to perform more reps at a given intensity, sustain higher intensities or outputs across a similar volume of work, and/or achieve similar adaptations with less fatigue (2,5,6).

There is a rapidly growing body of research to support the use of cluster sets in traditional training, however, only recently has it been explored as a potential method within flywheel training.
Like any new training method, there is usually a familiarization process when using flywheel training for the first time. There are some differences in the inertial loading offered by flywheel training compared to traditional methods, and this can take a few sessions to get used to before consistent and maximal outputs can be achieved (9).
Two recent studies by Ryan et al. (7,8) evaluated whether using cluster set structures can be an efficient method for familiarizing new users to flywheel training. The researchers also looked at whether providing instruction with external cues (“push the ground away”) or internal cues (“push through your heels”) influences the familiarization process.
The cluster set structure included 3 clusters of 5 reps and 45 seconds of intra-set rest for each set. The first two reps of each cluster were warm-up reps to build momentum into the flywheel followed by 3 high effort working reps. This was done with 4 different inertial loads, ranging from 0.025 to 0.10 kg.M² during each session.
Overall, they found that new users were able to become familiarized with flywheel training quarter-squats and RDLs within a couple of sessions using this cluster set approach, evidenced by consistent power outputs across reps and sets. This was particularly true when external focus cues were used.
Regardless of the reason, cluster sets appears to be useful when introducing new users to flywheel training, especially if proper technical instruction and cueing are provided as well.
The added intra-set rest during cluster sets can often help sustain higher velocities and power outputs across a set compared to traditional set structures (5). While less research has been done with cluster sets and flywheel training, the studies by Ryan et al. (7,8) found that a cluster set approach resulted in consistent power outputs across each cluster of reps in flywheel training squats and RDLs, despite the users being new to the training method.
This may be particularly useful when considering the user-defined resistance that flywheel training offers. The resistance provided during the eccentric phase is dictated by the effort the user exerts during the concentric phase. As fatigue sets in, the eccentric resistance will auto-regulate to the effort level of the user. This unique quality of flywheel training may be particularly useful when combined with cluster sets, as the added intra-set rest may minimize fatigue, allowing for greater efforts across the entire set and ultimately a higher quality stimulus!
I wanted to put this to the test, so I tracked some data (using the Exerfly app) across several training sessions. I used both traditional set and cluster set structures for multiple exercises in random order. The traditional set consisted of 2 warm-up reps to build momentum in the flywheel and 8 consecutive working reps with maximal concentric effort. The cluster set also had 8 working reps but divided into 4 rep clusters with 30-seconds of intra-set rest. Each cluster had 2 submaximal warm-up reps before the working reps.

Figure 2 displays the distribution of rep velocities measured with the Exerfly app across multiple sessions using either cluster sets or traditional sets. Over the course of several sessions, there were a few differences depending on the exercise. For the squat, I was consistently getting faster concentric velocities (which in turn provided a larger eccentric resistance) as shown by the cluster set density plot being further to the right. For the RDL and high-pull exercises, the biggest difference was that the velocities were more consistent with the cluster set, shown by the density plots being more condensed to a narrow range of velocities.
To dive deeper into what was happening with cluster sets vs traditional sets, I also looked at velocities across multiple sets of the same session rather than across all reps. An example of what I saw with the RDL can be observed in figure 3.

In this case, velocities were similar during set 1. The only small exception was a particularly slow rep in the traditional set, which occurred towards the end of the set. However, larger differences emerged during subsequent sets (following 2 minutes of rest). Velocity tended to drop off to a larger degree when using traditional sets than cluster sets. This suggests I may have needed longer rest between sets when using traditional set schemes with the Exerfly RDL and/or that cluster sets was less fatiguing and allowed for higher outputs across the sets.
It's important to note that this was my personal experience across a few individual sessions, but it does align with some of the research data on the advantages of cluster sets.
While there are many set structures that can be effective with flywheel training, cluster sets may be a viable option to consider. Particularly when the goal is to help familiarize new users or when attempting to maximize outputs across multiple sets and sessions. This does not mean cluster sets should be used in all circumstances, but rather that it offers an additional tool in the toolbox when programming flywheel training.
Get a monthly update with new research, blogs, and exclusive offers.