Cluster sets keep bar speed best — but in a 37-study meta-analysis, rest redistribution was close behind
In short
A Bayesian network meta-analysis of 37 studies found cluster sets preserved mean bar velocity better than traditional sets (SMD 0.60), with intra-set rest redistribution (0.41) and inter-set rest redistribution (0.42) also beating traditional sets. Cluster sets had the highest probability of ranking first, but head-to-head comparisons between the alternatives showed no clear difference, and neither exercise type nor relative intensity changed the result. Pick a structure based on what your gym and schedule allow, not on expected effect size.
Same load, same reps — and the bar still moves at different speeds depending on where you put the rest. A Bayesian network meta-analysis of 37 studies found cluster sets preserved mean velocity better than traditional sets by SMD 0.60. Rest-redistribution structures also beat traditional sets, at 0.41 to 0.42.
What is the difference between cluster sets and rest redistribution?
- Traditional sets — all reps back to back, rest only between sets.
- Cluster sets — the set is broken into small groups of reps with short added rests, such as a set of 5 done as 2+2+1.
- Intra-set rest redistribution — total rest stays the same; part of the long between-set rest is moved inside the set.
- Inter-set rest redistribution — total rest and total reps stay the same, but you do more, shorter sets, such as 8×3 instead of 4×6.
Which structure preserved velocity best?
On ranking probability (SUCRA), cluster sets led for both mean velocity (99.7%) and peak velocity (95.6%). Intra-set redistribution scored 46.9% and 65.1%; inter-set redistribution scored 46.3% and 26.9%. For peak velocity, only cluster sets (0.48) and intra-set redistribution (0.39) clearly beat traditional sets — the interval for inter-set redistribution crossed zero. Direct comparisons among the alternatives found no clear evidence of a meaningful difference.
Exercise type and relative intensity (percent of 1RM) did not moderate the effect, so the principle holds whether you are squatting or benching.
How should a lifter use this for squat, bench, and deadlift?
These are acute, within-session results. They do not show that cluster sets build a bigger 1RM over months. They do make cluster sets a useful tool for anyone whose technique breaks down late in heavy sets, or who is doing speed or technique work and wants every rep to look the same. When you log them, record total reps and total rest so the session stays comparable with your straight-set history. For when to stop a set based on speed loss, see 10% versus 20% velocity loss.
Most included studies were single-session experiments. Long-term hypertrophy and strength outcomes are outside what this meta-analysis can answer.
Frequently asked questions
What is a cluster set?
A cluster set breaks a set into small groups of reps with short rests in between. In a 37-study meta-analysis, cluster sets preserved mean bar velocity better than traditional sets by SMD 0.60.
How is rest redistribution different from cluster sets?
Rest redistribution keeps total rest the same and only changes where it goes — moving some between-set rest inside the set, or splitting the same total reps into more sets. Both versions preserved mean velocity better than traditional sets.
Do cluster sets build more strength?
This meta-analysis measured bar velocity within a session, so it does not answer whether cluster sets increase long-term strength. It does show they preserve speed and rep quality.
Do cluster sets work for every exercise and load?
Meta-regression found that exercise type and relative intensity did not change the size of the effect. Cluster sets preserved velocity better than traditional sets across lifts and loads.
Source: PubMed