Written by: Adam Tzur (primary author), Andrew Vigotsky, and Brandon Roberts.
Edited and reviewed by: Israel Halperin and Conrad Earnest
Acknowledgements: Greg Nuckols
Main article length: ~3000 words
Article length with tables: ~5500 words
Last updated: 04.05.2017

Summary

• We have reviewed the cluster set (CS), rest-pause (RP), and drop-set (DS) literature (including Schoenfeld’s new DS study!). You can find overviews of the studies in graphs, tables and study analyses.

• Cluster sets and rest-pause: CS/RP seems to be similar to traditional training for gaining strength and muscle mass. Cluster sets might be good for preventing form breakdown, reducing feelings of fatigue, and to build more volume with less discomfort. Long-term effects on power are unclear.

• Drop-sets: There are few DS studies and they show conflicting information for strength and hypertrophy. Practically speaking, DS help you finish your workouts quicker and can be a good tool in that regard.

• General conclusions: CS, DS, and RP can be good tools if programmed right.

• Main limitations: CS/RP programs in many of the studies may not have been challenging enough to create a stimulus for optimal gains. Furthermore, several studies selectively reported data and had other methodological issues. Many of the studies had small sample sizes on average, aka. few subjects (n = ~27 ± ~11).

Sign up for the newsletter and stay updated when new articles, videos, and e-books are published on SCI-FIT!

What are rest-pauses, cluster sets and drop-sets?

It’s interesting to see how the terms rest-pause and cluster set are used colloquially. If we go online to see how people define the terms, then we get contradictory opinions (1, 2, 3, 4). Though most people seem to agree that the rest-pause method involves training to failure, resting, and then doing some more reps. Cluster sets are about dividing volume into multiple clusters, aka sub-sets, and resting in-between them.

In the scientific literature, rest-pause has many different definitions. Sometimes, it can seem like rest-pause is very similar to cluster sets. Yet, rest-pause generally relies on training to failure (Tufano et al., 2017). Clusters and rest-pause are used in a lot of crossfit style workouts or competitions especially those that use AMRAPS for time.

It should be said that many researchers use the terms rest-pause and cluster set in different ways. The terminology isn’t properly standardized yet (Tufano et al., 2017). However, we’ve kept the definitions consistent in this article, even if it means redefining terms that are used in some of the studies.

Drop-sets (DS) involve training to failure, dropping the weight, and immediately continuing to do more reps to failure. How many sets you want to drop consecutively is up to you. The idea behind drop-sets is that they would lead to greater muscle mass gains because they cause a lot of local stress to the muscle (i.e. metabolite buildup). Schoenfeld suggests that drop-sets lead to extended metabolic stress which might be good for hypertrophy (Schoenfeld, 2011).

Check out the examples in the next section for details on how to do CS, RP, and DS.

Why would you want to do cluster sets, rest-pause, and drop-sets?

Cluster sets help you keep fatigue low. Some researchers suggest that by using cluster sets, you can focus on form while maintaining velocity, and power output (Nicholson et al., 2016c; Tufano et al., 2017).

Rest-pause and drop-sets are good ways of accumulating fatigue quickly. For example, if you are short on time in the gym you could drastically reduce training time and still get in the same amount of total training volume (Prestes et al., 2017). This may not be ideal if you want to work at a high intensity (%1RM), but for hypertrophy it could be enough.

There are other benefits as well, check out the practical applications for more details.

Examples of the different training methods

Traditional sets

Figure based on materials from Tufano et al., 2017

This is pretty straightforward. You do 4 repetitions and rest for 2 minutes between sets.

Cluster sets

Figure based on materials from Tufano et al., 2017

In the example above, you:

• Do 2 reps (sub-set #1 of the first cluster set)
• Rest for 15 seconds
• Do 2 reps (sub-set #2)
• Rest for 120 seconds (this separates cluster set 1 from cluster set 2)
• Do 2 reps (sub-set #1 of the second cluster set)
• Rest for 15 seconds
• Do 2 reps (sub-set #2)

You’ve now done a total of 2 cluster sets, 4 subsets, and 8 reps with the cluster set method. Remember that there is no rule for how many reps or clusters there should be in a cluster set, or how much rest you should use between clusters. Though, if you’re taking longer rests that last for minutes, you’re just doing regular sets. Reps are not generally done to failure and there is no specific cut off time for rests in the literature.

Rest-pause method

Figure based on materials from Tufano et al., 2017

In the first set from the example above, you:

• Do 8 reps to failure
• Rack the weight and rest for 30 seconds
• Do 2 reps to failure
• Rest for another 30s
• Do 1 rep

You’ve now done a total of 11 reps; 8 traditional reps + 3 rest-pause reps. Since you went to failure on the first sub-set, you hit your 8RM. RP allows you to do more reps after failure.

• [Then rest and do the next RP set]

Drop-sets

In the set from the example above, you:

• Do 12 reps to failure
• Drop to a lighter weight

• Do 8 reps to failure

• Drop to a lighter weight
• Do 5 reps to failure
• Drop to a lighter weight
• Do 3 reps to failure

You’ve now done a total of 28 reps; 1 traditional set + 3 drop-sets. Since you went to failure on the first set, you hit a 12RM. DS is a method that allows you to do more reps, at a lighter load, while still going to failure.

Inter-repetition rest sets

Figure based on materials from Tufano et al., 2017

With IRR sets, you take a short break in-between every repetition. This type of training is typically used to help focus on technique.

Graphs

Drop-sets: effects on strength

Drop-sets: effects on hypertrophy

Studies excluded from the graph, due to either major methodological issues and/or lacking outcome measurements:

• Giessing et al., 2014
• Fisher et al., 2016

You can find more information about the issues in the limitations section

Cluster sets and rest-pause: effects on strength

Studies excluded from the graph, due to major methodological issues:

• Rooney et al., 1994

• Giessing et al., 2016

You can find more information about the issues in the limitations section

CS and RP: effects on hypertrophy

CS and RP: effects on power

[themify_box]

Do you like this article? Check these out, too:

• Scientific recommendations for strength and hypertrophy training from 150+ studies (training frequency, training to failure, rest periods, isolation vs. compound exercise, exercise order, and ROM)

• The Science of Detraining: How long you can take a break from the gym before you lose muscle mass, strength, and endurance

[/themify_box]

Overall gains across studies

• In the drop-set studies, the weekly strength gains of the DS groups were between ~1.5% and ~3%.

• In the cluster-set studies, the CS groups gained somewhere between ~1.5% and ~4.5% per week.

• All of the traditional groups in both CS and DS studies gained between ~2% and ~5% strength per week (except for Lawton et al., 2014 which is an outlier).

• In terms of hypertrophy, CS and DS groups gained between ~0.2% and ~1.8% LBM or muscle CSA per week. However, this number is strongly influenced by Prestes et al., 2017 and their CSA measurement of the thigh musculature.

Interesting observations

Looking at this from a bird’s-eye-view, the traditional groups gained slightly more strength than the CS/DS groups. However, we should be careful about making definite conclusions because many of the studies had no statistically significant differences between groups even if the mean % changes were different. Also,  in studies with small groups, one participant can heavily influence the mean and few studies provided data about spread, outliers or individual data points.

An interesting observation is that the strength gains seemed to vary not only based on the training program, but also which exercise/muscle group was trained. For example, in Goto et al 2005, there was a large difference in gains made in knee extension (65.3% – Traditional; 38.1% – Cluster) compared to shoulder press(29.2% – Traditional; 21.2% – Cluster).

To add, in Prestes et al., 2017, the LBM gains of the traditional group were non-significantly larger than the CS group, but the thigh CSA changes were significantly larger for the CS group. This just goes to show how different hypertrophy measurements can give us different results. The thigh CSA change was the only significant hypertrophy change in that study. Strangely, the traditional group in the study had 12 kg less LBM than the CS group.

Should we train for power?

One of the first things we must ask ourselves when we do any type of training, is what is the purpose of the thing we’re doing? Why train for power? Is it to improve technique? Is it for a specific olympic lift? To maximize 1RM?

Fundamentally, we must question whether what we’re doing is an effective way to reach our goals.

Here’s what Nicholson et al., 2016b have to say about strength and power:

“(…) it would appear that the CL-1 regimen was not optimally designed for strength development since the ability to offset fatigue-induced reductions in repetition velocity and power did not translate to greater strength improvements. In some respects, this observation supports the specificity of training adaptations (Kaneko et al. 1983) since maximal strength assessments are not typically associated with higher velocities and power outputs. Instead, it seems logical that the slower repetitions in the STR and CL-2 regimens closely resembled the 1RM assessment and contributed to the larger strength improvements in these regimens”. (Nicholson et al., 2016b)

Is power important?

To understand if power is important, we must first understand the construct of mechanical power. Colloquially, the term power is often used to describe an action that is done forcefully, quickly, or explosively (Ayalon et al., 1974); for example, someone who can jump high or accelerate quickly is often described as “powerful”. However, mechanical power is different.

In mechanics, power, P, is the time-rate of doing work, W, (the rate at which work is done), and can be calculated as P = W/t. This formula can be rewritten as P = F⋅v, where F is force and v is velocity. Hence, power is equal to the component of force acting on a body that is in the same direction that it is moving times the velocity of that body.

This means that power, as used in research, is different from power as we use it in everyday speech. It’s important that we distinguish these two definitions so as to avoid confusion when talking about it.

When attempting to apply the purely mechanical definition of power, its usefulness becomes less clear than its colloquial meaning. Does someone need to apply a large force while traveling quickly in order to jump high? Does someone need to lift a heavy load quickly to be strong?

Fortunately, much has been written on misconceptions surrounding power (Knudson, 2009; Winter et al., 2015). For tasks like the vertical jump, impulse, rather than power, is what determines how high someone can jump (Ayalon et al., 1974). In fact, recent work has shown that jump height can improve without changing power output (Jiménez Reyes et al., 2017).

Of course, powerlifters lift loads slowly, and thus, ironically, display relatively low power output (Garhammer, 1993). That is not to say that power is totally useless, though. For sports like cycling, power is integral (Wilson, 2004). This is because drag (i.e., the force that resists one from moving forward) increases quadratically with the speed at which one travels, and thus, one must not only pedal faster, but also with increasing force; this is why gearing is so advantageous. In other words, tasks that involve high force output at high speeds require higher power.

Can we use power measurements that are taken during a training session to predict long-term power gains?

Most power research looks at short-term changes. This is is called acute research, and the studies we’ve reviewed usually last for less than a week. Sometimes the researchers train and test lifters for just one or two session. The issue with these studies is that they can’t predict long-term power changes. For example, a study might show that you can maintain better mean power output in one session when you use cluster sets. This doesn’t necessarily imply that the program they used will give you better power gains in the long-term.

Generally speaking, it’s hard to predict gains by looking at acute changes in, for example, hormones, mTOR, MPS, force, velocity, and power (1, 2, 3). However, this doesn’t mean that these variables are useless. It just means that we have to be careful about measuring something during or after a training session, and then concluding that the measurement will reflect long-term changes!

Therefore, we have to look at long-term power studies. Sadly, there aren’t many studies like that. If you want to look at the acute studies, you can find them listed here (spoiler alert: most acute studies find that cluster sets are better at maintaining power output compared to traditional sets). We haven’t included these studies in our analysis.

Our task here today is to inform you about how you can think of power. An in-depth discussion about power is a topic for another article, but like we’ve said, if you’re training for power then you should also justify why you are doing so, because it’s not given that more power translates to more strength, or better performance.

If you are interested in reading more about power, check out these articles

• Strength and Power: A Definition of Terms (Harman, 1993)

• ‘Correcting the use of the term “power” in the strength and conditioning literature (Knudson, 2009)

• Misuse of “Power” and Other Mechanical Terms in Sport and Exercise Science Research (Winter et al., 2015)

• Countermovement Jump Phase Characteristics of Senior and Academy Rugby League Players (McMahon et al., 2016)

SCR: “when athletes are making improvements from one level of sport to the next, jumping higher does not require them to increase the amount of force they exert in the jump. It requires them to produce a similar amount of force, but at a faster speed and over a longer distance.”

• Effectiveness of an Individualized Training Based on Force-Velocity Profiling during Jumping (Jiménez-Reyes et al., 2017)

• Relationships Among measurements of explosive strength and anaerobic power (Ayalon et al., 1974)

Practical applications for programming

 Should we use cluster sets and rest-pause?

Pros of cluster sets

• Cluster sets may be used to prevent from breakdown and maintain power and velocity (Hardee et al., 2012a; Nicholson et al., 2016c; Tufano et al., 2017)
• Cluster sets reduce feelings of fatigue, discomfort and perceived exertion. (Nicholson et al., 2016c; Tufano et al., 2017).
• This could allow you to use more volume without becoming fatigued during a workout (Tufano et al., 2017). However, the workout will be longer.

Cons of cluster sets

• Can be very time consuming, depending on how you program them (Nicholson et al., 2016c)

In terms of strength and hypertrophy, the evidence is very mixed for cluster sets and rest-pause. Some studies support CS/RP for strength and/or hypertrophy, and some studies do not. This is probably because studies used different exercises, programs, subjects, durations, and so on.

When programming cluster sets, it’s probably a good idea to add more volume or intensity (Nicholson et al., 2016b). If you just take your old sets and splice them in half with a rest in-between, your muscles will probably be understimulated. This was an issue in several of the studies.

Furthermore, it’s possible that cluster sets should only be something you temporarily add during a phase in your training program (Nicholson et al., 2016b).

Should we use drop-sets?

Pros of drop-sets

• Save time by making workouts shorte

Cons of drop-sets

• Possibly decrease strength/muscle endurance? (Fink et al., 2017)
• Fatiguing

There are few relevant studies on drop-sets and they have contradictory results for strength and hypertrophy. Hence, the jury is still out. However, drop-sets can be used strategically to increase training volume in a short amount of time. For example, some bodybuilders use drop-sets such as “running the racks” in order to accumulate volume or induce more fatigue. Yet, there is some debate over whether this technique accumulates “junk” volume or if it actually helps induce more hypertrophy.

There are currently no studies on this topic, but there is data showing that lifters can gain the same amount of muscle using high or low loads (Schoenfeld et al., 2016), so going to failure using a declining load during a drop-set likely induces hypertrophy. It could also be a good way to reduce the time you spend in the gym while still training to failure and getting enough volume.

If you want to get into the details of the DS studies, check out the DS table and single study analyses.

Table of CS and RP studies (long-term)

Legend

• Green background = this group was better than the other group for this variable (statistically significant between-group differences)

• Red background = this group did worse than the other group for this variable (statistically significant between-group differences)

• Grey background = result was statistically significant for group x time, but not between-group

• Purple background = no change or the result was not statistically significant (between-group or group x time)

• Empty cells = this variable was not measured in the study

Strength and muscle mass

Technique and power

Table of drop-set studies (long-term)

Acute cluster set studies

• Acute measurements are generally not predictive of long-term changes
• We’ve excluded acute studies from the article
• If you’re interested in the acute studies, you can check them out here

Limitations

Heterogeneity between studies

“Indeed, comparisons between studies are confounded by the use of single – and multi – set training regimens, contrasting progression strategies (i.e. periodised vs. non – periodised) and strength test protocols (i.e. 1RM vs. 6RM). In addition, it is difficult to compare studies that have used different load – repetition schemes, exercises (e.g. compound vs. ballistic), subjects (e.g. trained vs. untrained), training durations (5 – 12 weeks) and contrasting inter – set and intra-set rest intervals.” – Nicholson et al., 2016b

Ecological validity

• Some studies only used 1 set per session

• Studies examined pure CS training vs traditional training, which isn’t reflective of how CS would be used in real life. It’s quite possible that results would be different if CS only consisted of 10% of a training program.

• The CS programs were, in several of the studies, much easier than the traditional program. Real life approaches would probably use CS differently

Power testing

The test should be specific to the movement being trained in the program. Which raises questions why countermovement jump, squat jump, etc. are being used to test power gains:

“(…) for a valid assessment of strength gain from a resistance training regimen, the testing modality should closely resemble the training conditions (i.e. type of muscle action, etc.)”. – Nicholson et al., 2016b

Other issues

• Large variance in definitions for CS/RP

• A potential issue in several of the studies is that the CS/RP protocol was easier and less fatiguing. Less effort = less gains? In some of the studies, CS/RP programs required less effort compared to traditional training. Lifters in these groups may have been understimulated, hence less gains.

• Varying volume matching

• Poor reporting: Several of the studies did not report absolute values or standard deviations