Failure is defined as the inability to perform another full repetition of an exercise, due to your muscles being unable to produce the required force.
How close an individual gets to failure with their repetitions is probably one of the more hotly debated training variables.
Some propose training to failure is a must for muscle hypertrophy, while others may suggest leaving a certain number of repetitions in reserve can still sufficiently optimize muscle hypertrophy.
For those unaware, “repetitions in reserve” simply refers to how many further repetitions you likely could have been completed.
As an example, if you perform repetitions with a load you could perform a maximum of 12 repetitions with, performing 13 would result in failure, performing 12 would leave 0 repetitions in reserve, performing 11 would leave 1 repetition in reserve, performing 10 would leave 2 repetitions in reserve, and performing 9 would leave 3 repetitions in reserve (and so on and so forth).
In my opinion, probably one of the best ways to figure out how training to failure may compare to leaving repetitions in reserve is to assess the current relevant research.
Fortunately, there are a fair number of studies in this area. In this article, I’ll do my best to summarize this current research.
Let’s get into it.
Table of Contents
Training to Failure vs Stopping With a Few Reps in Reserve
4 well-designed papers have compared training to failure to stopping with a few repetitions in reserve.
Let’s briefly overview them.
A 2018 Brazillian study by Lacerda et al. had 10 previously untrained men train the unilateral leg extension for 3-4 sets with a 55-60% one-rep max load, 2-3 times per week for 14 weeks.
With one leg, subjects performed their repetitions to failure every set.
With their other leg, they ended up leaving 2 repetitions in reserve on their first set in a session, 1 repetition in reserve on their second set, and 0 repetitions in reserve on their final set.
By the end of the study, increases in rectus femoris and vastus lateralis cross-sectional area were similar between both legs, suggesting leaving 0 to 2 repetitions in reserve is as effective as training to failure for growth).
The second study by Nobrega et al., conducted in Brazil in 2017, allocated each leg of 32 previously untrained men into one of four conditions: a high load failure, high load non-failure, low load failure, and low load non-failure condition.
All four conditions involved training the unilateral leg extension for 3 sets per session, twice per week for 12 weeks.
The high load failure condition involved performing repetitions to failure every set with an 80% one-rep max load.
The high load non-failure condition, also with an 80% one-rep max load, involved performing repetitions to just before they felt they were going to fail. Based on the data provided, it’s likely this condition involved leaving 0 to 3 repetitions in reserve per set.
The low load failure condition involved performing repetitions to failure every set with a 30% one-rep max load.
The low load nonfailure condition, with a 30% one-rep max load, involved performing repetitions to just before they felt they were going to fail. Again, this condition likely ultimately left 0 to 3 repetitions in reserve per set.
By the end of the study, increases in vastus lateralis cross-sectional area were similar between all four conditions, indicating regardless of whether an 80% or 30% one-rep max load is used, leaving 0 to 3 repetitions in reserve is no less effective than training to failure for growth.
As a separate point, it’s worth noting this Nobrega study nicely demonstrates a point we’ve covered in other articles, that loads as light as 30% one-rep max can build muscle very effectively.
Moving on to the third paper, a 2015 Australian study by Sampson & Groeller had 28 previously untrained men perform a bicep curling-like exercise for 4 sets with an 85% one-rep max load, three times per week for 12 weeks.
One group of subjects performed all their sets with repetitions to failure with a 2-second lifting and 2-second lowering speed. This group ended up averaging 6 reps per set.
A second group performed only 4 repetitions per set, with a 2-second lowering and maximal lifting speed. Based on the numbers provided, this group would have left an average of 2 repetitions in reserve.
A third group also performed only 4 repetitions per set, but they used a maximal lowering and maximal lifting speed. This group would have also left an average of 2 repetitions in reserve.
Although not major, there are some slight limitations with this study.
As you likely noticed, the three different groups used different repetition speeds, which is potentially confounding. However, as described in our time under tension article, individual repetition speeds between 0.5-8 seconds appear similarly effective for muscle growth, so the use of different repetition speeds in this study is probably not a major concern.
Also, groups two and three, who both left around 2 repetitions in reserve, actually performed an additional one set of repetitions to failure each week. However, I doubt this would significantly impact the outcomes.
Anyway, by the end of the study, increases in elbow flexor cross-sectional area (consisting of the biceps and brachialis) were similar between all three groups, suggesting leaving 2 repetitions in reserve can be comparably effective to training to failure.
What About Research in Trained Individuals?
As a small summary thus far, these three studies overviewed indicate stopping 3 or fewer repetitions from failure can be equivalently effective for building muscle versus training to failure.
However, all three of these studies were conducted on previously untrained individuals. One could suggest these results should be expected in untrained individuals, as they would be highly responsive regardless of how close they get to failure with their repetitions on an exercise. They may suggest that different results may arise in trained individuals.
Fortunately, the fourth paper, a 2020 Brazillian study by Santanielo et al. was conducted on 14 men with an average of 5.1 years of training experience. They had an average body weight of 74.4kg and could leg press an average of 237.5kg before the study.
Subjects trained both the unilateral leg press and unilateral leg extension, twice per week for 10 weeks.
A great thing about this study was the researcher’s individualized set numbers. More precisely, the researchers noted how many sets per week the subjects performed for the quadriceps before the study, and then the subjects performed 20% more for the study’s duration.
For example, if a subject performed 15 weekly sets for the quadriceps before the study, they performed 18 weekly sets (a 20% increase) for the quadriceps in this study. These 18 weekly sets were equally divided between the leg press and leg extension.
With one leg, subjects performed every set with repetitions to failure.
With their other leg, they performed every set with repetitions to a point before failure. The data provided suggest they mostly ended up leaving 0 to 3 repetitions in reserve.
By the end of the study, increases in vastus lateralis cross-sectional area were similar between both legs, suggesting leaving 0 to 3 repetitions in reserve is still similarly effective to training to failure in trained individuals.
Considering the totality of the current literature, it appears in previously untrained and trained individuals, training to failure is not a must.
Leaving 0 to 3 repetitions in reserve appears equivalently capable of eliciting muscle hypertrophy.
However, what about leaving more than 3 repetitions in reserve? Could this still successfully optimize muscle hypertrophy?
Let us assess the evidence exploring this.
Training to Failure vs Leaving More Than 3 Repetitions in Reserve
I’m currently aware of 4 regular longitudinal studies that have technically examined how leaving more than 3 repetitions in reserve compares to training to failure for hypertrophy.
Martorelli et al. allocated 89 previously untrained women into one of three groups: a failure, non-failure equal set, or non-failure +1 set group.
All three groups trained the barbell biceps curl twice per week for 10 weeks.
The failure group performed the exercise with 3 sets of repetitions to failure with a 70% one-rep max load each session.
The non-failure equal set group performed 3 sets of 7 repetitions with a 70% one-rep max load each session.
With a 70% one-rep max load on the barbell biceps curl, the data provided in the study suggests subjects could perform a maximum of 12-13 repetitions.
Meaning this non-failure equal set group would have been leaving roughly 5 to 6 repetitions in reserve on their first set. However, their 2nd and 3rd sets, due to cumulative fatigue (the build-up of fatigue across sets), would have probably had them leave roughly 4-2 repetitions in reserve.
The non-failure +1 set group performed 4 sets of 7 repetitions with a 70% one-rep max load each session. This group would have also been training with roughly between 6 to 2 repetitions in reserve, the only difference being they performed an extra set per session.
All groups rested 2 minutes between their sets in a session. Moreover, one-rep max was retested on the 5th week to readjust the 70% one-rep max training loads used for all groups.
By the end of the study, increases in elbow flexor thickness seemed to be greatest for the failure group.
Therefore, this study suggests regardless of whether an individual performs the same number of sets, or performs one extra set, leaving more than 3 repetitions in reserve for the majority of your sets may be inferior for hypertrophy.
Do the other studies support this conclusion? Let’s find out.
The second study by Lasevicius et al. assigned 25 previously untrained men into either a high load or low load group.
All subjects performed the unilateral leg extension twice per week for 8 weeks.
The high load group used an 80% one-rep max load.
With one leg, which we’ll call the high load failure leg, they performed 3 sets of repetitions to failure every session. Ultimately, this leg performed 12 repetitions on average in each set.
With their other leg, which we’ll call the high load non-failure leg, they performed 60% of the average repetitions achieved each set with the high load failure leg. Moreover, they performed additional sets to equalize the overall volume load (the product of sets x reps x load) to the high load failure leg.
Ultimately, this leg performed an average of 5-6 sets of 7 repetitions, meaning they would have been leaving an average of 5 repetitions in reserve while performing 2-3 extra sets versus the high load failure leg.
The low load group used a 30% one-rep max load.
With one leg, which we’ll call the low load failure leg, they performed 3 sets of repetitions to failure every session. Ultimately, this leg performed 34 repetitions on average in each set.
With their other leg, which we’ll call the low load non-failure leg, they performed 60% of the average repetitions achieved each set with the low load failure leg, and they performed additional sets to equalize the overall volume load to the low load failure leg.
Ultimately, this leg performed an average of 5-6 sets of 20 repetitions, meaning they would have been leaving an average of 14 repetitions in reserve while performing 2-3 extra sets versus the low load failure leg.
All subjects rested 2 minutes between their sets in a session. Moreover, one-rep max was retested during the 4th week to readjust the 30% and 80% one-rep max training loads used for the respective groups.
By the end of the study, except for the low load non-failure leg, all the other legs experienced similar increases in quadriceps cross-sectional area.
As a result, given the low load-non failure leg experienced the least growth, this suggests leaving an average of 14 repetitions in reserve may be unable to optimize muscle hypertrophy, despite them performing 2-3 extra sets compared to the failure legs.
However, given the high load non-failure leg experienced similar growth to the high load and low load failure legs, this suggests leaving an average of 5 repetitions in reserve might actually be fine for muscle hypertrophy, so long as 2-3 extra sets are performed.
Stated another way, performing 5-6 sets while leaving roughly 5 repetitions in reserve may be as effective as performing 3 sets of repetitions to failure.
This is interesting, perhaps on a set equated basis, performing 3 sets of repetitions to failure would be superior to performing 3 sets of leaving 5 repetitions in reserve (as somewhat supported by the Martorelli study). However, performing extra sets while leaving 5 repetitions in reserve compensates for this.
Having said this, the third study by Karsten et al. questions this.
18 men with 2-5 years of training experience were assigned into a failure or non-failure group.
Both groups trained a range of exercises twice per week for 6 weeks.
The failure group aimed to perform every exercise each session with 4 sets of 10 repetitions to failure while using 2 minutes of rest between sets. When needed, the load was increased if they could do more than 10 repetitions on a set, or if they could not complete the 10 repetitions, 30 seconds of rest was given to enable them to complete this number.
The non-failure group aimed to perform every exercise each session with 8 sets of 5 repetitions that left roughly 5 repetitions in reserve, with 1 minute of rest between sets. When needed, this group adjusted the load each set to maintain roughly 5 repetitions in reserve.
By the end of the study, increases in thickness of the vastus medialis and elbow flexors tended to be greater for the failure group, whereas anterior deltoid thickness increases were similar between both groups.
As a result, this study conflicts with the Lasevicuis study, as it largely suggests despite performing extra sets (4 in this case), leaving roughly 5 repetitions in reserve was unable to optimize muscle hypertrophy.
Why the conflicting findings between these two studies?
For one, the Laseviciuis study used previously untrained men, while the Karsten study used men with 2-5 years of training experience. Maybe this implies leaving 5 repetitions in reserve, and performing extra sets, only works in untrained individuals and not those with greater training experience.
Secondly, as we noted, if the failure group in the Karsten study could not complete 10 repetitions on a set, they were given a 30-second break to enable them to then achieve this number. This is somewhat like rest-pause training. Frustratingly, the researchers did not state how frequently subjects did this, but if it was frequent, perhaps this could somehow explain the conflicting results between the two studies, as the Lasevicuis study merely had the failure groups perform their repetitions to failure (nothing more, nothing less).
Thirdly, in the Lasevicuis study, all subjects rested 2 minutes between their sets in a session. Opposingly, the Karsten study had the failure group rest 2 minutes between sets and the non-failure group rest only 1 minute between sets.
This potentially confounds the Karsten study, as it’s likely resting for 2 minutes between sets is better than resting for 1 minute between sets for hypertrophy, at least with compound exercises. Therefore, the largely inferior results for the non-failure group might not necessarily be because they left 5 repetitions in reserve, rather because they only rested 1 minute between sets.
Unfortunately, it’s difficult to say which of these three explanations, was it down to the differences in training experience, the potential rest-pause style in the Karsten study, or rest intervals that explains the conflicting findings of the Lasevicius and Karsten studies, more research is undoubtedly required.
The last regular longitudinal study I’m aware of that has examined how leaving more than 3 repetitions in reserve compares to training to failure comes from Carroll et al.
15 men with an average of 7.7 years of training experience were allocated into a relative intensity or rep max group.
Both groups trained three times per week for 10 weeks.
Quite an intricate periodized training plan was used for both groups. Feel free to check out the full study for more details, but I’ll simplify the main points.
Both groups performed the same exercises (mainly compound exercises, with some isolation work) for the same number of sets and roughly the same number of repetitions.
The rep max group used rep max loads for their sets. For example, they may have performed an exercise for 3 sets of 8-12 repetitions with a 10 repetition maximum load. The researchers stated this group reached failure on the last set of each exercise. Moreover, I think it’s probably likely their other sets would have left 3 or fewer repetitions in reserve, which we know is likely fine for muscle hypertrophy.
The relative intensity group used around 70-90% of the rep max loads for an exercise. For instance, if someone in the relative intensity group could perform 3 sets of 8-12 repetitions with a 100kg load on an exercise, and this 100kg load was their 10 rep-max load, using an 85% relative intensity resulted in them performing 3 sets of 8-12 repetitions with an 85kg load.
Based on some calculations, this relative intensity group would have been leaving around 2 to 6 repetitions in reserve on two of the training days per week, and between 4-14 repetitions in reserve on the 3rd day of each week.
By the end of the study, increases in vastus lateralis anatomical cross-sectional area favored the relative intensity group.
In other words, with all other training variables equal, largely leaving more than 3 repetitions in reserve (which is what the relative intensity group did) was more favorable for growth compared to training to or closer to failure (which is what the rep max group did).
Having said this, there is a potential confounding factor in this study.
Both groups performed additional sprint training twice per week. Training to failure is more fatiguing than leaving repetitions in reserve, and thus added sprint training may have impaired the rep max group’s recovery capacity, thereby possibly explaining why they experienced suboptimal growth versus the relative intensity group.
If true, perhaps this study indicates if an individual performs additional high-intensity training (like sprint training) or frequent cardiovascular training sessions of any type, leaving more than 3 repetitions in reserve might be better for managing fatigue and thus promoting muscle growth.
Although, I still feel more research is needed to verify this. I think it’d be useful to see simpler and cleaner designed studies comparing muscle growth between training with various proximities to failure (such as 3 repetitions in reserve, 5 repetitions in reserve, or failure training) while all subjects perform additional cardio or high-intensity training.
In summary, Martorelli et al. indicate with all other training variables equal, largely leaving more than 3 repetitions in reserve does not optimally build muscle in previously untrained women.
Lasevicius et al. indicate that provided an individual performs 2-3 more sets for an exercise than usual, leaving an average of 5 repetitions in reserve can build muscle optimally in previously untrained men.
However, a study by Karsten et al. suggests in men with 2-5 years of training experience, despite performing 4 extra sets, leaving 5 repetitions in reserve largely failed to optimize muscle growth. Although there were some confounders in this study.
Finally, Carroll et al. indicate when an individual performs additional sprint training sessions, training with more than 3 repetitions in reserve may be superior for promoting muscle growth compared to training to or closer to failure with all other training variables equal.
Overall, it’s clear to see this current evidence is limited, somewhat conflicting, and overall far from clear.
As a result, we unfortunately cannot conclude if leaving more than 3 repetitions in reserve can optimize muscle hypertrophy.
On the basis of this, as well as the detailed research, because it’s simply unclear if leaving more than 3 repetitions in reserve can still optimize muscle hypertrophy, I think it’s probably best to err on the side of caution and aim to leave 3 or fewer repetitions in reserve if your aim is to build muscle as best as you can.
Before moving on, if you’re curious about creating an effective training program for muscle hypertrophy, our high quality partner Alpha Progression can help. It can generate a highly effective program for you, track your workouts live with in-built progression recommendations, provide graphs displaying your long term progress, and it has a massive exercise database with more than 550 exercises.
Click HERE (the link opens in a new tab) to get a free 2 week trial of the apps features. If you like it and go beyond, the link also gives you 20% off a subscription!
We never promote trash at the House of Hypertrophy, so rest assured the app is high quality. The reviews speak to this, 4.8 starts (based on more than 7,000 reviews) on Google play, and 4.9 stars in Apple’s store (based on nearly 400 ratings).
Why Training to Failure May Not Be Necessary
Despite it being unclear if leaving more than 3 repetitions in reserve can optimize muscle hypertrophy, we’ve established leaving 0 to 3 repetitions in reserve appears to be no less effective than training to failure.
Why might this be?
I think it’s quite likely mechanical tension can explain this.
Mechanical tension is currently the best-categorized stimulus for muscle hypertrophy, and it’s quite likely leaving 0 to 3 repetitions in reserve elicits sufficient mechanical tension levels.
Let me explain.
Mechanical tension is equal to the force generated by muscle fibers.
Components within muscle fibers, called mechanosensors, detect forces produced by the muscle fiber and transduces these forces into a signaling cascade that results in muscle protein synthesis (which are the proteins that make our muscles bigger).
It’s worth exploring the concept of motor units to further understand overall mechanical tension and why training to failure may not be necessary.
A motor unit refers to a single motor neuron (that originates from the spine) and the multiple muscle fibers it innervates. When the single motor neuron sends an electrical signal through to the muscle fibers it innervates, those muscle fibers produce force.
Within a whole muscle, a variety of motor units with different characteristics exist.
In general, these characteristics likely lie across a spectrum, with low threshold motor units on one end, and high threshold motor units on the other end.
Typically, a low-threshold motor unit consists of a small motor neuron that innervates slow-twitch muscle fibers that produce low amounts of force but are highly fatigue resistant. Due to this, low-threshold motor units are described as highly fatigue resistant but low force-producing.
A high-threshold motor unit generally consists of a large motor neuron that innervates fast-twitch muscle fibers that produce high amounts of force but are highly fatigable. Therefore, high-threshold motor units are described as high-force producing but very fatigable.
For maximizing overall mechanical tension (and thus the stimulus for overall muscle growth), we’d want to recruit as many motor units as possible during an exercise (from the low-threshold motor units all the way up to the high-threshold motor units).
Furthermore, we’d likely want the individual motor units to produce at least decent amounts of their respective force capacities for a fair duration. The precise details of this will differ depending on the motor unit in question. Remember, low-threshold motor units have overall low-force producing capacities but can sustain these forces for a long time. Conversely, high-threshold motor units have overall high-force producing capacities but cannot sustain these forces for very long.
Here’s the main points:
Firstly, It’s not clear when you reach maximal motor unit recruitment, and the motor unit literature indicates this can differ between muscles.
However, the literature indicates this is likely achieved a fair while before failure.
For instance, one study suggests maximal motor unit recruitment for the shoulder and trap muscles may be obtained 3-5 repetitions before failure.
Secondly, by the time you’ve reached failure, it’s quite likely many motor units (particularly the high-threshold motor units) are past producing their highest forces and are instead just fatiguing.
Remember, it’s tension (analogous to force) that’s categorized as a prime stimulus for muscle hypertrophy. The idea that more motor unit fatigue is beneficial for muscle hypertrophy is not supported by the current literature.
A motor unit modeling study by Potvin and Fuglevand demonstrates our points nicely.
Their model included 7 different motor units, ordered in terms of thresholds.
Motor unit 1 was the lowest threshold motor unit, while motor unit 120 was the highest threshold motor unit.
The graph below depicts the contribution of each motor unit to a 50% maximal isometric contraction held until failure. The total contraction lasted 100 seconds.
Except for motor unit 120, all of the other motor units were recruited from the onset of contraction.
Motor units 1, 20, 40, and 60 produced steady low forces throughout the full contraction duration. This makes sense, these would be considered the low-threshold motor units (which produce low forces but are highly fatigue resistant).
Motor unit 80, a higher-threshold motor unit, produced its most force for the set at the start of the contraction, where it thereafter gradually fatigued.
Motor unit 100, a higher-threshold motor unit, continued to increase its force contribution up until the 65-second mark, and thereafter began to fatigue.
Motor unit 120, the highest-threshold motor unit, was recruited at the 70-ish second mark, and steeply increased its force contribution up until the 95-ish second mark, where it then steeply fatigued.
This data demonstrates motor unit recruitment, as well as the maintenance of at least decent amounts of respective force production from the individual motor units, is achieved before failure. At the point of failure, no additional motor units nor enhancements in individual motor units force occurred.
This probably explains why leaving 0 to 3 repetitions in reserve is sufficient for muscle growth.
I should restate this data was from modeling isometric contractions.
Recruitment patterns and strategies likely differ between isometric and the isotonic contractions that occur during your typical exercises.
However, I’m unaware of any reason to believe that the main point (that failure offers no additional motor unit recruitment or enhancements in individual motor units force) would not also be the same with isotonic contractions.
How to Gauge Repetitions in Reserve
In the real world, away from the laboratory in which researchers can verify proximity to failure, an individual may not actually be able to accurately estimate how close to failure they are.
For example, an individual may think they’ve left 3 repetitions in reserve, but in actuality, they may have left 5 or even more repetitions in reserve.
This is a very solid point.
How common is underestimating proximity to failure?
Moreover, how could an individual ensure greater accuracy in estimating how close to failure they are?
A recent 2020 meta-analysis by Halperin et al. answers these questions.
They combined the results of 12 studies that examined how well individuals estimate their proximity to failure.
Before detailing the findings, it’s worth briefly understanding the design of the included studies.
The 12 studies had a mix of designs.
The studies ranged from having subjects perform repetitions to failure with a 50 to 90% one-rep max load. Most of the studies used compound exercises, with the back squat and bench press being commonly used.
Although, other exercises like leg presses, chest presses, trap bar deadlifts, as well as biceps and triceps isolation exercises were also used in a few studies.
Some studies had subjects estimate how many repetitions they think they could perform with a given load on an exercise before performing the set, others had subjects estimate at some point during the set. For instance, a few studies had subjects stop at the 10th repetition with a 70% one-rep max load to report how many more repetitions they think they could achieve before they went onto complete repetitions until failure.
Furthermore, a few studies alternatively had subjects report once they felt they reached a certain number of repetitions in reserve. For example, subjects reported once they felt they hit the point of 3 repetitions in reserve before returning and performing repetitions to failure.
Let’s move on to the results.
Combining the 12 studies results, on average, subjects actually only underestimated their proximity to failure by 1 repetition. Although, as evidenced by the graph, there was quite some variation. A fair few subjects underestimated their proximity to failure by 1 to 3 repetitions. However, some subjects underestimated their proximity to failure by as much as 10 to 13 repetitions.
Let’s now explore some meta-regressions performed that can aid an individual in improving their accuracy at estimating proximity to failure.
Accuracy increased when subjects gave their predictions closer to failure. This means that studies in which subjects reported how many repetitions they thought they could perform before a set produced less accurate estimates compared to studies in which subjects reported estimates at some point during a set. Also, during a set, making a prediction closer to failure is better than making a prediction father from failure.
These findings imply that if an individual plans to train with repetitions in reserve, it’s likely best to avoid pre-planning how many repetitions you think will achieve this before the set. Instead, you should simply perform your repetitions and end the set only once you truly feel you’ve reached the desired proximity to failure.
Accuracy was much better when subjects trained with heavier loads that resulted in fewer repetitions. Using loads that resulted in more than 12 repetitions performed generally resulted in greater underestimations. These findings make sense, with higher repetitions (and thus lighter loads), there’s a lot more room for error. Additionally, lighter loads and higher repetitions have been shown to produce greater perceived exertion and discomfort compared to using heavier loads and fewer repetitions. These factors can mask an individual’s ability to truly delineate their proximity to failure.
For muscle hypertrophy, we know it’s quite likely loads between 30% to 85% one-rep max are similarly effective at building muscle. Therefore, it’s probably sensible for individuals to train on the heavier side of this if they wish to use repetitions in reserve.
Accuracy slightly improved during predictions in later sets compared to earlier sets.
This could be for a couple of reasons. Firstly, remember that technically the subjects of the studies took all their sets to failure, they just estimated their proximities to failure at certain time points before achieving failure. As a consequence, those earlier sets to failure may have acutely made them more accustomed and aware of the feeling of failure, thereby enhancing their prediction accuracy in later sets.
The potential implication of this is that if you’re someone who prefers to mix in failure and leaving repetitions in reserve in a session, it might be better to perform your failure sets first.
For example, let’s say you perform 4 sets with a certain load on the triceps skull crusher in a session. Performing the first 2 sets to failure may set you up for being more accurate if you were to use repetitions in reserves for the last two sets.
Accuracy was slightly worse with lower body exercises versus upper body exercises. It’s useful to be aware of this and recognize you may have to push slightly harder than you naturally anticipate if you use repetitions in reserve with lower body exercises.
Finally, and quite unexpectedly, training status minimally impacted accuracy. That is, on average, having multiple years of resistance training experience seemingly did not significantly enhance an individual’s ability to predict how close to failure they are.
It’s not clear why this is.
Although I’m not certain, it might simply be because many of the trained individuals in these studies may not have had a fair degree of experience with training to failure.
It seems reasonable to think if the researchers divided the trained individuals into those with minimal to no experience of training to failure and those with a fair degree of failure training experience, an effect might be seen such that those with failure training experience would be more accurate than those with no failure training experience.
If true, which I think it probably would be as it makes logical sense, gaining experience and familiarity with training to failure should enhance your accuracy when you use repetitions in reserve.
Perhaps you may perform a select few of your sets for a given exercise with repetitions to failure, or you could dedicate occasional blocks where most of your sets are performed to failure.
In summary, the current literature, although there is some variability, intriguing indicates individuals only underestimate their proximity to failure by one repetition on average.
Generally, it’s best not to pre-plan how many repetitions you should perform with a given load to reach a certain proximity to failure, rather simply perform repetitions with this load and stop only once you truly feel you’ve reached your desired proximity to failure.
Training with heavier loads and lower repetitions likely makes it easier to estimate your proximity to failure versus lighter loads and higher repetitions.
It’s worth being aware individuals tend to be more accurate with estimates in later sets, as well as with upper body exercises compared to lower body exercises.
Finally, it’s likely gaining a fair degree of experience with training to failure is beneficial for being accustomed and familiar with how failure feels to you, thereby plausibly enhancing your ability to estimate your closeness to failure when you do use repetitions in reserve.
Takeaway Points
- It seems leaving 0 to 3 repetitions in reserve produces similar muscle growth to training to failure, in both previously untrained and trained individuals.
- It’s not entirely clear if leaving more than 3 repetitions in reserve can still optimize muscle hypertrophy. The current evidence assessing this is conflicting, and not that well-designed.
- The reason why training to failure is probably not necessary is likely because it offers no additional benefit to mechanical tension levels (the best-categorized stimulus for muscle growth). Sufficient mechanical tension levels are likely obtained at the point of 3 repetitions in reserve.
- If an individual aims to leave repetitions in reserve, the below points may help them out:
- It’s best to avoid having a pre-planned number of repetitions in mind (only stop your repetitions with a given load once you truly feel you’ve reached your desired proximity to failure)
- Training with heavier loads (that result in fewer than 12 maximum repetitions) likely makes it easier to judge proximity to failure.
- Individuals tend to be more accurate judging proximity to failure with later sets on a given exercise, and with upper body exercises versus lower body exercises.
- Gaining a fair degree of experience with training to failure presumably enhances your ability to judge proximity to failure.
Remember to feel free to check out the Alpha Progression App if you’re interested. Also feel free to check out our free bench press e-book below.
Fantastic article! I have been wondering about the merit of “leaving a few or 2 in the tank” vs training to failure for a long time now. Various strength coaches and trainers have varying opinions on this, which has made it a confusing topic. This is the best analysis I have seen anywhere on it. I will definitely be checking out more of this site.